University of Southern California Dissertations and Theses

Trunk control during dynamic balance: effects of cognitive dual-task interference and a history of recurrent low back pain

(USC Thesis Other)

Trunk control during dynamic balance: effects of cognitive dual-task interference and a history of recurrent low back pain

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content Running head: TRUNK CONTROL DURING DYNAMIC BALANCE

TRUNK CONTROL DURING DYNAMIC BALANCE:
EFFECTS OF COGNITIVE DUAL-TASK INTERFERENCE
AND A HISTORY OF RECURRENT LOW BACK PAIN

By
K. Michael Rowley

A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(BIOKINESIOLOGY)

AUGUST 2018
TRUNK CONTROL DURING DYNAMIC BALANCE ii
DEDICATION

This work is dedicated to the giants on whose shoulders I stand.
Thank you to the human beings who dedicated themselves to free inquiry and the pursuit to
understand the beauty they saw in the world. From classic thinkers like Socrates and Galileo who
were executed or imprisoned for challenging the world-view of those in power, to contemporary
pioneers in the study of our own bodies and minds like Andreas Vesalius, Nikolai Bernstein,
A. V. Hill, David Winter, and Jacquelin Perry, I am indebted to their dedication and sacrifices.

“The true value of a man is not determined by his possession, supposed or real, of Truth, but
rather by his sincere exertion to get to the Truth. It is not possession of the Truth, but rather the
pursuit of Truth by which he extends his powers and in which his ever-growing perfectibility is
to be found. Possession makes one passive, indolent, and proud. If God were to hold all Truth
concealed in his right hand, and in his left only the steady and diligent drive for Truth, albeit with
the proviso that I would always and forever err in the process, and to offer me the choice, I
would with all humility take the left hand.” - Gotthold Lessing, Anti-Goeze (1778)

TRUNK CONTROL DURING DYNAMIC BALANCE iii

ACKNOWLEDGEMENTS
My utmost gratitude and admiration goes to my PhD Advisor, Dr. Kornelia Kulig. She
models what it looks like to not just do clinical science, but to be a clinical scientist. For her,
inquiry, experimentation, and clinical translation are not just a nine-to-five, but a lifestyle. Dr.
Kulig exemplifies what it means to be at the cutting edge between laboratory experimentation
and clinical application, leaping with ease from one to the other in a single lecture or discussion.
She values collaboration and is always open to seeking out diverse perspectives and experts to
help us in our study design and interpretation. She taught me that a stroll in the hills with
conversation or even just solitary thought can be as meaningful and enlightening as a day spent
tinkering in the lab. (Though those who have joined her and Jacob in the hills know that she
doesn’t stroll). She is a role model and mentor to countless doctoral, master’s, and physical
therapy students, and I have been honored to learn and grow under her wing.
My faculty committee is made up of both up-and-coming and experienced researchers in
human movement science. In my experience, Dr. Jim Gordon’s primary contribution not only to
my personal education and research but to the Division and to physical therapy as a whole is
simply, but impactfully – inspiration. I am unbelievably fortunate to be one of the handful of
people in the world to have taken his course – Classic Readings in Biokinesiology. In this course,
Dr. Gordon imparted upon us an appreciation and respect for the really innovative, meticulous,
and superbly-design human movement science that is so hard to come by today. These are the
days when researchers created unique tools and equipment designed to answer a carefully crafted
testable hypothesis. Dr. Gordon carried this passion for science through all stages of my
dissertation research from my study design to communicating my findings both at conferences
TRUNK CONTROL DURING DYNAMIC BALANCE iv

and in writing in clear and impactful ways. My most creative visuals and paragraphs are thanks
to his gently guiding suggestions and care.
Of my committee members, Dr. Carolee Winstein was the most intimidating. Her
reputation includes no-nonsense, strong opinions and a passion for meticulous study design and
interpretation. These things, of course, are all true. But through leaning on her expertise and
experience more and more as I designed and interpreted the cognitive dual-task interference
components of my dissertation work, I learned something else. Her method is to forge strong,
smart, and clinically influential scientists. One of the best meetings of my career was discussing
two associations between motor control measures and self-report measures highlighted in
Chapter V. I showed the figures, and Dr. Winstein explained them in way I completely disagreed
with. In my head, “Alright, should I argue? Or just say ‘yes, ma’am’ and go change my
Discussion section? Eh, it’s my last semester. Let’s go for it.” I told her I disagreed and began to
explain how I would interpret the data. Interrupting me, she said, “Okay, now what you should
say is, ‘If that were true, then…’” Aha! So, I shared additional data that I believed rebutted her
interpretation. She came back with a different explanation by pulling in a new theoretical
framework I had not considered, which appropriately reconciled all of our disagreements. So, of
course in the end, she was right and I re-wrote the Discussion section, but I am so glad I stood up
for what I thought. In addition to a much more interesting Chapter V, I learned crucial lessons in
how to move a scientific debate forward and structure my arguments. When we both stood our
ground was when I learned the most.
Dr. James Finley was one of the junior faculty on my committee. His fresh perspective on
motor control helped me put the concepts I was learning into a much broader and historical
context. I recall in one of our first meetings my “Aha! moment” when I realized these various
TRUNK CONTROL DURING DYNAMIC BALANCE v

concepts I was learning – internal models, intermittent control, equilibrium-point hypothesis, and
more – were each theories of how the brain plans and controls movement, each theory with its
own strengths and weaknesses. He helped me understand that these were not dictates on how
motor control works, but frameworks that help us to explain and then make predictions about
movement. In early phases of my dissertation work, Dr. Finley offered the most detailed help in
terms of coding and evaluating the robustness of my most important outcome measures.
Finally, my relationship with Dr. Joanne Armour Smith began as a colleague in our
shared office while she was in her last year in the program. Even before she officially stepped
into an advisor role, she was a mentor and role model in the lab. I learned many of the most
important research techniques from her including instrumenting participants with our motion
capture marker sets and surface and fine-wire electromyography. It has been an honor to build on
the stellar work she completed while here at USC. As a recent graduate, she has been the most
able to “commiserate” with me during the more stressful periods of this dissertation work. She
has a balanced perspective with a lot of experience presenting her work to different groups. Her
experience has been invaluable in helping me prepare for critiques, arguments, or questions
about some of my interpretations and conclusions.
Although I’ve worked most closely with my dissertation committee over the last few
years, there are countless faculty in the Division and at USC as a whole who I owe thanks to. In
the Division, Dr. Kathryn Havens, Dr. Lucinda Baker, Dr. Liz Poppert, and Dr. Robert Gregor
stand out as educators, guides, and mentors. These individuals contributed to the well-rounded
education I received at USC. Drs. Poppert and Havens model what it means to be a caring and
dedicated educator by always updating, improving, and sometimes obsessing over, the content
and techniques they use to convey difficult concepts to students. Dr. Gregor provided invaluable
TRUNK CONTROL DURING DYNAMIC BALANCE vi

contributions to my grant-writing training and experience, and Dr. Baker not only taught me the
technique of fine-wire electromyography but trained me in how to instruct others.
Thank you to all the students, past and present, of the Division and especially the
Musculoskeletal Biomechanics Research Lab. It would have been impossible to accomplish such
a challenging five years without caring and supportive colleagues, office-mates, and friends. I am
honored to have been able to learn and grow with stellar researchers and labmates like Jennifer
Bagwell, Eugene Chang, Kristamarie Pratt, Danielle Jarvis, Jennifer (Tzu-Chieh) Liao, Sharon
(Hsiang-Ling) Teng, Tulika Nandi, Matthew (Ming-Sheng) Chan, Andrea DuBois, Abbigail
Fietzer, Yo Shih, Steffi (Hai-Jung) Shih, Jonathan Lee, Jia Liu, Nicole Marcione, Jordan
Cannon, and Sara Almansouri. And, thank you to the staff of the Division, especially Tasha Hsu,
Janet Stevenson, Lydia Vazquez, Troy Lord, Veronica Perez, and Oshawa Smith. Our Division
supports us in ways other doctoral students around the country only wish!
Most importantly, I want to acknowledge my family. My parents were not thrilled when I
announced I was moving across the country for a graduate program, but they have been
supportive emotionally, financially, and more. Thank you to my dad, Kevin, for keeping me
connected to home and family. I moved at a time when his mother, Alice, was in failing health,
and he helped me stay in touch with her as her ability to answer the phone declined. I can always
count on his visits for a mental break and a fun time. Thank you to my mom, Lori, for sparking
and nurturing my curiosity and love of learning from a very early age. She was a single, working
mom for much of my childhood, and somehow between eight hours of work, making dinner,
taking care of the house, and visiting her parents as they aged and suffered from cancer and
stroke, she still found time to engage in creative educational activities with me. Loving,
supportive, and selfless parents is a privilege I didn’t earn, but wouldn’t have survived without.
TRUNK CONTROL DURING DYNAMIC BALANCE vii

My sister, Whitney, has been a sounding board and supportive friend for all of my adult
life. We didn’t become close until I moved out of our childhood home, and since then, she’s been
a best friend. She calls me out when I’m being impatient or “too much”, and sometimes she
seems like the only sane one in the family. In the last few years, Whitney has graduated from the
University of Pittsburgh with a degree in Athletic Training and began working as a trainer at
Point Park University. Our shared interest in human movement and health has made our
connection even stronger as we call each other for patient or research consultations at least
monthly and have fun conversations.
My fiancé, Preston Lopez, has been too important a part of the last four years of my life
to write in one paragraph. Not only has he supported and been patient with me as I worked
toward this degree, but he’s done something more important. He’s helped me not only stay
human but become more human throughout this process. Thank you for teaching me how to stay
connected – to myself and to you. Your heart is such a powerful force in the world that impacts
everyone fortunate enough to know and be loved by you. I really believe that impact already
resonates further and wider than any of the findings in this dissertation will. I feel your
commitment and it’s been a source of strength in these last few years.
Finally, I want to acknowledge organizations that helped fund this research including the
Division of Biokinesiology and Physical Therapy, the American Society of Biomechanics, and
the International Society of Biomechanics. These organizations have funding mechanisms to
support student research that this work would not have been possible without. And lastly, but
perhaps most relevant to this dissertation, I want to thank my research participants who gave up
their time to participate in this research study.
TRUNK CONTROL DURING DYNAMIC BALANCE viii

TABLE OF CONTENTS

DEDICATION............................................................................................................................... ii
ACKNOWLEDGEMENTS ........................................................................................................ iii
TABLE OF CONTENTS .......................................................................................................... viii
LIST OF TABLES ....................................................................................................................... xi
LIST OF FIGURES .................................................................................................................... xii
ABSTRACT ............................................................................................................................... xvii
CHAPTER I: OVERVIEW ......................................................................................................... 1
CHAPTER II: LITERATURE REVIEW .................................................................................. 5
INVESTIGATING THE TRUNK DURING POSTURAL CONTROL............................................................ 5
Discrete mechanical perturbations to posture ........................................................................ 6
Continuous mechanical perturbations to posture ................................................................. 13
Double- and single-limb standing balance............................................................................ 15
THE BALANCE-DEXTERITY TASK ............................................................................................... 17
TRUNK MUSCLES CONTRIBUTE TO STABILITY OF THE LUMBAR SPINE ......................................... 22
COGNITIVE CONTRIBUTIONS TO POSTURE ................................................................................... 27
Theoretical framework of attentional resources ................................................................... 28
Dual-task interference methodology ..................................................................................... 30
Dual-tasking and postural control ........................................................................................ 32
REINVESTMENT THEORY ............................................................................................................ 35
Theoretical framework .......................................................................................................... 35
Mechanism of reinvestment interference ............................................................................... 38
Measuring a tendency to reinvest .......................................................................................... 41
Factors that influence or invoke attention reinvestment in movement .................................. 44
Reinvestment and posture ...................................................................................................... 47
RECURRENT LOW BACK PAIN ...................................................................................................... 49
Epidemiology ......................................................................................................................... 49
Motor control dysfunction ..................................................................................................... 50
Psychosocial Factors............................................................................................................. 57
Effects of dual-task interference ............................................................................................ 59
SUMMARY .................................................................................................................................. 60
REFERENCES .............................................................................................................................. 62

TRUNK CONTROL DURING DYNAMIC BALANCE ix

CHAPTER III: CHARACTERIZING THE BALANCE-DEXTERITY TASK .................. 89
ABSTRACT .................................................................................................................................. 89
INTRODUCTION ........................................................................................................................... 90
METHODS ................................................................................................................................... 92
Participants and instrumentation .......................................................................................... 92
Procedures ............................................................................................................................. 92
Data analysis ......................................................................................................................... 94
RESULTS ..................................................................................................................................... 97
Task performance .................................................................................................................. 97
Trunk coordination .............................................................................................................. 100
DISCUSSION .............................................................................................................................. 104
REFERENCES ............................................................................................................................ 110
CHAPTER IV: TRUNK COUPLING IN PERSONS WITH RECURRENT LOW BACK
PAIN ........................................................................................................................................... 116
ABSTRACT ................................................................................................................................ 116
INTRODUCTION ......................................................................................................................... 117
METHODS ................................................................................................................................. 120
RESULTS ................................................................................................................................... 123
Participants ......................................................................................................................... 123
Task performance ................................................................................................................ 125
Trunk control ....................................................................................................................... 127
DISCUSSION .............................................................................................................................. 131
CONCLUSION ............................................................................................................................ 136
REFERENCES ............................................................................................................................ 137
CHAPTER V: INFLUENCE OF DUAL-TASK INTEREFERENCE ................................. 144
ABSTRACT ................................................................................................................................ 144
INTRODUCTION ......................................................................................................................... 145
METHODS ................................................................................................................................. 148
Participants and instrumentation ........................................................................................ 148
Procedures ........................................................................................................................... 149
Data analysis ....................................................................................................................... 151
RESULTS ................................................................................................................................... 153
Participants ......................................................................................................................... 153
Task prioritization hypothesis ............................................................................................. 154
Trunk control hypothesis ..................................................................................................... 157
Factors influencing trunk control ........................................................................................ 158
DISCUSSION .............................................................................................................................. 164
Confirmation of prioritization manipulation plus an unexpected facilitation effect ........... 165
Comparing task performance and prioritization manipulation between groups ................ 167
Trunk control differences between groups .......................................................................... 168
TRUNK CONTROL DURING DYNAMIC BALANCE x

Limitations and future directions ........................................................................................ 173
CONCLUSIONS .......................................................................................................................... 175
REFERENCES ............................................................................................................................ 176
CHAPTER VI: THE ROLE OF MOVEMENT-SPECIFIC REINVESTMENT ............... 182
ABSTRACT ................................................................................................................................ 182
INTRODUCTION ......................................................................................................................... 183
METHODS ................................................................................................................................. 186
Participants and instrumentation ........................................................................................ 186
Procedures ........................................................................................................................... 187
Data analysis ....................................................................................................................... 188
RESULTS ................................................................................................................................... 188
Participants ......................................................................................................................... 188
Task performance ................................................................................................................ 190
Trunk coordination .............................................................................................................. 193
DISCUSSION .............................................................................................................................. 195
Back-healthy control participants ....................................................................................... 195
Persons in remission from rLBP ......................................................................................... 197
Limitations and future directions ........................................................................................ 200
CONCLUSIONS .......................................................................................................................... 201
REFERENCES ............................................................................................................................ 202
CHAPTER VII: SUMMARY AND CONCLUSIONS .......................................................... 208
APPENDIX A: EXPLORATORY AIM ................................................................................ 217
PURPOSE ................................................................................................................................... 217
METHODS ................................................................................................................................. 217
RESULTS ................................................................................................................................... 217
DISCUSSION .............................................................................................................................. 222
APPENDIX B: METHODS .................................................................................................... 224
PARTICIPANT RECRUITMENT ..................................................................................................... 224
Inclusion and exclusion criteria .......................................................................................... 224
QUESTIONNAIRES AND PSYCHOMETRICS .................................................................................. 226
EXPERIMENTAL INSTRUMENTATION ......................................................................................... 229
PROCEDURES ............................................................................................................................ 231
DATA ANALYSIS ....................................................................................................................... 235
EXPLORATORY AIM – CASE SERIES FOLLOW-UP ....................................................................... 239
REFERENCES ............................................................................................................................ 240

TRUNK CONTROL DURING DYNAMIC BALANCE xi

LIST OF TABLES
Table IV.1. Participant demographics (mean ± standard deviation) .......................................... 125

Table V.1. Participant demographics (mean ± standard deviation) ............................................ 153

Table VI.1. Participant demographics (mean ± standard deviation) .......................................... 189

Table B.1. Participant demographics (mean ± standard deviation) ............................................ 226

TRUNK CONTROL DURING DYNAMIC BALANCE xii

LIST OF FIGURES
Figure I.1. The Balance-Dexterity Task. ........................................................................................ 2

Figure II.1. The Balance-Dexterity Task. ..................................................................................... 21

Figure II.2. Attentional resources model (adapted from Wickens, 2002). ................................... 29

Figure II.3. The Movement-Specific Reinvestment Scale (Masters and Maxwell, 2008). .......... 42

Figure III.1. The Balance-Dexterity Task with representative data showing balance outcome
measures including center of pressure (COP) measures (right) and dexterous vertical force
(vForce) control outcome measures including root-mean-squared error (RMSE), coefficient of
variation (CV) and median frequency (MDF) (left). .................................................................... 97

Figure III.2. (A) Submaximal, reproducible compression forces acquired from five practice trials
and presented as a target during stable block and Balance-Dexterity Task trials. Measures of
dexterous force control in the stable block and Balance-Dexterity Task conditions including root-
mean-squared-error (B), coefficient of variation (C), and median frequency (D). *p<0.05. ....... 98

Figure III.3. Center of pressure (COP) average resultant velocity in four conditions. ................. 99

Figure III.4. (A) Association between coefficient of variation (CV) of dexterous force control
and center of pressure (COP) average resultant velocity. (B) Association between participants’
self-reported assessment of task difficulty on a visual analog scale (VAS) and root-mean-
squared-error (RMSE) of dexterous force relative to the reproducible, submaximal compression
goal line. ...................................................................................................................................... 100

Figure III.5. Representative examples of high (left) and low (right) frontal plane trunk coupling
quantified with a coefficient of determination for the thorax and pelvis angle-angle plot (R
2
). 100

Figure III.6. Associations between trunk coupling quantified as the coefficient of determination
of a thorax-pelvis frontal plane angle-angle plot (R
2
) and (A-C) frontal plane segment and joint
excursions and (D) the percent of time spent in in-phase and in anti-phase coupling. (E) Trunk
coupling R
2
during single-limb stance and the Balance-Dexterity Task. ................................... 102

Figure III.7. (A) Muscle activation levels normalized to the stable block condition. MF:
multifidus, ES: erector spinae, IO: internal oblique, EO: external oblique, RA: rectus abdominis,
GMax: gluteus maximus, and GMed: gluteus medius. (B) Deep-to-superficial muscle activation
ratios. (C) Co-contraction ratios for deep muscles (MF and IO) and superficial muscles (ES and
EO). ............................................................................................................................................. 103
TRUNK CONTROL DURING DYNAMIC BALANCE xiii

Figure III.8. Associations between trunk coupling quantified as the coefficient of determination
of a thorax-pelvis frontal plane angle-angle plot (R
2
) and the ratio of multifidus to erector spinae
activation (MF:ES) normalized to maximum isometric voluntary contractions, shown with (solid
line) and without (dotted line) a potential outlier. ...................................................................... 104

Figure IV.1. Participant recruitment. rLBP = recurrent low back pain. ..................................... 124

Figure IV.2. Reproducible, submaximal compression force achieved during practice trials and set
as the goal for test trials of the Balance-Dexterity Task reported in Newtons (N) (A) and as a
percentage of body weight (%BW) (B). Self-report visual analog scale (VAS) measures of task
difficulty, participant confidence, and the amount of attention required to complete the task (C)
for persons with recurrent low back pain (rLBP) and back-healthy control participants (CNTRL).
..................................................................................................................................................... 126

Figure IV.3. Center of pressure (COP) average resultant velocity in double-limb stance, single-
limb stance, stable block, and Balance-Dexterity Task conditions. (A) Dexterous force control
measures including root-mean-squared-error (RMSE) (B) and coefficient of variation (CV) (C)
of the vertical force produced in the Balance-Dexterity Task for persons with recurrent low back
pain (rLBP) and back-healthy control participants (CNTRL). ................................................... 127

Figure IV.4. Coupling of thorax and pelvis frontal plane rotation (trunk coupling R
2
) (A) as well
as thorax, pelvis, and trunk frontal plane excursion (B) during the Balance-Dexterity Task in
persons with recurrent low back pain (rLBP) and back-healthy controls (CNTRL). *p<0.05 ... 128

Figure IV.5. Associations between trunk coupling (R
2
) and balance and dexterous force control
measures for persons with recurrent low back pain (rLBP) and back-healthy control participants
(CNTRL). Balance was quantified with center of pressure (COP) average resultant velocity (A),
and dexterous force control measures include root-mean-squared-error (RMSE) (B) and
coefficient of variation (CV) (C). ............................................................................................... 129

Figure IV.6. Associations between trunk coupling (R
2
) and psychometric scores including pain
catastrophizing scale (A), Tampa scale for kinesiophobia (B), and fear-avoidance beliefs
questionnaire (C) for persons with recurrent low back pain (rLBP) and back-healthy control
participants (CNTRL). ................................................................................................................ 129

Figure IV.7. (A) Muscle activation data during the Balance-Dexterity Task for multifidus (MF),
erector spinae (ES), internal oblique (IO), external oblique (EO), rectus abdominis (RA), gluteus
maximus (GMax), and gluteus medius (GMed). (B) Muscle activation ratios describing deep-to-
superficial paraspinal activity (MF:ES), deep-to-superficial abdominal activity (IO:EO), (C) deep
TRUNK CONTROL DURING DYNAMIC BALANCE xiv

muscle co-contraction (MF and IO), and superficial muscle co-contraction (ES and EO) in
persons with recurrent low back pain (rLBP) and back-healthy control participants (CNTRL).
EMG were normalized to the stable block condition. ................................................................. 130

Figure IV.8. Associations between trunk coupling (R
2
) and deep-to-superficial paraspinal muscle
activation ratio multifidus (MF) : erector spinae (ES) (A) and abdominal ratio internal oblique
(IO) : external oblique (EO) (B) in persons with recurrent low back pain (rLBP) and back-
healthy control participants (CNTRL). EMG were normalized to the stable block condition
before the ratios were calculated................................................................................................. 131

Figure V.1. Task performance measures for each group (back-healthy controls [CNTRL]: A, C,
E; persons with recurrent low back pain [rLBP]: B, D, F) in each instruction condition. To
quantify dexterous force control, root-mean-squared-error (RMSE) (A, B) is presented. To
quantify balance control, center of mass (COM) average resultant velocity is shown (C, D). To
quantify cognitive task performance, error variability is presented (E, F). *p<0.05 .................. 156

Figure V.2. Frontal plane trunk coupling, reported as a coefficient of determination (R
2
) of a
thorax-pelvis angle-angle plot for back-healthy controls (A) and persons with rLBP (B). *p<0.05
..................................................................................................................................................... 158

Figure V.3. Associations between frontal plane trunk coupling (R
2
) and deep-to-superficial trunk
muscle activation ratios. The ratio for paraspinal coordination is multifidus-to-erector spinae
(MF:ES) (top row) and the ratio for abdominal coordination is internal-to-external-oblique
(IO:EO) (bottom row). Conditions are the Balance-Dexterity single-task condition (STBal) (left
column), the dual-task condition with priority assigned to the cognitive task (DTCog) (middle
column) and the dual-task condition with priority assigned to the balance task (DTBal) (right
column). ...................................................................................................................................... 160

Figure V.4. Associations between the change in average muscle activation amplitudes for lumbar
multifidus (MF) (A), lumbar erector spinae (ES) (C), internal oblique (IO) (B), and external
oblique (EO) (D) and the change in frontal plane trunk coupling (R
2
) from single- to dual-task
cognitive priority conditions, where a positive ∆R
2
indicates a participant exhibited more
coupled trunk motion during the dual-task condition. Correlation shown in (C) is after removing
one participant with greatest decrease in ES activation amplitude and two participants with
greatest increases in trunk coupling. ........................................................................................... 161

Figure V.5. Associations between trunk coupling (R
2
) and self-report measure of cognitive task
difficulty for the Balance-Dexterity single-task condition (A), the dual-task condition with
priority assigned to the cognitive task (B) and the dual-task condition with priority assigned to
the balance task (C). .................................................................................................................... 163
TRUNK CONTROL DURING DYNAMIC BALANCE xv

Figure V.6. Associations between the change in frontal plane trunk coupling from single- to
dual-task cognitive priority conditions and self-report cognitive task difficulty (A) and recalled
pain during a symptomatic episode reported on a visual analog scale (VAS) (B). Correlation
statistics are shown for the full group as well as without one participant in (A) and two in (B)
that seem to follow a different pattern. ....................................................................................... 163

Figure V.7. Summary figure reporting effect sizes (ES) for statistically significant comparisons
of means and coefficients of determination (R
2
) for statistically significant associations. Arrows
oriented down indicate that a decrease in the given measure from single- to dual-task conditions
was associated with greater increases in trunk coupling for the group in remission from recurrent
low back pain (rLBP). ................................................................................................................. 165

Figure VI.1. Associations between center of mass (COM) average resultant velocity and
movement-specific attentional reinvestment (A-C) and fear-avoidance beliefs questionnaire (D-
F) in the single-task Balance-Dexterity Task (A,D), the dual-task condition with the priority on
the cognitive task (B,E), and the dual-task condition with the priority on balance (C,F). ......... 191

Figure VI.2. Associations between dexterous force control coefficient of variation (CV) and
movement-specific attentional reinvestment (A-C) and fear-avoidance beliefs questionnaire (D-
F) in the single-task Balance-Dexterity Task (A,D), the dual-task condition with the priority on
the cognitive task (B,E), and the dual-task condition with the priority on balance (C,F). ......... 192

Figure VI.3. Associations between cognitive task error variability and movement-specific
attentional reinvestment (A-C) and fear-avoidance beliefs questionnaire (D-F) in the cognitive
single-task condition (A,D), the dual-task condition with the priority on the cognitive task (B,E),
and the dual-task condition with the priority on balance (C,F). ................................................. 193

Figure VI.4. Associations between movement-specific attention reinvestment and trunk coupling
in single-task Balance Dexterity Task (A), dual-task cognitive priority (B), and balance priority
(C) conditions.............................................................................................................................. 194

Figure VI.5. Association between movement-specific attention reinvestment and the change in
trunk coupling from single- to dual-task cognitive priority conditions. ..................................... 194

Figure VII.1. Summary figure reporting effect sizes (ES) for statistically significant comparisons
of means and coefficients of determination (R
2
) for statistically significant associations. Arrows
oriented downward indicate that a decrease in the given measure was associated with greater
increases in trunk coupling for the group in remission from recurrent low back pain (rLBP). .. 208
TRUNK CONTROL DURING DYNAMIC BALANCE xvi

Figure A.0.1. Pain reported by participants from 0-10 every Monday morning at 9:00am via text
message from the week after initial testing until they were brought in for re-testing during a
painful episode. ........................................................................................................................... 219

Figure A.0.2. Trunk coupling (R
2
) (left) and thorax, pelvis, and trunk excursion measures (right)
for both groups. The three individuals who were tested during symptom remission (“No Pain”)
and during a painful episode (“Pain”) are shown with connecting lines. ................................... 220

Figure A.0.3. The relationship between trunk coupling (R
2
) and multifidus (MF) – to – erector
spine (ES) ratio (left) the relationship between trunk coupling (R
2
) and internal oblique (IO) – to
– external oblique (EO) ratio (right) for both groups. The three individuals who were tested
during symptom remission and during a painful episode (“Pain”) are shown with connecting
lines. Note that participant rLBP22 in blue did not undergo fine-wire EMG instrumentation and
so is shown on the y-axis. ........................................................................................................... 221

Figure A.0.4. Trunk coupling (R
2
) in two conditions – Balance-Dexterity Task (single-task)
condition and dual-task condition – for the back-healthy control group (left) and the group with
rLBP (right). The three individuals who were tested during symptom remission (“No Pain”) and
during a painful episode (“Pain”) are shown with connecting lines. .......................................... 222

Figure B.0.1. Participant recruitment. rLBP = recurrent low back pain. .................................... 225

Figure B.0.2. The Movement-Specific Reinvestment Scale (Masters and Maxwell, 2008). ..... 228

Figure B.0.3. Fine-wire electromyography insertion of the lumbar mulitifidus (left) and the
internal oblique (right). ............................................................................................................... 231

Figure B.0.4. The Balance-Dexterity Task with representative data showing balance outcome
measures including center of pressure (COP) measures (right) and dexterous vertical force
(vForce) control outcome measures including root-mean-squared error (RMSE), coefficient of
variation (CV) and median frequency (MDF) (left). .................................................................. 238

Figure B.0.5. Representative examples of high (left) and low (right) frontal plane trunk coupling
quantified with a coefficient of determination for the thorax and pelvis angle-angle plot (R
2
). 238
TRUNK CONTROL DURING DYNAMIC BALANCE xvii

ABSTRACT
Postural control is crucial for successful human movement. The consequences of a failure
in the postural control system range from an innocuous stumble to life-threatening falls and
detrimental conditions of chronic pain. The human movement system protects against such
failures by using multiple, diverse, and redundant interacting systems. By probing individual
systems through mechanical and cognitive perturbations to posture and balance we can learn
about roles and functions of these contributing components. These experimental approaches have
been impactful in the study and rehabilitation of persons with low back pain, a condition which
represents a tremendous cost to society in terms of quality of life, healthcare expenditure, and
work time loss. In persons suffering from recurrent episodes of low back pain (rLBP), specific
abnormal movement and muscle activation patterns have been identified that persist even during
symptom remission and are thought to contribute to the recurrence of pain. In order to translate
laboratory findings to rehabilitation of these patients, novel continuous, ecological balance tasks
must be developed and used to study typical balance control and how control may be
dysfunctional in persons suffering from rLBP. Chapter II provided a literature-informed
justification and rationale for the hypotheses tested in these dissertation studies and important
aspects of study designed including development of the Balance-Dexterity Task and the
cognitive dual-task paradigm. In addition, a justification for studying a clinical population in
remission from recurrent episodes of LBP as a model dysfunctional postural control system was
presented.
The purpose of Chapter III was to characterize the Balance-Dexterity Task as a novel
dynamic balance task investigate trunk control. The task combined aspects of single-limb
balance and the lower-extremity dexterity test by asking participants to stand on one limb while
TRUNK CONTROL DURING DYNAMIC BALANCE xviii

compressing an unstable spring with the contralateral limb to an individualized target force.
Nineteen back-healthy control participants completed the study, and performance measures for
the demands of each limb – balance and dexterous force control – as well as kinematic and
electromyographic measures of trunk control were collected. Given five practice trials,
participants achieved spring compression forces ranging from 100-139N (mean 121.2 ± 12.3N),
representing 14.4-23.0% of body weight (mean 18.7 ± 2.4%), which were then presented as
target forces during test trials. The tsask invoked larger center of pressure (COP) velocities than
double- and single-limb stance and more variable dexterous force control compared to a stable
step condition. Variability in dexterous force control was associated with variability in balance
control, indicated by a positive correlation between dexterous force coefficient of variability and
stance limb COP velocity (R=0.598, p=0.007). This finding in a concurrent bipedal task aligns
with similar work in bimanual control tasks where variability between limbs is also associated.
Trunk coupling, quantified as the coefficient of determination (R
2
) of a frontal plane thorax and
pelvis angle-angle plot, varied independently of any measure of balance or dexterous force
control. Trunk coupling R
2
was correlated with the percent of time in in-phase coupling and
oppositely correlated with the percent of time in anti-phase coupling, justifying use of this
simpler measure to capture trunk coordination. Standard segment and joint excursions, however,
seemed to quantify different aspects of trunk movement. Muscle activation levels normalized to
a stable block reference condition were greater in the Balance-Dexterity Task compared to the
stable block condition, but no muscle activations alone or in ratios were associated with trunk
coupling R
2
. The paucity of associations between individual muscle activations and trunk
coupling indicates that these non-disabled persons have redundant motor control processes
available to control balance, dexterous force control, and trunk coupling. When signals were
TRUNK CONTROL DURING DYNAMIC BALANCE xix

normalized to MVICs, however, a greater multifidus-to-erector spinae ratio was weakly
associated with higher trunk coupling R
2
(R=0.517, p=0.028) suggesting greater relative deep
paraspinal activation drove greater trunk coupling. In conclusion, the Balance-Dexterity Task is a
continuous, dynamic balance task where bipedal coordination, captured as measures of balance
and dexterous force control between which variability was associated, and trunk coupling can be
observed and studied.
The purpose of Chapter IV was to compare Balance-Dexterity Task performance and
trunk control during the task between persons with and without rLBP. Motor control dysfunction
persisting during symptom remission in persons with rLBP may contribute to the recurrence of
pain. No differences in task performance were expected between those with and without rLBP,
but it was hypothesized persons with rLBP would exhibit greater trunk coupling in line with a
trunk stiffening strategy. Persons with and without rLBP (n=19 per group) completed the
Balance-Dexterity Task. Persons with rLBP must have had at least two episodes of pain per year
but been in symptom remission during the time of testing. Task performance outcome measures
included COP velocity under the stance limb and vertical force variability under the spring.
Trunk coupling was quantified with the coefficient of determination (R
2
) of an angle-angle plot
of thorax-pelvis frontal plane motion. Fine-wire and surface electromyography captured
activations of paraspinals and abdominals. All participants were able to successfully complete
the Balance-Dexterity Task with no differences between groups in reproducible, submaximal
compression force goal, COP velocity, or dexterous force control measures. In both groups,
frontal plane trunk coupling varied independently of task performance measures. The group in
remission from rLBP exhibited reduced trunk coupling, or more dissociated thorax and pelvis
motion, compared to back-healthy controls (p=0.024). Trunk coupling in this group was
TRUNK CONTROL DURING DYNAMIC BALANCE xx

associated moderately with lumbar multifidus-to-erector spinae activation ratio (R=0.618,
p=0.006) and weakly with internal-to-external oblique ratio (R=0.476, p=0.046). Here, greater
deep trunk muscle activation relative to more superficial muscles resulted in higher trunk
coupling. The finding of more dissociated thorax and pelvis motion in those with rLBP was
counter to the hypothesis of increased coupling through adoption of a stiffening strategy. This
unsupported hypothesis was built on investigations of discrete perturbations to posture where
greater trunk co-contraction and stiffness has been observed in persons with rLBP. These
perturbations, however, involved delivering external perturbing forces either to a support surface
or directly to the trunk, which may invoke a stiffening strategy related to fear of movement or
pain. The Balance-Dexterity is a submaximal, internally-driven unstable balance task during
which more dissociated trunk motion was observed in persons in remission from rLBP. Findings
underscore the task-dependent nature of trunk control research and assessment in persons with
rLBP.
The purpose of Chapter V was to investigate effects of cognitive dual-task interference
and task prioritization instructions on task performance and trunk control during dynamic
balance in persons with and without rLBP. First, the ability to modulate task performance in
accord with prioritization instructions was tested. Second, it was hypothesized trunk control
strategies in persons with rLBP would rely more on cognitive resources, and therefore trunk
kinematics would be altered in this group under dual-task interference compared to back-healthy
control participants. Finally, individual factors explaining changes in trunk kinematics were
explored. Persons with and without rLBP (n=19 per group) completed the Balance-Dexterity
Task with and without a cognitive task utilizing verbal working memory. No differences between
groups were identified for task performance measures, indicating the groups did not modulate
TRUNK CONTROL DURING DYNAMIC BALANCE xxi

performance differently. Dexterous force control error increased in both dual-task interference
conditions compared to the single-task condition, but to a larger degree when the priority was
placed on the cognitive task. Center of mass (COM) velocity, a measure of balance control, also
changed dependent on priority instruction. When the priority was assigned to the cognitive task,
COM velocity was greater compared to both single-task and dual-task balance priority
conditions. Cognitive task performance during the dual-task conditions was also modulated by
reducing variability of errors when the cognitive task was the priority compared to when balance
was prioritized. These findings confirmed the methodological assumptions behind the attentional
interference and the prioritization manipulation. Trunk coupling was lower in the rLBP
compared to back-healthy controls in the single-task condition (p=0.024) and increased
significantly in the dual-task condition (p=0.002). The amount of increase was associated with
reductions in erector spinae activation (R=-0.574, p=0.020), lower self-reported cognitive task
difficulty (R=-0.497, p=0.036), and lower recalled pain (R=-0.642, p=0.005). Kinesiologically, a
reduction in superficial paraspinal activity from single- to dual-task conditions driving the
increase in trunk coupling up to the level of back-healthy controls suggests this increased trunk
coupling is beneficial. Those with rLBP who reported the cognitive task was easier and who
recalled less pain during a typical episode reaped this benefit. Two potential frameworks explain
these findings. Using a movement-specific attentional reinvestment framework, low attentional
loads interfered with memory-of-pain-related conscious processing of posture and resulted in
decreased erector spinae activation and improved trunk coupling, while higher attentional loads
potentially induced their own psychological stress. Using an action-specific perception
framework, cause and effect are reversed such that persons who increased trunk coupling
perceived the cognitive task as easier and their episodes as less painful.
TRUNK CONTROL DURING DYNAMIC BALANCE xxii

The purpose of Chapter VI was to investigate the role of movement-specific attention
reinvestment in task performance and trunk coordination during posture-cognition dual-tasking
in persons with and without rLBP. It was hypothesized persons who invest more attention in
their movement would exhibit greater trunk coupling within the framework of “locking down”
degrees of freedom influencing balance control, and cognitive dual-tasking would interrupt this
relationship. Persons with and without rLBP (n=19 per group) completed the Balance-Dexterity
Task with and without a cognitive task utilizing verbal working memory. Task performance
measures included COM velocity and vertical force variability under the spring. Trunk coupling
was quantified with the coefficient of determination (R
2
) of an angle-angle plot of thorax-pelvis
frontal plane motion. Psychometrics collected included the movement-specific reinvestment
scale (MSRS), the fear-avoidance beliefs questionnaire (FABQ), the pain catastrophizing scale,
and the Tampa scale for kinesiophobia. In the control group, persons who invested more
attention in their movements exhibited greater trunk coupling (R=0.647, p=0.003) and
experienced greater reductions in trunk coupling under dual-task interference (R=-0.537,
p=0.018). No associations between social-cognitive factors and trunk coupling were observed in
persons with rLBP. Those who scored higher on MSRS, however, had lower dexterous force
variability (R=0.532, p=0.019), and those who scored higher on FABQ had lower COM velocity
(R=-0.534, p=0.022). These relationships also disappeared under dual-task interference. Overall,
persons with rLBP modulated variables more directly related to task performance and balance
based on social-cognitive factors while control participants modulated trunk control based on
these factors. And, for both groups, dual-task interference interfered with these relationships.
More research needs to be done in order to understand links between these findings and the
recurrence of LBP.
TRUNK CONTROL DURING DYNAMIC BALANCE xxiii

Participants in these dissertation studies made up a convenience sample of persons with
and without rLBP recruited from student groups, classes, flyers, and university-affiliated
physical therapy clinics, and the clinical population was generally young (age 23.5 ± 2.8yrs),
minimally disabled (ODI 16.0 ± 18.7%), and all in pain remission at the time of testing (0.4 ± 0.4
out of 10 on VAS). Some associations in Chapter V were only statistically significant after
removing one, two, or three of the same subset of three participants who did not follow the
pattern. Two of these participants exhibited the greatest increases in trunk coupling from single-
to dual-task conditions and two of the highest reported recalled pain levels. We did not have a
clinical reason to exclude these participants, but the heterogeneity and presence of subgroups in a
LBP population are well-known, and applying findings from laboratory research on persons with
rLBP in the clinic should always be done with caution and with subject-specificity in mind. This
is why we decided to present the findings and acknowledge the limitation that there may be other
strategies at work in subgroups of persons with rLBP. Larger studies may allow for subgroup
analyses to help explain what different or additional mechanisms are at work in these three
persons with rLBP. Future work should test additional hypotheses raised in the current
dissertation studies. Importantly, explanations for the lower trunk coupling in the rLBP group
during the Balance-Dexterity Task hinge on task-specificity. This could be tested by taking the
same set of participants through a set of comparable balance tasks with different key features.
Findings from the dual-task interference study raised questions about how cognitive task
difficulty and pain recall influenced trunk coupling and posture-cognition dual-tasking. More
objective measures of these factors – using a math ability test, pain threshold assessments, etc –
may help to disambiguate objective measures and perceived measures. This would provide
evidence to clarify the role of action-specific perception in the present findings. As we learn
TRUNK CONTROL DURING DYNAMIC BALANCE xxiv

more about the role of task-specificity and cognitive processing in persons with rLBP,
interventions can be developed that will be more patient-specific and effective in reducing the
recurrence of pain.
TRUNK CONTROL DURING DYNAMIC BALANCE 1

CHAPTER I
OVERVIEW
These dissertation studies contribute to the body of knowledge about how the human
movement system maintains upright standing posture during challenging and dynamic balance
conditions. Specifically, findings help to elucidate the role of the trunk in postural control by
examining trunk coordination and muscle activation patterns during the Balance-Dexterity Task.
The influence of cognition on postural control and trunk coordination was investigated by
utilizing a cognitive dual-task interference paradigm and testing associations with specifically-
chosen social-cognitive factors. Persons in symptom remission from recurrent low back pain
(rLBP) were studied as a model system in which postural control dysfunction may be observed
and where findings from these studies are most applicable to rehabilitation.
Postural control is crucial for successful human movement. The consequences of a failure
in the postural control system range from an innocuous stumble to life-threatening falls and
detrimental conditions of chronic pain. The human movement system protects against such
failures by using multiple, diverse, and redundant interacting systems. By probing individual
systems during perturbations to posture and balance we can learn about roles and functions of
these contributing components. Using a novel continuous dynamic balance task – the Balance-
Dexterity Task (Figure I.1) – trunk coordination and muscle activation patterns were investigated
in back-healthy individuals as well as persons in symptom remission from rLBP. The role of
cognitive contributions to postural control was investigated using a dual-task interference
research paradigm to perturb cognitive resources. Findings were interpreted in the context of
movement-specific reinvestment theory. It is unknown how non-disabled persons vary in their
tendency to invest attention in their movements and how this tendency affects dual-task
TRUNK CONTROL DURING DYNAMIC BALANCE 2

performance and movement strategies during dynamic balance. Therefore, the primary purpose
of this study was to investigate the role of cognitive contributions to postural control, and
specifically trunk control, during dynamic balance.

Figure I.1. The Balance-Dexterity Task.

A history of falls or injury may induce an increase in attention devoted to movement.
Therefore, persons with rLBP were studied alongside healthy controls as a model condition in
which to investigate how a history of pain affects these phenomena. Specific abnormal
movement and muscle activation patterns that persist during symptom remission in this patient
population are thought to contribute to the recurrence of pain, but it is unknown how these relate
to cognitive processing or an individual’s tendency to invest attention in processing movement.
Findings offer insight into cognitive systems contributing to postural control in this population
and suggest activities with high risk of postural instability such as while attention is divided.
Conclusions also suggest activities beneficial for rehabilitation such as using mild cognitive tasks
to distract from fear-avoidance or kinesiophobia tendencies. Therefore, a secondary purpose was
to compare the role of cognitive contributions to postural control, and specifically trunk control,
between back-healthy individuals and persons in remission from rLBP.
TRUNK CONTROL DURING DYNAMIC BALANCE 3

I hypothesized there would be an association between a person’s baseline tendency to
invest attention in movement and the effect of dual-task interference on trunk coordination
during the Balance-Dexterity Task. Further, I hypothesized participants with a history of rLBP
would adopt highly coupled thorax and pelvic motion utilizing superficial trunk musculature, and
this control strategy would rely on more cognitive contributions making it less robust to dual-
task interference. In addition, it is largely unknown how these postural patterns differ within-
patients during painful episodes and during symptom remission. Therefore, an exploratory
purpose of these studies was to characterize these factors in patients during both symptom
remission and a painful episode.

Aim 1. Characterize the Balance-Dexterity Task and investigate associations between
performance on different aspects of the task – balance control, dexterous force control, and
trunk control. This aim elucidates how the trunk is coordinated in relation to performance on
dexterous control of the compression spring and balance control. I expected performance on
these two task demands to be associated, while trunk kinematics would vary independent of task
performance, indicating multiple potential strategies for trunk coordination.

Aim 2. Compare task performance and trunk control strategies during the Balance-
Dexterity Task in persons with and without a history of rLBP. This aim identifies trunk
control strategies during this dynamic balance task that persist during periods of symptom
remission. I expected there would be no differences in task performance, but that persons with
rLBP would exhibit greater trunk coupling associated with greater superficial trunk muscle
activity, in line with a “trunk stiffening” strategy.
TRUNK CONTROL DURING DYNAMIC BALANCE 4

Aim 3. Investigate effects of dual-task interference on task performance and trunk control
during the Balance-Dexterity Task and associations with social-cognitive factors. This aim
tests the hypothesis that trunk control strategies adopted by persons with rLBP rely on more
cognitive contributions and therefore exhibit greater changes under cognitive dual-task
interference. I expected to observe larger changes in trunk coupling and trunk muscle activity
under dual-task interference in this group compared to back-healthy controls. For all participants,
I expected the extent to which a person invests conscious attention in their movements (assessed
by the movement-specific reinvestment scale) would be associated with the degree of change
observed in trunk control measures under dual-task interference.

Exploratory Aim 4. Using a case series follow-up approach, compare trunk control in
persons with rLBP during both painful episodes and symptom remission. Three participants
were re-tested during a painful episode of LBP. Findings raise testable hypotheses to be
investigated in future work.
TRUNK CONTROL DURING DYNAMIC BALANCE 5

CHAPTER II
LITERATURE REVIEW
Investigating the trunk during postural control
The role of the trunk in balance dysfunction and rehabilitation has received less attention
compared to contributions from the lower extremities. It has long been understood, however, that
functional trunk control is necessary for any variety of postural control tasks – experimental or
ecological.
1–3
The importance of trunk control in standing balance and gait is supported by
observations that trunk muscle strength is associated with multiple balance measures and faller
status in older adults;
4
trunk extensor fatigue increases postural sway
5,6
even more so than fatigue
at the ankle or knee
7
and impairs recovery from a support surface translation;
8
trunk muscles
activate quickly and to a high level to compensate for a trip during walking;
9
and impairment in
recruiting deep abdominals is associated with functional disability in persons with chronic low
back pain.
10
In addition to an influence on global balance or functional performance, trunk
control during submaximal activities of daily living may play a role in more insidious overuse or
degenerative diseases like recurrent or chronic low back pain. Dysfunctional mechanical
behavior of the lumbar spine in particular, through motor control dysfunction and/or tissue
damage, is thought to play a role in degeneration through intervertebral instability
11
and/or
through increased load and decreased movement variability.
12–14
Persons with low back pain
exhibit greater intervertebral angular asymmetries associated with age and disc degeneration
during recumbent and standing trunk movements.
15
Clearly, the role of the trunk during postural
challenges of daily living needs to be rigorously investigated in order to learn more about healthy
human motor control as well as populations with chronic or recurrent pain in the low back. The
purpose of this literature review is to synthesize what is known about trunk function during
TRUNK CONTROL DURING DYNAMIC BALANCE 6

postural control tasks and to highlight recurrent low back pain as a model system in which to
study dysfunction of the trunk during postural control. This will be accomplished by
summarizing research from a variety of postural perturbation paradigms, presenting a foundation
for the experimental task used in these dissertation studies – the Balance-Dexterity Task – and
summarizing what is known about how attentional dual-tasking and psychometrics interact with
trunk postural control. Finally, recurrent low back pain will be discussed as a model system in
which to study dysfunction by presenting the high incidence of this condition and the specific
motor control impairments known to be present in persons suffering from recurrent low back
pain.
Discrete mechanical perturbations to posture
Using discrete perturbations to posture to study trunk control allows disambiguation of
cause and effect through carefully quantified inputs and outputs to the system. Discrete
perturbation paradigms discussed here will include both externally perturbed tasks – support
surface platform translations and tilts and mechanical perturbations to the trunk – and internally
driven tasks – voluntary limb movements.
Support surface platform perturbations
Support surface platform translations are accomplished by unexpectedly translating a
moveable platform on which subjects stand and quantifying measures of postural sway including
center of pressure and center of mass dynamics, body segment kinematics, and sometimes
muscle activation measures. One consistent finding is that center of pressure excursion usually
exceeds center of mass excursion in response to the perturbation.
16
The center of pressure moves
quickly toward the edge of the base of support to provide a force that can counteract the center of
TRUNK CONTROL DURING DYNAMIC BALANCE 7

mass movement.
1
Trunk muscles activate in response to any perturbation direction, but to a
larger amplitude when the perturbation is opposite the trunk muscle, i.e. paraspinals activated
more in response to forward perturbations and abdominals activated more in response to
backward perturbations.
17
In one study, lumbar extensor fatigue increased center of pressure
velocity among other measures in response to a translation.
8
van Drunen et. al. constructed a
device to specifically measure trunk responses to platform translations by translating an entire
semi-seated chair and quantified trunk stiffness using an inverted pendulum model with
kinematics and EMG.
18
With eyes open, subjects utilized only just enough trunk stiffness to
counteract the destabilizing perturbation forces (101% destabilizing forces), achieved by
decreasing reflex gains. With the eyes closed, participants increased trunk stiffness further
(114% destabilizing forces), potentially to have a greater margin for error at the cost of some
increased effort, but actually exhibited increased trunk motion, which was due to the greater
stiffness increasing forces transmitted to the center of mass. Overall, participants were able to
modulate intrinsic stiffness, damping, and feedback gains depending on sensory information
available and task instruction to behave naturally or to minimize trunk excursions. These
investigations provide evidence that the trunk is capable of contributing to whole-body balance
dynamics and could be targeted for balance interventions.
Mechanical external perturbations to the trunk
Trunk control has also been studied through direct mechanical perturbations to the trunk
both in this semi-seated posture seen previously and in more traditional whole-body standing
postures. The first group to conduct such a study investigated trunk muscle responses to a sudden
trunk release.
19
Participants held their trunk upright against 20% or 30% of maximum isometric
trunk exertion (responses averaged). Then, the magnetic cable was released, and muscle
TRUNK CONTROL DURING DYNAMIC BALANCE 8

activation timing was measured including the offset times for the agonists that were holding the
trunk upright and the onset times for antagonists. Participants turned off their agonists within 30-
60 ms of trunk release and turned on antagonists within 53-74 ms. These muscle latencies are
within the expected delays for a short-latency or monosynaptic reflex beginning to contribute 30-
60 ms after the perturbation. Such a reflex comes from the loop between primary muscle spindles
and alpha-motoneurons in the spinal cord, and is most directly measured in the lower extremity
using an H-reflex technique. In an ankle perturbation study, Kearney et. al. reports an EMG
response from this reflex about 40 ms after perturbation.
20
Another study estimates that the short-
latency reflex begins to contribute 20-45 ms after perturbation.
21
The normal value for soleus H-
reflex latency is 30.0 ms ± 2.1 ms.
22
Here, it appears participants react to the stretching of trunk
muscles during the trunk release quickly with contributions from this proprioceptive spinal
feedback loop. In order to assess effects of exertion on these trunk perturbation outcomes, one
group used a similar seated trunk flexion perturbation to model trunk stiffness parameters and
measure erector spinae reflex delays first 1-2 days after a triathlon (post-exercise) and then 4-5
days after (post-recovery).
23
Trunk stiffness was higher post-exercise than it was after recovery,
and paraspinal reflex delays were longer. The increased reflex delay after exercise may have
resulted from proprioceptive deficits in the fatigued muscles. Importantly, impairments in
proprioception or in these reflex loops may contribute to dysfunction. These researchers also
found that in 292 athletes, a delay in abdominal shut-off latency increased the odds of sustaining
a low back injury the following two to three years by 3% for every millisecond delay,
24

highlighting the importance of this quick reflexive control.
It is also important to note that multiplanar or asymmetric perturbations require more
complex responses. McGill with another group of researchers conducted a study with a
TRUNK CONTROL DURING DYNAMIC BALANCE 9

mechanical perturbation to the trunk in two conditions – the trunk was pulled posteriorly, and the
lateral trunk was pulled posteriormedially at a 30  angle. For the rotation perturbation, authors
report that the stiffness produced by trunk muscle preactivations may be less effective at
contributing to spine stabilization compared to the purely sagittal perturbation.
25
Studies of
isometric trunk exertions in different directions also report greater EMG variability in multi-
planar and asymmetric exertions.
26
Investigating this well-controlled trunk release paradigm has
allowed researchers to identify these effects of asymmetric exertion and reflex latencies.
Moving from seated to standing, researchers have found creative ways to perturb the
trunk in a more ecological posture. Chiang and Potvin used a standing posture with a strap
around the upper trunk that applies a perturbation laterally arguing that lateral loads require a
more complex interaction of trunk muscles since all muscles contributing to lateral bending also
influence moments in other planes.
27
In this task when participants pre-loaded the trunk with 5%-
15% of the perturbing force, they exhibited reduced trunk motion with greater preactivation of
muscles. Also, researchers report that the external oblique and lumbar erector spinae muscles
showed clear distinctions in activation when they were situated as agonists or antagonists. The
distinction was less obvious for the internal oblique, suggesting its role as a stabilizer
independent of perturbation direction. In another creative standing trunk perturbation,
participants stood holding a bucket in which water was dropped in a cued condition and an
uncued condition.
28
This study is the first described that utilized fine-wire EMG to instrument
deep fibers of the lumbar multifidus in a perturbation task like this. When the load was cued,
deep multifidus preactivated between 140 and 30 ms prior to load, followed by a quiet period,
and another activation period 50-90 ms after the load. When the load was not cued, there was no
preactivation but post-load activation remained between 50-100 ms. Both the amplitude and
TRUNK CONTROL DURING DYNAMIC BALANCE 10

duration of activity post-load were larger in the uncued condition. Superficial multifidus, in
contrast, did not preactivate in either condition, but had a similar response post-load. The
preactivation of the deep multifidus fits with theories that this muscle acts as a lumbar spine
stabilizer,
12,29
a concept not entirely supported by others.
11,30
The observation that the deep
multifidus shuts off 30 ms prior to load may be explained by electromechanical delay
31
and the
continuation of twitch force,
32
meaning muscle activation may have shut off but the force
application to the lumbar spine continued during the load.
Voluntary limb movements
The final type of discrete perturbation to posture discussed here may be the most
ecological – perturbations through voluntary limb movements. Early investigations into how
humans stabilize the body before a voluntary arm movement quantified the forces acting on the
body at the initiation of such a movement as well as the neural and motor preparation and
response to the movement. These types of investigations revealed that anticipatory postural
adjustments are highly dependent on the nature of the perturbing forces and that asymmetric
perturbations (e.g. a unilateral arm raise instead of a bilateral arm raise) requires larger and
longer duration anticipatory adjustments.
33
Activation levels of postural muscles (erector spinae
and hamstrings in one study
34
) were positively associated with the magnitude of arm
acceleration. For faster arm movements, the hamstrings activated prior to the anterior deltoid
with a latency correlated to arm acceleration. Together, postural muscle preactivations induce
extension of the lumbar spine prior to initiation of arm flexion, and more pre-movement trunk
extension resulted in less post-movement trunk flexion.
35
Rogers did a similarly structured
investigation into standing leg flexion and reported even more consistent findings that
preactivations of muscles in the supporting limb (postural muscles) were associated with
TRUNK CONTROL DURING DYNAMIC BALANCE 11

movement speed.
36
These seminal studies together conclude that primary movements and the
postural adjustments associated with them are parallel processes or part of the same motor
program, but are dependent on mechanical conditions (speed of movement, symmetry of
perturbing forces), sensory conditions (eyes open or closed), and behavioral conditions (self-
paced, instruction, feedback). It also appears that these relationships are more tightly maintained
for lower-extremity movements compared to upper-extremity, potentially due to the importance
of maintaining balance when the base of support is changing during lower-extremity movements.
Many studies have instrumented specific trunk muscles with surface or fine-wire EMG in
order to study how they preactivate or respond to contralateral voluntary limb movements. In
some older studies utilizing fine-wire EMG to instrument each abdominal layer – transverse
abdominis, internal oblique, and external oblique – researchers showed that the transverse
abdominis alone preactivated no matter which direction the arm is moved – flexion, abduction,
extension
37,38
– and preactivated independent of task complexity.
39
Another study reported that
the transverse abdominis and internal oblique preactivated prior to arm flexion more so in fast
movements than slow movements.
40
Later studies began to utilize fine-wire EMG in the
paraspinals to instrument deep fibers of the lumbar multifidus and found preactivations here
compared to superficial fibers of the multifidus, independent of arm movement direction in one
study,
38
but only prior to arm flexion in others.
41,42
When arm flexion is perturbed by adding
cable resistance in the first third of the movement, the first muscles to respond were the
transverse abdominis and erector spinae along with measured increases in intra-abdominal
pressure.
43
These responses were within 50 ms of the perturbation, indicating, again, the
involvement of a short-latency reflex loop. Other research groups contradict these findings by
instrumenting these muscles bilaterally and claiming that they are in fact activated in a direction-
TRUNK CONTROL DURING DYNAMIC BALANCE 12

dependent manner.
44
When it comes to lower-extremity movements, however, much like what
was seen in earlier seminal papers summarized above, preactivations are much more robust.
Hodges and Richardson report that the transverse abdominis and oblique muscles all preactivate
before contralateral hip flexion, abduction, and extension.
45
The lumbar multifidus preactivates
before prone hip extension.
46
Again, it is possible that the larger number of muscles preactivating
prior to multiple lower-extremity movement directions could be due to the higher consequences
of failed standing balance or more work needs to be done in lower extremity movements to
identify similar caveats or complexities found in arm movement perturbations.
These types of discrete perturbations have allowed researchers to identify specific
postural patterns – both feedforward and feedback – that allow humans to cope with postural
challenges of everyday life. Trunk muscle latencies indicate that short-latency reflex loops help
to stabilize the trunk both in sitting and standing. In preparation of voluntary arm movements,
primarily deep trunk muscles preactivate to counteract destabilizing forces, but with mixed
results. These findings are more robust when investigating lower-extremity movements. Many of
these discrete perturbation paradigms, however, are limited ecologically to the rare discrete
perturbations in daily life and do not reflect conditions where posture is challenged in a
continuous dynamic way as it is in most activities of daily living. More complex continuous
tasks – experimental and ecological – need to be studied in order to bring what has been learned
about postural control from discrete perturbation studies further into the realm of application.

