University of Southern California Dissertations and Theses

Bimanual coordination of goal-directed multi-joint rapid aiming movements following stroke

(USC Thesis Other)

Bimanual coordination of goal-directed multi-joint rapid aiming movements following stroke

PDF

Download

Open document

Flip pages

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content BIMANUAL COORDINATION OF GOAL-DIRECTED MULTI-JOINT RAPID
AIMING MOVEMENTS FOLLOWING STROKE
Copyright 2004
by
Dorian K. Rose
A Dissertation Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(BIOKINESIOLOGY)
December 2004
Dorian K. Rose
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
U M I N um ber: 3 1 5 5 4 6 9
Copyright 2004 by
Rose, Dorian K.
All rights reserved.
INFORMATION TO USERS
The quality of this reproduction is dependent upon the quality of the copy
submitted. Broken or indistinct print, colored or poor quality illustrations and
photographs, print bleed-through, substandard margins, and improper
alignment can adversely affect reproduction.
In the unlikely event that the author did not send a complete manuscript
and there are missing pages, these will be noted. Also, if unauthorized
copyright material had to be removed, a note will indicate the deletion.
®
UMI
UMI Microform 3155469
Copyright 2005 by ProQuest Information and Learning Company.
All rights reserved. This microform edition is protected against
unauthorized copying under Title 17, United States Code.
ProQuest Information and Learning Company
300 North Zeeb Road
P.O. Box 1346
Ann Arbor, Ml 48106-1346
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
ii
ACKNOWLEDGEMENTS
Although the title page suggests this document was the work of one person,
my desire is for these Acknowledgement pages to recognize the contribution
of many individuals in bringing this dissertation to fruition. Firstly, this
dissertation is the product of the sixty selfless individuals who cheerfully
volunteered their time and whose arm movements form the basis for the
information contained herein. I particularly want to acknowledge the thirty
stroke survivors who gave of their time for the future benefit of others.
I am indebted to my advisor, Dr. Carolee Winstein, for the time, energy and
instruction she has devoted to me and to the completion of this work. Her
inquisitive mind, devotion to detail and drive for excellence have been a
constant example to me of the characteristics underlying a true scientist and
superb academician. I am also grateful for the contributions of my committee
members: Dr. Stanley Azen, Dr. Lucinda Baker, Dr. James Gordon, Dr.
Stefan Schaal, and Dr. Steve Schreiber. This work is richer and stronger
because of their instruction and contribution.
I am thankful for the support and friendship of my colleagues in the Motor
Behavior and Neurorehabilitation Laboratory - for those who have gone
before - Dr. Lara Boyd and Dr. Pan Onla-or and for those who are following
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
iii
along this path - Jarugool Tretriluxana, Chein-Ho Lin and Jill Stewart. My
labmates Chelle Prettyman, Samantha Underwood and Tara Klassen have
infused this process with their energy, support and laughter. I have been
fortunate beyond measure to observe and benefit from the minds and hearts
of two dear mentors and friends, Dr. Beth Fisher and Dr. Kathy Sullivan.
They balance so beautifully the “ the science of healing” and “ the art of
caring” - both foundational to the profession of physical therapy. I have
learned and look forward to continuing to learn from them both.
I thank DPT students Jenn Penn and Chris Hahn for their assistance with
data analysis and Allan Wu, M.D. for assistance with lesion location analysis.
Joe Ye, Brenda Wessel and Sam Ward contributed their expertise to the
electronics and hardware used for this dissertation.
I am grateful to my community of faith, The Gathering Place, whose love,
support, and encouragement kept me grounded, balanced and moving
forward. I owe a debt of gratitude to my closest friend and dearest
confidante, Millie Anderson, who faithfully and tirelessly sustained me
through this process. Finally, I reserve my deepest gratitude and respect for
my parents for their unfailing and unconditional love and faithful support
throughout this endeavor from beginning to end. They stood by me and
never stopped believing in me. They are my friends and allies.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
iv
Funding for this work was provided by the Department of Biokinesiology and
Physical Therapy at the University of Southern California, the Neurology
Section of the American Physical Therapy Association (Mary Lou Barnes
Scholarship), the Foundation for Physical Therapy (PODS Scholarships
Levels I and II) and the Jacquelin Perry Scholarship.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
V
TABLE OF CONTENTS
Acknowledgements
List of Tables
List of Figures
Abstract
Chapter 1
Chapter 2
Chapter 3
Chapter 4
Chapter 5
THE QUESTION:
INTRODUCTION AND OVERVIEW
THE COORDINATION OF BIMANUAL
RAPID AIMING MOVEMENTS
FOLLOWING STROKE
THE COORDINATION OF BIMANUAL
ASYMMETRIC RAPID AIMING
MOVEMENTS: THE EFFECT OF
DIMINISHED CENTRAL DRIVE
TEMPORAL AND SPATIAL
CONSTRAINTS IN A SPATIALLY
ASYMMETRIC TASK: THE
EFFECT OF DIMINSHED CENTRAL
DRIVE
SUMMARY AND GENERAL DISCUSSION
VI
viii
xvi
1
8
48
101
156
References 166
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
vi
LIST OF TABLES
Chapter 2:
Table 2.1. Summary of Stroke Participant Characteristics 18
Table 2.2. Group means and standard deviations for the 22
primary outcome measures of Movement Time (MT),
Time to Peak velocity (TTP), Time after Peak velocity (TAP),
Peak Resultant Velocity (PRV), Reaction Time (RT)
Chapter 3:
Table 3.1. Experimental conditions 59
Table 3.2. Summary of stroke participants arranged 66
hierarchically according to Upper Extremity Fugl-Meyer
motor score
Table 3.3. Experiment I: Congruent Condition 69
(Paretic near/Nonparetic far; Left near/Right far for controls)
Group means and standard deviations for the primary outcome
measures of Reaction Time (RT), Movement Time (MT),
Time to Peak velocity (TTP), Time after Peak velocity (TAP),
Peak Resultant Velocity (PRV).
Table 3.4. Experiment II: Incongruent Condition 78
(Paretic far/Nonparetic near; Left far/Right near for controls)
Group means and standard deviations for the primary outcome
measures of Reaction Time (RT), Movement Time (MT),
Time to Peak velocity (TTP), Time after Peak velocity (TAP),
Peak Resultant Velocity (PRV).
Chapter 4:
Table 4.1. Seven experimental conditions 113
Table 4.2. Summary of stroke participants arranged 118
hierarchically according to Upper Extremity Fugl-Meyer
motor score
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
vii
Table 4.3. Group means and standard errors of percent
variance in nonbarrier (NB) limb behavior explained
by the barrier limb performance, derived from the
line of best fit for each participant
Table 4.4. Pearson correlation (r) between FM motor
score and interlimb coupling strength
127
143
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
viii
LIST OF FIGURES
Chapter 2:
Figure 2.1. Experimental set-up showing participant position 14
and target (overhead view). Planes of motion: x = medial/lateral;
y = fore/aft; z = vertical.
Figure 2.2. Line graphs of group means by aiming condition 20
For reaction time. Stroke (square) and control (circle) group by
aiming condition for each limb (paretic/left, dashed line;
nonparetic/right, solid line); error bars are SE. Nonparetic limb
reaction time was prolonged in the bimanual compared to the
unimanual condition (p< .05).
Figure 2.3. Line graphs of group means by aiming condition 21
For movement time. Stroke (square) and control (circle) group by
aiming condition for each limb (paretic/left, dashed line;
nonparetic/right, solid line); error bars are SE. Nonparetic limb
movement time prolonged in the bimanual compared to the
unimanual condition leaving only a 12 ms movement time difference
between the two limbs (Group x Limb x Aiming Type; p< .05).
Figure 2.4. Line graphs of group means by aiming condition 24
time after peak velocity. Stroke (square) and control (circle) group
by aiming condition for each limb (paretic/left, dashed line;
nonparetic/right, solid line); error bars are SE. Time after peak
velocity was generally longer for both groups and limbs in the
bimanual compared to the unimanual condition. This aiming
condition effect was most pronounced for the nonparetic limb,
leaving only a 15 msec difference between the two limbs in the
bimanual condition (Group x Limb x Aiming Type; p< .05).
Figure 2.5. Line graphs of group means by aiming condition 25
for peak resultant velocity. Stroke (square) and control (circle)
group by aiming condition for each limb (paretic/left, dashed line;
nonparetic/right, solid line); error bars are SE. Nonparetic limb
PRV was lower whereas paretic limb PRV was higher in the
bimanual compared to the unimanual aiming condition (p < .05).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
ix
Figure 2.6. Ensemble average time series of instantaneous 26
velocity plotted from movement onset to offset for stroke
participant (FM = 55). A. paretic limb profile for unimanual (solid line)
and bimanual (dashed line) aiming. Note greater paretic limb
PRV in bimanual compared to unimanual aiming (p<.05).
B. nonparetic limb profile for unimanual (solid line) and
bimanual (dashed line) aiming. Note greater nonparetic limb
PRV in unimanual compared to bimanual aiming (p<.05).
C. Bimanual velocity profile for paretic (solid line) and nonparetic
(dashed line) limb. Although the nonparetic limb reaches
peak velocity prior to the paretic limb, the deceleration phase
is adjusted to enable near simultaneous target hit.
Figure 2.1. Ensemble average time series of instantaneous 29
velocity plotted from movement onset to offset for stroke participant
(FM = 65). A. paretic limb profile, B. nonparetic limb profile for
unimanual (solid line) and bimanual (dashed line) aiming,
C. Bimanual velocity profile for paretic (solid line) and nonparetic
(dashed line) limb.
Figure 2.8. Ensemble average time series of instantaneous 31
Velocity plotted from movement onset to offset for control participant.
A. Left limb profile for unimanual (solid line) and bimanual
(dashed line) aiming, B. Right limb profile for unimanual (solid line)
and bimanual (dashed line) aiming, and C. Bimanual velocity
profile for left (solid line) and right (dashed line) limb.
Chapter 2: Postscript
Figure Postscript. 1 Time to Peak velocity for subgroup 45
of 19 stroke participants with greater paretic limb PRV in bimanual
compared to unimanual aiming. No difference in TTP between
the two aiming conditions (p > 0.05); error bars are SE.
Figure Postscript. 2 Peak resultant velocity for subgroup 46
of 19 stroke participants with greater paretic limb PRV in bimanual
than unimanual aiming (p < 0.05); error bars are SE.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
X
Figure Postscript. 3 Ensemble average time series of 46
instantaneous velocity plotted from movement onset to offset
for stroke participant (FM = 55). Paretic limb profile for unimanual
(solid line) and bimanual (dashed line) aiming. Greater PRV for
bimanual compared to unimanual aiming, but time to peak
velocity in each aiming condition is similar.
Chapter 3:
Figure 3.1. Congruent and incongruent aiming conditions 53
Figure 3.2. Experiment I: Congruent condition set-up 60
showing participant position and targets (overhead view) for
A. bimanual aiming: paretic near (left for controls) / nonparetic far
(right for controls), B. Unimanual aiming: paretic near
(left for controls), C. Unimanual aiming: nonparetic far
(right for controls). Planes of motion:x = medial/lateral; y = fore/aft;
z = vertical. P = paretic NP = nonparetic, L = left, R = right.
Figure 3.3. Experiment II: Incongruent condition set-up showing 61
participant position and targets (overhead view) for A. bimanual
aiming: paretic far (left for controls)/ nonparetic near
(right for controls),B. Unimanual aiming: paretic far
(left for controls), C. Unimanual aiming: nonparetic near
(right for controls). Planes of motion: x = medial/lateral; y = fore/aft;
z = vertical. P = paretic, NP = nonparetic, L = left, R = right.
Figure 3.4. Line graphs of group means by aiming condition 68
for reaction time. Stroke (square) and control (circle) group by
aiming condition for each limb (paretic/left, dashed line;
nonparetic/right, solid line); error bars are SE. RT prolonged for
bimanualaiming in stroke group alone(Group x Aiming condition
interaction p < 0.05; Group x Aiming condition x Limb, ns).
Figure 3.5. Line graphs of group means by aiming condition 71
for movement time. Stroke (square) and control (circle) group by
aiming condition for each limb (paretic/left, dashed line;
non paretic/right, solid line); error bars are SE. Nonparetic limb MT
was prolonged when paired with the paretic limb compared to the
unimanual condition to allow for just a 19 msec difference between
the limbs (Group x Aiming condition x Limb interaction, p < 0.05).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
xi
Figure 3.6. Line graphs of group means by aiming condition for 72
TTP. Stroke (square) and control (circle) group by aiming condition
for each limb (paretic/left, dashed line; nonparetic/right, solid line);
error bars are SE. Both limbs of the stroke group exhibited a similar
prolonged TTP in the bimanual compared to the unimanual
aiming condition to attain peak velocity nearly simultaneously
(Group x Aiming condition interaction p < 0.05;
Group x Aiming condition x Limb, ns).
Figure 3.7. Line graphs of group means by aiming condition for 73
TAP. Stroke (square) and control (circle) group by aiming condition
for each limb (paretic/left, dashed line; nonparetic/right, solid line);
error bars are SE. Increased TAP for control left limb aiming near
and stroke nonparetic limb aiming far
(Group x Aiming condition x Limb interaction, p < 0.05.)
Figure 3.8. Line graphs of group means by aiming condition for 75
PRV. Stroke (square) and control (circle) group by aiming condition
for each limb (paretic/left, dashed line; nonparetic/right, solid line);
error bars are SE. Nonparetic limb PRV decreased when paired
with the paretic limb compared to its unimanual performance
(Group x Aiming condition x Limb interaction, p < 0.05).
Figure 3.9. Line graphs of group means by aiming condition for RT. 79
Stroke (square) and control (circle) group by aiming condition for
each limb (paretic/left, dashed line; nonparetic/right, solid line);
error bars are SE. Nonparetic limb increased in the bimanual
condition when paired with the paretic limb aiming to the far
target (Group x Aiming condition x Limb interaction, p < 0.05.)
Figure 3.10. Line graphs of group means by aiming 80
condition for MT. Stroke (square) and control (circle) group
by aiming condition for each limb (paretic/left, dashed line;
nonparetic/right, solid line); error bars are SE. Nonparetic limb
increased in the bimanual condition when paired with the paretic
limb aiming to the far target (Group x Aiming condition x Limb
interaction, p < 0.05.)
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
xii
Figure 3.11. Line graphs of group means by aiming condition for 82
TTP. Stroke (square) and control (circle) group by aiming condition
for each limb (paretic/left, dashed line; nonparetic/right, solid line);
error bars are SE. Nonparetic limb TTP increased in the bimanual
compared to the unimanual condition
(Group x Aiming condition x Limb interaction, p < 0.05).
Figure 3.12. Line graphs of group means by aiming condition for 83
TAP. Stroke (square) and control (circle) group by aiming condition
for each limb (paretic/left, dashed line; nonparetic/right, solid line);
error bars are SE. Nonparetic limb TAP increased in the bimanual
condition when paired with the paretic limb aiming to the far target
(Group x Aiming condition x Limb interaction, p < 0.05).
Figure 3.13. Line graphs of group means by aiming condition 85
for PRV. Stroke (square) and control (circle) group by aiming
condition for each limb (paretic/left, dashed line;
nonparetic/right, solid line); error bars are SE. There was a
dissociation in peak resultant velocity, with nonparetic PRV
decreased, but paretic limb PRV increased in bimanual
compared to unimanual aiming
(Group x Aiming condition x Limb interaction, p < 0.05).
Figure 3.14. Nonparetic limb modifications were greater 89
in the incongruent compared to congruent bimanual
aiming condition for RT, MT, TAP and PRV (p < 0.05).
Figure 3.15. Time series of instantaneous velocity plotted 90
from movement onset to offset of a single trial for a
representative control participant. Left limb velocity profile in
the unimanual compared to bimanual aiming condition.
Prolongation of movement time occurs in the deceleration
phase only during bimanual aiming. Limb is aiming to a near
target in both conditions.
Figure 3.16. Time series of nonparetic limb instantaneous 91
velocity plotted from movement onset to offset of a single trial
for a representative stroke participant (#4). Prolongation of
movement time occurs in both the acceleration and deceleration
phases during bimanual (incongruent condition in this example)
aiming. Limb is aiming to near target in both conditions.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
xiii
Figure 3.17. Percentage of trials with simultaneous 93
Movement initiation (dark bar) and movement termination
(light bar) for control and stroke group. P = paretic,
NP = nonparetic, L = left, R = right.
Chapter 4:
Figure 4.1. Experimental set-up depicting participant 110
position. The three different height vertical barriers, and
respective targets for each limb (overhead view). Planes of
motion: x = medial/lateral; y = fore/aft; z = vertical
Figure 4.2. Single trial displacement profiles over time from 121
the no barrier plus the three asymmetric aiming conditions
from a representative control and stroke participant, A. control
(left as nonbarrier limb), B. stroke # 4 (nonparetic as nonbarrier
limb), C. stroke # 4 (paretic as nonbarrier limb). In A. and B. the
nonbarrier limb adjusts in time and space to the barrier limb
behavior. In C., adjustments are present, but not as distinctly
parameterized as in A. and B.
Figure 4.3. Nonbarrier and barrier limb movement time data 123
from all 3 asymmetric aiming conditions with resulting best-fit line
and regression equation from a representative control and stroke
participant A. control (right as nonbarrier limb), B. stroke # 29
(nonparetic as nonbarrier limb), C. stroke # 29 (paretic as
nonbarrier limb).
Figure 4.4. Individual participant (dashed) and group 126
average(solid) best-fit lines for A. control (right as nonbarrier
limb) and, B. control (left as nonbarrier limb). Equation is of
average best-fit line. (L= left, R = right).
Figure 4.5. Individual participant (dashed) and group 128
average (solid) best-fit lines for A. stroke (nonparetic as
nonbarrier limb) and B. stroke (paretic as nonbarrier limb).
Equation is of average best-fit line.
(P = paretic, NP = nonparetic)
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
xiv
Figure 4.6. Individual participant (dashed) and group 130
average (solid) best-fit lines for A. control (right as nonbarrier
limb), and B. control (left as nonbarrier limb). Equation is of
average best-fit line. (L= left, R = right).
Figure 4.7. Individual participant (dashed) and group 132
average (solid) best-fit lines for A. stroke (nonparetic as
nonbarrier limb) and B. stroke (paretic as nonbarrier limb).
Equation is of average best-fit line.
(P = paretic, NP = nonparetic).
Figure 4.8. Box and whisker plots of individual R2 values for 133
each limb in the nonbarrier role. A. Temporal Coupling,
B. Spatial Coupling, Data points within the box are from the
25th to 75th percentile of the group distribution. Error bars or
"whiskers" from the edge of the box are drawn to the 5th and
95th percentile. The solid horizontal line within each box is the
mean F^for the data set (L=left, R=right,
NP = nonparetic, P = paretic, NB=nonbarrier).
Figure 4.9. Bar graph of group mean R2 for temporal and 135
spatial measures. Overall interlimb coupling was stronger
for HC (.61+/-.03%) than for ST (,38+/-03%);
Main Effect Group (p < .05). HC: no difference between
temporal and spatial coupling; ST: temporal coupling
(R2 = .44+/-.03) > spatial coupling (R2 = .31+/-.04).
Group x Coupling Type interaction (p < .05).
Figure 4.10. Best-fit line with equation derived from 137
regression of MT on PVD across the three asymmetric
aiming conditions from a representative control participant
for A. barrier limb, B. nonbarrier limb. Greater slope for
nonbarrier compared to barrier limb. MT increase across
barrier heights similar in both limbs but with PVD increase
greater in the barrier than nonbarrier limb.
Figure 4.11. Best-fit line with equation derived from 138
regression of MT on PVD across the three asymmetric
aiming conditions from a representative stroke participant
for A. barrier limb (paretic), B. nonbarrier limb (nonparetic).
Greater slope for nonbarrier compared to barrier limb.
MT increase across barrier heights similar in both limbs but
with PVD increase greater in the barrier than nonbarrier limb.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
XV
Figure 4.12. Best-fit line with equation derived from 139
regression of MT on PVD across the three asymmetric aiming
conditions from a representative stroke participant for
A. barrier limb (nonparetic), B. nonbarrier limb (paretic).
Greater slope for nonbarrier compared to barrier limb.
MT increase across barrier heights similar in both limbs
but with PVD increase greater in the barrier than nonbarrier limb.
Figure 4.13. Bar plot showing mean and SEM of the 140
MT/PVD slope of each limb in the barrier and nonbarrier limb
role. Nonbarrier limb slope > Barrier limb slope for each of the
limb-barrier combinations(p < 0.05). (L= left, R = right,
P = paretic, NP = nonparetic, MT = movement time,
PVD = peak vertical displacement)
Figure 4.14. Single trial time series of instantaneous 141
resultant velocity plotted from movement onset to
movement offset from a representative control and stroke
participant, A. control participant (right as nonbarrier limb),
B. stroke participant # 21, (nonparetic as nonbarrier limb)
C. stroke participant # 24, (paretic as nonbarrier limb).
Solid line is barrier limb, dashed line is nonbarrier limb.
Figure 4.15. Plot of reaction time by task barrier for each 144
group. Main Effect of Group, Main Effect of Barrier (p < 0.05).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
ABSTRACT
XV!
Bimanual coordination is characterized by a strong temporal relationship
between the limbs as they perform similar or even dissimilar tasks. To probe
the stability of temporal synchrony in the presence of decreased central
motor drive due to unilateral stroke, 30 individuals with stroke and 30 controls
performed rapid aiming movements in three experimental paradigms. For two
(symmetrical aiming and asymmetrical aiming), participants aimed for the
target (s) bimanually and unimanually. In the third, a barrier was placed
midway in the path of one limb while the contralateral limb’s path remained
barrier-free. Psychometric and kinematic measures were used to
characterize movement planning and execution across participant groups,
aiming conditions, and limbs.
The nonparetic limb exhibited prolonged movement planning and execution
in bimanual compared to unimanual aiming, allowing for nearly simultaneous
target impact. The locus for the prolonged execution was primarily in the
deceleration phase. Paretic limb kinematics were modified as well with peak
velocity higher in bimanual compared to unimanual aiming for two-thirds of
the stroke cohort. Adjustments of both limbs were observed not only for
symmetrical aiming but also when they moved asymmetric distances. There
were two levels of congruency for asymmetric aiming. In the more
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
incongruent condition, the between-limb temporal difference in target hit was
two and a half times greater for the stroke participants than for the controls.
Although both limbs exhibited adjustments in the bilateral compared to the
unilateral conditions, the more incongruent condition revealed limitations in
the natural temporal constraint on bimanual movements. In the barrier
experiment, between-limb temporal coupling was more robust to the barrier
than spatial coupling for the stroke group with diminished central motor drive.
Overall, inter-limb temporal synchrony was relatively strong in this mildly
motor impaired stroke group, in part because of the distributed neural control
for bimanual coordination. Although the majority of reaction and movement
time adjustments for bimanual coordination occurred with the nonparetic
limb, there was also an increase in paretic limb peak resultant velocity. This
facilitation effect provides a scientific rationale for investigation of the efficacy
of bimanual training approaches to post-stroke rehabilitation of upper limb
function.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
1
CHAPTER 1
Introduction and Overview
The Question
Remarkably, much of the goal-directed movement that humans engage in
involves the beautifully orchestrated and coordinated actions of both hands.
Examples of such movements range from, what appear to be, the simple
tasks of catching a ball, kneading dough or tying shoelaces to the artistry of
playing the piano or communicating with sign language. This coordinated
behavior appears effortless in most cases, suggesting that the controller, the
central nervous system (CNS) is constrained by a set of rules that naturally
reduces this coordinated control problem. However, in other cases, the
coordination of the two upper limbs for goal achievement presents an
interesting problem for the CNS, which in turn, presents an intriguing
experimental paradigm for studying the motor control of these actions.
Although investigations of bimanual coordination are relatively young, the
majority occurring over only the past 25 years, initial observations were made
in the infant stages of the field of movement science over 100 years ago by
Woodworth, “It is common knowledge that movements with the left and right
hand are easy to execute simultaneously. We need hardly to try at all for
them to be nearly the same” (Woodworth, 1903).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2
Motor tasks can be characterized with various taxonomies, discrete (tasks
that have a definite beginning and end) versus continuous (tasks that have
no recognizable beginning or end), is one such categorization. Tasks can
also be described using temporal and spatial features. Bimanual coordination
paradigms have used both types of tasks (continuous and discrete) and have
described interlimb coordination with respect to both time and space.
Overall, the overwhelming finding is that when two limbs move together, the
behavior of one hand is affected by the task requirements of the contralateral
hand (i.e. assimilation effect). When the two hands perform tasks of differing
difficulty, there is generally an asymmetry to this assimilation, with the limb
performing the more difficult task affecting the limb performing the easier task
to a greater degree than the converse. A second phenomenon of bimanual
coordination is that the two limbs are more constrained to behave similarly in
the temporal domain than in the spatial domain.
Those with contralateral paresis from a hemispheric stroke bring a unique
level of complexity to the study of bimanual coordination. The control of two
effectors with disparate capabilities presents an even greater motor control
challenge for the CNS. Do the “rules” that govern bimanual coordination of
effectors with similar characteristics apply to effectors of disparate
characteristics? This dissertation was designed to answer that question.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
3
There have been six investigations of bimanual coordination in cases with
adult onset hemiparesis and these have appeared in the literature within the
past ten years, with one exception. A behavioral analysis of two individuals
with stroke appeared over 50 years ago (Cohn, 1951). When each arm
performed a cyclic and continuous pronation - supination motion together, an
obvious decrement in performance occurred for the nonparetic limb with
movement frequency dramatically slowed compared to unilateral
performance. In addition, although not as striking, the movement trajectories
increased slightly for the paretic limb in frequency and regularity in the
bilateral compared to the unilateral condition. Cohn (1951) did not quantify
these results, but the qualitative images provided an intriguing impression of
what might be possible with bimanual coordination in cases of hemiparesis.
More recently, Rice and Newell studied a continuous temporally symmetric
(2001) and continuous temporally asymmetric (2004) elbow flexion and
extension motion with 18 post-stroke individuals. For the temporally
symmetric task, participants were asked to maintain a “comfortable speed”
oscillation, initially determined for each limb separately. In the bilateral
situation, the nonparetic limb was unable to maintain its unilateral oscillation
frequency but rather was constrained to the slower oscillation frequency of
the paretic limb. For the temporally asymmetric task, participants were to
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
4
oscillate one limb at twice the frequency of the other. Those with stroke were
unable to perform this less stable temporally asymmetric task but rather
adopted a more stable in-phase movement pattern of the limbs - flexion and
extension of the limbs together at the same rate. Lewis and Byblow (2004)
examined interlimb temporal and spatial coordination of a continuous circle-
drawing task in 9 post-stroke hemiparetic individuals. Similar to Rice and
Newell (2001, 2004), the paretic limb influenced the behavior of the
nonparetic limb and no improvements in the hemiplegic limb could be elicited
with the bimanual nature of the task.
Dickstein and colleagues (1993) and Cunningham and colleagues (2002)
investigated bimanual coordination of a spatially symmetric discrete motor
task, both of these studies again limited to behavior about the elbow joint.
Dickstein and her collaborators reported a prolonged movement time for the
nonparetic limb during bilateral elbow flexion compared to the unilateral
condition. Although there was an increase in paretic limb movement time as
well (11%), the increase for the nonparetic limb was greater (18%). An
examination of discrete elbow extension movements (Cunningham et al.,
2002) revealed a slight facilitation of the paretic limb in the bimanual
compared to the unimanual condition as measured by a smoother elbow
extension velocity profile.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Five of these six investigations have been limited to the examination of single
joint movements, specifically of the elbow (Rice and Newell, 2001, 2004;
Dickstein et a!., 1993; Cunningham et al., 2002) or forearm (Cohn, 1951) with
the sixth an examination of a ciosed-chain, non-goal directed circle drawing
task. The majority of upper limb movements that humans make involve
multiple joints and are goal-directed. Investigations of inter limb coordination
that use more natural tasks could provide a more valid proxy for the study of
bimanual coordination.
The aforementioned studies focused exclusively on temporal coordination
between the two limbs. Data from healthy individuals suggest that the two
limbs demonstrate a spatial coordination as well (Franz et al., 1991; Franz,
1997) but this has yet to be studied in those with disparate effectors from
hemispheric stroke. An examination of both temporal and spatial coordination
of the same task is needed to contribute to an understanding of the control of
bimanual movements. The previous investigations in cases of hemispheric
brain damage from stroke, suggest that, to varying degrees, each limb is
influenced by the contralateral limb’s motor characteristics. However, without
quantification of this inter limb influence, a systematic comparison of inter
limb coordination is not possible over time, across tasks, or with practice.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
6
Therefore, this dissertation focused on three distinct yet interrelated
purposes (specific aims):
1) To determine the impact of hemispheric stroke on movement
planning and movement execution of spatially symmetric bimanual
multi-joint goal-directed aiming movements.
2) To probe the robustness of the temporal constraint on movement
planning and movement execution after hemispheric stroke
through the examination of spatially asymmetric bimanual mulit-
joint goal-directed aiming movements.
3) To examine the relative strength of bimanual temporal and spatial
coordination after hemispheric stroke through the examination of
spatially asymmetric bimanual mulit-joint goal-directed aiming
movements.
Overview
This dissertation is organized into three separate, yet related experimental
paradigms. Each will be presented separately with its own purpose, methods,
results and discussion. All three experiments examine bimanual coordination
in adult volunteers who sustained a hemispheric stroke at least 6 months
earlier and in a group of age-matched healthy individuals. Experiment 1
examines movement planning and execution of a bimanual spatially
symmetric aiming task (Chapter 2). Next, this paradigm is extended to study
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
7
the effects of an external task asymmetry, with different amplitude
requirements for each limb, in addition to the internal asymmetry of unilateral
hemiparesis (Chapter 3). The third experiment examines spatial and
temporal coordination after unilateral stroke during a spatially asymmetric
task and presents a quantification algorithm of coordination strength (Chapter
4).
Chapter 2 at the time of this writing is in press in Clinical Rehabilitation and is
reprinted here for completeness. Following Chapter 2 is a postscript
containing considerations that arose as the subsequent experiments were
analyzed. Chapters 3 and 4 are manuscripts in preparation. Following the
three experimental paradigms each with results and discussion, a general
discussion and summary of the dissertation with future directions is outlined
(Chapter 5).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