TRUNK CONTROL DURING DYNAMIC BALANCE 13

Continuous mechanical perturbations to posture
Mechanical external perturbations to the trunk
Many researchers have seen the need highlighted here for more continuous perturbations
to the trunk. Researchers have developed perturbation paradigms moving from discrete trunk
releases and pulls to the application of pseudorandom forces directly to the trunk in semi-
seated
47–49
or standing postures.
20,50
These types of continuous perturbations where subjects are
instructed either to resist or relax, allow us to quantify intrinsic and reflexive contributions to low
back stiffness with system identification analysis techniques. When instructed to resist the
continuous pseudorandom forces, participants increased muscle activation, trunk stiffness, and
reflex gains (by 73%).
48
In a modeling study where potential feedback signals were used to
predict trunk motion, which was then matched from experimental results, researchers showed
that intrinsic trunk stiffness and damping; length, velocity, and acceleration signals from muscle
spindles; and inhibition signals from Golgi tendon organs were required to recreate observed
trunk motion, but adding other signals including vestibular feedback did not enhance the fit.
49

Another group that moved this perturbation technique to a standing posture, also found short
latency reflexes (25-75 ms after perturbation) present in all six trunk muscle groups they
instrumented in response to the small, continuous perturbations.
50
Authors reported that the
iliocostalis and internal oblique responded with the fastest reflexes, and external oblique
responded through a long-latency reflex. They didn’t however, find any association between
reflex measures from EMG and modelled lumbar mechanical stability. It is possible that the
researchers discovered a limitation of applying what we learn from these semi-seated, well-
controlled perturbation experiments to more ecological balance conditions or in applying their
TRUNK CONTROL DURING DYNAMIC BALANCE 14

modeling procedures to these more complex body configurations, but findings remain clear that
short-latency reflexes are a crucial component to maintaining appropriate trunk stiffness.
Unstable sitting
These continuous perturbations discussed thus far have been applied directly to the trunk,
but more realistic continuous variable forces would likely come from the foot-ground interface
and therefore be delivered from the lower-extremity or pelvis up the trunk. Perhaps this is what
multiple research groups were thinking when they developed unstable sitting tasks. Some of
these devices allow instability in a single plane
51
and others in multiple planes.
52
Healthy
participants complete this task with about 40  of lumbar flexion and exhibit a positive correlation
between lumbar and thoracic sagittal plane rotation.
52
Muscle coordination influenced thoraco-
lumbar movement variability such that a higher ratio of longissimus (considered here an inter-
segmental muscle) to iliocostalis (a multi-segmental muscle) was associated with less movement
variability.
52
Quick trunk muscle reflexes are necessary in this task to maintain low trunk
displacements and moments.
53
As stiffness of an unstable seat decreases, activation of all
muscles functioning as agonists also increased and activation of the external and internal
obliques when functioning as antagonists also increased.
51
These findings, again, point toward
the obliques (transverse abdominis was not instrumented) and deeper inter-segmental paraspinals
as trunk stabilizers independent of perturbation direction, a finding seen now in discrete and
continuous perturbations.
There is now evidence from diverse trunk control investigations that the trunk is
controlled by short-latency reflexes activating quickly or primarily deeper trunk muscles
including the transverse abdominis, obliques, and deep fibers of the lumbar multifidus. Some
TRUNK CONTROL DURING DYNAMIC BALANCE 15

studies also add longissimus, as an inter-segmental paraspinal to this list. Clearly, however, these
conclusions are highly task-dependent with increased variability and complexity with
asymmetric tasks.
Double- and single-limb standing balance
Double- and single- limb balance are some of the most ecological postural demands we
face every day. We know that during prolonged standing, some persons will develop low back
pain and others will not,
54,55
and specific predictors for this have been identified including
lumbar lordosis and movement during standing.
56,57
These measures have not been useful,
however, in short bouts of unperturbed double-limb standing. In these short bouts, certain aspects
of balance discriminate between persons with and without low back pain such as increased
postural sway in the patient group.
58
But trunk control specifically is difficult to measure in
short-duration, unperturbed standing. Making balance more challenging by removing vision
through a blindfold and confounding proprioception with a foam surface results in increases in
deviation of the thoracic and lumbar spine segments in the frontal plane.
59
We also know that
fatiguing low back musculature has the largest effects on postural sway measures compared to
ankle, knee, and shoulder fatigue protocols.
7
Removing or perturbing sensation is useful in an
experimental setting but creates limitations to applying findings to real-world scenarios.
When it comes to muscle contributions to trunk control during standing balance, muscle
activations and forces are quite low (sometimes near-zero) in unperturbed double-limb stance.
One modeling study found that this position – double-limb standing in a neutral posture with
near-zero muscle force – was the most mechanically unstable for the lumbar spine, and
activation of the lumbar multifidus and erector spinae up to 1-3% maximum voluntary
contraction was necessary to restore stability.
11
But it is likely that stability is simply not required
TRUNK CONTROL DURING DYNAMIC BALANCE 16

in this very submaximal postural control task. When balance demands are increased by reducing
the stability of the platform on which participants stand, thickness (measured by ultrasound) of
the transverse abdominis, internal oblique, and external oblique all increase, with the deepest
layers increasing to the greatest extent.
60
Again, these findings rely on perturbations to balance –
either sensory (blindfold, foam pad) or mechanical (unstable support surface). It is difficult to
conduct research into trunk dysfunction when we want to use an ecological, submaximal,
continuous task with sensory signal intact like double-limb standing, but need to evoke some
response from the control system.
Single-limb balance represents an increase in difficulty and instability compared to
double-limb balance and may show some promise as a useful experimental and ecological task.
Single-limb balance tasks are typically done with the lifted leg relaxed,
61
but a variation called
the “stork task” with the lifted hip and knee flexed to 90° has also been used in the literature.
62

The instruction in these tasks is to maintain balance as long as possible, using 30 seconds as a
typical cutoff.
Many authors discuss balance strategies in terms of ankle or hip strategies.
63
Participants
using an ankle strategy act as an inverted pendulum with greatest joint motion at the ankle. This
strategy is effective for small perturbations or typical small center of pressure sway during
double-limb stance. Participants using a hip strategy move the hip and ankle out of phase to
maintain the center of mass over the base of support, and this strategy is necessary for larger
perturbations to balance. It also imparts more demands on the trunk, and researchers have
suggested that persons suffering from low back pain may avoid using a hip strategy for
balance.
64
Looking specifically at trunk motion, some researchers have developed complex
marker tracking and analysis techniques that allow them to report motion and motion variability
TRUNK CONTROL DURING DYNAMIC BALANCE 17

of specific spinal segments. In non-disabled controls, the “stability time” (quantified as time
below a certain threshold of resultant segment rotations) was greater in the lumbar spine segment
and lower thorax segment than in the upper thorax or the trunk as a whole.
65,66
With eyes closed,
stability was maintained in these two regions but dropped significantly in the upper thorax and
overall trunk. Single-limb stance certainly represents an increase in postural challenge compared
to double-limb stance, but high-functioning populations still perform the task with minimal trunk
motion and the role of the trunk in balance is difficult to ascertain.
Adding complexity and instability to standing postures has been a recent focus from
multiple groups in balance and trunk control literature. Donath et al. reported increasing postural
sway in a series of balance tasks in this order – double-limb stance with eyes open on foam,
double-limb stance with eyes closed, double-limb stance with eyes open on foam and one leg up
on a step, double-limb stance with eyes closed and one leg up on a step, and single-limb stance
with eyes open.
67
They also reported that young adults had more co-contraction at the trunk
while older adults had more co-contraction at the ankle in most tasks. Similar progressions of
balance task difficulty have been studied with comparable results.
68–70
Research groups around
the world have acknowledged the need for continuous balance tasks that are difficult enough to
elicit robust postural control responses.
The Balance-Dexterity Task
The Balance-Dexterity Task is a combination of single-limb stance and a modified lower-
extremity dexterity test (LED-test). First, it is important to discuss the motor control processes of
each of these tasks individually and how we might observe these processes with biomechanical
measures, starting with the LED-test. This test in its pure form involves the participant semi-
seated on a bicycle seat with the non-dominant leg on the ground and the dominant leg on a
TRUNK CONTROL DURING DYNAMIC BALANCE 18

small platform mounted on a spring attached to the ground.
71
The arms and trunk are supported
by a stable platform at elbow level with hand grips. The participant is instructed to compress the
spring, which is prone to buckling at low compression forces, and is provided visual feedback
about the achieved vertical compression force. Performance on the test, measured by the
maximum stable vertical compression force achieved by the participant, requires dynamic
control of limb endpoint force, and is meant to quantify the dexterous control ability of the
lower-extremity. Performance is not highly influenced by the participant’s lower-extremity
strength capability. This was validated in a study of adolescent soccer athletes where
performance on the LED-test was correlated with performance on the cross-agility test (R
2
=0.63)
but was not correlated with hip extensor strength, knee extensor strength, or knee flexor strength
(R
2
=0.036, R
2
=0.002, and R
2
=0.019, respectively).
71,72
The LED-test has been used to study a
variety of athletic populations including soccer athletes
73
and cross-country skiers.
74
A recent
study investigating contributions to balance from measures of static and dynamic single-limb
balance, the Y-balance test, vertical jump, and the LED-test using a principal component analysis
technique found that these assessments captured distinctly different aspects of balance control.
75

The test was designed based on the strength-dexterity test, which utilizes the same principle of
dynamic endpoint force control to compress an unstable spring with the fingers.
76
There are
important similarities between the mechanics and the neurophysiology of the tasks, but there are
also important differences. Therefore, to explore motor control processes in the LED-test, we can
use research into object manipulation tasks of the fingers while incorporating important
differences at the leg such as longer nerve conduction distances and different somatosensory
capabilities of the shod foot. Complex motor control processes are required to accomplish the
LED-test and are discussed here.
TRUNK CONTROL DURING DYNAMIC BALANCE 19

Like in precision grip and finger manipulation tasks, somatosensory information is the
most important intrinsic sensory component of the LED-test,
77
since other frequent contributors
to balance like vision and vestibular systems are constrained by the bicycle seat and trunk and
arm rest. In finger manipulation tasks like the strength-dexterity test or object manipulation
tasks, we know of four types of somatosensory receptors – fast-adapting and slow-adapting type
I and II afferents. These are used by the human body to sense specific manipulation events and
contact force magnitude and direction. In the LED-test, similar principles apply, but the shod
foot is not as sensitive, or at least does not have as high a concentration of these specific afferent
neurons, as the fingertips. That being said, acquiring specific information about the forces at the
foot-platform interface is crucial to the success of the task. In the analogous fingertip task, force
direction is communicated to the central nervous system by the activity of afferent neuron
populations preferentially tuned to specific directions, force magnitude is communicated by rate
coding of the afferents, and force timing is communicated by the timing of the firing of these
afferents.
77
In the LED-test where dynamic control of the limb endpoint force is required, all of
these – force direction, magnitude, and timing – are important sensory signals as well, but unlike
in the fingers where only tendons are crossing distal joints, the lower-extremity has uni- and bi-
articular muscles crossing the hip, knee, and ankle, all of which are positioned to acquire
information about these force parameters and limb position. Muscle spindles with length and
velocity receptors and Golgi tendon organs with force receptors convey this information from the
musculotendon units to the central nervous system. Even ligaments and capsular tissue in the
joints themselves convey information about joint position, though it is thought these provide the
most reliable information near joint end-ranges, not mid-ranges where the LED-test takes place.
Finally, it is worth mentioning that vision does play a role in this task but in extrinsic, not
TRUNK CONTROL DURING DYNAMIC BALANCE 20

intrinsic, feedback processing as the participant looks at a real-time trace of his or her task
performance – the vertical spring compression force.
The LED-test, like the strength-dexterity test and object manipulation tasks, is a task that
exemplifies the pairing of feedforward and feedback control requiring both the sensory and
motor processes described above. The participant begins and carries out the LED-test with a
feedforward goal of pure vertical force control and therefore pure vertical spring compression
with no buckling. Spring characteristics, however, are difficult to ascertain through visual
inspection. In pilot testing, the first time participants step on the spring, almost without fail, they
overestimate how much they will be able to compress the spring and buckle the springs. A
memory representation of the device is quickly updated, and the participant is usually successful
on the second attempt. What makes them successful during this continuous task is constant
comparison of the feedforward plan for pure vertical compression and the sensory information
discussed above about limb position and forces at the foot-platform interface. Comparison of
these signals allows for rapid error detection and context-dependent corrections. Long time
delays in sensorimotor control loops means dexterous manipulation is not possible without an
accurate feedforward plan, hence the requirement for accurate device characteristics and the
typical failure on the first attempt without them. Some of these feedback signals can be observed
by quantifying the foot-platform interaction forces with three-dimensional force measures (e.g.
measuring when and how much the spring buckles) and the stability of the vertical force goal
(e.g. using coefficient of variation).
As was described earlier, to move the field forward, investigations of trunk motor control
during balance must utilize continuous dynamic tasks. The most straightforward dynamic
postural control task is double-limb balance. This can and has been made more challenging by
TRUNK CONTROL DURING DYNAMIC BALANCE 21

revising to single-limb balance and/or balancing on a foam surface or with the eyes closed. Even
with these alterations, authors report that differences between healthy controls and patients with
low back pain can only be identified when the balance conditions are very challenging. Perhaps
this is because low back pain is typically a low-level chronic condition that may or may not
affect daily functioning to a large degree, i.e. patients are able to compensate through redundant
postural control mechanisms. One goal of these dissertation studies is to develop a motor task
that is sufficiently challenging while not blocking any motor control processes that participants
are able to use in their daily lives. In other words, we opt to make the task more challenging by
adding this concurrent dexterity task as opposed to removing vision, for example. Therefore, a
combination of these two tasks has been developed (Figure II.1). The bicycle seat and arm rests
were removed from the LED-test so as to not provide support to the trunk and pelvis. In addition,
the instructions and feedback were modified in order to change the purpose of the device from
quantifying maximal lower-limb dexterity to providing a challenging but submaximal dynamic
perturbing force to balance.

Figure II.1. The Balance-Dexterity Task.
TRUNK CONTROL DURING DYNAMIC BALANCE 22

Finally, activities of daily living require both single-limb balance ability and dexterous
control of endpoint forces between the foot and ground. The spring control component of this
task may be ecologically related to walking on uneven, changing, or slippery surfaces.
Combining these tasks links the endpoint force control of the limb controlling the spring with
that of the limb controlling the single-limb balance task because both share the responsibility of
supporting the body and controlling the center of mass, as is the case in many daily activities.
This may be analogous to bimanual tasks with different goals for each hand – stability and
dexterous manipulation – translated to the lower-extremity. While discrete perturbations teach us
something about muscle and control dysfunction and may point us in the direction of which
tissues to target, there is no meaningful way to rehabilitate the low back or postural control
dysfunction using just discrete perturbations. Clinicians need continuous tasks meant to mimic
real-life scenarios that require postural control, and the proposed combination of the modified
LED-test and single-limb balance is one such task.
Trunk muscles contribute to stability of the lumbar spine
Directly quantifying stiffness and stability of the lumbar spine is outside the scope of
these dissertation studies, but appropriate interpretation of kinematic and electromyographic data
rely on an understanding of how trunk muscles contribute to these phenomena. This becomes
especially important in the context of dysfunctional trunk control and injury risk. Starting with
“the problem”, it has long been understood that active contributions to lumbar spine stability are
necessary to prevent buckling, given that buckling occurs in cadaveric models at as low as 88 N
of vertical compression.
78
Stability has diverse meanings for different audiences. Generally,
stability refers to a system’s tendency to return to its unperturbed state after a perturbation.
79
For
a clinician, this is assessed visually by how a person looks when they perform a task. Aspects of
TRUNK CONTROL DURING DYNAMIC BALANCE 23

motion are examined like variability of the movement, speed of the movement, and the presence
of any out-of-plane or compensatory motions. “Stabilizers” may be identified by finding
muscles, cues, or assistive devices that reduce undesired motions. Mechanically, stability is
quantified much more specifically, usually through EMG-assisted modelling and a minimum
potential energy method.
11
This allows very specific assessments of the contributions to stability,
but is subject to the limitations of modelling-based approaches. In these dissertation studies, this
conceptualization of stability will be termed “mechanical stability”. Here, contributions to
lumbar spine stability will be summarized including contraction of deep paraspinal muscles,
intra-abdominal pressure, connective tissue coupling, and functional bracing.
One contributing mechanism to lumbar spine stability is contraction of deep paraspinal
muscles, sometimes called intrinsic, local, or inter-segmental paraspinal muscles. The lumbar
multifidus is an example of such a muscle, arising from vertebral laminae and spinous processes
very close to the axis of rotation and spanning between two and five vertebral levels.
80
The
structure and architecture of the lumbar multifidus support its role as a stabilizer. The small
moment arm of the muscle lends more toward intervertebral compression and mitigation of shear
forces more so than segmental rotations.
29
In addition, the muscle has a large cross-sectional area
and small fiber length suggesting its role in sustained force production.
81
The multifidus consists
of primarily type I muscle fibers,
82
and activation patterns of the multifidus suggest motor unit
firing is self-sustained and rotates among motor units to prevent fatigue,
83
all of which support
the postural role of these muscles. Functionally, there is evidence that humans preactivate the
deep fibers of the lumbar multifidus before voluntary limb movements
41
and expected postural
perturbations
28
independent of perturbation direction. In contrast, more superficial muscles,
sometimes called extrinsic, global, or multi-segmental paraspinal muscles, attach more laterally.
TRUNK CONTROL DURING DYNAMIC BALANCE 24

The erector spinae muscles attaching to lumbar transverse processes or skipping the lumbar
vertebrae all together and attaching from pelvis to ribcage are an example.
Another mechanism that contributes to lumbar spine stability is the regulation of intra-
abdominal pressure. In vivo, increased intra-abdominal pressure increases lumbar intervertebral
stiffness,
84
and in one cadaveric study, a physiologically realistic intra-abdominal pressure
implemented through balloons placed in the abdominal cavity provided enough force to
straighten a laterally bent trunk.
85
Again, while not direct measurements of mechanical stability,
both of these would contribute to a more mechanically stable lumbar spine in certain conditions.
All muscles around the abdominal cavity will contribute to intra-abdominal pressure including
the abdominals, pelvic floor, and diaphragm. The transversus abdominis, given its transversely-
oriented fibers and attachments to the thoracolumbar fascia, has the potential to be a primary
driver of intra-abdominal pressure. Activation of this muscle alone can increase intra-abdominal
pressure and lumbar intervertebral stiffness as shown in a porcine model.
86
And, similar to the
deep fibers of the lumbar multifidus, humans preactivate the transversus abdominis prior to a
voluntary limb movement or expected postural perturbations. It has been argued, however, that
other abdominal muscles contribute more to intra-abdominal pressure.
87
Comparing abdominal
hollowing, which activates primarily transversus abdominis and internal oblique, with abdominal
bracing, which activates those muscles in addition to external oblique, rectus abdominis, and
more muscles surrounding the abdominal cavity,
88
bracing has a much greater effect on intra-
abdominal pressure.
89
Importantly, many modeling studies investigating lumbar spine stability
do not include intra-abdominal pressure in the model.
11,30,87,90,91
This is an important limitation
when interpreting findings.
TRUNK CONTROL DURING DYNAMIC BALANCE 25

Abdominals and paraspinals both participate in another mechanism contributing to
lumbar spine stability termed here – “connective tissue balance”. Thoracolumbar fascia has a
posterior layer which arises from the midline of the lumbar spine over the spinous processes and
wraps around the posterior side of the paraspinal compartment as well as a middle layer which
arises from transverse processes and covers the anterior side of the paraspinal compartment
meeting the posterior layer laterally where they serve as attachments for the common tendon of
the transversus abdominis and the internal oblique muscles.
85
Cadaveric studies have examined
the influence of abdominal contraction (by pulling on the anterior portion of the fascia
anteriolaterally), intra-compartmental pressure (by increasing the size of the paraspinal
compartment simulating contraction of the erector spinae and/or multifidus), and intra-abdominal
pressure (by inflating balloons in an intact cadaveric abdominal cavity).
85,92
Tesh et. al. reported
that intra-compartmental and intra-abdominal pressure increases had similar effects on lumbar
spine stiffness in the sagittal plane, but with more local effects of intra-compartmental and more
diffuse effects of intra-abdominal pressure.
85
Intra-abdominal pressure through tensioning the
middle layer increased stiffness in the frontal plane.
85
Vleeming et. al. found that a balance of
intra-compartment pressure through paraspinal contraction and tension in the common tendon of
the transversus abdominis and internal oblique was necessary for “self-bracing” of the spine.
92

Richardson et. al. found some direct evidence of this by measuring increased sacroiliac joint
stiffness (measured with ultrasound Doppler) with contraction of the abdominals when lying
prone.
93

Thus far, we have considered smaller, deeper muscles focusing primarily on the lumbar
multifidus, transversus abdominis, and internal oblique, but larger, more superficial muscles may
also play a role in maintaining lumbar spine stability. More superficial muscles like the erector
TRUNK CONTROL DURING DYNAMIC BALANCE 26

spinae have larger moment arms and cross multiple vertebral levels, which means they have a
greater potential to stabilize the spine when activated symmetrically,
94
but also to destabilize the
spine when activated asymmetrically to a high level.
25,95
Modeling studies based on in vivo
kinematic, kinetic, and electromyographic data have shown that all major muscle groups
contribute to mechanical stability of the spine in large multiplanar movements, much more so
than passive structures.
11
In a model of the lumbar spine during isometric trunk exertions (20% -
60% maximum exertions), standing, and forward lifting tasks, one muscle at a time was
“knocked out” to quantify the effects on mechanical stability. No one muscle contributed more
than 30% to mechanical stability, and there was no uniform pattern of muscles’ contributions
when it came to inter- or multi-segmental muscles or muscle size.
30
Comparing the abdominal
bracing and hollowing techniques described above, researchers found that bracing increases
mechanical stability of the spine and reduces lumbar spine displacements in response to a
mechanical perturbation to the trunk, but also increases compression.
25,87,88
But, interestingly, a
hollowing maneuver increased sacroiliac joint stiffness more than a bracing maneuver in a prone
position.
93
These findings point toward the idea that more muscle activity increases mechanical
stability, at least in tasks that are multi-planar or loaded.
For tasks that are very submaximal or require asymmetric control, these conclusions
about which mechanisms contribute more or less to lumbar spine stability must be interpreted
with caution. Take, for instance, the Cholewicki and McGill model which showed that all major
muscle groups contribute to stability. In tasks that did not require large muscle force, they
reported that contributions from passive structures to mechanical stability played a more crucial
role.
11
The instances when the lumbar spine was most mechanically unstable were in neutral
postures with near-zero muscle forces in the multifidus and lumbar erector spinae, and activation
TRUNK CONTROL DURING DYNAMIC BALANCE 27

levels in these muscles needed to reach 1-3% maximum voluntary contraction to restore
stability.
11
McGill with another group of researchers conducted a study with a mechanical
perturbation to the trunk in two conditions – the trunk was pulled posteriorly, and the lateral
trunk was pulled posteriormedially at a 30  angle. Here, authors report that the stiffness produced
by trunk muscle preactivations may be less effective at contributing to spine stabilization
compared to the purely sagittal perturbation.
25

Clearly, muscle contributions to lumbar spine stability are complex and highly task-
dependent. This may be why McGill suggests “functional bracing” as the goal for effective and
safe trunk control.
96
Task demands must be considered before beginning to interpret how
measured muscle activations are contributing to lumbar spine stability. The Balance-Dexterity
Task, which takes place in a neutral or slightly flexed lumbar spine posture, is asymmetrical, and
submaximal. It stands to reason that the primary contributors to lumbar spine stability will be
contraction of deep paraspinals, regulation of intra-abdominal pressure, and connective tissue
balance. These are accomplished in this submaximal task with appropriate activation of deep and
superficial trunk musculature.
Cognitive contributions to posture
These dissertation studies aim not only to investigate performance and control strategies
used for this novel dynamic task but to probe mechanisms by perturbing cognitive systems
contributing to the neuromotor control of posture. Such a probe will not only provide unique
insight into the role and weighting of central processing contributions to posture in a healthy
population but also how this role may be different in persons suffering from recurrent low back
pain. In addition, postural control during activities of daily living is almost always accompanied
by posture-unrelated cognitive activity making conditions of cognitive task interference highly
TRUNK CONTROL DURING DYNAMIC BALANCE 28

ecological. Findings suggest activities with high risk of postural instability such as ambulation on
changing surfaces or under attentional interference e.g. while recalling a grocery or to-do list.
Effects of this cognitive perturbation will offer insight into dysfunction of the central systems
contributing to postural control in this population and probing mechanisms of dysfunction and
systems contributing to this dysfunction is necessary in developing targeted and effective
interventions for balance dysfunction.
Theoretical framework of attentional resources
The construct of attention, or more importantly, the construct of multiple attention
resources, has been outlined by Wickens’ multiple resource theory.
97
Attention is sometimes
discussed specifically as visual attention, but more broadly can be defined as the cognitive
process of selectively concentrating on a discrete aspect of information. Wickens’ multiple
resource theory adds that attention connotes awareness. The presence or ability to concentrate
attention is something everyone is familiar with, but Wickens presents a more nuanced
theoretical framework outlining a four-dimensional model of attention resources. In Wickens’
model, the four dimensions are: processing stages, perceptual modalities, visual channels, and
processing codes (Figure II.2).
97
Further, in order to make quantitative predictions about the
relative interference between tasks, Wickens developed a conflict matrix approach to sum
vectors representing task demands based on how many overlapping dimensions are utilized in
each task.
TRUNK CONTROL DURING DYNAMIC BALANCE 29

Figure II.2. Attentional resources model (adapted from Wickens, 2002).
Support for this construct of attention resources comes from Wickens’ strikingly accurate
predictions about task performance and time requirements in dual-task paradigms. Let's take the
construct of two processing codes – manual/spatial and verbal/vocal – since these are an aspect
of this model relevant to these dissertation studies. One dual-task study by Wickens and Liu
investigated straightforward interference of these two processing codes.
98
Participants completed
a cognitive task that required mental rotation of angles presented on a screen (in Figure II.2
above: manual/spatial codes throughout focal visual perception and cognition), a manual tracking
task with the dominant hand that required moving a joystick (in Figure II.2 above: manual/spatial
codes in cognition and response stages) and concurrently had to respond to an auditory cue as
quickly as possible in one of two conditions – either by hitting a keypad (in Figure II.2 above:
manual/spatial codes in cognition and response stages) or by speaking a command (in Figure II.2
above: verbal/vocal codes in cognition and response stages). As is predicted by Wickens’
conflict matrix model, participants increased reaction time and decreased accuracy of the
tracking task more-so under the manual response condition (hitting a keypad) than the vocal
TRUNK CONTROL DURING DYNAMIC BALANCE 30

response condition (speaking a command) because of the increased overlap of attention
resources. Even his extremely specific quantitative predictions of interference in easy and hard
conditions of each task variation were supported by the data. Similar support continues to be
found in more recent studies of this type like that of Azuma et. al. where subjects concurrently
performed a tone-choice task, cross motion recognition task, and motion response task.
99
This
model is even used in government policy to support recent laws passed requiring cell phone
voice commands to be used while driving as opposed to hand-held cell phone use.
The construct of attention processing codes intersects with the construct of working
memory. Working memory is defined as the holding and working with information in the mind
that is no longer perceptually present in the environment.
100
It is demarcated into verbal working
memory and visuo-spatial working memory and therefore intersects with Wickens’ model in the
cognition stage of these two processing codes. The concept of working memory is strongly
supported with bodies of work in psychology, imaging, development, skill learning, and other
diverse fields citing the construct. Using fMRI data, Eldreth et. al. identified areas in the
prefrontal cortex that increased in activity scaling with the difficulty of a working memory task
where participants had to reorder a set of letters.
101
These were distinctly different from areas
scaled during simple recall (not reorder) tasks indicating that the manipulation processes in
working memory is different from the recall processes thought to belong to short-term memory,
hence the construct name: “working” memory.
Dual-task interference methodology
Attentional resources are probed using dual-task paradigms, and these are generally
divided into dual-task probe paradigms and dual-task interference paradigms. Hui-Ting Goh
presents a wonderful description of these, summarized here.
102,103
Dual-task probe paradigms
TRUNK CONTROL DURING DYNAMIC BALANCE 31

involve participants completing two different continuous motor tasks (e.g. double- and single-
limb balance) and responding to the same discrete cognitive task (e.g. respond to a sound as
quickly as possible) in each condition. The difference in performance on the cognitive task (e.g.
longer reaction time in single-limb balance condition) reveals which motor task occupied more
attentional resources. In contrast, dual-task interference involves just one continuous motor task
completed with and without a concurrent continuous cognitive task (e.g. serial subtraction).
Here, performance on the motor task is compared between conditions (e.g. in double-limb
stance, postural sway changed under conditions of dual-task interference). Many studies utilize
dual-task interference paradigms to mimic ecological attention conditions. An important
limitation of these paradigms, however, is how to ensure participants are assigning appropriate
effort or priority to these two tasks. Here, researchers have developed a “primacy switch”
condition. So, in the most robust dual-task interference research studies, four conditions are
tested – (1) motor task, (2), cognitive task, (3) dual-task with the priority on the cognitive task,
and (4) dual-task with the priority on the cognitive task. This allows the effects of dual-task
interference on the motor task to be “double-checked” but putting the priority back on the motor
task and seeing if effects of interference are reversed.
102
In these dissertation studies, a dual-task
interference paradigm with a primacy switch is used to assess effects of interference on trunk
control during the Balance-Dexterity Task.
Importantly, cognitive tasks must be developed specific to the research questions being
asked. As outlined by the Wickens model,
97
choosing a cognitive task that probes similar
attentional resources as the motor task being studied will invoke the most dual-task interference.
The proposed studies also utilize a dual-task interference paradigm but with the specific intention
of interfering with conscious control of movement, described later in terms of Reinvestment
TRUNK CONTROL DURING DYNAMIC BALANCE 32

Theory. This specificity is accomplished by choosing a continuous verbal working memory
recall and manipulation task meant to interfere with verbal resources known to be used during
movement-specific attentional reinvestment.
104
The verbal working memory recall and
manipulation task used here involves participants remembering five numbers and operating on
them twice to produce five answers.
105
A constraint present in this particular research study is
that deep abdominals, which are instrumented with fine-wire electromyography, will fire with
verbalization.
106
Therefore, variations to cognitive tasks found frequently in literature had to be
made to prevent real-time verbalization of the answers and to hold all answers in working
memory until the completion of the trial.
Dual-tasking and postural control
Attentional resources utilized for postural control have been studied with dual-task
interference paradigms in non-disabled adults and diverse patient populations. A general theory
of posture or cognition prioritization has been proposed where weighting of these factors is
dependent on: the motor and cognitive state during a specific dual-task situation; postural reserve
and the compensatory capabilities of the individual; and individual characteristics like
personality, affect, and training.
107
In persons with high postural reserve and conditions of low
threat of falling, cognitive tasks can be prioritized, but a change in either of these factors
(dysfunctional postural control or injury; increase in fall risk) can induce a shift toward postural
task prioritization.
107
Understanding how trunk control, as a control strategy that can vary
independently from balance task performance in the Balance-Dexterity Task, is affected by these
various dual-task interference conditions will help reveal how cognition is involved in control of
the trunk during balance.
TRUNK CONTROL DURING DYNAMIC BALANCE 33

Investigations of healthy persons using dual-task interference paradigms to study posture
have produced mixed results, but by focusing on studies that have used recent methodological
guidelines for posture-cognition dual-tasking,
108
there are some conclusions that can be extracted
from the literature. In straight-forward double-limb standing, a concurrent working memory
cognitive task causes an increase in center of pressure frequency and a reduction in sway
amplitude.
109
Standing in a high threat condition (at the edge of a raised surface) resulted in these
same changes.
110
And, another study investigating interference of an auditory tone-counting task
on quiet standing reported decreases in center of pressure average velocity and anterior-posterior
standard deviation.
111
In standing, therefore, cognitive dual-tasking seems to “tighten” center of
pressure control within a smaller area and with higher frequency corrections. Through
perturbation research, however, it appears this is not as effective a postural control strategy.
When a platform perturbation is induced under dual-task interference conditions, center of
pressure peak displacements are larger.
112
These larger displacements are associated with
reductions in electroencephalography peaks related to sensory inputs for postural control.
112

Overall, it appears that cognitive dual-tasking, at least for working memory cognitive tasks,
reduces postural sway amplitude, processing of sensory inputs related to posture, and the ability
to control center of pressure displacements under mechanical perturbations to balance.
Moving the investigation of dual-task interference to trunk control, a few studies have
laid the foundations for these dissertation studies through discrete perturbations to posture.
Hodges conducted a voluntary arm flexion study similar to those discussed earlier, but with an
added condition that increased the complexity of the task.
39
In all conditions, participants saw
light cues about which way to move the arm, but in a complex condition the cues provided
ambiguous or incorrect information about the instructed direction forcing the participants to
TRUNK CONTROL DURING DYNAMIC BALANCE 34

increase cognitive processing. In response to this condition, Hodges reported an increase in
reaction time for the deltoid (prime mover) and superficial abdominals, but no change in reaction
time for the transverse abdominis. Similarly, Jones et. al. conducted a study where participants
sat on an exercise ball and flexed their hip in response to a cue with and without a concurrent
cognitive task.
113
This group also reported delays in the onset of the rectus femoris (prime
mover) but no change in latencies of trunk muscles. The feedforward postural muscle activations
appear to be less affected by dual-task interference compared to the voluntary activation of prime
movers, at least in non-disabled young adults.
Finally, investigating dual-task interference effects on gait allows a more continuous
functional task to be studied.
114
Work from our own lab reported decreased step length
variability, decreased trunk-pelvis coupling variability, and decreased hip motion in dual-task
interference during a walking pivot turn.
115
Participants in this study in symptom remission from
recurrent low back pain exhibited increased trunk frontal plane motion under dual-task
interference. Asai et. al. reports that a serial subtraction task done concurrently with locomotion
resulted in an increase in trunk motion, and this was more pronounced in a group of older adults
who scored high on a fear of falling survey.
116
In a functional continuous task like gait, dual-task
interference appears to reduce movement variability, potentially indicating less reliance on
sensory inputs similar to what was seen in standing above, and using a more stereotypical
movement pattern. In impaired populations, however, (e.g. low back pain and older adults with
fear of falling here) this could also result in increases in trunk motion.

TRUNK CONTROL DURING DYNAMIC BALANCE 35

Reinvestment Theory
Theoretical framework
The theory of reinvestment forms the underpinnings for hypotheses about trunk control
and the effects of dual-task interference in these dissertation studies. Movement-specific
reinvestment, defined by Masters and Maxwell as the “manipulation of conscious, explicit, rule
based knowledge, by working memory, to control the mechanics of one’s movements during
motor output.”
117
A critical appraisal of the theory is warranted and presented here. Appraising a
theory must include an evaluation of the constructs on which the theory is based and an
evaluation of the theory’s ability to predict meaningful differences in behavior or task
performance. The theory of reinvestment is no different from other psychological theories in that
it is based on hypothetical constructs that cannot be measured directly.
The first major construct to discuss is that of attention, which was summarized in the
Dual-task interference section above. The construct and model designed by Wickens are useful
for investigating task-dependent attention patterns, but only to describe multitasking in a relative
sense and only within one individual, pointing out an important weakness of the application of
the construct to reinvestment theory especially in studies investigating group or population
differences.
One more construct to introduce behind reinvestment theory is the construct of
declarative memory. This construct is typically discussed as a distinction of long-term memory
called explicit (or declarative) memory different from implicit (or non-declarative or procedural)
memory. Explicit memory is the deliberate or conscious retrieval of previous experiences as well
as conscious recall of factual knowledge about people, places, and things, whereas implicit
memory is an unconscious form of memory that is only evident in the performance of a task.
118

TRUNK CONTROL DURING DYNAMIC BALANCE 36

These memories held in long-term storage are brought to working memory when needed to
complete a task. Let’s take, as an example, one of the first times this distinction was discovered –
in the patient H.M. This patient, after a bilateral medial temporal lobe resection which resulted in
the removal of his hippocampus, was unable to form new explicit long-term memories.
Researchers working with H.M. discovered, however, that he could learn complex drawing tasks
and maintain skilled performance over multiple days even though he had no explicit memory of
the task.
118-119
The explanation of this performance is that H.M. was able to learn the task
implicitly, store that information in long-term memory, and retrieve it. Similarly, riding a bike is
an example of implicit skill knowledge many people store and retrieve. Knowing how to perform
the skill explicitly – titling the bike at an angle relative to the product of the curvature desired
and the speed squared – does not typically help people learn or remember the skill.
Reinvestment theory states that when reinvestment occurs, verbal, explicit working
memory is used during task performance. Evidence for this comes mainly from research where
participants who practice a task in a condition meant to evoke attention reinvestment to the
motor task (e.g. asking golfers to speak at different phases in a putting task) results in these
participants acquiring a greater volume of declarative task-relevant knowledge.
120
Even between
participants, those who have a greater tendency to reinvest accumulated more task-relevant
declarative knowledge during a golf putting learning protocol.
121
While this suggests that
reinvestment involves declarative memory, these self-report measures taken after task or practice
completion cannot confirm that this information was being used during the task. Evidence verbal
working memory resources are used during task performance is sparse. Some evidence at least
pointing in that direction comes from studies that have found a positive correlation between trait
verbal working memory capacity and a tendency to reinvest
122
and have linked it to
TRUNK CONTROL DURING DYNAMIC BALANCE 37

electroencephalography (EEG) collection during performance of a novel tennis task.
104
These
authors found a negative correlation between visuo-spatial working memory capacity and EEG
signals and a positive correlation between verbal working memory capacity and these EEG
signals. Together, this indicates that individuals with greater verbal working memory capacity,
which is correlated with a greater tendency to reinvest, used more verbal working memory
resources during the task. Stronger evidence is still desired to support this link.
Reinvestment theory fits into the motor skill learning timeline proposed by Fitts and
Posner and reiterated many times in literature by other authors states that motor skill learning
progresses from declarative or explicit working memory control during novice stages to
procedural or implicit, sometimes also called automatic, control as an expert.
123,124
This idea has
been tested and confirmed for decades and supported recently with EEG evidence showing
decreases in verbal working memory signals as a novel finger tapping task is learned.
125
This
timeline is where the “re-“ in reinvestment theory comes from. If task performers are expert,
their performance is automatic and they no longer require conscious attention focused on
declarative aspects of performance as they did when they were novices. If the performer
“reinvests” attention in this performance, however, then they are once again devoting verbal,
explicit attention resources to the task.
So far, support has been cited showing that reinvestment involves verbal working
memory attention resources devoted to task performance, but the last important construct here is
the idea of self-focus or self-regulation. Unlike in Wickens’ model of attention resources where
these resources can be directed at different tasks, reinvestment theory defines the direction of
these attention or working memory resources to control the mechanics of one’s own movements.
Duval et. al. provided a definition of self-focus or self-regulation as the deliberate self-control of
TRUNK CONTROL DURING DYNAMIC BALANCE 38

thoughts and emotions (and Masters and Maxwell
117
would add: movements) associated with
attaining personal goals.
126
A specific instance of self-focus is the practice of rumination,
continuously cycling thoughts around a common theme even when the stimulus for the thoughts
is not present. One might predict that persons with a tendency to ruminate will also have a
tendency to reinvest because of this link, and this has been found to be the case with a significant
positive correlation between these measures in a population of 312 participants.
127
This idea of
self-focus is also related to Wulf and colleagues’ discussion of internal and external focus of
attention. Their constrained action hypothesis that an internal focus of attention encourages
conscious control of movements and inhibits automatic control mechanisms is actually built
upon reinvestment theory.
128
In a standing task where postural sway was cued with an internal
and an external focus of attention, participants exhibited more EEG coherence associated with
verbal processing centers in the internal focus condition.
129
Here, an important weakness of the
Masters and Maxwell theory appears – there is an important difference between conscious
control and conscious monitoring of movement. We not aware of any work investigating how
independent these processes are or if conscious monitoring inexorably leads to conscious control.
Either way, the construct of self-focus has support from multiple research groups and has real
measurable effects on task performance.
Mechanism of reinvestment interference
The process by which reinvestment disrupts performance is called the progression-
regression hypothesis described by Fuchs.
130
This hypothesis was described earlier in terms of
the timeline of motor skill learning. After one has progressed from novice (performing using
declarative, explicit control) to expert (performing using automatic control) it is possible to
revert back or regress (the “re-” in reinvest) to explicit, verbal control. There are implications of
TRUNK CONTROL DURING DYNAMIC BALANCE 39

this regression seen in task performance, control mechanisms utilized, muscle activation patterns,
and joint dynamics.
First, the proposed mechanism of reinvestment’s action to interfere with skill
performance is an increase in verbal/analytical activity in the brain related to the increase in
conscious attention to movement. Evidence for this comes from EEG studies like those cited
previously where coherence between verbal-analytical regions of the left hemisphere and the
motor planning frontal region of the right hemisphere are high in conditions of reinvestment like
under high pressure performance situations.
104
It has even been proposed, but not empirically
supported, that handedness may play a role in the tendency to reinvest.
117
Since there is a less
direct link between the verbal/analytical center on the left hemisphere to the left hand, left-
handed performers may be protected against reinvestment because of the added delay in neural
communication.
During reinvestment from these verbal, explicit processes, tasks that have been
automatized are broken down again into their parts. This is most clearly explained using a serial
task with multiple degrees of freedom. As a person learns the task, the distinctness of each serial
step disappears, the transition times between the steps become shorter, and the task starts looking
like one fluid motion. If one regresses or reinvests attention in each serial step again, they regain
their distinctness and the result is a more staggered, slower movement. MacMahon and Masters
reported this exact phenomenon during the learning of a serial reaction time task as differences
in the task parts disappeared after twenty-one blocks of practice but reappeared once stress was
introduced into the task with a spotlight and video camera.
131

TRUNK CONTROL DURING DYNAMIC BALANCE 40

Wulf and her colleagues discuss changes in the control strategies used when evaluating
internal and external focus conditions, which were described earlier as predicated on
reinvestment theory, during a balance stabilometer task.
132
Participants exhibited higher
frequency corrections during the external focus condition indicating faster reflexive control of
balance as well as faster reaction times of a dual-task probe test indicating less cognitive
attention demand of the task during the external focus condition. Therefore it is reasoned that
during an internal focus condition, when attention is reinvested in movement, conscious higher-
level processing of movement is used and reflexive control is inhibited.
Changes during reinvestment be seen at the muscle and joint level as well. When
participants stand at the edge of a high platform, they score higher on reinvestment scales
133
and
exhibit a stiffer ankle strategy with more ankle muscle co-contraction.
134
Vance, Wulf, and
colleagues also report that electromyography recordings of muscle activity during an internal
focus condition of bicep curls indicate less efficient muscle recruitment patterns measured by
increased median frequency.
135
Finally, and perhaps most related to these dissertation studies, the
Hodges research group reported a positive correlation between trunk stiffness and score on a fear
of movement scale.
136
While authors here did not quantify reinvestment, taken in light of these
other studies it may be hypothesized that reinvestment was the intermediate step between fear of
movement and measured increased joint stiffness. Therefore, through mechanisms of increased
co-contraction, less efficient muscle recruitment, and increased joint stiffness, reinvestment has
clear observable effects on movement control.
All of these mechanistic changes in movement control coalesce to result in decreased task
performance. This is supported by a litany of studies cited by Masters and Maxwell and
conducted since. Some of these, in addition to those cited previously, are decreased golf putting
TRUNK CONTROL DURING DYNAMIC BALANCE 41

performance when participants are asked to identify putting movement phases as opposed to
identify external, unrelated auditory cues
120
and decreased acrobatic trampoline performance
when performers were given declarative cues.
137
Overall, through mechanisms of verbal
attention control, the breaking down of automated tasks, the inhibition of reflexive control, and
less efficient muscle activation and joint dynamics, reinvestment disrupts task performance. In
these dissertation studies, it was hypothesized that scores on reinvestment scales will be
positively correlated with these same outcomes including increased trunk kinematic coupling
(through the control system’s reduction in degrees of freedom), less efficient muscle activation,
and disruption from concurrent verbal dual-task interference.
Measuring a tendency to reinvest
Many studies cited thus far have referred to participants’ “tendency to reinvest” or “score
on reinvestment scales.” It is prudent, therefore, to discuss how reinvestment as a hypothetical
construct is quantified. The tool used to measure movement-specific reinvestment is the
Movement-Specific Reinvestment Scale (MSRS) (Figure II.3).
117
This scale has ten items, each
of which is rated using a six-point Lickert scale between ‘strongly disagree’ and ‘strongly agree’,
and is split into two components: Conscious Motor Processing (CMP) and Movement Self-
Consciousness (MSC).
TRUNK CONTROL DURING DYNAMIC BALANCE 42

Figure II.3. The Movement-Specific Reinvestment Scale (Masters and
Maxwell, 2008).
There are some distinct strengths and weaknesses of this scale. A major strength and
hypothesis in the proposed dissertation study is that these items probing reinvestment, especially
the CMP items, may be a mechanistic link between the high scores on scales used in low back
pain research like fear-avoidance beliefs, kinesiophobia, and pain catastrophizing and behavioral
outcomes measures. In other words, fear of pain or re-injury may lead to movement
reinvestment, which may lead to changes in motor performance. Discovering such a link would
provide therapists another effective means of intervention for suited patients. In addition to this,
however, reinvestment may also capture a phenomenon in healthy controls not motivated by fear
or pain. For example, a trained dancer likely will have high movement reinvestment due to the
high volume of internal focus cues given during training of certain dance tasks.
138
This healthy
dancer may score high on the MSRS without the influence of fear of pain. A scale that works
equally well for the patients and the healthy controls in a study is an important strength.
TRUNK CONTROL DURING DYNAMIC BALANCE 43