8
CHAPTER 2
The Coordination of Bimanual Rapid Aiming Movements Following
Stroke
Abstract
Objective: To determine the role of anticipatory and movement control
processes for the coordination of bimanual target aiming in individuals
post-stroke.
Subjects: Thirty adults with chronic stroke and thirty individuals without
stroke history.
Design: A two group (stroke, control) by 2 aiming type (unimanual,
bimanual) by 2 limb (paretic, nonparetic; left, right for controls), design with
repeated measures on the last two factors.
Outcome measures: Kinematic analyses of performance and psychometric
measures of reaction time, movement time, peak resultant velocity, time to
and after peak resultant velocity and interlimb timing for movement initiation
and target impact.
Results: Compared to unimanual aiming, the nonparetic limb exhibited a
prolonged movement time in the bimanual condition; the locus for
prolongation was primarily in the deceleration phase. This adaptive response
allowed for a nearly simultaneous (both limbs) target impact in 81 % of trials.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
9
Compared to the unimanual condition, the nonparetic limb exhibited a lower
peak velocity (10%) in the bimanual condition. Conversely, compared to the
unimanual condition, the paretic limb exhibited a higher peak velocity (4%) in
the bimanual condition. This dissociation between limb and condition was
observed for the stroke group but not the control group.
Conclusions: The interlimb coordination that emerged for the stroke group
revealed a complex and asymmetric contribution from each limb mediated
through anticipatory and motor control processes. We suggest that this
coordination may be harnessed for future bimanual intervention approaches
to rehabilitation of upper limb function after stroke.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
10
Introduction
Upper extremity rehabilitation protocols for stroke hemiparesis have focused
on the paretic limb with unilateral strengthening exercises, neuromuscular re
education and/or functional training (Davies, 1985; Sunderland et al., 1994;
Butefisch et al., 1995). One recent approach, constraint-induced movement
therapy, exploits this focus by physically constraining the nonparetic limb with
a sling or safety mitt (Morris et al., 1997; Taub et al.1999). However, many
daily tasks naturally require the coordinated participation of both hands,
providing a rationale for a bimanual approach to upper limb rehabilitation. A
small but growing number of investigations have provided evidence for the
potential of bilateral training on the recovery of the paretic limb post-stroke
(Mudie and Matyas, 2000; Whitall et al., 2000; Cauraugh and Kim, 2002).
Although these studies reported paretic limb changes with a bimanual
intervention approach, the focus was primarily on global measures of
movement without any kinematic analyses of the movement trajectory.
Previous systematic investigations of healthy individuals reveal that limb
performance differs between unimanual and bimanual aiming tasks when the
bimanual task is asymmetric (Kelso et al., 1979a; 1979b; 1983; Martenuik et
al., 1984; Fowler et al., 1991). Results for unimanual and symmetrical
bimanual aiming movements concur with predictions from Fitts’ Law (Fitts,
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
11
1954; Fitts and Peterson, 1964); movement time increases as movement
amplitude or precision requirements increase. However, for bimanual
asymmetric goal-directed aiming, Fitts’ Law is violated. Specifically, in these
investigations, and of importance to our research, movement time of the limb
directed towards the easy (i.e. large) target is prolonged and therefore,
similar to the movement duration for the difficult (i.e. small) target limb. A
naturalistic twist on Fitts’ paradigm occurs for persons after stroke. With
hemiparesis, the task constraint is internal (body-centered) rather than
external (target-centered). After stroke-hemiparesis, the constraint imposed
on the system is the disparate movement ability of the two limbs (i.e. paretic;
nonparetic), not the disparate targets (i.e. easy; difficult)
Previous motor control studies of those who are hemiparetic performing
bilateral elbow flexion (Dickstein et al., 1993) or reach and grasp
(Steenbergen et al., 1996) tasks, report temporal differences in both
anticipatory movement planning (reaction time) and movement control
(movement time) processes between unimanual and bimanual task
conditions, but lack behavioral kinematic information to determine the locus
of these differences. To date, there are no published reports of a systematic
psychometric and kinematic analysis of upper limb movements for persons
with stroke hemiparesis that directly compares the performance of rapid
unimanual and bimanual goal-directed aiming. The purpose of this
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
12
experiment is to investigate anticipatory planning and movement control
processes involved in bimanual aiming for individuals with mild stroke-
induced hemiparesis. We hypothesize that the nonparetic limb (easy task)
will exhibit temporal adaptation to that of the paretic limb (difficult task) in
bimanual symmetrical aiming, and that this temporal adaptation will be
expressed in both the anticipatory and motor control domains.
Method
Participants
Thirty adults post-stroke (Table 2.1) and thirty adults without stroke (mean
age: 67 yrs; range 49 - 86) participated. All were right-hand dominant
(Oldfield, 1971). Stroke participants were screened to verify: (1) intact visual
fields (Friedman, 1992); (2) intact upper limb proprioception (Fugl-Meyer et
al., 1975); (3) pain-free, active range of motion to at least 120° shoulder
flexion; at least 90° forearm supination and full motion of the elbow, wrist and
fingers into extension; and (4) Modified Ashworth Spasticity Scale
(Bohannon and Smith, 1987) scores for shoulder extensors, elbow flexors,
forearm supinators, wrist flexors < 2. Motor impairment was assessed with
the Fugl-Meyer Measurement of Physical Performance (FM) (Fugl-Meyer et
al., 1975) and grip and pinch strength (Mathiowetz et al., 1985).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
13
Task
Initial position for the seated participants was with shoulders in neutral
alignment (0° flexion, 0° abduction, 0° rotation), 90° elbow flexion, and 90°
degree forearm pronation. In response to an LED signal, participants were to
reach and aim forward with one hand (unimanual) or both hands (bimanual)
to hit a switch(es) mounted on the LED target, as soon and as fast as
possible. The movement required shoulder flexion and horizontal adduction
with elbow extension and forearm supination. The trunk was stabilized with a
chest strap so that task achievement was primarily accomplished through
upper limb movements.
Apparatus
The apparatus consisted of two sets of 2.5 cm2 switches; home-position (one
for each hand) and target-position (one for each hand). Each switch
consisted of a solid-state infrared photo light emitter and diode. A 0.5 x 1 cm
blackened mylar sheath was mechanically mounted to each of the four
switches. When the light beam from the photo emitter to diode was broken by
the mylar, the switches were considered on (home switches) or off (target
switches). The home-position switches were secured to the table surface,
aligned for each participant in midline of the long axis of the humerus. The
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
14
target was an LED embedded in the center of a 3 cm width wooden block
with the target-position switches located on either side. Target location was
determined individually in 3-D space: x-plane at body midline, y-plane such
that target hit required no more than 15° of full elbow extension, z-plane such
that target hit required 90° forward shoulder flexion (Figure 2.1). Aiming hand
end-point displacement was recorded from an electromagnetic sensor (Skill
Technologies, Inc.) attached to a rubberized paddle (10.5 cm diameter),
affixed to each hand. Movements were recorded with a “6-D Research”
electromagnetic motion capture and analysis system (Skill Technologies,
Inc.) using a 60 Hz sampling rate. Trial onset, LED onset, and movement
onset and offset of each hand were sampled on separate channels from
individual TTL pulses interfaced with a 400 MHz personal computer.
LED
z
Figure 2.1. Experimental set-up showing participant position and target for
Bimanual aiming (overhead view). Planes of motion: x = medial/lateral; y =
fore/aft; z = vertical.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
15
Procedure
Participants signed an informed-consent form approved by the Institutional
Review Board from the Health Sciences Campus of the University of
Southern California. Verbal instructions and demonstration of the task were
provided at the beginning of the session. Participants sat at a height-adjusted
table with elbows at 90° flexion, shoulders and wrists at 0° of flexion and
extension and abduction and adduction, and forearms in pronation at the
start position. Each participant performed one block of 5 practice trials before
each task condition. They were instructed to move “ as soon as possible and
as fast as possible” once the LED was illuminated, to hit the target from
either side. In the bimanual condition, no explicit instructions were provided
as to how the two limbs were to be coordinated (e.g., to initiate or terminate
the movements together).
Each trial began with a verbal ready signal followed by the LED presented
randomly within 500 to 1500 msec. The trial ended when the target
switch(es) was hit. Participants returned their hand(s) to the home position
and awaited the next “ready” signal. There were 3 aiming conditions: 1.
Bimanual, 2. Unimanual with the paretic arm (left for controls), 3. Unimanual
with the nonparetic arm (right for controls), presented in 3 separate, 28-trial
blocks (25 valid trials and 3 catch trials - a catch trial was one in which no
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
16
LED followed the “ready” signal). Intermittently, during each trial block,
reminders were provided to “move as soon and as fast as possible” in
response to the LED signal. The order of the three aiming conditions was
counterbalanced across participants with a 2 minute rest period provided
between each one.
Data reduction
For each trial, raw displacement data were filtered with a 0-lag Butterworth
filter (low-pass cutoff frequency of 20 Hz). Instantaneous velocity was
calculated by differentiating electromagnetic sensor displacement data using
a three-point moving average. Resultant velocity was calculated from the x, y
and z velocity values. Trial onset, LED onset, switch onset and offset, and
sensor resultant velocity were incorporated into one data file for each aiming
condition. Data files were analyzed with Datapac 2000 Laboratory
Applications System (RUN Technologies Co.). Movement onset was the time
value at which the sensor acceleration value exceeded 0.03 cm/sec2.
Movement offset was the time value when the target-position switch signal
exceeded 0 mV. Movement time was defined as the difference between time
of movement offset and time of movement onset, reaction time as the time
from LED onset to movement onset. For a trial to be considered for analysis
both hands had to initiate squarely from the onset switches and hit their
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
17
respective target switches. Additionally, trials with reaction times that were
greater or less than the mean +1-2 SD for that trial block were discarded. If
more than 20 acceptable trials remained following application of these
criteria, the middle 20 trials were used for analysis.
Statistical analysis
A participant mean was calculated per condition and limb for each dependent
measure: reaction time, movement time, time to peak velocity, time after
peak velocity, and peak resultant velocity. To test for effects of group (stroke,
control), aiming type (unimanual, bimanual), and limb (paretic, nonparetic;
left, right for controls) on these dependent measures, separate 2 (Group) x 2
(Aiming Type) x 2 (Limb) general linear model ANOVAs with repeated
measures on the last two factors were performed. Significance was set at p<
.05 for all comparisons. Post-hoc pairwise comparisons with a Bonferroni
correction determined the locus of any significant interaction.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
18
Results
Clinical Motor Impairment
Participants were considered mildly impaired, with the capability to isolate
elbow extension from shoulder flexion to complete the task (Duncan et al.,
1994). Group mean Fugl-Meyer motor score was 60+/- 5 out of 66 total
possible score. Grip and pinch strength of the paretic limb provided a more
sensitive measure of impairment, especially in this mildly-impaired group.
Although participants were able to perform the experimental task, upper limb
motor deficits remained (Table2.1).
Table 2.1. Demographic summary of stroke participants
Age
(yrs)
Sex Side
of
Lesion
Time
from
Onset
(mos)
FM
Motor
Score
Grip
(%)
LP
(%)
PP
(%)
Mean 63 18 M
12 F
18
LCVA
12
RCVA
29 60 65 71 63
Range 27-84 “ - 6-120 43-66 27-
100
27-
100
24-
100
Grip, Lateral Pinch (LP) and Palmar Pinch (PP) values represent percentage
strength of paretic (P) limb compared to limb, age and gender matched
normative data (Mathiowetz et al., 1985).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
19
Effect of Aiming Type
Reaction Time
The Group x Aiming Type x Limb interaction for reaction time did not reach
statistical significance (p= .09). However, because our primary interest
concerned the effects of aiming condition on reaction time, we conducted
within-group pairwise comparisons using reaction time from each aiming
condition. For the stroke group, nonparetic limb reaction time was prolonged
by 18 (+/-8) msec in the bimanual compared to the unimanual condition (p<
.05). In contrast, there were no aiming condition differences in reaction time
for the paretic limb of the stroke or either limb of the control group (Figure
2.2, Table 2.2).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2 0
300
o
- - © - ■ control left
— • — control right
- -E l- ■ stroke paretic
— ■ — stroke nonparetic
a >
E
c
150
UNIMANUAL BIMANUAL
Figure. 2.2. Line graphs of group means by aiming condition for reaction time
Stroke (square) and control (circle) group by aiming condition for each limb
(paretic/left, dashed line; nonparetic/right, solid line); error bars are SB.
Nonparetic limb reaction time was prolonged in the bimanual compared to
the unimanual condition (p< .05).
Movement Time
Movement Time was generally prolonged for both groups and limbs in the
bimanual compared to the unimanual aiming condition. However, this aiming
condition effect was most pronounced for the nonparetic limb in the stroke
group, resulting in a three-way interaction (Group x Limb x Aiming Type; p<
.05; Figure 2.3). Nonparetic limb movement time was prolonged by 21% in
the bimanual compared to the unimanual condition leaving only a 12 ms
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
21
movement time difference between the two limbs in the bimanual condition.
This result supports our hypothesis that the nonparetic limb exhibits temporal
adaptation to that of the paretic limb in a bimanual symmetrical aiming task.
This 12 msec difference stands in contrast to the 110 msec average
difference between the two limbs in the unimanual condition. Additionally,
the 113 (+/-14) msec increase in movement time for the nonparetic limb is 5
to 7 times greater than that observed for the paretic limb or for either limb of
the control group in the bimanual condition (Table 2.2).
control left
— • — control right
- - Q - - stroke paretic
— ■ — stroke nonparetic
UNIMANUAL BIMANUAL
Figure 2.3.Line graphs of group means by aiming condition for movement
time. Stroke (square) and control (circle) group by aiming condition for each
limb (paretic/left, dashed line; nonparetic/right, solid line); error bars are SE.
Nonparetic limb movement time prolonged in the bimanual compared to the
unimanual condition leaving only a 12 ms movement time difference between
the two limbs (Group x Limb x Aiming Type; p< .05).
575
¥
w
£
© 500
c
0
E
C D
>
O
425
350
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2 2
Table 2.2. Group means and standard deviations for the primary outcome
measures of Reaction Time (RT), Movement Time (MT), Time to Peak
velocity (TTP), Time after Peak velocity (TAP), Peak Resultant Velocity
(PRV).__________________________________________________________
Control
Left Right
Unimanual Bimanual Unimanual Bimanual
Mean SD Mean SD Mean SD Mean SD
MT 362 57 382 57 352 50 376 53
(ms)
TTP 179 36 180 31 170 29 174 31
(ms)
TAP 182 37 202 42 183 44 202 42
(ms)
PRV 195 40 196 35 192 32 189 30
(cm/s)
RT 202 24 206 27 197 27 209 24
(ms)
Stroke
Paretic Nonparetic
Unimanual Bimanual Unimanual Bimanual
Mean SD Mean SD Mean SD Mean SD
MT 526 150 541 145 417 75 529 142
(ms)
TTP 231 50 237 52 201 33 210 40
(ms)
TAP 294 139 304 126 222 77 319 120
(ms)
PRV 156 33 163 35 174 31 156 31
(cm/s)
RT 280 88 273 65 253 98 272 68
(ms)
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
23
To more closely examine aiming performance, velocity profiles of the hand
were parceled into two sub-phases: 1) acceleration phase, time to peak
velocity and 2) deceleration phase, time after peak velocity. Time to peak
velocity was generally prolonged for both groups in the bimanual compared
to the unimanual condition (p< .05), but this effect was similar across limbs
(Group x Limb x Aiming Type interaction; NS). In contrast, time after peak
velocity mirrored movement time; time after peak velocity was generally
longer for both groups and limbs in the bimanual compared to the unimanual
condition. This aiming condition effect was most pronounced for the
nonparetic limb. The 97 (+/-13) msec average time after peak velocity
prolongation for the nonparetic limb was 10 times longer than that of the
paretic limb and 5 times longer than either limb of the controls resulting in a
three-way interaction (Group x Limb x Aiming Type; p< .05; Figure 2.4).
Although overall movement time, on average, was 12 msec shorter for the
nonparetic compared to the paretic limb in the bimanual condition (Figure
2.3), time after peak velocity was, on average, 15 msec longer for the
non paretic limb (Figure 2.4, Table 2.2).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
24
o
C D
C /3
E,
J s < £
C D
C D
Q _
s _
C D
C D
C D
E
h-
350 -i
300
250
200 -
150
■ -©■ • control left
— 9 — control right
■-E-- stroke paretic
H i — stroke nonparetic
UNIMANUAL BIMANUAL
Figure 2.4. Line graphs of group means by aiming condition for time after
peakvelocity. Stroke (square) and control (circle) group by aiming condition
for each limb (paretic/left, dashed line; nonparetic/right, solid line); error bars
are SE. Time after peak velocity was generally longer for both groups and
limbs in the bimanual compared to the unimanual condition. This aiming
condition effect was most pronounced for the nonparetic limb, leaving only a
15 msec difference between the two limbs in the bimanual condition (Group x
Limb x Aiming Type; p< .05).
Velocity
Those with stroke demonstrated a dissociation in peak resultant velocity
between aiming conditions for each limb. Nonparetic limb peak resultant
velocity was 10% lower in the bimanual compared to the unimanual condition
(p < .05). In contrast, and surprisingly, paretic limb peak resultant velocity
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
25
was 4% higher in the bimanual compared to the unimanual aiming condition
(p < .05). Although this bimanual condition benefit in paretic limb peak
resultant velocity appears to be relatively small, it stands in contrast to no
significant difference in peak resultant velocity between the two conditions for
either limb of those in the control group (Group x Limb x Aiming condition
interaction; p < .05; Figure 2.5, Table 2.2).
225
o
0 )
C O
E
£•
o
o
0 )
>
c
03
C/3
C D
a:
T O
0
Q .
200
175
150 -
control left
— • — control right
--E I-- stroke paretic
■" stroke nonparetic
1 --------
UNIMANUAL BIMANUAL
Figure 2.5. Line graphs of group means by aiming condition for peak
resultant velocity. Stroke (square) and control (circle) group by aiming
condition for each limb (paretic/left, dashed line; nonparetic/right, solid line);
error bars are SE. Nonparetic limb PRV was lower whereas paretic limb PRV
was higher in the bimanual compared to the unimanual aiming condition (p <
.05).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
26
To more clearly illustrate the similarities and differences in kinematics
between aiming conditions, ensemble average velocity profiles from two
individuals with stroke and one without stroke are illustrated next. Two
examples from the stroke group are provided to demonstrate the
heterogeneity of upper limb motor control even within this select group of
individuals post-stroke. The individual presented in Figure 2.6A - C (FM =
55) is representative of the stroke group mean data with a higher peak
velocity for the paretic limb in the bimanual condition. Paretic limb movement
duration in the two aiming conditions is similar with velocity adjustments
evident in the deceleration phase for both conditions (Figure 2.6A).
Consistent with group data, peak velocity is higher in the bimanual than
unimanual condition.
Paretic Upper Extremity
250
Unimanual
200
Bimanual
150
100
0 100 200 300 400 500 600 700 800
Tim e (msec)
Figure 2.6A. Ensemble average time series of instantaneous velocity plotted
from movement onset to offset for stroke participant (FM = 55). Paretic limb
profile for unimanual (solid line) and bimanual (dashed line) aiming. Note
greater paretic limb PRV in bimanual compared to unimanual aiming (p<,05).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
27
For the nonparetic limb, time to peak velocity is similar in the two aiming
conditions, with the deceleration phase (time after peak velocity) prolonged
by 200 msec in the bimanual condition. Additionally, and consistent with the
group mean, nonparetic limb peak velocity is lower in the bimanual compared
to the unimanual condition (Figure 2.6B).
B.
Nonparetic Upper Extremity
o
f ll)
(/)
E
£*
o
o
C D
>
250
Unimanual
200
Bimanual
150
100
50
0
0 100 200 500 300 400 600 700 800
Time (msec)
Figure 2.6B. Ensemble average time series of instantaneous velocity plotted
from movement onset to offset for stroke participant (FM = 55). Nonparetic
limb profile for unimanual (solid line) and bimanual (dashed line) aiming.
Note greater nonparetic limb PRV in unimanual compared to bimanual
aiming (p< 0.05).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
28
Paretic and nonparetic profiles are plotted together in Figure 2.6C for the
bimanual condition. This figure illustrates that although the nonparetic limb
reaches peak velocity earlier than does the paretic limb, the deceleration
phase is prolonged to enable near simultaneous target hit. An additional
eighteen stroke participants produced similar profiles. In sum, approximately
63% of the sample (19/30) exhibited similar kinematics to that illustrated in
Fig 6A. -C.
C.
o
0
tn
E
S
"o
o
0
>
Bimanual Paretic vs Nonparetic
250
Paretic
200
Nonparetic
150
100
50
0
600 700 800 0 100 200 300 400 500
Time (msec)
Figure 2.6C. Ensemble average time series of instantaneous velocity plotted
from movement onset to offset for stroke participant (FM = 55). Bimanual
velocity profile for paretic (solid line) and nonparetic (dashed line) limb.
Although the nonparetic limb reaches peak velocity prior to the paretic limb,
the deceleration phase is adjusted to enable near simultaneous target hit.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
29
Velocity profiles represented in Figure 2.7k. - C. (FM = 65) differ from the
previous example in several ways. First, peak velocity for both the paretic
and nonparetic limb is very similar in both aiming conditions. Second, both
limbs exhibit a prolonged movement time in the bimanual compared to the
unimanual condition. However, consistent with the previous example,
movement time in the bimanual condition is similar for the two limbs to
accommodate a nearly simultaneous target hit (Figure 2.7C.). Ten additional
stroke participants produced this type of profile (11/30, 36% of the sample).
A.
o
03
C O
E
3
S
o
o
C D
>
Paretic Upper Extremity
250
Unimanual
200
Bimanual
150
100
50
0
200 600 700 800 0 100 300 400 500
Time (msec)
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
30
B.
Non-Paretic Upper Extremity 250
200
Unimanual
o
Q >
< f )
E
o
— Bimanual
150
8
CD
>
100
0 100 200 300 400 500 600 700 800
Tim e (msec)
C.
Bimanual Paretic vs Nonparetic
250
200
Paretic
Nonparetic
150
£
8
CD
>
100
0 100 200 300 400 500 600 700 800
Tim e (msec)
Figure 2.7. Ensemble average time series of instantaneous velocity plotted
from movement onset to offset for stroke participant (FM = 65). A.
nonparetic limb: unimanual (solid line) and bimanual aiming (dashed line), B.
paretic limb: unimanual (solid line) and bimanual (dashed line) aiming, C.
Bimanual velocity profile for paretic (solid line) and nonparetic (dashed line)
limb.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
31
There were no significant differences in upper limb Fugl-Meyer motor score
between the two subsets of stroke participants (those who exhibited a
greater paretic limb PRV in the bimanual vs. unimanual condition and those
who did not), suggesting that it is not simply the degree of motor impairment
that dictates bimanual coordination. Left and right limb velocity profiles of a
control participant are shown in Figure 2.8A. - 8C; each limb closely
mirroring the other in the two aiming conditions. This pattern is representative
of all those in the control group.
A.
250
Left Upper Extremity
Unimanuai
0
■ - ' Bimanual
0 100 200 300 400 500 600 700 800
Time (msec)
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
250 r
Right Upper Extremity
32
Unimanual
200
— Bimanual
150
100
50
0
0 100 200 300 400 600 700 500 800
Tim e (msec)
250
Bimanual Left vs Right
LEFT 200
o
0 3
to
E
o
RIGHT
150
£
'8
d )
>
100
400 600 800 0 100 200 300 500 700
Time (msec)
Figure 2.8. Ensemble average time series of instantaneous velocity plotted
from movement onset to offset for control participant. A. Left limb profile for
unimanual (solid line) and bimanual (dashed line) aiming, B. Right limb
profile for unimanual (solid line) and bimanual (dashed line) aiming, and C.
Bimanual velocity profile for left (solid line) and right (dashed line) limb.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Bimanual Movement Initiation and Termination
33
Movement initiation of the two limbs was considered simultaneous if it
occurred within 2 samples (< 34 msec) of each other. This same criterion
was used for movement termination. This criterion was established as a
conservative threshold given the temporal resolution of the electromagnetic
sensors and the sampling rate of 60 Hz. With this criterion, 95% of trials for
the control group were both initiated and terminated simultaneously. For the
stroke group, 80% of trials were initiated and 81% were terminated
simultaneously.
Discussion
We hypothesized that the nonparetic limb would exhibit temporal adaptation
to that of the paretic limb in both anticipatory and motor control domains for a
bimanual symmetrical aiming task. For the first time, this study provides a
kinematic analysis of the performance differences between unimanual and
bimanual aiming tasks, and with that, determines the locus of these effects
for persons post-stroke.
Compared to unimanual aiming, the nonparetic limb exhibited a prolonged
movement time in the bimanual condition; the locus for prolongation was
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
34
predominantly in the deceleration phase. Further, and compared to the
unimanual condition, the nonparetic limb exhibited a lower peak velocity in
the bimanual condition. Conversely, compared to the unimanual condition,
the paretic limb exhibited a higher peak velocity in the bimanual condition. It
is of particular interest that this dissociation between limb and task condition
was observed for the stroke group but not the control group.
A parallel can be drawn between previous studies of healthy adults in which
the hand going to the easy target slows down to become more like the hand
going to the hard target (Kelso et al., 1979a; 1979b; 1983; Martenuik et al.,
1984; Fowler et al., 1991) and the present results, in which the nonparetic
limb slows down to become more like the paretic limb. In both scenarios,
whether constraint to the system was external (disparate sized-targets) or
internal (central paresis), it appears that the goal of the nervous system is for
the two hands to reach their respective targets simultaneously.
Adjustments in Movement Control during Bimanual Movements
A prolonged movement time for bimanual compared to unimanual tasks has
been referred to in the literature as a bilateral deficit and a recognized
phenomenon for healthy adults (Wyke, 1969; Wyke 1971; Corcos, 1984;
Marteniuk et al., 1984; Fowler et al, 1991; Dickstein et al, 1993). The bilateral
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
35
deficit phenomenon suggests that there is some deterioration in the planning
and performance of limb movement for bimanual tasks. In this study,
nonparetic limb movement time in the bimanual condition was at least 5
times greater than that recorded for the paretic limb or for either limb of the
controls and thus, much greater than previous research has suggested would
be solely indicative of a bilateral deficit. Although a portion of the nonparetic
limb movement time adaptation could be attributed to a bilateral deficit
phenomenon, these pronounced results suggest that additional mechanisms
may contribute, such as those observed in studies in which target accuracy is
manipulated described earlier (Kelso et al., 1979a; 1979b; 1983; Martenuik et
al., 1984; Fowler et al., 1991).
Although both the acceleration and deceleration phases contributed to the
increase in movement time in the bimanual condition, the deceleration phase
did so to a greater extent. For some time, and across a variety of aiming
tasks, the deceleration phase has been shown to be the locus for visually
and proprioceptively mediated on-line adjustments and error-correction
(Woodworth, 1899; Soechting and Lacquaniti, 1981; Jeannerod, 1984;
Zelaznik et al., 1986). Limb position adjustments occur as the limb
approaches the target and these adjustments are typically described by
discontinuities in the velocity profile. This, together with the movement
termination data, would indicate that in the bimanual condition, the trajectory
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
36
of the nonparetic limb was not fully planned in advance, but rather, online
adjustments, particularly during the deceleration phase, mediated a nearly
simultaneous target impact. We infer that the controller monitors afferent and
efferent information of the paretic limb, and nonparetic limb adjustments are
made as a result. A similar mechanism has been invoked to explain the
pronounced temporal adjustments that evolve during movements in healthy
adults when performing an asymmetric bimanual task (Perrig et al., 1999;
Kazennikov et al., 2002).
These results can be interpreted as further evidence for the principle of motor
equivalence - invariant goal achievement with variable means - a hallmark
of goal-directed movement (Lashley, 1930). Goal invariance has also been
demonstrated in such diverse movements as speech gestures (Abbs and
Gracco, 1983) and precision grasp (Cole and Abbs, 1987). To maintain goal
invariance in the bimanual aiming condition, the nonparetic limb slowed down
so that the two limbs would arrive at the target goal nearly simultaneously.
Ecological motor control theory proposes that movement is a product of the
interaction between the organism and the environment (Gibson, 1977). The
environment provides the context for the emergence of unique movement
patterns. The results of this study suggest that an implicit task goal played an
important role in limb performance. Nonparetic limb movement time in the
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
37
bimanual condition was not simply the behavioral expression of that limb’s
performance capability (as expressed in the unimanual condition), but rather
the product of the apparent goal of the task - to reach the target with both
limbs at the same time. This emerged without explicit instructions to do so.
Only the nonparetic limb demonstrated movement time adjustments in the
bimanual compared to the unimanual condition. Nonparetic limb movement
time adjustments in a bimanual condition have also been reported for a
reciprocal elbow flexion and extension task (Rice and Newell, 2001). The
nonparetic limb slowed to adopt the cycle duration of the slower paretic limb
during this continuous task.
In contrast to the movement time results reported in our study, both limbs
exhibited an adjustment of peak resultant velocity, but in opposite directions.
Not only did the nonparetic limb exhibit a lower peak resultant velocity in the
bimanual task condition, but in contrast, the paretic limb, on average,
exhibited a higher peak resultant velocity, compared to that in the unimanual
task condition. Previous investigators have suggested that there is a
bimanual “ facilitation” effect after stroke (Whitall et al., 2000); however, this is
the first study to quantify and provide supporting kinematic evidence for such
an effect. If differences between movement conditions are described by
movement duration alone, it would appear that nonparetic movement time is
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
38
dictated by the slower paretic limb. However, a closer examination of the
kinematics reveals a facilitation (i.e. higher peak resultant velocity) of the
paretic limb in the bimanual condition (observed in 19/30 stroke participants)
that is not apparent from the movement time data alone. A similar though
different benefit was reported in a small-scale study of 6 individuals post
stroke, in which a smoother paretic elbow extension velocity profile was
observed in a bilateral compared to a unilateral condition (Cunningham et al.,
2002).
Adjustments in Movement Planning during Bimanual Movements
Despite no explicit instructions, the majority of those with stroke initiated
movements nearly simultaneously in the bimanual condition (i.e. same