Another strength of the MSRS lies in its concurrent and discriminant validity evaluated
by a four-part study conducted by Laborde et. al. in the European Union.
127
Comparing scores on
the MSRS with scores on a preference for intuition and deliberation inventory showed that
MSRS had convergent validity with a preference for deliberation, thinking through decision-
making step-by-step, and discriminant validity with a preference for intuition, making decisions
based on a quick judgment without much conscious awareness. In addition, MSRS score
correlated with trait characteristics like self-consciousness, some perfectionism measures, and a
tendency to ruminate as a means of working through issues. Discriminant validity was found
with trait anxiety and a tendency to use distraction as a means to work through issues. All of this
together with reported test-retest reliability of 0.71 (on CMP) and 0.67 (on MSC) by one study
117

and 0.78 (on CMP) and 0.61 (on MSC) by another
127
and with reported internal consistency
Cronbach’s alpha of 0.71 (on CMP) and 0.78 (on MSC) and of 0.67 (on CMP) and 0.73 (on
MSC) by these same two studies indicates the strength of this particularly valid and reliable
scale.
The scale has been used successfully in a number of patient populations providing
evidence not only for its theoretical strength but its practical utility. Orrell and colleagues
conducted a study with 162 survivors of stroke compared to 148 healthy controls and found not
only that stroke survivors scored higher on the MSRS but that CMP score and time spent in
rehabilitation were significant predictors of impairment following stroke, meaning that a
tendency to consciously process motor output could lead to worse functional outcomes for these
patients.
139
Another group of researchers found that duration of Parkinson disease was positively
correlated with an increased score on the MSRS.
140
And finally, a study of older adults found
TRUNK CONTROL DURING DYNAMIC BALANCE 44

that those who had fallen scored higher on both CMP and MSC components of the MSRS, and
there was a significant association between CMP score and faller status.
141

It is, of course, important to point out weaknesses of the scale, especially weaknesses or
limitations for the scale’s use in these dissertation studies. The biggest limitation here is that
persons with recurrent low back pain, especially during periods of symptom remission, do not
visibly move differently than their healthy counterparts. While they may still consciously process
their movements, reflected by their CMP score, it is unlikely they will score high on any of the
MSC items, for example “If I see my reflection in a shop window, I will examine my
movements” or “I am concerned about what people think about me when I am moving.” These
questions geared more toward the “style” of movement as opposed to the conscious
contemplation of movement may only be applicable for patient populations with visible
disability like in severe stroke or Parkinson disease conditions. In addition, two different studies
report lower test-retest reliability of the MSC compared to the CMP and only the CMP was
related to a motor imagery capability measure,
127
indicating that it measures something distinctly
different and more relevant to the current studies.
Factors that influence or invoke attention reinvestment in movement
Our appraisal of reinvestment theory thus far has included a review of the constructs on
which it is based, the mechanisms by which it affects movement, and the methods by which it is
measured. An important question left unanswered is: What causes someone to reinvest? Masters
and Maxwell provide a long list of such “contingencies that cause reinvestment,” which are
discussed here with some more recent examples and some specific applications and predictions
to these dissertation studies.
TRUNK CONTROL DURING DYNAMIC BALANCE 45

First, it is crucial to mention that reinvestment is entirely appropriate under circumstances
that require attention. For example, rain during a sporting event may make players pay more
attention to their movements, especially dangerous ones like running and cutting on slippery
ground. Another example is that a dancer may plateau in his or her dance ability and seek out a
new dancer teacher that has to “go back to the basics” and help the dancer unlearn certain habits
to build new ones. Such an attentional reinvestment may be detrimental in the short term but may
allow the dancer to surpass his or her previous performance plateau. While these instances of
appropriate reinvestment do exist, the connotation of reinvestment theory is that reinvestment
usually occurs involuntarily and results in decrements to task performance.
Personality traits are one such contingency that can involuntarily cause reinvestment to
occur. As has been described already, personality characteristics like perfectionism, self-
consciousness, preference for deliberation, and tendency to ruminate are all related to movement-
specific reinvestment.
127
Along with these characteristics is an individual’s tendency to
accumulate task-relevant declarative knowledge during skill learning. Those who accumulate
such knowledge are more likely to reinvest attention in their movements under pressure
conditions.
120

Reinvestment can also be induced by changes in performance conditions. One such
change that has been studied carefully is a change in sport equipment. Beilock et. al. reported
that expert golfers do not have a detailed episodic memory of their putting mechanics, as would
be expected for expert performers who are not reinvesting, but providing these golfers with an
‘S’ shaped putter resulted in a large increase in their knowledge and explicit memory of their
putting mechanics.
120
Changes in performer psychology because of performance conditions can
induce reinvestment as well. Performing under anxiety, pressure, or for an audience can cause
TRUNK CONTROL DURING DYNAMIC BALANCE 46

increased conscious processing of motor output. Many studies cited use techniques like
videotaping performance, having a judge present, adding a spotlight, setting a time limit, or
adding high monetary stakes to induce a high-pressure situation and measure corresponding
increases in attentional reinvestment.
122,131,133

Other psychological situations that can invoke reinvestment are related to practice
situations. If an athlete is practicing too much or has too much preparation time before a game or
performance, he or she may start “hypothesis-testing” his or her movements. This is a form of
attentional reinvestment and will likely have the same negative consequences for performance.
In fact, Masters and Maxwell recount a time-honored ruse of athletes to give the opponent extra
time before their play in the hopes of them taking that time to reinvest attention to their motor
plan. In the author’s eloquent words: “Cricketers will sometimes present a slow, full
toss…because the batter has time to think, ‘this is an easy one, I am expected to knock it to the
boundary – now, where should my front foot be…”
117
Such reinvestment will disrupt the batter’s
swing and play to the advantage of the pitcher.
Finally, and most importantly for these current studies, reinvestment can be caused by
injury or accident. A clear example of this is the studied cited earlier investigating tendencies to
reinvest in older adults, some who had fallen previously and other who had not. Those who had
fallen scored higher on reinvestment scales and there was a significant association between the
conscious motor processing component score and faller status.
141
In order to tease out cause and
effect a group of older adult non-fallers was studied.
142
During walking, participants had to
respond to a question either about their movement mechanics or about aspects of the
environment. Participants with a higher tendency to reinvest answered questions about their own
mechanics correctly more often and questions about the environment less often. Authors suggest
TRUNK CONTROL DURING DYNAMIC BALANCE 47

this could suggest a tendency to reinvest attention in movement is a causal factor for a future fall
because the ability to acquire important information about the environment during locomotion
may be impaired.
Pain or injury is the most obvious potential cause of reinvestment for participants with
recurrent low back pain in these dissertation studies. The hypothesis here is that patients are
reinvesting attentional resources to control their movements because of a fear of pain or re-
injury. It is hypothesized that all participants, both healthy controls and patients, who score high
on the movement-specific reinvestment scale, specifically the conscious motor processing
component, will show the predicted motor control characteristics of reinvestment described
previously. More specifically, however, patients with recurrent low back will have a positive
correlation between scores on fear-avoidance beliefs, kinesiophobia, and pain catastrophizing
and their scores on the movement-specific reinvestment scale. All participants who score high on
this scale will have a large change in outcome variables when performing the task with
concurrent verbal dual-task interference. This can be explained using different sets of constructs
described in this paper – these processes compete for the same attention resources; the verbal
dual-task invokes an external focus of verbal attention instead of the internal focus of movement
reinvestment; or the dual-task prohibits the focus and accumulation of declarative task-relevant
knowledge.
Reinvestment and posture
Reinvestment applied specifically to postural control research is a new and cutting edge
line of research. Searching these keywords together in PubMed reveals only eleven results, the
earliest of which is from 2009.
133
One study reported associations between score on the MSRS
and postural sway measures during double-limb standing.
111
Greater scores on MSRS were
TRUNK CONTROL DURING DYNAMIC BALANCE 48

associated with greater center of pressure medial-lateral standard deviation and reduced center of
pressure complexity (“more constrained,” quantified using a sample entropy method). These
associations disappeared in the dual-task interference condition where a tone-counting task was
done concurrently. The authors explain this in terms of Reinvestment Theory reporting that a
greater tendency to invest conscious attention to movement reduces movement degrees of
freedom contributing to center of pressure control – resulting in greater sway and less complex
trajectory – and that under the dual-task condition “secondary task loading reduces resources
available for conscious control.” In a population of healthy young adults completing three
standing postural sway tasks – one uncued, one with an internal focus of attention, and one with
an external focus – EEG coherence among verbal/analytical centers was greatest with the
internal focus of attention.
129
Importantly, however, there was no association between EEG
measures and MSRS scores in the uncued condition. Authors explain that the task – simple quiet
double-limb standing – may not have been challenging or perturbing enough to utilize a
participant’s tendency to reinvest. This strengthens the argument for a continuous balance task
with the added challenge of dexterous force control.
Though not directly measuring reinvestment, some studies using postural threat allow us
to make inferences about reinvestment because we know injury
117
and certain psychometrics
127

are related to reinvestment and standing in a high-threat condition increases scores on the
MSRS.
133
In response to the question, “What were you thinking about?” standing at a height
increased the amount of answers related to movement processes and self-regulatory strategies.
110

And, changes in postural sway were related to the change in attention devoted to movement
processes in the same direction as described previously – more attention to movement processes
associated with greater sway amplitude. Overall, research on the effects of movement-specific
TRUNK CONTROL DURING DYNAMIC BALANCE 49

reinvestment on postural control are lacking and effects on the role of the trunk in postural
control are completely absent.
Recurrent low back pain
Epidemiology
Low back pain, defined as pain between the lowest rib and the gluteal fold,
143
is an
incredibly prevalent condition with up to 80% of adults experiencing at least one episode during
their lifetime.
144
The global age-standardized point prevalence has been reported at 9.4%, but in
industrialized countries this is quite higher, for example 15.0% in western Europe.
145
Out of 291
conditions studied, low back pain ranked as the greatest contributor to global disability. The
annual prevalence of activity-limiting low back pain has been reported at 38%.
146
In a study of
US health care spending in 2013, low back and neck pain accounted for $87.6 billion – the
highest spending level for any musculoskeletal condition and the third highest of any health
condition behind diabetes and ischemic heart disease.
147
Unfortunately, despite decades of
research, healthcare spending on low back and neck pain has increased the most of any condition
studied, excluding diabetes, in the last eighteen years.
Persistent pain in the low back is all too common after an acute episode. A recent
systematic review analyzed studies that followed persons for a period of time after they
recovered from a first acute episode of low back pain and reported recurrence rates between
22.1% at three months and 77.1% at three years post-initial episode.
148
One very well-designed
study used latent class analysis statistical technique to analyze the healthcare records of 65,790
adults, focusing on healthcare utilization after an initial incident low back pain healthcare
visit.
149
These adults fell into four groups – (1) apparent recoverers who did not have a high
probability of utilizing additional healthcare services (53%), (2) recurrent sufferers who
TRUNK CONTROL DURING DYNAMIC BALANCE 50

maintained about a 40% chance of utilizing additional healthcare services (31%), (3) chronic
sufferers who maintained a >75% chance of utilizing additional healthcare services and primarily
used therapeutic services (8%), and (4) another group of chronic sufferers that primarily utilized
pain medications (7%). Much research has been done investigating patients with chronic and
symptomatic low back pain, but studying persons with recurrent episodes of pain in periods of
symptom remission has unique advantages. Here, the effect of current pain can be removed and
residual effects of psychological or motor control changes can be identified. Often, predictions
about how these effects contribute to the recurrence of pain are the focus of such investigations
in an attempt to apply the information learned to all pain presentations – acute, recurrent, and
chronic.
Motor control dysfunction
Muscle morphology, intervertebral instability, and proprioceptive impairments
Changes in trunk muscles have been characterized for decades, a summary of some
recent evidence of dysfunction is presented here. First, research has consistently shown that low
back musculature is atrophied in persons with low back pain. A recent and well-designed study
from our lab found that this atrophy was localized to the lumbar multifidus at the level of pain,
but was bilateral even in unilateral low back pain.
150
Other groups have reported increased fatty
infiltration, but one study only found fatty infiltration in a continuous chronic low back pain
group and not in a recurrent low back pain group.
151
The recurrent low back pain group,
however, had reduced metabolic measures in the lumbar multifidus.
151
In addition, these back
extensors are more quickly fatigued with differences in the electromyographic spectral pattern of
the multifidus suggesting faster dropout of type-II muscle fibers.
152
At the local intervertebral
level, a study of thirty-five persons with recurrent low back pain revealed increased L5/S1
TRUNK CONTROL DURING DYNAMIC BALANCE 51

translation; reduced flexion/extension motion at L1/L2, L2/L3, and L4/L5; and increased axial
rotation at L4/L5.
153
And, at L4/L5 and L5/S1 only, there were significant associations between
disc grade and measures of flexion/extension motion, axial rotation, and translation. Another
study reported greater intervertebral angular motion sharing inequality in persons with chronic
low back pain during standing flexion and recumbent flexion, extension, rotation, and side-
bending.
15
Here, also, this measure was associated with disc degeneration.
The human trunk contains an array of muscles that cross single spinal units including the
rotatores and interspinales, which are of such small cross-sectional area and have such a high
concentration of muscle spindles that they are thought to function mainly as position and motion
transducers.
154
These transducers along with sensory organs in other larger muscles are what
allow participants to sense the location of their trunk in space and detect trunk motion. Persons
with low back pain, however, when asked to reposition the trunk at a specified previous angle
produce greater repositioning error than healthy controls.
155
In addition, when the trunk is moved
slowly, patients with low back pain have a higher movement detection threshold than their
healthy counterparts.
156
Though not studied in persons with low back pain, trunk repositioning
errors were negatively associated with Berg Balance Score in a group of patients post-stroke,
157

meaning more errors was associated with poorer balance, indicating the potential influence of
impaired trunk proprioception on balance.
Findings from discrete perturbations to posture
In response to discrete support surface platform perturbations, persons with recurrent low
back pain in symptom remission exhibited reduced and delayed center of pressure movement,
but increased center of mass displacement.
16
Another study identified increased trunk muscle co-
TRUNK CONTROL DURING DYNAMIC BALANCE 52

contraction and reduced trunk peak torques.
17
Looking at a population currently in an episode of
low back pain, Jones et. al. reported no difference in trunk peak torques, along with increased
activity of prime movers (trunk muscles that were stretched by a particular perturbation
direction) and reduced activity of what the authors call stabilizers (trunk muscles that activated
regardless of perturbation direction in pain-free controls; in this study, the internal oblique was
the most consistent).
158
Along the same lines, MacDonald et. al. reported reduced or entirely
absent lumbar multifidus activation prior and in response to trunk loading.
28
This suggests that
episodes of pain are characterized by reduced deep muscle activity, but enough increase in prime
mover activity to maintain trunk peak torques. But then in periods of symptom remission a co-
contraction stiffening strategy is adopted to keep trunk peak torques low, at the expensive of fast
and appropriate center of pressure movement which are necessary to limit center of mass
displacement. It is also possible the reduced proprioception summarized above contributes to the
delayed and reduced center of pressure movement.
Mechanical perturbation directly to the trunk have also been tested in persons with
recurrent low back pain during symptom remission. Here, trunk muscle co-contraction patterns
were also identified. In response to a semi-seated trunk release, muscle shut-off and turn-on
latencies were delayed and more agonists remained active, compared to controls with no history
of low back pain.
19
The delayed shut-off latencies were associated with future low back pain
within two to three years, where the risk of low back pain increased 3% for every millisecond
abdominal shut-off latency delay and those who experienced low back pain during follow-up had
an average 14 ms delay.
24
This increase in co-contraction results in increased modeled trunk
mechanical stability
159
and explains the greater trunk stiffness seen in persons with recurrent low
back pain during unanticipated trunk loading,
160
and after recovery from exercise.
23

TRUNK CONTROL DURING DYNAMIC BALANCE 53

Many of the specific muscles with altered activation patterns were identified using
voluntary limb movements as a perturbation to posture. In anticipation of voluntary arm
movement, persons with recurrent low back pain in symptom remission delayed activation of the
deep lumbar multifidus
41
and transverse abdominis
37,40
in all directions of arm movement, and
the internal oblique
37,40
in certain directions. Experimentally induced muscle pain through
injection of hypertonic saline also reduced activity of the transverse abdominis and erector
spinae.
161
Though studied less frequently, perturbations through voluntary lower extremity
movements show similar results of delayed activation of transverse abdominis and lumbar
multifidus consistently, and internal oblique and erector spinae dependent on movement
direction.
46,162
The lack of these anticipatory activations likely coincides with kinematic findings
from another study where less preparatory trunk extension resulted in greater trunk flexion
displacement as an effect of arm flexion.
35
Though the response displacement was small (3.2 ), it
indicates there are kinematic consequences to the lack of anticipatory activation of these deep
trunk muscles. In addition, one study found that improvement in the delays in transverse
abdominis activation after a motor control intervention were associated with a reduction in
functional disability,
10
suggesting a functional influence of these preactivations as well.
Findings from continuous perturbations to posture
Continuous perturbations to posture, distinctly different from cyclic or periodic tasks like
gait, have been less frequently studied in persons with recurrent low back pain. One study
utilized continuous pseudorandom perturbations to the trunk in a semi-seated posture and
modeled reflex gains in “resist” and “relax” conditions. Persons with current low back pain
showed smaller changes between conditions, specifically less increases in reflex gains from the
“relax” to “resist” conditions compared to controls.
47,50
One group combined muscle onset delays
TRUNK CONTROL DURING DYNAMIC BALANCE 54

during a semi-seated trunk release with performance on an unstable seated balance task and
found that the greater delays in persons with chronic low back pain (pain 2.7/10 at time of
testing) were associated with poorer center of pressure control (greater RMS, maximum
displacement, and path length).
163
When modeling realistic increases in trunk muscle delays,
trunk displacement and trunk moments were increased, but not balance instability.
53
But other
studies found no difference
52
or reduced
164
center of pressure sway in persons with recurrent low
back pain in remission. Again, in persons with chronic low back pain, some studies report greater
thoraco-lumbar movements and reduced trunk stiffness associated with reduced deep-to-
superficial paraspinal muscle activation ratios
52
while others report increased trunk stiffness
associated with increased trunk muscle co-contraction in response to discrete release of the
unstable seat.
165–167
Clearly, there are mixed results without definite conclusions for persons with
recurrent low back pain in symptom remission, but this type of continuous unstable balance task
where perturbations to posture are delivered “bottom up” through the pelvis represents a
necessary step in moving toward ecological, continuous, and challenging balance tasks.
Standing balance
Balance ability in persons with recurrent low back pain in remission appears to be
reduced with lower Y-Balance scores
168
and increased postural sway.
58
More research has been
done, however, in persons currently in pain. For these patients, center of pressure measures are
more consistently increased.
61,169
In a well-designed study taking participants through a series of
more complex balance tasks – double-limb standing, tandem standing, and single-limb standing
with variations of eyes open and eyes closed – persons with chronic low back pain had greater
center of pressure sway measures and this differences was even larger as tasks became more
difficult.
68
Additionally, one study on persons with chronic low back pain found increased
TRUNK CONTROL DURING DYNAMIC BALANCE 55

thoracic and lumbar deviations in the frontal plane in standing with eyes closed.
59
Control from
longer-latency reflexes or voluntary/cognitive control may also be used more heavily. One group
of researchers tested this prediction that patients with recurrent low back pain in remission will
not highly weight proprioceptive information from the trunk during double-limb balance.
64

Healthy participants and participants with low back pain, in a crossed design, stood on a solid
surface and on a foam surface and had their plantarflexors and their low back vibrated in each of
the two stance conditions to interfere with muscle spindle sensation. In healthy participants,
larger increases in sway were seen when the ankles were vibrated in the solid surface condition
and when the back was vibrated in the foam condition. This indicated that participants utilized
ankle muscle spindles when information from the ankles was reliable (on a solid surface), and
utilized back muscle spindles when the information from the ankle was unreliable (on a foam
surface). Measurements from the low back pain group, however, showed they utilized ankle
muscle spindles in all conditions even when information from them was less reliable, indicating
that these patients failed to use sensory information from the back and supporting the idea that
this sensory information is less heavily weighted in this patient population.
Many authors suggest these effects are related to a reluctance of patients to utilize a hip
strategy to control balance and instead a reliance on ankle strategy, which will result in less
efficient control of center of pressure and mass. Motion at the hip also requires segmental trunk
motion, which would be inhibited if the trunk is stiffened through over-active superficial trunk
muscles. Evidence for this comes from a study on abdominal muscle activation during unstable
double-limb balance which showed decreased activity of transverse abdominis and internal
oblique in the low back pain group and increased activity of the external oblique.
60
This
reluctance to use a hip strategy then intuitively could be related to the increased stiffness of the
TRUNK CONTROL DURING DYNAMIC BALANCE 56

lumbar spine seen in the more controlled perturbation paradigms discussed earlier. In single-limb
standing in persons with current low back pain as well as in remission from pain, the lumbar
spine shows the greatest stability compared to other spinal regions.
62,66
Overall, it appears that
the trunk is stiffened in persons with recurrent low back pain through overactivity of superficial
trunk muscles and reduced activity of deeper trunk muscles, but the findings seem mixed and
highly task and population dependent.
Functional tasks
Finally, lumbar motion and muscle activation patterns during a few relevant functional
tasks are summarized. In a systematic review investigating effects of both current and a history
of low back pain on a variety of functional tasks, overall conclusions included reduced lumbar
range of motion and increase in movement variability.
170
A step-down task in persons in current
pain also showed reduced lumbar movement.
171
In gait, there are mixed findings depending on
the population studied. In persons in current pain, there are increased correlations of thorax and
pelvis motion,
172,173
but studies on persons with low back pain in symptom remission show no
difference in thorax-pelvis correlations.
174
A study in our lab of walking turns in persons in
remission from recurrent low back pain revealed no difference in movement variability or in-
phase trunk coupling, but an increase in the portion of time spent in trunk phase.
175
In addition,
control participants increased lumbar multifidus activation duration during the turn when
movement speed was increased, but persons with recurrent low back pain reduced the duration of
activity,
176
again highlighting dysfunction in these deeper paraspinals. Overall, findings are
mixed and specificity as to the plane of motion being observed, the task being used, and the pain
presentation and characteristics of the group studied are crucial.
TRUNK CONTROL DURING DYNAMIC BALANCE 57

Psychosocial Factors
Psychometrics were first introduced into the study and treatment of low back pain by
spine surgeons Waddell and colleagues.
177
These researchers discovered that a fear-avoidance
beliefs questionnaire predicted patients’ recovery after spine surgery and others have reported
similar findings for success of physical therapy interventions.
178
This questionnaire and others
are still used today by clinicians and researchers. George et. al. reported that many of these
different psychometrics are associated with each other, and the fear-avoidance beliefs about
physical activity and work were associated with pain intensity, disability, and physical
impairment in persons with chronic low back pain.
179
Kinesiophobia predicted scores on a timed-
up-and-go task in older adults with low back pain as well.
180
Patients with high fear-avoidance,
kinesiophobia, or pain catastrophizing will require some distinct interventions including
cognitive behavioral therapy to prevent pain or fear of pain from limiting progress. An
appropriate clinical trial has been done where a subgroup of patients scoring high on fear-
avoidance beliefs received a distinct intervention protocol.
181
The group with a matched
intervention had significantly better outcome measures than a group without a matched
intervention. These alterations in psychological factors influence cognitive processing and
attention demands, highlighted by one task using a dot-probe attention task to investigate effects
of seeing pain-related words or images on attention in persons with chronic low back pain.
182

Increased center of mass displacements in response to platform perturbations were
positively associated with pain-related fear and activity interference.
183
In this same study, P2
potentials measured by electroencephalography, which are thought to represent monitoring of
postural challenge, were negatively correlated with center of mass displacements. So, persons
with low back pain who monitor their posture more exhibited reduced center of mass
TRUNK CONTROL DURING DYNAMIC BALANCE 58

displacements after the perturbation. If we interpret these EEG measures in the reinvestment
theory framework, this highlights a potential benefit of reinvesting attention in posture for this
patient group, at least if prioritizing balance is desired, which may be task-dependent.
In contrast, unperturbed standing balance has less frequently shown associations with
psychometrics. Hooper et. al. reported no association between Y-Balance score and
psychometrics.
168
Mazaheri et. al. reported increased sway in persons with chronic and recurrent
low back pain but no association with kinesiophobia or or pain catastrophizing.
184
Sung, Silfies,
and colleagues reported no association between self-reported fear of movement and pain
characteristics and postural sway during unstable sitting.
185
Looking at global balance measures,
it appears psychometrics may play a more important role in perturbed balance than quiet
standing. This fits with previous findings suggesting different control strategies during externally
imposed mechanical perturbations in contrast to self-controlled unstable forces.
Though there are limited associations between psychometrics and balance measures, the
connections between psychometrics and trunk control specifically are much stronger. Research
into the mechanistic links between these psychological measures and motor control dysfunction
described thus far is cutting edge and controversial. Karayannis, Hodges, and colleagues
identified a positive correlation between score on a fear of movement scale and trunk stiffness
during their semi-seated discrete trunk release task.
136
In a repetitive trunk flexion-extension
task, persons who scored high on the pain catastrophizing scale increased trunk stability under
induced low back pain (injection of hypertonic saline) while those who score low actually had a
more unstable trunk under induced pain.
186
During gait, patients with chronic low back pain and
high pain catastrophizing increased trunk muscle activity in all muscles instrumented – rectus
abdominis, external oblique, erector spinae, and lumbar multifidus – more so than patients with
TRUNK CONTROL DURING DYNAMIC BALANCE 59

low pain catastrophizing.
187
Finally, the influence of pain expectation appears to be region-
specific since noxious stimuli presented to the elbow before a unilateral arm raise did not change
trunk muscle activation, while stimuli presented to the low back resulted in delayed transverse
abdominis and deep lumbar multifidus activation and increased amplitude of internal oblique and
superficial lumbar multifidus activity.
188
There are substantial gaps in the literature about how
psychometric measures of persons with recurrent low back pain in symptom remission interact
with the role of the trunk in balance.
Effects of dual-task interference
Dual-task interference has been used in research and rehabilitation of persons with
recurrent low back pain both because postural dual-tasking is ecological and because there are
specific information processing deficits that these studies aim to probe. Using a visual reaction
time test where responses must be made by clicking a mouse, researchers found that initially
reaction times for the persons with low back pain were longer, but then reduced to match
controls after rehabilitation.
189
The authors pull in information from other cited studies including
some psychometric findings described above to outline a theoretical framework for how pain and
disability influence short-term memory and motor output. This framework has been tested
through investigations of dual-task interference.
Much like research into the effects of psychometrics on trunk control during balance,
research into the effects of dual-task interference in persons with recurrent low back pain is
limited and findings are mixed. Hodges reports that an increased cognitive demand delays
transverse abdominis onset prior to arm flexion.
39
In controls, only the onset of the prime mover
(deltoid) was delayed, not the transverse abdominis, but for persons with recurrent low back pain
in remission, both the primary motor control and the associated anticipatory postural control
TRUNK CONTROL DURING DYNAMIC BALANCE 60

were delayed. This suggests that dysfunction in this patient group involves the automatic postural
control system being influenced by cognitive demands. Other groups have reported findings
from persons with chronic low back pain including delayed trunk muscle activation and reduced
transverse abdominis and internal oblique amplitudes with arm flexion under dual-task
interference,
190
decreased postural sway and thorax and pelvis deviations as well as decreased
trunk stiffness in unstable sitting under dual-task interference,
191
and increased gait variability
under dual-task interference.
192
Contrasting, one study reported less variable trunk motion during
dual-task gait
193
and another reported no differences in the effects of dual-task interference on
various functional balance tests between groups.
194
Notably, authors who provided a robust
analysis of task performance reported that persons with low back pain seemed to prioritize gait
over whatever cognitive task was being used,
193
aligning with theories described previously of
posture-cognition prioritization.
107
The limited research on dual-task interference in persons with
recurrent low back pain shows more postural instability during dual-tasking in unstable standing
balance
195
and decreased trunk coordination variability during a walking turn under dual-task
interference.
115
More research needs to be done to understand how dual-task interference affects
motor control in persons with recurrent low back pain in symptom remission.
Summary
The common themes and conclusions throughout this literature review are highlighted
here. The trunk plays an important role in standing balance and seems to be controlled primarily
with automatic and short-latency reflexes. Deeper trunk muscles act more as stabilizers to the
lumbar spine and superficial muscles seem to stiffen the trunk, but these effects are highly task
dependent. In submaximal, low effort tasks, these interpretations seem to be more robust. Dual-
task interference and various psychometric assessments seem to influence global balance
TRUNK CONTROL DURING DYNAMIC BALANCE 61

measures in a minor way, but trunk control in a more consistent way. Persons with recurrent low
back pain serve a model system in which to study dysfunction of the trunk during balance. This
patient groups serves as a good model because of the high prevalence of the condition and need
for research as well as the specific motor control deficits and biopsychosocial dysfunction that
have been identified.

TRUNK CONTROL DURING DYNAMIC BALANCE 62

References
1. Winter DA. Human balance and posture control during standing and walking. Gait
Posture. 1995;3(4):193-214.
2. Massion J. Postural control system. Curr Opin Neurobiol. 1994;4:877-887.
3. Oddsson L. Control of voluntary trunk movements in man. Mechanisms for postural
equilibrium during standing. Acta Physiol Scand Suppl. 1990;595:1-60.
4. Granacher U, Gollhofer A, Hortobágyi T, Kressig RW, Muehlbauer T. The importance of
trunk muscle strength for balance, functional performance, and fall prevention in seniors:
A systematic review. Sport Med. 2013;43(7):627-641. doi:10.1007/s40279-013-0041-1.
5. Davidson BS, Madigan ML, Nussbaum MA. Effects of lumbar extensor fatigue and
fatigue rate on postural sway. Eur J Appl Physiol. 2004;93(1-2):183-189.
doi:10.1007/s00421-004-1195-1.
6. Parreira RB, Amorim CF, Gil AW, Teixeira DC, Bilodeau M, da Silva RA. Effect of trunk
extensor fatigue on the postural balance of elderly and young adults during unipodal task.
Eur J Appl Physiol. 2013;113(8):1989-1996. doi:10.1007/s00421-013-2627-6.
7. Lin D, Nussbaum MA, Seol H, Singh NB, Madigan ML, Wojcik LA. Acute effects of
localized muscle fatigue on postural control and patterns of recovery during upright
stance: Influence of fatigue location and age. Eur J Appl Physiol. 2009;106(3):425-434.
doi:10.1007/s00421-009-1026-5.
8. Davidson BS, Madigan ML, Nussbaum MA, Wojcik LA. Effects of localized muscle
fatigue on recovery from a postural perturbation without stepping. Gait Posture.
TRUNK CONTROL DURING DYNAMIC BALANCE 63

2009;29(4):552-557. doi:10.1016/j.gaitpost.2008.12.011.
9. Van Der Burg JCE, Pijnappels M, van Dieën JH. Out-of-plane trunk movements and trunk
muscle activity after a trip during walking. Exp Brain Res. 2005;165(3):407-412.
doi:10.1007/s00221-005-2312-z.
10. Ferreira PH, Ferreira ML, Maher CG, Refshauge K, Herbert RD, Hodges PW. Changes in
recruitment of transversus abdominis correlate with disability in people with chronic low
back pain. Br J Sports Med. 2010;44(16):1166-1172. doi:10.1136/bjsm.2009.061515.
11. Cholewicki J, McGill SM. Mechanical stability of the in vivo lumbar spine. Clin Biomech.
1996;11(1):1-15.
12. van Dieën JH, Flor H, Hodges PW. Low-Back Pain Patients Learn to Adapt Motor
Behavior with Adverse Secondary Consequences. Exerc Sport Sci Rev. 2017:223-229.
doi:10.1249/JES.0000000000000121.
13. Hodges PW, Tucker K. Moving differently in pain: A new theory to explain the adaptation
to pain. Pain. 2011;152(SUPPL.3):S90-S98. doi:10.1016/j.pain.2010.10.020.
14. Hodges PW. Pain and motor control: From the laboratory to rehabilitation. J Electromyogr
Kinesiol. 2011;21(2):220-228. doi:10.1016/j.jelekin.2011.01.002.
15. Breen A, Breen A. Uneven intervertebral motion sharing is related to disc degeneration
and is greater in patients with chronic, non-specific low back pain: an in vivo, cross-
sectional cohort comparison of intervertebral dynamics using quantitative fluoroscopy.
Eur Spine J. 2018;27(1):145-153. doi:10.1007/s00586-017-5155-y.
16. Henry SM, Hitt JR, Jones SL, Bunn JY. Decreased limits of stability in response to
TRUNK CONTROL DURING DYNAMIC BALANCE 64

postural perturbations in subjects with low back pain. Clin Biomech. 2006;21:881-892.
doi:10.1016/j.clinbiomech.2006.04.016.
17. Jones SL, Henry SM, Raasch CC, Hitt JR, Bunn JY. Individuals with non-specific low
back pain use a trunk stiffening strategy to maintain upright posture. J Electromyogr
Kinesiol. 2012;22(1):13-20. doi:10.1016/j.jelekin.2011.10.006.
18. van Drunen P, van der Helm FCT, van Dieën JH, Happee R. Trunk stabilization during
sagittal pelvic tilt: From trunk-on-pelvis to trunk-in-space due to vestibular and visual
feedback. J Neurophysiol. 2015:jn.00867.2015. doi:10.1152/jn.00867.2015.
19. Radebold A, Cholewicki J, Panjabi MM, Patel TC. Muscle response pattern to sudden
trunk loading in healthy individuals and in patients with chronic low back pain. Spine
(Phila Pa 1976). 2000;25(8):947-954. doi:10.1097/00007632-200004150-00009.
20. Larivière C, Ludvig D, Kearney R, Mecheri H, Caron J-M, Preuss R. Identification of
intrinsic and reflexive contributions to low-back stiffness: medium-term reliability and
construct validity. J Biomech. 2015;48(2):254-261. doi:10.1016/j.jbiomech.2014.11.036.
21. Pruszynski JA, Kurtzer I, Scott SH. The long-latency reflex is composed of at least two
functionally independent processes. J Neurophysiol. 2011;106(1):449-459.
doi:10.1152/jn.01052.2010.
22. Burke D, Hallett M, Fuhr P, Pierrot-Deseilligny E. H refexes from the tibial and median
nerves. In: Deuschl G, Eisen A, eds. Recommendations for the Practice of Clinical
Neurophysiology: Guidelines of the IFCN. Elsevier Science B.V.; 1999:259-262.
23. Miller EM, Bazrgari B, Nussbaum MA, Madigan ML. Effects of exercise-induced low
TRUNK CONTROL DURING DYNAMIC BALANCE 65

back pain on intrinsic trunk stiffness and paraspinal muscle reflexes. J Biomech.
2013;46(4):801-805. doi:10.1016/j.jbiomech.2012.11.023.
24. Cholewicki J, Silfies SP, Shah RA, et al. Delayed trunk muscle reflex responses increase
the risk of low back injuries. Spine (Phila Pa 1976). 2005;30(23):2614-2620.
doi:10.1097/01.brs.0000188273.27463.bc.
25. Vera-Garcia FJ, Brown SHM, Gray JR, McGill SM. Effects of different levels of torso
coactivation on trunk muscular and kinematic responses to posteriorly applied sudden
loads. Clin Biomech. 2006;21(5):443-455. doi:10.1016/j.clinbiomech.2005.12.006.
26. Talebian S, Mousavi SJ, Olyaei GR, Sanjari MA, Parnianpour M. The effect of exertion
level on activation patterns and variability of trunk muscles during multidirectional
isometric activities in upright posture. Spine (Phila Pa 1976). 2010;35(11):443-451.
doi:10.1097/BRS.0b013e3181bc34b5.
27. Chiang J, Potvin JR. The in vivo dynamic response of the human spine to rapid lateral
bend perturbation: Effects of preload and step input magnitude. Spine (Phila Pa 1976).
2001;26(13):1457-1464. doi:10.1097/00007632-200107010-00016.
28. MacDonald D, Moseley GL, Hodges PW. People With Recurrent Low Back Pain
Respond Differently to Trunk Loading Despite Remission From Symptoms. Spine (Phila
Pa 1976). 2010;35(7):818-824. doi:10.1097/BRS.0b013e3181bc98f1.
29. Macintosh JE, Bogduk N. The biomechanics of the lumbar multifidus. Clin Biomech.
1986;1(4):205-213. doi:10.1016/0268-0033(86)90147-6.
30. Cholewicki J, VanVliet IV JJ. Relative contribution of trunk muscles to the stability of the
TRUNK CONTROL DURING DYNAMIC BALANCE 66

lumbar spine during isometric exertions. Clin Biomech. 2002;17(2):99-105.
doi:10.1016/S0268-0033(01)00118-8.
31. van Dieën JH, Thissen C, van de Ven A, Toussaint H. The electro-mechanical delay of the
erector spinae muscle: influence of rate of force development, fatigue and electrode
location. Eur J Appl Physiol Occup Physiol. 1991;63(3-4):216-222.
32. Burke R, Levine D, Tsairis P, Zajac F. Physiological types and histochemical profiles in
motor units of the cat gastrocnemius. J Physiol. 1973;234:723-748.
33. Bouisset S, Zattara M. Biomechanical study of the programming of anticipatory postural
adjustments associated with voluntary movement. J Biomech. 1987;20(8):735-742.
34. Lee W, Buchanan T, Rogers M. Effects of arm acceleration and behavioral conditions on
the organization of postural adjustments during arm flexion. Exp Brain Res.
1987;66(2):257-270.
35. Mok NW, Brauer SG, Hodges PW. Failure to use movement in postural strategies leads to
increased spinal displacement in low back pain. Spine (Phila Pa 1976).
2007;32(19):E537-43. doi:10.1097/BRS.0b013e31814541a2.
36. Rogers MW. Influence of task dynamics on the organization of interlimb responses
accompanying standing human leg flexion movements. Brain Res. 1992;579(2):353-356.
doi:10.1016/0006-8993(92)90075-K.
37. Hodges P, Richardson C. Inefficient muscular stabilization of the lumbar spine associated
with low back pain A motor control evaluation of transversus abdominis. Spine (Phila Pa
1976). 1996;21(22):2640-2650.
TRUNK CONTROL DURING DYNAMIC BALANCE 67

38. Moseley GL, Hodges PW, Gandevia SC. Deep and superficial fibers of the lumbar
multifidus muscle are differentially active during voluntary arm movements. Spine (Phila
Pa 1976). 2002;27(2):E29-36.
39. Hodges PW. Changes in motor planning of feedforward postural responses of the trunk
muscles in low back pain. Exp Brain Res. 2001;141(2):261-266.
doi:10.1007/s002210100873.
40. Hodges PW, Richardson CA. Altered trunk muscle recruitment in people with low back
pain with upper limb movement at different speeds. Arch Phys Med Rehabil.
1999;80(9):1005-1012. doi:10.1016/S0003-9993(99)90052-7.
41. MacDonald D, Moseley GL, Hodges PW. Why do some patients keep hurting their back?
Evidence of ongoing back muscle dysfunction during remission from recurrent back pain.
Pain. 2009;142(3):183-188. doi:10.1016/j.pain.2008.12.002.
42. Abiko T, Shimamura R, Ogawa D, et al. Difference in the electromyographic onset of the
deep and superficial multifidus during shoulder movement while standing. PLoS One.
2015;10(4):1-11. doi:10.1371/journal.pone.0122303.
43. Hodges P, Cresswell A, Thorstensson A. Perturbed upper limb movements cause short-
latency postural responses in trunk muscles. Exp Brain Res. 2001;138(2):243-250.
doi:10.1007/s002210100693.
44. Morris SL, Lay B, Allison GT. Transversus abdominis is part of a global not local muscle
synergy during arm movement. Hum Mov Sci. 2013;32(5):1176-1185.
doi:10.1016/j.humov.2012.12.011.
TRUNK CONTROL DURING DYNAMIC BALANCE 68

45. Hodges PW, Richardson C. Contraction of the abdominal muscles associated with
movement of the lower limb. Phys Ther. 1997;77(2):132-142.
46. Suehiro T, Mizutani M, Ishida H, Kobara K, Osaka H, Watanabe S. Individuals with
chronic low back pain demonstrate delayed onset of the back muscle activity during prone
hip extension. J Electromyogr Kinesiol. 2015;25(4):675-680.
doi:10.1016/j.jelekin.2015.04.013.
47. Griffioen M, Maaswinkel E, Zuurmond WWA, van Dieën JH, Perez RSGM. Trunk
stabilization estimated using pseudorandom force perturbations, a reliability study. J
Biomech. 2016;49(2):244-251. doi:10.1016/j.jbiomech.2015.12.007.
48. van Drunen P, Maaswinkel E, van der Helm FCT, van Dieën JH, Happee R. Identifying
intrinsic and reflexive contributions to low-back stabilization. J Biomech.
2013;46(8):1440-1446. doi:10.1016/j.jbiomech.2013.03.007.
49. van Dieën JH, van Drunen P, Happee R. Sensory contributions to stabilization of trunk
posture in the sagittal plane. J Biomech. 2018;70:219-227.
doi:10.1016/j.jbiomech.2017.07.016.
50. Ludvig D, Larivière C. Trunk muscle reflexes are elicited by small continuous
perturbations in healthy subjects and patients with low-back pain. J Electromyogr
Kinesiol. 2016;30:111-118. doi:10.1016/j.jelekin.2016.06.007.
51. Oomen NMCW, Reeves NP, Priess MC, van Dieën JH. Trunk muscle coactivation is
tuned to changes in task dynamics to improve responsiveness in a seated balance task. J
Electromyogr Kinesiol. 2015;25(5):765-772. doi:10.1016/j.jelekin.2015.07.001.
TRUNK CONTROL DURING DYNAMIC BALANCE 69

52. Willigenburg NW, Kingma I, van Dieën JH. Center of pressure trajectories, trunk
kinematics and trunk muscle activation during unstable sitting in low back pain patients.
Gait Posture. 2013;38(4):625-630. doi:10.1016/j.gaitpost.2013.02.010.
53. Reeves NP, Cholewicki J, Narendra KS. Effects of reflex delays on postural control
during unstable seated balance. J Biomech. 2009;42(2):164-170.
doi:10.1016/j.jbiomech.2008.10.016.
54. Sorensen CJ, Johnson MB, Callaghan JP, George SZ, Van Dillen LR. Validity of a
Paradigm for Low Back Pain Symptom Development During Prolonged Standing. Clin J
Pain. 2015;31(7):652-659. doi:10.1097/AJP.0000000000000148.
55. Nelson-Wong E, Callaghan JP. Transient Low Back Pain Development During Standing
Predicts Future Clinical Low Back Pain in Previously Asymptomatic Individuals. Spine
(Phila Pa 1976). 2014;39(6):E379-E383. doi:10.1097/BRS.0000000000000191.
56. Sorensen CJ, Norton BJ, Callaghan JP, Hwang CT, Van Dillen LR. Is lumbar lordosis
related to low back pain development during prolonged standing? Man Ther.
2014;20(4):553-557. doi:10.1016/j.math.2015.01.001.
57. Stewart DM, Gregory DE. The use of intermittent trunk flexion to alleviate low back pain
during prolonged standing. J Electromyogr Kinesiol. 2016;27:46-51.
doi:10.1016/j.jelekin.2016.01.007.
58. Mazaheri M, Coenen P, Parnianpour M, Kiers H, van Dieën JH. Low back pain and
postural sway during quiet standing with and without sensory manipulation: A systematic
review. Gait Posture. 2013;37(1):12-22. doi:10.1016/j.gaitpost.2012.06.013.
TRUNK CONTROL DURING DYNAMIC BALANCE 70

59. Schelldorfer S, Ernst MJ, Rast FM, Bauer CM, Meichtry A, Kool J. Low back pain and
postural control, effects of task difficulty on centre of pressure and spinal kinematics. Gait
Posture. 2015;41(1):112-118. doi:10.1016/j.gaitpost.2014.09.004.
60. Ehsani F, Arab AM, Jaberzadeh S, Salavati M. Ultrasound measurement of deep and
superficial abdominal muscles thickness during standing postural tasks in participants with
and without chronic low back pain. Man Ther. 2016;23:98-105.
doi:10.1016/j.math.2016.01.003.
61. da Silva RA, Vieira ER, Carvalho CE, Oliveira MR, Amorim CF, Neto EN. Age-related
differences on low back pain and postural control during one-leg stance: a case–control
study. Eur Spine J. 2016;25(4):1251-1257. doi:10.1007/s00586-015-4255-9.
62. Sung PS, Leininger PM. A kinematic and kinetic analysis of spinal region in subjects with
and without recurrent low back pain during one leg standing. Clin Biomech.
2015;30(7):696-702. doi:10.1016/j.clinbiomech.2015.05.003.
63. Ting LH. Dimensional reduction in sensorimotor systems: a framework for understanding
muscle coordination of posture. Prog Brain Res. 2007;165:299-321. doi:10.1016/S0079-
6123(06)65019-X.
64. Brumagne S, Janssens L, Knapen S, Claeys K, Suuden-Johanson E. Persons with recurrent
low back pain exhibit a rigid postural control strategy. Eur Spine J. 2008;17(9):1177-
1184. doi:10.1007/s00586-008-0709-7.
65. Sung PS, Yoon B, Lee DC. Lumbar spine stability for subjects with and without low back
pain during one-leg standing test. Spine (Phila Pa 1976). 2010;35(16):753-760.
TRUNK CONTROL DURING DYNAMIC BALANCE 71

doi:10.1097/BRS.0b013e3181d53b9c.
66. Ham YW, Kim DM, Baek JY, Lee DC, Sung PS. Kinematic analyses of trunk stability in
one leg standing for individuals with recurrent low back pain. J Electromyogr Kinesiol.
2010;20(6):1134-1140. doi:10.1016/j.jelekin.2010.05.011.
67. Donath L, Kurz E, Roth R, Zahner L, Faude O. Leg and trunk muscle coordination and
postural sway during increasingly difficult standing balance tasks in young and older
adults. Maturitas. 2016;91:60-68. doi:10.1016/j.maturitas.2016.05.010.
68. da Silva RA, Vieira ER, Fernandes KBP, et al. People with chronic low back pain have
poorer balance than controls in challenging tasks. Disabil Rehabil. 2018;40(11):1294-
1300. doi:10.1080/09638288.2017.1294627.
69. Calatayud J, Borreani S, Martin J, Martin F, Flandez J, Colado JC. Core muscle activity in
a series of balance exercises with different stability conditions. Gait Posture.
2015;42(2):186-192. doi:10.1016/j.gaitpost.2015.05.008.
70. Mientjes M, Frank J. Balance in chronic low back pain patients compared to healthy
people under various conditions in upright standing. Clin Biomech. 1999;14(10):710-716.
doi:10.1016/S0268-0033(99)00025-X.
71. Lyle MA, Valero-Cuevas FJ, Gregor RJ, Powers CM. The lower extremity dexterity test
as a measure of lower extremity dynamical capability. J Biomech. 2013;46(5):998-1002.
doi:10.1016/j.jbiomech.2012.11.058.
72. Lyle MA, Valero-Cuevas FJ, Gregor RJ, Powers CM. Lower extremity dexterity is
associated with agility in adolescent soccer athletes. Scand J Med Sci Sports.
TRUNK CONTROL DURING DYNAMIC BALANCE 72

2013;25(1):81-88. doi:10.1111/sms.12162.
73. Lyle MA, Valero-Cuevas FJ, Gregor RJ, Powers CM. Control of dynamic foot-ground
interactions in male and female soccer athletes: females exhibit reduced dexterity and
higher limb stiffness during landing. J Biomech. 2014;47(2):512-517.
doi:10.1016/j.jbiomech.2013.10.038.
74. Krenn O, Werner I, Lawrence E, Valero-Cuevas F. The lower extremity dexterity test
quantifies sensorimotor control for cross-country skiing. Sci Ski. 2015;6:439-445.
75. Lawrence EL, Cesar GM, Bromfield MR, Peterson R, Valero-Cuevas FJ, Sigward SM.
Strength, Multijoint Coordination, and Sensorimotor Processing Are Independent
Contributors to Overall Balance Ability. Biomed Res Int. 2015;2015.
doi:10.1155/2015/561243.
76. Valero-Cuevas FJ, Smaby N, Venkadesan M, Peterson M, Wright T. The strength–
dexterity test as a measure of dynamic pinch performance. J Biomech. 2003;36(2):265-
270. doi:10.1016/S0021-9290(02)00340-8.
77. Johansson RS, Flanagan JR. Coding and use of tactile signals from the fingertips in object
manipulation tasks. Nat Rev Neurosci. 2009;10(5):345-359. doi:10.1038/nrn2621.
78. Crisco JJ, Panjabi MM, Yamamoto I, Oxland TR. Euler stability of the human
ligamentous lumbar spine. Part II: Experiment. Clin Biomech. 1992;7(1):27-32.
doi:10.1016/0268-0033(92)90004-N.
79. Reeves NP, Narendra KS, Cholewicki J. Spine stability: the six blind men and the
elephant. Clin Biomech (Bristol, Avon). 2007;22(3):266-274.
TRUNK CONTROL DURING DYNAMIC BALANCE 73

doi:10.1016/j.clinbiomech.2006.11.011.
80. Macintosh JE, Valencia F, Bogduk N, Munro RR. The morphology of the human lumbar
multifidus. Clin Biomech. 1986;1(4):196-204. doi:10.1016/0268-0033(86)90146-4.
81. Ward SR, Kim CW, Eng CM, et al. Architectural analysis and intraoperative
measurements demonstrate the unique design of the multifidus muscle for lumbar spine
stability. J Bone Joint Surg Am. 2009;91(1):176-185. doi:10.2106/JBJS.G.01311.
82. Rantanen J, Rissanen A, Kalimo H. Lumbar muscle fiber size and type distribution in
normal subjects. Eur Spine J. 1994;3(6):331-335.
83. Lothe LR, Raven TJL, Eken T. Single-motor-unit discharge characteristics in human
lumbar multifidus muscle. J Neurophysiol. 2015:jn.00010.2014.
doi:10.1152/jn.00010.2014.
84. Hodges PW, Eriksson EM, Shirley D, Gandevia SC. Intra-abdominal pressure increases
stiffness of the lumbar spine. J Biomech. 2005;38(9):1873-1880.
doi:10.1016/j.jbiomech.2004.08.016.
85. Tesh KM, Dunn JS, Evans JH. The abdominal muscles and vertebral stability. Spine
(Phila Pa 1976). 1987;12(5):501-508.
86. Hodges P, Holm AK, Holm S, et al. Intervertebral stiffness of the spine is increased by
evoked contraction of transversus abdominis and the diaphragm: in vivo porcine studies.
Spine (Phila Pa 1976). 2003;28(23):2594-2601.
87. Grenier SG, McGill SM. Quantification of Lumbar Stability by Using 2 Different
Abdominal Activation Strategies. Arch Phys Med Rehabil. 2007;88(1):54-62.
TRUNK CONTROL DURING DYNAMIC BALANCE 74

doi:10.1016/j.apmr.2006.10.014.
88. Vera-Garcia FJ, Elvira JLL, Brown SHM, McGill SM. Effects of abdominal stabilization
maneuvers on the control of spine motion and stability against sudden trunk perturbations.
J Electromyogr Kinesiol. 2007;17(5):556-567. doi:10.1016/j.jelekin.2006.07.004.
89. Tayashiki K, Takai Y, Maeo S, Kanehisa H. Intra-abdominal Pressure and Trunk
Muscular Activities during Abdominal Bracing and Hollowing. Int J Sports Med.
2015;37(2):134-143. doi:10.1055/s-0035-1559771.
90. Brown SHM, Potvin JR. Exploring the geometric and mechanical characteristics of the
spine musculature to provide rotational stiffness to two spine joints in the neutral posture.
Hum Mov Sci. 2007;26(1):113-123. doi:10.1016/j.humov.2006.09.006.
91. Christophy M, Faruk Senan NA, Lotz JC, O’Reilly OM. A musculoskeletal model for the
lumbar spine. Biomech Model Mechanobiol. 2012;11(1-2):19-34. doi:10.1007/s10237-
011-0290-6.
92. Vleeming A, Schuenke MD, Danneels L, Willard FH. The functional coupling of the deep
abdominal and paraspinal muscles: The effects of simulated paraspinal muscle contraction
on force transfer to the middle and posterior layer of the thoracolumbar fascia. J Anat.
2014:447-462. doi:10.1111/joa.12227.
93. Richardson CA, Snijders CJ, Hides JA, Damen L, Pas MS, Storm J. The relation between
the transversus abdominis muscles, sacroiliac joint mechanics, and low back pain. Spine
(Phila Pa 1976). 2002;27(4):399-405. doi:10.1097/00007632-200202150-00015.
94. Crisco J, Panjabi M. The intersegmental and multisegmental muscles of the lumbar spine:
TRUNK CONTROL DURING DYNAMIC BALANCE 75

A biomechanical model comparing lateral stabilizing potential. Spine (Phila Pa 1976).
1991;16(7):793-799.
95. Brown SHM, McGill SM. Muscle force-stiffness characteristics influence joint stability:
A spine example. Clin Biomech. 2005;20(9):917-922.
doi:10.1016/j.clinbiomech.2005.06.002.
96. McGill SM. Low back stability: From formal description to issues for performance and
rehabilitation. Exerc Sport Sci Rev. 2001;29(1):26-31. doi:10.1097/00003677-200101000-
00006.
97. Wickens CD. Multiple resources and performance prediction. Theor Issues Ergon Sci.
2002;3(2):159-177. doi:10.1080/14639220210123806.
98. Wickens CD, Liu Y. Codes and modalities in multiple resources: a success and a
qualification. Hum Factors. 1988;30(5):599-616. doi:10.1177/001872088803000505.
99. Azuma R, Prinz W, Koch I. Dual-task slowing and the effects of cross-task compatibility.
Q J Exp Psychol. 2004;57(4):693-713. doi:10.1080/02724980343000459.
100. Diamond A. Executive functions. Annu Rev Psychol. 2013;64:135-168.
doi:10.1146/annurev-psych-113011-143750.
101. Eldreth DA, Patterson MD, Porcelli AJ, Biswal BB, Rebbechi D, Rypma B. Evidence for
multiple manipulation processes in prefrontal cortex. Brain Res. 2006;1123(1):145-156.
doi:10.1016/j.brainres.2006.07.129.
102. Goh H-T, Sullivan KJ, Gordon J, Wulf G, Winstein CJ. Dual-task practice enhances motor
learning: a preliminary investigation. Exp Brain Res. 2012;222(3):201-210.
TRUNK CONTROL DURING DYNAMIC BALANCE 76

doi:10.1007/s00221-012-3206-5.
103. Goh H-T, Gordon J, Sullivan KJ, Winstein CJ. Evaluation of attentional demands during
motor learning: validity of a dual-task probe paradigm. J Mot Behav. 2014;46(August
2014):95-105. doi:10.1080/00222895.2013.868337.
104. Buszard T, Farrow D, Zhu FF, Masters RSW. The relationship between working memory
capacity and cortical activity during performance of a novel motor task. Psychol Sport
Exerc. 2016;22:247-254. doi:10.1016/j.psychsport.2015.07.005.
105. Redick TS, Lindsey DRB. Complex span and n-back measures of working memory: A
meta-analysis. Psychon Bull Rev. 2013;20(6):1102-1113. doi:10.3758/s13423-013-0453-9.
106. De Troyer A, Estenne M, Ninane V, Gansbeke D Van, Gorini M. Transversus abdominis
muscle function in humans. J Appl Physiol. 1990;68(3):1010-1016.
107. Yogev-Seligmann G, Hausdorff JM, Giladi N. Do we always prioritize balance when
walking? Towards an integrated model of task prioritization. Mov Disord.
2012;27(6):765-770. doi:10.1002/mds.24963.
108. Fraizer E, Mitra S. Methodological and interpretive issues in posture-cognition dual-
tasking in upright stance. Gait Posture. 2008;27(2):271-279.
doi:10.1016/j.gaitpost.2007.04.002.
109. Dault MC, Frank JS, Allard F. Influence of a visuo-spatial, verbal and central executive
working memory task on postural control. Gait Posture. 2001;14(2):110-116.
doi:10.1016/S0966-6362(01)00113-8.
110. Zaback M, Carpenter MG, Adkin AL. Threat-induced changes in attention during tests of
TRUNK CONTROL DURING DYNAMIC BALANCE 77

static and anticipatory postural control. Gait Posture. 2016;45:19-24.
doi:10.1016/j.gaitpost.2015.12.033.
111. Uiga L, Capio CM, Ryu D, Wilson MR, Masters RSW. The role of conscious control in
maintaining stable posture. Hum Mov Sci. 2017;(May):0-1.
doi:10.1016/j.humov.2017.10.008.
112. Little CE, Woollacott M. EEG measures reveal dual-task interference in postural
performance in young adults. Exp Brain Res. 2014;233(1):27-37. doi:10.1007/s00221-
014-4111-x.
113. Jones P, Sorinola I, Strutton PH. Effect of dual tasking on postural responses to rapid
lower limb movement while seated on an exercise ball. Gait Posture. 2014;40(2):297-304.
doi:10.1016/j.gaitpost.2014.04.185.
114. Amboni M, Barone P, Hausdorff JM. Cognitive contributions to gait and falls: Evidence
and implications. Mov Disord. 2013;28(11):1520-1533. doi:10.1002/mds.25674.
115. Smith JA, Gordon J, Kulig K. The influence of divided attention on walking turns: Effects
on gait control in young adults with and without a history of low back pain. Gait Posture.
2017;58(September):498-503. doi:10.1016/j.gaitpost.2017.09.019.
116. Asai T, Misu S, Doi T, Yamada M, Ando H. Effects of dual-tasking on control of trunk
movement during gait: Respective effect of manual- and cognitive-task. Gait Posture.
2014;39(1):54-59. doi:10.1016/j.gaitpost.2013.05.025.
117. Masters R, Maxwell J. The theory of reinvestment. Int Rev Sport Exerc Psychol.
2008;1(2):160-183. doi:10.1080/17509840802287218.
TRUNK CONTROL DURING DYNAMIC BALANCE 78

118. Schacter D, Wagner A. Learning and Memory. In: Kandel E, Schwartz J, Jessell T,
Siegelbaum S, Hudspeth A, eds. Principles of Neural Science. 5th ed. New York, NY:
McGraw Hill Medical; 2013:1441-1460.
119. Blakemore C. Mechanics of the Mind. Cambridge, MA: Cambridge University Press;
1977.
120. Beilock SL, Wierenga SA, Carr TH. Expertise, attention, and memory in sensorimotor
skill execution: impact of novel task constraints on dual-task performance and episodic
memory. Q J Exp Psychol. 2002;55(4):1211-1240. doi:10.1080/02724980244000170.
121. Maxwell JP, Masters RS, Eves FF. From novice to no know-how: a longitudinal study of
implicit motor learning. J Sports Sci. 2000;18(October 2012):111-120.
doi:10.1080/026404100365180.
122. Buszard T, Farrow D, Kemp J. Examining the influence of acute instructional approaches
on the decision-making performance of experienced team field sport players. J Sports Sci.
2013;31(3):238-247. doi:10.1080/02640414.2012.731516.
123. Fitts P, Posner M. Human Performance. Belmont, CA: Brooks/Cole; 1967.
124. Logan GD. Toward an instance theory of automatization. Psychol Rev. 1988;95(4):492-
527. doi:10.1037/0033-295X.95.4.492.
125. Zhu FF, Maxwell JP, Hu Y, et al. EEG activity during the verbal-cognitive stage of motor
skill acquisition. Biol Psychol. 2010;84(2):221-227. doi:10.1016/j.biopsycho.2010.01.015.
126. Duval TS, Duval VH, Mulilis JP. Effects of self-focus, discrepancy between self and
standard, and outcome expectancy favorability on the tendency to match self to standard
TRUNK CONTROL DURING DYNAMIC BALANCE 79

or to withdraw. J Pers Soc Psychol. 1992;62(2):340-348. doi:10.1037/0022-
3514.62.2.340.
127. Laborde S, Musculus L, Kalicinski M, Klämpfl MK, Kinrade NP, Lobinger BH.
Reinvestment: Examining convergent, discriminant, and criterion validity using
psychometric and behavioral measures. Pers Individ Dif. 2015;78:77-87.
doi:10.1016/j.paid.2015.01.020.
128. Kal EC, Van Der Kamp J, Houdijk H. External attentional focus enhances movement
automatization: A comprehensive test of the constrained action hypothesis. Hum Mov Sci.
2013;32(4):527-539. doi:10.1016/j.humov.2013.04.001.
129. Ellmers TJ, Machado G, Wong TWL, Zhu F, Williams AM, Young WR. A validation of
neural co-activation as a measure of attentional focus in a postural task. Gait Posture.
2016;50:229-231. doi:10.1016/j.gaitpost.2016.09.001.
130. Fuchs AH. The progression-regression hypotheses in perceptual-motor skill learning. J
Exp Psychol. 1962;63(2):177-182. doi:10.1037/h0041569.
131. MacMahon K, Masters R. From novice to expert and back again: Chunking and
dechunking in motor skills. J Sports Sci. 1999;17:606-607.
132. Wulf G, McNevin N, Shea CH. The automaticity of complex motor skill learning as a
function of attentional focus. Q J Exp Psychol. 2001;54(4):1143-1154.
doi:10.1080/0272498014300011.
133. Huffman JL, Horslen BC, Carpenter MG, Adkin AL. Does increased postural threat lead
to more conscious control of posture? Gait Posture. 2009;30(4):528-532.
TRUNK CONTROL DURING DYNAMIC BALANCE 80

doi:10.1016/j.gaitpost.2009.08.001.
134. Carpenter MG, Frank JS, Silcher CP, Peysar GW. The influence of postural threat on the
control of upright stance. Exp Brain Res. 2001;138(2):210-218.
doi:10.1007/s002210100681.
135. Vance J, Wulf G, Tollner T, McNevin N, Mercer J. EMG activity as a function of the
performer’s focus of attention. J Mot Behav. 2004;36(4):450-459.
136. Karayannis N V, Smeets RJEM, van den Hoorn W, Hodges PW. Fear of Movement Is
Related to Trunk Stiffness in Low Back Pain. PLoS One. 2013;8(6).
doi:10.1371/journal.pone.0067779.
137. Hardy L, R M. Effect of Task-Relevant Cues and State Anxiety on Motor Performance.
Percept Mot Skills. 2001;92(3):943-946. doi:10.2466/PMS.92.3.943-946.
138. Guss-West C, Wulf G. Attentional Focus in Classical Ballet: A survey of professional
dancers. J Danc Med Sci. 2016;20(1):23-29.
139. Orrell AJ, Masters RSW, Eves FF. Reinvestment and Movement Disruption Following
Stroke. 2009:177-183.
140. Masters RSW, Pall HS, MacMahon KMA, Eves FF. Duration of Parkinson disease is
associated with an increased propensity for “reinvestment”. Neurorehabil Neural Repair.
2007;21(2):123-126. doi:10.1177/1545968306290728.
141. Wong WL, Masters RSW, Maxwell JP, Abernethy AB. Reinvestment and falls in
community-dwelling older adults. Neurorehabil Neural Repair. 2008;22(4):410-414.
doi:10.1177/1545968307313510.
TRUNK CONTROL DURING DYNAMIC BALANCE 81

142. Uiga L, Capio CM, Wong TWLL, Wilson MR, Masters RSWW. Movement specific
reinvestment and allocation of attention by older adults during walking. Cogn Process.
2015;16(1):421-424. doi:10.1007/s10339-015-0685-x.
143. Deyo RA, Dworkin SF, Amtmann D, et al. Report of the NIH Task Force on research
standards for chronic low back pain. J Pain. 2014;15(6):569-585.
doi:10.1016/j.jpain.2014.03.005.
144. Rubin DI. Epidemiology and risk factors for spine pain. Neurol Clin. 2007;25(2):353-371.
doi:10.1016/j.ncl.2007.01.004.
145. Hoy D, March L, Brooks P, et al. The global burden of low back pain: Estimates from the
Global Burden of Disease 2010 study. Ann Rheum Dis. 2014;73(6):968-974.
doi:10.1136/annrheumdis-2013-204428.
146. Hoy D, Bain C, Williams G, et al. A systematic review of the global prevalence of low
back pain. Arthritis Rheum. 2012;64(6):2028-2037. doi:10.1002/art.34347.
147. Dieleman JL, Baral R, Birger M, et al. US Spending on Personal Health Care and Public
Health, 1996-2013. Jama. 2016;316(24):2627. doi:10.1001/jama.2016.16885.
148. da Silva T, Mills K, Brown BT, Herbert RD, Maher CG, Hancock MJ. Risk of Recurrence
of Low Back Pain: A Systematic Review. J Orthop Sport Phys Ther. 2017;47(5):305-313.
doi:10.2519/jospt.2017.7415.
149. Norton G, McDonough CM, Cabral HJ, Shwartz M, Burgess JF. Classification of patients
with incident non-specific low back pain: implications for research. Spine J. 2015:1-10.
doi:10.1016/j.spinee.2015.08.015.
TRUNK CONTROL DURING DYNAMIC BALANCE 82

150. Beneck GJ, Kulig K. Multifidus atrophy is localized and bilateral in active persons with
chronic unilateral low back pain. Arch Phys Med Rehabil. 2012;93(2):300-306.
doi:10.1016/j.apmr.2011.09.017.
151. Goubert D, De Pauw R, Meeus M, et al. Lumbar muscle structure and function in chronic
versus recurrent low back pain: a cross-sectional study. Spine J. 2017;17(9):1285-1296.
doi:10.1016/j.spinee.2017.04.025.
152. Beneck GJ, Baker LL, Kulig K. Spectral analysis of EMG using intramuscular electrodes
reveals non-linear fatigability characteristics in persons with chronic low back pain. J
Electromyogr Kinesiol. 2013;23(1):70-77. doi:10.1016/j.jelekin.2012.07.001.
153. Basques BA, Espinoza Orías AA, Shifflett GD, et al. The Kinematics and Spondylosis of
the Lumbar Spine Vary Depending on the Levels of Motion Segments in Individuals With
Low Back Pain. Spine (Phila Pa 1976). 2017;42(13):E767-E774.
doi:10.1097/BRS.0000000000001967.
154. Panjabi MM. The stabilizing system of the spine: Part I. Function, dysfunction,
adaptation, and enhancement. J Spinal Disord. 1992;5(4):383-389.
doi:10.1097/00002517-199212000-00001.
155. Osthoff A-KR, Ernst MJ, Rast FM, et al. Measuring Lumbar Reposition Accuracy in
Patients With Unspecific Low Back Pain: Systematic Review and Meta-analysis. Spine
(Phila Pa 1976). 2015;40(2):E97-E111. doi:10.1097/BRS.0000000000000677.
156. Lee AS, Cholewicki J, Reeves NP, Zazulak BT, Mysliwiec LW. Comparison of trunk
proprioception between patients with low back pain and healthy controls. Arch Phys Med
TRUNK CONTROL DURING DYNAMIC BALANCE 83

Rehabil. 2010;91(9):1327-1331. doi:10.1016/j.apmr.2010.06.004.
157. Ryerson S, Byl NN, Brown DA, Wong RA, Hidler JM. Altered trunk position sense and
its relation to balance functions in people post-stroke. J Neurol Phys Ther. 2008;32(1):14-
20.
158. Jones SL, Hitt JR, DeSarno MJ, Henry SM. Individuals with non-specific low back pain in
an active episode demonstrate temporally altered torque responses and direction-specific
enhanced muscle activity following unexpected balance perturbations. Exp Brain Res.
2012;221(4):413-426. doi:10.1007/s00221-012-3183-8.
159. van Dieën JH, Cholewicki J, Radebold A. Trunk muscle recruitment patterns in patients
with low back pain enhance the stability of the lumbar spine. Spine (Phila Pa 1976).
2003;28(8):834-841. doi:10.1097/01.BRS.0000058939.51147.55.
160. Hodges P, van den Hoorn W, Dawson A, Cholewicki J. Changes in the mechanical
properties of the trunk in low back pain may be associated with recurrence. J Biomech.
2009;42(1):61-66. doi:10.1016/j.jbiomech.2008.10.001.
161. Hodges PW, Moseley GL, Gabrielsson A, Gandevia SC. Experimental muscle pain
changes feedforward postural responses of the trunk muscles. Exp Brain Res.
2003;151(2):262-271. doi:10.1007/s00221-003-1457-x.
162. Hodges P, Richardson C. Delayed postural contraction of transversus abdominis in low
back pain associated with movement of the lower limb. J Spinal Disord. 1998;11(1):46-
56.
163. Radebold A, Cholewicki J, Polzhofer GK, Greene HS. Impaired postural control of the
TRUNK CONTROL DURING DYNAMIC BALANCE 84

lumbar spine is associated with delayed muscle response times in patients with chronic
idiopathic low back pain. Spine (Phila Pa 1976). 2001;26(7):724-730.
doi:10.1097/00007632-200104010-00004.
164. van Dieën JH, Koppes L, Twisk J. Low back pain hisoty and postural sway in unstable
sitting. Spine (Phila Pa 1976). 2010;35(7):812-817.
165. Freddolini M, Strike S, Lee RYW. Stiffness properties of the trunk in people with low
back pain. Hum Mov Sci. 2014;36:70-79. doi:10.1016/j.humov.2014.04.010.
166. Freddolini M, Strike S, Lee RYW. The role of trunk muscles in sitting balance control in
people with low back pain. J Electromyogr Kinesiol. 2014;24(6):947-953.
doi:10.1016/j.jelekin.2014.09.009.
167. Freddolini M, Strike S, Lee R. Dynamic stability of the trunk during unstable sitting in
people with low back pain. Spine (Phila Pa 1976). 2014;39(10):785-790.
doi:10.1097/BRS.0000000000000296.
168. Hooper TL, James CR, Brismée JM, et al. Dynamic balance as measured by the Y-
Balance Test is reduced in individuals with low back pain: A cross-sectional comparative
study. Phys Ther Sport. 2016;22:29-34. doi:10.1016/j.ptsp.2016.04.006.
169. Caffaro RR, França FJR, Burke TN, Magalhães MO, Ramos LAV, Marques AP. Postural
control in individuals with and without non-specific chronic low back pain: A preliminary
case-control study. Eur Spine J. 2014;23(4):807-813. doi:10.1007/s00586-014-3243-9.
170. Laird RA, Gilbert J, Kent P, Keating JL. Comparing lumbo-pelvic kinematics in people
with and without back pain: a systematic review and meta-analysis. BMC Musculoskelet
TRUNK CONTROL DURING DYNAMIC BALANCE 85

Disord. 2014;15(1):229. doi:10.1186/1471-2474-15-229.
171. Hernandez A, Gross K, Gombatto S. Differences in lumbar spine and lower extremity
kinematics during a step down functional task in people with and people without low back
pain. Clin Biomech. 2017;47(May):46-52. doi:10.1016/j.clinbiomech.2017.05.012.
172. van den Hoorn W, Bruijn SM, Meijer OG, Hodges PW, van Dieen JH. Mechanical
coupling between transverse plane pelvis and thorax rotations during gait is higher in
people with low back pain. J Biomech. 2012;45(2):342-347.
doi:10.1016/j.jbiomech.2011.10.024.
173. Ebrahimi S, Kamali F, Razeghi M, Haghpanah SA. Comparison of the trunk-pelvis and
lower extremities sagittal plane inter-segmental coordination and variability during
walking in persons with and without chronic low back pain. Hum Mov Sci. 2017;52:55-66.
doi:10.1016/j.humov.2017.01.004.
174. Crosbie J, de Faria Negrão Filho R, Nascimento DP, Ferreira P. Coordination of Spinal
Motion in the Transverse and Frontal Planes During Walking in People With and Without
Recurrent Low Back Pain. Spine (Phila Pa 1976). 2013;38(5):E286-E292.
doi:10.1097/BRS.0b013e318281de28.
175. Smith JA, Kulig K. Trunk-pelvis coordination during turning: A cross sectional study of
young adults with and without a history of low back pain. Clin Biomech. 2016;36:58-64.
doi:10.1016/j.clinbiomech.2016.05.011.
176. Smith J, Kulig K. Altered multifidus recruitment during walking in young asymptomatic
individuals with a history of low back pain. J Orthop Sports Phys Ther. 2016;27:46-51.
TRUNK CONTROL DURING DYNAMIC BALANCE 86

177. Waddell G, Newton M, Henderson I, Somerville D, Main CJ. A Fear-Avoidance Beliefs
Questionnaire (FABQ) and the role of fear-avoidance beliefs in chronic low back pain and
disability. Pain. 1993;52(2):157-168. doi:10.1016/0304-3959(93)90127-B.
178. George SZ, Fritz JM, Childs JD. Investigation of elevated fear-avoidance beliefs for
patients with low back pain: a secondary analysis involving patients enrolled in physical
therapy clinical trials. J Orthop Sport Phys Ther. 2008;38(2):50-58.
doi:10.2519/jospt.2008.2647.
179. George SZ, Valencia C, Beneciuk JM. A Psychometric Investigation of Fear-Avoidance
Model Measures in Patients With Chronic Low Back Pain. J Orthop Sport Phys Ther.
2010;40(4):197-205. doi:10.2519/jospt.2010.3298.
180. Ishak NA, Zahari Z, Justine M. Performances among Older Persons with Low Back Pain.
2017;2017.
181. George SZ, Fritz JM, Bialosky JE, Donald DA. The effect of a fear-avoidance-based
physical therapy intervention for patients with acute low back pain: results of a
randomized clinical trial. Spine (Phila Pa 1976). 2003;28(23):2551-2560.
doi:10.1097/01.BRS.0000096677.84605.A2.
182. Roelofs J, Peters ML, Fassaert T, Vlaeyen JWS. The role of fear of movement and injury
in selective attentional processing in patients with chronic low back pain: A dot-probe
evaluation. J Pain. 2005;6(5):294-300. doi:10.1016/j.jpain.2004.12.011.
183. Jacobs J V., Roy CL, Hitt JR, Popov RE, Henry SM. Neural mechanisms and functional
correlates of altered postural responses to perturbed standing balance with chronic low
TRUNK CONTROL DURING DYNAMIC BALANCE 87

back pain. Neuroscience. 2016;339:511-524. doi:10.1016/j.neuroscience.2016.10.032.
184. Mazaheri M, Heidari E, Mostamand J, Negahban H, van Dieën JH. Competing Effects of
Pain and Fear of Pain on Postural Control in Low Back Pain? Spine (Phila Pa 1976).
2014;39(25):E1518-E1523. doi:10.1097/BRS.0000000000000605.
185. Sung W, Abraham M, Plastaras C, Silfies SP. Trunk motor control deficits in acute and
subacute low back pain are not associated with pain or fear of movement. Spine J.
2015;15(8):1772-1782. doi:10.1016/j.spinee.2015.04.010.
186. Ross GB, Sheahan PJ, Mahoney B, Gurd BJ, Hodges PW, Graham RB. Pain
catastrophizing moderates changes in spinal control in response to noxiously induced low
back pain. J Biomech. 2017;58:64-70. doi:10.1016/j.jbiomech.2017.04.010.
187. Pakzad M, Fung J, Preuss R. Pain catastrophizing and trunk muscle activation during
walking in patients with chronic low back pain. Gait Posture. 2016;49:73-77.
doi:10.1016/j.gaitpost.2016.06.025.
188. Moseley GL, Nicholas MK, Hodges PW. Does anticipation of back pain predispose to
back trouble? Brain. 2004;127(10):2339-2347. doi:10.1093/brain/awh248.
189. Luoto S, Taimela S, Hurri H, Alaranta H. Mechanisms explaining the association between
low back trouble and deficits in information processing: A controlled study with follow-
up. Spine (Phila Pa 1976). 1999;24(3):255-261.
190. Larivière C, Butler H, Sullivan MJL, Fung J. An exploratory study on the effect of pain
interference and attentional interference on neuromuscular responses during rapid arm
flexion movements. Clin J Pain. 2013;29(3):265-275.
TRUNK CONTROL DURING DYNAMIC BALANCE 88

191. Van Daele U, Hagman F, Truijen S, Vorlat P, Van Gheluwe B, Vaes P. Decrease in
postural sway and trunk stiffness during cognitive dual-task in nonspecific chronic low
back pain patients, performance compared to healthy control subjects. Spine (Phila Pa
1976). 2010;35(5):583-589. doi:10.1097/BRS.0b013e3181b4fe4d.
192. Hamacher D, Hamacher D, Schega L. A cognitive dual task affects gait variability in
patients suffering from chronic low back pain. Exp Brain Res. 2014;232(11):3509-3513.
doi:10.1007/s00221-014-4039-1.
193. Lamoth CJC, Stins JF, Pont M, Kerckhoff F, Beek PJ. Effects of attention on the control
of locomotion in individuals with chronic low back pain. J Neuroeng Rehabil. 2008;5:1-8.
doi:10.1186/1743-0003-5-13.
194. Hemmati L, Rojhani-Shirazi Z, Malek-Hoseini H, Mobaraki I. Evaluation of Static and
Dynamic Balance Tests in Single and Dual Task Conditions in Participants With
Nonspecific Chronic Low Back Pain. J Chiropr Med. 2017;16(3):189-194.
doi:10.1016/j.jcm.2017.06.001.
195. Sherafat S, Salavati M, Takamjani IE, et al. Effect of dual-tasking on dynamic postural
control in individuals with and without nonspecific low back pain. J Manipulative Physiol
Ther. 2014;37(3):170-179. doi:10.1016/j.jmpt.2014.02.003.

TRUNK CONTROL DURING DYNAMIC BALANCE 89

CHAPTER III
CHARACTERIZING THE BALANCE-DEXTERITY TASK
TO INVESTIGATE TRUNK CONTROL DURING DYNAMIC BALANCE
Abstract
The purpose of the study was to characterize the Balance-Dexterity Task as a means to
investigate trunk control during dynamic balance. The task combines aspects of single-limb
balance and the lower-extremity dexterity test by asking participants to stand on one limb while
compressing an unstable spring with the contralateral limb to an individualized target force.
Nineteen control participants completed the study, and performance measures for the demands of
each limb – balance and dexterous force control – as well as kinematic and electromyographic
measures of trunk control were collected. Given five practice trials, participants achieved
compression forces ranging from 100-139N (mean 121.2 ± 12.3N), representing 14.4-23.0% of
body weight (mean 18.7 ± 2.4%), which were then presented as target forces during test trials.
Dexterous force control coefficient of variation and center of pressure (COP) average resultant
velocity were associated such that greater variability in force control was accompanied by
greater COP velocity (R=0.598, p=0.007). Trunk coupling, quantified as the coefficient of
determination (R
2
) of a frontal plane thorax and pelvis angle-angle plot, varied independently of
any measure of balance or dexterous force control. Trunk coupling was positively associated
with the ratio of lumbar multifidus to erector spinae activation amplitude ratio (R=0.517,
p=0.028) where greater trunk coupling was accompanied by greater deep paraspinal muscle
activation relative to more superficial paraspinal activation. The Balance-Dexterity Task is a
continuous, dynamic balance task where bipedal coordination, captured as measures of balance
TRUNK CONTROL DURING DYNAMIC BALANCE 90

and dexterous force control between which variability was associated, and trunk coupling can be
observed and studied.
Introduction
The purpose of this study was to characterize the Balance-Dexterity Task and evaluate its
use in investigating trunk control during dynamic balance. The role of the trunk in balance
dysfunction and rehabilitation has received less attention than contributions from the lower
extremities even though functional trunk control is necessary for any variety of postural control
tasks – experimental or ecological (Massion, 1994; Oddsson, 1990; Winter, 1995). Laboratory
tasks used to investigate coordination of the trunk must move from classic discrete mechanical
perturbations, such as platform perturbations, (Henry et al., 2006) trunk loading, (MacDonald et
al., 2010; Radebold et al., 2000) or discrete voluntary limb movement, (Hodges & Richardson,
1997; Moseley et al., 2002; Suehiro et al., 2015) toward continuous, dynamic postural control.
Though much has been learned through discrete perturbation paradigms, these do not reflect
conditions where posture is challenged in a continuous dynamic way as in most activities of
daily living. In addition, perturbations that involve an external force imparted on the body have
the potential to effect motor control responses through the fear they induce, especially in certain
patient populations. Many authors acknowledge this limitation in the discussion of their results,
(Griffioen et al., 2016; Luoto et al., 1999) and associations exist between trunk control measures
and psychometrics such as fear-avoidance beliefs and kinesiophobia (Karayannis et al., 2013).
Researchers have developed continuous balance tasks that begin to address these needs.
Novel unstable seated balance tasks (Freddolini et al., 2014; Lee & Granata, 2008; Reeves et al.,
2009; Van Daele et al., 2010) and adjustments to standing like on a step with a foam pad (Donath
et al., 2016) or on rockers and foam pads with perturbing forces at the arms (Calatayud et al.,
TRUNK CONTROL DURING DYNAMIC BALANCE 91

2015) have been used to study trunk control. Even here, however, only when the task is
challenging enough – with eyes closed or with lower-extremities unsupported – do differences
appear between controls and high-functioning patient groups, and authors cite fear of pain as a
contributing factor (Van Daele et al., 2010). Along the same lines, double- and single-limb
balance, though they are continuous tasks, make it difficult to identify control differences
between groups, (Penney et al., 2014) likely because high-functioning healthy and clinical
populations have redundant control systems compensating for dysfunction in submaximal tasks.
A sufficiently challenging, continuous balance task is needed to study postural control in high-
functioning healthy persons and patient groups.
The Balance-Dexterity Task was developed by combining single-limb balance with the
lower-extremity dexterity test (LED-test), (Lyle et al., 2013a) and may serve as an ideal task to
observe aspects of submaximal, continuous postural control. The traditional LED-test involves
compression of an unstable spring while semi-seated on a bicycle seat with arms resting on a
support surface and quantifies lower-limb dexterity since the compression force achieved is
associated with performance on the cross-agility test (R
2
=0.63) but not hip extensor strength
(R
2
=0.04), knee extensor strength (R
2
<0.01), or knee flexor strength (R
2
=0.02) (Lyle et al.,
2013a; Lyle et al., 2013b). Adding this dexterous force control demand to the balance demands
of single-limb stance (Lawrence et al., 2015) can be viewed as a lower-extremity bimanual task
of sorts and allows us to study motor control processes involved in successful task execution.
The purpose of this study was to characterize a novel Balance-Dexterity Task (Figure
III.1). A characterization framework started by quantifying and evaluating performance measures
for the demands of each limb – balance and dexterous force control – then continued by
examining relationships between these measures. Next, trunk coordination was quantified and
TRUNK CONTROL DURING DYNAMIC BALANCE 92

associations between trunk coordination and task performance measures were tested. Finally,
factors potentially contributing to trunk coordination were explored including muscle activation
data.
Methods
Participants and instrumentation
Nineteen non-disabled participants with no back or lower-extremity injury or pain in the
last year and no conditions which would affect balance were recruited for the study with
Institutional Review Board approval and informed consent (12 females, 7 males; 23.9±3.3yrs;
169.1±10.4cm; 67.1±10.8kg; BMI 23.3±1.8). Participants were instrumented with a full-body
retroreflective marker set as well as surface electromyography (EMG) of the external oblique
(EO), rectus abdominis (RA), gluteus maximus (GMax), and gluteus medius (GMed) and fine-
wire EMG of the internal oblique (IO), lumbar multifidus (MF), and erector spinae (ES) at the
level of L4 (Noraxon Wireless EMG; Scottsdale, AZ; 3000Hz). Surface EMG were collected
with bipolar silver/silver chloride electrodes with an interelectrode distance of 22mm placed per
guidelines from SENIAM (Hermens et al., 2006), and fine-wire EMG were collected with a pair
of 50μm nickel-chromium alloy wires insulated with nylon with distal 2mm exposed and loaded
into a 25-gauge hypodermic needle and sterilized. Insertions were done under ultrasound
guidance, and protocols were adapted from Perotto (Perotto et al., 2011). All muscles were
instrumented on the side contralateral to the participant’s preferred kicking limb, hereafter
referred to as the stance side. Motion data were captured with an 11-camera Qualisys Oqus
System (Gothenburg, Sweden; 250Hz), and kinetic data were captured with Advanced Medical
Technology Inc. force plates (Watertown, MA; 3000Hz).
Procedures
TRUNK CONTROL DURING DYNAMIC BALANCE 93

Participants completed a 30s trial of double-limb standing (preferred stance width) and
three 30s trials of single-limb standing on the stance side. Participants were introduced to the
Balance-Dexterity Task, which used a custom device made by mounting polyvinyl chloride
(PVC) adaptors to boards with a spring between them (spring characteristics: outside diameter
1.750in, inside diameter 1.336in, free length 12.0in, rate 28.0 lbs/in, wire diameter 0.207in, and
total coils 27.5; Compression Spring #805, Century Spring Corp., Commerce, CA). A similar
instrumented device is available from Neuromuscular Dynamics, LLC (La Crescenta, CA).
Participants were shown real-time feedback of the vertical force under the spring, and instructed:
“While standing on one leg, compress this spring so that the line is first as high, then as stable as
possible” (Figure III.1). Each trial lasted 20-25s. After one familiarization trial and five practice
trials, the mean of the middle 50% of the last three practice trials were used to calculate an
individual’s “reproducible, submaximal compression”. This value was called reproducible,
submaximal compression because (1) the goal of the Balance-Dexterity Task, in contrast to the
LED-test’s goal of measuring maximum dexterous control ability, is to use submaximal
dexterous force control to perturb balance; (2) for the LED-test, at least 20-25 attempts are
required to produce a stable maximum indicating the compression described here is not maximal
(Lyle et al., 2013); (3) in pilot testing it was found that, without the seat and arm rests, giving
subjects more than five practice trials to achieve a stable maximum led to creative but
confounding strategies sometimes including a deep squat with the stance leg or wedging the
spring into a contorted shape; and (4) a subset of five participants who repeated testing
reproduced similar compression values between days with an intraclass correlation coefficient
(ICC) of 0.875. After practice, participants used a visual analog scale (VAS) to report how
TRUNK CONTROL DURING DYNAMIC BALANCE 94

difficult the task was, how confident they were they could complete the task successfully, and
how much attention the task required.
Participants then completed five trials where a dotted line indicating their reproducible,
submaximal compression was shown as a target with the instructions: “While standing on one
leg, compress this spring so that the line is as stable as possible directly over the dotted goal
line.” Three trials were interspersed where the spring was replaced with a stable block of the
same height, and the same goal line instructions were given. Five participants were brought in
for re-testing after the initial collection to assess test-retest reliability of outcomes measures.
Results are reported as ICCs for absolute agreement.
Data analysis
All trials were trimmed so the middle 50% of the task was analyzed. Kinematic and force
plate data were low-pass filtered with cutoff frequencies of 12Hz and 50Hz, respectively.
Surface and fine-wire EMG data were band-pass filtered between 20 and 500 Hz and 20 and
1000 Hz, respectively, all using a dual-pass 4
th
order Butterworth filter. EMG data were rectified
and smoothed with a moving weighted average window of 500ms. Signals were normalized both
to maximum voluntary isometric contractions (MVICs) and to averaged signal amplitude during
the stable block trials. Here, EMG amplitude was reported as a percentage of activation during
the stable block condition, allowing activation to be interpreted as a response to adding dexterous
force control, not to body position or vertical force production. Note one subject was excluded
when reporting MVIC-normalized EMG because of corrupted MVIC data.
Balance demands were quantified with center of pressure (COP) measures from the
stance limb including average COP resultant velocity and COP area measured with a 95%
confidence ellipse, as well as average center of mass (COM) resultant velocity (Figure
TRUNK CONTROL DURING DYNAMIC BALANCE 95

III.1,Equation A, Equation B). These measures had excellent reliability between-days
(ICC≥0.927). Dexterous force control was measured using the vertical force produced under the
spring and quantified as root-mean-squared error (RMSE) from the reproducible, submaximal
compression goal line; coefficient of variation (CV); and median frequency (MDF) of the
detrended force (Figure III.1). These measures had excellent reliability between-days
(ICC≥0.911). Muscle activation data were averaged to acquire a mean activation amplitude for
each trial. Muscle activation ratios were calculated in a frame-by-frame manner including deep-
to-superficial ratios for the paraspinals (MF:ES) and abdominals (IO:EO) and co-contraction
ratios for the deep trunk muscles (MF, IO) and superficial trunk muscles (ES, EO) (Equation C).
Equation A. Average resultant center of pressure (COP) velocity.

Equation B. Average resultant center of mass (COM) velocity.

Equation C. Trunk muscle activation ratio.
For deep-to-superficial ratios, Muscle 1 was multifidus (MF) for paraspinals and
internal oblique (IO) for abdominals; and for co-contraction ratios, the muscle
whose mean amplitude was lower was Muscle 1.

TRUNK CONTROL DURING DYNAMIC BALANCE 96

Trunk control was quantified by tracking thorax and pelvis motion relative to global
coordinates. Using an angle-angle plot of thorax and pelvis frontal plane rotation, a coefficient of
determination (R
2
) was calculated where a high R
2
would indicate highly coupled thorax and
pelvis motion and a low R
2
would indicate more dissociated or independent motion of the thorax
and pelvis. This metric has been used to distinguish participants with and without low back pain
through frontal and transverse plane trunk coupling during gait (Crosbie et al., 2013; van den
Hoorn et al., 2012) Also from this angle-angle plot, instantaneous coupling angles were
calculated and the percentages of time spent in in-phase coupling and in anti-phase coupling
were extracted, all per equations presented in Needham (Needham et al., 2015). In addition,
more traditional range-of-motion metrics were acquired including trunk (thorax relative to
pelvis), thorax, and pelvis angular excursions. These measures had moderate to excellent
reliability between days (ICC≥0.684). Given the characterization nature of this study,
relationships among variables were explored using descriptive statistics, bivariate Pearson
correlations, and repeated measures analysis of variance (ANOVA) with Bonferroni-corrected
post-hoc tests where appropriate with α=0.05 (PASW Statistics, IBM Corp., Armonk, NY).
TRUNK CONTROL DURING DYNAMIC BALANCE 97

Figure III.1. The Balance-Dexterity Task with representative data showing
balance outcome measures including center of pressure (COP) measures
(right) and dexterous vertical force (vForce) control outcome measures
including root-mean-squared error (RMSE), coefficient of variation (CV)
and median frequency (MDF) (left).
Results
Task performance
Given five practice trials, all participants achieved submaximal, but sufficiently
perturbing dexterous control of the spring with compression forces ranging from 100-139N
(mean 121.2±12.3N), representing 14.4-23.0% of body weight (mean 18.7±2.4%) (Figure
III.2A). A handful of participants stumbled in the first trial, before tactile information about
spring stiffness was acquired, but after practice all participants were able to complete the task
safely and successfully.
Measures of performance of the two primary demands in the Balance-Dexterity Task –
balance and dexterous force control – were evaluated to select the most sensitive measure(s) to
TRUNK CONTROL DURING DYNAMIC BALANCE 98

the increased postural demands of the task. Dexterous force control was quantified using RMSE
(6.54±2.5N), CV (3.60±1.2%), and MDF (2.42±0.5Hz), and none of these measures were
associated with each other (0.346≥p≥0.129) (Figure III.2B-D). Comparing these measures
between the Balance-Dexterity Task and the stable block condition, where force control was
required but without instability of a spring, using a repeated-measures ANOVA revealed a main
effect of task for all three measures – RMSE (F(3,18)=16.033, p=0.001), CV (F(3,18)=8.320,
p=0.010), and MDF (F(3,18)=87.004, p<0.001). RMSE and CV were taken forward as the
primary measures of dexterous force control demands.

Figure III.2. (A) Submaximal, reproducible compression forces acquired
from five practice trials and presented as a target during stable block and
Balance-Dexterity Task trials. Measures of dexterous force control in the
stable block and Balance-Dexterity Task conditions including root-mean-
squared-error (B), coefficient of variation (C), and median frequency (D).
*p<0.05.
Measures of balance demands were all associated (COM velocity and COP velocity:
R=0.457, p=0.057; COM velocity and COP area: R=0.505, p=0.033; COP velocity and COP
area: R=0.476, p=0.040). Using a repeated-measures ANOVA, a main effect of task was found
for all three measures – COM velocity (F(3,15)=24.5, p<0.001), COP velocity (F(3,16)=116.6,
TRUNK CONTROL DURING DYNAMIC BALANCE 99

p<0.001), and COP area (F(3,16)=33.4, p<0.001). Only for COP velocity, however, were all
conditions significantly different from one another (Figure III.3). Given that these measures
capture similar outcomes and COP velocity was the most sensitive to different conditions, COP
velocity was taken forward as the primary measure of balance demands.

Figure III.3. Center of pressure (COP) average resultant velocity in four
conditions.
The relationships between these two demands were investigated by testing associations
between performance variables. Greater dexterous force CV was associated with greater COP
velocity (R=0.598, p=0.007) (Figure III.4A). In addition, those who reported greater perceived
task difficulty rated on a VAS also exhibited greater RMSE relative to the target compression
force (R=0.539, p=0.017) (Figure III.4B).
TRUNK CONTROL DURING DYNAMIC BALANCE 100

Figure III.4. (A) Association between coefficient of variation (CV) of
dexterous force control and center of pressure (COP) average resultant
velocity. (B) Association between participants’ self-reported assessment of
task difficulty on a visual analog scale (VAS) and root-mean-squared-error
(RMSE) of dexterous force relative to the reproducible, submaximal
compression goal line.
Trunk coordination

Figure III.5. Representative examples of high (left) and low (right) frontal
plane trunk coupling quantified with a coefficient of determination for the
thorax and pelvis angle-angle plot (R
2
).
TRUNK CONTROL DURING DYNAMIC BALANCE 101

The primary outcome measure for trunk kinematic coordination – trunk coupling R
2

(Figure III.5) – was compared to traditional segment excursion measures and measures derived
from thorax and pelvis coupling angle. Trunk coupling R
2
was not associated with thorax
(R=0.230, p=0.354), pelvis (R=0.330, p=0.174), or trunk (R=0.190, p=0.440) excursions (Figure
III.6A-C). This measure was correlated with the percent of time spent in in-phase coupling
(R=0.850, p<0.001) and was negatively associated with the percent of time spent in anti-phase
coupling (R=-0.800, p<0.001) (Figure III.6D). A wider range of trunk coupling R
2
values were
observed during the Balance-Dexterity Task compared to single-limb stance (Figure III.6E).
Trunk coupling R
2
was not associated with any of the measures of task demands described above
(0.872 ≥ p ≥ 0.279). Taken together, trunk coupling R
2
captures something different than simple
measures of excursion, but similar to a phase analysis of segment coupling angle. Also, trunk
coupling is modulated independent of task performance.
TRUNK CONTROL DURING DYNAMIC BALANCE 102

Figure III.6. Associations between trunk coupling quantified as the
coefficient of determination of a thorax-pelvis frontal plane angle-angle plot
(R
2
) and (A-C) frontal plane segment and joint excursions and (D) the
percent of time spent in in-phase and in anti-phase coupling. (E) Trunk
coupling R
2
during single-limb stance and the Balance-Dexterity Task.
Muscle activation levels during the Balance-Dexterity normalized to the stable block
condition were summarized (Figure III.7). The mean activation levels for all muscles were
greater than 100% with some participants increasing activation of muscles up to five-times
(500%) the stable block condition. There were no significant associations between trunk
coupling R
2
and any of the individual muscle activations or muscle activation ratios normalized
to the stable block condition (0.992 ≥ p ≥ 0.217). When normalized to MVICs, trunk coupling R
2
was positively associated with MF:ES ratio (R=0.517, p=0.028) and negatively associated with
GMed activation (R=-0.503, p=0.033). When all variables were tested in a stepwise multiple
linear regression to predict trunk coupling only the MF:ES ratio entered the model (β=0.333,
TRUNK CONTROL DURING DYNAMIC BALANCE 103

p=0.028) (Figure III.8). Greater trunk coupling was accompanied by greater deep paraspinal
muscle activation relative to more superficial paraspinal activation.

Figure III.7. (A) Muscle activation levels normalized to the stable block
condition. MF: multifidus, ES: erector spinae, IO: internal oblique, EO:
external oblique, RA: rectus abdominis, GMax: gluteus maximus, and
GMed: gluteus medius. (B) Deep-to-superficial muscle activation ratios. (C)
Co-contraction ratios for deep muscles (MF and IO) and superficial muscles
(ES and EO).
TRUNK CONTROL DURING DYNAMIC BALANCE 104

Figure III.8. Associations between trunk coupling quantified as the
coefficient of determination of a thorax-pelvis frontal plane angle-angle plot
(R
2
) and the ratio of multifidus to erector spinae activation (MF:ES)
normalized to maximum isometric voluntary contractions, shown with (solid
line) and without (dotted line) a potential outlier.
Discussion
The Balance-Dexterity Task is a submaximal, continuous balance task useful for clinical
and basic research in human postural control. The task invokes larger COP velocities than
double- and single-limb stance and more variable dexterous force control compared to a stable
block condition. The large COP velocity utilized in the task was likely necessary to control COM
motion, (Winter, 1995) as we saw that COP velocity was higher in the Balance-Dexterity Task
than in single-limb standing, but COM velocity was not different between these conditions. A
TRUNK CONTROL DURING DYNAMIC BALANCE 105

novel finding here is that variability in dexterous force control was associated with variability in
balance control, indicated by a positive correlation between dexterous force CV and stance limb
COP velocity. This finding in a concurrent bipedal lower-extremity task aligns with similar work
in bimanual control tasks, which show that spatiotemporal measures of movement control,
including reaction time, movement time, and error variability, are strongly coordinated between
the two hands in discrete tapping and reaching tasks (Marteniuk et al., 1984). Additionally, in
continuous tasks such as bimanual force tracking, the degree of variability has been shown to
increase with increased force demands (Kennedy et al., 2016; Lodha et al., 2012). While
analogous in some ways, a caveat in drawing this parallel between bipedal and bimanual control
is that asymmetries in bimanual control are known to be driven by the motor dominant hand
(Marteniuk et al., 1984). The present study did not test both sides in the Balance-Dexterity Task,
but future work could evaluate the possible effect of lower extremity dominance in this bipedal
control task. In addition, dexterous force RMSE was associated with participants’ perceived task
difficulty. RMSE is most directly related to the visual error signal provided to the participants as
they view vertical force in reference to their reproducible, submaximal target line, which could
explain its association with a self-report measure of task difficulty. In contrast, COP measures
were associated with CV, not RMSE, because CV more accurately quantifies force variability
and is not sensitive to an offset from the goal line.
Frontal plane trunk coupling was altered independently of task performance. Frontal
plane motion was analyzed for two reasons – there was minimal sagittal plane motion during the
task, and frontal plane motion is frequently the target of investigation and intervention in single-
limb balance tasks, as frontal plane hip control has received much attention in its contributions to
single-limb balance (Winter, 1995; Winter et al., 1996), and frontal plane control is most
TRUNK CONTROL DURING DYNAMIC BALANCE 106

perturbed in dual-tasking (Asai et al., 2013). Trunk coupling R
2
was correlated with the percent
of time in in-phase coupling and oppositely correlated with the percent of time in anti-phase
coupling, justifying use of this simpler measure to capture trunk coordination. Segment and joint
excursions, however, seemed to quantify different aspects of trunk movement. A wider range of
trunk coupling values were observed during the Balance-Dexterity Task compared to single-limb
stance, and trunk coupling was not associated with any of the balance or dexterous force control
demand measures. Added demands of dexterous force control allowed a wide range of trunk
coupling strategies, modulated independently of task performance, to emerge. Trunk
coordination, therefore, may prove to be a useful target in this task for observing or modifying
movement patterns in various patient, aging, or athletic populations.
Muscle activation levels were greater in the Balance-Dexterity Task compared to the
stable block condition. No muscle activations alone or in ratios were associated with trunk
coupling R
2
. The paucity of associations between individual muscle activations and trunk
coupling indicates that these non-disabled persons have redundant motor control processes
available to control balance, dexterous force control, and trunk coupling. It is hypothesized that
stronger associations would be present in patient populations with dysfunctional systems or in
non-disabled controls when one or more sensory inputs are knocked out or perturbed. When
signals were normalized to MVICs, however, a greater MF:ES ratio was associated with higher
trunk coupling R
2
. Careful readers will have noticed the significant association between MF:ES
ratio and trunk coupling R
2
seems contingent on one participant with the highest ratio. Without
this participant, the association retains marginal statistical significance (R=0.428, p=0.086).
Given that no errors or abnormalities were found in this participant’s trials, marginal significance
TRUNK CONTROL DURING DYNAMIC BALANCE 107

remains without the participant, and the known variability in EMG-based measurements the
authors report this as a real finding.
This relationship between trunk coupling and coordination of deeper and more superficial
paraspinal musculature, when combined with anatomical and functional evidence about these
muscle groups, forms a foundation for many testable hypotheses. Deeper paraspinals, like the
deep lumbar MF instrumented here with fine-wire EMG, have the capacity to limit intervertebral
motion since each crosses few spine segments (Ward et al., 2009) with small moment arms
(Macintosh et al., 1986; Macintosh and Bogduk, 1986), and they function as stabilizers by
preactivating prior to voluntary limb movements in any direction (Hodges and Richardson, 1996;
Moseley et al., 2002). More superficial musculature like the ES attach further from vertebral axis
of rotations (Macintosh et al., 1986), activate in a direction-dependent manner in response to
mechanical perturbations (Hodges and Richardson, 1996; Moseley et al., 2002), and have the
potential both to stabilize the spine when activated symmetrically and to induce trunk rotation.
MacDonald et. al. summarized the evidence for this conceptual differentiation (2006). The
present findings support a hypothesis that greater MF relative to ES activation stabilizes the
lumbar spine, resulting in greater coupling between thorax and pelvis segments. In the context of
this task, which is both asymmetrical and induces very low trunk excursions (< 5°), this leap
from intervertebral stability to thorax and pelvis coupling is feasible. In fact, a seminal modeling
study by Cholewicki and McGill reported at least 1-3% MVC activation of MF and ES was
required to achieve sufficient lumbar spine stability in a neutral, standing posture (Cholewicki
and McGill, 1996). Our findings support this model’s predictions and add that the relative
activation between these muscle groups is important. Future work should more rigorously test
how MF:ES governs trunk coupling through lumbar spine stabilization using a wider variety of
TRUNK CONTROL DURING DYNAMIC BALANCE 108

tasks and analysis techniques. A population with known impairment in deep trunk muscle
activation and with reduced intervertebral stability like persons with recurrent low back pain
should be tested, hypothesizing lower trunk coupling and MF:ES ratios. If supported by more
research, the Balance-Dexterity Task could serve as a means to characterize coordination of
deep-to-superficial paraspinal musculature by observing trunk coupling.
In addition to testing these proposed hypothesis, future work utilizing the Balance-
Dexterity Task can address limitations of the current study. This proof-of-concept and task
characterization study was conducted on a convenience sample of nineteen healthy young adults.
Future studies should utilize larger sample sizes to present more robust normative values as well
as study more diverse populations including athletes, older adults, and patient groups. The
participants in the current study had no problems completing the task, and none fell during the
duration of the study. Studying populations with balance impairments may require some
adjustments like adding a light touch cue, handlebars, or harness. The authors believe, however,
that since an individual’s reproducible, submaximal compression force is presented as a target,
the task is individualized enough to allow anyone to complete the task safely. The focus on trunk
control in the present study was motivated by our lab’s research into trunk control impairments
and persons suffering from low back pain. The Balance-Dexterity Task is also well-suited for an
investigation into hip- and ankle-strategy approaches to balance. Adding EMG signals from
muscles around these joints and utilizing more intensive inverse dynamics analyses will answer
questions about which strategies are used in different conditions and populations.
In conclusion, the Balance-Dexterity Task is a challenging, continuous dynamic balance
task from which many fruitful measures of movement and motor control processes may be
observed. These include measures of balance and dexterous force control, between which
TRUNK CONTROL DURING DYNAMIC BALANCE 109

variability was positively associated, and trunk coupling, which was modulated independent of
task performance. Trunk coupling was positively associated with the MF:ES paraspinal ratio,
indicating that greater deep paraspinal activation relative to more superficial paraspinals
increases trunk coupling in this task.