reaction time). These results are consistent with those reported for
individuals with cerebral palsy (Steenbergen et al., 1996). In contrast, for
unimanual aiming preparation, nonparetic limb reaction time was shorter than
that for the paretic limb. Two concomitant processes were used to achieve
near simultaneity of movement onset: nonparetic reaction time was
lengthened and paretic reaction time was shortened in the bimanual
compared to the unimanual condition. Although this latter difference was not
statistically significant (p= .16) it was trending in the opposite direction
(shorter as opposed to longer reaction time) to that of the nonparetic limb
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
and of both limbs of the controls. Additionally, a somewhat shorter reaction
time in the bimanual condition is contrary to what would be expected for a
bilateral deficit effect described earlier. In contrast, the lengthened reaction
time for the nonparetic limb (compared to the unimanual condition) supports
a hypothesis for temporal adaptation in anticipatory processes.
We have demonstrated that two limbs with disparate movement abilities are
constrained to function with features of a single unit in a bimanual
symmetrical task. It is primarily the nonparetic limb, however, that alters its
movement strategy when paired with the paretic limb for the bimanual task
condition. The adjustment of nonparetic limb movement time occurred
primarily during the deceleration phase of the movement indicating that limb
performance may be driven by the perceived goal of temporal synchrony.
Participation in this study required a relatively high level of motor control,
specifically the ability to extend the elbow with simultaneous flexion of the
shoulder, a movement combination difficult for many individuals post-stroke.
Therefore, the generalizability of these results to a more impaired population
of stroke survivors is not known. Specific brain lesion location was not
reported here. However, the control of bimanual movements is not allocated
to one neuroanatomical structure alone, but rather appears to be the result of
a distributed network involving both cortical and subcortical areas (Donchin
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
40
et al., 1998; Kermadi et al, 2000; Debaere et al., 2001; Swinnen, 2002).
Lesion location information, therefore, may not have added to the
interpretation of our results.
This study examined bimanual aiming in a relatively simplistic laboratory
task. The translation of these results to tasks encountered in everyday life is
not known. Examination of limb movements when the task demands are
different would further our understanding of bimanual motor control after
stroke. With the blocked design of this paradigm, participants were not
required to make a choice in limb movement (i.e. unimanual, bimanual); they
were not burdened by the attention demands required for response selection.
Adjustments in movement planning could be further probed with a random
condition experimental design.
The clinical application of these results, that a bimanual approach to upper
extremity rehabilitation may be beneficial, stands in contrast to constraint-
induced movement therapy, in which bimanual activities are not permitted -
the therapeutic intervention aimed toward the paretic arm alone. While we
understand the rationale underlying constraint-induced movement therapy,
we believe that by prohibiting an entire class of actions (those that are
naturally bimanual) constraint-induced movement therapy does not permit
stroke survivors to benefit from the full scope of therapeutic opportunities.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
41
These data presented here provide convincing support for the inclusion of
bimanual activities as part of rehabilitation intervention for those with mild
post-stroke upper limb hemiparesis.
Conclusions
This study provides evidence that peak velocity of the paretic limb can be
facilitated through interlimb effects of the nonparetic limb for a rapid bimanual
aiming task. This facilitation, however, may be damped through a drive for
motor equivalence. Future research might attempt to exploit the goal for
temporal synchrony and further facilitate paretic limb movement by entraining
the nonparetic limb to adopt a unimanual movement strategy during
bimanual tasks.
Clinical messages
• Individuals with mild paresis post-stroke retain a level of bimanual
coordination. This ability may be harnessed for rehabilitation
interventions to benefit recovery of function.
• The nonparetic limb may slow down in a bimanual task, however this
slowing may well be accompanied by an actual facilitation of the peak
speed in the paretic limb.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Postscript
42
The preceding was submitted as a manuscript for publication in Clinical
Rehabilitation and is currently in press. Since then, and as a result of
analysis of the subsequent two experiments (Chapter 3 and Chapter 4),
several additional considerations have arisen related to the data in Chapter 2
and its interpretation. In order to preserve the manuscript in the form it was
accepted for publication, these additional considerations are presented as a
postscript to Chapter 2. The two primary purposes for this postscript were 1)
to acknowledge the role of proprioception in bimanual aiming and 2 ) to
further examine the control of paretic limb velocity in the subset of 19
participants who demonstrated a greater paretic limb PRV in the bimanual
compared to the unimanual aiming condition.
Role of proprioception in bimanual coordination
It is important to note that proprioception was intact at the shoulder, elbow,
and wrist for all stroke participants as measured by the Fugl-Meyer
Measurement of Physical Performance (Fugl-Meyer et al., 1975). Jackson
and colleagues (1999) have suggested that a sensorimotor mechanism,
based upon proprioceptive coding of limb position and motion maintains
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
43
interlimb coordination during movement execution. Stroke participants in this
study had proprioception available to guide interlimb coordination.
Two participants exhibited impaired proprioception at the thumb
interphalangeal joint. Participants wore paddles on their hands, holding the
digits static as the proximal joints were engaged to aim toward the targets.
Targets were hit with the entire paddle; no individual digit movement
required. All joints that were actively involved in aiming to the target
possessed intact proprioception and therefore any observed coordination
deficits could not be attributed any proprioceptive impairment, but rather to
the diminished central motor drive.
Velocity
The stroke group average increase in paretic limb PRV for the bimanual
condition was of considerable interest secondary to the potential clinical
implications for rehabilitation. There are a number of factors that can
contribute to an increase in peak resultant velocity of a movement. An
extensive literature supports greater PRV with increased movement
amplitudes (Bootsma et al., 1994; Jakobson and Goodale, 1991; Marteniuk
et al., 1987; Servos et al., 1992). This cannot be used as an argument here,
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
44
however. Movement amplitude, from home position to target, was identical
for both aiming conditions.
Verbal instructions to participants, “move as soon and as fast as you can”
were identical and the order of conditions was counterbalanced across
participants such that a fatigue factor (if unimanual aiming always followed
bimanual aiming) likely did not contribute to the greater bimanual than
unimanual PRV. The pairing with the nonparetic limb, therefore, remains the
most likely explanation for the greater paretic limb PRV in the bimanual
compared to the unimanual condition.
Nineteen of 30 stroke participants demonstrated a greater paretic limb PRV
in the bimanual aiming condition (9% PRV increase for this subset compared
to 4% increase for entire stroke cohort). For these 19 participants, paretic
limb time to peak velocity was the same for both unimanual and bimanual
aiming conditions (Figure Postscript-1), despite the greater PRV in the
bimanual condition (Figure Postscript-2). With a greater PRV, one might
predict a longer rise time to attain that peak. This, however, was not the
case. The logical conclusion, then, is that the greater PRV in the bimanual
condition was attained by the CNS controlling the rate of rise of velocity
(acceleration) while maintaining rise time similar to that in the unimanual
condition. This is illustrated for one stroke participant in Figure Postscript-3
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
45
below; time to peak velocity was similar in both aiming conditions despite a
greater PRV in the bimanual condition. This was due to a greater rate of rise
in velocity (acceleration) in the bimanual compared to the unimanual
condition.
250 -|
UNIMANUAL BIMANUAL
Figure Postscript-1. Paretic limb T IP velocity for subgroup of 19 stroke
participants with greater paretic limb PRV in bimanual compared to
unimanual aiming. No difference in TTP between the two aiming conditions
(p > 0.05); error bars are SE.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
UNIMANUAL
1 ----
BIMANUAL
Figure Postscript-2. Peak resultant velocity for subgroup of 19 stroke
participants with greater paretic limb PRV in bimanual than unimanual aiming
(p < 0.05); error bars are SE.
Paretic Upper Extremity
250
Unimanual
200
Bimanual
o
0
t/i
E
a
150
-S'
o
o
ai
>
100
400 700 800 0 100 200 300 500 600
Time (msec)
Figure Postscript-3 Ensemble average time series of instantaneous velocity
plotted from movement onset to offset for stroke participant (FM = 55).
Paretic limb profile for unimanual (solid line) and bimanual (dashed line)
aiming. Greater PRV for bimanual compared to unimanual aiming, but time to
peak velocity in each aiming condition is similar.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
47
Greater paretic limb PRV in bimanual aiming without a concomitant increase
in time to attain that peak was accomplished in 63% of stroke participants;
there was not an associated increase in rise time as one might predict.
Rather, rise time rate appeared to be the controlled variable. Although this is
an interesting finding, one then must consider what purpose this regulation of
rate rather than time served.
In bimanual aiming, Schmidt and colleagues (1979) have proposed that a
similar temporal structure is prescribed by the CNS for both limbs. In this
subset of 19 participants, time to peak velocity was similar for both limbs in
the bimanual aiming condition (paretic limb: 225 +/- 47 msec, nonparetic limb
199 +/- 38 msec;p > 0.05). It appears the CNS adopted a pulse-height
control strategy to accommodate the increase in PRV of the paretic limb.
Pulse-height control policy, originally used to characterize CNS control of
force pulses with varying magnitudes (Freund and Budingen, 1978; Ghez,
1978; Ghez and Vicario, 1978; Gordon and Ghez, 1987), describes, in this
case, control of the rate of paretic limb rise time (increase in acceleration) as
overall rise time is constrained by the single central plan for the two limbs.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
48
CHAPTER 3
The Coordination of Bimanual Asymmetric Rapid Aiming Movements:
The Effect of Diminished Central Drive
Abstract
When moving two limbs different distances, there is a tendency of the limb
moving the shorter distance to prolong its movement time to become similar
to the limb moving the longer distance. This temporal constraint, existent for
an asymmetric task, has not been tested when the two limbs performing this
task are asymmetric in their motor ability. Twenty-nine individuals post-stroke
with intact sensation but mild unilateral paresis and twenty-eight age-
matched controls performed unimanual and asymmetric bimanual discrete,
rapid multi-joint aiming movements in response to a lighted target.
Psychophysical and kinematic measurements were recorded via hand
switches and an electromagnetic sensor attached to each hand. There were
two bimanual conditions: one designated as congruent in which (for the
stroke participants) the nonparetic limb aimed to a far target as the paretic
limb aimed to a near target and a second, incongruent condition in which the
paretic limb aimed far as the non paretic limb aimed near. For the control
group, in both bimanual conditions, the limb aiming to the near target slowed
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
49
its movement time compared to its unilateral performance to allow for a
nearly simultaneous target hit. The locus of this slowing was in the
deceleration phase of the movement. For the stroke group, movement time
prolongation of the near-aiming limb was most striking in the incongruent
condition when the nonparetic limb was paired with the paretic limb aiming
far. Differential performance of the paretic limb was also observed. Peak
resultant velocity was greater during bimanual compared to unimanual
aiming but only for the incongruent condition. This peak velocity facilitation
may depend not only on the bimanual nature of the task but also the aiming
amplitude requirements of the paretic limb. An examination of the
acceleration and deceleration phase of the movement profile provides
evidence that the CNS exerts a temporal constraint on the two limbs and that
this constraint is robust to the sequelae of decreased central motor drive
following stroke.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
50
Introduction
Previous systematic investigations reveal considerable differences in limb
performance between unimanual and bimanual aiming tasks with healthy
individuals (Kelso et al., 1979a, 1979b, 1983; Martenuik et al., 1984; Fowler
et al., 1991). Temporal results for unimanual and symmetrical bimanual
aiming concur with the speed-accuracy trade-off relationship initially
proposed by Woodworth (1899) and formalized into a mathematical law by
Fitts (Fitts, 1954; Fitts and Peterson, 1964) that is, movement time increases
as movement amplitude or precision (task difficulty) requirements increase.
However, for bimanual asymmetric goal-directed aiming, Fitts’ Law is not
upheld. Specifically, in these investigations, and of importance to our
research, movement time of the limb directed towards the less difficult (i.e.
near or large) target is prolonged (compared to its unimanual performance)
and very similar to the movement duration for the limb directed towards the
more difficult (i.e. far or small) target. In the typical Fitts’ aiming task, the
distance between targets and target size specify task difficulty, and therefore
difficulty is an inherent constraint of the task. A naturalistic twist on Fitts’
paradigm occurs for persons after stroke. With hemiparesis, the task
constraint is internal (body-centered) rather than external (target-centered);
the constraint imposed on the system is the disparate movement ability of the
two limbs (i.e. paretic; nonparetic).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
51
A limited number of investigations have examined the coordination and
control of bimanual spatially symmetric tasks in individuals with unilateral
hemiparesis (Dickstein et al., 1993; Cunningham et al., 2002). Both of these
studies were limited to uni-joint elbow flexion (Dickstein et al., 1993) or elbow
extension (Cunningham et al., 2002) movements. Dickstein and colleagues’
results focused on the increased movement duration of the nonparetic limb
when paired with the paretic limb compared to unimanual performance,
whereas Cunningham’s group highlighted adaptations of the paretic limb - a
smoother velocity profile was produced when paired with the nonparetic limb
compared to paretic limb unimanual performance.
We previously examined the effects of this internal, body-centered constraint
in subjects with mild post-stroke hemiparesis who performed a multi-joint
bimanual symmetric aiming task for which the spatial requirements were
identical as each limb aimed to a single target (Rose and Winstein, in press).
We observed a temporal assimilation with adjustments primarily in the
nonparetic limb. Nonparetic limb movement time was prolonged when paired
with the paretic limb and compared to the unimanual condition resulting in
similar movement times. This finding is analogous to those of non-paretic
healthy controls in which movement time for the limb directed towards the
easy (i.e. near) target is prolonged in the asymmetric bimanual condition,
and, therefore, comparable to that of the limb directed towards the difficult
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
52
(i.e. far) target (Kelso et al., 1979a, 1979b, 1983; Martenuik et al., 1984;
Corcos, 1984; Fowler et al., 1991). In summary, previous bimanual
coordination investigations of healthy controls have suggested that the CNS
exerts a strong temporal constraint on the limbs, specifying a common
movement time, despite asymmetric task demands (Kelso, et al., 1979a,
1979b; Sherwood, 1989; Sherwood, 1991). Previous bimanual coordination
investigations of those with unilateral paresis performing a symmetric task
have also demonstrated the CNS drive to temporally constrain two limbs of
asymmetric ability (Dickstein, et al., 1993; Rice and Newell, 2001, Rose and
Winstein, in press). We designed this experiment to probe the robustness of
this temporal constraint by combining an internal, body-centered asymmetry
(unilateral paresis) and an external, target-centered asymmetry (targets
located at a near or far distance) in the same experiment. Based on previous
results from both healthy controls and those with stroke, we hypothesize that
the temporal constraint will be robust to the presence of both an internal and
external manipulation. The CNS will adapt both movement preparation and
movement execution of the nonparetic limb to maintain bimanual
coordination.
Unilateral paresis and asymmetric target location afford four internal-external
constraint combinations (in order of increasing difficulty): 1. nonparetic
limb/near target, 2. nonparetic limb/far target, 3. paretic limb/near target and
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
53
4. paretic limb/far target. These constraint combinations, in turn, afford two
bimanual asymmetric aiming combinations: a) nonparetic limb far
target/paretic limb near target and b) paretic limb far target/nonparetic limb
near target.
We suggest that these two aiming combinations are not equivalent. The first
combination pairs the paretic limb performing an easy task with the
nonparetic limb performing a difficult task. We define this aiming condition as
congruent in that the two task-limb combinations are relatively similar in
difficulty. This is in contrast to the second combination, which pairs the
paretic limb performing a difficult task with the nonparetic limb performing an
easy task. We define this aiming condition as incongruent in that the two
task-limb combinations are more disparate in level of difficulty (Figure 3.1).
1. nonparetic /near ......
2 . nonparetic /far
' > incongruent
3. paretic /near
..............► congruent
4. paretic /far
Figure 3.1. Congruent and incongruent aiming conditions
Investigations of bimanual aiming with target-centered external constraints
show that temporal synchrony is achieved through adjustments of the limb
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
54
performing the less difficult task (Kelso, et al., 1979a, 1979b). In this
experimental design, the nonparetic limb performs the easier task in the
incongruent condition. We hypothesize, therefore, that the CNS will specify
larger adjustments in movement planning and execution for the nonparetic
limb (compared with its unimanual performance) in the incongruent
compared to the congruent experimental condition.
In our previous study with a bimanual symmetric task, the paretic limb
exhibited a greater peak resultant velocity when paired with the nonparetic
limb and compared to the unimanual condition in a subgroup (63%) of
participants. Both limbs, in the terminology of the present experiment, aimed
a far distance. The present experimental design, with two bimanual aiming
combinations, affords further investigation of the potential mechanism
underlying this small, but significant increase in paretic limb peak resultant
velocity (PRV) during bimanual aiming. If greater paretic limb PRV were due
solely to pairing with the nonparetic limb, we would predict a greater paretic
limb PRV in both congruent and incongruent aiming conditions (compared to
its unimanual performance). Conversely, this greater PRV may also be
dependent on the task demands imposed upon the limb itself. In our previous
study, the paretic limb’s movement amplitude was comparable to the far
aiming distance in this experiment. If the far movement amplitude is also a
critical factor that promotes a greater PRV, in addition to bimanual
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
coordination, we would predict a greater bimanual, compared to unimanual
PRV in the incongruent (paretic far/nonparetic near) but not the congruent
(paretic near/nonparetic far) condition.
Method
Participants
Twenty-nine adults post-stroke (age 27 - 84 yr; M = 64; 17 left-brain lesion,
12 right-brain lesion; 6-120 mos post-stroke; M= 29 mos) and twenty-eight
adults without stroke (age 49 - 86 yr; M = 68) participated (Table 3.2). These
were the same participants as in Chapter 2 with the omission of data from
one stroke and two control participants secondary to incomplete data sets.
All were right-hand dominant (Oldfield, 1971). Stroke participants were
screened to verify: (1) intact visual fields (Friedman, 1992); (2) intact
shoulder, elbow and wrist proprioception (Fugl-Meyer et al., 1975); (3) pain-
free, active range of motion to at least 120° shoulder flexion; at least 90°
forearm supination and full motion of the elbow, wrist and fingers into
extension; and (4) Modified Ashworth Spasticity Scale scores (Bohannon
and Smith, 1987) for shoulder extensors, elbow flexors, forearm supinators,
wrist flexors < 2. Motor impairment was assessed with the Fugl-Meyer
Measurement of Physical Performance (FM) (Fugl-Meyer et al., 1975) and
grip and pinch strength (Mathiowetz et al., 1985).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
56
Task
Initial position for the seated participants was with shoulders in neutral
alignment, 90° elbow flexion, and 90° degree forearm pronation. In response
to an LED signal, participants were to reach and aim forward with one hand
(unimanual) or both hands (bimanual) to hit a switch mounted on an LED
target, as soon and as fast as possible. The movement required shoulder
flexion with elbow extension and forearm supination. For the bimanual
condition, the limbs moved to separate targets, located at different distances
in the y plane relative to the participant. The trunk was stabilized with a chest
strap so that task performance was accomplished primarily with upper limb
movements.
Apparatus
The apparatus consisted of two sets of 2.5 cm2 switches; home-position (one
for each hand) and target-position (one for each hand). Each switch
consisted of a solid-state infrared photo light emitter and diode. A 0.5 x 1 cm
blackened mylar sheath was mechanically mounted to each of the four
switches. When the light beam from the photo emitter to diode was broken by
the mylar, the switches were considered on (home switches) or off (target
switches). The home-position switches were secured to the table surface,
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
57
aligned for each participant in midline of the long axis of the humerus. For the
bimanual condition, there were separate targets for each hand, an LED
embedded in the center of a 3 cm width wooden block with the target-
position switches located on the outer side of each block. Target location was
determined individually in 3-D space: x-plane in line with long axis of the
humerus such that at target hit the humerus was at 0° of horizontal adduction
and abduction, y-plane such that target hit required no more than 15° of full
elbow extension (far condition) or 50% of the far position (near condition), z-
plane such that target hit required 90° forward shoulder flexion (Figures 3.2
and 3.3). Aiming hand end-point displacement was recorded from an
electromagnetic sensor attached to a rubberized paddle (10.5 cm diameter),
affixed to each hand. Movements were recorded with a “ 6-D Research”
electromagnetic motion capture and analysis system (Skill Technologies,
Inc.) using a 60 Hz sampling rate. Trial onset, LED onset, and movement
onset and offset of each hand were sampled on separate channels from
individual TTL pulses interfaced with a 400 MHz personal computer.
Procedure
Participants signed an informed-consent form approved by the Institutional
Review Board from the Health Sciences Campus of the University of
Southern California. Verbal instructions and demonstration of the task were
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
58
provided at the beginning of the session. Participants sat at a height-adjusted
table with elbows at 90° flexion, shoulders and wrists at 0° of flexion and
extension and abduction and adduction, and forearms in pronation at the
start position. Each participant performed one block of 5 practice trials before
each task condition. They were instructed to move “as soon as possible and
as fast as possible” once the LED was illuminated, to hit the target from
either side. In the bimanual condition, no explicit instructions were provided
as to how the two limbs were to be coordinated (e.g., to initiate or terminate
the movements together).
Each trial began with a verbal ready signal followed by the LED presented
randomly within 500 to 1500 msec to avoid anticipation. The trial ended when
the target switch(es) was hit. Participants returned their hand(s) to the home
position and awaited the next “ready” signal. There were two separate
bimanual vs. unimanual experiments, each with one bimanual and two
unimanual aiming conditions. Experiment I: 1. Bimanual Congruent (paretic
near/nonparetic far), 2. Unimanual: paretic near, 3. Unimanual: nonparetic far
(Table 3.1, Figure 3.2), and Experiment II: 1. Bimanual Incongruent (paretic
far/nonparetic near), 2. Unimanual: paretic far, 3. Unimanual: nonparetic near
(Table 3.1, Figure 3.3). For Experiment I, the paretic limb aimed to a target
located one-half the distance from the home-position relative to the
nonparetic limb’s target. For example, if the non paretic limb’s target was
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
59
located 37 cm from its home position, the paretic limb’s target was placed
18.5 cm from its home position. For Experiment II, the converse was true.
The control group performed each bimanual and unimanual condition with
their left and right arms. For the control group, we acknowledge that the
designation of “congruent” and “incongruent” does not readily apply (although
an examination of hand dominance influences awaits future study), however,
for consistency, we will use this terminology in discussing the results for both
experimental groups. Each condition was presented in separate, 28-trial
blocks (25 valid trials and 3 catch trials - a catch trial was one in which no
LED followed the “ready” signal). Intermittently, during each trial block,
reminders were provided to “move as soon and as fast as possible” in
response to the LED signal. The order of the two experiments was
counterbalanced across participants with a 10-minute rest period provided
between each one.
Table 3.1. Experimental Conditions____________________________________
Experiment Bimanual Unimanual
I Congruent: Paretic near /nonparetic far 1. paretic near
2 . nonparetic far
II Incongruent: Paretic far/nonparetic near 1. paretic far
2 . nonparetic
near
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
60
LED
LED
NP/R
P/L
Figure 3.2A. Experiment I: Congruent condition set-up showing participant
position and targets for bimanual aiming (overhead view). Paretic near (left for
controls) / Nonparetic far (right for controls). Planes of motion: x = medial/lateral;
y = fore/aft; z = vertical. P = paretic, NP = nonparetic, L = left, R = right.
LED
LED
NP/R
P/L
LED
LED
NP/R
P/L
Figure 3.2B. Experiment I: Unimanual Figure 3.2C. Experiment I:
Paretic near (left for controls). Unimanual Nonparetic far (right
for controls).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
61
LED
LED
NP/R P/L
Figure 3.3A. Experiment II: Incongruent condition set-up showing
participant position and targets for bimanual aiming (overhead view).
Paretic far (left for controls) /Nonparetic near (right for controls). Planes
of motion: x = medial/lateral; y = fore/aft; z = vertical. P = paretic, NP =
nonparetic, L = left, R = right.
LED
LED
NP/R P/L
LED
LED
NP/R
P/L
Figure 3.3B. Experiment II: Unimanual Figure 3.3C. Experiment II:
Paretic far (left for controls). Unimanual Nonparetic near (right
for controls).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
62
Data reduction
For each trial, raw displacement data were filtered with a 0-lag Butterworth
filter (low-pass cutoff frequency of 20 Hz). Instantaneous velocity was
calculated by differentiating electromagnetic sensor displacement data using
a three-point moving average. Resultant velocity was calculated from the x, y
and z velocity values. Trial onset, LED onset, switch onset and offset, and
sensor resultant velocity were incorporated into one data file for each aiming
condition. Data files were analyzed with Datapac 2000 Laboratory
Applications System (RUN Technologies Co.). Movement onset was the time
value at which the sensor acceleration value exceeded 0.03 cm/sec2.
Movement offset was the time value when the target-position switch signal
exceeded 0 mV. Movement time was defined as movement offset -
movement onset, reaction time as the time from LED onset to movement
onset. For a trial to be considered for analysis both hands had to initiate
squarely from the onset switches and hit their respective target switches.
Additionally, trials with reaction times that were greater or less than the mean
+1-2 SD for that trial block were discarded. If more than 20 acceptable trials
remained following application of these criteria, the middle 20 trials were
used for analysis.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
63
Statistical analysis
A participant mean was calculated per condition and limb for each dependent
measure: reaction time, movement time, time to peak velocity, time after
peak velocity, and peak resultant velocity. To test for effects of group (stroke,
control), aiming type (unimanual, bimanual), and limb (paretic, nonparetic;
left, right for controls) on these dependent measures, separate 2 (Group) x 2
(Aiming Type) x 2 (Limb) general linear model ANOVAs with repeated
measures on the last two factors were performed. This analysis was
conducted independently for Experiments I and II.
Comparison of movement initiation and termination across groups and
aiming conditions allowed a further examination of temporal adaptation. A
participant mean was calculated for each experiment and event. To test for
effects of group (stroke, control), experiment (congruent, incongruent for
stroke; left near/right far, left far/right near for controls), and event (initiation,
termination) a 2 (Group) x 2 (Experiment) x 2 (Event) general linear model
ANOVAs with repeated measures on the last two factors was performed.
For each participant, and for each dependent measure the raw difference in
nonparetic limb performance between the two aiming conditions was
converted to a percent difference score [(bimanual - unimanual)/unimanual].
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
64
This percent difference score was subjected to a one-way ANOVA to test for
differences in nonparetic limb performance between the two experiments.
Post-hoc pairwise comparisons with a Bonferroni correction determined the
locus of any significant interaction. Significance was set at p< 0.05 for all
comparisons. Statistical procedures were conducted using SPSS statistical
software (v. 11.5).
Results
Clinical Impairment
Participants were considered mildly impaired with the capability to isolate
elbow extension from shoulder flexion to complete the task (Duncan et al.,
1994). Group mean Fugl-Meyer motor score was 60+/- 5 out of 66 total
possible score. Grip and pinch strength of the paretic limb provided a more
sensitive measure of impairment, especially in this mildly impaired group.
Although participants were able to perform the experimental task, upper limb
motor deficits remained (Table 3.2). Stroke group means, expressed as a
percentage of age, gender and limb-matched norms were all less than
normal: grip: 65+/-24%; lateral pinch: 71+/-23%; palmer pinch: 63+/-18%
(Mathiowetz, et al., 1985). Although participants were able to perform the
experimental task, upper extremity motor deficits were evident from the grasp
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
65
and pinch data. Proprioception was intact at the shoulder, elbow, and wrist
for all stroke participants as measured by the Fugl-Meyer Measurement of
Physical Performance (Fugl-Meyer et al., 1975; Table 3.2). One participant
exhibited impaired proprioception at the thumb interphalangeal joint.
Participants wore paddles on their hands, holding the digits static as the
proximal joints were engaged to aim toward the targets. Targets were hit with
the entire paddle; no individual digit movement required. Therefore, impaired
thumb proprioception should not have influenced task performance for this
participant.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
66
Table 3.2. Summary of stroke participants arranged hierarchically according
to Upper Extremity Fugl-Meyer motor score (M) (total possible=66)._______
No. Age Lesion
Side
M P Grip
(%)
LP
(%)
PP
(%)
Lesion
type
Lesion Location
1 62 L 43 8 59 29 39 I Corona radiata
2 67 R 50 8 55 44 24 I CT negative
3 43 L 54 8 39 34 24 I Fronto-parietal ctx
subcortical areas
4 66 R 55 8 38 45 42 I Parietal cortex
5 68 L 56 8 59 100 54 I Thalamus
6 53 R 57 8 94 94 52 I CT negative
7 51 L 58 8 27 56 55 I Fronto-parieto-
temporal ctx
8 49 L 58 7 27 48 37 H Fronto-parietal ctx
9 80 L 59 8 60 63 71 I Pons
10 69 R 59 8 66 27 55 I Posterior limb IC
11 72 L 60 8 28 # # I CT negative
12 64 R 60 8 48 78 52 H Pons
13 71 R 60 8 79 86 68 I CT negative
14 69 R 60 8 72 90 80 I Pons