TRUNK CONTROL DURING DYNAMIC BALANCE 110

References
Asai, T., Doi, T., Hirata, S., Ando, H., 2013. Dual tasking affects lateral trunk control in healthy
younger and older adults. Gait Posture 38, 830–6.
https://doi.org/10.1016/j.gaitpost.2013.04.005
Calatayud, J., Borreani, S., Martin, J., Martin, F., Flandez, J., Colado, J.C., 2015. Core muscle
activity in a series of balance exercises with different stability conditions. Gait Posture 42,
186–192. https://doi.org/10.1016/j.gaitpost.2015.05.008
Cholewicki, J., McGill, S.M., 1996. Mechanical stability of the in vivo lumbar spine. Clin.
Biomech. 11, 1–15.
Crosbie, J., de Faria Negrão Filho, R., Nascimento, D.P., Ferreira, P., 2013. Coordination of
Spinal Motion in the Transverse and Frontal Planes During Walking in People With and
Without Recurrent Low Back Pain. Spine (Phila. Pa. 1976). 38, E286–E292.
https://doi.org/10.1097/BRS.0b013e318281de28
Donath, L., Kurz, E., Roth, R., Zahner, L., Faude, O., 2016. Leg and trunk muscle coordination
and postural sway during increasingly difficult standing balance tasks in young and older
adults. Maturitas 91, 60–68. https://doi.org/10.1016/j.maturitas.2016.05.010
Freddolini, M., Strike, S., Lee, R., 2014. Dynamic stability of the trunk during unstable sitting in
people with low back pain. Spine (Phila. Pa. 1976). 39, 785–90.
https://doi.org/10.1097/BRS.0000000000000296
Griffioen, M., Maaswinkel, E., Zuurmond, W.W.A., van Dieën, J.H., Perez, R.S.G.M., 2016.
Trunk stabilization estimated using pseudorandom force perturbations, a reliability study. J.
Biomech. 49, 244–251. https://doi.org/10.1016/j.jbiomech.2015.12.007
TRUNK CONTROL DURING DYNAMIC BALANCE 111

Henry, S.M., Hitt, J.R., Jones, S.L., Bunn, J.Y., 2006. Decreased limits of stability in response to
postural perturbations in subjects with low back pain. Clin. Biomech. 21, 881–892.
https://doi.org/10.1016/j.clinbiomech.2006.04.016
Hermens, H., Freriks, B., Merletti, R., Rau, G., Disselhorst-Klug, C., Stegeman, D., Hagg, G.,
2006. The Surface ElectroMyoGraphy for the Non-Invasive Assessment of Muscles Project
(SENIAM) [WWW Document]. URL seniam.org
Hodges, P., Richardson, C., 1996. Inefficient muscular stabilization of the lumbar spine
associated with low back pain A motor control evaluation of transversus abdominis. Spine
(Phila. Pa. 1976). 21, 2640–50.
Hodges, P.W., Richardson, C., 1997. Contraction of the abdominal muscles associated with
movement of the lower limb. Phys. Ther. 77, 132–142.
Karayannis, N. V, Smeets, R.J.E.M., van den Hoorn, W., Hodges, P.W., 2013. Fear of
Movement Is Related to Trunk Stiffness in Low Back Pain. PLoS One 8.
https://doi.org/10.1371/journal.pone.0067779
Kennedy, D.M., Boyle, J.B., Wang, C., Shea, C.H., 2016. Bimanual force control: cooperation
and interference? Psychol. Res. 80, 34–54. https://doi.org/10.1007/s00426-014-0637-6
Lawrence, E.L., Cesar, G.M., Bromfield, M.R., Peterson, R., Valero-Cuevas, F.J., Sigward,
S.M., 2015. Strength, Multijoint Coordination, and Sensorimotor Processing Are
Independent Contributors to Overall Balance Ability. Biomed Res. Int. 2015.
https://doi.org/10.1155/2015/561243
Lee, H., Granata, K.P., 2008. Process stationarity and reliability of trunk postural stability. Clin.
TRUNK CONTROL DURING DYNAMIC BALANCE 112

Biomech. (Bristol, Avon) 23, 735–42. https://doi.org/10.1016/j.clinbiomech.2008.01.008
Lodha, N., Coombes, S.A., Cauraugh, J.H., 2012. Bimanual isometric force control: Asymmetry
and coordination evidence post stroke. Clin. Neurophysiol. 123, 787–795.
https://doi.org/10.1016/j.clinph.2011.08.014
Luoto, S., Taimela, S., Hurri, H., Alaranta, H., 1999. Mechanisms explaining the association
between low back trouble and deficits in information processing: A controlled study with
follow-up. Spine (Phila. Pa. 1976). 24, 255–261.
Lyle, M.A., Valero-Cuevas, F.J., Gregor, R.J., Powers, C.M., 2013. The lower extremity
dexterity test as a measure of lower extremity dynamical capability. J. Biomech. 46, 998–
1002. https://doi.org/10.1016/j.jbiomech.2012.11.058
Lyle, M.A., Valero-Cuevas, F.J., Gregor, R.J., Powers, C.M., 2013. Lower extremity dexterity is
associated with agility in adolescent soccer athletes. Scand. J. Med. Sci. Sports 25, 81–8.
https://doi.org/10.1111/sms.12162
MacDonald, D.A., Lorimer Moseley, G., Hodges, P.W., 2006. The lumbar multifidus: Does the
evidence support clinical beliefs? Man. Ther. 11, 254–263.
https://doi.org/10.1016/j.math.2006.02.004
MacDonald, D., Moseley, G.L., Hodges, P.W., 2010. People With Recurrent Low Back Pain
Respond Differently to Trunk Loading Despite Remission From Symptoms. Spine (Phila.
Pa. 1976). 35, 818–824. https://doi.org/10.1097/BRS.0b013e3181bc98f1
Macintosh, J.E., Bogduk, N., 1986. The biomechanics of the lumbar multifidus. Clin. Biomech.
1, 205–213. https://doi.org/10.1016/0268-0033(86)90147-6
TRUNK CONTROL DURING DYNAMIC BALANCE 113

Macintosh, J.E., Valencia, F., Bogduk, N., Munro, R.R., 1986. The morphology of the human
lumbar multifidus. Clin. Biomech. 1, 196–204. https://doi.org/10.1016/0268-
0033(86)90146-4
Marteniuk, R.G., MacKenzie, C.L., Baba, D.M., 1984. Bimanual movement control: Information
processing and interaction effects. Q. J. Exp. Psychol. 36, 355–265.
Massion, J., 1994. Postural control system. Curr. Opin. Neurobiol. 4, 877–887.
Moseley, G.L., Hodges, P.W., Gandevia, S.C., 2002. Deep and superficial fibers of the lumbar
multifidus muscle are differentially active during voluntary arm movements. Spine (Phila.
Pa. 1976). 27, E29-36.
Needham, R.A., Naemi, R., Chockalingam, N., 2015. A new coordination pattern classification
to assess gait kinematics when utilising a modified vector coding technique. J. Biomech. 48,
3506–3511. https://doi.org/10.1016/j.jbiomech.2015.07.023
Oddsson, L., 1990. Control of voluntary trunk movements in man. Mechanisms for postural
equilibrium during standing. Acta Physiol. Scand. Suppl. 595, 1–60.
Penney, T., Ploughman, M., Austin, M.W., Behm, D.G., Byrne, J.M., 2014. Determining the
activation of gluteus medius and the validity of the single leg stance test in chronic,
nonspecific low back pain. Arch. Phys. Med. Rehabil. 95, 1969–1976.
https://doi.org/10.1016/j.apmr.2014.06.009
Perotto, A.O., Delagi, E.F., Iazzetti, J., Morrison, D., 2011. Anatomical Guide for the
Electromyographer, 5th ed. Charles C Thomas Publisher, LTD., Springfield, Illinois.
Radebold, A., Cholewicki, J., Panjabi, M.M., Patel, T.C., 2000. Muscle response pattern to
TRUNK CONTROL DURING DYNAMIC BALANCE 114

sudden trunk loading in healthy individuals and in patients with chronic low back pain.
Spine (Phila. Pa. 1976). 25, 947–954. https://doi.org/10.1097/00007632-200004150-00009
Reeves, N.P., Cholewicki, J., Narendra, K.S., 2009. Effects of reflex delays on postural control
during unstable seated balance. J. Biomech. 42, 164–70.
https://doi.org/10.1016/j.jbiomech.2008.10.016
Suehiro, T., Mizutani, M., Ishida, H., Kobara, K., Osaka, H., Watanabe, S., 2015. Individuals
with chronic low back pain demonstrate delayed onset of the back muscle activity during
prone hip extension. J. Electromyogr. Kinesiol. 25, 675–680.
https://doi.org/10.1016/j.jelekin.2015.04.013
Van Daele, U., Hagman, F., Truijen, S., Vorlat, P., Van Gheluwe, B., Vaes, P., 2010. Decrease in
postural sway and trunk stiffness during cognitive dual-task in nonspecific chronic low back
pain patients, performance compared to healthy control subjects. Spine (Phila. Pa. 1976).
35, 583–9. https://doi.org/10.1097/BRS.0b013e3181b4fe4d
van den Hoorn, W., Bruijn, S.M., Meijer, O.G., Hodges, P.W., van Dieen, J.H., 2012.
Mechanical coupling between transverse plane pelvis and thorax rotations during gait is
higher in people with low back pain. J. Biomech. 45, 342–347.
https://doi.org/10.1016/j.jbiomech.2011.10.024
Ward, S.R., Kim, C.W., Eng, C.M., Gottschalk, L.J., Tomiya, A., Garfin, S.R., Lieber, R.L.,
2009. Architectural analysis and intraoperative measurements demonstrate the unique
design of the multifidus muscle for lumbar spine stability. J. Bone Joint Surg. Am. 91, 176–
185. https://doi.org/10.2106/JBJS.G.01311
Winter, D.A., 1995. Human balance and posture control during standing and walking. Gait
TRUNK CONTROL DURING DYNAMIC BALANCE 115

Posture 3, 193–214.
Winter, D.A., Prince, F., Frank, J.S., Powell, C., Zabjek, K.F., 1996. Unified theory regarding
A/P and M/L balance in quiet stance. J. Neurophysiol. 75, 2334–2343.

TRUNK CONTROL DURING DYNAMIC BALANCE 116

CHAPTER IV
TRUNK COUPLING IN PERSONS WITH RECURRENT LOW BACK PAIN
DURING THE BALANCE-DEXTERITY TASK
Abstract
Motor control dysfunction persisting during symptom remission in persons with recurrent
low back pain (rLBP) may contribute to the recurrence of pain. The purpose was to investigate
trunk control in persons in remission from rLBP using a dynamic, internally-driven balance task.
No differences in task performance were expected between those with and without rLBP, but it
was hypothesized persons with rLBP would exhibit greater trunk coupling in line with a trunk
stiffening strategy. Persons with and without rLBP (n=19 per group) completed the Balance-
Dexterity Task, which involved balancing on one limb while compressing an unstable spring
with the other. Task performance outcome measures included center-of-pressure velocity under
the stance limb and vertical force variability under the spring. Trunk coupling was quantified
with the coefficient of determination (R
2
) of an angle-angle plot of thorax-pelvis frontal plane
motion. Fine-wire and surface electromyography captured activations of paraspinals and
abdominals. There were no differences between groups for any task performance outcome
measure. The group in remission from rLBP exhibited reduced trunk coupling, or more
dissociated thorax and pelvis motion, compared to back-healthy controls (p=0.024). Trunk
coupling in this group was associated moderately with lumbar multifidus-to-erector spinae
activation ratio (R=0.608, p=0.007) and weakly with internal-to-external oblique ratio (R=0.473,
p=0.048). The Balance-Dexterity Task is a submaximal, internally-driven unstable balance task
during which more dissociated trunk motion was observed in persons in remission from rLBP.
TRUNK CONTROL DURING DYNAMIC BALANCE 117

Findings underscore the task-dependent nature of trunk control research and assessment in
persons with rLBP.
Introduction
The impact of recurrent low back pain (rLBP) on society has long been recognized. In a
recent global health study, out of 291 conditions studied, LBP ranked as the greatest contributor
to global disability.
29
The annual prevalence of activity-limiting LBP has been reported at 38%.
28

In a study of US health care spending in 2013, LBP and neck pain accounted for $87.6 billion –
the highest spending level for any musculoskeletal condition and the third highest of any health
condition behind diabetes and ischemic heart disease.
11
Unfortunately, despite decades of
research, spending on LBP and neck pain has increased the most of any condition studied,
excluding diabetes, in the last eighteen years. A major contributor to this continued impact is
persistent pain in the low back after the first episode. A recent systematic review analyzed
studies that followed persons after recovery from a first acute episode of LBP and reported
recurrence rates between 22.1% at three months and 77.1% at three years follow-up.
45

Researching persons with rLBP in periods of symptom remission allows effects of current pain
to be removed and residual aspects of motor control dysfunction or psychosocial factors to be
identified.
Motor control dysfunction in this population may contribute to the recurrence of pain.
Using mechanical perturbations to posture through support surface translation or trunk release
paradigms, multiple groups have reported increased trunk muscle co-contraction
30,43
and
stiffness
9,20
in persons with rLBP in symptom remission. Prospectively, delayed shut-off
latencies in response to a trunk release were associated with future LBP within two to three
years, where the risk of LBP increased 3% for every millisecond abdominal shut-off latency
TRUNK CONTROL DURING DYNAMIC BALANCE 118

delay.
6
It has been suggested that increasing trunk stiffness in the short-term after an acute
episode of LBP is beneficial to protect damaged tissue, but that residual increased stiffness after
symptoms subside could increase spine loading and decrease loading variability contributing to
further degeneration and pain recurrence.
23,26

Findings supporting a stiffening strategy are mixed, however, and seem to be dependent
on task and symptom status. In persons with symptomatic LBP performing unstable seated
balance tasks, we observe increased trunk stiffness associated with increased trunk muscle co-
contraction.
15–17
In a systematic review mixing research in persons with both current and a
history of LBP performing a variety of functional tasks, overall conclusions included reduced
lumbar range of motion, suggesting increased trunk stiffness.
32
During gait, findings depend on
the population studied. In persons in current pain, there are increased correlations of thorax and
pelvis motion, a kinematic measure used to describe stiffness,
12,27
but persons in remission
exhibited no difference in thorax-pelvis correlations.
7
In studies where a stiffening strategy is
observed, muscle activation patterns often include increased superficial trunk muscle activity.
Unstable standing balance revealed decreased activity of transverse abdominis and internal
oblique but increased activity of the more superficial external oblique in the LBP group.
13

Modeling of a discrete trunk release revealed that the increased co-activation adopted by
symptomatic persons stabilized the lumbar spine.
9
These findings suggest a pattern of overactive
superficial trunk musculature resulting in trunk stiffening in persons with rLBP, but this may be
more consistent while symptomatic or during large external perturbations to posture.
Other investigational tasks, however, have evoked reduced trunk stiffness associated with
impaired deep trunk muscle recruitment. In anticipation of and response to sudden trunk loading,
MacDonald et. al. reported reduced or entirely absent lumbar multifidus activation in persons in
TRUNK CONTROL DURING DYNAMIC BALANCE 119

remission from rLBP.
36
In an unstable sitting task, greater thoraco-lumbar movements and
reduced trunk stiffness were associated with reduced deep-to-superficial paraspinal muscle
activation ratios.
49
A study of walking turns in our lab revealed no difference in movement
variability or in-phase trunk coupling between groups, suggesting no difference in trunk
stiffness.
46
When walking speed was increased, control participants increased lumbar multifidus
activation duration, but persons in remission from rLBP reduced the duration of activity.
47

Conflicting conclusions about alterations in trunk control in this population contribute to
confusion in research and clinical practice.
These conflicts may be partially reconciled, however, by synthesizing research findings
recognizing task-dependency. A seminal modeling study by Cholewicki and McGill suggests
that lower-effort tasks place the lumbar spine at injury risk due to intervertebral instability but
higher-effort tasks place the lumbar spine at risk due to tissue failure.
5
This fits with findings
summarized thus far where increased demands on postural control, especially through externally
perturbed posture, evoked increases in trunk stiffness, but lower-demand continuous balance
tasks showed mixed findings depending on current symptoms. While it has long been recognized
that participant heterogeneity has been problematic for LBP research, here we also emphasize the
failure to critically evaluate investigational tasks for mechanical, physiological, and psychosocial
characteristics as well as the need to develop and characterize novel tasks that are well-tuned to
evoke dysfunctional trunk control strategies in ecological contexts along a continuum of task-
demand levels. In this context, laboratory research tasks must be carefully evaluated to address
where they fit on this task-dependency continuum, and new tasks should be developed with this
task-dependency in mind.
TRUNK CONTROL DURING DYNAMIC BALANCE 120

The Balance-Dexterity Task, (Chapter III) designed by combining single-limb balance
with the lower-extremity dexterity test
35
serves as an ideal task to observe aspects of
submaximal, unstable postural control. Adding the challenge of dexterous force control of the
lower limb to single-limb balance provides increased postural demands and evokes greater trunk
motion, making observing trunk control strategies more feasible than single-limb stance.
33

(Chapter III) The purpose of this study was to investigate trunk control in persons with rLBP in
symptom remission and back-healthy control participants using the Balance-Dexterity Task. It
was hypothesized that there would be no differences in task performance, but that persons with
rLBP would exhibit greater trunk coupling associated with greater superficial trunk muscle
activity, in line with a trunk stiffening strategy.
Methods
This cross-sectional controlled laboratory study was reviewed by USC’s Health Sciences
Campus Institutional Review Board. After informed consent, participants with non-specific rLBP
and matched back-healthy control participants with no history of LBP in the past year
39
were
recruited from student groups, flyers, and physical therapy clinics. Participants with rLBP must
have had at least two episodes of pain, localized to the area between the lower rib cage and
horizontal gluteal fold, per year for at least one year, but experienced pain less than half of the
days in the previous six months (distinguishes chronicity and recurrence
8
). Painful episodes were
severe enough to limit function based on questions in the NIH Task Force recommended
minimum dataset
8
and Oswestry Disability Index.
14
Participants were in symptom remission at
the time of testing and for the preceding seven days (pain <1.5 out of 10 on a visual analog scale
(VAS)
4
). Exclusion criteria included age > 45 years; low back surgery; a radiological or clinical
diagnosis of spinal stenosis, scoliosis, malignancy, infection, or radiculopathy; current or
TRUNK CONTROL DURING DYNAMIC BALANCE 121

previous musculoskeletal injury or surgery affecting locomotion or balance; a history of diabetes
mellitus, rheumatic joint disease, any blood-clotting disorder or current anti-coagulant therapy,
or polyneuropathy; or current pregnancy. A sample size of nineteen per group was determined
from a pre-hoc power analysis after four pilot subjects in each group were collected.
Participants were instrumented with a full-body retroreflective marker set as well as
surface electromyography (EMG) of the external oblique (EO), rectus abdominis (RA), gluteus
maximus (GMax), and gluteus medius (GMed) and fine-wire EMG of the internal oblique (IO),
lumbar multifidus (MF), and erector spinae (ES) at the level of L4 (Noraxon Wireless EMG;
Scottsdale, AZ; 3000Hz). Surface EMG were collected with bipolar silver/silver chloride
electrodes with an interelectrode distance of 22mm placed per guidelines from SENIAM
19
, and
fine-wire EMG were collected with a pair of 50μm nickel-chromium alloy wires insulated with
nylon with distal 2mm exposed and loaded into a 25-gauge hypodermic needle and sterilized.
Insertions were done under ultrasound guidance, and protocols were adapted from Perotto.
42
All
muscles were instrumented on the side contralateral to the participant’s preferred kicking limb,
hereafter referred to as the stance side. Motion data were captured with an 11-camera Qualisys
Oqus System (Gothenburg, Sweden; 250Hz), and force data were captured with two Advanced
Medical Technology Inc. force plates (Watertown, MA; 3000Hz).
The Balance-Dexterity Task device and procedures have been described previously
(Chapter III). Briefly, participants completed a 30s trial of double-limb standing (preferred
stance width) and three 30s trials of single-limb standing on the stance side. Participants were
introduced to the Balance-Dexterity Task, which used a custom device with a spring mounted
between two boards (Compression Spring #805, Century Spring Corp., Commerce, CA).
Participants were shown real-time feedback of the vertical force under the spring, and instructed:
TRUNK CONTROL DURING DYNAMIC BALANCE 122

“While standing on one leg, compress this spring so that the line is first as high, then as stable as
possible” (Figure III.1). Each trial lasted 20-25s. After one familiarization trial and five practice
trials, the mean of the middle 50% of the last three practice trials was used to calculate an
individual’s reproducible, submaximal compression. After practice, participants used a VAS to
report how difficult the task was, how confident they were they could complete the task
successfully, and how much attention the task required. Participants then completed five trials
where a dotted line indicating this reproducible, submaximal compression was shown as a goal
with the instructions: “While standing on one leg, compress this spring so that the line is as
stable as possible directly over the dotted goal line.” Three trials were interspersed where the
spring was replaced with a stable block of the same height, and the same instructions were given.
Trials were trimmed so the middle 50% of the task was analyzed. Kinematic and force
plate data were low-pass filtered with cutoff frequencies of 12Hz and 50Hz, respectively.
Surface and fine-wire EMG data were band-pass filtered between 20 and 500 Hz and 20 and
1000 Hz, respectively, all using a dual-pass 4
th
order Butterworth filter. After filtering, EMG
data were rectified and smoothed with a moving weighted average window of 500ms. Signals
were normalized to averaged signal amplitude during the stable block trials. Here, EMG
amplitude could be reported as a percentage of activation during the stable block, allowing
activation to be interpreted as a response to the added challenge of dexterous force control, not to
the body position or vertical force production.
Average center of pressure (COP) resultant velocity from the stance limb was calculated.
Dexterous force control was measured using the vertical force produced under the spring and
quantified as root-mean-squared error (RMSE) from the reproducible, submaximal compression
goal line and coefficient of variation (CV). Trunk control was quantified by tracking thorax and
TRUNK CONTROL DURING DYNAMIC BALANCE 123

pelvis motion relative to global coordinates. Using an angle-angle plot of thorax and pelvis
frontal plane rotation, a coefficient of determination (R
2
) was calculated where a high R
2
would
indicate highly coupled thorax and pelvis motion and a low R
2
would indicate more dissociated
or independent motion of the thorax and pelvis. This particular metric has been used during gait
to distinguish participants with and without LBP through frontal and transverse plane trunk
coupling.
7,27
In addition, more traditional range of motion metrics were acquired including trunk
(thorax relative to pelvis), thorax, and pelvis angular excursions. Muscle activation data were
averaged to acquire a mean activation amplitude for each trial. Muscle activation ratios were
calculated in a frame-by-frame manner including deep-to-superficial ratios for the paraspinals
(MF:ES) and abdominals (IO:EO) and co-contraction ratios for the deep trunk muscles (MF, IO)
and superficial trunk muscles (ES, EO) with the muscle of lower average amplitude in the
numerator. Outcome variables were statistically analyzed using two-way analysis of variance
(ANOVA) and post-hoc Bonferroni corrections when testing task (double-limb stance, single-
limb stance, stable block, and Balance-Dexterity Task) and group main effects, and using paired
t-tests when testing group effects alone. Associations between outcome measures were tested
with bivariate Pearson correlations with α=0.05 for all tests (PASW Statistics, IBM Corp.,
Armonk, NY).
Results
Participants
Nineteen participants with rLBP and no sign of neurological involvement and nineteen
matched back-healthy control participants participated in the study (Figure IV.1) (Table IV.1).
Participants with rLBP were in symptom remission at the time of testing with mean VAS pain
TRUNK CONTROL DURING DYNAMIC BALANCE 124

rating 0.4±0.4 out of 10. Note that one participant with rLBP is missing MF and IO data due to
failed fine-wire EMG insertion.

Figure IV.1. Participant recruitment. rLBP = recurrent low back pain.

TRUNK CONTROL DURING DYNAMIC BALANCE 125

Table IV.1. Participant demographics (mean ± standard deviation). BMI =
body mass index; VAS = visual analog scale; ODI = Oswestry Disability
Index; PCS = pain catastrophizing scale; TSK = tampa scale for
kinesiophobia; FABQ = fear-avoidance belief questionnaire; PA = physical
activity subset; W = work subset.

Task performance
Measures of balance and dexterous force control demands were not different between
persons with rLBP and back-healthy controls. All participants were able to complete the
Balance-Dexterity Task safely with reproducible, submaximal compression forces of 100-139N
(mean 121.2±12.3N), representing 14.4-23.0% of body weight (mean 18.7±2.4%) for back-
healthy controls and 102-159N (mean 123.5±15.4N) representing 13.8-22.2% of body weight
(mean 18.6±2.7%) for persons with rLBP (t(18)=-0.544, p=0.593, Cohen’s d=0.125) (Figure
IV.2A-B). There were also no differences between groups in self-report measures
(0.612>p>0.452) (Figure IV.2C). Dexterous force control error quantified as RMSE (t(18)=-
Table 1
rLBP CNTRL p
Age (yrs) 23.5 ± 2.8 23.9 ± 3.3 0.679
Sex 7 M, 12 F 7 M, 12 F
Leg Dominance 18 R, 1 L 18 R, 1 L
Height (cm) 170.4 ± 8.4 169.1 ± 10.4 0.692
Weight (kg) 68.7 ± 10.3 67.1 ± 10.8 0.661
BMI 23.6 ± 2.4 23.3 ± 1.8 0.714
Baecke Physical Activity Scale (vector sum) 4.8 ± 1.0 4.9 ± 0.6 0.537
Episodes per year 3.4 ± 1.2
Pain during episodes (recall, VAS 0-10) 4.9 ± 2.2
Pain at time of testing (VAS 0-10) 0.4 ± 0.4
ODI (recall, %) 16.0 ± 18.7
PCS (0-52) 8.3 ± 8.5 6.5 ± 5.1 0.438
TSK (17-68) 31.3 ± 6.5 30.5 ± 6.0 0.706
FABQ (0-96) 20.2 ± 10.7
FABQ-PA (0-66) 12.2 ± 7.7
FABQ-W (0-30) 8.1 ± 6.7
Demographics of Participants
TRUNK CONTROL DURING DYNAMIC BALANCE 126

0.476, p=0.640, Cohen’s d=0.109) or CV (t(18)=0.011, p=0.991, Cohen’s d=0.003) was not
different between groups (Figure IV.3B-C). COP velocity was analyzed as the primary measure
of task demands given all task conditions were significantly different for this measure only,
suggesting better sensitivity to balance demands in these tasks. There was no significant task-by-
group interaction effect for COP velocity (F(3,16)=1.036, p=0.403, 𝜂 𝑝 2
=0.163) or group main
effect (F(1,16)=0.416, p=0.526, 𝜂 𝑝 2
=0.023), but there was a significant task main effect
(F(3,16)=152.525, p<0.001, 𝜂 𝑝 2
=0.966). COP velocity increased from double-limb stance, to
stable block condition, to single-limb stance, and was greatest in the Balance-Dexterity Task, and
there were no group differences in any condition (Figure IV.3A). Task performance, therefore,
was not affected by a history of rLBP.

Figure IV.2. Reproducible, submaximal compression force achieved during
practice trials and set as the goal for test trials of the Balance-Dexterity Task
reported in Newtons (N) (A) and as a percentage of body weight (%BW) (B).
Self-report visual analog scale (VAS) measures of task difficulty, participant
confidence, and the amount of attention required to complete the task (C) for
persons with recurrent low back pain (rLBP) and back-healthy control
participants (CNTRL).
TRUNK CONTROL DURING DYNAMIC BALANCE 127

Figure IV.3. Center of pressure (COP) average resultant velocity in double-
limb stance, single-limb stance, stable block, and Balance-Dexterity Task
conditions. (A) Dexterous force control measures including root-mean-
squared-error (RMSE) (B) and coefficient of variation (CV) (C) of the
vertical force produced in the Balance-Dexterity Task for persons with
recurrent low back pain (rLBP) and back-healthy control participants
(CNTRL).
Trunk control
Participants with rLBP exhibited reduced frontal plane trunk coupling during the
Balance-Dexterity Task compared to control participants (t(18)=2.457, p=0.024, Cohen’s
d=0.564) but no differences in thorax, pelvis, or trunk frontal plane excursions (t(18)=-1.058,
p=0.304, Cohen’s d=0.243; t(18)=-1.414, p=0.174, Cohen’s d=0.324; t(18)=-1.333, p=0.199,
Cohen’s d=0.306, respectively) (Figure IV.4). There were no associations between trunk
coupling and COP velocity (CNTRL: R=0.260, p=0.282; rLBP: R=0.025, p=0.919) or dexterous
force control RMSE (CNTRL: R=0.262, p=0.279; rLBP: R=0.158, p=0.517) or CV (CNTRL:
R=0.146, p=0.551; rLBP: R=0.191, p=0.434) (Figure IV.5). There were also no associations
between trunk coupling and pain catastrophizing (CNTRL: R=0.07, p=0.776; rLBP: R=-0.103,
p=0.675), kinesiophobia (CNTRL: R=-0.133, p=0.588; rLBP: R=0.062, p=0.800), or fear-
avoidance beliefs (rLBP: R=0.041, p=0.867) (Figure IV.6). For both groups, trunk coupling
TRUNK CONTROL DURING DYNAMIC BALANCE 128

seemed to vary independently of balance and dexterous force control demands or social-
cognitive factors, and persons with rLBP exhibited reduced trunk coupling or more dissociated
motion of the thorax and pelvis.

Figure IV.4. Coupling of thorax and pelvis frontal plane rotation (trunk
coupling R
2
) (A) as well as thorax, pelvis, and trunk frontal plane excursion
(B) during the Balance-Dexterity Task in persons with recurrent low back
pain (rLBP) and back-healthy controls (CNTRL). *p<0.05
TRUNK CONTROL DURING DYNAMIC BALANCE 129

Figure IV.5. Associations between trunk coupling (R
2
) and balance and
dexterous force control measures for persons with recurrent low back pain
(rLBP) and back-healthy control participants (CNTRL). Balance was
quantified with center of pressure (COP) average resultant velocity (A), and
dexterous force control measures include root-mean-squared-error (RMSE)
(B) and coefficient of variation (CV) (C).

Figure IV.6. Associations between trunk coupling (R
2
) and psychometric
scores including pain catastrophizing scale (A), Tampa scale for
kinesiophobia (B), and fear-avoidance beliefs questionnaire (C) for persons
with recurrent low back pain (rLBP) and back-healthy control participants
(CNTRL).
Muscle activation data normalized to the stable block condition revealed no significant
differences in any individual EMG signal or ratio of EMG signals between groups
(0.972>p>0.242) (Figure IV.7). No muscles alone were associated with trunk coupling, but ratios
of deep-to-superficial paraspinals and abdominals were significantly positively associated with
TRUNK CONTROL DURING DYNAMIC BALANCE 130

trunk coupling only in the rLBP group (R=0.618, p=0.006 and R=0.476, p=0.046, respectively)
(Figure IV.8). Greater deep trunk muscle (MF, IO) activation relative to more superficial trunk
muscles (ES, EO) resulted in greater trunk coupling, while less relative deep muscle activation
resulted in more dissociated thorax and pelvis frontal plane motion.

Figure IV.7. (A) Muscle activation data during the Balance-Dexterity Task
for multifidus (MF), erector spinae (ES), internal oblique (IO), external
oblique (EO), rectus abdominis (RA), gluteus maximus (GMax), and gluteus
medius (GMed). (B) Muscle activation ratios describing deep-to-superficial
paraspinal activity (MF:ES), deep-to-superficial abdominal activity (IO:EO),
(C) deep muscle co-contraction (MF and IO), and superficial muscle co-
contraction (ES and EO) in persons with recurrent low back pain (rLBP)
and back-healthy control participants (CNTRL). EMG were normalized to
the stable block condition.
TRUNK CONTROL DURING DYNAMIC BALANCE 131

Figure IV.8. Associations between trunk coupling (R
2
) and deep-to-
superficial paraspinal muscle activation ratio multifidus (MF) : erector
spinae (ES) (A) and abdominal ratio internal oblique (IO) : external oblique
(EO) (B) in persons with recurrent low back pain (rLBP) and back-healthy
control participants (CNTRL). EMG were normalized to the stable block
condition before the ratios were calculated.
Discussion
Participants with and without rLBP were able to successfully complete five trials of the
Balance-Dexterity Task with no differences in reproducible, submaximal compression force
goal, COP velocity, or dexterous force control. In both groups, frontal plane trunk coupling
varied independently of any task performance measure or any psychometric collected. Persons
with rLBP had, on average, lower trunk coupling, meaning more dissociated thorax and pelvis
motion, compared to back-healthy control participants. And, in this group, trunk coupling was
associated with deep-to-superficial paraspinal and abdominal EMG ratios, where greater deep
muscle activation relative to more superficial muscles resulted in higher trunk coupling.
Performance on the Balance-Dexterity Task did not distinguish between persons with and
without rLBP in symptom remission. Findings from other standing balance tasks are mixed on
TRUNK CONTROL DURING DYNAMIC BALANCE 132

differences between groups, but a majority of studies report greater COP sway measures in LBP
groups,
38,44
but this effect is more robust for persons with symptomatic LBP unlike in the present
study. A recent attempt to use a clinical single-limb balance test to distinguish participants with
and without chronic LBP reported that even though reduced GMed strength was identified in the
patient group, this did not translate to impaired single-limb stance.
41
Given these mixed findings,
it is reasonable, therefore, no differences were observed for this group of young, minimally
disabled persons in remission from rLBP. A group of older participants with longer duration or
currently symptomatic LBP would potentially exhibit task performance differences. While task
performance was successful and measures of balance demands and dexterous force control were
not distinguishable between groups, redundant control mechanisms allowed different strategies
to be adopted.
Traditional trunk segment excursion measures were similar between groups, but the
coupling of frontal plane thorax and pelvis rotations was on average lower in persons with rLBP,
indicating more dissociated thorax and pelvis motion. This was counter to the hypothesis of
increased coupling in the clinical population through adoption of a stiffening strategy. This
unsupported hypothesis was built on investigations of discrete perturbations to posture where
greater trunk co-contraction
30,43
and stiffness
9,20
has been observed in persons with rLBP. These
perturbations, however, involved delivering external perturbing forces either to a support surface
or directly to the trunk, which may invoke a stiffening strategy related to fear of movement or
pain. In fact, one study showed persons with higher kinesiophobia and fear-avoidance beliefs
exhibited greater trunk stiffness in one of these discrete trunk release tasks.
31
In the present
study, there was no such association between trunk coupling and any psychometric measure of
fear of pain or movement. In a continuous functional task like gait, there are mixed findings
TRUNK CONTROL DURING DYNAMIC BALANCE 133

depending on the population studied. When in pain, increased correlations of thorax and pelvis
motion are observed,
12,27
but a study on persons in symptom remission show no difference
between groups.
7
Taking this young, minimally disabled population in pain remission with this
submaximal, low-range-of-motion, volitionally-driven unstable balance task, it is likely we are
observing thorax and pelvis dissociation in a low-effort task as opposed to the trunk stiffening
seen in many external perturbation paradigms or higher-effort tasks. This harkens back to a
framework proposed by Cholewicki and McGill where causes of LBP in low-effort tasks are
related to instability and in high-effort tasks are more related to tissue failure.
5
The Balance-
Dexterity Task serves as an unstable balance task to observe dissociated trunk motion in this
population at one end of this continuum of task demands.
Further evidence linking reduced trunk coupling to aspects of deep trunk muscle
dysfunction was found in the associations between trunk coupling and deep-to-superficial trunk
muscle ratios. Trunk muscles, most consistently the deep lumbar MF, transverse abdominis, and
IO, exhibit dysfunction in persons with rLBP. Lumbar MF is atrophied at the level of pain
2
and
has reduced metabolic measures.
18
In anticipation of voluntary arm movement, persons with
rLBP exhibit delayed activation of the deep lumbar MF
37
and transverse abdominis
22,25
in all
directions of arm movement, and the IO
22,25
in certain directions. Experimentally induced muscle
pain through injection of hypertonic saline also reduced activity of the transverse abdominis and
ES.
24
Though studied less frequently, perturbations through voluntary lower extremity
movements show similar results of delayed activation of transverse abdominis and lumbar MF
consistently, and of IO and ES dependent on movement direction.
21,48
Mechanical modeling
studies suggest that every trunk muscle contributes to lumbar stability in large and multi-planar
movements, but that passive structures and the MF and ES primarily control stability in neutral
TRUNK CONTROL DURING DYNAMIC BALANCE 134

posture tasks with near-zero muscle forces.
5
Although, intra-abdominal pressure modulated
through abdominal activation was not modeled in the study. The findings from the present study
support these previous investigations and add that, in fact, the relative activation of MF and ES
as well as IO and EO predict trunk coupling in a rLBP population. This relationship only existed
in the clinical population, potentially because impaired motor control processes, including
reduced trunk proprioception,
34,40
paraspinal atrophy,
2
and increased trunk extensor
fatiguability,
1
meant trunk coupling was more strongly or more exclusively influenced by muscle
coordination as opposed to back-healthy control participants with many redundant control
mechanisms contributing to coupling. This confirms hypotheses suggested in our previous work
where MF:ES, with signals normalized to maximum contractions, was weakly associated with
trunk coupling in back-healthy controls (Chapter III).
The reduced trunk coupling associated with less relative deep trunk muscle activation has
the capacity to contribute to lumbar spine degeneration and the recurrence of pain in this clinical
population. Dysfunctional mechanical behavior of the lumbar spine, through motor control
dysfunction and/or tissue damage, is thought to play a role in degeneration through intervertebral
instability
3,5
and/or through increased load and decreased movement variability.
10,23,26
Other
groups have proposed a link between the stiffening strategy, characterized by increased
superficial muscle activation and trunk stiffness, in persons with rLBP and a reduction in
movement variability leading to increased load of the same tissues. Here, we are observing the
opposite end of the spectrum – dissociated trunk motion suggesting reduced lumbar stability
associated with reduced relative deep trunk muscle activity. The Balance-Dexterity Task itself
may serve as a means to characterize these factors during patient assessment and cue and train
rehabilitation of trunk control. Taking a closer look at individuals in Figure IV.8 reveals that a
TRUNK CONTROL DURING DYNAMIC BALANCE 135

few participants in both groups increased their deep-to-superficial trunk muscle ratios far above
the mean in order to achieve trunk coupling similar to most back-healthy controls – consider four
subjects with MF:ES ratios above 2.0 and two with IO:EO ratios above 2.0. It is possible these
individuals represent a subgroup that has learned a strategy for maintaining appropriate trunk
coupling, at the expense of efficient deep-to-superficial muscle activation ratios. Given the high
prevalence of LBP, the back-healthy control participants with these distinctly higher ratios may
be at risk for future development of pain, but this conclusion is speculative.
The present findings fit well into what is currently known about motor control
dysfunction in persons with rLBP when viewed in a task-dependent way, but limitations to
applying these findings must be considered. Participants made up a convenience sample of
persons with and without rLBP recruited from student groups, classes, flyers, and university-
affiliated physical therapy clinics. Persons with rLBP were generally young (age 23.5±2.8yrs),
minimally disabled (ODI 16.0±18.7%), and all in pain remission at the time of testing (0.4±0.4
out of 10 on VAS). Generalizing findings to persons with rLBP in a painful episode or persons
with acute or chronic LBP may be inappropriate. The usual limitations of EMG methodology
apply here as well including the potential for cross-talk from surface EMG and potential errors in
fine-wire placement. Insertions were done, however, under ultrasound guidance and confirmed
through electrical stimulation so substantial errors are unlikely.
Future work should continue to clarify the task-dependency of what is commonly
discussed about motor control dysfunction in persons with various LBP etiologies. Robust
studies taking the same set of participants through a variety of controlled laboratory postural
tasks – both externally-perturbed and internally-driven – will help elucidate the influence of
postural demands, muscle coordination, and psychosocial factors on trunk control and pain
TRUNK CONTROL DURING DYNAMIC BALANCE 136

recurrence in ecological tasks of varying effort. Adding cognitive perturbations through dual-task
interference can add to the ecological validity of the Balance-Dexterity Task and may reveal how
attentional demands influence trunk control and muscle coordination. Applied clinical research
could also test the effects of using the Balance-Dexterity Task in patient assessment and
rehabilitation.
Conclusion
Participants in symptom remission from rLBP did not perform the Balance-Dexterity
Task differently than back-healthy controls, but exhibited reduced frontal plane trunk coupling
indicating more dissociated thorax and pelvis motion. In the rLBP only, lower trunk coupling
was associated with lower deep-to-superficial paraspinal (MF:ES) and abdominal (IO:EO)
muscle activation ratios. The Balance-Dexterity Task is a submaximal, internally-driven unstable
balance task that induces sufficient perturbation to postural control to observe trunk control
deficits in this minimally-disabled population in remission from pain.

TRUNK CONTROL DURING DYNAMIC BALANCE 137

References
1. Beneck GJ, Baker LL, Kulig K. Spectral analysis of EMG using intramuscular electrodes
reveals non-linear fatigability characteristics in persons with chronic low back pain. J
Electromyogr Kinesiol. 2013;23(1):70-77. doi:10.1016/j.jelekin.2012.07.001.
2. Beneck GJ, Kulig K. Multifidus atrophy is localized and bilateral in active persons with
chronic unilateral low back pain. Arch Phys Med Rehabil. 2012;93(2):300-306.
doi:10.1016/j.apmr.2011.09.017.
3. Breen A, Breen A. Uneven intervertebral motion sharing is related to disc degeneration
and is greater in patients with chronic, non-specific low back pain: an in vivo, cross-
sectional cohort comparison of intervertebral dynamics using quantitative fluoroscopy.
Eur Spine J. 2018;27(1):145-153. doi:10.1007/s00586-017-5155-y.
4. Carlsson AM. Assessment of chronic pain. I. Aspects of the reliability and validity of the
visual analogue scale. Pain. 1983;16(1):87-101. doi:10.1016/0304-3959(83)90088-X.
5. Cholewicki J, McGill SM. Mechanical stability of the in vivo lumbar spine. Clin
Biomech. 1996;11(1):1-15.
6. Cholewicki J, Silfies SP, Shah RA, et al. Delayed trunk muscle reflex responses increase
the risk of low back injuries. Spine (Phila Pa 1976). 2005;30(23):2614-2620.
doi:10.1097/01.brs.0000188273.27463.bc.
7. Crosbie J, de Faria Negrão Filho R, Nascimento DP, Ferreira P. Coordination of Spinal
Motion in the Transverse and Frontal Planes During Walking in People With and Without
Recurrent Low Back Pain. Spine (Phila Pa 1976). 2013;38(5):E286-E292.
doi:10.1097/BRS.0b013e318281de28.
TRUNK CONTROL DURING DYNAMIC BALANCE 138

8. Deyo RA, Dworkin SF, Amtmann D, et al. Report of the NIH Task Force on research
standards for chronic low back pain. J Pain. 2014;15(6):569-585.
doi:10.1016/j.jpain.2014.03.005.
9. van Dieën JH, Cholewicki J, Radebold A. Trunk muscle recruitment patterns in patients
with low back pain enhance the stability of the lumbar spine. Spine (Phila Pa 1976).
2003;28(8):834-841. doi:10.1097/01.BRS.0000058939.51147.55.
10. van Dieën JH, Flor H, Hodges PW. Low-Back Pain Patients Learn to Adapt Motor
Behavior with Adverse Secondary Consequences. Exerc Sport Sci Rev. 2017:223-229.
doi:10.1249/JES.0000000000000121.
11. Dieleman JL, Baral R, Birger M, et al. US Spending on Personal Health Care and Public
Health, 1996-2013. Jama. 2016;316(24):2627. doi:10.1001/jama.2016.16885.
12. Ebrahimi S, Kamali F, Razeghi M, Haghpanah SA. Comparison of the trunk-pelvis and
lower extremities sagittal plane inter-segmental coordination and variability during
walking in persons with and without chronic low back pain. Hum Mov Sci. 2017;52:55-
66. doi:10.1016/j.humov.2017.01.004.
13. Ehsani F, Arab AM, Jaberzadeh S, Salavati M. Ultrasound measurement of deep and
superficial abdominal muscles thickness during standing postural tasks in participants with
and without chronic low back pain. Man Ther. 2016;23:98-105.
doi:10.1016/j.math.2016.01.003.
14. Fairbank JC, Pynsent PB. The Oswestry Disability Index. Spine (Phila Pa 1976).
2000;25(22):2940-52; discussion 2952. doi:10.1097/00007632-200011150-00017.
TRUNK CONTROL DURING DYNAMIC BALANCE 139

15. Freddolini M, Strike S, Lee R. Dynamic stability of the trunk during unstable sitting in
people with low back pain. Spine (Phila Pa 1976). 2014;39(10):785-790.
doi:10.1097/BRS.0000000000000296.
16. Freddolini M, Strike S, Lee RYW. Stiffness properties of the trunk in people with low
back pain. Hum Mov Sci. 2014;36:70-79. doi:10.1016/j.humov.2014.04.010.
17. Freddolini M, Strike S, Lee RYW. The role of trunk muscles in sitting balance control in
people with low back pain. J Electromyogr Kinesiol. 2014;24(6):947-953.
doi:10.1016/j.jelekin.2014.09.009.
18. Goubert D, De Pauw R, Meeus M, et al. Lumbar muscle structure and function in chronic
versus recurrent low back pain: a cross-sectional study. Spine J. 2017;17(9):1285-1296.
doi:10.1016/j.spinee.2017.04.025.
19. Hermens H, Freriks B, Merletti R, et al. The Surface ElectroMyoGraphy for the Non-
Invasive Assessment of Muscles Project (SENIAM). seniam.org. Published 2006.
20. Hodges P, van den Hoorn W, Dawson A, Cholewicki J. Changes in the mechanical
properties of the trunk in low back pain may be associated with recurrence. J Biomech.
2009;42(1):61-66. doi:10.1016/j.jbiomech.2008.10.001.
21. Hodges P, Richardson C. Delayed postural contraction of transversus abdominis in low
back pain associated with movement of the lower limb. J Spinal Disord. 1998;11(1):46-
56.
22. Hodges P, Richardson C. Inefficient muscular stabilization of the lumbar spine associated
with low back pain A motor control evaluation of transversus abdominis. Spine (Phila Pa
TRUNK CONTROL DURING DYNAMIC BALANCE 140

1976). 1996;21(22):2640-2650.
23. Hodges PW. Pain and motor control: From the laboratory to rehabilitation. J Electromyogr
Kinesiol. 2011;21(2):220-228. doi:10.1016/j.jelekin.2011.01.002.
24. Hodges PW, Moseley GL, Gabrielsson A, Gandevia SC. Experimental muscle pain
changes feedforward postural responses of the trunk muscles. Exp Brain Res.
2003;151(2):262-271. doi:10.1007/s00221-003-1457-x.
25. Hodges PW, Richardson CA. Altered trunk muscle recruitment in people with low back
pain with upper limb movement at different speeds. Arch Phys Med Rehabil.
1999;80(9):1005-1012. doi:10.1016/S0003-9993(99)90052-7.
26. Hodges PW, Tucker K. Moving differently in pain: A new theory to explain the adaptation
to pain. Pain. 2011;152(SUPPL.3):S90-S98. doi:10.1016/j.pain.2010.10.020.
27. van den Hoorn W, Bruijn SM, Meijer OG, Hodges PW, van Dieen JH. Mechanical
coupling between transverse plane pelvis and thorax rotations during gait is higher in
people with low back pain. J Biomech. 2012;45(2):342-347.
doi:10.1016/j.jbiomech.2011.10.024.
28. Hoy D, Bain C, Williams G, et al. A systematic review of the global prevalence of low
back pain. Arthritis Rheum. 2012;64(6):2028-2037. doi:10.1002/art.34347.
29. Hoy D, March L, Brooks P, et al. The global burden of low back pain: Estimates from the
Global Burden of Disease 2010 study. Ann Rheum Dis. 2014;73(6):968-974.
doi:10.1136/annrheumdis-2013-204428.
30. Jones SL, Henry SM, Raasch CC, Hitt JR, Bunn JY. Individuals with non-specific low
TRUNK CONTROL DURING DYNAMIC BALANCE 141

back pain use a trunk stiffening strategy to maintain upright posture. J Electromyogr
Kinesiol. 2012;22(1):13-20. doi:10.1016/j.jelekin.2011.10.006.
31. Karayannis N V, Smeets RJEM, van den Hoorn W, Hodges PW. Fear of Movement Is
Related to Trunk Stiffness in Low Back Pain. PLoS One. 2013;8(6).
doi:10.1371/journal.pone.0067779.
32. Laird RA, Gilbert J, Kent P, Keating JL. Comparing lumbo-pelvic kinematics in people
with and without back pain: a systematic review and meta-analysis. BMC Musculoskelet
Disord. 2014;15(1):229. doi:10.1186/1471-2474-15-229.
33. Lawrence EL, Cesar GM, Bromfield MR, Peterson R, Valero-Cuevas FJ, Sigward SM.
Strength, Multijoint Coordination, and Sensorimotor Processing Are Independent
Contributors to Overall Balance Ability. Biomed Res Int. 2015;2015.
doi:10.1155/2015/561243.
34. Lee AS, Cholewicki J, Reeves NP, Zazulak BT, Mysliwiec LW. Comparison of trunk
proprioception between patients with low back pain and healthy controls. Arch Phys Med
Rehabil. 2010;91(9):1327-1331. doi:10.1016/j.apmr.2010.06.004.
35. Lyle MA, Valero-Cuevas FJ, Gregor RJ, Powers CM. The lower extremity dexterity test
as a measure of lower extremity dynamical capability. J Biomech. 2013;46(5):998-1002.
doi:10.1016/j.jbiomech.2012.11.058.
36. MacDonald D, Moseley GL, Hodges PW. People With Recurrent Low Back Pain
Respond Differently to Trunk Loading Despite Remission From Symptoms. Spine (Phila
Pa 1976). 2010;35(7):818-824. doi:10.1097/BRS.0b013e3181bc98f1.
TRUNK CONTROL DURING DYNAMIC BALANCE 142

37. MacDonald D, Moseley GL, Hodges PW. Why do some patients keep hurting their back?
Evidence of ongoing back muscle dysfunction during remission from recurrent back pain.
Pain. 2009;142(3):183-188. doi:10.1016/j.pain.2008.12.002.
38. Mazaheri M, Coenen P, Parnianpour M, Kiers H, van Dieën JH. Low back pain and
postural sway during quiet standing with and without sensory manipulation: A systematic
review. Gait Posture. 2013;37(1):12-22. doi:10.1016/j.gaitpost.2012.06.013.
39. Norton G, McDonough CM, Cabral HJ, Shwartz M, Burgess JF. Classification of patients
with incident non-specific low back pain: implications for research. Spine J. 2015:1-10.
doi:10.1016/j.spinee.2015.08.015.
40. Osthoff A-KR, Ernst MJ, Rast FM, et al. Measuring Lumbar Reposition Accuracy in
Patients With Unspecific Low Back Pain: Systematic Review and Meta-analysis. Spine
(Phila Pa 1976). 2015;40(2):E97-E111. doi:10.1097/BRS.0000000000000677.
41. Penney T, Ploughman M, Austin MW, Behm DG, Byrne JM. Determining the activation
of gluteus medius and the validity of the single leg stance test in chronic, nonspecific low
back pain. Arch Phys Med Rehabil. 2014;95(10):1969-1976.
doi:10.1016/j.apmr.2014.06.009.
42. Perotto AO, Delagi EF, Iazzetti J, Morrison D. Anatomical Guide for the
Electromyographer. 5th ed. Springfield, Illinois: Charles C Thomas Publisher, LTD.;
2011.
43. Radebold A, Cholewicki J, Panjabi MM, Patel TC. Muscle response pattern to sudden
trunk loading in healthy individuals and in patients with chronic low back pain. Spine
(Phila Pa 1976). 2000;25(8):947-954. doi:10.1097/00007632-200004150-00009.
TRUNK CONTROL DURING DYNAMIC BALANCE 143

44. da Silva RA, Vieira ER, Fernandes KBP, et al. People with chronic low back pain have
poorer balance than controls in challenging tasks. Disabil Rehabil. 2018;40(11):1294-
1300. doi:10.1080/09638288.2017.1294627.
45. da Silva T, Mills K, Brown BT, Herbert RD, Maher CG, Hancock MJ. Risk of Recurrence
of Low Back Pain: A Systematic Review. J Orthop Sport Phys Ther. 2017;47(5):305-313.
doi:10.2519/jospt.2017.7415.
46. Smith JA, Kulig K. Trunk-pelvis coordination during turning: A cross sectional study of
young adults with and without a history of low back pain. Clin Biomech. 2016;36:58-64.
doi:10.1016/j.clinbiomech.2016.05.011.
47. Smith J, Kulig K. Altered multifidus recruitment during walking in young asymptomatic
individuals with a history of low back pain. J Orthop Sports Phys Ther. 2016;27:46-51.
48. Suehiro T, Mizutani M, Ishida H, Kobara K, Osaka H, Watanabe S. Individuals with
chronic low back pain demonstrate delayed onset of the back muscle activity during prone
hip extension. J Electromyogr Kinesiol. 2015;25(4):675-680.
doi:10.1016/j.jelekin.2015.04.013.
49. Willigenburg NW, Kingma I, van Dieën JH. Center of pressure trajectories, trunk
kinematics and trunk muscle activation during unstable sitting in low back pain patients.
Gait Posture. 2013;38(4):625-630. doi:10.1016/j.gaitpost.2013.02.010.

TRUNK CONTROL DURING DYNAMIC BALANCE 144

CHAPTER V
INFLUENCE OF DUAL-TASK INTEREFERENCE ON TRUNK COUPLING
IN PERSONS WITH AND WITHOUT RECURRENT LOW BACK PAIN
DURING DYNAMIC BALANCE
Abstract
The purpose of the study was to investigate effects of cognitive dual-task interference and
task prioritization instructions on task performance and trunk control during dynamic balance in
persons with and without recurrent low back pain (rLBP). First, the ability to modulate task
performance in accord with prioritization instructions was tested. Second, it was hypothesized
trunk control strategies in persons with rLBP would rely more on cognitive resources, and
therefore trunk kinematics would be altered in this group under dual-task interference compared
to back-healthy control participants. Finally, individual factors explaining changes in trunk
kinematics were explored. Persons with and without rLBP (n=19 per group) completed the
Balance-Dexterity Task, which involved balancing on one limb while compressing an unstable
spring with the other, with and without a cognitive task utilizing verbal working memory. Trunk
coupling was quantified with the coefficient of determination (R
2
) of an angle-angle plot of
thorax-pelvis frontal plane motion. Fine-wire and surface electromyography captured activations
of paraspinals and abdominals. No differences between groups were identified for task
performance measures, indicating the groups did not modulate performance differently. Trunk
coupling was lower in the rLBP compared to back-healthy controls in the single-task condition
(p=0.024) and increased significantly in the dual-task condition (p=0.002). The amount of
increase was associated with reductions in erector spinae activation (R=-0.580), lower self-
reported cognitive task difficulty (R=-0.497), and lower recalled pain (R=-0.642). Interpreting
TRUNK CONTROL DURING DYNAMIC BALANCE 145

findings in perspectives of movement-specific reinvestment and action-specific perception
theories sheds light on cognitive contributions to trunk control in persons with rLBP.
Introduction
Attentional resources utilized for postural control have been studied with dual-task
interference paradigms in non-disabled adults and diverse patient populations. Understanding
alterations in posture-cognition prioritization dynamics plays an important role in designing
interventions for persons suffering from chronic or recurrent low back pain (rLBP) as
highlighted by recent calls for investigations into the role of attentional processing and
psychometrics in this clinical population.
1,2
Research into the effects of dual-task interference in
persons with rLBP is limited and findings are mixed. We believe this is, in part, due to dual-
tasking being used in research as an ecological test condition, which certainly has merit, but
often lacks a solid theoretical foundation and suffers from simplistic assumptions about dual-task
interference. Other research groups have called for more careful study design and dual-task
methodology in posture-cognition research as well.
3

From some well-designed studies, features of cognitive processing in persons with rLBP
have been characterized. From a motor control point of view, recruitment of muscles as part of
primary and automatic postural control systems seems to be more linked. Hodges et. al. report
that an increased cognitive demand delayed transverse abdominis onset prior to arm flexion.
4
In
controls only the onset of the prime mover (deltoid) was delayed, but for persons with rLBP in
remission, muscles recruited as part of both primary motor control and the associated
anticipatory postural control were delayed, suggesting cognitive demands influenced both
systems together in this clinical population. Kinematic analyses of the effects of dual-task
interference on trunk control in persons with symptomatic LBP are mixed with some studies
TRUNK CONTROL DURING DYNAMIC BALANCE 146

reporting decreases in trunk stiffness (measured by decreased trunk and pelvis segment
correlations during unstable sitting
5
and by increased trunk motion variability during gait
6
) and
another reporting no differences between groups in functional balance tasks.
7
The type of task
used to challenge balance and trunk control may play a role in these mixed findings in
symptomatic individuals.
There is even more limited research on dual-task interference during continuous
functional tasks in persons in remission from rLBP. One study using support surface translations
and a concurrent Stroop task reported different effects on balance dynamics dependent on the
intensity of the perturbation.
8
Work from our lab studying a walking pivot turn reported
increases in consistency of step length and trunk-pelvis coupling (suggesting increased trunk
stiffness) under dual-task interference.
9
Participants in this study in remission from rLBP also
exhibited increased trunk frontal plane motion with dual-task interference. While there are
clearly distinct influences of cognitive load in persons in remission from rLBP, more research
needs to be done to understand mechanisms behind how dual-task interference affects motor and
cognitive performance. Performing the continuous unstable balance task used in the present
study – the Balance-Dexterity Task – persons in remission from rLBP exhibited more dissociated
frontal plane thorax and pelvis motion (lower trunk coupling suggesting reduced stiffness
compared to back-healthy controls) (Chapter IV). If, in fact, this abnormal movement pattern
reflects a less automatic postural control strategy using more cognitive contributions in this
population compared to back-healthy controls, one hypothesis is that under a dual-task
interference condition which consumes cognitive resources the rLBP group’s motor behavior
will assume a more natural trunk coupling strategy, similar to the back-healthy control group. A
robust investigation to test this hypothesis is warranted.
TRUNK CONTROL DURING DYNAMIC BALANCE 147

Perturbing attentional resources through dual-task paradigms is generally divided into
dual-task probe paradigms and dual-task interference paradigms. Hui-Ting Goh presents a
wonderful description of these, summarized here,
10,11
where dual-task probe paradigms involve
responding to a discrete cognitive task while participants complete two different continuous
motor tasks, while dual-task interference involves just one continuous motor task completed with
and without a concurrent continuous cognitive task. Studies utilize dual-task interference
paradigms to mimic ecological attention conditions. An important limitation of these paradigms,
however, is how to ensure participants are assigning appropriate effort or priority to these two
tasks. To address this, researchers have developed a “primacy switch” condition where
prioritization instructions are given to prioritize the motor task in one dual-task condition and the
cognitive task in another. This allows the effects of dual-task interference on the motor task to be
“double-checked” but putting the priority back on the motor task and seeing if effects of
interference are reversed.
10
In the present study, a dual-task interference paradigm with a
primacy switch is used to assess effects of interference on task performance and trunk control
during the Balance-Dexterity Task.
Importantly, cognitive tasks must be developed specific to the research questions being
asked. As outlined by the Wickens attention model,
12
choosing a cognitive task that interferes
with similar attentional resources as the motor task being studied will invoke the most dual-task
interference. The present study uses a cognitive task with the specific intention of interfering
with verbal working memory contributions to movement, which are elevated in conditions of
pain or injury
13
and in persons with LBP.
14,15
This specificity was accomplished by adapting a
continuous verbal working memory recall and manipulation task designed to interfere with
verbal attention resources.
16

TRUNK CONTROL DURING DYNAMIC BALANCE 148

The purpose of the study was to investigate effects of cognitive dual-task interference and
task prioritization instructions on task performance and trunk control during dynamic balance in
persons with and without rLBP. First, the task prioritization hypothesis states that the ability to
modulate task performance in accord with prioritization instructions would be impaired in
persons with rLBP, and this would be expressed as an attenuated response to prioritization
manipulation compared to back-healthy control participants. Second, the trunk control
hypothesis states that trunk control strategies in persons with rLBP would rely more on cognitive
resources, and therefore under a dual-task interference condition, trunk kinematics would be
altered in this group compared to back-healthy control participants. Finally, individual factors
that could explain these alterations in trunk control were explored including self-report measures
of cognitive task difficulty, pain recall, and trunk muscle activation patterns.
Methods
Participants and instrumentation
Participants with rLBP who have had at least two episodes of pain, localized to the area
between the lower rib cage and horizontal gluteal fold, per year for at least one year, but
experienced pain less than half of the days in the previous six months (distinguishes chronicity
and recurrence
17
) and back-healthy control participants with no history of LBP in the past year
18

matched by sex, leg dominance, age within five years, BMI category, and Baecke activity scale
vector sum within two points were recruited for the study, per Institutional Review Board
approval and with informed consent. Painful episodes were severe enough to limit function based
on questions in the NIH Task Force recommended minimum dataset
17
and Oswestry Disability
Index.
19
Participants were in symptom remission at the time of testing and for the preceding
seven days (pain <1.5 out of 10 on a visual analog scale (VAS)
20
). Exclusion criteria included
TRUNK CONTROL DURING DYNAMIC BALANCE 149

age > 45 years; low back surgery; a radiological or clinical diagnosis of spinal stenosis, scoliosis,
malignancy, infection, or radiculopathy; current or previous musculoskeletal injury or surgery
affecting locomotion or balance; a history of diabetes mellitus, rheumatic joint disease, any
blood-clotting disorder or current anti-coagulant therapy, or polyneuropathy; or current
pregnancy. A sample size of nineteen per group was determined from a pre-hoc power analysis
after data from four pilot subjects in each group were assessed.
Participants were instrumented with a full-body retroreflective marker set as well as
surface electromyography (EMG) of the external oblique (EO) and fine-wire EMG of the
internal oblique (IO), lumbar multifidus (MF), and erector spinae (ES) at the level of L4
(Noraxon Wireless EMG; Scottsdale, AZ; 3000Hz). Surface EMG were collected with bipolar
silver/silver chloride electrodes with an interelectrode distance of 22mm placed per guidelines
from SENIAM
21
, and fine-wire EMG were collected with a pair of 50μm nickel-chromium alloy
wires insulated with nylon with distal 2mm exposed and loaded into a 25-gauge hypodermic
needle and sterilized. Insertions were done under ultrasound guidance, and protocols were
adapted from Perotto.
22
All muscles were instrumented on the side contralateral to the
participant’s preferred kicking limb, hereafter referred to as the stance side. Motion data were
captured with an 11-camera Qualisys Oqus System (Gothenburg, Sweden; 250Hz), and force
data were captured with two Advanced Medical Technology Inc. force plates (Watertown, MA;
3000Hz). Self-report information about participants’ pain presentation was acquired including
reporting pain during a typical episode, pain at time of testing, and episodes of pain per year.
Procedures
The Balance-Dexterity Task device and procedures have been described previously (Chapter
III). Briefly, participants were introduced to the Balance-Dexterity Task, which used a custom
TRUNK CONTROL DURING DYNAMIC BALANCE 150

device with a spring mounted between two boards (Compression Spring #805, Century Spring
Corp., Commerce, CA). Participants were shown real-time feedback of the vertical force under
the spring, and instructed: “While standing on one leg, compress this spring so that the line is
first as high, then as stable as possible” (Figure III.1). Each trial lasted 20-25s. After one
familiarization trial and five practice trials, the mean of the middle 50% of the last three practice
trials was used to calculate an individual’s reproducible, submaximal compression. Next,
participants were introduced to the cognitive task which was a modified verbal working memory
recall and manipulation task requiring participants remember a list of five random digits from
zero to ten, do an arithmetic operation to those digits twice during the trial, and remember five
answers. The development of this task was constrained by the EMG instrumentation of deep
abdominals, which activate with verbalization.
23
Therefore, the task had to be modified from
similar tasks in the literature to avoid real-time verbalization of answers. After practice of each
of these tasks, participants used a VAS to report how difficult each task was, how confident they
were they could complete each task successfully, how much attention each task required, and
finally how confident they were they could do both tasks concurrently. Then, in a randomized
order, participants completed the following task conditions, each consisting of five trials:
 STCog: single-task condition for the cognitive task, conducted while seated;
 STBal: single-task condition for the Balance-Dexterity Task, where a dotted line
indicating the participant’s reproducible, submaximal compression was shown as a goal
with the instructions, “While standing on one leg, compress this spring so that the line is
as stable as possible directly over the dotted goal line”;
TRUNK CONTROL DURING DYNAMIC BALANCE 151

 DTCog: dual-task condition with a cognitive priority, where the instructions given were,
“In this trial, it is most important that you get the five numbers correct, and it is less
important the line is as stable as possible”;
 DTBal: dual-task condition with a balance priority, where the instructions given were, “In
this trial, it is most important that you keep the line as stable as possible, and it is less
important that you get the five numbers correct”;
 and finally, three trials where the spring was replaced with a stable block of the same
height, and the same instructions as in STBal were given.
Data analysis
Trials were trimmed so the middle 50% of the task was analyzed. Kinematic and force
plate data were low-pass filtered with cutoff frequencies of 12Hz and 50Hz, respectively.
Surface and fine-wire EMG data were band-pass filtered between 20 and 500 Hz and 20 and
1000 Hz, respectively, all using a dual-pass 4
th
order Butterworth filter. After filtering, EMG
data were rectified and smoothed with a moving weighted average window of 500ms. Signals
were normalized to averaged signal amplitude during the stable block trials. Here, EMG
amplitude could be reported as a percentage of activation during the stable block, allowing
activation to be interpreted as a response to the added challenge of dexterous force control, not to
the body position or vertical force production. EMG signals were averaged to acquire a mean
activation amplitude for each trial. Muscle activation ratios were calculated in a frame-by-frame
manner including deep-to-superficial ratios for the paraspinals (MF:ES) and abdominals
(IO:EO).
TRUNK CONTROL DURING DYNAMIC BALANCE 152

To test the task prioritization hypothesis, measures of task performance were collected
including average center of mass (COM) resultant velocity to quantify balance control and root-
mean-squared error (RMSE) of the vertical force produced under the spring relative to the
reproducible, submaximal compression goal line to quantify dexterous force control. Cognitive
task performance was measured as variability of errors in the five trials of each condition, a
measure more sensitive than average absolute error.
24,25
The hypothesis was tested using a two-
way analysis of variance (ANOVA) for each outcome measure – RMSE, COM velocity, and
cognitive task error variability. Two groups (control and rLBP) and three conditions (for RMSE
and COM velocity: STBal, DTCog, and DTBal; for cognitive task error variability: STCog,
DTCog, and DTBal) were included as factors in the ANOVA. Significant interaction effects for
one or more of these outcome measures would support the hypothesis that these groups are
prioritizing the tasks differently. Statistical tests were performed with α=0.05 and Bonferroni
corrections for multiple comparisons (PASW Statistics, IBM Corp., Armonk, NY).
To test the trunk control hypothesis, trunk coupling was quantified by tracking thorax and
pelvis motion relative to global coordinates. Using an angle-angle plot of thorax and pelvis
frontal plane rotation, a coefficient of determination (R
2
) was calculated where a high R
2
would
indicate highly coupled thorax and pelvis motion and a low R
2
would indicate more dissociated
or independent motion of the thorax and pelvis. This particular metric has been used during gait
to distinguish participants with and without LBP through frontal and transverse plane trunk
coupling.
26,27
The hypothesis was tested using a two-way ANOVA with trunk coupling as the
outcome measure where, again, a significant interaction effect between group (control and rLBP)
and condition (STBal, DTCog, DTBal) would support the hypothesis. To investigate factors
influencing trunk control differences between groups and conditions, muscle activation data and
TRUNK CONTROL DURING DYNAMIC BALANCE 153

self-report measures were analyzed. Associations between trunk coupling and these outcomes
were tested with bivariate Pearson correlations. Statistical tests were performed with α=0.05 and
Bonferroni corrections for multiple comparisons. More details on all these measures can be
found in our previous work (Chapter III).
Results
Participants
Nineteen participants with rLBP and nineteen matched back-healthy control participants
with no history of LBP in the last year participated in the study (Table V.1). Participants with
rLBP were in symptom remission at the time of testing with an average VAS pain rating of
0.4±0.4 out of 10. Characterizing the group with self-report measures of task difficulty,
confidence, and attention demands revealed no differences between groups except that persons
with rLBP reported lower confidence in their ability to dual-task (t(18)=2.358, p=0.030). Note
one participant in each group is missing COM data due to marker occlusion during collection;
one participant with rLBP is missing MF and IO data due to failed fine-wire EMG insertion.
Table V.1. Participant demographics (mean ± standard deviation). BMI =
body mass index; VAS = visual analog scale; ODI = Oswestry Disability
Index.

Table 1
rLBP CNTRL p
Age (yrs) 23.5 ± 2.8 23.9 ± 3.3 0.679
Sex 7 M, 12 F 7 M, 12 F
Leg Dominance 18 R, 1 L 18 R, 1 L
Height (cm) 170.4 ± 8.4 169.1 ± 10.4 0.692
Weight (kg) 68.7 ± 10.3 67.1 ± 10.8 0.661
BMI 23.6 ± 2.4 23.3 ± 1.8 0.714
Baecke Physical Activity Scale (vector sum) 4.8 ± 1.0 4.9 ± 0.6 0.537
Episodes per year 3.4 ± 1.2
Pain during episodes (recall, VAS 0-10) 4.9 ± 2.2
Pain at time of testing (VAS 0-10) 0.4 ± 0.4
ODI (recall, %) 16.0 ± 18.7
PCS (0-52) 8.3 ± 8.5 6.5 ± 5.1 0.438
TSK (17-68) 31.3 ± 6.5 30.5 ± 6.0 0.706
FABQ (0-96) 20.2 ± 10.7
FABQ-PA (0-66) 12.2 ± 7.7
FABQ-W (0-30) 8.1 ± 6.7
Demographics of Participants
TRUNK CONTROL DURING DYNAMIC BALANCE 154

Task prioritization hypothesis
Using measures of task performance – dexterous force control RMSE, balance control
COM velocity, and cognitive task error variability – the hypothesis that persons with and without
rLBP prioritize dual-task demands differently was tested with a two-way ANOVA for each of
these outcome measures. A significant interaction effect for these tests would support the
hypothesis. While not the specific hypothesis being tested, a significant main effect of condition
in these tests would support the methodological assumption that the prioritization manipulation
worked as expected.
First, dexterous force control performance during STBal, DTCog, and STBal was
quantified with an RMSE from the target spring compression force. There was no interaction
effect (F(2,17)=0.489, p=0.622, 𝜂 𝑝 2
=0.054) or main effect of group (F(1,18)=0.072, p=0.791,
𝜂 𝑝 2
=0.004) on RMSE. There was a main effect of condition (F(2,17)=17.957, p<0.001,
𝜂 𝑝 2
=0.679), and all conditions were significantly different from one another (p≤0.034) (Figure
V.1A-B). Dexterous force control was modulated as predicted such that error increased from
STBal to DTCog and decreased when the priority was switched from DTCog to DTBal, though
not back to the level of single-task performance (STBal).
Second, balance control was quantified using COM average resultant velocity. There was
no interaction effect (F(2,15)=1.905, p=0.183, 𝜂 𝑝 2
=0.203) or main effect of group
(F(1,16)=0.019, p=0.891, 𝜂 𝑝 2
=0.019) on COM velocity. There was a main effect of condition
(F(2,15)=5.719, p=0.014, 𝜂 𝑝 2
=0.433), where there was significantly greater COM velocity in the
DTCog condition compared to STBal and DTBal conditions (p≤0.031). The lack of a significant
interaction effect here does not necessarily warrant testing effects within each group, but given
TRUNK CONTROL DURING DYNAMIC BALANCE 155

the effect size was larger than the other outcome measures (though still “small”
28
at 𝜂 𝑝 2
=0.203),
within-group between-condition tests are reported here. The same pattern observed when
collapsing across groups was true for the back-healthy control group alone where COM velocity
was greater in DTCog compared to STBal (Cohen’s d=0.482) and DTBal (Cohen’s d=0.262), but
there were no differences between conditions for the rLBP group (Cohen’s d≤0.132) (Figure
V.1C-D). These moderate pair-wise effect sizes for the back-healthy control group suggest
balance control was modulated with prioritization instruction only in the back-healthy control
group and not between-conditions in the rLBP group, but the study may have been
underpowered to detect this in the omnibus ANOVA.
Finally, variability of errors on the cognitive task was used to capture cognitive task
performance. Again, there was no interaction effect (F(2,17)=1.098, p=0.356, 𝜂 𝑝 2
=0.114) or main
effect of group (F(1,18)=0.001, p=0.973, 𝜂 𝑝 2
<0.001) on cognitive task error variability. There
was a main effect of condition (F(2,18)=6.521, p=0.008, 𝜂 𝑝 2
=0.434), and collapsing across
groups, there was significantly lower cognitive task error variability during DTCog condition
compared to STCog and DTBal conditions where cognitive task error variability was greater
(p≤0.044) (Figure V.1E-F). This did not support the hypothesis of different prioritization
between groups, but, of particular interest, revealed a facilitation effect of dual-tasking where
cognitive task error variability decreased from STCog to DTCog conditions. Error variability
increased again when the priority was switched back to balance (from DTCog to DTBal).
TRUNK CONTROL DURING DYNAMIC BALANCE 156

Figure V.1. Task performance measures for each group (back-healthy
controls [CNTRL]: A, C, E; persons with recurrent low back pain [rLBP]: B,
D, F) in each instruction condition. To quantify dexterous force control, root-
mean-squared-error (RMSE) (A, B) is presented. To quantify balance
control, center of mass (COM) average resultant velocity is shown (C, D). To
quantify cognitive task performance, error variability is presented (E, F).
*p<0.05
TRUNK CONTROL DURING DYNAMIC BALANCE 157

Trunk control hypothesis
The trunk control hypothesis was tested using frontal plane trunk coupling reported as the
coefficient of determination (R
2
) of a thorax-pelvis angle-angle plot. The hypothesis that the
rLBP group’s trunk coupling would be affected by dual-task interference while back-healthy
control trunk coupling would not be affected would be supported by a significant interaction
effect of the group-by-condition (STBal, DTCog, DTBal) ANOVA. For trunk coupling, there
was in fact a significant interaction effect (F(2,17)=6.904, p=0.006, 𝜂 𝑝 2
=0.448) but no main
effects of group (F(1,18)=1.713, p=0.207, 𝜂 𝑝 2
=0.087) or condition (F(2,17)=1.908, p=0.179,
𝜂 𝑝 2
=0.183). In the back-healthy control group there were no differences between conditions
(p≥0.453) (Figure V.2A). In the rLBP group, participants increased trunk coupling in both dual-
task interference conditions – DTCog (p=0.006) and DTBal (p=0.008) – compared to STBal,
supporting the hypothesis that dual-task inference regardless of prioritization instruction affects
trunk control in persons with rLBP (Figure V.2B). Looking in more detail, the groups were
significantly different during single-task performance such that the rLBP group had lower trunk
coupling, or more dissociated thorax and pelvis motion (p=0.024), and this was reported
previously (Chapter IV). The difference between groups that existed during the single-task
condition disappeared as the rLBP group increased trunk coupling from STBal to DTCog
(p=0.002). In summary, the rLBP group increased trunk coupling from STBal to both dual-task
inference conditions. Starting with more dissociated trunk motion compared to the control group
in the single-task condition, this increase brought the average trunk coupling in the rLBP group
back up to a level comparable with that observed in the back-healthy control group.

TRUNK CONTROL DURING DYNAMIC BALANCE 158

Figure V.2. Frontal plane trunk coupling, reported as a coefficient of
determination (R
2
) of a thorax-pelvis angle-angle plot for back-healthy
controls (A) and persons with rLBP (B). *p<0.05
Factors influencing trunk control
The goal of this exploration was to explain trunk coupling changes from STBal to DTCog
conditions in the group with rLBP from both kinesiological and psychological perspectives.
First, trunk muscle activation measures were investigated to explain changes in trunk coupling
from a kinesiological point of view. In STBal, the ratio of deep-to-superficial trunk muscle
activation amplitude – for paraspinals MF:ES and for abdominals IO:EO – was associated with
trunk coupling in the rLBP group, where greater relative deep muscle activation resulted in more
trunk coupling. These associations disappeared under conditions of dual-task interference
(DTCog and DTBal) suggesting they are mutable rather than fixed (Figure V.3). No single
muscle, muscle activation ratio, or combination of muscles predicted trunk coupling in any
conditions for the back-healthy control group or in DTCog or DTBal for the rLBP group.
Visualizing associations between the change in trunk coupling and the change in muscle
activation amplitudes from STBal to DTCog conditions revealed heterogeneity in the back-
healthy control group but a more consistent pattern for those with rLBP. First, from STBal to
TRUNK CONTROL DURING DYNAMIC BALANCE 159

DTCog, the majority of participants with rLBP increased or maintained trunk coupling (18/19
participants) while decreasing MF activation (14/18), ES activation (19/19), and IO activation
(17/18). Compare this to the back-healthy control group where more evenly distributed effects
were observed from STBal to DTCog – 9/19 increased or maintained trunk coupling; 11/19
decreased MF activation; 12/19 decreased ES activation; and 13/19 decreased IO activation
(Figure V.4). Second, excluding three participants with rLBP who exhibited a different pattern
than the rest of the group, a significant association between the change in ES activation and
change in trunk coupling from STBal to DTCog was observed. Here, greater decreases in ES
activation from STBal to DTCog resulted in greater increases in trunk coupling (R= -0.574,
p=0.020) (Figure V.4C). Summarizing, the rLBP increased trunk coupling from STBal to
DTCog while decreasing MF, ES, and IO activation amplitudes, and the amount of trunk
coupling increase was associated with the amount of ES decrease when three outliers were
removed.

TRUNK CONTROL DURING DYNAMIC BALANCE 160

Figure V.3. Associations between frontal plane trunk coupling (R
2
) and deep-
to-superficial trunk muscle activation ratios. The ratio for paraspinal
coordination is multifidus-to-erector spinae (MF:ES) (top row) and the ratio
for abdominal coordination is internal-to-external-oblique (IO:EO) (bottom
row). Conditions are the Balance-Dexterity single-task condition (STBal) (left
column), the dual-task condition with priority assigned to the cognitive task
(DTCog) (middle column) and the dual-task condition with priority assigned
to the balance task (DTBal) (right column).

TRUNK CONTROL DURING DYNAMIC BALANCE 161

Figure V.4. Associations between the change in average muscle activation
amplitudes for lumbar multifidus (MF) (A), lumbar erector spinae (ES) (C),
internal oblique (IO) (B), and external oblique (EO) (D) and the change in
frontal plane trunk coupling (R
2
) from single- to dual-task cognitive priority
conditions, where a positive ∆R
2
indicates a participant exhibited more
coupled trunk motion during the dual-task condition. Correlation shown in
(C) is after removing one participant with greatest decrease in ES activation
amplitude and two participants with greatest increases in trunk coupling.
Next, self-report measures of task difficulty and pain presentation were explored to help
elucidate a psychological mechanism for the observed kinesiologic changes. In the rLBP group,
trunk coupling was associated with self-report measures of cognitive task difficulty and
TRUNK CONTROL DURING DYNAMIC BALANCE 162

confidence under conditions of dual-task interference (both DTCog and DTBal). Persons who
rated the cognitive task as more difficult exhibited lower trunk coupling in both dual-task
conditions (DTCog: R= -0.512, p=0.025; DTBal: R= -0.522, p=0.022) compared to those who
rated the task as easier (Figure V.5). Persons who reported less confidence on the cognitive task
also exhibited lower trunk coupling (DTBal: R=0.476, p=0.040), and trends in the same direction
were observed for those who reported that the cognitive task required more attention (DTCog:
R= -0.416, p=0.076; DTBal: R= -0.438, p=0.061). Persons with rLBP who reported higher
cognitive task difficulty, lower task confidence, and greater task attention demands exhibited
lower trunk coupling during dual-task interference conditions.
In addition, the change in trunk coupling from STBal to DTCog in the rLBP group was
associated with both cognitive task difficulty rating and participants’ recalled pain during a
typical painful episode (Figure V.6). Those who reported less difficulty with the cognitive task or
less pain during a typical LBP episode had the greatest increases in trunk coupling from STBal to
DTCog. Stated with cause and effect switched, which is possible for this association and
discussed later, persons who increased trunk coupling the most from STBal to DTCog perceived
the cognitive task as easier and remembered their episodes of LBP as less painful. Note that one
and two participants, respectively, seemed to follow a different pattern and were excluded from
the correlation analysis. Of particular note, these two participants exhibited the greatest increases
in trunk coupling from STBal to DTCog and two of the highest reported recalled pain levels.
These were in the same subset of three outliers who were removed from the ES analysis
described previously, and limitations to interpretation are discussed later.
TRUNK CONTROL DURING DYNAMIC BALANCE 163

Figure V.5. Associations between trunk coupling (R
2
) and self-report
measure of cognitive task difficulty for the Balance-Dexterity single-task
condition (A), the dual-task condition with priority assigned to the cognitive
task (B) and the dual-task condition with priority assigned to the balance
task (C).

Figure V.6. Associations between the change in frontal plane trunk coupling
from single- to dual-task cognitive priority conditions and self-report
cognitive task difficulty (A) and recalled pain during a symptomatic episode
reported on a visual analog scale (VAS) (B). Correlation statistics are shown
for the full group as well as without one participant in (A) and two in (B) that
seem to follow a different pattern.
TRUNK CONTROL DURING DYNAMIC BALANCE 164

Discussion
This study tested two hypotheses about the effect of a history of rLBP – (1) the ability to
modulate task performance in accord with prioritization instructions would be impaired in
persons with rLBP compared to back-healthy control participants, and (2) trunk control strategies
in persons with rLBP would rely more on cognitive resources, and therefore under a dual-task
interference condition, trunk coupling would be altered in this group compared to back-healthy
control participants. The first hypothesis was not supported, as there were no significant
interaction effects on any of the three task performance outcome measures. The second
hypothesis was supported by a significant interaction effect of group-by-condition on trunk
coupling. Persons with rLBP increased frontal plane trunk coupling from the single-task to the
dual-task conditions, independent of priority instruction, while there was no effect of dual-task
interference on trunk coupling in the back-healthy control group. No muscle activation levels
alone predicted trunk coupling in any condition within each group, but those with greater deep-
to-superficial trunk muscle activation ratios had greater trunk coupling in the rLBP group in the
single-task condition. The change in trunk coupling in this group from STBal to DTCog was
associated with the change in ES activation amplitude, self-reported cognitive task difficulty, and
recalled pain during a typical symptomatic episode (Figure V.7).
TRUNK CONTROL DURING DYNAMIC BALANCE 165

Figure V.7. Summary figure reporting effect sizes (ES) for statistically
significant comparisons of means and coefficients of determination (R
2
) for
statistically significant associations. Arrows oriented down indicate that a
decrease in the given measure from single- to dual-task conditions was
associated with greater increases in trunk coupling for the group in
remission from recurrent low back pain (rLBP).
Confirmation of prioritization manipulation plus an unexpected facilitation effect
First, the methodological assumptions behind the prioritization manipulation were
confirmed by a significant main effect of condition on all three task performance outcome
measures. Collapsing across groups, dexterous force control error (quantified by RMSE of the
vertical compression of the spring) increased from STBal to DTCog, then decreased again when
the priority was switched (DTBal). The same was true for balance control (quantified by average
resultant COM velocity). Cognitive task error variability, meant to capture cognitive task
performance, was greater when the priority was placed on balance (DTBal) compared to when
the priority was placed on the cognitive task (DTCog). In all these conditions, performance on a
TRUNK CONTROL DURING DYNAMIC BALANCE 166

given task was worse in the DT condition where that condition was not prioritized compared to
where it was prioritized or compared to the single-task condition.
Notably absent from the comparisons summarized in the previous paragraph was the
STCog condition, conducted in a seated posture. Interestingly, a dual-task facilitation effect was
observed when comparing STCog to DTCog where cognitive task error variability decreased
when the concurrent Balance-Dexterity Task was added. Dual-task facilitation was not expected,
but has been observed in previous posture-cognition dual-task research in healthy adults,
9,29

persons with chronic LBP,
30
and persons with Parkinson’s disease.
31
Since it is known that not
only pain,
32
but even simply the threat of pain
33
interrupts attention, it is possible that persons
with rLBP experienced this interruption in the STCog condition but were distracted from it under
the increased attentional load of dual-tasking. That does not, however, explain the facilitation
effect for the back-healthy control group. Recent work by Huang et al. supports a mechanism for
posture-cognition dual-task facilitation that may explain the facilitation effect observed in both
groups in the current study.
29
Researchers analyzed EEG signals collected while healthy
participants completed a force-matching task during two different stance conditions – level
ground and a stabilometer. Facilitation could be explained from observed changes in EEG
measures from the simplier to the more difficult postural task through three primary neural
mechanisms – “(1) increased effectiveness of information transfer, (2) an anterior shift of
processing resources toward frontal executive function, and (3) cortical dissociation of control
hubs in the parietal-occipital cortex for neural economy.”
29
These mechanisms explain how
adding postural demands can induce more efficient and effective neural communication utilizing
more anterior frontal executive function and result in observed dual-task facilitation.

TRUNK CONTROL DURING DYNAMIC BALANCE 167

Comparing task performance and prioritization manipulation between groups
There were no main effects of group on any of the three task performance outcome
measures, which indicates that persons with or without rLBP did not perform the two elements of
the Balance-Dexterity Task – dexterous force control or balance – or the cognitive task
differently. This allows us to interpret differences observed in trunk coupling between groups, a
variable controlled independently of task performance measures, (Chapter III) as an effect of a
history of LBP and dual-task interference and not simply an effect of different task performance.
The task prioritization hypothesis tested here, which stated the ability to modulate task
performance in accord with prioritization instructions would be impaired in persons with rLBP
compared to back-healthy control participants, was not supported by the data. There were no
significant interaction effects of group-by-condition on any of the three task performance
outcomes measures. The interaction effect for balance control (quantified by average resultant
COM velocity), however, had the largest effect size ( 𝜂 𝑝 2
=0.203) of the variables tested. It is
possible the current study was underpowered to detect a statistically significant interaction effect
for this measure, especially because the study was powered pre-hoc based on trunk coupling
measures, not COM velocity specifically. Although a significant interaction effect on balance
control was not detected, the larger effect sizes in pair-wise between-condition tests for the back-
healthy control group compared to effect sizes between-conditions in the rLBP group would
suggest persons with rLBP have less ability or are choosing not to modulate balance control in
conditions of different attention demands and priority instructions. A general theory of posture-
cognition prioritization has been proposed where weighting of these factors during dual-task
performance is dependent on the motor and cognitive state during a task performance, postural
reserve and the compensatory capabilities of the individual, and individual characteristics like
TRUNK CONTROL DURING DYNAMIC BALANCE 168

personality, affect, and training.
34
In persons with high postural reserve and conditions of low
threat of falling (e.g. the back-healthy controls in the present study) cognitive tasks can be
prioritized and more postural instability can be tolerated. But a change in any of these factors
(e.g. injury or dysfunctional postural control, like those suffering from rLBP) can induce a shift
toward postural task prioritization. A significant interaction effect was seen in a dual-task gait
study on persons with chronic LBP
30
and may be confirmed in a follow-up study here with more
participants as persons with rLBP may maintain COM velocity in all conditions while back-
healthy control participants modulated COM velocity between conditions.
Trunk control differences between groups
The trunk control hypothesis, which stated that trunk control strategies in persons with
rLBP would rely more on cognitive resources, and therefore under a dual-task interference
condition, trunk coupling would be altered in this group compared to back-healthy control
participants, was supported by the current study. There was a significant interaction effect of
group-by-condition on trunk coupling, where the rLBP group exhibited lower trunk coupling
compared to back-healthy control participants in STBal but then increased trunk coupling from
STBal to DTCog and DTBal. The more general attentional load therefore, not the specific
priority instruction, was what appeared to drive this more tightly coupled trunk motion seen in
the dual-task compared to the single-task conditions. This general, not specific, effect was further
supported by the finding that self-report measures of cognitive task difficulty predicted trunk
coupling in the rLBP group in both DTCog and DTBal conditions. The lack of specific responses
for different priority manipulations is a novel finding for a population of minimally disabled
(89% of participants with rLBP had ODI scores in the “minimally disabled” range
19
) persons in
remission from rLBP. Back-healthy control participants, in contrast, showed no differences in
TRUNK CONTROL DURING DYNAMIC BALANCE 169

trunk coupling between conditions indicating trunk coupling was not influenced by attentional
load or priority manipulation in this group.
The increase in trunk coupling in the rLBP group from STBal to DTCog could be
interpreted as a benefit, since it “corrected” the reduced trunk coupling observed in this
population during STBal compared to back-healthy control participants. This conclusion is
premature, however, until we explore potential mechanisms – both kinesiological and
psychological – for this increase. Kinesiologically, the present study and previous reports
(Chapter IV) state that in single-task performance, persons with rLBP exhibit trunk coupling
proportional to deep-to-superficial trunk muscle activation ratios where those who have greater
ratios exhibit greater trunk coupling. In dual-task conditions, this relationship disappeared, and in
fact, most participants decreased deep trunk muscle (MF and IO) activation amplitudes from
what was observed in the single-task condition. So, the increased trunk coupling from STBal to
DTCog cannot be explained by increased deep trunk muscle activity. Looking at changes in ES
activity, however, a trend is identified where participants with rLBP who decreased ES
activation more from STBal to DTCog also increased trunk coupling more. Synthesizing these
findings leads to the following conclusion. Persons with rLBP exhibited more dissociated trunk
motion compared to back-healthy control participants during submaximal dynamic balance,
driven in part by higher ES activation levels (actually lower MF:ES ratios). From STBal to
DTCog conditions, reductions in ES activation were associated with increases in trunk coupling,
bringing the average trunk coupling of the rLBP up to the average level of back-healthy controls.
Given that these findings – more ES activation associated with lower trunk coupling in STBal
and reduced ES activations associated with increased trunk coupling from STBal to DTCog – are
in agreement and help to tell a complete story, we interpret increased trunk coupling from STBal
TRUNK CONTROL DURING DYNAMIC BALANCE 170

to DTCog in persons with rLBP as a benefit driven by a reduction in overactive superficial
paraspinal activity between these conditions. But we emphasize that our conclusions are specific
only to this novel Balance-Dexterity Task (Chapter IV) as greater trunk coupling and associated
trunk stiffness in other externally-perturbed tasks is thought to contribute to the recurrence of
pain.
35,36
These other frameworks, however, explain the link between greater trunk stiffness and
pain recurrence through greater muscle activity and lumbar loading. Here, increases in trunk
coupling from STBal to DTCog were associated with reductions in superficial paraspinal
activity, not increases, which further supports the interpretation of increased trunk coupling as a
benefit. The next step is to explore why some persons with rLBP reaped this benefit while others
did not through psychological science-informed mechanisms.
Self-report measures of cognitive task difficulty and pain during a typical symptomatic
episode begin to explain this heterogeneity. Persons with rLBP who thought the cognitive task
was more difficult maintained lower trunk coupling during dual-task performance compared to
those who reported the task as easier. In addition, self-reported cognitive task difficulty was
associated with the change in trunk coupling where those who thought the task was easier
increased trunk coupling the most from STBal to DTCog. Also, those who recalled high pain
levels during a typical symptomatic episode did not reap the benefit of increased trunk coupling,
while those who reported lower pain recall increased trunk coupling more from STBal to
DTCog. These findings can be explained using two conceptual frameworks.
In terms of the movement-specific attentional reinvestment framework, the benefit of
increased trunk coupling under dual-task interference is credited to the distraction effect of the
cognitive task. Theories of movement-specific attentional reinvestment predict that memory of
recurrent episodes of pain and a fear of pain recurrence cause more attentional resources and
TRUNK CONTROL DURING DYNAMIC BALANCE 171

conscious motor processing to be devoted to movement.
13
While movement-specific attentional
reinvestment has not been studied in this population, we know that COM displacements in
response to platform perturbations in persons with rLBP are associated with pain-related fear and
activity interference as well as EEG measures of postural monitoring,
15
and trunk stiffness in
response to a discrete trunk release is associated with higher scores on a fear of movement
scale.
37
In patients with symptomatic LBP, those with high pain catastrophizing had greater trunk
muscle activity in the lumbar ES and MF compared to patients with low pain catastrophizing.
38

Persons in the present study with rLBP who reported the cognitive task as less difficult may have
experienced a lower attentional load. The effect of this low attentional load was to interfere with
conscious motor processing induced by fear or memory of pain in the trunk, and this reduced ES
activation amplitude from STBal to DTCog and increased trunk coupling up to the level of
controls. The reduction in ES activation associated with more tightly coupled trunk motion is
interpreted as a result of interfering with these psychological processes. For persons with rLBP
who experienced the cognitive task as a large attentional load, however, there was little change
in trunk coupling from STBal to DTCog. Potentially, the difficult cognitive task induced its own
psychological stress which replaced, as opposed to distracted from, conscious motor processing.
Also, persons who remembered their previous episodes as more painful did not reap the benefit,
suggesting conscious motor processing induced by more painful episodes may be more resistant
to distraction.
The cause and effect for these associations between trunk coupling and self-report task
difficulty and pain could be reversed where, instead of the relationship described above where
perceived task difficulty and pain influenced the response of dual-task interference, the response
to dual-task interference could have influenced perceived task difficulty and pain. Reversing
TRUNK CONTROL DURING DYNAMIC BALANCE 172

cause and effect for these associations can be justified with an action-specific perception
framework. In the movement-specific attentional reinvestment framework, there was an
underlying assumption that participants accurately and objectively assessed the difficulty of the
cognitive task and the intensity of their previous painful episodes. Recent work in the field of
action-perception, however, acknowledges a two-way relationship between these phenomena
with Jessica Witt writing, “people perceive the surrounding environment in terms of their ability
to act in it.”
39
One study on persons with LBP and leg pain found these individuals perceived
equivalent walking distances as further than their pain-free counterparts perceived them,
supporting this two-way action-perception relationship in this population.
40
Flipping cause and
effect in Figure V.6 suggests persons with rLBP who increased trunk coupling from STBal to
DTCog perceived the cognitive task as easier and recalled their episodes as less painful. Again, if
we accept that increasing trunk coupling from STBal to DTCog is an appropriate and beneficial
response to the increased challenge of cognitive dual-tasking, this explains why persons who
increased trunk coupling reported these more favorable self-report measures. Perceiving the
cognitive task as easier and perceiving less pain, or at least less remembered pain, could
contribute to continued increasing of trunk coupling under challenging posture-cognition
conditions for some and getting “stuck in a rut” for others, which suggests a role of
reinforcement learning in the persistence of these motor control patterns during periods of
symptom remission. In addition to motor control training to address trunk control, interventions
that train dual-tasking ability, cognitive behavioral interventions that address pain perception and
coping, or the use of biofeedback to educate patients about their motor behavior may help to
decouple these phenomena.

TRUNK CONTROL DURING DYNAMIC BALANCE 173

Limitations and future directions
The present findings add clarity to the dual-task and rLBP literature and suggest
improvements for future work. Cognitive tasks should be chosen such that they can be scaled or
calibrated to account for individual variations in education level, working memory capacity, or
other factors. This would allow a uniform attentional load to be delivered to all participants.
Also, by the nature of needing to design a cognitive task without real-time verbalization, the
measure of cognitive task performance was likely not very sensitive. This would explain why
variable error showed significant differences but not absolute error. We encourage future
investigations to develop novel dual-task methodology that can avoid real-time verbalization for
trunk control research since verbalization itself influences abdominal and diaphragm activation,
intra-abdominal pressure, and therefore trunk control. In addition, the cognitive task was
designed to interfere with verbal working memory resources because of the assumption that
these resources are utilized more in conditions of pain or injury.
13
Findings and conclusions may
be different if cognitive tasks designed to interfere with visuo-spatial attentional resources were
utilized instead.
Future work should be designed to test the two frameworks discussed. The first could be
tested using scales that assess conscious motor processing like the Movement-Specific
Reinvestment Scale.
13
Careful application of cognitive tasks with low and high attentional loads
by calibrating attentional load and intentionally delivering varied loads to the participants could
also test the distraction effects of mild cognitive tasks and the potentially stress-inducing effects
of more difficult tasks. The influence of action-specific perception could be tested by using more
objective measures of task difficulty, such as a math ability test or psychophysiological
TRUNK CONTROL DURING DYNAMIC BALANCE 174

techniques like skin conductance, heart rate variability, or EEG measures, to help disambiguate
objective task demands and perceived task demands.
Participants made up a convenience sample of persons with and without rLBP recruited
from student groups, classes, flyers, and university-affiliated physical therapy clinics, and the
clinical population was generally young (age 23.5±2.8yrs), minimally disabled (ODI
16.0±18.7%), and all in pain remission at the time of testing (0.4±0.4 out of 10 on VAS). Future
research could include different LBP presentations and larger samples. The usual limitations of
EMG methodology apply here as well including the potential for cross-talk from surface EMG
and potential errors in fine-wire placement. Insertions were done, however, under ultrasound
guidance and confirmed through electrical stimulation so substantial errors are unlikely.
The associations between change in trunk coupling and change in ES activation from
STBal to DTCog, cognitive task difficulty, and pain recall were only statistically significant after
removing one, two, or three of the same subset of three participants who did not follow a
consistent pattern. Two of these participants stood out by exhibiting the greatest increases in
trunk coupling from STBal to DTCog and two of largest reported recalled pain intensity. Persons
suffering from chronic pain have many motivators influencing VAS pain ratings including
insurance and worker’s compensation coverage, access and trust in healthcare services, and other
psychological factors, some of which may explain these outliers.
41
We did not have a clinical
reason to exclude these participants, but the heterogeneity and presence of subgroups in a LBP
population specifically are well-known,
18,42
and applying findings from laboratory research on
persons with rLBP in the clinic should always be done with caution and with subject-specificity
in mind. This is why we decided to present the findings and acknowledge the limitation that
there may be other strategies at work in subgroups of persons with rLBP. The greater variability
TRUNK CONTROL DURING DYNAMIC BALANCE 175

in this group may have also resulted in reduced power to detect significant differences in some
measures. Larger studies may allow for subgroup analyses to help explain what different or
additional mechanisms are at work in these three persons with rLBP.
Conclusions
While dual-tasking with the Balance-Dexterity Task and a working memory recall and
manipulation task, persons with and without rLBP modulated task performance in accord with
prioritization manipulation and there were no significant differences between groups in task
performance modulation. Almost all participants with rLBP increased frontal plane trunk
coupling from single-task to dual-task conditions. Increases in trunk coupling from single- to
dual-task conditions were associated with decreases in ES activation, with lower cognitive task
difficulty ratings, and with lower recalled pain in most participants. Two potential theoretical
frameworks explain these findings. Using a movement-specific attentional reinvestment
perspective, low attentional loads interfered with memory-of-pain-related conscious processing
of posture and resulted in decreased ES activation and improved trunk coupling when the dual-
task was added compared to the single-task condition, while higher attentional loads potentially
induced their own psychological stress. Using an action-specific perception perspective, persons
who increased trunk coupling from single- to dual-task conditions perceived the cognitive task as
easier and their episodes as less painful, suggesting a role of reinforcement learning in the
persistence of motor control patterns during periods of symptom remission.

TRUNK CONTROL DURING DYNAMIC BALANCE 176

References
1. George SZ, Fritz JM, Bialosky JE, Donald DA. The effect of a fear-avoidance-based
physical therapy intervention for patients with acute low back pain: results of a
randomized clinical trial. Spine (Phila Pa 1976). 2003;28(23):2551-2560.
doi:10.1097/01.BRS.0000096677.84605.A2.
2. Riipinen M, Niemistö L, Lindgren KA, Hurri H. Psychosocial differences as predictors for
recovery from chronic low back pain following manipulation, stabilizing exercises and
physician consultation or physician consultation alone. J Rehabil Med. 2005;37(3):152-
158. doi:10.1080/16501970410022426.
3. Fraizer E, Mitra S. Methodological and interpretive issues in posture-cognition dual-
tasking in upright stance. Gait Posture. 2008;27(2):271-279.
doi:10.1016/j.gaitpost.2007.04.002.
4. Hodges PW. Changes in motor planning of feedforward postural responses of the trunk
muscles in low back pain. Exp Brain Res. 2001;141(2):261-266.
doi:10.1007/s002210100873.
5. Van Daele U, Hagman F, Truijen S, Vorlat P, Van Gheluwe B, Vaes P. Decrease in
postural sway and trunk stiffness during cognitive dual-task in nonspecific chronic low
back pain patients, performance compared to healthy control subjects. Spine (Phila Pa
1976). 2010;35(5):583-589. doi:10.1097/BRS.0b013e3181b4fe4d.
6. Hamacher D, Hamacher D, Schega L. A cognitive dual task affects gait variability in
patients suffering from chronic low back pain. Exp Brain Res. 2014;232(11):3509-3513.
doi:10.1007/s00221-014-4039-1.
TRUNK CONTROL DURING DYNAMIC BALANCE 177

7. Hemmati L, Rojhani-Shirazi Z, Malek-Hoseini H, Mobaraki I. Evaluation of Static and
Dynamic Balance Tests in Single and Dual Task Conditions in Participants With
Nonspecific Chronic Low Back Pain. J Chiropr Med. 2017;16(3):189-194.
doi:10.1016/j.jcm.2017.06.001.
8. Etemadi Y, Salavati M, Arab AM, Ghanavati T. Balance recovery reactions in individuals
with recurrent nonspecific low back pain: Effect of attention. Gait Posture. 2016;44:123-
127. doi:10.1016/j.gaitpost.2015.11.017.
9. Smith JA, Gordon J, Kulig K. The influence of divided attention on walking turns: Effects
on gait control in young adults with and without a history of low back pain. Gait Posture.
2017;58(September):498-503. doi:10.1016/j.gaitpost.2017.09.019.
10. Goh H-T, Sullivan KJ, Gordon J, Wulf G, Winstein CJ. Dual-task practice enhances motor
learning: a preliminary investigation. Exp Brain Res. 2012;222(3):201-210.
doi:10.1007/s00221-012-3206-5.
11. Goh H-T, Gordon J, Sullivan KJ, Winstein CJ. Evaluation of attentional demands during
motor learning: validity of a dual-task probe paradigm. J Mot Behav. 2014;46(August
2014):95-105. doi:10.1080/00222895.2013.868337.
12. Wickens CD. Multiple resources and performance prediction. Theor Issues Ergon Sci.
2002;3(2):159-177. doi:10.1080/14639220210123806.
13. Masters R, Maxwell J. The theory of reinvestment. Int Rev Sport Exerc Psychol.
2008;1(2):160-183. doi:10.1080/17509840802287218.
14. Roelofs J, Peters ML, Fassaert T, Vlaeyen JWS. The role of fear of movement and injury
TRUNK CONTROL DURING DYNAMIC BALANCE 178

in selective attentional processing in patients with chronic low back pain: A dot-probe
evaluation. J Pain. 2005;6(5):294-300. doi:10.1016/j.jpain.2004.12.011.
15. Jacobs J V., Roy CL, Hitt JR, Popov RE, Henry SM. Neural mechanisms and functional
correlates of altered postural responses to perturbed standing balance with chronic low
back pain. Neuroscience. 2016;339:511-524. doi:10.1016/j.neuroscience.2016.10.032.
16. Buszard T, Farrow D, Zhu FF, Masters RSW. The relationship between working memory
capacity and cortical activity during performance of a novel motor task. Psychol Sport
Exerc. 2016;22:247-254. doi:10.1016/j.psychsport.2015.07.005.
17. Deyo RA, Dworkin SF, Amtmann D, et al. Report of the NIH Task Force on research
standards for chronic low back pain. J Pain. 2014;15(6):569-585.
doi:10.1016/j.jpain.2014.03.005.
18. Norton G, McDonough CM, Cabral HJ, Shwartz M, Burgess JF. Classification of patients
with incident non-specific low back pain: implications for research. Spine J. 2015:1-10.
doi:10.1016/j.spinee.2015.08.015.
19. Fairbank JC, Pynsent PB. The Oswestry Disability Index. Spine (Phila Pa 1976).
2000;25(22):2940-52; discussion 2952. doi:10.1097/00007632-200011150-00017.
20. Carlsson AM. Assessment of chronic pain. I. Aspects of the reliability and validity of the
visual analogue scale. Pain. 1983;16(1):87-101. doi:10.1016/0304-3959(83)90088-X.
21. Hermens H, Freriks B, Merletti R, et al. The Surface ElectroMyoGraphy for the Non-
Invasive Assessment of Muscles Project (SENIAM). seniam.org. Published 2006.
22. Perotto AO, Delagi EF, Iazzetti J, Morrison D. Anatomical Guide for the
TRUNK CONTROL DURING DYNAMIC BALANCE 179

Electromyographer. 5th ed. Springfield, Illinois: Charles C Thomas Publisher, LTD.;
2011.
23. De Troyer A, Estenne M, Ninane V, Gansbeke D Van, Gorini M. Transversus abdominis
muscle function in humans. J Appl Physiol. 1990;68(3):1010-1016.
24. Schmidt RA, Lee TD. Motor Control and Learning: A Behavioral Emphasis. 4th ed.
Champaign, IL: Human Kinetics; 2005.
25. Chiviacowsky S, Campos T, Domingues MR. Reduced frequency of knowledge of results
enhances learning in persons with Parkinson’s disease. Front Psychol. 2010;1(DEC):1-6.
doi:10.3389/fpsyg.2010.00226.
26. van den Hoorn W, Bruijn SM, Meijer OG, Hodges PW, van Dieen JH. Mechanical
coupling between transverse plane pelvis and thorax rotations during gait is higher in
people with low back pain. J Biomech. 2012;45(2):342-347.
doi:10.1016/j.jbiomech.2011.10.024.
27. Crosbie J, de Faria Negrão Filho R, Nascimento DP, Ferreira P. Coordination of Spinal
Motion in the Transverse and Frontal Planes During Walking in People With and Without
Recurrent Low Back Pain. Spine (Phila Pa 1976). 2013;38(5):E286-E292.
doi:10.1097/BRS.0b013e318281de28.
28. Cohen J. Statistical Power Analysis for the Behavioral Sciences. Second. New Jersey:
Lawrence Erlbaum Associates; 1988.
29. Huang C-Y, Chang G-C, Tsai Y-Y, Hwang I-S. An Increase in Postural Load Facilitates
an Anterior Shift of Processing Resources to Frontal Executive Function in a Postural-
TRUNK CONTROL DURING DYNAMIC BALANCE 180

Suprapostural Task. Front Hum Neurosci. 2016;10(August):1-16.
doi:10.3389/fnhum.2016.00420.
30. Lamoth CJC, Stins JF, Pont M, Kerckhoff F, Beek PJ. Effects of attention on the control
of locomotion in individuals with chronic low back pain. J Neuroeng Rehabil. 2008;5:1-8.
doi:10.1186/1743-0003-5-13.
31. Hazamy AA, Altmann LJP, Stegemöller E, et al. Improved cognition while cycling in
Parkinson’s disease patients and healthy adults. Brain Cogn. 2017;113:23-31.
doi:10.1016/j.bandc.2017.01.002.
32. Moore DJ, Keogh E, Eccleston C. The interruptive effect of pain on attention. Q J Exp
Psychol. 2012;65(3):565-586. doi:10.1080/17470218.2011.626865.
33. Moore DJ, Keogh E, Eccleston C. The effect of threat on attentional interruption by pain.
Pain. 2013;154(1):82-88. doi:10.1016/j.pain.2012.09.009.
34. Yogev-Seligmann G, Hausdorff JM, Giladi N. Do we always prioritize balance when
walking? Towards an integrated model of task prioritization. Mov Disord.
2012;27(6):765-770. doi:10.1002/mds.24963.
35. Hodges P, van den Hoorn W, Dawson A, Cholewicki J. Changes in the mechanical
properties of the trunk in low back pain may be associated with recurrence. J Biomech.
2009;42(1):61-66. doi:10.1016/j.jbiomech.2008.10.001.
36. Hodges PW, Tucker K. Moving differently in pain: A new theory to explain the adaptation
to pain. Pain. 2011;152(SUPPL.3):S90-S98. doi:10.1016/j.pain.2010.10.020.
37. Karayannis N V, Smeets RJEM, van den Hoorn W, Hodges PW. Fear of Movement Is
TRUNK CONTROL DURING DYNAMIC BALANCE 181

Related to Trunk Stiffness in Low Back Pain. PLoS One. 2013;8(6).
doi:10.1371/journal.pone.0067779.
38. Pakzad M, Fung J, Preuss R. Pain catastrophizing and trunk muscle activation during
walking in patients with chronic low back pain. Gait Posture. 2016;49:73-77.
doi:10.1016/j.gaitpost.2016.06.025.
39. Witt JK. Action’s effect on perception. Curr Dir Psychol Sci. 2011;20(3):201-206.
doi:10.1177/0963721411408770.
40. Witt JK, Linkenauger SA, Bakdash JZ, Augustyn JS, Cook A, Proffitt DR. The long road
of pain: Chronic pain increases perceived distance. Exp Brain Res. 2009;192(1):145-148.
doi:10.1007/s00221-008-1594-3.
41. Reynolds Losin EA, Anderson SR, Wager TD. Feelings of Clinician-Patient Similarity
and Trust Influence Pain: Evidence From Simulated Clinical Interactions. J Pain.
2017;18(7):787-799. doi:10.1016/j.jpain.2017.02.428.
42. Mistry D, Patel S, Hee SW, Stallard N, Underwood M. Evaluating the quality of subgroup
analyses in randomized controlled trials of therapist-delivered interventions for
nonspecific low back pain: a systematic review. Spine (Phila Pa 1976). 2014;39(7):618-
629. doi:10.1097/BRS.0000000000000231.