15 66 L 62 8 62 61 77 H Lateral thalamus,
insular cortex
16 63 L 62 8 64 81 69 I Pons
17 73 R 62 8 65 86 49 I CT neg.
18 60 L 63 8 78 100 74 I Internal capsule
19 61 L 63 8 75 70 77 I CT negative
20 71 L 63 8 100 96 74 I Pons
21 55 L 64 8 84 96 80 I Parietal cortex,
corona radiata
22 62 R 65 8 100 94 78 I Pons
23 80 L 65 8 90 79 64 I Posterior IC
24 54 L 66 8 33 50 53 I Thalamus
25 73 R 66 8 90 74 73 I Basal Ganglia
26 73 L 66 8 27 34 60 I Corona Radiata
27 84 R 66 8 100 92 74 I Thalamus
28 69 L 66 8 100 95 100 I Tempo-parietalctx,
subcortical areas
29 27 R 66 8 76 80 84 I CT negative
Grip, Lateral Pinch (LP) and Palmar Pinch (PP) values represent percentage
strength of paretic (P) limb compared to limb, age and gender matched
normative data (Mathiowetz et al., 1985), P = UE proprioception score (Fugl-
Meyer et al., 1975), H = hemorrhage, I = infarct, ctx = cortex, IC = internal
capsule, # data not available
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
67
Trials were excluded from analysis if RT was outside of pre-established
boundaries, movement was not initiated squarely from the start switches, or
the target was missed. For the control group, 0.6 % (13/2240) of unimanual
trials and 9% (99/1120) of bimanual trials were excluded. For the stroke
group, 9% (212/2320) of unimanual trials and 24% (280/1160) of bimanual
trials were excluded based on these criteria. The majority of these 280
excluded trials were due to target miss by the paretic limb aiming far in the
incongruent condition. Results for each bimanual condition will be presented
separately.
Experiment I: Congruent Condition
Reaction Time
Both limbs of the stroke group exhibited an 8% increase in RT (20 msec) in
the bimanual compared to the unimanual aiming condition. This increase was
in contrast to a small but not significant difference in either limb between the
two aiming conditions for the control group (Group x Aiming condition
interaction p < 0.05; Group x Aiming condition x Limb, ns; Figure 3.4, Table
3.3).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
6 8
control left near
control right far
stroke paretic near
stroke nonparetlc far
Figure 3.4. Line graphs of group means by aiming condition for reaction time.
Stroke (square) and control (circle) group by aiming condition for each limb
(paretic/left, dashed line; nonparetic/right, solid line); error bars are SE. RT
prolonged for bimanual aiming in stroke group alone (Group x Aiming
condition interaction p < 0.05; Group x Aiming condition x Limb, ns).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
300 -i
< D
O'
150
UNIMANUAL BIMANUAL
69
Table 3.3. Experiment I: Congruent Condition (Paretic near/Nonparetic far;
Left near/Right far for controls) Group means and standard deviations for the
primary outcome measures of Movement Time (MT), Time to Peak velocity
(TTP), Time after Peak velocity (TAP), Peak Resultant Velocity (PRV),
Reaction Time (RT),______________
Control
Left Near Right Far
Unimanual Bimanual Unimanual Bimanual
Mean SD Mean SD Mean SD Mean SD
MT 300 40 343 51 349 49 363 49
(ms)
TTP 163 21 169 25 181 28 185 32
(ms)
TAP 137 30 175 48 168 34 178 40
(ms)
PRV 183 39 174 36 195 33 190 29
(cm/s)
RT 202 26 212 26 199 27 210 24
(ms)
Stroke
Paretic Near Nonparetic Far
Unimanual Bimanual Unimanual Bimanual
Mean SD Mean SD Mean SD Mean SD
MT 415 106 441 99 401 59 460 101
(ms)
TTP 219 67 239 64 206 37 232 68
(ms)
TAP 196 90 202 93 195 52 228 84
(ms)
PRV 157 32 152 30 180 28 164 28
(cm/s)
RT 262 55 284 79 240 39 258 52
.M s )..
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
70
Movement Time
Movement time analysis produced a Group x Aiming condition x Limb
interaction (p < 0.05; Figure 3.5, Table 3.3). Post-hoc analysis revealed a
significant increase in MT for the nonparetic limb in the bimanual compared
to the unimanual condition. Nonparetic limb MT was prolonged by 15% when
paired with the paretic limb compared to the unimanual condition (401 +/-10
msec vs. 460 +/-15 msec). A smaller (6%), but also significant increase
occurred for the paretic limb in the bimanual condition. The interlimb MT
difference was just 19 +/- 5 msec. An effect size of .19 suggests this
difference is small. The left limb of the control group, aiming to the near
target, prolonged its MT by 14% when paired with the right limb aiming to the
far target, creating a MT difference between the two limbs of just 20 +/-5
msec (Figure 3.5). It is interesting that the average inter-limb difference for
the stroke group performing the asymmetric task with two disparate effectors
was no different than that for the control group.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
71
575 -
cT
< D
| 500 -
< s >
E
425 -
-m
c
a >
E
o >
g 350 -
275
Figure 3.5. Line graphs of group means by aiming condition for movement
time. Stroke (square) and control (circle) group by aiming condition for each
limb (paretic/left, dashed line; nonparetic/right, solid line); error bars are SE.
Nonparetic limb MT was prolonged when paired with the paretic limb
compared to the unimanual condition to allow for just a 19 msec difference
between the limbs (Group x Aiming condition x Limb interaction, p < 0.05).
To more closely examine aiming performance, velocity profiles of the hand
were parceled into two sub-phases: 1) acceleration phase, time to peak
velocity (TTP) and 2) deceleration phase, time after peak velocity (TAP).
Both limbs of the stroke group exhibited a similar prolonged TTP in the
bimanual compared to the unimanual aiming condition (paretic limb: 9%
increase; nonparetic limb: 13% increase) and attained peak velocity nearly
simultaneously. In contrast there were no differences for either limb of the
X
T
X
— 0- - control left near
— # — control right far
- Q - stroke paretic near
— ■ — stroke nonparetic far
UNIMANUAL BIMANUAL
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
72
control group between the two conditions; Group x Aiming condition
interaction p < 0.05; Group x Aiming condition x Limb, ns; Figure 3.6, Table
3.3).
300
O
C D
w
E
&
8
< D
>
£
i-
250
200
150
— & - control left near
— # — control right far
- B- ■ stroke paretic near
— ■ — stroke nonparetlc far
UNIMANUAL
I
BIMANUAL
Figure 3.6. Line graphs of group means by aiming condition for TTP. Stroke
(square) and control (circle) group by aiming condition for each limb
(paretic/left, dashed line; nonparetic/right, solid line); error bars are SE. Both
limbs of the stroke group exhibited a similar prolonged TTP in the bimanual
compared to the unimanual aiming condition to attain peak velocity nearly
simultaneously (Group x Aiming condition interaction p < 0.05; Group x
Aiming condition x Limb, ns).
For time after peak velocity, the nonparetic limb (aiming to the far target)
increased by 33 +/- 9 msec in the bimanual condition whereas there was no
difference in the paretic limb between the two conditions. For the control
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
73
group, the left limb (aiming to the near target) increased by 37 msec in the
bimanual condition providing a nearly identical time after peak velocity for
both limbs. This increase in TAP of the limb aiming near (left) for the controls
but for the limb aiming far (nonparetic) for the stroke group resulted in a
Group x Aiming condition x Limb interaction (p < 0.05; Figure 3.7, Table 3.3).
350
o
0 )
to
E
£
o
o
< D
>
«
0 )
C L
L_
0 )
t o
< D
E
i-
225
x
- 0 - ■
control left near
control right far
- a -
stroke paretic near
-m - stroke nonparetic far
100
UNIMANUAL BIMANUAL
Figure 3.7. Line graphs of group means by aiming condition for TAP. Stroke
(square) and control (circle) group by aiming condition for each limb
(paretic/left, dashed line; nonparetic/right, solid line); error bars are SE.
Increased TAP for control left limb aiming near and stroke nonparetic limb
aiming far (Group x Aiming condition x Limb interaction, p < 0.05.)
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
74
Velocity
Analysis of peak resultant velocity also revealed a Group x Aiming condition
x Limb interaction. Post-hoc analysis revealed that the greatest difference
between the two aiming conditions was observed in the nonparetic limb of
those with stroke (Figure 3.8, Table 3.3). The nonparetic limb demonstrated
a 9% decrease (180 +/- 6 cm/sec vs. 164 +/- 5 cm/sec) in PRV when paired
with the paretic limb, compared to its unimanual performance. A smaller
(5%), but also significant decrease in the bimanual aiming condition occurred
in the left limb of the control group when it was paired with the right limb
aiming a far distance. There was no difference in paretic limb PRV between
the two aiming conditions.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
75
225
&
-W
E
o
g
8
C D
>
c
£
3
O T
0 >
D U
jx t
03
C D
CL
175
— 0 - - control left near
— • — control right far
- B - stroke paretic near
— ■ — stroke nonparetlc far
125
UNIMANUAL
I
BIMANUAL
Figure 3.8. Line graphs of group means by aiming condition for PRV. Stroke
(square) and control (circle) group by aiming condition for each limb
(paretic/left, dashed line; nonparetic/right, solid line); error bars are SE.
Nonparetic limb PRV decreased when paired with the paretic limb compared
to its unimanual performance (Group x Aiming condition x Limb interaction, p
< 0.05).
Experiment I: Congruent Condition Summary
For each dependent measure, nonparetic limb behavior was significantly
different in the bimanual compared to the unimanual aiming condition.
Specifically, reaction time and movement time (including both time to- and
time after peak velocity) were prolonged and peak resultant velocity was
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
76
lower in bimanual compared to the unimanual aiming. The paretic limb
demonstrated prolonged reaction time and movement time (specifically in
time to peak velocity) in the bimanual compared to unimanual condition. This
prolongation in movement time provided for only a 19 +/-5 msec difference
between the two limbs in the bimanual condition, an inter-limb difference not
unlike that observed for the control group. Both limbs attained peak velocity
nearly simultaneously, suggesting that the two limbs are under control of the
same motor program.
Modifications of the limb aiming to the near target in the bimanual compared
to the unimanual condition were also present for the control group.
Movement time, specifically time to peak velocity, was prolonged, and PRV
was lower for the left limb aiming to the near target when paired with the right
limb aiming to the far target. Although these differences were significant
compared to performance in the unimanual condition, they were not as great
as those for the nonparetic limb of the stroke group.
Experiment II: Incongruent Condition
Reaction Time
The largest difference in reaction time between the unimanual and bimanual
aiming condition was for the nonparetic limb of those with stroke. Post-hoc
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
analysis revealed that RT of the limb increased by 14% (238 +/- 7 msec vs.
272 +/- 8 msec) in the bimanual condition when paired with the paretic limb
aiming to the far target (Group x Aiming Type x Limb, p < 0.05; Figure 3.9,
Table 3.4). This 34 +/- 5 msec increase resulted in a nearly identical RT for
each limb; there was only a 7 +/- 5 msec mean RT difference between the
limbs (265 +/- 8 msec vs. 272 +-/ 8 msec). A small (5%) yet significant
increase also was observed in the right limb of the controls when paired with
the left limb aiming the long distance (p < 0.05).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
78
Table 3.4. Experiment II: Incongruent Condition (Paretic far/Nonparetic near;
Left far/Right near for controls) Group means and standard deviations for the
primary outcome measures of Reaction Time (RT), Movement Time (MT),
Time to Peak velocity (TTP), Time after Peak velocity (TAP), Peak Resultant
Velocity (PRV)._____________________________________________________
Control
Left Far Right Near
Unimanual Bimanual Unimanual Bimanual
Mean SD Mean SD Mean SD Mean SD
MT 356 46 357 46 298 36 340 52
(ms)
TTP 183 31 195 46 161 23 170 44
(ms)
TAP 173 36 162 36 137 32 170 50
(ms)
PRV 200 37 201 39 178 28 164 29
(cm/s)
RT 203 28 202 31 202 25 212 24
(ms)
Stroke
Paretic Far Nonparetic
Near
Unimanual Bimanual Unimanual Bimanual
Mean SD Mean SD Mean SD Mean SD
MT 516 130 519 131 338 45 468 95
(ms)
TTP 250 75 247 53 185 22 216 57
(ms)
TAP 267 120 273 113 152 44 252 98
(ms)
PRV 158 31 164 31 165 24 138 21
(cm/s)
RT 259 53 265 51 238 43 272 54
(ms)
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
79
300
o
< D
(0
E
< D
E
F
c
o
4 -J
U
8
0£
250 -
200
T
- 0- ■ control left far
— # — control right near
- Q ■ stroke paretic far
1 - stroke nonparetlc near
150
1 ------
UNIMANUAL
1 -----
BIMANUAL
Figure 3.9. Line graphs of group means by aiming condition for reaction time.
Stroke (square) and control (circle) group by aiming condition for each limb
(paretic/left, dashed line; nonparetic/right, solid line); error bars are SE.
Nonparetic limb increased in the bimanual condition when paired with the
paretic limb aiming to the far target (Group x Aiming condition x Limb
interaction, p < 0.05.)
Movement Time
The greatest movement time difference between the two aiming conditions
was again observed in the nonparetic limb of those with stroke (Group x
Aiming Type x Limb interaction, p < 0.05; Figure 3.10, Table 3.4). Post-hoc
analysis revealed that in the bimanual condition there was a 38% increase in
MT from the unimanual condition (338 +- 8 msec vs. 468 +/-14 msec). This
prolongation of MT resulted in a 51 +/-11 msec difference, on average,
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
80
between the paretic limb aiming to the far target and the nonparetic limb
aiming half that distance to the near target. Although this inter-limb difference
is larger than that observed in the congruent aiming condition, the 51 msec
difference between the two limbs when they moved together is considerably
less than 179 +/-17 msec average between-limb difference observed for
unimanual aiming (Figure 3.10). The right limb of the controls had a smaller
(14%) yet significant prolongation in the bimanual condition when paired with
the left limb aiming to the far target (p < 0.05), allowing for just an 18 + /-11
msec inter-limb difference in the bimanual condition (Figure 3.10).
©
E
c
©
E
©
§
575
500
425 -
350
275
T
e-
i
- © •
-B
control left far
control right near
stroke paretic far
stroke nonparetic near
UNIMANUAL BIMANUAL
Figure 3.10. Line graphs of group means by aiming condition for movement
time. Stroke (square) and control (circle) group by aiming condition for each
limb (paretic/left, dashed line; nonparetic/right, solid line); error bars are SE.
Nonparetic limb increased in the bimanual condition when paired with the
paretic limb aiming to the far target (Group x Aiming condition x Limb
interaction, p < 0.05.)
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
81
Movement time was again examined more closely by separately analyzing
the acceleration and deceleration phases of movement. The only significant
difference in time to peak velocity between the two aiming conditions was for
the nonparetic limb of those with stroke (Figure 3.11, Table 3.4). Post-hoc
analysis revealed that time to peak velocity increased by 16% in the
bimanual compared to the unimanual condition (185 +/- 4 msec vs. 216 +/- 9
msec, Group x Condition x Limb interaction, p = 0.05). With this increase, the
limbs attained peak velocity within 31 msec of one another. It is of interest to
note that for the control group, the difference in time to peak velocity for the
two limbs was 25 msec, similar to that observed for the stroke participants,
suggesting that the asymmetric nature of the task and not simply the
asymmetric ability of the limbs contributes to this non-simultaneity in time to
peak velocity for both limbs.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
82
300
o
0 )
(fl
E
'w '
£
o
o
©
>
to
©
C L
250 -
200 -
©
£
h - 150
I
I
T
a
i
- 0 - • control left far
control right near
- B - stroke paretic far
—■— stroke nonparetic near
UNIMANUAL BIMANUAL
Figure 3.11. Line graphs of group means by aiming condition for TTP. Stroke
(square) and control (circle) group by aiming condition for each limb
(paretic/left, dashed line; nonparetic/right, solid line); error bars are SE.
Nonparetic limb TTP increased in the bimanual compared to the unimanual
condition (Group x Aiming condition x Limb interaction, p < 0.05).
For time after peak velocity, again, the greatest difference between the two
aiming conditions was observed in the nonparetic limb of those with stroke
(Group x Condition x Limb, p = 0.05; Figure 3.12, Table 3.4). Post-hoc
comparisons revealed that nonparetic limb TAP was prolonged by 66% (152
+/- 7 msec vs. 252 +/-14 msec) in the bimanual condition when paired with
the paretic limb aiming to the far target. This 100 +/-12 msec increase
permitted the two limbs’ TAP’s to be within 21 msec of each other (252 +/- 14
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
83
msec vs. 273 +-/16 msec). The right limb of the controls had a smaller
(24%), but significant prolongation in the bimanual condition when paired
with the left limb aiming to the far target (p < 0.05).
350
o
< D
W
E
■w
£
o
o
< D
>
w
Q )
C L
< D
O J
< D
E
i-
225
-©■ ■ control left far
♦ control right near
- 0 • stroke paretic far
- i - stroke nonparetlc near
100
U N I M A N U A L BIMANUAL
Figure 3.12. Line graphs of group means by aiming condition for TAP. Stroke
(square) and control (circle) group by aiming condition for each limb
(paretic/left, dashed line; nonparetic/right, solid line); error bars are SE.
Nonparetic limb TAP increased in the bimanual condition when paired with
the paretic limb aiming to the far target (Group x Aiming condition x Limb
interaction, p < 0.05).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
84
Velocity
The nonparetic limb of those with stroke again demonstrated the greatest
difference between the two aiming conditions (Group x Reach Type x Limb, p
< 0.05; Figure 3.13, Table 3.4). Post-hoc analysis revealed a significant
decrease (17%) in PRV for the nonparetic limb in the bimanual compared to
the unimanual aiming condition (165 +/- 4 cm/sec vs. 138 +/- 4 cm/sec). The
right limb of the control participants demonstrated a smaller yet significant
decrease (8%) in PRV when paired with the left limb, compared to its
unimanual performance.
In contrast to the nonparetic limb’s decrease in PRV in the bimanual
condition, the paretic limb demonstrated a 4% increase (158 +/- 6 cm/sec vs.
164 +/- 7 cm/sec) in PRV when aiming simultaneously with the nonparetic
limb. This represents a dissociation in peak resultant velocity between aiming
conditions for each limb, similar to our observation in a bimanual symmetric
experiment (Rose and Winstein, in press). Although the increase in paretic
limb PRV during bimanual aiming was small, it was statistically significant,
and stands in contrast to both the nonparetic PRV decrease and the lack of
difference in PRV for either limb of the control group.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
85
225 -
o
©
m
~ .. .
£
u
’ w'
S'
o
o
c
2
3
m
©
E
ro
(L 125
U N I M A N U A L BIMANUAL
Figure 3.13. Line graphs of group means by aiming condition for PRV. Stroke
(square) and control (circle) group by aiming condition for each limb
(paretic/left, dashed line; nonparetic/right, solid line); error bars are SE.
There was a dissociation in peak resultant velocity, with nonparetic PRV
decreased, but paretic limb PRV increased in bimanual compared to
unimanual aiming (Group x Aiming condition x Limb interaction, p < 0.05).
As this group average increase in PRV in the bimanual condition was small,
but of considerable interest, we examined this measure with individual
participant data. Seventeen of 29 stroke participants demonstrated a greater
PRV in bimanual compared to unimanual aiming. With a greater PRV, one
might predict a longer rise time to attain that peak. This, however, was not
the case for these 17; paretic limb time to peak velocity was not different for
the two aiming conditions (p > 0.05). One interpretation for this finding is that
l
i
I
1
j.
- & • control left far
—t — control right near
- B • stroke paretic far
1 stroke nonparetic near
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
the greater bimanual condition PRV was attained through control of the rate
of rise of velocity (acceleration) while maintaining a similar rise time to that in
the unimanual condition. An examination of MT for this subset of seventeen
participants revealed a shorter paretic limb MT in the bimanual than
unimanual aiming conditions (537 +/- 27 vs. 557 +/- 26 msec). Although this
difference was not statistically significant, we believe it is noteworthy as this
20 msec shorter MT stands in contrast to the overall group finding of no
difference in MT between the two aiming conditions.
Experiment II: Incongruent Condition Summary
As in Experiment I, nonparetic limb behavior was significantly different for
each dependent measure in the bimanual compared to the unimanual aiming
condition; reaction time and movement time were prolonged and peak
resultant velocity was lower. The prolonged MT provided for just a 51 +/-11
msec inter-limb difference between the nonparetic limb aiming to a far target
and the paretic limb aiming to a near target. This difference yielded an effect
size of .44, considered not even a moderate difference. Although TTP and
TAP velocity both contributed to prolonged MT in the bimanual condition, the
contribution of TAP was greater (66% vs. 16%, p < 0.05). In the bimanual
condition, the nonparetic limb demonstrated an ability to adapt to both the
internal constraint of aiming with the paretic limb and the external constraint
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
87
of aiming to a near target while the slower paretic limb aimed to a far target.
Although the right limb of the control group, when paired with the left limb
aiming to the far target, also demonstrated changes compared to its
unimanual performance, these changes were not as large as those for the
nonparetic limb. Paretic limb peak resultant velocity was greater in bimanual
compared with unimanual aiming. Specifically, 17/29 (59%) stroke
participants demonstrated this increase.
Comparison of nonparetic limb modifications across congruent and
incongruent conditions
As we hypothesized, the performance of the nonparetic limb was significantly
modified between the unimanual and bimanual aiming conditions in both
experiments. We then determined if these modifications were similar or
different across the two conditions.
Nonparetic limb modifications in reaction time, movement time, time after
peak velocity and peak resultant velocity between the unimanual and
bimanual aiming conditions were greater in the incongruent than the
congruent experiment (Figure 3.14). Nonparetic RT was prolonged by 15 +/-
3% in the incongruent compared to an 8 +/-2% increase in the congruent
experiment, this difference trending towards significance (p = 0.06).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
8 8
Nonparetic limb MT was prolonged by 39 +1-4% in the incongruent
experiment, more than twice the increase observed for the congruent
condition (15 +/-3%) (p < 0.05).
Parceling MT into its two components, TTP and TAP, again revealed that
adjustments of the NP limb are made primarily in the deceleration phase of
movement. Although prolonged compared to unimanual aiming for both
bimanual conditions, there was no significant difference in nonparetic limb
TTP prolongation regardless of whether the paretic limb was aiming to the far
or near target (17 +1-6% vs 13 +1-5%, p > 0.05). In contrast, nonparetic limb
TAP behavior did vary in the two bimanual conditions. Nonparetic limb TAP
was prolonged by 70 +/-11 % (compared to unimanual TAP) in the
incongruent experiment, much greater than the 18 +1-6% TAP increase in the
congruent experiment (p < .05). Nonparetic limb PRV decreased in both
bimanual conditions, but this decrease was larger in the incongruent
experiment (16 +1-2% vs. 8 +/-2%; p < 0.05). In summary, for both movement
planning (RT) and execution (MT, TAP, PRV) the modifications of the
nonparetic limb between the unimanual and bimanual aiming conditions were
dependent on both the internal and external constraints of the contralateral
limb.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
89
H I Incongruent (P far/NP near)
Congruent (P near/NP far)
RT MT TTP TAP PRV
Figure 3.14. Nonparetic limb modifications were greater in the incongruent
compared to congruent bimanual aiming condition for RT, MT, TAP and PRV
(p < 0.05), but not TTP.
A similarity of nonparetic limb behavior in both bimanual conditions
(congruent and incongruent) was the significant prolongation in both phases
of the velocity profile (TTP and TTP). This is in contrast to the control group
where adjustments only occurred in the deceleration phase of movement.
Data from a representative control and stroke participant illustrate this
difference in behavior. In the control example, prolongation of movement
time in the bimanual aiming condition occurred primarily in the deceleration
phase of the movement (unimanual TAP: 166 msec, bimanual TAP: 235
msec), time to peak velocity was not significantly different in the two aiming
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
90
conditions (unimanual TTP: 166 msec, bimanual TTP: 166 msec) (Figure
3.15). In contrast, velocity profiles from the stroke participant reveal that in
bimanual aiming, nonparetic limb adjustments took place in both the
acceleration (unimanual TTP: 183 msec, bimanual TTP: 216 msec) and
deceleration (unimanual TAP: 216 msec, bimanual TAP: 466 msec) phases
of movement (Figure 3.16). For the stroke group, the asymmetric aiming
conditions necessitated nonparetic limb adjustments not only in the
deceleration phase, but adaptation to be prescribed for the acceleration
phase as well.
Control
200
Bimanual
Unimanual
150
100
50
0
200 300 400 500 0 100
Time (msec)
Figure 3.15. Time series of instantaneous velocity plotted from movement
onset to offset of a single trial for representative control participant. Left limb
velocity profile in the unimanual compared to bimanual aiming condition.
Prolongation of movement time occurs in the deceleration phase only during
bimanual aiming. Limb is aiming to a near target in both conditions.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
91
Stroke
#4
Bimanual 200
Unimanual
150
100
50
0
0 100 200 300 400 500 600 700
Time (msec)
Figure 3.16. Time series of nonparetic limb instantaneous velocity plotted
from movement onset to offset of a single trial for representative stroke
participant (#4). Prolongation of movement time occurs in both the
acceleration and deceleration phases during bimanual (incongruent condition
in this example) aiming. Nonparetic limb is aiming to the near target in both
conditions shown here.
Comparison of paretic limb PRV across congruent and incongruent
conditions
Alteration of paretic limb PRV in the bimanual compared to the unimanual
aiming conditions was unique to the incongruent condition. Paretic limb PRV
was greater in the bimanual (164 +/- 7 cm/sec) compared to unimanual
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
92
condition (158 +/- 6 cm/sec) when the limb aimed a far distance while
nonparetic limb aimed a near distance. This increase in paretic limb PRV is
in contrast to no difference between the two aiming conditions in the
congruent aiming condition (unimanual: 157 +/- 7 cm/sec, bimanual: 152 +/-
6 cm/sec; p > 0.05).
To determine if it was simply aiming distance that contributed to the greater
paretic limb PRV we then compared paretic limb PRV unimanual aiming in
the near and far aiming conditions. One-way ANOVA revealed no difference
in PRV for the two unimanual aiming distances (near distance: 157 +1-7
cm/sec, far distance: 158 +/- 6 cm/sec; p > 0.05). These additional results
lead us to conclude that both paretic limb aiming distance and bimanual
coordination of the task contributed to the greater PRV in the incongruent
condition.
Bimanual Movement Initiation and Termination
Temporal adaptation was also examined by determining how frequently the
asymmetric amplitude movements were initiated and terminated together.
Movement initiation of the two limbs was considered simultaneous if it
occurred within 2 samples (< 34 msec) of each other. This same criterion
was used for movement termination. This criterion was established as a
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
93
conservative threshold given the temporal resolution of the electromagnetic
sensors and the sampling rate of 60 Hz. Overall, those with stroke had a
smaller percentage of trials with simultaneous movement initiation or
termination compared to the control group (89 +/- 3% vs. 72 +/- 3%, Main
Effect Group, p < 0.05; Figure 3.17). Post-hoc analysis revealed that for the
stroke group, the percentage of trials with simultaneous termination was
lower in the incongruent experiment (60 +/-5%) compared to the congruent
experiment (81+/- 4%) (Group x Experiment x Event interaction, p < 0.05).
— Initiation
A /A Termination
L far/R near R far/L near P near/NP far P far/N P near
Congruent Incongruent
Control Stroke
Figure 3.17. Percentage of trials with simultaneous movement initiation (dark
bar) and movement termination (light bar) for control and stroke group. P =
paretic, NP = nonparetic, L = left, R = right.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
94
The smaller percentage of trials with simultaneous offset in the incongruent
condition compared to the congruent condition is not surprising. The
congruent condition affords the limbs to be more equitable; the paretic limb
performing the easy task (aiming to near target) is paired with the nonparetic
limb performing the hard task (aiming to far target). In this more congruent
combination there is a greater percentage of simultaneous target hits than in
the incongruent combination. In the incongruent condition, both the internal
(limb paresis) and external (far movement amplitude) “difficult” constraints
are paired with the “easy” internal (no limb paresis) and external (near
movement amplitude) constraints. It is noteworthy, therefore, that despite this
incongruence at two levels of constraint, more than 60% of all trials were
terminated simultaneously, despite no explicit directions to do so. This would
suggest that temporal synchrony is a goal of the CNS that is achieved
despite both internal and external constraints to the system.
Discussion
These results extend our previous work pertaining to bimanual coordination
in the presence of decreased central motor drive due to stroke. This study
investigated the CNS’ temporal constraint on the planning and execution of a
bimanual task in the presence of both internal (body-centered) and external
(task-centered) asymmetries in healthy controls and those with unilateral
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
95
paresis. We examined the temporal adaptations of both limbs as they
performed a bimanual aiming task at different levels of congruency: 1)
Congruent (paretic limb aiming near/nonparetic limb aiming far) and 2)
Incongruent (paretic limb aiming far/nonparetic limb aiming near) and then
compared these temporal adaptations across the two levels to determine if
they were planned and executed using similar or dissimilar strategies. For
the first time, this study provides a kinematic analysis of performance
differences between unimanual and two bimanual aiming tasks that afford
different levels of congruency. We determined the locus of these effects and
interpret these results in light of the control of bimanual coordination.
Our primary finding was that the temporal constraint on bimanual movements
was robust to the addition of an external asymmetry (disparate target
location) to the internal asymmetry of unilateral stroke-induced paresis. The
nonparetic limb prolonged movement planning and execution time when
paired with the paretic limb, compared to its unimanual performance.
Furthermore, the nervous system demonstrated an adaptability to the
external task demands in that the magnitude of prolongation was a function
of paretic limb aiming distance; nonparetic limb movement time increase was
greater when it paired with the paretic limb aiming the far compared to the
near distance. This adaptation of nonparetic limb performance permitted for a
nearly simultaneous bimanual target hit in both aiming conditions.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
96
Impressively, in the congruent condition the inter-limb movement time
difference for the stroke group was no different than that of the control group.
Even for the incongruent condition, in which the limb/task combination was
even more disparate, the effect size for the inter-limb movement time
difference was small (ES = .44). The remarkable decrease in inter-limb
movement time difference when the limbs move together compared to when
they move separately strongly suggests that a different control mechanism is
operative in bimanual compared to unimanual movements.
To gain insight into the CNS control of bimanual aiming we parceled
movement time into its component parts. For the control participants,
movement time of the limb aiming the near distance increased in the
bimanual condition when paired with the limb aiming the far distance. This
result is consistent with results from similar asymmetric bimanual aiming
conditions (Kelso et a!., 1979a, 1979b; 1983; Marteniuk et al., 1984; Corcos
1984; Fowler et al., 1991). Despite asymmetric distances, time to peak
velocity was similar for both limbs (in congruent and incongruent conditions),
indicating a strong coupling and providing support for a specified movement
plan common to both limbs.
For the stroke group, time to peak velocity was nearly identical for both limbs
in the congruent condition (Fig. 6), and similar to that for the control group.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
97
For the incongruent condition, time to peak velocity was not identical for the
two limbs, due to a longer time to peak for the paretic limb (Fig. 11), most
likely due to the impaired force production of the limb and the asymmetric
nature of the task, limiting the behavioral expression of temporal synchrony
observed for symmetrical aiming in Chapter 2.
For the control group, movement time slowing was specific to the
deceleration phase of the movement, time after peak velocity. There was no