TRUNK CONTROL DURING DYNAMIC BALANCE 182

CHAPTER VI
THE ROLE OF MOVEMENT-SPECIFIC REINVESTMENT IN DYNAMIC BALANCE
IN PERSONS WITH AND WITHOUT RECURRENT LOW BACK PAIN
Abstract
The purpose was to investigate the role of movement-specific attention reinvestment in
task performance and trunk coordination during dynamic balance in persons with and without
recurrent low back pain (rLBP). It was hypothesized persons who invest more attention in their
movement would exhibit greater trunk coupling within the framework of “locking down”
degrees-of-freedom influencing balance control, and cognitive dual-tasking would interrupt this
relationship. Persons with and without rLBP (n=19 per group) completed the Balance-Dexterity
Task, which involved balancing on one limb while compressing an unstable spring with the
other, with and without a cognitive task utilizing verbal working memory. Task performance
measures included center-of-mass velocity and vertical force variability under the spring. Trunk
coupling was quantified with the coefficient of determination (R
2
) of an angle-angle plot of
thorax-pelvis frontal plane motion. Psychometrics collected included movement-specific
reinvestment scale (MSRS), fear-avoidance beliefs questionnaire (FABQ), pain catastrophizing
scale, and Tampa scale for kinesiophobia. In the control group, persons who invested more
attention in their movements exhibited greater trunk coupling (R=0.647) and experienced greater
reductions in trunk coupling under dual-task interference (R=-0.537). No associations between
social-cognitive factors and trunk coupling were observed in persons with rLBP. Those who
scored higher on MSRS, however, had lower dexterous force variability (R=0.532), and those
who scored higher on FABQ had lower center-of-mass velocity (R=-0.534). These relationships
also disappeared under dual-task interference. Persons with rLBP modulated variables more
TRUNK CONTROL DURING DYNAMIC BALANCE 183

directly related to task performance and while control participants modulated trunk coupling
based on social-cognitive factors. And, for both groups, dual-tasking interfered with these
relationships.
Introduction
Postural control is influenced by perturbations to cognition and attention, which indicates
a role of cognitive processing in maintaining upright posture.
1,2
A general theory of posture-
cognition prioritization has been proposed where weighting of these factors is dependent on the
motor and cognitive state during a specific dual-task situation; postural reserve and the
compensatory capabilities of the individual; and individual characteristics like personality, affect,
and training.
3
In a support surface translation perturbation, psychometrics capturing fear of pain
and activity were associated with center of mass (COM) displacements, and additional
associations with specific electroencephalography (EEG) sites linked to cognitive postural
monitoring suggested a mechanism for this association.
4
This mechanistic link between posture
and cognition can be explained by Movement-Specific Reinvestment Theory, which introduces
movement-specific attentional investment as the “manipulation of conscious, explicit, rule-based
knowledge, by working memory, to control the mechanics of one’s movements during motor
output.”
5
Here, more movement-specific reinvestment could lead to this increased postural
monitoring captured by EEG and associated with COM displacements, and other studies have
supported this framework with EEG-based evidence.
6–8

Reinvestment can be induced by a number of factors including personality traits,
9

changes in performance conditions or equipment,
10
high-pressure or stressful performance
environments,
6
and, most relevant to the current study, conditions of pain or injury. A
psychometric tool that measures the movement-specific reinvestment construct is the Movement-
TRUNK CONTROL DURING DYNAMIC BALANCE 184

Specific Reinvestment Scale (MSRS).
5
The scale has been used successfully in a number of
patient populations providing evidence not only for its theoretical strength but its practical utility
in populations of older adults who survived a stroke,
11
persons with Parkinson’s disease,
12
and
fallers.
13
Only a few studies have investigated associations between movement-specific
reinvestment and posture. Giving an internal focus cue for a postural sway task increased EEG
coherence measures associated with verbal processing in young adults.
8
One study reported
associations between scores on the MSRS and postural sway measures during double-limb
standing in young adults but not in older adults.
14
Findings supported the framework’s prediction
that greater attention investment in movement “locks down” degrees of freedom which, in this
case, led to greater sway.
5
This association was not present in the dual-task interference
condition where a tone-counting task was done concurrently, supporting the idea that these
cognitive processes are perturbed under dual-task interference.
One population in which an investigation into cognitive contributions to posture is
prudent is persons suffering from recurrent episodes of low back pain (rLBP). The annual
prevalence of activity-limiting LBP has been reported at 38%,
15
and in a study of US health care
spending in 2013, low back and neck pain accounted for $87.6 billion – the highest spending
level for any musculoskeletal condition.
16
Unfortunately, despite decades of research, spending
on LBP and neck pain has increased the most of any condition studied, excluding diabetes, in the
last eighteen years. Dual-task interference paradigms have been used in research and
rehabilitation of persons with rLBP both because postural dual-tasking is ecological and because
there are specific information processing deficits, especially in trunk control, that these studies
probe.
17,18
(Chapter V)
TRUNK CONTROL DURING DYNAMIC BALANCE 185

While the influence of movement-specific reinvestment on trunk control has not been
investigated, some groups have reported findings using other social-cognitive factors that help us
form testable hypotheses. Karayannis, et al. identified persons with greater fear of movement
scores exhibited more trunk stiffness during a semi-seated discrete trunk release task.
19
During
gait, patients with chronic LBP and high pain catastrophizing increased trunk muscle activity
more so than patients with low pain catastrophizing.
20
Finally, the influence of expected pain
appears to be region-specific since noxious stimuli presented to the elbow before a unilateral arm
raise did not change trunk muscle activation, while stimuli presented to the low back resulted in
delayed trunk muscle activations.
21
While these findings have been important in supporting a
biopsychosocial model of rLBP, there are substantial gaps in the literature about how social-
cognitive factors in persons with rLBP in symptom remission interact with the role of the trunk
in standing balance.
The purpose of the study was to investigate the role of movement-specific attention
reinvestment in task performance and trunk coordination during the Balance-Dexterity Task
(Chapter III). It was hypothesized that no associations would exist between task performance
measures and movement-specific attention reinvestment, but that persons who invest more
attention in their movement will exhibit greater trunk coupling within the framework of “locking
down” degrees of freedom influencing balance control. Persons with rLBP would be at one end
of this spectrum with higher MSRS scores and greater trunk coupling in line with a “trunk
stiffening strategy”. It was also hypothesized that cognitive dual-task interference would
interrupt this relationship.

TRUNK CONTROL DURING DYNAMIC BALANCE 186

Methods
Participants and instrumentation
Participants with rLBP and matched back-healthy control participants with no history of
LBP in the past year
22
were recruited for the study, per Institutional Review Board approval and
with informed consent. Participants with rLBP had at least two episodes of pain per year for at
least one year, but experienced pain less than half of the days in the previous six months
(distinguishes chronicity and recurrence
23
). Painful episodes were severe enough to limit
function based on questions in the NIH Task Force recommended minimum dataset
23
and
Oswestry Disability Index.
24
Participants were in symptom remission at the time of testing and
for the preceding seven days (pain <1.5 out of 10 on a visual analog scale (VAS)
25
). A sample
size of nineteen per group was determined from a pre-hoc power analysis after four pilot subjects
in each group were collected. Participants were instrumented with a full-body retroreflective
marker set, and motion data were captured with an 11-camera Qualisys Oqus System
(Gothenburg, Sweden; 250Hz). Force data were captured with two Advanced Medical
Technology Inc. force plates (Watertown, MA; 3000Hz). Self-report information about
participants’ pain presentation was collected including reporting episodes of pain per year, pain
during a typical episode, and pain at time of testing on a VAS. Psychometrics were collected
including the movement-specific reinvestment scale (MSRS),
5
fear-avoidance beliefs
questionnaire (FABQ),
26
pain catastrophizing scale (PCS),
27
and Tampa scale for kinesiophobia
(TSK).
28

TRUNK CONTROL DURING DYNAMIC BALANCE 187

Procedures
The Balance-Dexterity Task device and procedures were described previously (Chapter
III). The Balance-Dexterity Task uses a custom device with a spring mounted between two
boards (Compression Spring #805, Century Spring Corp., Commerce, CA). Participants were
shown real-time feedback of the vertical force under the spring, and instructed: “While standing
on one leg, compress this spring so that the line is first as high, then as stable as possible”
(Figure III.1). After one familiarization trial and five practice trials each lasting 20-25s, the mean
of the middle 50% of the last three practice trials was used to calculate an individual’s
reproducible, submaximal compression. Next, participants were introduced to the cognitive task
which was a modified verbal working memory recall and manipulation task requiring
participants remember a list of five random digits from zero to ten, do an arithmetic operation to
those digits twice during the trial, and remember five answers. After practice of these tasks,
participants used a VAS to report how difficult each task was, how confident they were they
could complete each task successfully, how much attention each task required, and finally how
confident they were they could do both tasks concurrently. In a randomized order, participants
completed the following task conditions – five trials of the cognitive task while seated
(“cognitive single-task”); five trials where a dotted line indicating their reproducible,
submaximal compression was shown as a goal with the instructions, “While standing on one leg,
compress this spring so that the line is as stable as possible directly over the dotted goal line”
(“Balance-Dexterity Task”); five dual-task trials with a cognitive priority where the instructions
given were, “In this trial, it is most important that you get the five numbers correct, and it is less
important the line is as stable as possible” (“dual-task cognitive priority”); five dual-task trials
with a balance priority where the instructions given were, “In this trial, it is most important that
TRUNK CONTROL DURING DYNAMIC BALANCE 188

you keep the line as stable as possible, and it is less important that you get the five numbers
correct” (“dual-task balance priority”); and, finally, three trials where the spring was replaced
with a stable block of the same height with the same instructions as the Balance-Dexterity Task
condition.
Data analysis
The middle 50% of the task was analyzed. Kinematic and force plate data were low-pass
filtered with cutoff frequencies of 12Hz and 50Hz, respectively, using a dual-pass 4
th
order
Butterworth filter. Measures of task performance were collected including average center of
mass (COM) resultant velocity to quantify balance performance and coefficient of variation (CV)
of the vertical force produced under the spring to quantify dexterous force control. Trunk control
was quantified by tracking thorax and pelvis motion relative to global coordinates. Using an
angle-angle plot of thorax and pelvis frontal plane rotation, a coefficient of determination (R
2
)
was calculated where a high R
2
would indicate highly coupled thorax and pelvis motion and a
low R
2
would indicate more dissociated or independent motion of the thorax and pelvis. This
metric has been used during gait to distinguish participants with and without LBP through frontal
and transverse plane trunk coupling.
29,30
More details on these measures can be found in our
previous work (Chapter III). Associations between outcome measures and psychometric scores
were tested with bivariate Pearson correlations with α=0.05 for all tests (PASW Statistics, IBM
Corp., Armonk, NY).
Results
Participants
Nineteen participants with rLBP and nineteen matched back-healthy control participants
with no history of LBP in the last year participated in the study (Table VI.1). Participants with
TRUNK CONTROL DURING DYNAMIC BALANCE 189

rLBP were in symptom remission at the time of testing with an average VAS pain rating of
0.4±0.4 out of 10. There were no differences between groups on any of the psychometric scores
collected. Movement-specific attention reinvestment was associated with fear-avoidance beliefs
in persons with rLBP (R=0.563, p=0.012) but was not associated with any other social-cognitive
measure. Note, moving forward, that one participant in each group is missing COM data due to
marker occlusion during collection.
Table VI.1. Participant demographics (mean ± standard deviation). BMI =
body mass index; VAS = visual analog scale; ODI = Oswestry Disability
Index; MSRS = movement-specific reinvestment scale; CMP = conscious
motor processing subset; MSC = movement self-consciousness subset; PCS =
pain catastrophizing scale; TSK = Tampa scale for kinesiophobia; FABQ =
fear-avoidance belief questionnaire; PA = physical activity subset; W = work
subset.

Table 1
rLBP CNTRL p
Age (yrs) 23.5 ± 2.8 23.9 ± 3.3 0.679
Sex 7 M, 12 F 7 M, 12 F
Leg Dominance 18 R, 1 L 18 R, 1 L
Height (cm) 170.4 ± 8.4 169.1 ± 10.4 0.692
Weight (kg) 68.7 ± 10.3 67.1 ± 10.8 0.661
BMI 23.6 ± 2.4 23.3 ± 1.8 0.714
Baecke Physical Activity Scale (vector sum) 4.8 ± 1.0 4.9 ± 0.6 0.537
Episodes per year 3.4 ± 1.2
Pain during episodes (recall, VAS 0-10) 4.9 ± 2.2
Pain at time of testing (VAS 0-10) 0.4 ± 0.4
ODI (recall, %) 16.0 ± 18.7
MSRS (0-50) 25.6 ± 4.8 24.8 ± 9.7 0.793
CMP (0-25) 14.9 ± 4.8 13.8 ± 4.5 0.484
MSC (0-25) 10.7 ± 5.5 10.9 ± 5.9 0.891
PCS (0-52) 8.3 ± 8.5 6.5 ± 5.1 0.438
TSK (17-68) 31.3 ± 6.5 30.5 ± 6.0 0.706
FABQ (0-96) 20.2 ± 10.7
FABQ-PA (0-66) 12.2 ± 7.7
FABQ-W (0-30) 8.1 ± 6.7
Demographics of Participants
TRUNK CONTROL DURING DYNAMIC BALANCE 190

Task performance
Associations between psychometric scores and measures of task performance – balance
control COM velocity, dexterous force control CV, and cognitive task error variability – were
examined. COM velocity was associated with FABQ in persons with rLBP such that persons
who had more fear-avoidance beliefs had lower COM velocity which indicates more
conservative control of postural sway. This association was strong in single-task (R=-0.534,
p=0.022) and dual-task balance priority (R=-0.513, p=0.029) conditions and retained marginal
significance in the dual-task cognitive priority condition (R=-0.421, p=0.082) (Figure VI.1).
Dexterous force control CV was associated with MSRS in persons with rLBP such that persons
who devoted more attention to their movements had lower variability of the vertical force. This
association was strong in single-task (R=-0.532, p=0.019) and dual-task balance priority (R=-
0.470, p=0.042) conditions but was not present when the priority was placed on the cognitive
task (R=-0.073, p=0.767) (Figure VI.2). Cognitive task error variability was associated with
MSRS in persons with rLBP only in the dual-task balance-priority condition where persons who
scored higher on MSRS had lower cognitive task error (R=-0.514, p=0.024) (Figure VI.3). In
summary, no associations between social-cognitive factors and task performance measures were
found for back-healthy control participants, while persons with rLBP modulated task
performance based on these factors. Balance control was associated with fear-avoidance beliefs,
and both dexterous force control and cognitive task performance were associated with
movement-specific attention reinvestment. These relationships weakened or disappeared under
dual-task interference with priority on the cognitive task.

TRUNK CONTROL DURING DYNAMIC BALANCE 191

Figure VI.1. Associations between center of mass (COM) average resultant
velocity and movement-specific attentional reinvestment (A-C) and fear-
avoidance beliefs questionnaire (D-F) in the single-task Balance-Dexterity
Task (A,D), the dual-task condition with the priority on the cognitive task
(B,E), and the dual-task condition with the priority on balance (C,F).
TRUNK CONTROL DURING DYNAMIC BALANCE 192

Figure VI.2. Associations between dexterous force control coefficient of
variation (CV) and movement-specific attentional reinvestment (A-C) and
fear-avoidance beliefs questionnaire (D-F) in the single-task Balance-
Dexterity Task (A,D), the dual-task condition with the priority on the
cognitive task (B,E), and the dual-task condition with the priority on balance
(C,F).
TRUNK CONTROL DURING DYNAMIC BALANCE 193

Figure VI.3. Associations between cognitive task error variability and
movement-specific attentional reinvestment (A-C) and fear-avoidance beliefs
questionnaire (D-F) in the cognitive single-task condition (A,D), the dual-task
condition with the priority on the cognitive task (B,E), and the dual-task
condition with the priority on balance (C,F).
Trunk coordination
Frontal plane trunk coupling was examined as a control strategy variable that could be
altered independent of task performance (Chapter III). In back-healthy control participants,
persons who invested more attention in their movement exhibited more tightly coupled thorax
and pelvis motion (R=0.647, p=0.003), but this relationship disappeared under both conditions of
dual-task interference (Figure VI.4). In addition, MSRS score was associated with the change in
trunk coupling from single- to dual-task cognitive priority conditions (R=-0.537, p=0.018)
(Figure VI.5) where persons who invested more attention in their movements decreased trunk
coupling when dual-task interference was introduced. Persons who did not invest as much
attention in their movement increased or did not change their trunk coupling. In persons with
rLBP, there were no associations between MSRS and trunk coupling or between MSRS and
TRUNK CONTROL DURING DYNAMIC BALANCE 194

change in trunk coupling. In summary, MSRS had no influence on trunk coupling in persons
with rLBP, but an association existed for back-healthy controls where persons who devoted more
attention to their movements had greater trunk coupling in the single-task condition and
decreased trunk coupling more under dual-task interference.

Figure VI.4. Associations between movement-specific attention reinvestment
and trunk coupling in single-task Balance Dexterity Task (A), dual-task
cognitive priority (B), and balance priority (C) conditions.

Figure VI.5. Association between movement-specific attention reinvestment
and the change in trunk coupling from single- to dual-task cognitive priority
conditions.
TRUNK CONTROL DURING DYNAMIC BALANCE 195

Discussion
Findings supported hypotheses for back-healthy control participants but did not support
hypotheses for persons in remission from rLBP. In the control group, persons who invested more
attention in their movements exhibited greater trunk coupling and experienced greater reductions
in trunk coupling under dual-task interference. In conditions of dual-task interference, the
relationship between MSRS and trunk coupling disappeared. We expected persons with rLBP to
exhibit the same relationship but to be at one end of the continuum. Instead, there were no
associations between social-cognitive factors and trunk coupling in persons with rLBP. In this
group, however, persons who scored higher on the MSRS had lower dexterous force control CV
and cognitive task error variability, and those who scored higher on FABQ had lower average
COM velocity. These relationships also disappeared under dual-task interference when the
priority was placed on the cognitive task. Overall, persons with rLBP modulated variables more
directly related to task performance and balance based on social-cognitive factors while control
participants modulated trunk control based on these factors. And, for both groups, dual-task
interference interfered with these relationships.
Back-healthy control participants
Back-healthy control participants behaved in a way consistent with Movement-Specific
Reinvestment Theory. Investing more conscious attention in movement results in “locking
down” of degrees of freedom and increasing joint stiffness.
5,31,32
In a series of studies where
postural threat was induced in a group of healthy young adults by having persons stand at a
height, persons increased their scores on a modified MSRS
31
and exhibited increased ankle joint
stiffness.
32
In another study on reinvestment and posture, greater scores on MSRS were
associated with greater center of pressure variability and reduced center of pressure complexity
TRUNK CONTROL DURING DYNAMIC BALANCE 196

(“more constrained,” quantified using a sample entropy method).
14
These associations also
disappeared in a dual-task interference condition with a concurrent tone-counting task. The
authors explain that a greater tendency to invest conscious attention to movement reduces
movement degrees of freedom that contribute to center of pressure control, resulting in greater
sway and a less complex trajectory, and under the dual-task condition “secondary task loading
reduces resources available for conscious control.”
14
While studies investigating associations
between MSRS and postural control are sparse, the narrative is consistent that greater attentional
investment in movement reduces movement degrees of freedom, and this relationship breaks
down under dual-task interference.
Our findings agree with these published investigations into reinvestment with some
caveats. We observed an influence of MSRS not on measures directly related to task
performance or postural control but on trunk coupling, which is a movement variable modulated
independently of COM velocity, dexterous force CV, and other performance measures in this
task (Chapter III). The direction of the relationship was as expected – persons who invest more
attention in movement exhibited more coupled thorax and pelvis frontal plane motion or less
independent motion of these trunk segments. Associations between movement and movement-
specific reinvestment in healthy young adults were only seen in movement variables that could
be controlled independently of task performance, and we recommend this strategy when
investigating effects of movement-specific reinvestment in future work. When dual-task
interference was added, persons who invest more attention in movement had greater reductions
in trunk coupling or, in other words, allowed more independent motion of these trunk segments.
We agree with interpretations by Uiga et. al.
14
that secondary task loading reduced cognitive
resources available for movement-specific reinvestment allowing more automatic control to
TRUNK CONTROL DURING DYNAMIC BALANCE 197

emerge, but only in persons who had higher baseline movement-specific attention investment
tendencies.
Persons in remission from rLBP
The story was different for persons with a history of rLBP. For this group, there were no
associations between MSRS and trunk coupling. Combining these findings with reports that
older adults both who had and had not fallen previously did not show an association between
MSRS and posture while younger adults did
14
suggests that groups affected by pathology, pain,
or aging show altered relationships between social-cognitive factors and motor control. In our
previous studies, we proposed a testable hypothesis that could explain trunk coupling changes
under dual-task interference in persons with rLBP (Chapter V). We hypothesized movement-
specific reinvestment would be associated with trunk coupling at baseline but would be
interrupted under dual-task interference. This turned out to be true for the control group, but not
for persons with rLBP, meaning this hypothesis was not supported.
There were, however, associations between MSRS and dexterous force control variability
and between FABQ and average COM velocity in persons with rLBP during the single-task
condition and the dual-task interference condition where the priority was placed on balance.
Those who scored higher on FABQ or MSRS had lower COM velocity and dexterous force CV,
respectively. Interpreting lower COM velocity as prioritizing balance
33
and lower dexterous
force CV as reducing variability of the perturbations from the unstable spring (Chapter III), those
with rLBP who think about their movements more or have more fear-avoidance beliefs exhibited
more tightly controlled balance and dexterous force control. Other research supports this
interpretation including a study by Jacobs et al. reporting greater EEG measures of postural
monitoring were associated with less COM displacement in response to a platform perturbation.
4

TRUNK CONTROL DURING DYNAMIC BALANCE 198

Our previous work on task performance under dual-task interference suggested that persons with
rLBP prioritize posture by maintaining low COM velocity in all conditions while control
participants tolerated an increased COM velocity under dual-task interference (Chapter V). This
is partially explained here by the presence of these associations between social-cognitive factors
and task performance for the rLBP only. This group’s attention reinvestment is associated with
task performance, while in control participants, attention investment may be free to influence
trunk control instead.
Previous work has suggested links between trunk control and social-cognitive factors in
this population. Karayannis et. al. reported an association between trunk stiffness and
kinesiophobia in a discrete trunk release perturbation and significantly greater trunk stiffness in
those who scored higher on the FABQ.
19
Based on findings here, we propose this association
was present in the Karayannis study because trunk stiffness was a direct performance outcome of
a seated trunk release with the pelvis fixed. In tasks like the Balance-Dexterity Task where other
task performance variables more directly influence postural control or balance, these other
variables were associated with MSRS and FABQ while trunk coupling was not. The pattern
supported here, which should be tested in future research, is that persons with a history of rLBP
invest attention in a way that modulates task performance while back-healthy control participants
invest attention in a way that modulates movement degrees of freedom indirectly related to
postural control or balance.
It must be noted, however, that conflicting evidence exists in the literature. Unperturbed
standing balance has less frequently shown associations with psychometric scores. Hooper et. al.
reported no association between Y-Balance score and psychometric scores.
34
Mazaheri et. al.
reported increased sway in persons with chronic and rLBP but no association with kinesiophobia
TRUNK CONTROL DURING DYNAMIC BALANCE 199

or pain catastrophizing.
35
Sung et. al. reported no association between self-reported fear of
movement or pain characteristics and postural sway during unstable sitting.
36
The important
differences from the present study, however, are two-fold – (1) balance demands were quantified
here from COM movement which may more closely reflect balance control as opposed to center
of pressure trajectories, and (2) it appears social-cognitive factors may play a more important
role in challenging or perturbed balance tasks than in quiet standing, given that the Balance-
Dexterity Task involves continuous control of a dynamic unstable force under the spring
(Chapter III). We believe these key features of the present study help explain conflicting findings
reported in the literature, and we recommend careful consideration of motor task demands in
future studies.
This influence of attention on task performance may or may not be an appropriate or
beneficial response to pain or injury. Frameworks of LBP recurrence suggest that motor control
adaptions in the short-term may benefit healing and reduce tissue injury risk, but anatomical or
structural lesions are not always present, and these adaptations could be contributing to the
recurrence or persistence of symptoms in the long-term.
37
It makes sense that investing attention
in direct contributors to postural instability and in how to prioritize cognitive and postural
demands may be beneficial for persons with impairments in automatic control, as it is known that
persons with rLBP have impairments in trunk proprioception
38,39
and processing sensory signals
related to balance.
40
From this perspective, closer investment of attention may be necessary to
successfully control balance and posture, or may be necessary in a certain subgroup of persons
with greater neurophysiological impairments. On the other hand, the persistence of this
phenomenon during periods of symptom remission may be an indicator of maladaptive motor
control and cognitive processing strategies and could explain the more dissociated trunk coupling
TRUNK CONTROL DURING DYNAMIC BALANCE 200

in this group during single-task performance (Chapter IV). Future longitudinal work
investigating which of these factors best predict pain recurrence and including a large enough
sample size to identify predictive factors and subgroups is necessary.
Limitations and future directions
The present study contributes to areas of postural control, rLBP research, and movement-
specific reinvestment constructs but has limitations and suggests improvements for future work.
The choice of cognitive task may have a large effect on findings. For the present study, a verbal
working memory cognitive task was chosen specifically to cause the greatest interference with
movement-specific reinvestment, which utilizes verbal working memory resources.
6,7,10,41
Using
visuo-spatial working memory tasks will interfere with different aspects of cognition and motor
control and could be used to answer different research questions.
42
Also, the difficulty of the
cognitive task may influence findings. Researchers recommend scaling or calibrating cognitive
tasks to account for individual variations in working memory capacity, education, or other
factors. Objective effects of dual-task interference on cognition could be more physiologically
tested with tools like EEG
4,8
or skin conductance.
Participants in this study were a convenience sample of persons with and without rLBP
recruited from student groups, classes, flyers, and university-affiliated physical therapy clinics.
The clinical population was young (age 23.5±2.8yrs), minimally disabled due to LBP (ODI
16.0±18.7%), and all in pain remission at the time of testing (0.4±0.4 out of 10 on VAS).
Including different LBP presentations and larger samples in future research could help broaden
the interpretation and application of these findings. The heterogeneity and presence of subgroups
in a LBP population are well-known.
22,43
In the present study through a visual inspection of
Figure VI.4, one might differentiate seven participants with rLBP who seem to lie right along the
TRUNK CONTROL DURING DYNAMIC BALANCE 201

line-of-best-fit seen in the back-healthy control group from twelve participants with rLBP who
have lower trunk coupling than their MSRS score would predict. We investigated these potential
subgroups but did not detect any differences between groups for any demographics, pain
presentation measures, or task performance measures, except that dexterous force CV was
significantly lower in the subgroup that did not follow the pattern (p=0.029) suggesting a
difference in how they controlled the instability of the spring. Given that this subgroup division
created small sample sizes, it is possible that we were underpowered to detect real differences,
and a larger sample with more robust subgroup analysis techniques is recommended. As always,
applying findings from laboratory research on persons with rLBP in the clinic should always be
done with caution and with subject-specificity in mind.
Conclusions
While back-healthy control participants modulated trunk control based on their tendency
to invest attention in movement, persons in remission from rLBP instead modulated task
performance associated with movement-specific attention reinvestment and fear-avoidance
beliefs. These associations were no longer present under conditions of cognitive dual-task
interference. More research needs to be done in order to understand links between these findings
and the recurrence of LBP.

TRUNK CONTROL DURING DYNAMIC BALANCE 202

References
1. Fraizer E, Mitra S. Methodological and interpretive issues in posture-cognition dual-
tasking in upright stance. Gait Posture. 2008;27(2):271-279.
doi:10.1016/j.gaitpost.2007.04.002.
2. Dault MC, Frank JS, Allard F. Influence of a visuo-spatial, verbal and central executive
working memory task on postural control. Gait Posture. 2001;14(2):110-116.
doi:10.1016/S0966-6362(01)00113-8.
3. Yogev-Seligmann G, Hausdorff JM, Giladi N. Do we always prioritize balance when
walking? Towards an integrated model of task prioritization. Mov Disord.
2012;27(6):765-770. doi:10.1002/mds.24963.
4. Jacobs J V., Roy CL, Hitt JR, Popov RE, Henry SM. Neural mechanisms and functional
correlates of altered postural responses to perturbed standing balance with chronic low
back pain. Neuroscience. 2016;339:511-524. doi:10.1016/j.neuroscience.2016.10.032.
5. Masters R, Maxwell J. The theory of reinvestment. Int Rev Sport Exerc Psychol.
2008;1(2):160-183. doi:10.1080/17509840802287218.
6. Buszard T, Farrow D, Zhu FF, Masters RSW. The relationship between working memory
capacity and cortical activity during performance of a novel motor task. Psychol Sport
Exerc. 2016;22:247-254. doi:10.1016/j.psychsport.2015.07.005.
7. Buszard T, Farrow D, Kemp J. Examining the influence of acute instructional approaches
on the decision-making performance of experienced team field sport players. J Sports Sci.
2013;31(3):238-247. doi:10.1080/02640414.2012.731516.
TRUNK CONTROL DURING DYNAMIC BALANCE 203

8. Ellmers TJ, Machado G, Wong TWL, Zhu F, Williams AM, Young WR. A validation of
neural co-activation as a measure of attentional focus in a postural task. Gait Posture.
2016;50:229-231. doi:10.1016/j.gaitpost.2016.09.001.
9. Laborde S, Musculus L, Kalicinski M, Klämpfl MK, Kinrade NP, Lobinger BH.
Reinvestment: Examining convergent, discriminant, and criterion validity using
psychometric and behavioral measures. Pers Individ Dif. 2015;78:77-87.
doi:10.1016/j.paid.2015.01.020.
10. Beilock SL, Wierenga SA, Carr TH. Expertise, attention, and memory in sensorimotor
skill execution: impact of novel task constraints on dual-task performance and episodic
memory. Q J Exp Psychol. 2002;55(4):1211-1240. doi:10.1080/02724980244000170.
11. Orrell AJ, Masters RSW, Eves FF. Reinvestment and Movement Disruption Following
Stroke. 2009:177-183.
12. Masters RSW, Pall HS, MacMahon KMA, Eves FF. Duration of Parkinson disease is
associated with an increased propensity for “reinvestment”. Neurorehabil Neural Repair.
2007;21(2):123-126. doi:10.1177/1545968306290728.
13. Uiga L, Capio CM, Wong TWLL, Wilson MR, Masters RSWW. Movement specific
reinvestment and allocation of attention by older adults during walking. Cogn Process.
2015;16(1):421-424. doi:10.1007/s10339-015-0685-x.
14. Uiga L, Capio CM, Ryu D, Wilson MR, Masters RSW. The role of conscious control in
maintaining stable posture. Hum Mov Sci. 2017;(May):0-1.
doi:10.1016/j.humov.2017.10.008.
TRUNK CONTROL DURING DYNAMIC BALANCE 204

15. Hoy D, Bain C, Williams G, et al. A systematic review of the global prevalence of low
back pain. Arthritis Rheum. 2012;64(6):2028-2037. doi:10.1002/art.34347.
16. Dieleman JL, Baral R, Birger M, et al. US Spending on Personal Health Care and Public
Health, 1996-2013. Jama. 2016;316(24):2627. doi:10.1001/jama.2016.16885.
17. Etemadi Y, Salavati M, Arab AM, Ghanavati T. Balance recovery reactions in individuals
with recurrent nonspecific low back pain: Effect of attention. Gait Posture. 2016;44:123-
127. doi:10.1016/j.gaitpost.2015.11.017.
18. Smith JA, Gordon J, Kulig K. The influence of divided attention on walking turns: Effects
on gait control in young adults with and without a history of low back pain. Gait Posture.
2017;58(September):498-503. doi:10.1016/j.gaitpost.2017.09.019.
19. Karayannis N V, Smeets RJEM, van den Hoorn W, Hodges PW. Fear of Movement Is
Related to Trunk Stiffness in Low Back Pain. PLoS One. 2013;8(6).
doi:10.1371/journal.pone.0067779.
20. Pakzad M, Fung J, Preuss R. Pain catastrophizing and trunk muscle activation during
walking in patients with chronic low back pain. Gait Posture. 2016;49:73-77.
doi:10.1016/j.gaitpost.2016.06.025.
21. Moseley GL, Nicholas MK, Hodges PW. Does anticipation of back pain predispose to
back trouble? Brain. 2004;127(10):2339-2347. doi:10.1093/brain/awh248.
22. Norton G, McDonough CM, Cabral HJ, Shwartz M, Burgess JF. Classification of patients
with incident non-specific low back pain: implications for research. Spine J. 2015:1-10.
doi:10.1016/j.spinee.2015.08.015.
TRUNK CONTROL DURING DYNAMIC BALANCE 205

23. Deyo RA, Dworkin SF, Amtmann D, et al. Report of the NIH Task Force on research
standards for chronic low back pain. J Pain. 2014;15(6):569-585.
doi:10.1016/j.jpain.2014.03.005.
24. Fairbank JC, Pynsent PB. The Oswestry Disability Index. Spine (Phila Pa 1976).
2000;25(22):2940-52; discussion 2952. doi:10.1097/00007632-200011150-00017.
25. Carlsson AM. Assessment of chronic pain. I. Aspects of the reliability and validity of the
visual analogue scale. Pain. 1983;16(1):87-101. doi:10.1016/0304-3959(83)90088-X.
26. Waddell G, Newton M, Henderson I, Somerville D, Main CJ. A Fear-Avoidance Beliefs
Questionnaire (FABQ) and the role of fear-avoidance beliefs in chronic low back pain and
disability. Pain. 1993;52(2):157-168. doi:10.1016/0304-3959(93)90127-B.
27. Sullivan M, Bishop S, Pivik J, Plvik J. The pain catastrophizing scale: development and
validation. Psychol Assess. 1995;7(4):432-524. doi:10.1037/1040-3590.7.4.524.
28. Vlaeyen JWS, Kole-Snijders AMJ, Rotteveel AM, Ruesink R, Heuts PHTG. The role of
fear of movement/(re) injury in pain disability. J Occup Rehabil. 1995;5(4):235-252.
29. van den Hoorn W, Bruijn SM, Meijer OG, Hodges PW, van Dieen JH. Mechanical
coupling between transverse plane pelvis and thorax rotations during gait is higher in
people with low back pain. J Biomech. 2012;45(2):342-347.
doi:10.1016/j.jbiomech.2011.10.024.
30. Crosbie J, de Faria Negrão Filho R, Nascimento DP, Ferreira P. Coordination of Spinal
Motion in the Transverse and Frontal Planes During Walking in People With and Without
Recurrent Low Back Pain. Spine (Phila Pa 1976). 2013;38(5):E286-E292.
TRUNK CONTROL DURING DYNAMIC BALANCE 206

doi:10.1097/BRS.0b013e318281de28.
31. Huffman JL, Horslen BC, Carpenter MG, Adkin AL. Does increased postural threat lead
to more conscious control of posture? Gait Posture. 2009;30(4):528-532.
doi:10.1016/j.gaitpost.2009.08.001.
32. Carpenter MG, Frank JS, Silcher CP, Peysar GW. The influence of postural threat on the
control of upright stance. Exp Brain Res. 2001;138(2):210-218.
doi:10.1007/s002210100681.
33. Winter DA. Human balance and posture control during standing and walking. Gait
Posture. 1995;3(4):193-214.
34. Hooper TL, James CR, Brismée JM, et al. Dynamic balance as measured by the Y-
Balance Test is reduced in individuals with low back pain: A cross-sectional comparative
study. Phys Ther Sport. 2016;22:29-34. doi:10.1016/j.ptsp.2016.04.006.
35. Mazaheri M, Heidari E, Mostamand J, Negahban H, van Dieën JH. Competing Effects of
Pain and Fear of Pain on Postural Control in Low Back Pain? Spine (Phila Pa 1976).
2014;39(25):E1518-E1523. doi:10.1097/BRS.0000000000000605.
36. Sung W, Abraham M, Plastaras C, Silfies SP. Trunk motor control deficits in acute and
subacute low back pain are not associated with pain or fear of movement. Spine J.
2015;15(8):1772-1782. doi:10.1016/j.spinee.2015.04.010.
37. Hodges PW, Tucker K. Moving differently in pain: A new theory to explain the adaptation
to pain. Pain. 2011;152(SUPPL.3):S90-S98. doi:10.1016/j.pain.2010.10.020.
38. Osthoff A-KR, Ernst MJ, Rast FM, et al. Measuring Lumbar Reposition Accuracy in
TRUNK CONTROL DURING DYNAMIC BALANCE 207

Patients With Unspecific Low Back Pain: Systematic Review and Meta-analysis. Spine
(Phila Pa 1976). 2015;40(2):E97-E111. doi:10.1097/BRS.0000000000000677.
39. Lee AS, Cholewicki J, Reeves NP, Zazulak BT, Mysliwiec LW. Comparison of trunk
proprioception between patients with low back pain and healthy controls. Arch Phys Med
Rehabil. 2010;91(9):1327-1331. doi:10.1016/j.apmr.2010.06.004.
40. Brumagne S, Janssens L, Knapen S, Claeys K, Suuden-Johanson E. Persons with recurrent
low back pain exhibit a rigid postural control strategy. Eur Spine J. 2008;17(9):1177-
1184. doi:10.1007/s00586-008-0709-7.
41. Maxwell JP, Masters RS, Eves FF. From novice to no know-how: a longitudinal study of
implicit motor learning. J Sports Sci. 2000;18(October 2012):111-120.
doi:10.1080/026404100365180.
42. Wickens CD. Multiple resources and performance prediction. Theor Issues Ergon Sci.
2002;3(2):159-177. doi:10.1080/14639220210123806.
43. Mistry D, Patel S, Hee SW, Stallard N, Underwood M. Evaluating the quality of subgroup
analyses in randomized controlled trials of therapist-delivered interventions for
nonspecific low back pain: a systematic review. Spine (Phila Pa 1976). 2014;39(7):618-
629. doi:10.1097/BRS.0000000000000231
TRUNK CONTROL DURING DYNAMIC BALANCE 208

CHAPTER VII
SUMMARY AND CONCLUSIONS

Figure VII.1. Summary figure reporting effect sizes (ES) for statistically
significant comparisons of means and coefficients of determination (R
2
) for
statistically significant associations. Arrows oriented downward indicate that
a decrease in the given measure was associated with greater increases in
trunk coupling for the group in remission from recurrent low back pain
(rLBP).
Multiple diverse and redundant interacting systems contributing to postural control allow
successful human movement. By probing individual systems during challenging perturbations to
posture and balance we can learn about roles and functions of these contributing components.
Chapter II laid out a literature-informed justification and rationale for the hypotheses tested in
these dissertation studies and important aspects of study designed including development of the
Balance-Dexterity Task and the cognitive dual-task paradigm. In addition, a justification for
studying a clinical population in remission from recurrent episodes of low back pain (rLBP) as a
model dysfunctional postural control system was presented.
TRUNK CONTROL DURING DYNAMIC BALANCE 209

Chapter III presented a careful characterization of the Balance-Dexterity Task and
features of the task relevant to postural control and trunk coordination. The Balance-Dexterity
Task is a submaximal, continuous balance task useful for clinical and basic research in human
postural control. The task invokes larger center of pressure (COP) velocities than double- and
single-limb stance and more variable dexterous force control compared to a stable block
condition. Variability in dexterous force control was associated with variability in balance
control. This finding in a concurrent bipedal lower-extremity task aligns with similar work in
bimanual control tasks. Trunk coordination was quantified with a frontal plane trunk coupling
measure. Trunk coupling was correlated with the percent of time in in-phase coupling and
oppositely correlated with the percent of time in anti-phase coupling, justifying use of this
simpler measure to capture trunk coordination. Segment and joint excursions, however, seemed
to quantify different aspects of trunk movement. Trunk coupling was not associated with any of
the balance or dexterous force control demand measures, suggesting it was modulated
independent of task performance. Muscle activation levels were greater in the Balance-Dexterity
Task compared to the stable block condition, but no muscle activations alone or in ratios were
associated with trunk coupling R
2
. The paucity of associations between individual muscle
activations and trunk coupling indicates that these non-disabled persons have redundant motor
control processes available to control balance, dexterous force control, and trunk coupling. When
signals were normalized to maximum voluntary isometric contractions, however, a greater
multifidus-to-erector spinae ratio was weakly associated with higher trunk coupling R
2

suggesting greater relative deep paraspinal activation drove greater trunk coupling in this task. In
conclusion, the Balance-Dexterity Task is a challenging, continuous dynamic balance task from
which many fruitful measures of movement and motor control processes may be observed.
TRUNK CONTROL DURING DYNAMIC BALANCE 210

Chapter IV compared Balance-Dexterity Task performance and trunk control measures
between persons with and without rLBP. All participants were able to successfully complete the
Balance-Dexterity Task with no differences between groups in reproducible, submaximal
compression force goal, COP velocity, or dexterous force control. In both groups, frontal plane
trunk coupling varied independently of any task performance measure. Persons with rLBP had,
on average, lower trunk coupling, meaning more dissociated thorax and pelvis motion, compared
to back-healthy control participants. And, in this group, trunk coupling was associated with deep-
to-superficial paraspinal and abdominal EMG ratios, where greater deep muscle activation
relative to more superficial muscles resulted in higher trunk coupling.
The finding of more dissociated thorax and pelvis motion was counter to the hypothesis
of increased coupling in the clinical population expected through adoption of a stiffening
strategy. This unsupported hypothesis was built on investigations of discrete perturbations to
posture where greater trunk co-contraction and stiffness has been observed in persons with rLBP.
These perturbations, however, involved delivering external perturbing forces either to a support
surface or directly to the trunk, which may invoke a stiffening strategy related to fear of
movement or pain. In the present studies, there was no such association between trunk coupling
and any psychometric measure of fear of pain or movement. Taking this young, minimally
disabled population in pain remission with this submaximal, low-range-of-motion, volitionally-
driven unstable balance task, it is likely we are observing thorax and pelvis dissociation in a low-
effort task as opposed to the trunk stiffening seen in many external perturbation paradigms or
higher-effort tasks. This harkens back to a framework proposed by Cholewicki and McGill
where causes of LBP in low-effort tasks are related to instability and in high-effort tasks are
more related to tissue failure.
1
The Balance-Dexterity Task serves as an unstable balance task to
TRUNK CONTROL DURING DYNAMIC BALANCE 211

observe dissociated trunk motion in this population at one end of this continuum of task
demands.
Lower trunk coupling was associated with lower deep-to-superficial paraspinal and
abdominal muscle activation ratios in the rLBP group. These findings support previous
mechanical modeling studies that suggest every trunk muscle contributes to lumbar stability in
large and multi-planar movements, but that passive structures and the MF and ES primarily
control stability in neutral posture tasks. Here we add that, in fact, the relative activation of MF
and ES as well as IO and EO predict trunk coupling in a rLBP population. This relationship only
existed in the clinical population, potentially because impaired motor control processes,
including reduced trunk proprioception,
2,3
paraspinal atrophy,
4
and increased trunk extensor
fatiguability,
5
meant trunk coupling was more strongly or more exclusively influenced by muscle
coordination as opposed to back-healthy control participants with many redundant control
mechanisms contributing to coupling.
In Chapter V, it was shown that persons with and without rLBP did not modulate task
performance differently when dual-tasking with the Balance-Dexterity Task and a working
memory recall and manipulation task. Dexterous force control error increased in both dual-task
interference conditions compared to the single-task condition, but to a larger degree when the
priority was placed on the cognitive task. COM velocity, a measure of balance control, was also
modulated dependent on priority instruction. When the priority was assigned to the cognitive
task, then COM velocity was greater compared to both single-task and dual-task balance priority
conditions. Both groups also modulated cognitive task performance by reducing variability of
errors when the cognitive task was the priority compared to the balance priority dual-task
TRUNK CONTROL DURING DYNAMIC BALANCE 212

condition. These findings confirmed the methodological assumptions behind the attentional
interference and the prioritization manipulation.
The effects of dual-task interference on trunk control were different between groups
where the back-healthy control group had variable changes in trunk coupling from single- to
dual-task conditions, but almost all participants with rLBP increased trunk coupling from single-
to dual-task conditions. First, recall that persons with rLBP exhibited more dissociated trunk
motion compared to back-healthy control participants during the single-task (Balance-Dexterity
Task) condition, driven in part by higher erector spinae activation levels (actually lower
multifidus-to-erector spinae ratios). From single- to dual-task conditions, reductions in erector
spinae activation were associated with increases in trunk coupling, bringing the average trunk
coupling of the rLBP up to the average level of back-healthy controls. Given that these findings
– more erector spinae activation associated with lower trunk coupling in the single-task condition
and reduced erector spinae activations associated with increased trunk coupling from single- to
dual-task conditions – are in agreement and help to tell a complete story, we interpret increased
trunk coupling from single- to dual-task conditions in persons with rLBP as a benefit driven by a
reduction in overactive superficial paraspinal activity between these conditions. But we
emphasize that our conclusions are specific only to this novel Balance-Dexterity Task.
Modulating this effect, persons with rLBP who thought the cognitive task was more
difficult maintained lower trunk coupling during dual-task performance. Also, those who
recalled higher pain levels during a typical symptomatic episode did not reap the benefit of
increased trunk coupling, while those who reported lower pain recall increased trunk coupling
from single- to dual-task conditions. Two potential frameworks explain these findings. Using a
movement-specific attentional reinvestment framework, low attentional loads interfered with
TRUNK CONTROL DURING DYNAMIC BALANCE 213

memory-of-pain-related conscious processing of posture and resulted in decreased superficial
paraspinal activation and improved trunk coupling, while higher attentional loads potentially
induced their own psychological stress. From the perspective of an action-specific perception
framework, persons who increased trunk coupling perceived the cognitive task as easier and
remembered their episodes as less painful.
The investigation described in Chapter VI partially tested these two frameworks by
examining associations between scores on psychometrics including the movement-specific
reinvestment scale (MSRS) and trunk coupling. In the control group, persons who invested more
attention in their movements exhibited greater trunk coupling in line with movement-specific
reinvestment theory, which predicts a “locking down” of degrees of freedom. These individuals
also experienced greater reductions in trunk coupling under dual-task interference. In conditions
of dual-task interference, the relationship between MSRS and trunk coupling disappeared. We
expected persons with rLBP to exhibit the same relationship but to be at one end of the
continuum. Instead, there were no associations between social-cognitive factors including
movement-specific reinvestment and trunk coupling in persons with rLBP. In this group,
however, persons who scored higher on the MSRS had lower dexterous force control variability
and cognitive task error variability, and those who scored higher on FABQ had lower average
COM velocity. These relationships also disappeared under dual-task interference when the
priority was placed on the cognitive task. Overall, persons with rLBP modulated variables more
directly related to task performance and balance based on social-cognitive factors while control
participants modulated trunk control based on these factors. And, for both groups, dual-task
interference interfered with these relationships. Clearly, there are changes in the role of
TRUNK CONTROL DURING DYNAMIC BALANCE 214

psychosocial factors on trunk control in this group, but more research needs to be done in order
to understand links between these findings and the recurrence of LBP.
It is important to note that participants made up a convenience sample of persons with
and without rLBP recruited from student groups, classes, flyers, and university-affiliated
physical therapy clinics, and the clinical population was generally young (age 23.5 ± 2.8yrs),
minimally disabled (ODI 16.0 ± 18.7%), and all in pain remission at the time of testing (0.4 ± 0.4
out of 10 on VAS). The associations between change in trunk coupling and change in erector
spinae activation, cognitive task difficulty, and pain recall were only statistically significant after
removing one, two, or three of the same subset of three participants who did not follow the
pattern. Two of these participants exhibited the greatest increases in trunk coupling from single-
to dual-task conditions and two of the highest reported recalled pain levels. We did not have a
clinical reason to exclude these participants, but the heterogeneity and presence of subgroups in a
LBP population are well-known,
6,7
and applying findings from laboratory research on persons
with rLBP in the clinic should always be done with caution and with subject-specificity in mind.
This is why we decided to present the findings and acknowledge the limitation that there may be
other strategies at work in subgroups of persons with rLBP. The greater variability in this group
may have also resulted in reduced power to detect significant differences in some measures.
Larger studies may allow for subgroup analyses to help explain what different or additional
mechanisms are at work in these three persons with rLBP.
Future work should test additional hypotheses raised in the current dissertation studies.
Importantly, explanations for the lower trunk coupling in the rLBP during the Balance-Dexterity
Task hinge on task-specificity. A very useful study would take the same set of participants
through a set of comparable balance tasks with different key features. More and less dynamic
TRUNK CONTROL DURING DYNAMIC BALANCE 215

control could be tested by varying stiffness parameters of the spring used in the Balance-
Dexterity Task. Effects of internally-driven vs externally-perturbed balance could be tested by
adding a platform-translation-type perturbation to either the stance limb or the spring itself. And,
different cues could be tested to direct attention to dexterous force control, to balance, or to the
trunk specifically. Next, findings from the dual-task interference study raised questions about
how cognitive task difficulty and pain recall influenced trunk coupling and posture-cognition
dual-tasking. More objective measures of these factors – using a math ability test, pain threshold
measures, etc – may help to disambiguate objective measures and perceived measures. This
would provide evidence to clarify the role of action-specific perception in the present findings.
As we learn more about the role of task-specificity and cognitive processing in persons with
rLBP, interventions can be developed that will be more patient-specific and effective in reducing
the recurrence of pain.