prolongation of the acceleration phase. In contrast, for the stroke
participants, prolongation of both the acceleration and deceleration phase of
the nonparetic limb’s movement occurred when paired with the paretic limb
although the deceleration phase was greater. The modifications in the
acceleration phase provides evidence that overall movement duration is
specified to some degree in advance and appears to be common to both
limbs, suggestive of a task-specific motor program (Schmidt et al., 1979).
Our data suggest that the CNS exerts a temporal constraint on the system
even in the presence of both an internal and external asymmetry. The
asymmetric aiming conditions used in this experiment necessitated
nonparetic limb adjustments not only in the deceleration phase (similar to
what we reported for the bimanual symmetric task, Chapter 2), but it appears
that the CNS prescribed an adaptation for the acceleration phase as well.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
98
Our control group data, similar to that of previous investigators, suggest that
when the CNS is faced with controlling multiple degrees of freedom, as with
bimanual aiming, it solves the problem by constraining the two limbs to act as
a single unit. We demonstrate here, for the first time, that after unilateral
stroke with subsequent mild limb paresis, this invariant temporal structure
remains largely intact. It is the limb paresis that limits the full expression of
the common temporal structure prescribed for both limbs.
Results from this experiment provide further insight regarding potential
requisite task conditions for a facilitation effect of the paretic limb in bimanual
aiming. Our findings suggest that it may not be the bimanual task alone that
affords this facilitation. Task requirements for the paretic limb may also be
important. Increase in PRV for bimanual aiming only occurred when the
paretic limb aimed a far distance (and only for a subgroup of participants),
not when it aimed a near distance. However, there was no difference in
unimanual PRV for near compared to far aiming, providing evidence as well
that it is not simply aiming distance underlying the larger paretic limb PRV.
Both a far aiming distance and the bimanual task may be requisite for the
expression of a greater PRV. This paretic limb PRV increase in bimanual
aiming corroborates results from our previous work (Rose and Winstein, in
press). In that study, the paretic limb also exhibited a small, but significant
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
99
increase in PRV when paired with the nonparetic limb compared to its
performance in a unimanual aiming condition. It is noteworthy that in that
experiment the paretic limb aimed a distance comparable to the “ far”
distance in the current study. Common to both experiments in which a
greater paretic limb PRV was observed in a bimanual compared to a
unimanual aiming condition was not only the bimanual task but also a
movement amplitude that could be described as challenging for the paretic
limb.
In this experiment we limited our external constraint manipulation to
asymmetric amplitude requirements. Asymmetric precision requirements
(i.e., differences in target width) could also be imposed to further test the
robustness of the CNS’ temporal constraint in the presence of both internal
and external asymmetry. The dependent measures used to assess bimanual
coordination were within the temporal domain only. An examination of limb
spatial adjustments may also be of interest although it has been consistently
shown that the temporal constraint to bimanual coordination is stronger than
the spatial constraint (Schmidt et al., 1979; Sherwood, 1989,1991, see also
Chapter 4 in this dissertation).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
100
The psychophysical and kinematic data of those with unilateral paresis
presented here fit the theoretical perspective described by a coordinative
structure. The hallmark of a coordinative structure, defined by Turvey and
colleagues, (1978) is that when a group of muscles is constrained to act as a
unit, some temporal relationship is preserved invariantly over changes in the
magnitude of the activity. Despite the presence of asymmetric decreased
central motor drive and asymmetric task demands, these data demonstrate
that temporal coupling does persist.
At the same time, our data show that there are limits to this temporal
constraint. Those with stroke, aiming in the incongruent condition indeed
tested the limits of this proposed coordinative structure such that the
temporal coupling between the limbs was not as rigid as observed for the
stroke group aiming in the congruent condition, nor for the control group
performance. Nevertheless, the temporal constraint of the CNS remains
impressive and the data presented here, from those with unilateral paresis,
strongly supports a theoretical model of a common motor plan for the two
limbs that characterizes bimanual motor control.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
CHAPTER 4
101
T e m p o r a l a n d S p a t i a l C o n s t r a i n t s i n a S p a t i a l l y A s y m m e t r i c T a s k : T h e
Effect of Diminished Central Motor Drive
Abstract
Inter-limb temporal and spatial coupling are significant features of human
motor control, with evidence from healthy adults that the CNS exerts a global
constraint on the timing properties of bimanual movements. Unilateral brain
damage from stroke can be considered a perturbation to the CNS and
therefore, provides a unique model to study the neural control of bimanual
movements. The relative strength of the temporal and spatial coupling was
investigated by comparing inter-limb performance of individuals with
unilateral brain damage to that of healthy controls. Thirty individuals post
stroke with intact sensation but mild unilateral paresis and twenty-nine age-
matched healthy adults performed 20-trial blocks of unimanual and
asymmetric bimanual discrete, rapid multi-joint aiming movements in
response to a lighted target in 7 conditions: no barrier between the home and
target position, and with a 10,15 or 20 cm vertical height-barrier midway in
the path of one limb while the path of the contralateral limb remained barrier-
free. Reaction time and movement time were determined from photo-emitting
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
1 0 2
diode switches at the home and target positions. Maximum vertical
displacement was captured via electromagnetic sensors attached to each
hand. Nonbarrier limb movement time or maximum vertical displacement was
regressed on the corresponding barrier limb dependent measure for each
participant to determine the strength of temporal and spatial inter-limb
coupling. Nonbarrier limb temporal and spatial behaviors were predicted by
performance of the contralateral barrier limb with the between-limb temporal
coupling more robust than spatial coupling to the decreased central motor
drive of the stroke group. Within-limb spatiotemporal relationships and
velocity profile analysis provide corroboration for the theory of a single
temporal structure specified by the CNS simultaneously for both limbs.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
103
Introduction
Bimanual coordination is requisite for many of our everyday activities. The
neural control of this coordination is likely distributed; the various loci not yet
clearly defined (Tanji et al. 1988; Weisendanger et al. 1994; Weisendanger
et al. 1996; Donchin et al. 1998; Swinnen, 2002). Despite this distributed
nature of bimanual motor control, studies of bimanual rapid discrete
movements in post-callosotomy patients have provided some useful insights
(Preilowski, 1972; Zaidel and Sperry, 1977; Tuller and Kelso, 1989; Franz et
al., 1996; Eliassen et al., 1999; Ivry et al., 1999; Kennerley et al., 2002). The
corpus callosum is thought to mediate the spatial and temporal coupling
between the hemispheres for continuous movements, whereas a subcortical
locus, possibly the cerebellum, orchestrates the temporal coupling of discrete
movements.
A unilateral hemispheric stroke offers a more common and natural
perturbation to the central nervous system (CNS) than callosotomy for the
study of bimanual control. Following this CNS insult there is diminished
central drive to the effectors with a resulting asymmetric paresis. With the
most common form of stroke that affects the areas within the distribution of
the middle cerebral artery (MCA), the corpus callosum and cerebellum are
not usually directly damaged. Rather, paresis of the contralateral upper
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
104
extremity (UE) results secondary to direct and indirect damage of the
corticospinal pathways. Such an asymmetric paresis affords a unique model
to probe the neural control of bimanual movements, and is the focus of this
investigation.
Humans can dissociate movements of the two limbs quite easily for tasks
requiring asymmetric actions such as stabilizing a jar with one hand and
removing the lid with another (Guiard, 1987). However, it is nearly impossible
to dissociate the movements of two limbs for tasks that require similar
muscle activation, but with different spatial or temporal requirements such as
when moving different amplitudes or rates. In the latter case, a behavioral
assimilation between the two limbs often occurs (Martenuik et al., 1984).
Assimilation, often referred to as coupling, (Franz et al., 1991; Franz, 1997;
Steglich, et al., 1999; Peper & Carson, 1999; Weigelt & Cardoso de Oliveira,
2003) describes the interaction of the movements between each limb.
Amplitude assimilation, for example, is the effect of each arm’s movement
amplitude on the amplitude of the contra lateral arm.
Assimilation in bimanual control can be expressed in both spatial and
temporal domains. There is considerable evidence that the CNS exerts a
global temporal constraint on the timing properties of bimanual movements
(Schmidt et al., 1979; Heuer, 1985; Sherwood, 1989). For example,
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
105
examination of bimanual tapping (Peters 1977; Klapp 1979; Yamanishi et ai.
1980) and timing tasks of anatomically distinct systems, such as speech
production and finger tapping (Klapp 1981; Smith et al. 1986; Chang and
Hammond, 1987) or clapping and walking (Muzii et al. 1984) reveal poor
coordination for two different timing patterns, particularly those that are not
simple harmonic derivatives. These findings provide evidence for tight
temporal coupling through the emergence of a strong behavioral assimilation
that is quite difficult to resist. Spatial constraints also play a role in both
continuous and discrete bimanual skilled actions (Franz et al., 1991; Franz et
al., 1996; Franz et al, 1997; Franz et al., 2001; Sherwood, 1991). In all
cases, the phenomenon of spatial coupling appears to result in the trajectory
of one limb taking on some features of the contralateral limb’s trajectory. For
example, Franz et al., (1991) had participants perform a continuous circle
drawing task with one hand while the other performed a continuous vertical
line drawing task. There was a clear tendency for the movement path of the
circle task to become more ‘line-like’ and the movement path for the line task
to become more ‘circle-like’.
While investigations into the control of bimanual movements have been quite
extensive in healthy individuals, the study of those with disparate effector
ability (i.e. unilateral stroke) has been limited to symmetric tasks in the
temporal domain (Dickstein et al., 1993; Rice & Newell, 2001, 2004;
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
106
Cunningham et al., 2002; Rose & Winstein, in press). For a continuous elbow
flexion and extension task, the nonparetic limb made adjustments to its
movement rate in a bilateral condition compared to its unilateral rate, that
approached the paretic limb movement rate (Rice & Newell, 2001). In a
small-scale study of 6 individuals post-stroke, a smoother paretic elbow
extension velocity profile was observed when paired with the nonparetic limb
compared to a unilateral condition (Cunningham et al., 2002).
We previously examined bimanual coordination using a symmetric aiming
task with identical spatial requirements in a population of mildly impaired
individuals post-stroke. Temporal assimilation occurred with adjustments
primarily in the nonparetic limb in both the planning for and execution of this
bimanual rapid aiming movement toward a single target. The nonparetic limb
exhibited prolonged reaction time and movement time when paired with the
paretic limb in the bimanual condition compared to that in a unimanual
condition. This resulted in similar planning and execution times for both limbs
(Rose and Winstein, in press). Although these previous studies of bimanual
movements provide evidence that temporal assimilation does persist to some
extent despite disparate effector ability, the tasks used did not allow for an
examination of spatial coupling. Quantification of the strength of this
preserved coupling and its comparison to that of healthy controls has yet to
be investigated.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
107
The purpose of this experiment is to extend previous work in three important
ways: 1) bimanual movement planning and control are explored in a spatially
asymmetric target-aiming task for a select group of participants with
unilateral diminished central motor drive, 2 ) inter limb coupling strength is
quantified using regression analyses to determine the degree to which one
limb’s performance predicts the performance of the contra lateral limb, and 3)
the relative strength of the temporal and spatial interlimb coupling between
groups and across conditions is compared within the same experiment.
We hypothesize that temporal and spatial coupling strength will be
diminished in those with motor stroke compared to controls, however,
considering previous evidence (i.e., Schmidt et al., 1979; Heuer, 1985;
Sherwood, 1989), temporal interlimb coupling will be more robust than spatial
interlimb coupling to diminished central motor drive; spatial interlimb coupling
will be more susceptible to stroke-induced effector paresis than temporal
coupling. Further, quantification of the coordination between limb
displacement and movement time will reveal the spatial and temporal
adjustments through which coupling is achieved during an asymmetric task.
Examination of movement planning can also provide a window into the CNS’
organization of movement. Reaction time (RT) of both upper limbs is
prolonged following stroke (Dickstein, et al., 1993). Previous studies of
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
108
healthy individuals report a longer RT for the limb performing the “ easier”
movement compared to the contra lateral limb in a bimanual asymmetric task
(Kelso et al, 1983; Goodman et al., 1983; Boessenkool et al., 1999). We
hypothesize that although overall RT for those with stroke will be prolonged
compared to controls, again and due to the global temporal constraint
exerted by the CNS on bimanual movements, the relative relationship of limb
RT will be robust to diminished central motor drive; as for control participants,
we expect the limb performing the “ easier” task will demonstrate a relatively
prolonged RT compared to the contra lateral limb.
Method
Participants
Thirty adults post-stroke (age 27 - 84 yrs; M = 63; 18 left-brain lesion, 12
right-brain lesion; 6-120 mos post-stroke; M= 29 mos) and twenty-nine age-
matched adults without stroke (range 49 - 86 yrs; M = 67) participated (Table
4.2). Participants were the same as in Chapters 2 and 3 with the omission of
data from one control participant due to an incomplete data set. All were
right-hand dominant (Oldfield, 1971). Stroke participants were screened to
verify: (1) intact visual fields (Friedman, 1992); (2) intact shoulder, elbow and
wrist proprioception (Fugl-Meyer et al., 1975); (3) active pain-free functional
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
109
range of motion in shoulder flexion and abduction, elbow flexion and
extension, and wrist and finger extension; and (4) Modified Ashworth
Spasticity Scale (Bohannon et al., 1987) scores for shoulder extensors,
elbow flexors, forearm supinators, and wrist flexors < 2. Motor impairment
was assessed with the Fugl-Meyer Measurement of Physical Performance
(FM) (Fugl-Meyer) as well as with grip, lateral pinch and palmer pinch
dynamometry (Mathiowetz et al., 1985). Specific information regarding
lesion location was obtained from MRI or CT films for all 30 stroke
participants, although due to the poor sensitivity of CT in the acute stage, CT
scans were negative for 7 of the participants. In spite of the negative CT
scan, these 7 participants manifested clear clinical signs of hemiparesis.
Task
From a seated position and in response to an LED signal, participants were
to move both hands laterally, away from midline in the frontal plane to hit a
switch mounted on a target, as soon and as fast as possible. The movement
required shoulder abduction with elbow extension. A vertical barrier was
placed mid-way between the home and target position and only on one side
(Figure 4.1). The task was adapted from previous work by Kelso and
colleagues (1983) and Goodman and colleagues (1983).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
110
LED
target target
Nonbarrier limb
Barrier limb
LED
target target
Nonbarrier limb
Barrier limb
Figure 4.1. Experimental set-up depicting participant position, the three
different height vertical barriers, and respective targets for each limb
(overhead view).
Apparatus
The apparatus consisted of two sets of 2.5 cm2 switches: home-position (one
for each hand) and target-position (one for each hand). Each switch
consisted of a solid-state infrared photo light emitter and diode. A 0.5 x 1 cm
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
111
blackened mylar sheath was mechanically mounted to each of the four
switches. When the light beam from the photo emitter to diode was broken by
the mylar, the switches were considered “on” (home switches) or “ off’ (target
switches). The home-position switches were secured to the table surface,
aligned for each participant in midline of the long axis of the humerus. Target
switches were affixed to a 5 cm width wooden block located 24 cm to the left
and to the right from their respective home switches. The “ go” signal was a
LED mounted within a box and located directly in front and midline. Effector
end-point displacement was recorded via an electromagnetic sensor
attached to a rubberized paddle (10.5 cm diameter), affixed to each hand.
Movements were recorded with a “ 6-D Research” electromagnetic motion
capture and analysis system (Skill Technologies, Inc.) at a 60 Hz sampling
rate. Trial onset, LED onset, and movement onset and offset of each hand
were sampled on separate channels from individual TTL pulses interfaced
with a 400 MHz personal computer.
Procedure
Participants signed an informed-consent form approved by the Institutional
Review Board at the University of Southern California. Verbal instructions
and demonstration of the task were provided at the beginning of the session.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
112
Participants performed one block of 5 practice trials before each
experimental condition. Participants sat at a height-adjusted table with
elbows at 90° flexion, shoulders and wrists in neutral, and forearms in full
pronation at the start position. They were instructed to move “as soon as
possible, and as fast as possible” to hit the targets once the “go” LED was
illuminated. No explicit instructions were provided to initiate or terminate each
limb’s movement together.
Each trial began with a verbal “ready” followed by the LED “ go” presented
randomly within a variable foreperiod that was 500 to 1500 msec. The trial
ended when the target switches were hit squarely. Participants then returned
their hands to the home position and awaited the next “ready” signal. There
were 7 bimanual aiming conditions: Condition 1: each limb’s path was
barrier-free (the only symmetric condition); Conditions 2-4: non paretic limb
traversed a 10, 15, or 20 cm barrier while the paretic limb’s path was barrier-
free; Conditions 5-7: paretic limb traversed a 10,15, or 20 cm barrier while
the nonparetic limb’s path was barrier-free (Table 4.1). Each condition was
presented in separate, 28-trial blocks (25 valid trials plus 3 catch trials - no
“ go” LED followed verbal “ready”). Frequently, during each trial block, the
experimenter reminded the participant to move as soon and as fast as
possible, in response to the LED signal. The symmetric aiming condition
(Condition 1) was always presented first. Conditions 2-4 and 5-7 were
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
113
blocked with the order of each block counterbalanced across participants in
each group.
Table 4.1. Seven experimental conditions
Condition Paretic Limb Barrier Nonparetic Limb Barrier
___________ (left for controls) (right for controls)_______
1 none none
2 none 10 cm
3 none 15 cm
4 none 20 cm
5 10 cm none
6 15 cm none
7 20 cm none
Data reduction
Trial onset, LED onset, switch onset and offset were incorporated into one
data file for each aiming condition and analyzed offline with Datapac 2000
Laboratory Applications System (RUN Technologies Co.). For each limb,
movement onset was defined as the time value at which hand sensor
acceleration value exceeded 0.03 cm/sec2. Movement offset was the time
when the target-position switch signal exceeded 0 mV. Movement time (MT)
was defined as movement offset - movement onset, reaction time (RT) as
the time from LED onset to movement onset. Limb peak vertical
displacement (PVD) was the maximum vertical displacement each hand
sensor attained within a trial. All trials were visually screened and only
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
114
acceptable trials were selected for analysis. For a trial to be considered for
analysis both hands had to initiate squarely from the onset switches and hit
their respective target switches. Additionally, trials with RTs that were greater
or less than the participant’s mean +1-2 SD for that trial block were discarded.
If more than 20 acceptable trials remained following application of these
criteria, the middle 20 trials were selected for analysis.
Data analysis
Interlimb Coupling. All 60 data points (20 trials x 3 barrier heights) were used
to create a line of best fit for each participant. Specifically, MT (or PVD) of the
nonbarrier limb was regressed on MT (or PVD) of the contralateral limb for
conditions using the 10 cm, 15 cm and 20 cm barriers. The primary
dependent measures for temporal and spatial coordination are movement
time (MT) and peak vertical displacement (PVD), respectively. The slope of
the regression line represents the relationship between each limb. Interlimb
coupling was defined as a slope for the best-fit line that was significantly
different from zero (p < .05). We used the squared correlation coefficient
from the regression analysis to quantify the strength of interlimb coupling.
The interlimb linear correlation coefficients for temporal and spatial
parameters were calculated first. The squared correlation coefficient provides
a direct estimate of the contribution of the common variability to the total
variability observed in each movement parameter. In this way, r provides an
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
115
estimate of the magnitude of interlimb coupling, and J R 2 provides an estimate
of the strength of interlimb coupling during this asymmetric bimanual task.
(Regression analysis assumes that the independent variable is constant.
Barrier limb behavior was used as the independent variable in our regression
analyses, although we cannot assume constant behavior from trial to trial.
However, this misuse of regression analysis should not detract from the
group comparison of R2, which was the ultimate aim for these analyses.)
To test our hypotheses that interlimb coupling would be diminished in those
with stroke compared to controls and that temporal coupling would be more
robust than spatial coupling to diminished central drive a 2 (Group) X 2
(Coupling Variable) X 2 (Limb) general linear model ANOVA with repeated
measures on the last two factors was performed. All correlation coefficients
were converted to Fisher z scores prior to statistical analysis.
To investigate limb adjustments in displacement and/or time to maintain
coupling in this asymmetric task each limb’s MT was regressed on its PVD.
To test for the effects of group (control, stroke), barrier condition (present,
absent), and limb (paretic, nonparetic) on the MT/PVD regression line slope,
we conducted a 2 (Group) X 2 (Barrier Condition) X 2 (Limb) general linear
model ANOVA with repeated measures on the last two factors. As a follow-
up to this analysis, we parceled MT into two sub-phases: 1) acceleration
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
116
phase, time to peak (TIP) velocity and 2) deceleration phase, time after peak
(TAP) velocity. A 2 (Group) x 2 (Barrier Condition) x 2 (MT subcomponent)
general linear model ANOVA with repeated measures on the last two factors
was performed.
Movement Planning. Reaction time was collapsed across the three barrier
heights (10,15, 20) to obtain a mean RT per limb and barrier condition for
each participant. To test for the effects of group (stroke, control), barrier
condition (present, absent), and limb (paretic, nonparetic) on RT, we
conducted a 2 (Group) X 2 (Barrier Condition) X 2 (Limb) general linear
model ANOVA with repeated measures on the last two factors. Significance
was set at p< 0.05 for all comparisons. Post-hoc pairwise comparisons with a
Bonferroni correction determined the locus of any significant interaction.
Statistical procedures were conducted using SPSS statistical software (v.
11.5).
Results
Clinical Impairment
Group mean FM was 60 (range 43-66) out of 66 total possible. Six subjects
had a normal FM score; however, grip and pinch strength of the paretic limb
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
117
provides a more sensitive measure of motor impairment, especially in this
mildly-impaired group. Stroke group means, expressed as a percentage of
age, gender and limb-matched norms (Mathiowetz, et al., 1985) were all less
than normal: grip: 65+/-24%; lateral pinch: 71+/-23%; palmer pinch: 63+/-
18%. Although participants were able to perform the experimental task,
upper extremity motor deficits remained as evident by these force
measurements. Proprioception was intact at the shoulder, elbow, and wrist
for all stroke participants. Two participants exhibited impaired proprioception
at the thumb interphalangeal joint (Table 4.2). Participants wore paddles on
their hands, holding the digits static as the proximal joints were engaged to
aim toward the targets. Targets were hit with the entire paddle; no individual
digit movement required. Therefore, impaired thumb proprioception would
not have been expected to influence task performance.
Lesion location was unknown for seven participants secondary to a negative
CT image. Of the 23 participants with identifiable lesions, 17 had lesions in
the distribution of the MCA and 6 had pontine lesions (posterior cerebral
artery distribution). Although 6 participants had lesions involving the parietal
cortex, upper extremity sensation was intact except for the two participants
noted previously (Table 4.2).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
118
Table 4.2. Summary of stroke participants arranged hierarchically according
to Upper Extremity Fugl-Meyer motor score (M) (total possible=66)._______
No. Age Lesion
side
M P Grip
(%)
LP
(%)
PP
(%)
Lesion
type
Lesion Location
1 62 L 43 8 59 29 39 I Corona radiata
2 67 R 50 8 55 44 24 I CT negative
3 43 L 54 8 39 34 24 I Frontoparietal ctx
subcortical areas
4 66 R 55 8 38 45 42 I Parietal ctx
5 68 L 56 8 59 100 54 I Thalamus
6 53 R 57 8 94 94 52 I CT negative
7 51 L 58 8 27 56 55 I Fronto-parieto-
temporal ctx
8 49 L 58 7 27 48 37 H Frontoparietal ctx
9 80 L 59 8 60 63 71 I Pons
10 69 R 59 8 66 27 55 I Posterior limb IC
11 72 L 60 8 28 # # I CT negative
12 64 R 60 8 48 78 52 H Pons
13 71 R 60 8 79 86 68 I CT negative
14 69 R 60 8 72 90 80 I Pons
15 42 L 62 7 45 66 75 I Corona radiata
16 66 L 62 8 62 61 77 H Lateral thalamus,
insular ctx
17 63 L 62 8 64 81 69 I Pons
18 73 R 62 8 65 86 49 I CT negative
19 60 L 63 8 78 100 74 I IC
20 61 L 63 8 75 70 77 I CT negative
21 71 L 63 8 100 96 74 I Pons
22 55 L 64 8 84 96 80 I Parietal ctx,
corona radiata
23 62 R 65 8 100 94 78 I Pons
24 80 L 65 8 90 79 64 I Posterior limb IC
25 54 L 66 8 33 50 53 I Thalamus
26 73 R 66 8 90 74 73 I Basal Ganglia
27 73 L 66 8 27 34 60 I Corona Radiata
28 84 R 66 8 100 92 74 I Thalamus
29 69 L 66 8 100 95 100 I Tempoparietal
ctx, subcortical
30 27 R 66 8 76 80 84 I CT negative
# data not available
Grip, Lateral Pinch (LP) and Palmar Pinch (PP) values represent percentage
strength of paretic (P) limb compared to limb, age and gender matched normative
data (Mathiowetz et al., 1985), P = UE proprioception score (Fugl-Meyer et al.,
1975), I = infarct, H = hemorrhage, IC = internal capsule, ctx = cortex.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
General group differences in movement time and reaction time
119
Baseline differences between groups are defined using mean MT and RT for
the bimanual symmetrical condition (Table 4.1, Condition 1). Of all
symmetrical aiming trials, 10% (57/580) of the control and 14% (86/600) of
the stroke group trials were excluded from analysis due to RT outside of pre-
established boundaries, movement not initiated squarely from start switches,
or target miss. Movement time and RT were approximately 100 ms and 60
ms longer, respectively, for the stroke (MT = 367+/-14 msec; RT = 290 +/ 8
msec) compared to the control group (MT = 259 +/-14 msec; RT = 233+/- 8
msec) (Group Main Effect, p < 0.05). Additionally, and not surprisingly, there
were differences between limbs in the stroke but not the control group
(Group X Limb interaction, p < 0.05). Both MT (P: 379 +/- 16 msec; NP: 355
+/-13 msec) and RT (P: 295 +/- 8 msec; NP: 284 +/- 9 msec) were
significantly longer for the paretic compared to the nonparetic limb but not
different between the right and left limbs for the control group. The between-
limb differences in MT and RT for the stroke group were slightly larger than
those reported for the bimanual symmetric task described in Chapter 2
(Table 2.2), most probably due to differences in task demands. Although both
were symmetric, the task in Chapter 2 involved both limbs aiming to a single
target, whereas the symmetric task described here required the two limbs to
aim towards different targets.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
1 2 0
L i m b t r a j e c t o r i e s
The time series limb trajectory profiles reveal how the nonbarribr limb
assimilated to the barrier limb in displacement and time as a function of
barrier height. Limb trajectory profiles from a representative control and
stroke participant illustrate the temporal and spatial adjustments exhibited by
the nonbarrier limb to maintain interlimb coupling, despite asymmetric task
demands (Figure 4.2). In general, the left limb (similar results for the right
limb) of the control participant (Figure 4.2A) and the nonparetic limb of the
stroke participant (Figure 4.2B) exhibited systematic increases in peak
vertical displacement and movement time as the contralateral limb traversed
each progressively higher barrier. When the paretic limb was in the
nonbarrier role (Figure 4.2C), overall, vertical displacement increased and
MT was prolonged as barrier height increased, but these adjustments were
not as systematically scaled compared to the opposite condition. It is
noteworthy, however, that the paretic limb demonstrates clear assimilation,
albeit imperfectly, with the nonparetic limb. The magnitude and strength of
assimilation is quantified in the next sections.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Control
A .
E
o,
c
a >
E
0)
a
ro
Q .
C fl
" O
1
•e
a >
>
40
32
24
/ y " \ \ 16
8
0
500 100 300 -300 100 -500
Nonbarrier Limb
(Left)
Time (msec)
Stroke
Participant #4
Barrier Limb
(Right)
B.
E
3*
-3^
c
0)
£
8
J S
Q .
< n
1
'■ e
C D
>
— no barrier
— 10 cm
— 15 cm
— 20 cm
30
24
18
12
6
0
900 300 600 -600 -300 0 -900
0 barrier
10 cm barrier
15 cm barrier
20 cm barrier
Nonbarrier Limb T im o frrtQ A fA earner Limi
(Nonparetic) 1 l m e (Paretic)
Barrier Limb
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
122
C.
E
c
0
E
8
n s
C L
( O
T3
1
t
0
>
Stroke
Participant #4
30
24
18
12
/ A \
6
0
0 300 600 900 -900 -600 -300
■” no barrier
— 10 cm
■" 15 cm
— 20 cm
Nonbarrier Limb Time (rnsec) Barrier Limb
(Paretic) (Nonparetic)
Figure 4.2. Single trial displacement profiles over time from the no barrier
plus the three asymmetric aiming conditions from a representative control
and stroke participant, A. control (left as nonbarrier limb), B. stroke # 4
(nonparetic as nonbarrier limb), C. stroke # 4 (paretic as nonbarrier limb). In
A. and B. the nonbarrier limb adjusts in time and space to the barrier limb
behavior. In C., adjustments are present, but not as distinctly parameterized
as in A. and B.
Quantification of Interlimb Coupling Strength
The best-fit line derived from each participant’s data depicts the relationship
between each limb in asymmetric aiming conditions (2-7). Of all trials across
the three barrier heights, 8% (139/1740) of the control and 17% (310/1800)
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
123
of the stroke group trials were excluded from analysis due to RT outside of
pre-established boundaries, movement not initiated squarely from start
switches, or target miss. Figure 4.3 provides examples from individual
participant data for the between-limb analyses with MT for one control and
two stroke participants. Although the slopes vary, all three examples
demonstrate that as MT of the barrier limb (x-axis) increased there was an
associated increase in nonbarrier limb MT (y-axis). Individual best-fit lines
for all participants and the group means are shown in Figure 4.4 using MT
and Figure 4.5 using PVD as the dependent measure.
A.
Control
800
R 2 = .89
o
a >
W )
E
600
H
2
Q L
400 X >
E
Q )
'£
r a
X2
c
o
Z
200
200 400 600 800 0
Barrier Limb (L) MT (msec)
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
124
Stroke (#29)
800
R 2 = .54
o
0
to
E
600
Q.
Z
400
-Q
E
L _
0 )
E
C O
4 0
c
o
z
200
0 200 400 600 800
Barrier Limb (P) MT (msec)
Stroke (#29)
800
O
0
O T
E
600
D l
400 J Q
E
0
'E
E
c
o
z
200
0 200 400 600 800
Barrier Limb (NP) MT (msec)
Figure 4.3. Nonbarrier and barrier limb movement time data from all 3
asymmetric aiming conditions with resulting best-fit line and regression
equation from a representative control and stroke participant A. control (right
as nonbarrier limb), B. stroke # 29 (nonparetic as nonbarrier limb), C. stroke
# 29 (paretic as nonbarrier limb).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
125
Temporai Coupling. The slope of the line formed by the regression of
nonbarrier on barrier MT, with the right limb as the nonbarrier limb (Table 4.1,
Conditions 5-7), was significantly different from zero for 27/29 control
participants (r = 0.79, range: -0.06- 1.38; p < .05), providing evidence for
interlimb temporal coupling in 93% of the control group. Specifically, as
barrier limb MT increased, there was a concomitant increase in nonbarrier
limb MT. On average, 64+/- 5% of the variance in right nonbarrier limb MT
could be explained by left barrier limb MT. In the converse conditions (Table
4.1, Conditions 2-4) for the control group, with the left limb as the nonbarrier
limb, the regression line slope was significantly different from zero for 28/29
(97%) control participants (r = 0.72, range 0.0 - 1.0, p < .05). On average
57+/-4% of the variance, could be explained by right barrier limb MT (Table
4.3, Figure 4A. and B.).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
126
A.
F? = .6 4 + /-.0 5
Y = 79x + 42.09
800 -
B O O -
400 -
200
400 600 800 1000
Barrier Limb (L) MT (m sec)
B.
FP = .57+/-.04
Y = ,72x + 77.78
800 -
I—
£
600 -
_i
■Q
E
_I
400
200 -
c
o
z
400 600 800 1000
Barrier Limb (R) M T (m sec)
Figure 4.4. individual participant (dashed) and group average (solid) best-fit
lines for A. control (right as nonbarrier limb), B. control (left as nonbarrier
limb). Equation is of average best-fit line (L= left, R = right).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
127
Table 4.3. Group means and standard errors of percent variance in
nonbarrier (NB) limb behavior explained by the barrier limb performance,
derived from the line of best fit for each participant
Temporal Spatial
Control Right NB Left NB Right NB Left NB
Mean(SEM) 64(5) 57(4) 61(5) 62(5)
Range 0-96 0-92 0-94 0-94
Stroke Nonparetic NB Paretic NB Nonparetic NB Paretic NB
Mean(SEM) 45(5) 42(4) 35(5) 28(5)
Range 0 - 9 4 0-89 0-85 0-74
Across Conditions 5-7, the slope of the line formed by the regression of
nonbarrier limb (nonparetic) MT on barrier limb (paretic) MT, was significantly
different than zero for 26/30 stroke participants (r= 0.59, range: -0.18-1.05;
p < .05). On average, 45+/- 5% of the variance in nonbarrier, nonparetic limb
MT could be explained by paretic barrier limb MT. In the converse conditions
(2-4), 28/30 had regression lines significantly different from zero (r = 0.57,
range: .06 - 1.0, p < .05); 42+/- 4% of the variance in paretic nonbarrier limb
MT could be explained by nonparetic barrier limb MT. Group average
temporal coupling strength was significantly lower for the stroke group
compared to controls and there were no group by limb interactions, (Group
Main Effect, p< 0.05; Group x Limb interaction, ns; Table 4.3, Figure 4.5A
and B).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
1 2 8
1000
F?= .4 5 + /-.0 5
o
59x + 160.39
i/S
£ 800
t—
2
600
-Q
£
400
200
800 800 1000 200 400
Barrier Limb (P) MT (msec)
B.
o
0 3
if,
jE
S
oT
a »
'E
cs
xs
£ =
o
1000
F ? = .42+/-.04
58x + 178.17
800
600
400
200
0
800 1000 400 B O O 200
Barrier Limb (NP) MT (m sec)
Figure 4.5. Individual participant (dashed) and group average (solid) best-fit
lines for A. stroke (nonparetic as nonbarrier limb) and B. stroke (paretic as
nonbarrier limb). Equation is of average best-fit line. (P = paretic, NP =
nonparetic)
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
129
Spatial Coupling. In this analysis, the slope of the regression line represents
the relationship between limbs for peak vertical displacement. With the right
limb as the nonbarrier limb (conditions 5-7), the slope of the line formed by
the regression was significantly different from zero for all 29 control
participants (r = .38, range: -0.29 - 1.32, p< 0.05). On average, 61+/- 5% of
the variance in right nonbarrier limb peak vertical displacement was
explained by left barrier limb peak vertical displacement. One participant’s
vertical displacement data resulted in a negative slope, such that as the
barrier limb increased in vertical displacement, the nonbarrier limb decreased
in vertical displacement. In the converse conditions (2-4), PVD data from
27/29 control participants produced regression line slopes significantly
different from zero (r = .39, range: -0.06 -0.83, p< 0.05). On average, 62+/-
5% of the variance in left nonbarrier limb peak vertical displacement could be
explained by right barrier limb peak vertical displacement (Table 4.3, Figure
4.6A and 6B).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
130
A 30 -| F ? = .61+/-.05
Y = ,30x + 2.52
E
a
g 20 -
CL
B e
20 30
Barrier Limb (L) PVD (cm)
B .
F?= .62+/-.05
Y = .39X + 3.16
E
U
Q
>
CL
3
O J
30 20
Barrier Limb (R) PVD (cm)
Figure 4.6. Individual participant (dashed) and group average (solid) best-fit
lines for A. control (right as nonbarrier limb), B. control (left as nonbarrier
limb). Equation is of average best-fit line (L= left, R = right).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
131
Across conditions 5-7 (Table 4.1), the slope of the line formed by the
regression of nonbarrier limb (nonparetic) PVD on barrier limb (paretic) PVD,
was significantly different than zero for 22/30 stroke participants (r = 0.32,
range: -0.10-1.15; p < .05); 35+/- 5% of the variance in nonparetic nonbarrier
peak vertical displacement could be explained by paretic barrier limb peak
vertical displacement. The markedly steeper slope for two stroke participants
(Figure 4.7A) may indicate an actual overcompensation in nonbarrier limb
vertical displacement. In the converse conditions (2-4, Table 4.1), 22/30 had
regression lines significantly different from zero (r = 0.23, range: -0.14 - 0.55,
p < .05); 28+/- 5% of the variance in paretic nonbarrier limb PVD could be
explained by nonparetic barrier limb PVD. Similar to that for temporal
coupling, group average spatial coupling strength was lower for those with
stroke compared to controls with no group by limb interaction (Group Main
Effect, p< 0.05; Group x Limb interaction, ns; Table 4.3, Figure 4.7A and
4.7B).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
132
A.
F?= .35+/-.05
on
Y = ,31x+ 5.54
E
u
Q
>
CL
k _
CD
10 20 30 40
B arrier Lim b (P) PVD (cm )
F? = .28+/-.05 30
Y = ,23x + 5.O S
C J
£ 20
Q _
S O
« 3
■g 10
o
0
1 0 20 30
Barrier Lim b (NP) PVD (cm )
Figure 4.7. Individual participant (dashed) and group average (solid) best-fit
lines for A. stroke (nonparetic as nonbarrier limb) and B. stroke (paretic as
nonbarrier limb). Equation is of average best-fit line.P = paretic, NP =
nonparetic).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
133
Temporal vs. Spatial Coupling Strength across Groups
Group differences in temporal and spatial interlimb coupling strength and
between subject variability in both groups, are most readily illustrated with
box and whisker plots that display each participant R2 within the group
distribution for each functional condition (Figure 4.8A and B). Individual
temporal and spatial coupling strength was widely distributed within both
groups, but clearly different between groups.
V C
J D
J
CD
. Q
* o
0
• 4 —
o
ID
0
i—
Q .
J D
J
C O
z
C
c
05
'C
05
>
1.00
0.80
0.60
0.40
0.20
0.00
Temporal Analysis
ii
is...
LN B
Control
R N B NP NB P NB
Stroke
Figure 4.8A. Box and whisker plots of individual R2 values for each limb in
the nonbarrier role. Data points within the box are from the 25th to 75th
percentile of the group distribution. Error bars or "whiskers" from the edge of
the box are drawn to the 5th and 95th percentile. The solid horizontal line
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
134
within each box is the mean R2 for the data set. (L=left, R=right, NR
nonparetic, P = paretic, NB=nonbarrier)
Spatial Analysis
■ Q
E
Q Q
> .
n
T 3
C D
■ o
< B
! —
Q.
■ Q
|
m
CD
O
C
T O
" i_
r o
>
1.00
0.80
0.60 -
0.40 -
0.20
0.00
L NB R NB NP NB P NB
Control Stroke
Figure 4.8B. Box and whisker plots of individual R2 values for each limb in
the nonbarrier role. Data points within the box are from the 25th to 75th
percentile of the group distribution. Error bars or "whiskers" from the edge of
the box are drawn to the 5th and 95th percentile. The solid horizontal line
within each box is the mean R2 for the data set (L=left, R=right, NP =
nonparetic, P = paretic, NB=nonbarrier).
Once inter limb coupling strength was quantified, we were able to compare
temporal and spatial coupling strength across both groups of participants.
Temporal and spatial coupling were similar for the control group (Temporal
61 +/- 4%; Spatial: 62 +/- 4%). This was not the case for the stroke group.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
135
Spatial interlimb coupling strength was significantly lower than temporal
coupling strength (Temporal 44 +/- 3%; Spatial: 31 +/- 4%; Group x Coupling
Variable interaction; p < .05, Figure 4.9).
1.00
Temporal Coupling
Control
Stroke
0.80 -
S 0.60 -
0
(5 0.40
0.20 -
Spatial Coupling
Figure 4.9. Bar graph of group mean R2 for temporal and spatial measures.
Overall interlimb coupling stronger for HC (.61+/-.03%) than for ST (.38+/-
03%); Main Effect Group (p < .05). HC: no difference between temporal and
spatial coupling; ST: temporal coupling (R2 = .44+/-.03) > spatial coupling
(f?2 = .31+/-.04). Group x Coupling Type interaction (p < .05).
To probe the mechanisms underlying interlimb coupling, we compared the
slope of the line formed between MT and PVD for the barrier limb and
nonbarrier limb within each functional condition and group. Data from one
representative control and two representative stroke participants
demonstrate the motor control strategy. In each case, the MT/PVD slope was
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
99
136
significantly greater for the nonbarrier limb compared to the barrier limb
(Figure 4.10).
In the healthy control example, barrier limb vertical displacement ranged from
11-32 cm (21 cm range) as the limb traversed progressively higher barriers.
This range in displacement was associated with a range of MT from
approximately 300-500 msec (Figure 4.10A). In contrast, vertical
displacement of the nonbarrier limb was significantly lower and did not attain
the same PVD as the barrier limb. Maximum vertical displacement ranged
from 5-15 cm (10 cm range). Despite this discrepancy in vertical
displacement, the MT’s were similar between the two limbs. As for the barrier
limb, MT range for the nonbarrier limb was approximately 300-500 msec.
This resulted in a steeper slope for the nonbarrier compared to the barrier
limb (Figure 4.10B).
The stroke participants portrayed a similar pattern whether the nonparetic
(Figure 4.11) or the paretic limb (Figure 4.12) was in the nonbarrier role:
despite a lower vertical displacement, the nonbarrier limb MT was similar to
that of the barrier limb. Overall, across both groups and both limbs, the
MT/PVD slope for the nonbarrier limb regression line (b = 13.4+/-.90) was
significantly greater than that for the barrier limb (b = T.9+/-.42). Taken
together, the unit increase in MT per unit increase in vertical displacement
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
137
was greater for the nonbarrier compared to the barrier limb for this
asymmetric aiming task (Condition Main Effect, p <0.05; Group Main Effect,
ns; Group x Limb, ns, Figure 4.13).
Barrier Limb
700
y=5.84x+310.04
600
V )
E
500
< D
E
h -
I
400
A A
>
2
300
200
1 0 2 0 30 40 0
Vertical Displacement (cm)
Nonbarrier Limb
B. 700 -I y=17.53x+257.69
600
C O
~ 500
to
E
i-
c 400
§ !
o 300
200
0 1 0 2 0 30 40
Vertical Displacement (cm)
Figure 4.10. Best-fit line with equation derived from regression of MT on PVD
across the three asymmetric aiming conditions from a representative control
participant for A. barrier limb, B. nonbarrier limb. Greater slope for nonbarrier
compared to barrier limb. MT increase across barrier heights similar in both
limbs but with PVD increase greater in the barrier than nonbarrier limb.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
138
Barrier Limb
A. (Paretic)
700
600
C O
E
M ,
500
0 )
E
1 -
c
400
< u
>
o
5
300
200
0 10 20 30 40
Vertical Displacement (cm)
Nonbarrier Limb
(Nonparetic)
700
y=16.53x+275.85
„ 600
E.
| 500
F
| 400
1
2 300
A.
200
1 0 30 40 0 20
Vertical Displacement (cm)
Figure 4.11. Best-fit line with equation derived from regression of MT on PVD
across the three asymmetric aiming conditions from a representative stroke
participant for A. barrier limb (paretic), B. nonbarrier limb (nonparetic).
Greater slope for nonbarrier compared to barrier limb. MT increase across
barrier heights similar in both limbs but with PVD increase greater in the
barrier than nonbarrier limb.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
139
Barrier Limb
(Nonparetic)
700
▲
600
w
E
500
0
E
t —
1 400
E
>
| 300
200
0 10 20 30 40
Vertical Displacement (cm)
Nonbarrier Limb
(Paretic)
700
y=10.72x+357.34
600
A
A.
C O
E
500
o
E
h -
C
400
0
>
300
200
0 10 20 30 40
Vertical Displacement (cm)
Figure 4.12. Best-fit line with equation derived from regression of MT on PVD
across the three asymmetric aiming conditions from a representative stroke
participant for A. barrier limb (nonparetic), B. nonbarrier limb (paretic).
Greater slope for nonbarrier compared to barrier limb. MT increase across
barrier heights similar in both limbs but with PVD increase greater in the
barrier than nonbarrier limb.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
140
MT/PVD Slope
o
>
C L
C D
Q .
O
C O
Barrier
Nonbarrier
L R R L P NP NP P
Control Stroke
Figure 4.13. Bar plot showing mean and SEM of the MT/PVD slope of each
limb in the barrier and nonbarrier limb role. Nonbarrier limb slope > Barrier
limb slope for each of the limb-barrier combinations (p < 0.05). (L= left, R =
right, P = paretic, NP = nonparetic, MT = movement time, PVD = peak
vertical displacement)
We used an analysis of the velocity profiles to probe the locus of the MT/
PVD effect for the nonbarrier limb. Although TAP velocity was longer than
TTP for both the barrier (267+/- msec vs. 212+/- msec ) and nonbarrier
(236+/- msec vs. 184+/- msec) limbs (Main Effect Barrier Condition, Main
Effect MT subcomponent, p < .05) there were no differences in the
relationship of the two subcomponents between barrier conditions (MT
subcomponent x Barrier Condition interaction, ns, p < .05). Each velocity
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
141
subcomponent of nonbarrier limb MT closely mirrored that of the barrier limb.
Velocity profiles from one representative control participant and two stroke
participants illustrate that the dynamic changes in nonbarrier limb resultant
velocity mirrors closely that of the barrier limb throughout the movement
(Figure 4.14A - C) and appear to be relatively independent of central paresis.
a. Control
300
Barrier Limb (L)
Nonbarrier Limb (R)
% 200
C/3
E
o
o
o
0 3
>
100
500 100 200 300 400 600 0
Time (msec)
Figure 4.14A. Single trial time series of instantaneous resultant velocity
plotted from movement onset to movement offset from a representative
control participant as left limb traverses 20 cm barrier. Solid line is barrier
limb, dashed line is nonbarrier limb.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
142
Stroke
300 -i
Barrier Limb (P)
Non barrier Limb (NP)
o
o
( /}
E
o
> ,
o
o
< D
>
100 -
0 100 200 300 400 500 600
Time (msec)
Stroke
300 -i
200
Barrier Limb (NP)
Nonbarrier Limb (P)
3)
>
100
100 200 300 400
Time (msec)
500 600
Figure 4.14B-C. Single trial time series of instantaneous resultant
velocity plotted from movement onset to movement offset from two
representative stroke participants: B. stroke participant # 21,
(nonparetic as nonbarrier limb) C. stroke participant # 24, (paretic as
nonbarrier limb). The second, smaller velocity peak near the end of the
movement represents a ballistic plunge just prior to target impact.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
143
Coupling Strength and Motor Impairment
For the stroke participants we then determined if there was a relationship
between temporal or spatial coupling strength and impairment of the paretic
limb. A separate correlation was calculated between the FM motor score and
each of the four R2 values determined for each participant (paretic,
nonparetic limb in nonbarrier role; temporal and spatial coupling strength).
For each case there was not a significant relationship between FM motor
score and coupling strength. However, the largest correlation, and thereby
the one closest to reaching significance was between FM motor score and
spatial coupling strength with the paretic limb as the nonbarrier limb (Table
4.4). This corroborates our earlier finding that spatial coupling with the paretic
limb in the nonbarrier role resulted in the lowest group R2 value. Spatial
coupling appears to be more closely related to effector impairment.
Table 4.4. Pearson correlation (r) between FM motor score and inter-limb
coupling strength_______________________________________
Nonparetic Nonbarrier Paretic Nonbarrier
Temporal •21 (p = .27) .23 (p = .23)
Spatial .27 ( id = .16) .30 (p = .11)
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
144
Movement Planning
Overall, RT was longer for the stroke group (299 +/-8 msec) than the control
group (231 +/- 8 msec) (Group Main Effect, p < .01). Additionally, there was
an overall effect of Barrier Limb, with RT for the nonbarrier limb 12 +1-2 msec
longer than that for the barrier limb (271 +/6 msec vs 259 +/-6 msec). Both
the stroke and control groups demonstrated a prolonged RT for the
nonbarrier limb (Group x Barrier interaction, p = NS, Figure 4.15), consistent
with that of previous work with younger control subjects (Kelso, et al, 1983;
Goodman et al, 1983).
350
310
270 -
230
O
0)
w
E
o >
E
i_
c
o
o
C O
£T 190
150
T
--S--
control left
control right
stroke paretic
stroke nonparetic
Barrier
1 ------
Nonbarrier
Figure 4.15. Plot of reaction time by barrier condition for each group.
Reaction time longer for nonbarrier compared to barrier limb (Main Effect of
Group, Main Effect of Barrier, p < 0.05).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
145
D i s c u s s i o n
This study examined the human central nervous system’s capability to,
despite diminished unilateral central motor drive, plan and coordinate two
simultaneous aiming movements that entailed disparate task requirements.
We provide, for the first time, a quantitative analysis of the strength of the
relationship between the paretic and nonparetic limbs of individuals post
stroke performing a spatially asymmetric task. Both spatial and temporal
coupling were examined and compared to probe the nervous system’s
solution to this complex bimanual task with disparate effectors.
Interlimb coupling strength
These results provide evidence that in the presence of unilateral diminished
central motor drive with subsequent mild hemiparesis, a temporal and spatial
relationship persists between the two limbs, although the strength of this
relationship is weakened compared with that for age-matched controls. For
the majority of stroke participants, a significant portion of the variance in the
nonbarrier limb’s movement could be explained by the barrier limb’s
behavior; the two limbs are clearly not independent of one another.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
146
Proprioceptive information is vital for controlling naturalistic movements
involving multiple joints (Ghez and Sainburg, 1995) and is considered to play
an even larger role in bimanual movements. Jackson et al, (1999) have
argued that a sensorimotor mechanism, based upon proprioceptive coding of
limb position and motion exists to maintain interlimb coordination during
movement execution. Stroke participants in this study had proprioception
available to guide interlimb coordination. Any observed coordination deficits
compared to controls, therefore, could not be attributed to proprioceptive
impairment, but must be attributed to the diminished central motor drive.
Temporal and spatial interlimb coupling strength varied widely across all
participants. Even within the control group, a more motorically homogeneous
sample than the stroke group, interlimb coupling strength ranged between
0% and 96% variance in the nonbarrier limb explained by the barrier limb.
Although the two limbs of the majority of control group participants did
demonstrate a clear relationship in both the temporal and spatial domain,
there were cases in which the nonbarrier limb movement behavior did not
appear related to or predicted by that of the barrier limb. These exceptions
did not seem to be related to a participant’s profession or leisure pursuits
(determined by informal exit interview) that would require a dissociation
between the two limbs. Numerous previous investigations of interlimb
coupling for neurologically intact adults have reported similar between-
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
147
subject variability (Kelso et al., 1979a, 1983; Sherwood, 1989; Martenuik et
al., 1984; Fowler et al., 1991).
The lowest group average R2 across experimental conditions occurred in the
spatial domain for the nonparetic limb-barrier/paretic limb-nonbarrier
combination. In this instance, the weaker ability of barrier limb (nonparetic)
peak vertical displacement to predict nonbarrier limb (paretic) displacement
may be a reflection of limb paresis. If not constrained to clear a vertical
barrier, the paretic limb is less likely to mirror the vertical displacement
required of the nonparetic-barrier limb. In this case, the drive for energy
conservation or movement efficiency may be stronger than the tendency for
the nervous system to couple the limbs spatially.
Mechanisms of interlimb coupling
This study provides insight into how the CNS controls a bimanual task with
asymmetric limb movement requirements. The temporal-spatial relationship
of the barrier limb followed Fitts’ Law of Movement - there was a positive
linear relationship between movement amplitude and movement time; as
amplitude increased, movement time increased. The nonbarrier limb vertical
displacement was lower than that of the barrier limb. Following Fitts’ Law, a
smaller movement amplitude in the nonbarrier limb, should be produced in a
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
shorter movement time. Surprisingly, this was not the case. Although
amplitude was smaller, nonbarrier limb MT was similar to barrier limb MT.
This implies that the CNS exerts a stronger temporal constraint than spatial
constraint for interlimb coupling particularly during bimanual asymmetric
tasks; this perspective is also consistent with that previously suggested in
motor control studies with healthy adults (Schmidt et al., 1979; Sherwood,
1989). We found a strong temporal coupling between limbs even in the
presence of mild unilateral central motor paresis (> 40% variance explained).
Similar to control participants, the nonbarrier limb of those with stroke,
whether it be nonparetic or paretic, demonstrated a greater MT/PVD slope
than the contralateral barrier limb; MT was prolonged in the nonbarrier limb
(to assimilate to that of the barrier limb), despite the lack of a proportional
increase in limb vertical displacement.
Schmidt, and co-workers (1979) proposed a model that provides one
explanation for our findings. In their bimanual motor control model, the motor
program specifies the common properties (essential variables) across each
limb (e.g., the overall duration of the movement); then the parameters that
control the specific movement characteristics of each limb are specified, such
as distance or direction (nonessential variables). Further, they suggest that
these limb-specific parameters are selected independently for each limb. Our
findings provide evidence that a common movement duration is specified for
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
149
both limbs (although subjects were given no explicit instructions to initiate
and terminate limb movements simultaneously at each target), whereas limb-
specific vertical displacements are specified relatively independently for each
limb. Additionally, the nonbarrier limb velocity profile closely mirrors that of
the barrier limb throughout the movement. This would suggest that a similar
temporal movement plan drives the performance of each limb. There is also
clear evidence for on-line velocity adjustments of both limbs that appear to
mirror one another during the deceleration phase of movement as the targets
are approached. Jackson and colleagues (2000) have suggested that
proprioceptive signals operate especially during the deceleration phase to
maintain coordination.
We provide evidence that the nonbarrier limb is constrained temporally more
so than spatially in this asymmetric bimanual task. Indeed, these results are
consistent with those of others (Marteniuk and MacKenzie, 1980; Kelso et al.,
1983; Corcos, 1984), that timing is a global constraint of the system,
controlled jointly for the two limbs. This does not mean, however, that an
interlimb spatial relationship does not exist. For the majority of participants,
control and stroke, the slope of the best fit line for interlimb vertical
displacement was significantly different from zero. The spatial topology of the
barrier limb, constrained by the physical barrier, did influence the topology of
the contralateral (nonbarrier) limb. Our results concur with those of Franz and
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
150
colleagues (1991). Adults performing a continuous drawing task (of different
shapes) simultaneously with each limb demonstrated a relative spatial
coordination between the limbs, also known as a ‘spatial magnet effect’.
Movement Planning
Reaction time reflects the time needed for information processing and
planning of a forthcoming motor response. Accordingly, following the classic
work of Henry and Rogers’ (1960) numerous authors have provided evidence
that RT is increased with increased task complexity (Christina and Rose,
1985; Siegel, 1986; Carlton et al. 1987; Van Donkelaar and Franks, 1991). In
our paradigm, negotiating a barrier to reach a target would appear to be a
more complex task ecologically than the no barrier condition and hence a
longer RT for the barrier than the nonbarrier limb would be predicted. Our
results were opposite to what would be predicted with a task complexity
hypothesis, with a longer RT for the nonbarrier than the barrier limb. These
findings, however, are consistent with those of other investigators who also
studied bimanual control in asymmetric aiming in healthy adults (Kelso et al.
1983; Goodman et al. 1983). Boessenkool et al. 1999 demonstrated in a
bimanual asymmetric task of different movement amplitudes, that the limb
with a longer movement requirement, initiated movement first (i.e., shorter
RT). Recently, in a bimanual key pressing paradigm, Obhi and Haggard
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
151
(2004) report a longer RT for the finger performing a simple key press motion
than for the analogous finger on the contralateral hand required to perform a
more complex key press action. Our study with a larger sample of healthy
adults and individuals post-stroke corroborates the findings of this and the
earlier studies.
One hypothesis for a prolonged nonbarrier limb RT pertains to the attention
demands. In an asymmetric bimanual task, the nervous system may focus
attention on the limb faced with the more complex task and commence
movement of that limb first. Once that movement has begun, the nonbarrier
limb follows, enslaved to the barrier limb with the more complex task goal.
Obhi and Haggard (2004) suggested that the CNS withholds the more
automatic response while it devotes processing resources to organizing and
executing the more complex action. Movement time of the barrier limb was
longer than for the nonbarrier limb across both groups and limbs. The shorter
RT for the barrier limb may be a compensatory measure in order for both
limbs to achieve similar total response times. A third hypothesis, suggested
by Boessenkool et al., (1999) is that the limb with the larger movement
requirement initiates first in order to minimize any discrepancy in the timing of
peak velocity between the two limbs, although the rationale for this
synchronization is not clear. In our study, movement onset and offset was
considered to be simultaneous if it occurred within 2 samples (< 34 msec) of
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
152
each other. This criterion was established as a conservative threshold given
the temporal resolution of the electromagnetic sensors and the sampling rate
of 60 Hz. The two limbs (barrier and nonbarrier) attained peak velocity
simultaneously, within our pre-established conservative measure of
synchronization.
Neuroanatomical substrates
Early studies identified the SMA as the primary brain region responsible for
interlimb coordination (Travis, 1955, Brinkman, 1984; Brinkman & Porter,
1979). Subsequent research on human and non-human primates has led to
the conclusion that control of bimanual movements is not allocated to any
one neuroanatomical structure alone but rather appears to be the result of a
distributed network involving both cortical and subcortical areas. In addition
to the SMA, other structures active during bimanual movements include the
primary motor and sensory cortices (Donchin et al., 1998; Toyokura et al.,
1999; Donchin et al., 2001), the premotor cortex (Sadato et al., 1997;
Kermadi et al., 2000), and the cingulate cortex (Stephan et al., 1999;
Kermadi et al., 2000). Two subcortical structures, the basal ganglia and the
cerebellum have also been suggested to play a role in this distributed
network (Franz et al, 1996; Ivry & Hazeltine, 1999).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
153
Because of this distributed network we did not have an a priori hypothesis
regarding location of lesion insult and performance on our asymmetrical
bilateral aiming task. The wide distribution of control of inter-limb coordination
across several cortical areas (and possibly also sub-cortical regions) is likely
to make it robust to a restricted lesion affecting one or another motor area,
whose role may be subsumed by other areas also normally involved in such
a task. Initial reorganization would have occurred in the participants in this
study, as all were at least six months post-stroke. Additionally, lesion location
in this group of individuals post-stroke was heterogeneous, not permitting an
analysis of the relationship between lesion location and performance on this
bilateral asymmetric task.
Insight into the potential neurophysiological mechanisms that underlie
bimanual coordination are provided by the examination of correlations
between local field potentials (LFP) in each hemisphere of nonhuman
primates performing unimanual, bimanual symmetrical and bimanual
asymmetrical tasks (Cardoso de Oliveira et al., 2001). Interhemispheric LFP
correlations were related to the degree of behavioral bimanual coupling.
Symmetric bimanual movements were accompanied by significantly stronger
correlation increases than asymmetric bimanual or unimanual movements,
while behaviorally, the correlations of the two arms themselves were also
highest for bimanual symmetric movements of the same amplitude. One may
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
conclude that these interhemispheric correlations contribute to interlimb
coupling and aid in the production of bimanually symmetric movements.
Although interhemispheric correlations are lower during asymmetric bimanual
movements, the fact that they are observed at al! may be a neural correlate
of the difficulty in producing completely asymmetric movements.
Supporting this hypothesis is the finding that in post-callosotomy patients
(lacking interhemispheric connections), the ability to perform strongly
asymmetric tasks is better than control individuals (Eliassen et al., 1999).
Seven of the participants in our study had lesions involving the frontal and or
parietal cortex. However, given the mild residual motor impairment in these
individuals, and with radiographic evidence, we conclude these lesions were
small, thereby leaving the majority of the interhemispheric connections intact,
and not permitting a dissociation between the two limbs.
These findings extend our previous work pertaining to bimanual coordination
following unilateral brain damage. Our earlier study used a symmetric aiming
paradigm, and focused on the adjustments of the nonparetic limb in bimanual
coordination. This study revealed that the paretic limb remains responsive to
the temporal constraint of the motor plan in the completion of a bimanual
asymmetric aiming task. Our results not only demonstrate a relationship
between the two limbs of differing motor ability following stroke, but also
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
155
provide a quantifiable description of the strength of that relationship. This
approach may be a tool that can be applied to future experiments, i.e. more
complex paradigms, to test the nervous system’s control of bimanual
coordination as well as quantify changes with practice and training.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
156
CHAPTER 5
Summary and General Discussion
Introduction
Most physical tasks can be performed an infinite number of ways. Consider
the coordination of the two upper limbs towards a common or towards
independent targets. At one extreme of the coordination continuum, the limbs
could be aimed sequentially, moving completely independently of one
another. At the other extreme of this coordination continuum, they could
move in synchrony, initiating and terminating the movement together.
Previous experiments with healthy adults reveal the synchronous
coordination scenario as the over riding choice for bimanual aiming tasks,
despite no explicit instructions to choose such coordination (Kelso, 1979a,
1979b, 1983; Martenuik et al., 1983; Sherwood, 1989; Fowler et al., 1991.
The individual experiments comprising this dissertation probed the stability of
this bimanual coordination phenomenon with our choice of participant cohort
and our manipulation of the task requirements. The results of these group
and task probes are interpreted and discussed in the next sections of this
chapter in the context of the existing behavioral, neurophysiological, and
neuroscientific literatures of bimanual coordination. The final two sections
consider the potential clinical implications for this work and areas for further
research study.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
157
H o w r o b u s t i s t h e i n t r i n s i c t e m p o r a l s y n c h r o n y c o n s t r a i n t o n b i m a n u a l
m o v e m e n t s ?
In the simplest task (Chapter 2, symmetrical aiming), control participants
initiated and terminated movements simultaneously, corroborating the results
from previous investigations (Kelso, 1979a, 1979b, 1983; Martenuik et al.,
1983; Fowler et al., 1991). Those with stroke demonstrated that, despite the
presence of asymmetric motor drive, movements were initiated and
terminated nearly simultaneously. For this to be accomplished, adjustments
were made in both movement planning and movement execution. Time
allotted for movement planning was increased for the nonparetic and
decreased for the paretic limb when they moved together compared to the
unimanual condition. This mutual adjustment permitted simultaneous
movement initiation. A nearly coincident target impact occurred as a result of
prolonged movement duration for the nonparetic limb, compared to its
unimanual performance. This increase was five times that of either limb of
the control participants, and suggestive of more than a bilateral deficit effect -
a deterioration of performance, compared to single limb use, from the
simultaneous action of two limbs (Wyke, 1969; 1971; Corcos, 1984; Dickstein
et a., 1993; Ohtsuki, 1994).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
158
We probed this tendency for simultaneity of movement initiation and
termination with the addition of disparate task requirements to the internal
asymmetry caused by asymmetric central drive (Chapter 3). For control
participants, in both external asymmetric paradigms, the limb aiming to the
near target prolonged its movement duration when paired with the
contralateral limb aiming to the far target, permitting the limbs to reach their
targets nearly simultaneously, despite the 2:1 target distance ratio. These
adjustments in movement duration were greater than those observed in the
symmetric aiming paradigm (Chapter 2), providing further evidence for the
saliency of movement synchronicity.
In contrast to the control group, stroke participants’ performance differed in
the two asymmetric aiming conditions (Chapter 3). Temporal difference in
target hit between limbs was small and comparable to that observed in the
controls for the congruent (paretic near/nonparetic far) condition. For the
incongruent condition (paretic far/nonparetic near), target-hit difference was
two and one-half times greater than in the congruent condition and for either
condition of the controls. Despite this difference in target hit in the presence
of asymmetric motor drive and asymmetric target distance, the temporal
constraint on bimanual coordination was impressive (nonparetic limb
movement duration forty percent longer compared to unimanual
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
159
performance). Thus, a temporal constraint is present in the presence of
decreased central motor drive to one limb, but does have limitations.
The design of the barrier experiment (Chapter 4) also probed the tendency
for temporal synchrony in bimanual aiming movements. The limbs were
constrained temporally more so than spatially as they performed disparate
movement trajectories. This result finds support in a bimanual coordination
model proposed by Schmidt and colleagues (1979). In their model, a single
movement duration is specified for both limbs. This common specification
implies that movement duration is an invariant characteristic. In contrast,
movement distance is specified independently for each limb. The model
predicts a stronger relationship between variables that are - invariant
characteristics applied to both limbs compared to variables specified
independently for each limb.
Motor equivalence, a hallmark of goal-directed movement, provides a
framework for the interpretation of the results from all three of our
experimental manipulations. Nearly 30 years ago, Hughes and Abbs (1976)
defined motor equivalence as, “the capacity of a motor system to achieve the
same end product with considerable variation in the individual components
that contribute to them.” As our data and others’ consistently demonstrate,
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
160
the end product in the case of bimanual aiming is the coincident timing of
target impact.
Coincident time in bimanual movements is a desirable end product for at
least two central reasons. First, it is usually important in functional tasks for
timing of the two limbs to be synchronized. In an asymmetric bimanual task
described by Wiesendanger and colleagues (1996; see also Perrig et al.,
1999) participants pulled open a drawer with one hand to grasp a small rod
from inside the drawer with the opposite hand. When an additional load was
imposed on the pulling hand, prolonging the pull phase, a delay occurred for
the grasping hand in reaching the goal. Load-induced changes in one hand
were accommodated for by an equal prolongation of the non-loaded hand.
Goal invariance was preserved even though the changed constraints had a
direct influence on only one limb.
In our work, the paretic limb with diminished central motor drive is analogous
to the “loaded” limb of the drawer- pulling task. To maintain goal invariance,
the nonparetic limb slows down so the two limbs reach the target goal
simultaneously. Similarly, in the barrier experiment, the nonbarrier limb
maintains similar movement duration to that of the barrier limb, despite
differences in limb trajectory requirements.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
161
Secondly, coincident timing simplifies control of movement. Our results
suggest that a common timing mechanism exists across limbs and
movement amplitudes and underlies this drive for motor equivalence. It
appears that the limbs are temporally constrained to act as a single unit. This
resembles a coordinative structure, a group of muscles that is constrained to
act as a single unit (Turvey et al., 1978). The CNS simplifies management of
both limbs by controlling them as a unit rather than individually. Controlling
the two limbs as a single coordinative structure reduces the large number of
degrees of freedom inherent in this multi-joint bimanual task and provides a
solution to the problem of managing them simultaneously (Bernstein, 1967).
The behavioral outcome of this simplification scheme is motor equivalence.
What are the clinical implications of these results?
Although the temporal constraint on bimanual movements was manifest
primarily through nonparetic limb adjustments, the bimanual nature of the
task influenced paretic limb behavior as well. In both symmetrical (Chapter 2)
and asymmetrical (Chapter 3) bimanual aiming, over 60% of the stroke
participants exhibited a greater paretic limb peak resultant velocity in the
bimanual than the paretic unimanual aiming conditions. Those who exhibited
greater bimanual peak velocity possessed, as a group, a ten percent lower
grip strength, (although not statistically significant, p = .33; ES = .38), a
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
162
known reliable and sensitive measure of neurological recovery (Heller et al.,
1987; Sunderland et al., 1989; Boissy et al., 1999), suggestive of a greater
impairment in this subgroup.
In an experiment with healthy participants, Walter and Swinnen (1990)
discovered a relationship between limb peak velocity and force exerted by
the contralateral limb during a bimanual coordination task. Participants
performed a unidirectional elbow flexion movement with one arm while the
contralateral arm performed a sequential elbow flexion/extension/flexion
movement under either a “no load” or “loaded” condition. A greater peak
velocity was observed in the unidirectional movement arm when the
contralateral arm was “loaded” This suggests that loading of the nonparetic
limb in a bimanual task may facilitate the paretic limb.
Despite decreased central motor drive, the individual’s post-stroke studied
here retained the ability to temporally synchronize bimanual movements; this
basic mechanism of motor control remained intact. Bimanual control is
subserved by a widely distributed neuronal system with not one brain
structure solely responsible for coordinating the two limbs. The wide
distribution of control of bimanual coordination across several cortical areas
- primary motor, supplementary motor (Kermadi et al., 1998; Donchin et al.,
1998; 1999), premotor, cingulate motor, and posterior parietal cortices
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
163
(Kermadi et al., 2000) - as well as subcortical areas - basal ganglia
(Weisendanger et al., 1996) and cerebellum (Serrien and Wiesendanger,
2000) is likely to make it robust, as our data demonstrated, to a specific
lesion affecting one or another motor area.
Bimanual tasks are ubiquitous in everyday life yet have not received much
attention in post-stroke rehabilitation. The focus has either been teaching the
individual how to use their nonparetic arm to compensate for their weakened
arm or on strengthening and functional tasks for the paretic arm alone. We
have demonstrated that following stroke, bimanual movements retain the
same underlying structure as in health. There is an entire class of actions
that have been virtually ignored and that indeed may be useful for facilitating
recovery post-stroke.
Limitations and Suggestions for Future Work
Inherent to participation in these series of experiments was a relatively mild
upper limb impairment. Task demands necessitated shoulder, elbow and
forearm movements that many post-stroke volunteers are unable to perform
and we only tested those with intact proximal upper limb proprioception.
Proprioception plays an important role in the completion of accurate
bimanual movements (Duncan 1984; Duncan et al., 1997; Jackson et al.,
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
164
2000; Franz & Ramachandran, 1998). The extension of our results to those
with a greater degree of motor or sensory impairment is not known.
Vision was available prior to and during the unimanual and bimanual aiming
movements, and is known to play an important role in the planning, control
and guidance of motor acts. Visual feedback during the early phase of a
movement may influence the execution of an ongoing movement if the
movement lasts long enough. The minimal time for visual feedback to
influence ongoing movements has been estimated at approximately 120
msec (Paillard, 1996; Desmurget & Grafton, 2000). Movements made in our
experimental paradigms were longer than this so a role for visual feedback
cannot be excluded.
The visual representation of bimanual movements plays a role in the
coordination of both continuous (Mechsner et al., 2000; Bogaerts et al.,
2003) and discrete (Weigelt et al., 2003) tasks. The experiments outlined
here were not designed to specifically determine the contribution of vision in
the coordination between the limbs. An important extension of these
experiments would include bimanual aiming in a no-vision condition and then
a comparison of those results with the current results obtained with vision
available.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
165
As in any study of motor control, one must consider if these kinematic
observations translate into a meaningful functional difference. Our
experimental design required ballistic, discrete movements. We do not know
if these apparent kinematic results will transfer into meaningful differences in
the life situation.
The intriguing trend for paretic limb kinematic benefits in bimanual aiming for
those with greater impairment is motivating for future work. A more sensitive
measure of motor impairment may be helpful in further characterizing this
interesting subset. A related extension, then, would be to determine if those
with greater impairment would indeed benefit most from a bimanual training
protocol.
This dissertation initiated the process of elucidating the critical task features
that afford changes in paretic limb kinematic behavior. Not only the bimanual
task but the specific movement requirements for the paretic limb are also
important. Systematically testing additional task features will provide more
information on the conditions that could be used to exploit this beneficial
effect for those with mild post-stroke hemiparesis.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
166
REFERENCES
Abbs, J.H., & Gracco, V.L. (1983). Sensorimotor actions in the control of
multi-movement speech gestures. Trends in Neuroscience, 6 , 391-395.
Bernstein, N. (1967). The coordination and regulation of movements. Oxford,
England: Pergamon Press.
Boessenkool, J.J., Nijhof, E.J., & Erkelens, C.J. (1999). Variability and
correlations in bimanual pointing movements. Human Movement Science,
18, 525-552.
Bogaerts, H., Buekers, M.J., Zaal, F.T., Swinnen, S.P. (2003). When visuo-
motor incongruence aids motor performance: the effect of perceiving motion
structures during transformed visual feedback on bimanual coordination.
Behavioral Brain Research, 138, 45-57.
Boissy, P., Bourbonnais, D., Gravel, D., Arsenault, B.A. (1999). Maximal grip
force in chronic stroke subjects and its relationship to global upper extremity
function. Clinical Rehabilitation, 13, 354-362.
Bohannon, R.W., & Smith, M.B. (1987). Inter rater reliability of a modified
Ashworth scale of muscle spasticity. Physical Therapy, 67, 206-207.
Bootsma, R.J., Marteniuk, R.G., MacKenzie, C.L., & Zaal, F.T.J.M. (1994).
The speed-accuracy trade-off in manual prehension: effects of movement
amplitude, object size, and object width on kinematic characteristics.
Experimental Brain Research, 98, 535-541.
Brinkman, C., Porter, R. (1979). Supplementary motor area in the monkey:
activity of neurons during performance of a learned motor task.
Journal of Neurophysiology, 42, 681-709.
Brinkman, C. (1984). Supplementary motor area of the monkey’s cerebral
cortex: short- and long-term deficits after unilateral ablation and the effects of
subsequent callosal section. Journal of Neuroscience, 4, 918-929.
Butefisch, C.M., Hummelsheim, H., Denzler, P., & Mauritz, K.H. (1995).
Repetitive training of isolated movements improves the outcome of motor
rehabilitation of the centrally paretic hand. Journal of Neurological
Sciences, 130, 59-68.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
167
Byblow, W.D., Lewis, G.N. (2004). Bimanual coordination dynamics in
poststroke hemiparetics. Journal of Motor Behavior, 36, 174-188.
Cardoso de Oliveira, S., Gribova, A., Donchin, O., Bergman, H., & Vaadia, E.
(2001). Neural interactions between motor cortical hemispheres during
bimanual and unimanual arm movements. European Journal of
Neuroscience, 14, 1881-1896.
Carlton, L.G., Carlton, M.J., & Newell, K.M. (1987). Reaction time and
response dynamics. Quarterly Journal Experimental Psychology, 39, 337-
360.
Cauraugh, J. & Kim, S. (2002). Two coupled motor recovery protocols are
better than one. Stroke, 33,1589-1594.
Chang, P. & Hammond, G.R. (1987). Mutual interactions between speech
and finger movements. Journal of Motor Behavior, 19, 265-274.
Christina, R.W., & Rose, D.J. (1985). Premotor and motor reaction time as a
function of response complexity. Research Quarterly of Exercise and Sport,
56, 306-315.
Cohn R. (1951). Interaction in bilaterally simultaneous voluntary motor
function. A.M.A. Archives of Neurology and Psychiatry, 65, 472-476.
Cole, K.J., Abbs, J.H. (1987). Kinematic and electromyographic responses to
perturbation of a rapid grasp. Journal of Neurophysiology, 57, 1498-1510.
Corcos, D.M. (1984). Two-handed movement control. Research Quarterly,
55, 117-122.
Cunningham, C.L., Phillips-Stoykov M.E., & Walter, C.B. (2002). Bilateral
facilitation of motor control in chronic hemiplegia. Acta Psychologica, 110,
321-337.
Davies, P.M. (1985). Steps to Follow. New York: Springer-Verlag.
Debaere, F., Swinnen, S.P., Beatse, E., Sunaert, S., Van Hecke, P., &
Duysens, J. (2001). Brain areas involved in interlimb coordination: a
distributed network. Neuroimage, 14, 947-958.
Desmurget, M., Grafton, S. (2000). Forward modeling allows feedback
control for fast reaching movements. Trends in Cognitive Sciences, 4, 423-
431.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
168
Dickstein, R., Hocherman, S., Amdor, G., & Pillar, T. (1993). Reaction and
movement times for patients with hemiparesis for unilateral and bilateral
elbow flexion. Physical Therapy, 73, 374-385.
Donchin, O., Gribova, A., Steinberg, H., Bergman, H., & Vaadia, E. (1998).
Primary motor cortex is involved in bimanual coordination. Nature, 395, 274-
278.
Donchin, O., Cardoso de Oliveira, S., & Vaadia, E. (1999). Who tells one
hand what the other is doing: the neurophysiology of bimanual movements.
Neuron, 23, 15-18.
Donchin, O., Gribova, A., Steinberg, O., Bergman, H., Cardoso de Oliveira,
S., & Vaadia, E. (2001). Local field potentials related to bimanual movements
in the primary and supplementary motor cortices. Experimental Brain
Research, 140, 46-55.
Duncan, J. (1984). Selective attention and the organization of visual
information. Journal of Experimental Psychology: General, 113, 501-517.
Duncan, J. Humphreys, G., & Ward, R. (1997). Competitive brain activity in
visual attention. [Review]. Current Opinion in Neurobiology, 7, 255-261.
Duncan, P.W., Goldstein, L.B., Horner, R.D., Landsman, P.B., Samsa, G.P.,
& Matchar, D.B. (1994). Similar motor recovery of upper and lower
extremities after stroke. Stroke, 25,1181-1188.
Eliassen, J.C., Banes, K., & Gazzaniga, M.S. (1999). Direction information
coordinated via the posterior third of the corpus callosum during bimanual
movements. Experimental Brain Research ,128, 573-577.
Fitts, P.M. (1954). The information capacity of the human motor system in
controlling the amplitude of movement. Journal Experimental Psychology,
47, 381-391.
Fitts, P.M., & Peterson, J.R. (1964). Information capacity of discrete motor
responses. Journal Experimental Psychology, 67,103-112.
Fowler, B., Duck, T., Mosher, M., & Mathieson, B. (1991). The coordination
of bimanual aiming movements: evidence for progressive desynchronization.
Quarterly Journal Experimental Psychology, 43A, 205-221.
Franz, E.A., Zelaznik, H.N., & McCabe, G. (1991). Spatial topological
constraints in a bimanual task. Acta Psychologica, 7 7 ,137-151.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
169
Franz, E.A., Eliassen, J.C., Ivry, R.B., & Gazzaniga, M.S. (1996).
Dissociation of spatial and temporal coupling in the bimanual movements of
callosotomy patients. Psychological Science, 7, 306-310.
Franz, E.A. (1997). Spatial coupling in the coordination of complex actions.
Quarterly Journal Experimental Psychology, 50A, 684-704.
Franz, E.A., & Ramachandran, V.S. (1998). Bimanual coupling in amputees
with phantom limbs. Nature Neuroscience, 1, 443-444.
Franz, E.A., Zelaznik, H.N., Swinnen, S., & Walter, C. (2001). Spatial
conceptual influences on the coordination of bimanual actions: when a dual
task becomes a single task. Journal Motor Behavior, 33, 103-112.
Friedman, P.J. (1992). The star cancellation test in acute stroke. Clinical
Rehabilitation, 6 , 23-30.
Freund, H.J., & Budingen, H.J. (1978). The relationship between speed and
amplitude of the fastest voluntary contractions of human arm muscles.
Experimental Brain Research, 31,1-12.
Fugl-Meyer, A.R., Jaasko, L., Leyman, I., Olsson, S., & Steglind, S. (1975).
The post-stroke hemiplegic patient. 1. A method for evaluation of physical
performance. Scandinavian Journal of Rehabilitation Medicine, 7,13-31.
Ghez, C, & Vicario, D. (1978). The control of rapid limb movement in the cat.
II. Scaling of isometric force adjustments. Experimental Brain Research, 33,
191-202.
Ghez, C. (1979). Contributions of central programs to rapid limb movement in
the cat. In H. Asanuma & V.J. Wilson (Eds.), Integration in the nervous
system (pp.305-320). New York:lgakou-Shoin.
Ghez, C. & Sainburg, R. (1995). Proprioceptive control of interjoint
coordination. Canadian Journal of Physiological Pharmacology, 73, 273-84.
Gibson, J.J. (1977). The theory of affordances. In R. Shaw & J. Bransford
(Eds.), Perceiving, acting and knowing.Toward an ecological psychology.
Hillsdale, NJ: Erlbaum.
Goodman, D.A., Kobayashi, R.B., & Kelso, J.A.S. (1983). Maintenance of
symmetry as a constraint in motor control. Canadian Journal Applied Sport
Sciences, 8 , 238.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
170
Gordon, J.G., & Ghez, C. (1987). Trajectory control in targeted force
impulses. II. Pulse height control. Experimental Brain Research, 67, 241-252.
Guiard, Y. (1987). Asymmetric division of labor in human skilled bimanual
action: The kinematic chain as a model. Journal Motor Behavior, 19, 486-
517.
Haaland, K.Y., Harrington, D., & Yeo, R. (1987). The effects of task
complexity in left and right CVA patients. Neuropsychologia, 25, 783-94.
Henry, F.M., & Rogers, D.E. (1960). Increased response latency for
complicated movements and a “memory drum” theory of neuromotor
reaction. Research Quarterly, 31, 448-458.
Heller, A., Wade, D., Wood, V., Sunderland, A., Hewer, R., Ward, E. (1987).
Arm function after stroke: measurement and recovery over the first three
months. Journal of Neurology Neurosurgery Psychiatry, 50, 714-719.
Heuer, H. (1985). Intermanual interactions during simultaneous execution
and programming of finger movements. Journal Motor Behavior, 17, 335-
354.
Hughes, O.M. & Abbs, J.H. (1976). Labial-mandibular coordination in the
production of speech: implications for the operation of motor equivalence.
Phonetica, 33,199-221.
Ivry, R.B., & Hazeltine, E. (1999). Subcortical locus of temporal coupling in
the bimanual movements of a callosotomy patient. Human Movement
Science, 18, 345-375.
Jackson, G.M., Jackson, G.M., & Kritikos, A. (1999). Attention for action:
coordinating bimanual reach-to-grasp movements. British Journal of
Psychology, 90, 247-270.
Jackson, G.M., Jackson, S.R., Husain, M., Harvey, M., Kramer, T., & Dow, L.
(2000). The coordination of bimanual prehension movements in a centrally
deafferented patient. Brain, 123, 380-393.
Jakobson, L.S., & Goodale, M.A. (1991). Factors affecting higher-order
movement planning: a kinematic analysis of human prehension.
Experimental Brain Research, 86,199-208.
Jeannerod, M. (1984). The timing of natural prehension movements. Journal
Motor Behavior, 16, 235-254.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
171
Kazennikov, O., Perrig, S., & Wiesendanger, M. (2002). Kinematics of a
coordinated goal-direct bimanual task. Behavoral Brain Research, 134, 83-
91.
Kelso, J.A.S., Southard, D.L., & Goodman, D. (1979a). On the coordination
of two-handed movements. Journal Experimental Psychology, 5, 229-238.
Kelso, J.A.S., Southard, D.L., & Goodman, D. (1979b). On the nature of
human interlimb coordination. Science, 203,1029-1031.
Kelso, J.A.S., Putnam, C.A., & Goodman, D. (1983). On the space-time
structure of human interlimb coordination. Quarterly Journal Experimental
Psychology, 35A, 347-375.
Kennerley, S.W., Diedrichsen, J., Hazeltine, E., Semjen, A., & Ivry, R.B.
(2002). Callosotomy patients exhibit temporal uncoupling during continuous
bimanual movements. Nature Neuroscience, 5, 376-381.
Kermadi, I., Liu, Y., Tempini, E., Calciati, E., & Rouiller. (1998). Neuronal
activity in the primate supplementary motor area and the primary motor
cortex in relation to spatio-temporal bimanual coordination. Somatosensory
Motor Research, 15, 287-308.
Kermadi, I., Liu, Y., & Rouiller, E.M. (2000). Do bimanual motor actions
involved the dorsal premotor, cingulated and posterior parietal cortices?
Comparison with primary and supplementary motor cortical areas.
Somatosensory Motor Research, 17, 255-271.
Klapp, S.T. (1979). Doing two things at once: the role of temporal
compatibility. Memory and Cognition, 7, 375-381.
Klapp, S.T. (1981). Doing two things at once: the role of temporal
compatibility. Memory and Cognition, 9, 398-401.
Lashley, K.S. (1930). Basic neural mechanisms in behavior. Psychological
Review, 37, 1-58.
Marteniuk, R.G., & MacKenzie, C.L. (1980). Information processing in
movement organization and execution. In R.Nickerson (Ed.), Attention and
Performance VIII (pp.29-57). Hillsdale,NJ: Erlbaum.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
172
Marteniuk, R.G., MacKenzie, C.L., & Baba, D.M. (1984). Bimanual
movement control: information processing and interaction effects. Quarterly
Journal of Experimental Psychology, 36A, 335-365.
Marteniuk, R.G., MacKenzie, C.L., Jeannerod, M., Athenes, S., & Dugas, C.
(1987). Constraints on human arm movement trajectories. Canadian Journal
Psychology, 41, 365-387.
Mathiowetz, V., Kashman, N., Volland, G., Weber, K., Dowe, M., & Rogers,
S. (1985). Grip and pinch strength: normative data for adults. Archives
Physical Medicine Rehabilitation, 66, 69-74.
Mechsner, F., Kerzel, D., Knoblich, G., Prinz, W. (2001). Perceptual basis of
bimanual coordination. Nature, 414, 69-73.
Morris, D.M., Crago, J.E., DeLuca, S.C., Pidikiti, R.D., & Taub, E. (1997).
Constraint-induced movement therapy for motor recovery after stroke.
NeuroRehabilitation, 9, 29-43.
Mudie, H.M. & Matyas, T.A. (2000). Can simultaneous bilateral movement
involve the undamaged hemisphere in reconstruction of neural networks
damage by stroke? Disability and Rehabilitation, 22, 23-37.
Muzii, R.A., Warburg, C.L., & Gentile, A.M. (1984). Coordination of the upper
and lower extremities. Human Movement Science, 3, 337-354.
Obhi, S.S., & Haggard, P. (2004). The relative effects of external spatial and
motoric factors on the bimanual coordination of discrete movements.
Experimental Brain Research, 154, 399-402.
Ohtsuki, T. (1994). Changes in strength, speed, and reaction time induced by
simultaneous bilateral muscular activity. In S.P. Swinnen, H. Heuer, J.
Massion, P. Casaer (Eds.), Interlimb Coordination: Neural, Dynamical and
Cognitive Constraints (pp. 179-207) San Diego: Academic.
Oldfield, R.C. (1971). The assessment and analysis of handedness: the
Edinburgh inventory. Neuropsychologia, 9, 97-113.
Paillard, J. (1996). Fast and slow feedback loops for the visual correction of
spatial errors in a pointing task: a re-appraisal. Canadian Journal of
Physiological Pharmacology, 74: 401-417.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
173
Peper, C.E. & Carson, R.G. (1999). Bimanual coordination between isometric
contractions and rhythmic movements: an asymmetric coupling.
Experimental Brain Research, 129, 417-432.
Perrig, S., Kazennikov, O., & Wiesendanger, M. (1999). Time structure of a
goal-directed bimanual skill and its dependence on task constraints.
Behavioral Brain Research, 103, 95-104.
Peters, M. (1977). Simultaneous performance of two motor activities: the
factor of timing. Neuropsychologia, 15, 461-465.
Peters, M. (1985). Constraints in the performance of bimanual tasks and their
expression in unskilled and skilled subjects. Quarterly Journal Experimental
Psychology, 37A, 171-196.
Preilowski, B. (1972). Possible contribution of the anterior forebrain
commissures to bilateral motor coordination. Neuropsychologia, 10, 267-277.
Preilowski, B. (1975). Bilateral motor interaction: perceptual-motor
performance of partial and complete “split-brain” patients. In: K.S. Zulich, O.
Creutzfeldt, & G.C. Galbraith (Eds.), Cerebral localization (pp. 115-132). New
York: Springer.
Rice, M.S., & Newell, K.M. (2001). Interlimb coupling and left hemiplegia
because of right cerebral vascular accident. Occupational Therapy Journal of
Research, 21, 12-28.
Rice, M.S., & Newell, K.M. (2004). Upper-extremity interlimb coupling in
persons with left hemiplegia due to stroke. Archives of Physical Medicine
and Rehabilitation, 85, 629-634.
Rose, D.K., & Winstein, C.J. The coordination of bimanual symmetrical rapid
aiming movements following stroke. Clinical Rehabilitation (in press).
Sadato, N. (1997). Role of the supplementary motor area and the right
premotor cortex in the coordination of bimanual finger movements. Journal of
Neuroscience, 17, 9667-9674.
Schmidt, R.A., Zelaznik, H., Hawkins, B., Frank, J.S., & Quinn, Jr., J.T.
(1979). Motor-Output variability: a theory for the accuracy of rapid motor acts.
Psychological Review, 86, 415-451.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
174
Serrien, D., Wiesendanger, M. (2000).Temporal control of a bimanual task in
patients with cerebellar dysfunction. Neuropsychologia, 38, 558-65.
Servos, P., Goodale, M.A., & Jakobson, L.S. (1992). The role of binocular
vision in prehension: a kinematic analysis. Vision Research, 32,1513-1521.
Sherwood, D.E. (1989). The coordination of simultaneous actions. In S.A.
Wallace (Ed) Perspectives on the Coordination of Movement (pp. 303-327).
North-Holland: Elsevier.
Sherwood, D.E. (1991). Distance and location assimilation effects in rapid
bimanual movement. Research Quarterly Exercise and Sport, 62, 302-308.
Siegel, D. (1986). Movement duration, fractionated reaction time, and
response programming. Research Quarterly Exercise and Sport, 57,128-
131.
Smith, A., McFarland, D.H., & Weber, C.M. (1986). Interactions between
speech and finger movements: an exploration of the dynamic pattern
perspective. Journal of Speech and Hearing Research, 29, 471-480.
Soechting, J.F., & Lacquaniti, F. (1981). Invariant characteristics of pointing
movement in man. Journal of Neuroscience, 1, 710-720.
Steenbergen, B., Hulstijn, W., de Vries, A., & Berger, M. (1996). Bimanual
movement coordination in spastic hemiparesis. Experimental Brain
Research, 110, 91-98.
Steglich, C., Heuer, H., Spijkers.W., & Kleinsorge, T. (1999). Bimanual
coupling during the specification of isometric forces. Experimental Brain
Research. 129, 302-316.
Stephan, K.M., Binkofski, F., Halsband, U., Dohle, C., Wunderlich, G.,
Schnitzler, A., et al. (1999). The role of ventral medial wall motor areas in
bimanual co-ordination: a combined lesion and activation study. Brain, 122,
351-368.
Sunderland, A., Tinson, D., Bradley, L, Hewer, R. (1989). Arm function after
stroke. An evaluation of grip strength as a measure of recovery and a
prognostic indicator. Journal of Neurology Neurosurgery and Psychiatry, 52,
1267-1272.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
175
Sunderland, A., Fletcher, D., Bradley, L , Tinson, D., Hewer, R.L., & Wade,
D.T. (1994). Enhanced physical therapy for arm function after stroke: a one-
year follow-up study. Journal of Neurology Neurosurgery and Psychiatry, 57,
856-858.
Swinnen, S.P. (2002). Intermanual coordination: from behavioural principles
to neural-network interactions. Nature Review Neuroscience, 3, 350-361.
Tanji, J., Okano, K., Sato, K.C. (1988). Neuronal activity in cortical motor
areas related to ipsilateral, contralateral and bimanual digit movements of the
monkey. Journal of Neurophysiology, 60, 325-343.
Taub, E., Uswatte, G., & Pidikiti, R. (1999). Constraint-induced movement
therapy: a new family of techniques with broad application to physical
rehabilitation - a clinical review. Journal Rehabilitation Research and
Development, 36, 237-251.
Toyokura, M., Muro, I., Komiya, T., & Obara, M. (1999). Relation of bimanual
coordination to activation in the sensorimotor cortex and supplementary
motor area: analysis using functional magnetic resonance imaging. Brain
Research Bulletin, 48, 211-217.
Travis, A.M. (1955). Neurological deficiencies following supplementary motor
area lesions in Macaca mulatta. Brain, 78,174-201.
Tuller, B. & Kelso, J.A.S. (1989). Environmentally-specified patterns of
movement coordination in normal and split-brain subjects. Experimental
Brain Research, 75, 306-316.
Turvey, M.T., Shaw, R.E., & Mace, W. (1978). Issues in the theory of actions:
Degrees of freedom, coordinative structures, and coalitions. In J. Requin
(Ed.), Attention and Performance VII (pp. 557-595). Hillsdale, NJ: Erlbaum.
Van Donkelaar, P., Franks, I.M. (1991). The effects of changing movement
velocity and complexity on response preparation: evidence from latency,
kinematic, and EMG measures. Experimental Brain Research, 83, 618-632.
Walter, C.B., & Swinnen, S.P. (1990a). Asymmetric interlimb interference
during performance of a dynamic bimanual task. Brain and Cognition, 14,
185-200.
Weigelt, C. & Cardoso de Oliveira, S. (2003). Visuomotor transformations
affect bimanual coupling. Experimental Brain Research, 48, 439-50.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
176
Wiesendanger, M., Wicki, U..Rouiller, E. (1994). Are there unifying structures
in the brain responsible for interlimb coordination? In S.P. Swinnen, H.
Heuer, J. Massion, P. Casaer (Eds.), Interlimb Coordination: Neural,
Dynamical and Cognitive Constraints (pp. 179-207) San Diego: Academic.
Weisendanger, M., Kazennikov, O., Perrig, S., & Kaluzny, P. (1996). In A.M.
Wing, P. Haggard, R. Flanagan (Eds.), Hand and Brain (pp. 283-300).
Oxford: Oxford University Press.
Whitall, J., Waller, S.M., Silver, K.H.C., & Macko, R.F. (2000). Repetitive
bilateral arm training with rhythmic auditory cueing improves motor function
in chronic hemiparetic stroke. Stroke, 31, 2390-2395.
Woodworth, R.S. (1899). The accuracy of voluntary movements.
Psychological Review, 3, 1-114.
Woodworth, R.S. (1903). Le mouvement. Paris: Doin.
Wyke, M. (1969). Influence of direction on the rapidity of bilateral arm
movements. Neuropsychologia, 7 , 125-134.
Wyke, M. (1971). The effects of brain lesions on the performance of bilateral
arm movements. Neuropsychologia, 9, 33-42.
Yamanishi, J., Kawato, M., & Suzuki, R. (1980). Two coupled oscillators as a
model for the coordinated finger tapping by both hands. Biological
Cybernetics, 37, 219-225.
Zaidel, D. & Sperry, R.W. (1977). Some long-term motor effects of cerebral
commissurotomy in man, Neuropsychologia, 15,193-204.
Zelaznik, H.N., Schmidt, R.A., & Gielen, S. (1986). Kinematic properties of
rapid aimed hand movements. Journal of Motor Behavior, 18, 353-372.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Functional brain correlates for premovement planning and compensatory adjustments in rapid aimed movement

PDF

Human and environmental factors contributing to slip events during walking

PDF

Data-driven derivation of skills for autonomous humanoid agents

PDF

Hemispheric specialization of reach-to-grasp actions

PDF

Factors contributing to patellofemoral joint stress: a comparison of persons with and without patellofemoral pain

PDF

Biomechanical and neuromuscular aspects of non-contact ACL injuries: The influence of gender, experience and training

PDF

Contextual interference in motor skill learning: an investigation of the practice schedule effect using transcranial magnetic stimulation (TMS)

PDF

At what age are gait characteristics mature? Evaluation of gait kinematics, kinetics, and intersegmental dynamics in 7 year-old children

PDF

Planning of unconstrained reach actions after unilateral sensorimotor stroke

PDF

Deficits and rehabilitation of upper-extremity multi-joint movements in individuals with chronic stroke

PDF

Relationship between brain structure and motor behavior in chronic stroke survivors

PDF

A comparative study of environmental factors associated with multiple sclerosis in disease-discordant twin pairs

PDF

Development and evaluation of standardized stroke outcome measures in a population of stroke patients in rural China

PDF

An intervention and program evaluation to determine the effectiveness of public health reforms on primary prevention practices by chiropractic interns

PDF

Design of a stealth liposome delivery system for a novel glycinamide ribonucleotide formyltransferase inhibitor

PDF

Relationship between interhemispheric inhibition and bimanual coordination in musicians

PDF

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

PDF

Enabling clinically based knowledge discovery in pharmacy claims data: An application in bioinformatics

PDF

Modeling human reaching and grasping: cortex, rehabilitation and lateralization

PDF

Learning reaching skills in non-disabled and post-stroke individuals

Asset Metadata

Creator Rose, Dorian K. (author)

Core Title Bimanual coordination of goal-directed multi-joint rapid aiming movements following stroke

School Graduate School

Degree Doctor of Philosophy

Degree Program Biokinesiology

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

Tag health sciences, rehabilitation and therapy,OAI-PMH Harvest

Language English

Contributor Digitized by ProQuest (provenance)

Advisor Winstein, Carolee (committee chair), Azen, Stanley (committee member), Baker, Lucinda (committee member), Gordon, James (committee member), Schaal, Stefan (committee member), Schreiber, Steven (committee member)

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c16-423735

Unique identifier UC11335773

Identifier 3155469.pdf (filename),usctheses-c16-423735 (legacy record id)

Legacy Identifier 3155469.pdf

Dmrecord 423735

Document Type Dissertation

Rights Rose, Dorian K.

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 au...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus, Los Angeles, California 90089, USA