TRUNK CONTROL DURING DYNAMIC BALANCE 216

References
1. Cholewicki J, McGill SM. Mechanical stability of the in vivo lumbar spine. Clin
Biomech. 1996;11(1):1-15.
2. Osthoff A-KR, Ernst MJ, Rast FM, et al. Measuring Lumbar Reposition Accuracy in
Patients With Unspecific Low Back Pain: Systematic Review and Meta-analysis. Spine
(Phila Pa 1976). 2015;40(2):E97-E111. doi:10.1097/BRS.0000000000000677.
3. Lee AS, Cholewicki J, Reeves NP, Zazulak BT, Mysliwiec LW. Comparison of trunk
proprioception between patients with low back pain and healthy controls. Arch Phys Med
Rehabil. 2010;91(9):1327-1331. doi:10.1016/j.apmr.2010.06.004.
4. Beneck GJ, Kulig K. Multifidus atrophy is localized and bilateral in active persons with
chronic unilateral low back pain. Arch Phys Med Rehabil. 2012;93(2):300-306.
doi:10.1016/j.apmr.2011.09.017.
5. Beneck GJ, Baker LL, Kulig K. Spectral analysis of EMG using intramuscular electrodes
reveals non-linear fatigability characteristics in persons with chronic low back pain. J
Electromyogr Kinesiol. 2013;23(1):70-77. doi:10.1016/j.jelekin.2012.07.001.
6. Mistry D, Patel S, Hee SW, Stallard N, Underwood M. Evaluating the quality of subgroup
analyses in randomized controlled trials of therapist-delivered interventions for
nonspecific low back pain: a systematic review. Spine (Phila Pa 1976). 2014;39(7):618-
629. doi:10.1097/BRS.0000000000000231.
7. Norton G, McDonough CM, Cabral HJ, Shwartz M, Burgess JF. Classification of patients
with incident non-specific low back pain: implications for research. Spine J. 2015:1-10.
doi:10.1016/j.spinee.2015.08.015.

TRUNK CONTROL DURING DYNAMIC BALANCE 217

APPENDIX A:
EXPLORATORY AIM – COMPARING INDIVIDUALS WITH RECURRENT LOW
BACK PAIN DURING SYMPTOM REMISSION AND A PAINFUL EPISODE
Purpose
The purpose of the Exploratory Aim was to build testable hypotheses about the
differences in trunk control and dual-tasking between periods of symptom remission and painful
episodes in persons with recurrent episodes of low back pain (rLBP). Testing these hypotheses in
future studies will provide insight into the effects of current pain on motor control and posture-
cognition dual-tasking.
Methods
Participants in the rLBP group were asked if they would come in for re-testing during an
episode of pain. Participants who agreed were contacted weekly via text message for 30 weeks
and asked to report their average pain for the past week between 0 and 10. Any persons who
reported pain greater than 4.0 out of 10, or in the range of their typical painful episodes, during
the 30-week follow-up were brought in to repeat testing as described in Appendix B. Five
participants were followed and three participants were brought in for re-testing.
Results
Three participants were brought in for re-testing during or shortly after a painful episode.
They are described here.
rLBP05 (RED): Participant rLBP05 was a 27-year-old female with height 168cm, weight 59.9kg,
and BMI 21.2. She suffered from rLBP for more than five years, averaging three episodes of
pain per year, and the largest Oswestry Disability Index of any participant studied – 80%. This
TRUNK CONTROL DURING DYNAMIC BALANCE 218

participant was followed for 28 weeks after initial testing during symptom remission (Figure
A.0.1A). The participant reported a painful episode between weeks 17-20, but was unable to
come in for testing. She reported an episode again in weeks 27 and 28, after returning from a
visit home which required sitting for a long amount of time on a flight. Here, she reported pain
up to 5/10 at its worst, which was one week before testing. She reported limping for a few days
from the pain, and pain was still 2/10 the week of testing.
rLBP06 (GREEN): Participant rLBP06 was a 28-year-old female with height 173cm, weight
63.5kg, and BMI 21.2. She suffered from rLBP for one year and reported two episodes of pain
per year. Her Oswestry Disability Index was 12%, minimally disabled. This participant was
followed for 14 weeks after initial testing (Figure A.0.1B). She reported an episode at week 14
where pain was up to 4/10 at its worst, which was 10 days before testing. She reported that after
a day of prolonged standing, pain started in the evening and was severe for three days (up to 7
days before testing). On the day of testing she reported pain at 0.6/10 but more pain with
movement, specifically 1/10 with side-bending both directions and 3/10 with extension.
rLBP22 (BLUE): Participant rLBP22 was a 29-year-old male with height 182cm, weight 80.5kg,
and BMI 24.1. This participant was actually tested as a back-healthy control participant during
pilot testing. During the year and a half while the dissertation data collections were being done,
he developed a first episode of LBP. Testing was re-done during this episode, during which he
reported pain at 7.2/10. He did not undergo fine-wire EMG instrumentation in either session.
Since re-testing, he has had two more episodes of LBP.
TRUNK CONTROL DURING DYNAMIC BALANCE 219

Figure A.0.1. Pain reported by participants from 0-10 every Monday
morning at 9:00am via text message from the week after initial testing until
they were brought in for re-testing during a painful episode.
The effect of pain on trunk coupling was different for each individual (Figure A.0.2).
rLBP05 increased trunk coupling dramatically during a painful episode compared to symptom
remission. rLBP06 decreased trunk coupling during a painful episode. rLBP22 started as a back-
healthy control and had greater trunk coupling than any participant in remission from rLBP. In
his first painful episode, his trunk coupling decreased compared to before he had pain. Next, the
influence of trunk muscle activity on trunk coupling was explored in these two conditions
(Figure A.0.3). Recall participant rLBP22 did not undergo fine-wire EMG, so his data points are
shown on the y-axis of the plot. The relationship in the full group of persons in remission from
rLBP was as follows – those with greater deep-to-superficial trunk muscle ratios (multifidus-to-
erector spinae [MF:ES] for paraspinals, internal-to-external oblique [IO:EO] for abdominals)
TRUNK CONTROL DURING DYNAMIC BALANCE 220

exhibited more tightly coupled trunk motion. The relationship between MF:ES and trunk
coupling seemed to remain the same for persons in and out of pain, but individuals moved along
this relationship differently. From remission to a painful episode, rLBP05 increased trunk
coupling along with her MF:ES ratio while rLBP06 decreased trunk coupling and her MF:ES
ratio. The same was not true for the IO:EO ratio, as both individuals decreased this value when
in pain.

Figure A.0.2. Trunk coupling (R
2
) (left) and thorax, pelvis, and trunk
excursion measures (right) for both groups. The three individuals who were
tested during symptom remission (“No Pain”) and during a painful episode
(“Pain”) are shown with connecting lines.
TRUNK CONTROL DURING DYNAMIC BALANCE 221

Figure A.0.3. The relationship between trunk coupling (R
2
) and multifidus
(MF) – to – erector spine (ES) ratio (left) the relationship between trunk
coupling (R
2
) and internal oblique (IO) – to – external oblique (EO) ratio
(right) for both groups. The three individuals who were tested during
symptom remission and during a painful episode (“Pain”) are shown with
connecting lines. Note that participant rLBP22 in blue did not undergo fine-
wire EMG instrumentation and so is shown on the y-axis.
The effect of dual-task interference was tested in each pain condition as well (Figure
A.0.4). As was stated above, the effect of pain on trunk coupling was variable – two participants
decreased trunk coupling when in pain compared to when in symptom remission (rLBP06 and
rLBP22) and one participant increased trunk coupling (rLBP05). The same direction change
occurred in the dual-task interference condition, but to a smaller magnitude. Stated specifically,
rLBP05 increased trunk coupling in the single-task condition from about 0.16 during remission
to 0.70 when in pain, an increase of 0.54. She only increased trunk coupling in the dual-task
condition from 0.25 during remission to 0.61 when in pain, an increase of 0.36. rLBP06
decreased trunk coupling in the single-task condition from about 0.46 during remission to 0.20
when in pain, a decrease of 0.26. She only decreased trunk coupling in the dual-task condition
from 0.44 during remission to 0.37 when in pain, a decrease of 0.07. Finally, rLBP22 decreased
TRUNK CONTROL DURING DYNAMIC BALANCE 222

trunk coupling in the single-task condition from about 0.73 during remission to 0.49 when in
pain, a decrease of 0.24. He only decreased trunk coupling in the dual-task condition from 0.78
during remission to 0.61 when in pain, a decrease of 0.17. The average magnitude of change
from testing in remission to in pain in the single-task condition was 0.35 while the average
magnitude of change from testing in remission to in pain in the dual-task condition was 0.20.

Figure A.0.4. Trunk coupling (R
2
) in two conditions – Balance-Dexterity
Task (single-task) condition and dual-task condition – for the back-healthy
control group (left) and the group with rLBP (right). The three individuals
who were tested during symptom remission (“No Pain”) and during a painful
episode (“Pain”) are shown with connecting lines.
Discussion
This three-person case series raises important testable hypotheses for future work. Effects
of current pain on trunk coupling appear to be different for different individuals. A testable
hypothesis for future work is that there are two subgroups of persons with rLBP – one that
TRUNK CONTROL DURING DYNAMIC BALANCE 223

increases trunk coupling when in pain and one that adopts more dissociated trunk motion when
in pain. Given that we captured one participant’s first episode of pain, we can propose a
hypothesis about the cause-and-effect relationship between trunk coupling and pain recurrence.
At baseline this participant had larger trunk coupling than any other person in remission from
rLBP, and this trunk coupling decreased during his first episode of pain. This could be the start
of the “vicious cycle” where reduced trunk coupling (observed in the group as a whole) could
have been induced by an initial episode of pain and will persist during symptom remission.
Testing a group of back-healthy individuals longitudinally and bringing those who develop pain
back in for re-testing would provide evidence to support or refute this hypothesis.
The effect of dual-task interference for all three of these participants was to mitigate the
effects of pain. Stated another way, the effect of pain on trunk coupling was less in the dual-task
interference condition than it was in the single-task condition. This fits with the perspective
discussed in Chapter V where the cognitive task serves as a distraction from memory or fear of
pain. Here, it appears the cognitive task is distracting from current pain as well. This could be
tested in future work implementing dual-task interference both in conditions of current pain and
symptom remission.
TRUNK CONTROL DURING DYNAMIC BALANCE 224

APPENDIX B:
METHODS
Participant recruitment
Participants were recruited from the local community. Back-healthy control participants
were recruited from USC-affiliated academic programs and student organizations as well as
personal networks of the investigator. Participants with recurrent low back pain (rLBP) were
recruited from student groups, academic programs, and university-affiliated physical therapy
clinics. Active recruitment methods included making presentations to local clinicians and
classes, distributing flyers, shadowing physical therapists at USC-affiliated physical therapy
clinics, and by word-of-mouth.
Inclusion and exclusion criteria
All participants were between the ages of 18 and 45.
1,2
Back-healthy control participants
must have had no LBP in the past year, a washout period supported by epidemiological evidence
of healthcare utilization.
3
Participants in remission from rLBP had at least two episodes of pain,
localized to the area between the lower posterior margin of the rib cage and the horizontal gluteal
fold, per year for at least one year and had pain less than half of the days in the previous six
months (an NIH definition that distinguishes chronicity and recurrence
4
). Painful episodes were
severe enough to limit function based on questions in the NIH Task Force recommended
minimum dataset
4
and Oswestry Disability Index.
5
Pain was less than 1.5 out of 10 on a visual
analog scale (VAS)
6
at the time of testing and for the preceding seven days. Exclusion criteria
included history of low back surgery, history of leg pain below the knee, a radiological or
clinical diagnosis of spinal stenosis, a radiological or clinical diagnosis of spinal scoliosis, spinal
malignancy, spinal infection, lumbar radiculopathy, current or previous musculoskeletal injury or
TRUNK CONTROL DURING DYNAMIC BALANCE 225

surgery affecting locomotion or balance, a history of diabetes mellitus, rheumatic joint disease,
any blood-clotting disorder or current anti-coagulant therapy, polyneuropathy, or current
pregnancy. Individuals in each group were matched based on the following criteria: sex, leg
dominance, age within five years, BMI in the same category, and Baecke Activity Scale vector
sum within two points.

Figure B.0.1. Participant recruitment. rLBP = recurrent low back pain.

TRUNK CONTROL DURING DYNAMIC BALANCE 226

Table B.1. Participant demographics (mean ± standard deviation). BMI =
body mass index; VAS = visual analog scale; ODI = Oswestry Disability
Index; MSRS = movement-specific reinvestment scale; CMP = conscious
motor processing subset; MSC = movement self-consciousness subset; PCS =
pain catastrophizing scale; TSK = Tampa scale for kinesiophobia; FABQ =
fear-avoidance belief questionnaire; PA = physical activity subset; W = work
subset.

Questionnaires and psychometrics
Participants completed a series of questionnaires. Before traveling to the lab for testing, a
Screening Informed Consent document and set of screening questionnaires were sent via email.
Participants completed the recommended minimum dataset provided by the NIH Task Force on
Research Standards for Chronic LBP drawing heavily from the PROMIS and STarT Back
methodology,
4
the Oswestry Disability Index,
5
Baecke Physical Activity Questionnaire,
7
and
demographic information. If eligible, participants were brought in for testing. Once in the lab,
Table 1
rLBP CNTRL p
Age (yrs) 23.5 ± 2.8 23.9 ± 3.3 0.679
Sex 7 M, 12 F 7 M, 12 F
Leg Dominance 18 R, 1 L 18 R, 1 L
Height (cm) 170.4 ± 8.4 169.1 ± 10.4 0.692
Weight (kg) 68.7 ± 10.3 67.1 ± 10.8 0.661
BMI 23.6 ± 2.4 23.3 ± 1.8 0.714
Baecke Physical Activity Scale (vector sum) 4.8 ± 1.0 4.9 ± 0.6 0.537
Episodes per year 3.4 ± 1.2
Pain during episodes (recall, VAS 0-10) 4.9 ± 2.2
Pain at time of testing (VAS 0-10) 0.4 ± 0.4
ODI (recall, %) 16.0 ± 18.7
MSRS (0-50) 25.6 ± 4.8 24.8 ± 9.7 0.793
CMP (0-25) 14.9 ± 4.8 13.8 ± 4.5 0.484
MSC (0-25) 10.7 ± 5.5 10.9 ± 5.9 0.891
PCS (0-52) 8.3 ± 8.5 6.5 ± 5.1 0.438
TSK (17-68) 31.3 ± 6.5 30.5 ± 6.0 0.706
FABQ (0-96) 20.2 ± 10.7
FABQ-PA (0-66) 12.2 ± 7.7
FABQ-W (0-30) 8.1 ± 6.7
Demographics of Participants
TRUNK CONTROL DURING DYNAMIC BALANCE 227

they completed a fear-avoidance beliefs questionnaire (FABQ),
8
Tampa scale for kinesiophobia
(TSK),
9
pain catastrophizing scale (PCS),
10
and the movement-specific reinvestment scale
(MSRS).
11

Due to the central role the construct of movement-specific reinvestment holds in the
dissertation aims, hypotheses, and analyses, it is worth discussing details of this construct’s
measurement tool – the MSRS (Figure B.0.2). This scale has ten items, each of which is rated
using a six-point Lickert scale between ‘strongly disagree’ and ‘strongly agree’, and is split into
two components: Conscious Motor Processing (CMP) and Movement Self-Consciousness
(MSC). A major strength of the scale and hypothesis in the proposed dissertation studies is that
these items probing reinvestment, especially the CMP items, may be a mechanistic link between
the high scores on scales used in LBP research like fear-avoidance beliefs, kinesiophobia, and
pain catastrophizing and behavioral outcomes measures. In other words, fear of pain or re-injury
may be one construct that leads to movement reinvestment, which in turn may lead to changes in
motor performance. In addition to this, however, reinvestment may also capture a phenomenon
in healthy controls not motivated by fear or pain. For example, a trained dancer likely will have
high movement reinvestment due to the high volume of internal focus cues given during training
of certain dance tasks.
12
This healthy dancer may score high on the MSRS without the influence
of fear of pain. A scale that works equally well for the patients and the healthy controls in a study
is a strength of the psychometric tool. Another strength of the MSRS lies in its concurrent and
discriminant validity evaluated by a four-part study conducted by Laborde et. al. in the European
Union.
13
Comparing scores on the MSRS with scores on a preference for intuition or deliberation
inventory showed that MSRS had convergent validity with a preference for deliberation and
discriminant validity with a preference for intuition, defined as making decisions based on a
TRUNK CONTROL DURING DYNAMIC BALANCE 228

quick judgment without much conscious awareness. In addition, MSRS score correlated with
trait characteristics like self-consciousness, certain perfectionism measures, and a tendency to
ruminate as a means of working through issues. Discriminant validity was found with trait
anxiety and a tendency to use distraction to work through issues.

Figure B.0.2. The Movement-Specific Reinvestment Scale (Masters and
Maxwell, 2008).
The scale has been used successfully in a number of patient populations providing
evidence not only of its theoretical strength but its practical utility. Orrell and colleagues
conducted a study with 162 survivors of stroke compared to 148 healthy controls and found not
only that stroke survivors scored higher on the MSRS but that CMP score and time spent in
TRUNK CONTROL DURING DYNAMIC BALANCE 229

rehabilitation were significant predictors of impairment following stroke, meaning that a
tendency to consciously process motor output could lead to worse functional outcomes for these
patients.
14
Another group of researchers found that duration of Parkinson’s disease was
positively correlated with an increased score on the MSRS.
15
And finally, a study of older adults
found that those who had fallen scored higher on both CMP and MSC components of the MSRS,
and there was a significant association between CMP score and faller status.
16
It is, of course,
important to point out weaknesses of the scale, especially weaknesses or limitations for the
scale’s use in these dissertation studies. The biggest limitation here is that persons with rLBP,
especially during periods of symptom remission, do not visibly move differently than their
healthy counterparts. While they may still consciously process their movements, reflected by
their CMP score, it is unlikely they will score high on any of the MSC items, for example “If I
see my reflection in a shop window, I will examine my movements” or “I am concerned about
what people think about me when I am moving.” These questions geared more toward the “style”
of movement as opposed to the conscious contemplation of movement may only be applicable
for patient populations with visible disability like in severe stroke or Parkinson’s disease
conditions. In addition, two different studies report lower test-retest reliability of the MSC
compared to the CMP and only the CMP was related to a motor imagery capability measure,
13

indicating that it measures something distinctly different and more relevant to the proposed
study.
Experimental instrumentation
3D motion capture, force plates, and surface and fine-wire electromyography (EMG)
were used to accomplish the proposed aim. Participants were fitted with a full-body set of 14 mm
retroreflective markers following a modified Cleveland Clinic marker set. Data were collected
TRUNK CONTROL DURING DYNAMIC BALANCE 230

with an 11-camera Qualisys Oqus System (Gothenburg, Sweden) sampled at 250 Hz. Ground
reaction force and center of pressure data were collected using two force plates (Advanced
Medical Technology Inc., Watertown, MA) sampled at 3000 Hz. EMG data were collected using
a wireless Noraxon system (Scottsdale, AZ) sampled at 3000 Hz. Surface EMG were placed on
the stance leg (defined as the leg contralateral to the preferred kicking leg) and same side of the
trunk. After skin preparation including shaving and alcohol wipe, bipolar silver/silver chloride
electrodes with an interelectrode distance of 22 mm were placed over the following muscles
according to guidelines from Surface ElectroMyoGraphy for the Non-Invasive Assessment of
Muscles (SENIAM)
17
: the stance leg gluteus maximus (GMax) and gluteus medius (GMed); the
rectus abdominis (RA); and the external oblique (EO). Fine-wire EMG were collected using a
pair of 50 μm nickel-chromium alloy wires with nylon insulation with the distal 2 mm stripped
of insulation and bent into a hook loaded into 25 gauge hypodermic needles and sterilized. The
following muscles were instrumented on the same side as the stance leg: deep fibers of the
lumbar multifidus at the level of L4 (MF), the lumbar erector spinae at the level of L4 (ES) with
contributions likely from both longissimus and iliocostalis thoracis pars lumborum, and the
internal oblique (IO). Insertions were done under ultrasound guidance (Siemens Medical
Solutions USA, Inc., Malvern, PA) (Figure B.0.3). The cannula was removed immediately while
the wires remained in the muscle and were “set” using isometric contraction followed by slow
movement throughout the available range of motion. Electrode placement was confirmed using
mild electrical stimulation to produce contraction, which was visualized using ultrasound
imaging. All insertions were done under the supervision of a certified kinesiological
electromyographer. To minimize infection risk, sterile electrodes and ultrasound gel, and single-
use needles, ultrasound probe covers, and gloves were used.
TRUNK CONTROL DURING DYNAMIC BALANCE 231

Figure B.0.3. Fine-wire electromyography insertion of the lumbar mulitifidus
(left) and the internal oblique (right).
Procedures
On the day of testing, participants gave Informed Consent and completed the series of
questionnaires described. They then were instrumented with a full-body marker set and
completed a series of baseline tests including active standing trunk range-of-motion measures, a
single question about which (if any) movement directions caused pain or discomfort, a 30s
double-leg stance trial, and three 30s single-leg stance trials. They were then instrumented with
surface and fine-wire EMG as described. Next, participants were introduced to the Balance-
Dexterity Task. The Balance-Dexterity Task was developed by combining single-limb balance
with the lower-extremity dexterity test (LED-test),
18
and serves as an ideal task to observe
aspects of submaximal, continuous postural control. The traditional LED-test involves
compression of an unstable spring while semi-seated on a bicycle seat with arms resting on a
support surface. It quantifies lower-limb dexterity since the compression force achieved is
associated with performance on the cross-agility test (R
2
=0.63) but not hip extensor strength
(R
2
=0.04), knee extensor strength (R
2
<0.01), or knee flexor strength (R
2
=0.02).
18,19
Adding this
dexterous force control demand to the balance demands of single-limb stance
20
can be viewed as
TRUNK CONTROL DURING DYNAMIC BALANCE 232

a lower-extremity bimanual task of sorts and allows us to study motor control processes involved
in successful task execution. The Balance-Dexterity Task used a custom device made by
mounting polyvinyl chloride (PVC) adaptors to clipboards with a spring between them (spring
characteristics: outside diameter 1.750in, inside diameter 1.336in, free length 12.0in, rate 28.0
lbs/in, wire diameter 0.207in, and total coils 27.5; Compression Spring #805, Century Spring
Corp., Commerce, CA). A similar instrumented device is available from Neuromuscular
Dynamics, LLC (La Crescenta, CA).
Participants completed five practice trials of the Balance-Dexterity Task lasting 20-25s each.
Participants received visual feedback about the vertical force produced under the compressed
spring and were given the instruction “While standing on one leg, compress this spring so that
the blue line is first as high and then as stable as possible”. The middle 50% of the last three of
these trials were then used to calculate the individual’s reproducible, submaximal compression.
This value was called reproducible, submaximal compression because (1) the goal of the
Balance-Dexterity Task is to use submaximal dexterous force control to perturb balance; (2) for
the LED-test, at least 20-25 attempts are required to produce a stable maximum indicating the
compression described here is not maximal;
18
(3) in pilot testing it was found that, without the
seat and arm rests used in the LED-test, giving subjects more than five practice trials to achieve a
stable maximum led to creative but confounding strategies sometimes including a deep squat
with the stance leg or wedging the spring into a contorted shape; and (4) a subset of five
participants who repeated testing reproduced similar compression values between days with an
ICC of 0.875. Given five practice trials, all back-healthy control participants achieved
submaximal, but sufficiently perturbing dexterous control of the spring with compression forces
ranging from 100-139N (mean 121.2 ± 12.3N), representing 14.4-23.0% of body weight (mean
TRUNK CONTROL DURING DYNAMIC BALANCE 233

18.7 ± 2.4%). After these practice trials, participants were asked the following set of questions
and marked their answer on a VAS from 0-10.
(1) How difficult is the balance task?
0 Anchor: “Not difficult at all”
10 Anchor: “Extremely difficult”
(2) How much attention does the balance task require?
0 Anchor: “No attention at all”
10 Anchor: “All my attention”
(3) How confident are you that you can perform this balance task?
0 Anchor: “Not confident at all”
10 Anchor: “Extremely confident”
Next, participants were introduced to the cognitive task. This task was a modified verbal
working memory recall and manipulation task requiring participants remember a list of five
random digits from zero to ten, do an arithmetic operation to those digits twice during each trial,
and remember five answers. The development of this task was constrained by the EMG
instrumentation of deep abdominals, which activate with verbalization.
21
Therefore, the task had
to be modified from similar tasks in the literature to avoid real-time verbalization of answers.
Participants practiced the task while seated until comfortable. After practice, participants were
asked the following set of questions and marked their answer on a VAS from 0-10.
(1) How difficult is the cognitive task?
0 Anchor: “Not difficult at all”
10 Anchor: “Extremely difficult”
TRUNK CONTROL DURING DYNAMIC BALANCE 234

(2) How much attention does the cognitive task require?
0 Anchor: “No attention at all”
10 Anchor: “All my attention”
(3) How confident are you that you can perform this cognitive task?
0 Anchor: “Not confident at all”
10 Anchor: “Extremely confident”
(4) How confident are you that you can perform both the balance task and the cognitive task
together?
0 Anchor: “Not confident at all”
10 Anchor: “Extremely confident”
Then, in a randomized order, participants completed five different task conditions.
(1) Participants completed five trials of the cognitive task while seated (“cognitive single-task”).
(2) Participants completed five trials where a dotted line indicating their reproducible,
submaximal compression was shown as a goal with the instructions: “While standing on one
leg, compress this spring so that the line is as stable as possible directly over the dotted goal
line” (“Balance-Dexterity Task”).
(3) Participants completed five dual-task trials with a cognitive priority where the instructions
given were: “In this trial, it is most important that you get the five numbers correct, and it is
less important the line is as stable as possible” (“dual-task cognitive priority”).
(4) Participants completed another five dual-task trials with a balance priority where the
instructions given were: “In this trial, it is most important that you keep the line as stable as
TRUNK CONTROL DURING DYNAMIC BALANCE 235

possible, and it is less important that you get the five numbers correct” (“dual-task balance
priority”).
(5) Finally, three trials were interspersed where the spring was replaced with a stable block of
the same height, and the same instructions as in the Balance-Dexterity Task were given.
At the end of data collection, participants performed a rate-controlled full trunk flexion
movement in order to identify the presence or absence of a flexion-relaxation response. Finally,
participants were given the MSRS again with wording changes to reference the Balance-
Dexterity Task specifically. Five back-healthy control participants were brought in for re-testing
1-4 days after the initial collection to assess test-retest reliability of outcomes measures.
Data analysis
All trials were trimmed so the middle 50% of the task was analyzed. Trials were screened
to assure stable force was reached by the start of the 50% window. Kinematic and force plate
data were low-pass filtered with cutoff frequencies of 12 Hz and 50 Hz, respectively. Surface
and fine-wire EMG data were band-pass filtered between 20 and 500 Hz and 20 and 1000 Hz,
respectively, all using a dual-pass 4
th
order Butterworth filter. After filtering, EMG data were
rectified and smoothed with a moving weighted average window of 500 ms. Signals were
normalized both to maximum voluntary isometric contractions (MVICs) and to averaged signal
amplitude during the stable block trials. Here, EMG amplitude could be reported as a percentage
of activation during the stable block, allowing activation to be interpreted as a response to the
added challenge of dexterous force control, not to the body position or vertical force production.
Note one subject was excluded when reporting MVIC-normalized EMG because of corrupted
TRUNK CONTROL DURING DYNAMIC BALANCE 236

MVIC data. And, one participant with rLBP is missing MF and IO muscle activation data due to
failed fine-wire insertions.
Measures of task performance were analyzed. To quantify balance control demands,
center of pressure (COP) measures from the stance limb were calculated including average COP
resultant velocity and COP area measured with a 95% confidence ellipse, as well as average
center of mass (COM) resultant velocity (Figure B.0.4, Equation D, Equation E). Note that one
participant in each group is missing COM data due to marker occlusion during collection.
Dexterous force control demands were measured using the vertical force produced under the
spring and quantified as root-mean-squared error (RMSE) from the reproducible, submaximal
compression goal line; coefficient of variation (CV); and median frequency (MDF) of the
detrended vertical force (Figure B.0.4). Cognitive task performance was quantified as variability
of errors in the five trials of each condition, a measure more sensitive than absolute error.
22,23

Muscle activation data were averaged to acquire a mean activation amplitude for each trial.
Muscle activation ratios were calculated in a frame-by-frame manner including deep-to-
superficial ratios for the paraspinals (MF:ES) and abdominals (IO:EO) and co-contraction ratios
for the deep trunk muscles (MF, IO) and superficial trunk muscles (ES, EO) (Equation F).
Equation D. Average resultant center of pressure (COP) velocity.

Equation E. Average resultant center of mass (COM) velocity.

TRUNK CONTROL DURING DYNAMIC BALANCE 237

Equation F. Trunk muscle activation ratio.
For deep-to-superficial ratios, Muscle 1 was multifidus (MF) for paraspinals and
internal oblique (IO) for abdominals; and for co-contraction ratios, the muscle
whose mean amplitude was lower was Muscle 1.

Trunk control was quantified by tracking thorax and pelvis motion relative to global
coordinates. Using an angle-angle plot of thorax and pelvis frontal plane rotation, a coefficient of
determination (R
2
) was calculated where a high R
2
would indicate highly coupled thorax and
pelvis motion and a low R
2
would indicate more dissociated or independent motion of the thorax
and pelvis (Figure B.0.5). This particular metric has been used during gait to distinguish
participants with and without low back pain through frontal and transverse plane trunk
coupling.
24,25
Also from this angle-angle plot, instantaneous coupling angles were calculated and
metrics were extracted including the percent of time spent in in-phase coupling and in anti-phase
coupling, all per equations presented in Needham et al.
26
Finally, a principal component analysis
(PCA) was run and the percent of variance explained by the first PCA vector was extracted. In
addition, more traditional range-of-motion metrics were acquired including thorax, pelvis and
trunk (thorax relative to pelvis) frontal plane angular excursions.

TRUNK CONTROL DURING DYNAMIC BALANCE 238

Figure B.0.4. The Balance-Dexterity Task with representative data showing
balance outcome measures including center of pressure (COP) measures
(right) and dexterous vertical force (vForce) control outcome measures
including root-mean-squared error (RMSE), coefficient of variation (CV)
and median frequency (MDF) (left).

Figure B.0.5. Representative examples of high (left) and low (right) frontal
plane trunk coupling quantified with a coefficient of determination for the
thorax and pelvis angle-angle plot (R
2
).
TRUNK CONTROL DURING DYNAMIC BALANCE 239

Exploratory aim – Case series follow-up
Participants in the rLBP group were asked if they would come in for re-testing during an
episode of pain. Participants who agreed were contacted weekly via text message for 30 weeks
and asked to report their average pain for the past week between 0 and 10. Any persons who
reported pain greater than 4.0 out of 10, or in the range of their typical painful episodes, during
the 30-week follow-up were brought in to repeat testing as described above. Three participants
were brought in for re-testing.

TRUNK CONTROL DURING DYNAMIC BALANCE 240

References
1. Beneck GJ, Baker LL, Kulig K. Spectral analysis of EMG using intramuscular electrodes
reveals non-linear fatigability characteristics in persons with chronic low back pain. J
Electromyogr Kinesiol. 2013;23(1):70-77. doi:10.1016/j.jelekin.2012.07.001.
2. Beneck G, Kulig K, Landel R, Powers C. The relationship between lumbar segmental
motion and pain response produced by a posterior-to-anterior force in persons with
nonspecific low back pain. J Orthop Sports Phys Ther. 2005;35:203-209.
3. Norton G, McDonough CM, Cabral HJ, Shwartz M, Burgess JF. Classification of patients
with incident non-specific low back pain: implications for research. Spine J. 2015:1-10.
doi:10.1016/j.spinee.2015.08.015.
4. Deyo RA, Dworkin SF, Amtmann D, et al. Report of the NIH Task Force on research
standards for chronic low back pain. J Pain. 2014;15(6):569-585.
doi:10.1016/j.jpain.2014.03.005.
5. Fairbank JC, Pynsent PB. The Oswestry Disability Index. Spine (Phila Pa 1976).
2000;25(22):2940-52; discussion 2952. doi:10.1097/00007632-200011150-00017.
6. Carlsson AM. Assessment of chronic pain. I. Aspects of the reliability and validity of the
visual analogue scale. Pain. 1983;16(1):87-101. doi:10.1016/0304-3959(83)90088-X.
7. Baecke J, Burema J, Frijters J. A short questionanaire for the measuremnet of habitual
physical activity in epidemiological studies. Am J Clin …. 1982;36(November):936-942.
http://ajcn.nutrition.org/content/36/5/936.short.
8. Waddell G, Newton M, Henderson I, Somerville D, Main CJ. A Fear-Avoidance Beliefs
Questionnaire (FABQ) and the role of fear-avoidance beliefs in chronic low back pain and
disability. Pain. 1993;52(2):157-168. doi:10.1016/0304-3959(93)90127-B.
TRUNK CONTROL DURING DYNAMIC BALANCE 241

9. Vlaeyen JWS, Kole-Snijders AMJ, Rotteveel AM, Ruesink R, Heuts PHTG. The role of
fear of movement/(re) injury in pain disability. J Occup Rehabil. 1995;5(4):235-252.
10. Sullivan M, Bishop S, Pivik J, Plvik J. The pain catastrophizing scale: development and
validation. Psychol Assess. 1995;7(4):432-524. doi:10.1037/1040-3590.7.4.524.
11. Masters R, Maxwell J. The theory of reinvestment. Int Rev Sport Exerc Psychol.
2008;1(2):160-183. doi:10.1080/17509840802287218.
12. Guss-West C, Wulf G. Attentional Focus in Classical Ballet: A survey of professional
dancers. J Danc Med Sci. 2016;20(1):23-29.
13. Laborde S, Musculus L, Kalicinski M, Klämpfl MK, Kinrade NP, Lobinger BH.
Reinvestment: Examining convergent, discriminant, and criterion validity using
psychometric and behavioral measures. Pers Individ Dif. 2015;78:77-87.
doi:10.1016/j.paid.2015.01.020.
14. Orrell AJ, Masters RSW, Eves FF. Following Stroke. Neurorehabil Neural Repair.
2009;23(2):177-183.
15. Masters RSW, Pall HS, MacMahon KMA, Eves FF. Duration of Parkinson disease is
associated with an increased propensity for “reinvestment”. Neurorehabil Neural Repair.
2007;21(2):123-126. doi:10.1177/1545968306290728.
16. Wong WL, Masters RSW, Maxwell JP, Abernethy AB. Reinvestment and falls in
community-dwelling older adults. Neurorehabil Neural Repair. 2008;22(4):410-414.
doi:10.1177/1545968307313510.
17. Hermens H, Freriks B, Merletti R, et al. The Surface ElectroMyoGraphy for the Non-
Invasive Assessment of Muscles Project (SENIAM). seniam.org. Published 2006.
18. Lyle MA, Valero-Cuevas FJ, Gregor RJ, Powers CM. The lower extremity dexterity test
TRUNK CONTROL DURING DYNAMIC BALANCE 242

as a measure of lower extremity dynamical capability. J Biomech. 2013;46(5):998-1002.
doi:10.1016/j.jbiomech.2012.11.058.
19. Lyle MA, Valero-Cuevas FJ, Gregor RJ, Powers CM. Lower extremity dexterity is
associated with agility in adolescent soccer athletes. Scand J Med Sci Sports.
2013;25(1):81-88. doi:10.1111/sms.12162.
20. Lawrence EL, Cesar GM, Bromfield MR, Peterson R, Valero-Cuevas FJ, Sigward SM.
Strength, Multijoint Coordination, and Sensorimotor Processing Are Independent
Contributors to Overall Balance Ability. Biomed Res Int. 2015;2015.
doi:10.1155/2015/561243.
21. De Troyer A, Estenne M, Ninane V, Gansbeke D Van, Gorini M. Transversus abdominis
muscle function in humans. J Appl Physiol. 1990;68(3):1010-1016.
22. Schmidt RA, Lee TD. Motor Control and Learning: A Behavioral Emphasis. 4th ed.
Champaign, IL: Human Kinetics; 2005.
23. Chiviacowsky S, Campos T, Domingues MR. Reduced frequency of knowledge of results
enhances learning in persons with Parkinson’s disease. Front Psychol. 2010;1(DEC):1-6.
doi:10.3389/fpsyg.2010.00226.
24. van den Hoorn W, Bruijn SM, Meijer OG, Hodges PW, van Dieen JH. Mechanical
coupling between transverse plane pelvis and thorax rotations during gait is higher in
people with low back pain. J Biomech. 2012;45(2):342-347.
doi:10.1016/j.jbiomech.2011.10.024.
25. Crosbie J, de Faria Negrão Filho R, Nascimento DP, Ferreira P. Coordination of Spinal
Motion in the Transverse and Frontal Planes During Walking in People With and Without
Recurrent Low Back Pain. Spine (Phila Pa 1976). 2013;38(5):E286-E292.
TRUNK CONTROL DURING DYNAMIC BALANCE 243

doi:10.1097/BRS.0b013e318281de28.
26. Needham RA, Naemi R, Chockalingam N. A new coordination pattern classification to
assess gait kinematics when utilising a modified vector coding technique. J Biomech.
2015;48(12):3506-3511. doi:10.1016/j.jbiomech.2015.07.023.

Abstract (if available)

Abstract Postural control is crucial for successful human movement. The consequences of a failure in the postural control system range from an innocuous stumble to life-threatening falls and detrimental conditions of chronic pain. The human movement system protects against such failures by using multiple, diverse, and redundant interacting systems. By probing individual systems through mechanical and cognitive perturbations to posture and balance we can learn about roles and functions of these contributing components. These experimental approaches have been impactful in the study and rehabilitation of persons with low back pain, a condition which represents a tremendous cost to society in terms of quality of life, healthcare expenditure, and work time loss. In persons suffering from recurrent episodes of low back pain (rLBP), specific abnormal movement and muscle activation patterns have been identified that persist even during symptom remission and are thought to contribute to the recurrence of pain. In order to translate laboratory findings to rehabilitation of these patients, novel continuous, ecological balance tasks must be developed and used to study typical balance control and how control may be dysfunctional in persons suffering from rLBP. Chapter II provided a literature-informed justification and rationale for the hypotheses tested in these dissertation studies and important aspects of study designed including development of the Balance-Dexterity Task and the cognitive dual-task paradigm. In addition, a justification for studying a clinical population in remission from recurrent episodes of LBP as a model dysfunctional postural control system was presented. ❧ The purpose of Chapter III was to characterize the Balance-Dexterity Task as a novel dynamic balance task investigate trunk control. The task combined aspects of single-limb balance and the lower-extremity dexterity test by asking participants to stand on one limb while compressing an unstable spring with the contralateral limb to an individualized target force. Nineteen back-healthy control participants completed the study, and performance measures for the demands of each limb—balance and dexterous force control—as well as kinematic and electromyographic measures of trunk control were collected. Given five practice trials, participants achieved spring compression forces ranging from 100-139N (mean 121.2 ± 12.3N), representing 14.4-23.0% of body weight (mean 18.7 ± 2.4%), which were then presented as target forces during test trials. The task invoked larger center of pressure (COP) velocities than double- and single-limb stance and more variable dexterous force control compared to a stable step condition. Variability in dexterous force control was associated with variability in balance control, indicated by a positive correlation between dexterous force coefficient of variability and stance limb COP velocity (R=0.598, p=0.007). This finding in a concurrent bipedal task aligns with similar work in bimanual control tasks where variability between limbs is also associated. Trunk coupling, quantified as the coefficient of determination (R²) of a frontal plane thorax and pelvis angle-angle plot, varied independently of any measure of balance or dexterous force control. Trunk coupling R² was correlated with the percent of time in in-phase coupling and oppositely correlated with the percent of time in anti-phase coupling, justifying use of this simpler measure to capture trunk coordination. Standard segment and joint excursions, however, seemed to quantify different aspects of trunk movement. Muscle activation levels normalized to a stable block reference condition were greater in the Balance-Dexterity Task compared to the stable block condition, but no muscle activations alone or in ratios were associated with trunk coupling R². The paucity of associations between individual muscle activations and trunk coupling indicates that these non-disabled persons have redundant motor control processes available to control balance, dexterous force control, and trunk coupling. When signals were normalized to MVICs, however, a greater multifidus-to-erector spinae ratio was weakly associated with higher trunk coupling R² (R=0.517, p=0.028) suggesting greater relative deep paraspinal activation drove greater trunk coupling. In conclusion, the Balance-Dexterity Task is a continuous, dynamic balance task where bipedal coordination, captured as measures of balance and dexterous force control between which variability was associated, and trunk coupling can be observed and studied. ❧ The purpose of Chapter IV was to compare Balance-Dexterity Task performance and trunk control during the task between persons with and without rLBP. Motor control dysfunction persisting during symptom remission in persons with rLBP may contribute to the recurrence of pain. No differences in task performance were expected between those with and without rLBP, but it was hypothesized persons with rLBP would exhibit greater trunk coupling in line with a trunk stiffening strategy. Persons with and without rLBP (n=19 per group) completed the Balance-Dexterity Task. Persons with rLBP must have had at least two episodes of pain per year but been in symptom remission during the time of testing. Task performance outcome measures included COP velocity under the stance limb and vertical force variability under the spring. Trunk coupling was quantified with the coefficient of determination (R²) of an angle-angle plot of thorax-pelvis frontal plane motion. Fine-wire and surface electromyography captured activations of paraspinals and abdominals. All participants were able to successfully complete the Balance-Dexterity Task with no differences between groups in reproducible, submaximal compression force goal, COP velocity, or dexterous force control measures. In both groups, frontal plane trunk coupling varied independently of task performance measures. The group in remission from rLBP exhibited reduced trunk coupling, or more dissociated thorax and pelvis motion, compared to back-healthy controls (p=0.024). Trunk coupling in this group was associated moderately with lumbar multifidus-to-erector spinae activation ratio (R=0.618, p=0.006) and weakly with internal-to-external oblique ratio (R=0.476, p=0.046). Here, greater deep trunk muscle activation relative to more superficial muscles resulted in higher trunk coupling. The finding of more dissociated thorax and pelvis motion in those with rLBP was counter to the hypothesis of increased coupling through adoption of a stiffening strategy. This unsupported hypothesis was built on investigations of discrete perturbations to posture where greater trunk co-contraction and stiffness has been observed in persons with rLBP. These perturbations, however, involved delivering external perturbing forces either to a support surface or directly to the trunk, which may invoke a stiffening strategy related to fear of movement or pain. The Balance-Dexterity is a submaximal, internally-driven unstable balance task during which more dissociated trunk motion was observed in persons in remission from rLBP. Findings underscore the task-dependent nature of trunk control research and assessment in persons with rLBP. ❧ The purpose of Chapter V was to investigate effects of cognitive dual-task interference and task prioritization instructions on task performance and trunk control during dynamic balance in persons with and without rLBP. First, the ability to modulate task performance in accord with prioritization instructions was tested. Second, it was hypothesized trunk control strategies in persons with rLBP would rely more on cognitive resources, and therefore trunk kinematics would be altered in this group under dual-task interference compared to back-healthy control participants. Finally, individual factors explaining changes in trunk kinematics were explored. Persons with and without rLBP (n=19 per group) completed the Balance-Dexterity Task with and without a cognitive task utilizing verbal working memory. No differences between groups were identified for task performance measures, indicating the groups did not modulate performance differently. Dexterous force control error increased in both dual-task interference conditions compared to the single-task condition, but to a larger degree when the priority was placed on the cognitive task. Center of mass (COM) velocity, a measure of balance control, also changed dependent on priority instruction. When the priority was assigned to the cognitive task, COM velocity was greater compared to both single-task and dual-task balance priority conditions. Cognitive task performance during the dual-task conditions was also modulated by reducing variability of errors when the cognitive task was the priority compared to when balance was prioritized. These findings confirmed the methodological assumptions behind the attentional interference and the prioritization manipulation. Trunk coupling was lower in the rLBP compared to back-healthy controls in the single-task condition (p=0.024) and increased significantly in the dual-task condition (p=0.002). The amount of increase was associated with reductions in erector spinae activation (R=-0.574, p=0.020), lower self-reported cognitive task difficulty (R=-0.497, p=0.036), and lower recalled pain (R=-0.642, p=0.005). Kinesiologically, a reduction in superficial paraspinal activity from single- to dual-task conditions driving the increase in trunk coupling up to the level of back-healthy controls suggests this increased trunk coupling is beneficial. Those with rLBP who reported the cognitive task was easier and who recalled less pain during a typical episode reaped this benefit. Two potential frameworks explain these findings. Using a movement-specific attentional reinvestment framework, low attentional loads interfered with memory-of-pain-related conscious processing of posture and resulted in decreased erector spinae activation and improved trunk coupling, while higher attentional loads potentially induced their own psychological stress. Using an action-specific perception framework, cause and effect are reversed such that persons who increased trunk coupling perceived the cognitive task as easier and their episodes as less painful. ❧ The purpose of Chapter VI was to investigate the role of movement-specific attention reinvestment in task performance and trunk coordination during posture-cognition dual-tasking in persons with and without rLBP. It was hypothesized persons who invest more attention in their movement would exhibit greater trunk coupling within the framework of “locking down” degrees of freedom influencing balance control, and cognitive dual-tasking would interrupt this relationship. Persons with and without rLBP (n=19 per group) completed the Balance-Dexterity Task with and without a cognitive task utilizing verbal working memory. Task performance measures included COM velocity and vertical force variability under the spring. Trunk coupling was quantified with the coefficient of determination (R²) of an angle-angle plot of thorax-pelvis frontal plane motion. Psychometrics collected included the movement-specific reinvestment scale (MSRS), the fear-avoidance beliefs questionnaire (FABQ), the pain catastrophizing scale, and the Tampa scale for kinesiophobia. In the control group, persons who invested more attention in their movements exhibited greater trunk coupling (R=0.647, p=0.003) and experienced greater reductions in trunk coupling under dual-task interference (R=-0.537, p=0.018). No associations between social-cognitive factors and trunk coupling were observed in persons with rLBP. Those who scored higher on MSRS, however, had lower dexterous force variability (R=0.532, p=0.019), and those who scored higher on FABQ had lower COM velocity (R=-0.534, p=0.022). These relationships also disappeared under dual-task interference. Overall, persons with rLBP modulated variables more directly related to task performance and balance based on social-cognitive factors while control participants modulated trunk control based on these factors. And, for both groups, dual-task interference interfered with these relationships. More research needs to be done in order to understand links between these findings and the recurrence of LBP. ❧ Participants in these dissertation studies made up a convenience sample of persons with and without rLBP recruited from student groups, classes, flyers, and university-affiliated physical therapy clinics, and the clinical population was generally young (age 23.5 ± 2.8yrs), minimally disabled (ODI 16.0 ± 18.7%), and all in pain remission at the time of testing (0.4 ± 0.4 out of 10 on VAS). Some associations in Chapter V were only statistically significant after removing one, two, or three of the same subset of three participants who did not follow the pattern. Two of these participants exhibited the greatest increases in trunk coupling from single- to dual-task conditions and two of the highest reported recalled pain levels. We did not have a clinical reason to exclude these participants, but the heterogeneity and presence of subgroups in a LBP population are well-known, and applying findings from laboratory research on persons with rLBP in the clinic should always be done with caution and with subject-specificity in mind. This is why we decided to present the findings and acknowledge the limitation that there may be other strategies at work in subgroups of persons with rLBP. Larger studies may allow for subgroup analyses to help explain what different or additional mechanisms are at work in these three persons with rLBP. Future work should test additional hypotheses raised in the current dissertation studies. Importantly, explanations for the lower trunk coupling in the rLBP group during the Balance-Dexterity Task hinge on task-specificity. This could be tested by taking the same set of participants through a set of comparable balance tasks with different key features. Findings from the dual-task interference study raised questions about how cognitive task difficulty and pain recall influenced trunk coupling and posture-cognition dual-tasking. More objective measures of these factors—using a math ability test, pain threshold assessments, etc.—may help to disambiguate objective measures and perceived measures. This would provide evidence to clarify the role of action-specific perception in the present findings. As we learn more about the role of task-specificity and cognitive processing in persons with rLBP, interventions can be developed that will be more patient-specific and effective in reducing the recurrence of pain.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

The footprint of pain: investigating persistence of altered trunk control in recurrent low back pain

PDF

Trunk neuromechanics during turning: a window into recurrent low back pain

PDF

The structure of kinematic variability during hopping: a window into motor control factors in overuse injury

PDF

Development of automaticity for discrete movements: A systematic investigation of the dual-task probe paradigm

PDF

Manipulating cognitive intensity during aerobic exercise: clinical proof of concept

PDF

The role of the upper extremities in the recovery of balance during a slip perturbation

PDF

Behavioral and neurophysiological studies of hand dexterity in health and Parkinson's disease

PDF

Hemisphere-specific deficits in the control of bimanual movements after stroke

PDF

Development of postural strategies for the control of forward momentum

PDF

The influence of tibiofemoral kinematics and knee extensor mechanics on patellar tendon stress: a comparison of persons with and without patellar tendinopathy

PDF

Control and dynamics of turning tasks with different rotation and translation requirements

PDF

Pathomechanics of femoroacetabular impingement syndrome: utilizing subject-specific modeling approaches to investigate the influence of hip joint morphology and neuromuscular control

PDF

Whole body mechanics of running turn maneuvers: relationship to injury and performance

Asset Metadata

Creator Rowley, Kevin Michael, Jr. (author)

Core Title Trunk control during dynamic balance: effects of cognitive dual-task interference and a history of recurrent low back pain

School School of Dentistry

Degree Doctor of Philosophy

Degree Program Biokinesiology

Publication Date 06/18/2018

Defense Date 05/07/2018

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag attention,Balance,dual-task,low back pain,movement-specific attention reinvestment,OAI-PMH Harvest,trunk control

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Kulig, Kornelia (committee chair), Armour Smith, Joanne (committee member), Finley, James (committee member), Gordon, James (committee member), Winstein, Carolee (committee member)

Creator Email krowley@usc.edu,rowleykmichael@gmail.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c40-508024

Unique identifier UC11267820

Identifier etd-RowleyKevi-6338.pdf (filename),usctheses-c40-508024 (legacy record id)

Legacy Identifier etd-RowleyKevi-6338.pdf

Dmrecord 508024

Document Type Dissertation

Format application/pdf (imt)

Rights Rowley, Kevin Michael, Jr.

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA