University of Southern California Dissertations and Theses

Infant exploratory learning: influence on leg coordination

(USC Thesis Other)

Infant exploratory learning: influence on leg coordination

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content Running head: INFANT LEG COORDINATION i

INFANT EXPLORATORY LEARNING:

INFLUENCE ON LEG COORDINATION

by

Barbara A. Sargent

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

May 2013

Copyright 2013 Barbara A. Sargent
(continued)
INFANT LEG COORDINATION ii
DEDICATION

To Jess, Troy, and Luke, for their unconditional love and support…
and to all the children and families who have inspired this research

INFANT LEG COORDINATION iii
ACKNOWLEDGEMENTS
This dissertation is the result of the support, guidance, and assistance of many
individuals. I would like to thank my advisor, Dr. Linda Fetters. She has provided me with the
training, guidance, and resources necessary to investigate my questions about infant action. In
addition, she has mentored me in the skills necessary to continue developing as a researcher,
including systematic thought, professional writing, and pursuing expert guidance. She has
accomplished this with a calm, confident presence that I sincerely appreciate.
I would also like to thank the members of my dissertation committee and other
individuals who have contributed significantly to the research. Dr. Nina Bradley contributed to
my understanding of the neuroscience which implements infant action. Dr. Nicolas
Schweighofer contributed to the design of the learning paradigm and encouraged me to think
about infant learning from a computational approach, which revealed new avenues for
investigating infant action. He also assisted in providing technical resources for the development
of the infant-activated mobile system by recruiting Hyeshin Park, Jeremy Welch, Yukikazu
Hidaka, and Younggeun Choi. Without their technical expertise this research would not have
been possible. Dr. John Scholz was instrumental in elevating the methods in the lab so that
infant movement could be studied using a three-dimensional (3D) kinematic and kinetic
approach. In addition, the depth of his knowledge of biomechanics allowed analysis of data and
interpretation of results which would not have been possible without his expertise. Dr. Hendrik
Reimann from the Instutut für Neuroinformatik in Bochum, Germany, developed the torque
computational algorithms for the 3D kinetic analysis. Dr. Masayoshi Kubo from the Niigata
University of Health and Welfare in Niigata, Japan, contributed to the development of
INFANT LEG COORDINATION iv
computational algorithms to analyze the 3D kinematic and kinetic data. Dr. Stanley Azen
provided critical statistical support. Chantelle Takata illustrated Figures 3.1 and 5.2.
I would also like to thank the members of the Division of Biokinesiology and Physical
Therapy. The members of the Motor Control Development Lab, Motor Behavior and
Neurorehabilitation Lab, Neuroplasticity and Imaging Lab, and Computational Neuro-
Rehabilitation and Learning Lab at the University of Southern California (USC) have welcomed
me to present at their lab meetings and provided the critical feedback necessary to improve this
research. Shaveonte Graham, Ya-Yun Lee, Dr. Hsiang-Han Huang, and Dr. Sue Duff assisted
with data collection. Matt Sandusky, David Donaldson, and Chad Louie provided technological
support which included moving the location of the lab and problem solving video system issues.
I would like to acknowledge the support of my family throughout my graduate study.
My husband, Jess, has always supported my quest for lifelong learning. My teenage children,
Troy and Luke, encouraged me throughout the process and also assisted me with managing the
computer during data collection and editing videos. My father, Robert Vogtmann, assisted
frequently with data collection at the initial stages of this research. In addition, my extended
family of parents, brothers, and sister-in-laws all provided critical supported at different stages of
the process.
This research would not have been possible without generous financial support from the
Division of Biokinesiology and Physical Therapy at USC; Promotion of Doctoral Scholarship
(PODS) I and II funding from the Foundation for Physical Therapy; and an Adopt-A-Doc
Scholarship from the Section of Education of the American Physical Therapy Association
(APTA).
INFANT LEG COORDINATION v
Most importantly, I would like to thank the parents who brought their children to
participate in the research studies during the first three months of their child’s life. This is a time
of significant transition in the life of a family, and these parents demonstrated extreme generosity
to participate in this research at such a pivotal time. Thank you.
INFANT LEG COORDINATION vi
TABLE OF CONTENTS
DEDICATION ii
ACKNOWLEDGEMENTS iii
LIST OF TABLES viii
LIST OF FIGURES ix
ABSTRACT xi
CHAPTER I: OVERVIEW 1
Background and Significance 1
Specific Aims, Hypotheses, and Significance of Each Study 3

CHAPTER II: THE QUANTITATIVE STUDY OF INFANT LEG MOVEMENTS FROM
SPONTANEOUS KICKING TO TASK-SPECIFIC LEG ACTION 8
Abstract 8
Introduction 8
Method 10
Results 11
Discussion 27

CHAPTER III: CHANGES IN INFANT LEG COORDINATION: INSIGHTS FROM A 3D
KINEMATIC AND KINETIC APPROACH 30
Abstract 30
Introduction 30
Method 35
Results 45
Discussion 59

CHAPTER IV: CHANGES IN INFANT LEG COORDINATION: INFLUENCE OF
PREMATURITY 63
Abstract 63
Introduction 64
Method 70
Results 79
Discussion 99

CHAPTER V: INFANT EXPLORATORY LEARNING: INFLUENCE ON LEG
COORDINATION 102
Abstract 102
Introduction 103
Method 107
Results 119
INFANT LEG COORDINATION vii
Discussion 132

CHAPTER VI: SUMMARY AND GENERAL DISCUSSION 139

REFERENCES 148
INFANT LEG COORDINATION viii
LIST OF TABLES

Table 3.1 Full-Term Infants: Participant Characteristics 36

Table 3.2 Full-Term Infants: Joint Correlations 48

Table 3.3 Full-Term Infants: Relative Phase 50

Table 3.4 Full-Term Infants: Normalized Torques 52

Table 4.1 Preterm Infants: Participant Characteristics 71

Table 4.2 Preterm Infants Compared to Full-Term Infants: Joint Correlations 81

Table 4.3 Preterm Infants Compared to Full-Term Infants: Relative Phase 84

Table 4.4 Preterm Infants: Normalized Torques 90

Table 4.5 Preterm Infants Compared to Full-Term Infants: Relation of Kick Coordination
and Percent Interval 97

Table 4.6 Preterm Infants Compared to Full-Term Infants: Relation of Kick Coordination
and Torque Impulse 98

Table 5.1 Full-Term and Preterm Infants: Participant Characteristics 109

Table 5.2 Number of Kicks Analyzed During each Condition 123

Table 5.3 Relative Phase of Hip-Knee Pair by Interval 128

Table 5.4 Relative Phase of Hip-Knee Pair by Condition 129

INFANT LEG COORDINATION ix
LIST OF FIGURES

Figure 3.1 Experimental Set-Up 37

Figure 3.2 Five Time Points in each Kick 41

Figure 3.3 Full-Term Infants: Joint Correlations 47

Figure 3.4 Full-Term Infants: Mean Partitioned Torque Percent Interval Contribution to Net
Knee Torque 54

Figure 3.5 Full-Term Infants: Mean Partitioned Torque Impulse Contribution to Net Knee
Torque 54

Figure 3.6 Full-Term Infants: Relation between Kick Coordination and Percent Interval
Across Ages 56

Figure 3.7 Full-Term Infants: Relation between Kick Coordination and Torque Impulse
Across Ages 57

Figure 4.1 Preterm Infants: Joint Correlations 80

Figure 4.2 Preterm Infants: Mean Partitioned Torque Percent Interval Contribution to Net
Knee Torque 92

Figure 4.3 Preterm Infants: Mean Partitioned Torque Impulse Contribution to Net Knee
Torque 92

Figure 4.4 Preterm Infants: Relation between Kick Coordination and Percent Interval Across
Ages 94

Figure 4.5 Preterm Infants: Relation between Kick Coordination and Torque Impulse Across
Ages 95

Figure 5.1 Infant Mobile Task Set-Up 106

Figure 5.2 Infant Mobile Experimental Set-Up 110

Figure 5.3 Infant Mobile Experimental Protocol 111

Figure 5.4 Raw Data from One Representative Participant 114

Figure 5.5 Full-Term Infants: Mean Percent Reinforced Leg Action by Interval 119

Figure 5.6 Full-Term Infants: Mean Percent Reinforced Leg Action by Condition 120

INFANT LEG COORDINATION x
Figure 5.7 Preterm Infants at 3 months: Mean Percent Reinforced Leg Action by Interval
121

Figure 5.8 Preterm Infants at 4 months: Mean Percent Reinforced Leg Action by Interval
122

Figure 5.9 Learners and Non-Learners: Fisher Z Correlation Coefficients of Hip-Knee Pair by
Interval 124

Figure 5.10 Learners and Non-Learners: Fisher Z Correlation Coefficients of Hip-Knee Pair by
Condition 126

Figure 5.11 Learners and Non-Learners: Relation of Change in Reinforced Leg Action to
Change in Fisher Z Correlation Coefficients 131
INFANT LEG COORDINATION xi
ABSTRACT
Infants born preterm with very low birth weight are at increased risk for developing
spastic cerebral palsy, which is characterized by walking limitations due to a reduced ability to
move the joints of the leg in an out-of-phase pattern; e.g., flexing the hip while extending the
knee. In typical development, the spontaneous kicks of newborn infants are dominated by an in-
phase coordination pattern of synchronous flexion or extension of the hip, knee, and ankle joints.
Within the first months of life, infants demonstrate a greater variety of intralimb joint
coordination patterns which is thought to support the progress to more functional coordination
patterns of the lower extremities necessary for independent mobility. PT infants also
demonstrate this developmental change from in-phase to a greater variety of joint coordination
patterns, however, some PT infants exhibit prolonged in-phase joint coordination which places
them at increased risk of developing cerebral palsy. The purpose of this dissertation is to gain a
better understanding of the contribution of torque changes to the early changes in leg joint
coordination of typically-developing full-term infants, clarify the differences between leg joint
coordinations and torques of preterm and full-term infants, and determine whether preterm
infants have the potential for a greater amount of out-of-phase leg joint coordination when
interacting with an infant-activated mobile that reinforces out-of-phase leg joint coordination.
Three studies were conducted to answer these questions. The first study investigated the
contribution of torque changes to the early changes in leg joint coordination of typically-
developing full-term infants. We analyzed kicking actions within 10 full-term infants and
between 6 and 15-weeks of age using a three-dimensional kinematics and kinetics approach. We
found that although 73% of joint angle pairs demonstrated a change from an in-phase intralimb
coordination pattern at 6-weeks to an increased variety of intralimb joint coordination patterns at
INFANT LEG COORDINATION xii
15-weeks, there was not an obvious developmental change in net joint torques or intersegmental
dynamics. Further analysis supported that a greater variety of hip-knee joint coordinations from
6 to 15-weeks of age was associated with a decreased influence of knee muscle torques which
allowed passive knee gravitational and motion dependent torques to have a greater influence on
the coordination of the kick, contributing to a greater variety of hip-knee joint coordinations.
The second study investigated the contribution of torque changes to the early changes in
leg joint coordination of preterm infants. From 6 to 15-weeks, preterm infants demonstrated less
in-phase coordination in 30% of joint angle pairs, versus 73% of joint angle pairs in full-term
infants. At 6-weeks preterm infants as compared to full-term infants demonstrated less in-phase
coordination in 4 of 15 joint angle pairs, but this difference resolved at 15-weeks. Similar to
full-term infants, there was not an obvious developmental change in net joint torques or
intersegmental dynamics from 6 to 15-weeks. Although PT infants exhibited a greater variety of
hip-knee coordinations from 6 to 15-weeks of age, unlike full-term infants, no differences were
noted in the intersegmental torques. However, the PT infants demonstrated a smaller change in
hip-knee coordination from 6 to 15-weeks which may have been insufficient to document the
relation between intersegmental dynamics and joint coordination.
The final study investigated whether preterm infants could generate a greater amount of
out-of-phase leg joint coordination when participating in an innovative learning task which
utilized an infant-activated mobile. Fourteen full-term infants and 6 preterm infants participated
at 3 - 4 months corrected age. Each infant participated in 2 sessions of mobile reinforcement on
consecutive days. During each session, the infant was positioned supine under an overhead infant
mobile. Day 1 consisted of a 2-minute non-reinforcement condition (spontaneous kicking)
followed by a 6-minute reinforcement condition (the infant mobile rotated and played music
INFANT LEG COORDINATION xiii
when the infant moved either foot vertically across a virtual threshold). Day 2 consisted of a 2-
minute non-reinforcement condition, 6-minute reinforcement condition, and 2-minute non-
reinforcement condition. The full-term group, but not the preterm group, increased the
percentage of mobile activation to meet performance criteria the 2
nd
day. Neither group met
learning criteria. However, both the full-term and preterm groups included infants that learned
the contingency between their leg action and mobile activation. Infants were separated into
infants that learned the contingency and infants that did not learn the contingency. Infants who
learned the contingency demonstrated a greater amount of out-of-phase hip-knee joint
coordination during the reinforcement condition on the second day as compared to spontaneous
kicking during the initial non-reinforcement condition on the first day. This coordination change
was not demonstrated by the group of infants that did not learn the contingency. These results
indicate that some full-term and preterm infants can demonstrate a greater amount of out-of-
phase hip-knee joint coordination when participating in a task in which their leg actions are
reinforced with mobile activation.
The dissertation concludes with a discussion of each research study highlighting novel
findings, implications for clinical practice, and opportunities for future research. The results
from this dissertation will be an essential step toward developing more relevant and successful
screening, intervention, and prevention programs for preterm infants at risk for impairments in
motor coordination.
INFANT LEG COORDINATION 1
CHAPTER I

OVERVIEW

Background and Significance
Approximately 500,000 preterm (PT) births occur in the United States each year; this is
twelve percent of total births (Hamilton, Martin, & Ventura, 2011). PT infants are at high risk
for white matter damage (WMD), which may increase the risk of developing spastic cerebral
palsy (CP), a severe dysfunction of motor coordination (Himpens, Van den Broeck, Oostra,
Calders, & Vanhaesebrouck, 2008; Oskoui, Coutinho, Dykeman, Jetté, & Pringsheim, 2013;
Rosenbaum et al., 2007). Even PT infants who do not sustain WMD are at high risk for
impairments in motor coordination, which increase in severity with decreasing gestational age
and birth weight (Aarnoudse-Moens, Weisglas-Kuperus, van Goudoever, & Oosterlaan, 2009; de
Kieviet, Piek, Aarnoudse-Moens, & Oosterlaan, 2009; Delobel-Ayoub et al., 2009; Larroque et
al., 2008; Pearsall-Jones, Piek, & Levy, 2010).
Early therapeutic intervention is shown to improve motor skill acquisition of PT infants
at risk for impairments in motor coordination (Heathcock & Galloway, 2009; Heathcock, Lobo,
& Galloway, 2008). The plasticity of the brain is considerable during the first months of life
which necessitates that intervention be initiated early, to optimize the neural circuitry which
implements motor coordination (de Graff-Peters & Hadders-Algra, 2006; Lobo, Harbourne,
Dusing, & McCoy, 2013; Ulrich, 2010). The early identification of motor coordination
impairments with the potential for early intervention may be critical to optimize the motor skill
acquisition of PT infants during the first months of life.
INFANT LEG COORDINATION 2
Early impairments in the motor coordination of PT infants may be apparent in the joint
coordination of infant spontaneous kicking. The spontaneous kicks of newborn infants are
dominated by in-phase intralimb coordinations of synchronous flexion or extension of the hip,
knee, and ankle joints (Thelen & Fisher, 1983b). Within the first months of life, infants
demonstrate a greater variety of intralimb coordinations (Fetters, Chen, Jonsdottir, & Tronick,
2004; Fetters, Sapir, Chen, Kubo, & Tronick, 2010; Heriza, 1988a; Jeng, Chen, & Yau, 2002;
Thelen & Fisher, 1983b; Vaal, van Soest, Hopkins, Sie, & van der Knapp, 2000), which is
thought to support the progress to more functional coordination of the lower extremities
necessary for optimal motor skill acquisition (Fetters et al., 2010; Thelen & Fisher, 1983b). PT
infants also demonstrate this developmental change in joint coordination, however, some PT
infants demonstrate prolonged in-phase joint coordination which places them at increased risk of
developing CP. The purpose of this dissertation is to gain a better understanding of the
contribution of torque changes to the early changes in leg joint coordination of typically-
developing full-term infants, clarify differences between the leg joint coordination and torques of
preterm and full-term infants, and determine whether preterm infants have the potential for a
greater amount of out-of-phase leg joint coordination when interacting with an infant-activated
mobile that reinforces out-of-phase leg joint coordination.
The dissertation is organized into the following chapters. Chapter One, The qualitative
study of infant leg movements: from spontaneous kicking to task-specific leg action, is a
literature review which summarizes the processes and mechanisms which contribute to
developmental changes in infant leg movements over the first four months of life, the early
emergence of task-specific leg action, and differences in leg action between typically-developing
infants and infants at risk for disability. Chapter Two, Changes in infant leg coordination:
INFANT LEG COORDINATION 3
insights from a three-dimensional (3D) kinematic and kinetics approach, is a longitudinal
research study which investigates the contribution of torque changes to the early changes in leg
joint coordination of typically developing FT infants from 6 to 15-weeks of age. Chapter Three,
Changes in infant leg coordination: influence of prematurity, is a longitudinal research study
which investigates differences in joint coordination and torques between PT and FT infants from
6 to 15-weeks of age. Chapter Four, Infant exploratory learning: influence on leg coordination,
is a research study which investigates the ability of 3 to 4-month old FT and PT infants to
demonstrate a greater amount of out-of-phase joint coordination when participating in an
innovative learning task using an infant-activated mobile.
The results from this dissertation will provide critical information on the contribution of
torque changes to the early changes in leg joint coordination of typically-developing full-term
infants, clarify differences between the leg joint coordination and torques of preterm and full-
term infants, and determine whether PT infants have the potential for a greater amount of out-of-
phase leg joint coordination when interacting with an infant-activated mobile that reinforces out-
of-phase leg joint coordination. This information will be an essential step toward developing
more relevant and successful screening, intervention, and prevention programs for PT infants at
risk for impairments in motor coordination.
Specific Aims, Hypotheses and Significance of Each Study
Changes in Infant Leg Coordination: Insights from a 3D Kinematic and Kinetics Approach
Specific aims and hypotheses.
1) To characterize early changes in joint coordination between all joint angle pairs in FT
infants from 6 to 15-weeks of age.
INFANT LEG COORDINATION 4
a) We hypothesize that all joint angle pairs will change from a predominately in-
phase coordination pattern at 6-weeks to an increased variety of joint coordination
patterns at 15-weeks.
2) To identify changes in intersegmental dynamics of infant spontaneous kicks in FT
infants from 6 to 15-weeks of age.
a) We hypothesize that hip NET joint torque will increase and knee and ankle
NET torques will decrease from 6 to 15-weeks of age.
3) To describe the relation between changes in joint coordination and intersegmental
dynamics in FT infants from 6 to 15-weeks through the analysis of the coordination
of the hip and knee into flexion and extension.
a) We hypothesize that a greater variety of hip-knee joint coordinations will be
associated with either:
(1) an increased influence of knee muscle (MUS) torque to dampen the passive
knee gravitational (GRA) and motion-dependent (MDT) torques which are
associated with more in-phase hip-knee joint coordination, or
(2) a decreased influence of knee MUS torque which allow passive knee GRA
and MDT torques to have a greater influence on the coordination of the kick,
contributing to a greater variety of hip-knee joint coordinations.
Significance. This study is a unique contribution to the literature since it is the first time
that infant leg actions have been investigated using a 3D kinematic and kinetic approach,
as well as the first time that these two analyses have been combined to investigate
coordination changes in young infants.

INFANT LEG COORDINATION 5
Changes in Infant Leg Coordination: Influence of Prematurity
Specific aims and hypotheses.
1) To characterize early changes in joint coordination between all joint angle pairs in
PT infants from 6 to 15-weeks of age and compare with the FT group from the
previous study.
a) We hypothesize that, similar to FT infants, the majority of joint angle pairs
will exhibit a greater variety of joint coordinations between 6 and 15-weeks of
age, however, at both 6-weeks and 15-weeks PT infants, as compared to FT
infants, will exhibit a greater variety of joint coordinations between joint angle
pairs.
2) To identify changes in intersegmental dynamics of infant spontaneous kicks in PT
infants from 6 to 15-weeks of age and compare with the FT group.
a) We hypothesize that between 6 and 15-weeks of age PT infants will
demonstrate similar intersegmental dynamics as the FT infants; net joint torque
and partitioned joint torque data normalized to the mass of the leg of each
individual participant will remain relatively consistent from 6 to 15-weeks of age.
3) To describe the relation between changes in joint coordination and intersegmental
dynamics in PT infants from 6 to 15-weeks through the analysis of the coordination
of the hip and knee into flexion and extension.
a) We hypothesize that, similar to FT infants, a greater variety of joint
coordinations will be associated with a decreased influence of knee MUS forces
which allow passive knee GRA and MDT to have a greater influence on the
INFANT LEG COORDINATION 6
coordination of the kick, contributing to a greater variety of hip-knee joint
coordinations.
Significance. This study is a unique contribution to the literature since it is the first time
that a 3D kinematic and kinetic approach has been used to investigate joint coordination
and intersegmental dynamics in a population of infants with known differences in joint
coordination from typically-developing FT infants. An understanding of the similarities
and differences in the changes in joint coordination and intersegmental dynamics
between FT and PT infants may further clarify the processes underlying the early
coordination changes of young infants.
Infant Exploratory Learning: Influence on Leg Coordination
Specific aims and hypotheses.
1) To determine the ability of 3-month old FT and PT infants to learn, through discovery,
the contingency between leg action and mobile activation using a virtual threshold.
(a) We hypothesize that
(1) the FT group will demonstrate performance of the contingency on the
first day and both memory and learning of the contingency on the second
day,
(2) the PT group will not demonstrate performance of the contingency on
the first day or learning and memory on the second day, but will
demonstrate performance on the second day.
2) To determine whether FT and PT infants who learn the contingency exhibit more out-
of-phase hip-knee joint coordination patterns when leg actions are reinforced with
mobile activation.
INFANT LEG COORDINATION 7
(a) Since both FT and PT infant groups are expected to include infants that
learned the task, infants will be separated into those that learned and did not learn
the task based on an individualized learning criteria. We hypothesize that:
(3) the FT and PT infants who learn the contingency, as compared to
infants who did not learn the contingency, will exhibit more out-of-phase
hip-knee joint coordination patterns when leg actions are reinforced with
mobile activation.
Significance. This study is a unique contribution to the literature since it is the first time
that the discovery learning process has been explicitly investigated in young FT and PT
infants. We believe that allowing infants to independently discover that their leg actions
activate the mobile more closely approximates the infant learning environment and may
contribute to our understanding of the infant learning process. In addition, this is the first
time that an out-of-phase coordination pattern will be reinforced in PT infants at risk for
impairments in motor coordination. If some PT infants can generate a greater amount of
out-of-phase joint coordination while interacting with the mobile, this information can be
used to better specify intervention protocols for PT infants that have difficulty generating
out-of-phase joint coordination.

INFANT LEG COORDINATION 8
CHAPTER II
THE QUANTITATIVE STUDY OF INFANT LEG MOVEMENTS:
FROM SPONTANEOUS KICKING TO TASK-SPECIFIC LEG ACTION

Abstract
The purpose of this review is to summarize the research on the quantitative study of the
leg movements of both typically developing infants and infants at risk for disability. The study
of infant leg movements provides a productive means of investigating the processes and
mechanisms which contribute to developmental changes in spontaneous kicking over the first
months of life, the early emergence of task-specific leg action, and the differences in leg
movements between typically-developing infants and infants at risk for disability. Initial
research focused on understanding the role of the central nervous system as the primary
controller of infant leg movements. Current research focuses on understanding the confluence of
systems for emergent infant action. We explore the influence of changing theory on research
questions, experimental paradigms, and interpretation of the results. We conclude with
implications for clinical practice and opportunities for future research.
Introduction
The purpose of this review is to summarize the research on the quantitative study of
infant leg movements in order to gain an understanding of the processes and mechanisms which
contribute to the emergence of spontaneous kicking and early task-specific leg action. An
understanding of how functional action emerges in typically developing (TD) children and
children at risk for motor disability is an essential foundation for pediatric neurorehabilitation.
Of particular interest to pediatric physical therapists is an understanding of how leg action
INFANT LEG COORDINATION 9
develops within the first 4-6 months of life as a precursor to early independent mobility such as
combat crawling, creeping, and walking.
Infant leg movements have been studied quantitatively for the past 30 years (Harris &
Heriza, 1987). The quantitative approach uses video digitization or motion capture technology
to precisely compute the spatial and temporal parameters of infant leg movements, such as joint
angle excursions, joint angle velocities, interjoint coordination, and interlimb coordination.
Using the quantitative approach, infant leg movements can be precisely represented during the
experimental manipulation of environmental conditions. This allows for the investigation of the
processes and mechanisms which contribute to developmental changes in spontaneous kicking
over the first months of life (Thelen, 1985; Thelen & Fisher, 1983b), the emergence of task-
specific leg action (Angulo-Kinzler, 2001; Angulo-Kinzler & Horn, 2001; Angulo-Kinzler,
Ulrich, & Thelen, 2002; Chen, Fetters, Holt, & Saltzman, 2002; Tiernan & Angulo-Barroso,
2008; Watanabe & Taga, 2009), and differences in leg movements between TD infants and
infants at risk for motor disability (Fetters et al., 2004; Fetters et al., 2010; Heriza, 1988a, 1988b;
McKay & Angulo-Barroso, 2006; Rademacher, Black, & Ulrich, 2008; Ulrich & Ulrich, 1995;
Vaal, van Soest, Hopkins, et al., 2000). This is distinguished from qualitative assessments of
general movements (Einspieler, Prechtl, Bos, Ferrari, & Cioni, 2004) and standardized
assessments of developmental milestones (Bayley, 1969; Gesell, Thompson, & Amatruda, 1934)
which are used respectively to characterize whole-body movement patterns and motor skill
development, but cannot be used to precisely quantify changes in leg action due to specific
experimental manipulations.
Historically, the quantitative study of infant leg movements has been influenced by a
change in conceptual framework from a neuromaturational framework to a systems framework.
INFANT LEG COORDINATION 10
This has resulted in a shift in research questions, experimental paradigms, and interpretation of
results. This review describes this transition and provides implications for clinical practice and
opportunities for future research.
Method
Four databases (CINAHL, Cochrane, MEDLINE, and PEDro) were searched for
literature published in English between 1981 and December 2012, using the key words infant,
kicking, coordination. References lists of key articles were scanned for additional material. The
initial date limit was set because this is the year that the first quantitative study of infant leg
movements appeared in the research literature. Studies were selected for inclusion if: (1) the
study population included infants under 6 months of age, (2) the methods included the use of
three-dimensional (3D) motion capture technology to quantify the kinematics of infant leg
movements, and (3) the methods included the use of manual digitization of video or two-
dimensional (2D) motion capture technology if the research question had not been studied using
a 3D approach. Studies were excluded if infant leg movements were elicited using methods that
were not considered to be ecologically valid, such as supporting an infant on a motorized
treadmill which mechanically moves the infant’s legs away from his body. The leg movements
that are elicited from this experimental manipulation are not within the natural repertoire of the
child, rather the treadmill either entrains the infant’s legs to move in a reciprocal leg pattern that
is not typically observed at this age or the treadmill destabilizes the infant’s postural stability to
the point that compensatory leg movements are elicited.
The titles and abstracts of the results were scanned by the first author applying the
inclusion and exclusion criteria described. Articles potentially meeting the criteria were read in
full.
INFANT LEG COORDINATION 11
Results
The initial search produced 84 results. Forty-eight articles were excluded because they
did not meet inclusion criteria. Thirty-six articles were used to inform this review: 9 addressed
spontaneous kicking of TD infants, 21 addressed task-specific leg action of TD infants, 8
addressed preterm infants, 5 addressed preterm infants at risk for cerebral palsy (CP), 2
addressed infants with myelomeningocele (MMC), and 3 addressed infants with Down syndrome
(DS). Some articles were used to inform more than one category.
This review is organized into three sections: the quantitative study of infant spontaneous
kicks, the quantitative study of infant leg action, and the quantitative study of the leg movements
of infants at risk for disability. In each section we explore the influence of conceptual
framework on research questions, experimental paradigms, and interpretation of the results. The
review concludes with clinical implications and future research directions.
The Quantitative Study of Infant Spontaneous Kicks
Neuromaturational framework. Quantification of the spatial and temporal parameters
of spontaneous infant kicks led to the observation that the spontaneous kicks of infants were
highly organized and demonstrated a relatively consistent developmental progression among
infants (Thelen, 1985). Traditionally, this consistent developmental progression was attributed
entirely to prescriptive motor programs resulting from hierarchical central nervous system (CNS)
maturation (Bayley, 1969; Gesell et al., 1934; McGraw, 1945). It was proposed that a central
pattern generator (CPG) which is being progressively regulated by higher cortical centers over
the course of development may be the sole contributor to infant spontaneous kicks.
The locomotor CPG has been conceptualized as a spinal neuronal network which
generates the commands specifying the basic locomotor pattern for rhythmical leg movements
INFANT LEG COORDINATION 12
(Grillner, 1975). It has been proposed that the locomotor CPG is organized in flexibly arranged
modules, such that unit CPGs control a group of synergistic muscles at each joint, limb CPGs
control the interjoint coordination within each leg, and the linking of two limb CPGs controls the
interlimb coordination between legs.
If the locomotor CPG is considered to be the sole contributor to infant leg movements,
three assumptions emerge when applied to the development of infant leg movements. First, all
lower extremity movements (spontaneous supine kicks, neonatal stepping, mature walking)
result from the same locomotor CPG and therefore should demonstrate strong similarity in terms
of muscle activations, interjoint coordination, and interlimb coordination. Second, spontaneous
infant kicks should demonstrate a direct progression toward a mature walking pattern during the
first year of life. Third, the developmental progression of spontaneous infant kicks should
correlate to the increased regulation of a spinal CPG by cortical and brainstem structures, such as
the myelination of the corticospinal tract (Forssberg, 1999; Yang et al., 2004).
The strongest support for a locomotor CPG that was common to spontaneous supine
kicking, neonatal stepping, and mature walking came from the early quantitative analysis of
newborn leg movements. It was observed that spontaneous supine kicks and newborn stepping
were characterized by: (1) flexion and extension phase durations which approximated 300 ms,
(2) nearly synchronous hip, knee, and ankle flexion and extension within each limb, and (3) a
predominance of alternating movements between limbs (Thelen, Bradshaw, & Ward, 1981;
Thelen & Fisher, 1983b). In addition, the interjoint and interlimb coordination patterns of supine
kicks and newborn stepping were similar to mature walking. However, further quantitative
research did not demonstrate the results expected if infant leg movements were entirely
controlled by a locomotor CPG. This led researchers to speculate that the control of leg
INFANT LEG COORDINATION 13
movements may be influenced by biomechanical and environmental factors (Fetters et al., 2004;
Fetters et al., 2010; Heriza, 1988a; Jeng et al., 2002; Piek, 1996; Thelen, 1985; Thelen & Fisher,
1982, 1983b).
Infant kicks that were solely regulated by a locomotor CPG were expected to be
characterized by alternating phasic activation of the leg flexors and extensors. The analysis of
infant kicks using surface electromyography (EMG) both confirmed and refuted this assumption.
As expected, 2- and 4-week old infants demonstrated consistencies in the muscle activations
used to initiate the flexion phase of kicks: strong, phasic activation of hip and ankle flexor
muscles (Thelen & Fisher, 1983b). Unexpectedly, the hip and ankle extensors were co-activated
during the flexion phase of kicks, and minimal if any muscle activity was observed during the
initiation of the extensor phase. Researchers hypothesized that passive forces resulting from the
effects of gravity and the viscoelastic properties of the leg might contribute to the termination of
the flexion phase and be responsible for the entire extensor phase of infant kicks (Thelen, 1985;
Thelen & Fisher, 1983b).
It was also expected that there would be a direct progression from in-phase interjoint
coordination to out-of phase interjoint coordination as the spinal CPG was regulated by higher
cortical structures. Analysis of infant kicks over the first 5 months of life seemed to support this
assumption. The spontaneous kicks of typically-developing newborn infants were dominated by
strong in-phase movements of the hip, knee, and ankle joints into flexion and extension (Heriza,
1988a; Thelen & Fisher, 1983b) which became more out-of-phase over the first months of life
starting with the hip-ankle joints at 1 to 2 months of age (Fetters et al., 2004; Jeng et al., 2002;
Piek, 1996; Thelen, 1985; Vaal, van Soest, Hopkins, et al., 2000) followed by the hip-knee and
knee-ankle joints over the next few months (Thelen, 1985; Vaal, van Soest, Hopkins, et al.,
INFANT LEG COORDINATION 14
2000). Unexpectedly, a return to more in-phase movement of all joints was noted at 5 to 8
months as well as an increase in the velocity of leg movements (Fetters et al., 2010; Thelen,
1985). One hypothesis was that the observed increase in the velocity of leg movements
constrained the joints of the leg to move in a more in-phase manner due to biomechanical and
environmental factors such as the effect of gravity, ratio of leg mass and muscle strength, and
elastic properties of the leg (Fetters et al., 2010).
Last, it was expected that the alternating kicks observed during the spontaneous supine
kicks of newborns would provide a foundation for the alternating leg movements required for
independent walking. However, although newborns demonstrated a predominance of alternating
kicks with relatively tight spatial and temporal coordination (Thelen, 1985; Thelen, Ridley-
Johnson, & Fisher, 1983), from 1 to 4 months of age a predominance of unilateral kicks was
observed with little or no apparent interlimb coordination (Jeng et al., 2002; Thelen, 1985;
Thelen et al., 1983). This was followed by a predominance of bilateral symmetrical kicks at 4 to
6 months of age which definitely did not provide the prerequisite interlimb coordination
expected to support independent walking (Jeng et al., 2002; Thelen, 1985; Vaal, van Soest,
Hopkins, et al., 2000).
Summary. Although a spinal CPG that is being progressively regulated by higher
cortical centers may contribute to the control of early leg movements, its contribution has been
difficult to quantify since the developmental progression of infant kicks does not demonstrate the
spatial and temporal organization consistent with this explanation. In addition, a strong
correlation has not been established between changes in infant kicking patterns and the
regulation of higher cortical structures on a spinal CPG. Consequently, the quantitative study of
infant spontaneous kicking has increased our understanding of the typical changes in
INFANT LEG COORDINATION 15
spontaneous kicking that occur during the first months of life, but has not explicitly identified the
precise processes or mechanisms which contribute to these changes.
The Quantitative Study of Infant Leg Action
Systems framework. Contemporary theories of motor development, such as Dynamic
Systems Theory (Heriza, 1991; Scholz, 1990; Thelen & Smith, 1994) and Perception-Action
Theory (Fetters & Ellis, 2006; Gibson & Pick, 2000; Newell, 1986) interpret a change in motor
behavior as resulting not only from CNS factors, but also biomechanical and environmental
factors. These theories support what we will refer to as a systems framework since they share
the premise that motor skills emerge from the self-organizing properties of multiple subsystems
of the infant interacting with a given task and environmental context. The CNS is not considered
to be the sole contributor to motor development, rather each human action is supported or
constrained by the interaction between the environment and the many contributing infant
subsystems including: perceptual, cognitive, affective, neuromuscular, and skeletal. The infant’s
subsystems, as well as the environmental conditions under which action emerges, provide
constraints, or boundaries, for emergent action (Fetters & Ellis, 2006; Gibson & Pick, 2000;
Newell, 1986). Research within a systems framework investigates the relation between emergent
action and these multiple constraints.
Research on infant leg movements focuses on the relation of leg action and: (a) the body
constraints associated with moving in a gravitational environment (Jensen, Schneider, Ulrich, &
Zernicke, 1994; Thelen, Skala, & Kelso, 1987) and (b) the causal relations discovered through
perceptual-motor exploratory leg action (Angulo-Kinzler, 2001; Thelen & Fisher, 1983a; Tiernan
& Angulo-Barroso, 2008). Innovative experimental paradigms are used to investigate the
relation between emergent action and specific infant-environmental constraints. In a typical
INFANT LEG COORDINATION 16
paradigm, infants explore environmental constraints in relation to their dynamics. For example,
an infant’s orientation in space has been manipulated to determine the effect of gravity on
spontaneous kicking (Jensen et al., 1994). A significant advantage of the systems perspective is
the wide range of potential independent variables for study and the ability to directly manipulate
these variables to explicitly identify their relative influence on infant action. In the neuro-
maturational framework, the only independent variable for study was the degree of CNS
maturation and this variable could not be directly manipulated or even definitively assessed in
human infants due to limitations of neuroimaging and the inability to use invasive methods.
Manipulation of body constraints while moving in a gravitational environment.
Unilateral weighting. The influence of changes to body constraints on leg action has
been studied by weighting one leg of TD infants. Since the legs are considered part of a linked
neuromuscular system, it was expected that unilateral weighting might result in an alteration of
the kicking rate of both legs and that additional properties might emerge from the dynamic
nature of the linked system. When the kicking of 6-week old infants was assessed during
unilateral weighting with 185 grams, the infants altered the kicking rate of both legs: the kicking
rate of the weighted leg decreased and the kicking rate of the unweighted leg increased such that
the overall kicking rate was preserved (Thelen et al., 1987). Moreover, the amplitude and
velocity of the weighted leg were maintained at baseline levels, but significantly increased in the
unweighted leg as compared to baseline. It was proposed that the infants perceived the load
perturbation and increased their overall neural activation levels resulting in a consistent overall
kick rate, consistent amplitude and velocity of the weighted leg, and increased amplitude and
velocity of the unweighted leg (Thelen et al., 1987). This finding was interpreted as evidence for
a dynamical mass-spring model of leg control. In this model, the CNS regulates the overall
INFANT LEG COORDINATION 17
stiffness of the infant’s legs to compensate for the heavier load, but the spatial and temporal
parameters of the leg movement emerge from the intrinsic biomechanics of the legs moving in a
gravitational environment.
Age-related effects of unilateral weighting were investigated longitudinally in infants
from 6 to 26 months of age by weighting one leg with one third of leg mass (Vaal, van Soest, &
Hopkins, 2000). As reported previously, 6-week old infants decreased kicking rate of the
weighted leg and increased kicking rate of the unweighted leg, whereas 12-week old infants
increased the kicking rate of both legs (Vaal, van Soest, & Hopkins, 2000). Eighteen and 26-
week old infants demonstrated no change in the kicking rate or kinematic measures of either leg
in response to unilateral weighting (Vaal, van Soest, & Hopkins, 2000). The precise infant
constraints which drove these changes in kick frequency have not been systematically
investigated, however, insufficient force-generating capacity of the muscles to counteract the
additional mass has been proposed as a constraint when kicking frequency decreases in the
weighted leg. In addition, enhanced tactile/proprioceptive input has been proposed when kicking
frequency increases in either leg, and an improved ability to move the legs in a task-specific
manner has been proposed when kicking frequency remains unchanged in infants over 4 months
of age (Vaal, van Soest, & Hopkins, 2000).
Orientation to gravity. The influence of gravity on leg action has been studied by
changing the orientation of the infant to gravity. Kicking frequency was predicted to decrease
when infants were oriented upright since it was expected that the force-generating capacity of
young infants would be insufficient to counteract gravitational constraints. As expected, 3-
month old TD infants demonstrated more in-phase hip and knee joint movement with decreased
amplitude of kicks in the vertical as compared to the supine or angled positions in relation to
INFANT LEG COORDINATION 18
gravity (Jensen et al., 1994). In addition, infants demonstrated decreased velocity and amplitude
of kicks from an angled versus supine position (Chapman, 2002). Decreasing the amplitude and
velocity of movement as well as coupling hip and knee movement into an in-phase synergy may
emerge when the capacity to generate the necessary forces is insufficient for task constraints.
Influence of the causal relations the infant discovers through exploration.
Infant-activated mobile reinforcement. The influence of learning on leg action has been
studied by providing contingent reinforcement to specific leg movements using an overhead
infant mobile. Three-month old infants differentially increased the kicking rate of one leg when
movement of that leg was reinforced with a moving mobile (Heathcock, Bhat, Lobo, &
Galloway, 2004, 2005; Rovee-Collier & Gekoski, 1979; Rovee & Rovee, 1969; Thelen & Fisher,
1983a). Three-month old infants also exhibited precise control of their leg when mobile
reinforcement was contingent on a specified interjoint coordination pattern, such as crossing 85°
of knee flexion, crossing 35° of knee extension, or synchronously extending the hip and knee
(Angulo-Kinzler, 2001; Angulo-Kinzler & Horn, 2001; Angulo-Kinzler et al., 2002). When
crossing these joint angles, infants either demonstrated a movement-based strategy consisting of
increasing kicking frequency or a posture-based strategy consisting of small movements around
the required joint angle. These studies support the view that 3-month old infants not only
perceive the causal relation between their leg movement and mobile activation, but also have the
ability to precisely control their leg actions to elicit an expected environmental result.
When many different coordination patterns were reinforced by mobile movement, 3 and
4-month old infants preferentially demonstrated the interjoint or interlimb coordination pattern
that most efficiently elicited mobile reinforcement. For example, infants demonstrated an out-of-
phase interjoint coordination pattern of hip flexion with knee extension when it provided a more
INFANT LEG COORDINATION 19
direct means to elicit mobile reinforcement (Chen et al., 2002). Infants also altered their
preferred interlimb coordination pattern of either unilateral or alternating kicks, to simultaneous
bilateral kicks when simultaneous kicks elicited increased mobile reinforcement (Thelen, 1994).
Results from these studies highlight the ability of young infants to explore differences in
reinforcement based on their leg action and tune their intrinsic dynamics to optimize the rate of
reinforcement.
The ability to control the arms as well as the legs also supports this perception and
control. Limb action during the mobile paradigm changes during early infancy from general
movement of all four limbs at 2-months of age to more specific movement of only the reinforced
limb at 4-months of age (Watanabe & Taga, 2006). When mobile reinforcement was contingent
on the movement of one arm, 2-month old infants increased the movement of all four limbs, 3-
month old infants increased the movement of the arms as compared to the legs, and 4-month old
infants increased the movement of the reinforced arm only as compared to the unreinforced
limbs (Watanabe & Taga, 2006). The differential activation of interlimb coordination patterns
as a function of age was interpreted as resulting from differences in resources for perception and
action although the study did not determine the precise infant and environmental constraints that
drove the change from general to more specific limb action.
Environmental information. The type and amount of environmental information has
been studied as a constraint to infant action. It was expected that enhanced, contingent
information would result in improved learning of the causal relation between leg action and
environmental consequences. As expected, 3-month old infants demonstrated enhanced learning
of the association between leg action and mobile activation when contingent, multimodal
auditory and visual information was provided by the mobile versus unimodal auditory, unimodal
INFANT LEG COORDINATION 20
visual, or non-contingent information (Tiernan & Angulo-Barroso, 2008). Four-month old
infants also demonstrated enhanced learning during the mobile paradigm when proprioceptive
information was augmented with the use of leg weights (Chen et al., 2002).
Experiential history. Experiential history has been investigated as a possible constraint
to action during the mobile paradigm by manipulating the reinforced limb. It was expected that
prior learning, including using the other leg, would decrease the amount of time necessary to
learn the causal relation between leg movement and mobile activation. As expected, when the
right leg was reinforced first followed by the left leg, 3-month old infants learned the association
between leg movement and mobile activation more quickly with the left leg as compared to the
right leg (Angulo-Kinzler, 2001). In addition, the effect of prior learning on emergent action has
been shown to be highly influenced by the strategy used during the prior learning. When only
the movement of one arm was reinforced with mobile activation, 3-month old infants increased
movement of both arms. When this was followed by reinforcement of the movement of one leg,
the infants decreased arm movement and increased leg movement (Watanabe & Taga, 2009).
However, in a second group of 3-month old infants in which a leg was reinforced first followed
by an arm, infants increased movement of all limbs when the leg was reinforced and continued
moving all limbs when the arm was reinforced (Watanabe & Taga, 2009). These findings
support that 3-month old infants use different strategies to activate the mobile based on whether
an arm or leg is reinforced, and that this previous strategy influences future performance even
when a different limb is reinforced.
End effector. Differences in the use of the hands and feet as end effectors has been
studied by placing toys within reach of an infant’s hands and feet (Galloway & Thelen, 2004).
Infants as young as 8 weeks old reached toward a toy with their feet on average 4 weeks earlier
INFANT LEG COORDINATION 21
than they reached toward a toy with their hands. This finding was explained by differences in
the anatomy and movement patterns of the legs and arms which reduces the amount of
movement that needs to be controlled during foot reaching as compared to hand reaching.
Summary. Infant leg action results from the interaction of body constraints associated
with moving in a gravitational environment and the causal relations discovered through
exploratory leg action. The CNS is no longer considered to be solely responsible for the control
of movement via prescriptive motor programs, rather movement is softly assembled and
dynamically controlled through the self-organization of multiple subsystems. The research
reviewed thus far supports that leg action is influenced by: (1) biomechanics of the limb
segments, (2) force-generating capacity in relation to task, (3) enhanced tactile/proprioceptive
information, (4) perception of causal relations, (5) experiential history, and (6) multi-modal
information. The possible influence of these constraints informs our understanding of the control
of typical human action and provides an expanded foundation for the study of infants at risk for
disability.
The Quantitative Study of the Leg Action of Infants at Risk for Disability
Infants at risk for disability provide a unique sample to investigate infant constraints that
contribute to task-specific changes in action. The underlying pathology that places infants at risk
contributes to differences in leg action due to primary impairments, such as decreased sensation,
decreased motor unit accessibility, and decreased force-generating capacity of muscles; as well
as secondary impairments, such as muscle contractures and osteopenia, which result from
moving atypically in a gravitational environment. Infants at risk may also demonstrate
differences in leg action as a result of a reduced ability to extract the relevant information for
action.
INFANT LEG COORDINATION 22
Initial research with infants at risk for disability focused on identifying differences in
movement parameters between at risk infants and TD infants with specific differences in
movement parameters becoming associated with specific at risk populations (Fetters et al., 2004;
Fetters et al., 2010; Heriza, 1988a; Jeng et al., 2002; McKay & Angulo-Barroso, 2006;
Rademacher et al., 2008; Ulrich & Ulrich, 1995; Vaal, van Soest, Hopkins, et al., 2000). For
example, infants with brain damage, at risk for CP, demonstrate kicks with increased in-phase
interjoint coordination possibly due to damage to the corticospinal tracts (Fetters et al., 2004;
Vaal, van Soest, Hopkins, et al., 2000). Whereas infants with DS demonstrate more time spent
in low intensity leg activity possibly due to decreased force-generating capacity of muscles
(McKay & Angulo-Barroso, 2006). Once distinguishing movement parameters had been
established for an at risk population, research questions rapidly began to focus on identifying
possible constraints that influence leg action and how leg action could be influenced by
environmental factors, such as unilateral weighting (Ulrich, Ulrich, & Angulo-Kinzler, 1997;
Vaal, van Soest, Hopkins, & Sie, 2002), changing the infant’s orientation to gravity (Chapman,
2002), and mobile reinforcement (Heathcock et al., 2004).
Preterm infants. Infants born preterm (PT) are at risk for impairments in motor
coordination which tend to increase with decreasing gestational age and birth weight (de Kieviet
et al., 2009; Stephens & Vohr, 2009). PT infants exhibit a more extended leg posture than TD
infants and demonstrate differences in early task-specific leg action. The spontaneous kicking of
PT infants at 1-month corrected age (CA) has been characterized by a more extended leg posture
than full-term (FT) infants possibly due to the early influence of gravity on their musculoskeletal
system (Fetters et al., 2004). By 4 to 5-months CA PT infants as compared to FT infants
demonstrate more out-of-phase interjoint coordination, however, differences in spontaneous
INFANT LEG COORDINATION 23
kicking between PT and FT infants have largely resolved (Fetters et al., 2010; Jeng et al., 2002).
A decreased rate of walking attainment of PT infants, as compared to FT infants, has been
associated with parameters of early spontaneous kicking, including a high hip-knee correlation at
2 months CA, high kick frequency at 4 months CA, and short intra-kick pause together with a
low variability in interlimb coordination at 2 and 4 months CA (Jeng, Chen, Tsou, Chen, & Luo,
2004).
Differences in early task-specific action between PT and FT infants have been
investigated using the mobile paradigm. Age-corrected 3-month old PT infants required more
time to learn the contingency between leg action and mobile reinforcement (two versus one
session) and didn’t exhibit retention a week later (Gekoski, Fagen, & Pearlman, 1984). In
addition, when movement of one leg was reinforced with mobile movement, 3-month old FT
infants increased the kicking frequency of only the reinforced leg, whereas age-corrected 3-
month old PT infants increased the kicking frequency of both legs, even after a 6-week
intervention (Heathcock et al., 2004, 2005). It was hypothesized that the PT infants did not learn
the association between mobile movement and movement of the reinforced leg due to poor visual
attention, poor regulation of arousal level, and tight interlimb coordination (Heathcock et al.,
2004). Difficulties in regulating arousal level has been documented in PT infants during the
mobile paradigm and has been associated with slower rates of contingency learning (Haley,
Grunau, Oberlander, & Weinberg, 2008; Haley, Weinberg, & Grunau, 2006). Specifically, PT
infants, as compared to FT infants, looked less at the mobile, demonstrated higher basal heart
rates, showed greater increases in heart rate responses to the contingency, and demonstrated
slower rates of contingency learning (Haley et al., 2008). The researchers suggested that the PT
infants began the paradigm in a heightened state of arousal, as supported by the higher basal
INFANT LEG COORDINATION 24
heart rates, then became over-aroused during the mobile paradigm, exhibiting greater increases
in heart rate responses to the contingency, and looked away from the mobile in an attempt to
regulate their arousal (Haley et al., 2008). They speculated that this strategy to regulate arousal
contributed to the slower rate of learning of the PT group and may be one factor underlying the
learning difficulties of PT infants.
Preterm infants at risk for cerebral palsy. CP describes a group of permanent
disorders of movement and posture that are attributed to nonprogressive disturbances which
occurred in the developing fetal or infant brain (Rosenbaum et al., 2007). PT infants are at
increased risk for developing CP due to a high incidence of cerebral lesions, specifically white
matter damage (WMD) of the corticospinal tracts (Bax, Tydeman, & Flodmark, 2006; Beaino et
al., 2010; Han, Bang, Lim, Yoon, & Kim, 2002; Volpe, 2009). Differences in interjoint
coordination appear to distinguish between preterm infants with white matter damage (PTWMD)
and TD infants (Chen & Fetters, 2002). Increased in-phase joint coordination has been identified
in PTWMD infants as early as 1-month CA (Fetters et al., 2004). This difference reduces by 5
months (Fetters et al., 2010), but some PTWMD infants continue to demonstrate excessive in-
phase joint coordination which places them at increased risk of developing spastic CP (Vaal, van
Soest, Hopkins, et al., 2000). It has been proposed that it may be possible to detect newborns at
risk for developing CP using eight parameters derived from the analysis of spontaneous kicking
using quantitative motion analysis (Meinecke et al., 2006). These parameters include: the
skewness of the feet’s velocity, the cross-correlation in acceleration between the left and right
foot, the periodicity in the feet’s trajectory and velocities, and the area in which the feet’s
trajectory and speed profiles are out of the standard deviation of the moving average and the area
differs from the moving average.
INFANT LEG COORDINATION 25
PTWMD infants, as compared to TD infants, demonstrated differences in their response
to unilateral weighting at 26-weeks CA (Vaal et al., 2002). The two groups of infants
demonstrated similar tightness of interlimb coordination patterns during unilateral weighting,
however, PTWMD infants with the most severe WMD increased the kicking rate of both legs,
whereas the PTWMD infants with less significant WMD and the TD infants demonstrated no
change in kicking rate. The precise constraints that contributed to the differences in response to
unilateral weighting were not identified.
Myelomeningocele. MMC is a congenital defect of the spinal cord that results in
impairments of motor and sensory function below the level of the lesion. Infants with MMC as
compared to TD infants demonstrated shorter movement durations, more symmetrical leg
movements, and decreased frequency of movement during spontaneous kicking (Chapman,
2002; Rademacher et al., 2008). However, similar to TD infants, infants with MMC
demonstrated decreased amplitudes and velocities of leg movement when positioned in more
vertical orientations possibly due to the greater force required to counteract gravity (Chapman,
2002). These findings support that action of infants with MMC may be influenced by the
infant’s constraints (decreased frequency of movement due to decreased sensation and force-
generating capacity), but the emergent actions under constraints remain (decreasing amplitudes
and velocities of leg movement when force-generating requirements increase beyond strength
capacity).
Down syndrome. DS is a genetic disorder characterized by cognitive and sensorimotor
impairments including hypotonia, ligamentous laxity, decreased force-generating capacity of
muscles, and decreased sensitivity to somatosensory information. Four to 6-month old infants
with DS and TD infants increased the frequency of their leg movement in response to unilateral
INFANT LEG COORDINATION 26
weightings of 25%, 50% and 100% of their estimated calf mass (Ulrich et al., 1997). However,
infants with DS tended to increase their frequency of leg movement at higher weights than TD
infants or not at all possibly due to decreased sensitivity to somatosensory information (Ulrich et
al., 1997). Three to 6-month old infants with DS, as compared to TD infants, demonstrated
similar frequency of leg movement, however, they spent more time in low intensity activity
when movement was analyzed over 24 hours using accelerometers (McKay & Angulo-Barroso,
2006). This is supported by data analyzed from video in which infants with Down syndrome
produce more foot rubs (low intensity activity) and less kicks (high intensity activity) than TD
infants (Ulrich & Ulrich, 1995). This may have significant implications for independent walking
since infants with DS who demonstrated less frequent kicking (high intensity activity) in the
early months of life walked later than those who demonstrated a greater frequency of kicking
(Lloyd, Burghardt, Ulrich, & Angulo-Barroso, 2010; McKay & Angulo-Barroso, 2006; Ulrich &
Ulrich, 1995).
Summary. The study of infants at risk for disability has provided a unique opportunity
to investigate infant constraints that contribute to spontaneous kicking and task-specific leg
action. In particular, the use of environmental perturbations and innovative task-contexts to
discern the influence of infant constraints on leg action has proven to be a rich avenue of
research. The research reviewed supports that the leg actions of infants are influenced not only
by the constraints listed previously for TD infants but also by: (1) the ability to regulate affect,
(2) parasympathetic activity, and (3) sensitivity to somatosensory information. Further
investigation of infants with well-defined impairments as they interact in specific task-contexts,
is necessary to more closely examine the relative contributions of an infant’s perceptual
capabilities, resources for action, and environmental considerations on emergent action. This
INFANT LEG COORDINATION 27
line of research clarifies the constraints that influence leg action of TD infants and infants at risk
for disability, and may also provide a stronger foundation for the role of the environment in
optimizing the leg action of infants at risk for disability.
Discussion
Clinical Implications
The quantitative study of infant leg movements has increased our understanding of the
typical changes in spontaneous kicking and task-specific leg action that occur during the first
months of life. In addition, it has highlighted the importance of early identification and
intervention of atypical kicking patterns and task-specific leg action. During therapeutic
assessment, clinicians can use this information to distinguish between typical and atypical
spontaneous kicking patterns and identify the subsystems that will need to be assessed in order to
determine their contribution to atypical leg action. Subsystems may include: biomechanics of
the limb segments, force-generating capacity in relation to task, perception of relevant sensory
information, and perception of causal relations. During therapeutic intervention, this information
can be used to encourage age appropriate interjoint and interlimb leg coordination patterns
through: addressing the subsystems that are contributing to atypical leg action, enhancing tactile
and proprioceptive information, and utilizing multi-modal contingent reinforcement of specific
leg action.
Future Research Directions
The quantitative study of infant leg movements has provided insight into the infant and
environmental constraints that influence spontaneous kicking and task-specific leg action.
However, the precise influence of specific infant and environmental constraints on leg action has
yet to be determined. Future research of infant leg action must accurately quantify and
INFANT LEG COORDINATION 28
systematically manipulate infant and environmental constraints to explicitly determine their
influence on leg action. For example, the force-generating capacity of infants is often proposed
as a constraint to infant leg action and has practical implications for the rehabilitation of infants
at risk for motor disability; however, strength has not been quantified or systematically
manipulated to determine its precise influence on infant leg action in specific task-contexts.
These specific relations must be clarified in order to understand the contribution of specific
constraints to infant leg action, as well as to provide a stronger foundation for the role of the
environment in optimizing the leg action of infants at risk for motor disability.
The quantitative study of infant leg movements has also contributed to our understanding
of the process by which infants learn. As infants spontaneously move their legs they perceive
the causal relation between their leg action and environmental outcomes, they then demonstrate
either exploratory leg action to generate more information about possible environmental
outcomes or exploitive actions to elicit an expected environmental response. The quantitative
study of leg action provides evidence that infants can generate very specific leg action to exploit
an expected environmental response. However, it is currently unknown the extent to which
exploratory strategies influence learning. Future research is needed to identify how exploration
strategies influence learning, as well as the optimal exploration level that maximizes the speed of
learning and the rate of reinforcement in specific task-contexts. From this research,
individualized reinforcement schedules for specific task-contexts can be developed to optimize
the leg action of TD infants and infants at risk for motor disability.
To develop a principled account of the neuromotor control mechanisms which implement
spontaneous kicking and emergent leg action, the integration of a wide range of infant
constraints, environmental constraints, and infant-environmental relations is required.
INFANT LEG COORDINATION 29
Computational modeling may be an appropriate means to augment the study of early leg action.
A computational model is a mathematical simulation which can be used to explicitly test
particular aspects of a hypothesis or theory. The quantitative approach lends itself well to
computational modeling since infant leg action in specific task-contexts can be precisely
described. The dynamical mass-spring model of infant leg control is one example of a
computational model that has been used in the infant kicking research (Thelen et al., 1987).
Future research would benefit from the expanded use of computational models to investigate the
relation between emergent action and specific infant-environmental constraints, the relation
between exploratory and exploitive leg action, the critical factors which contribute to the
emergence of task-specific leg action, and the process by which the primary impairments of
infants at risk for motor disability lead to future secondary impairments and functional
limitations.

INFANT LEG COORDINATION 30
CHAPTER III
CHANGES IN INFANT LEG COORDINATION:
INSIGHTS FROM A 3D KINEMATIC AND KINETIC APPROACH

Abstract
The purpose of this study is to investigate the contribution of torque changes to early changes in
infant leg joint coordination. We analyzed kicking actions within 10 full-term infants and
between 6 and 15-weeks of age using a three dimensional kinematics and kinetics approach. The
majority of joint angle pairs demonstrated a change from an in-phase intralimb coordination
pattern at 6-weeks to an increased variety of intralimb joint coordination patterns at 15-weeks.
Unlike joint coordination, an obvious developmental change was not observed in joint torques.
Net joint torques and partitioned joint torque profiles normalized to the mass of the leg remained
relatively consistent across ages. Further analysis supported that the increased variety of joint
coordination patterns observed between the hip and knee joints from 6 to 15-weeks of age is
associated with a decreased influence of knee muscle torques which allow passive knee
gravitational and motion dependent torques to have a greater influence on the coordination of the
kick.
Introduction
Human infants experience continual and profound changes in all body systems during the
first year of life. They begin to master control of their bodies in relation to the world while their
central and peripheral nervous systems, as well as their musculoskeletal system, are in dynamic
development. Control of actions such as kicking, requires not only the generation of active
muscle forces, but the anticipation and control for the effects of gravity and for the emergent
INFANT LEG COORDINATION 31
passive forces that are generated as a result of active muscle force generation (Bernstein, 1967;
Jensen et al., 1994; Schneider, Zernicke, Ulrich, Jensen, & Thelen, 1990). Mastery of action is
achieved through exploration and discovery (Gibson, 1988) and we have begun to understand the
nature of this discovery process (Adolph & Anthony, 2000; Chen et al., 2002; Thelen & Fisher,
1983a) and the properties of actions that emerge through discovery.
Functional coordination patterns of the upper extremities are apparent very early in
development. Infants attempt to coordinate their upper limbs as toys and objects are within sight
and reach (Bhat, Heathcock, & Galloway, 2005; Bhat, Lee, & Galloway, 2007; Lee, Bhat,
Scholz, & Galloway, 2008). The functional coordination patterns of the lower extremities do not
become apparent until months later when infants start crawling and creeping. The joint
coordination of spontaneous kicking, however, changes significantly during the early months of
life, even without an apparent task. The spontaneous kicks of newborn infants are dominated by
an in-phase intralimb coordination pattern of synchronous flexion and extension of the hip, knee,
and ankle joints. Within the first months of life, infants demonstrate a greater variety of
intralimb joint coordination patterns (Fetters et al., 2004; Fetters et al., 2010; Heriza, 1988a; Jeng
et al., 2002; Thelen & Fisher, 1983b; Vaal, van Soest, Hopkins, et al., 2000), which is thought to
support the progress to more functional coordination patterns of the lower extremities necessary
for independent mobility, such as creeping and walking (Fetters et al., 2010; Thelen & Fisher,
1983b).
This transition in leg joint coordination patterns has been studied using kinematics
(Fetters et al., 2004; Jeng et al., 2002; Piek, 1996; Thelen, 1985; Vaal, van Soest, Hopkins, et al.,
2000). In terms of the motion of flexion and extension, the transition from a predominately in-
phase intralimb coordination pattern to the emergence of a greater variety of intralimb
INFANT LEG COORDINATION 32
coordination patterns is observed to begin with the hip-ankle joints at 1 to 2 months of age
(Fetters, et al., 2004; Jeng, et al., 2002; Piek, 1996; Thelen, 1985; Vaal, van Soest, Hopkins, et
al., 2000) followed by the hip-knee and knee-ankle joints over the next few months (Thelen,
1985; Vaal, van Soest, Hopkins, et al., 2000). Coordination between joint angle pairs which
include the rotational aspects of infant leg motion, such as hip abduction/adduction and hip
external/internal rotation, have not yet been investigated due to limitations in the method used to
collect and analyze position data from infant spontaneous kicks. To date, position data has been
collected from the spatial three dimensional (3D) coordinates of individual markers placed on
anatomic landmarks, typically the lateral midline of the trunk, greater trochanter of the hip, knee
joint line, lateral malleolus, and fifth metatarsal (Fetters et al., 2004; Fetters et al., 2010; Jeng et
al., 2002; Vaal, van Soest, Hopkins, et al., 2000). Although this has allowed for the computation
of relatively accurate sagittal plane estimates of flexion and extension angles at the hip, knee,
and ankle, accurate estimates of all joint angles in all planes of motion have not been computed,
i.e. hip abduction/adduction, hip external/internal rotation and ankle eversion/inversion.
Results from research using electromyography (EMG) support that 2- and 4-week old
infants initiate kicks using strong, phasic activation of the hip and ankle flexors, often with co-
contraction of the hip and ankle extensors (Thelen & Fisher, 1983b). These strong activations
may contribute to the in-phase character of the kinematics. However, minimal, if any, muscle
activity reverses the hip joint from flexion to extension (Thelen & Fisher, 1983b). It has been
hypothesized that gravitational forces acting with the viscoelastic properties of the leg contribute
to hip joint reversal and the entire extensor phase of infant kicks (Thelen, 1985; Thelen & Fisher,
1983b).
INFANT LEG COORDINATION 33
To clarify the contribution of muscle (MUS) torque and passive gravitational (GRA) and
motion dependent (MDT) torques to hip joint reversal, Schneider et al. compared the
intersegmental dynamics of one representative high-intensity kick, defined through accelerations,
and one representative low-intensity kick from 3-month old infants (Schneider et al., 1990).
They proposed that with changing passive torques produced from the differing intensity of kicks,
the MUS torque was adjusted to produce a net (NET) torque to reverse the kicking motion.
However, this study was limited by the analysis of only two kicks and the utilization of a two-
dimensional (2D) dynamics approach to compute the intersegmental dynamics. The utilization
of a 2D approach may lead to incorrect estimates of torques during actions, like infant kicking,
which do not occur primarily in the sagittal plane. Valid estimates of intersegmental torques for
infant kicking actions require a 3D dynamics approach, yet this approach has not been used in
previous research to investigate infant kicking actions.
We are interested in the relation of intersegmental dynamics to the developmental change
in leg joint coordination which occurs during the first months of life, from a predominately in-
phase intralimb coordination pattern to an increase in the variety of intralimb coordination
patterns. The association between changes in joint coordination and changes in intersegmental
dynamics has not been studied. It is unknown whether the transition to a greater variety of joint
coordination patterns is associated with an increase or decrease in MUS torque. It is also
unknown whether passive GRA and MDT torques contribute to the dominance of an in-phase
pattern or to an increase in the variety of joint coordination patterns.
In summary, previous research supports that over the first months of life the coordination
pattern of infant spontaneous kicks change from a predominately in-phase coordination pattern
of synchronous flexion and extension of the hip, knee, and ankle joints to an increase in the
INFANT LEG COORDINATION 34
variety of joint coordination patterns. The coordination of other joint angle pairs remains
unstudied, as well as changes in intersegmental dynamics, and the contribution of intersegmental
dynamics to the developmental transition to a greater variety of joint coordination patterns.
The objectives of this study are to 1) characterize the early changes in joint coordination
between all joint angle pairs, 2) identify changes in intersegmental dynamics of infant
spontaneous kicks, and 3) describe the relation between changes in joint coordination and
intersegmental dynamics during a time period of change in the coordination of kicking actions.
To achieve these objectives, standard 3D kinematics were used to compute complete 3D
kinematic data (including sagittal, frontal, and transverse plane motion) and Lagrangian
mechanics were used to determine the 3D kinetics of infant spontaneous kicking within subjects
and between 6 and 15-weeks of age. We first investigate changes in joint coordination between
all joint angle pairs from 6 to 15-weeks of age. Our hypothesis is that all joint angle pairs will
change from a predominately in-phase coordination pattern at 6-weeks to an increased variety of
joint coordination patterns at 15-weeks. We then investigate the intersegmental dynamics of
infant spontaneous kicks during this period. We hypothesize that hip NET joint torque will
increase from 6 to 15-weeks of age as infants gain greater control of their hip motion and hip co-
contraction decreases. We also hypothesize that knee and ankle NET torques will decrease as
muscle co-activation decreases. We realize that since our experimental paradigm does not
include EMG we cannot investigate hip co-contraction and knee and ankle co-activation directly,
however, here we simply provide a rationale for how we think the torques will change from 6 to
15-weeks of age. Finally, we investigate the relation between changes in coordination and
changes in intersegmental dynamics through the analysis of the coordination of the hip and knee
into flexion and extension. Since it is unknown whether a greater variety of hip-knee joint
INFANT LEG COORDINATION 35
coordination patterns from 6 to 15-weeks of age is associated with an increased or decreased
influence of passive GRA and MDT torques, we held two opposing hypotheses. We
hypothesized that a greater variety of hip-knee joint coordination patterns would be associated
with either: (1) an increased influence of knee MUS torque to dampen the passive knee GRA and
MDT torques which are associated with more in-phase joint coordination, or (2) a decreased
influence of knee MUS torque which allow passive knee GRA and MDT torques to have a
greater influence on the coordination of the kick, contributing to a greater variety of coordination
patterns between the hip and the knee joints. This study is a unique contribution to the literature
since it is the first time that infant leg actions have been investigated using a 3D kinematic and
kinetic approach, as well as the first time that these two analyses have been combined to
investigate changes in coordination in young infants.
Method
Participants
Ten infants participated in this longitudinal study at both 6 and 15-weeks of age (±1
week). This is the developmental period during which a greater variety of leg joint coordination
patterns emerge and infants kick spontaneously in supine without attempts to roll over. All
participants were healthy, full-term infants, without known sensorimotor impairments, as
confirmed by their parents and their scores on the motor subtest of the Bayley Scales of Infant
Motor Development, 3
rd
Edition, (Bayley, 2006) conducted at 15-weeks of age. Participant
characteristics are listed in Table 3.1. Parents signed a consent form prior to participation in the
study, and families received a small gift for their participation. The Institutional Review Board
at the University of Southern California approved the study.

INFANT LEG COORDINATION 36
Table 3.1

Full-Term Infants: Participant Characteristics

Participant
Number
Gender Age

Ponderal
Index

Bayley III
Motor
Subtest

Number of
Kicks for
Kinematic
Analysis
Number of
Kicks for
Kinetics
Analysis
days kg/m
3
percentile rank
FT 1 F 43 26.7 33 0
107 24.1 42 161 14

FT 2 F 49 24.2 23 3
103 24.0 79 91 4

FT 3 M 45 26.8 284 64
109 27.7 58 264 49

FT 4 F 44 24.6 82 3
109 25.2 58 191 40

FT 5 F 49 23.2 199 30
112 22.1 68 157 0

FT 6 M 38 25.8 221 17
100 21.7 84 444 183

FT 7 F 36 25.1 280 83
114 23.4 88 301 8

FT 8 M 41 22.4 455 203
100 22.9 88 111 13

FT 9 F 40 24.1 450 139
104 23.6 84 278 114

FT 10 M 48 24.3 156 2
112 25.3 68 192 6

Note. F=female, FT = full-term, M=male

INFANT LEG COORDINATION 37
Procedure
Experimental set-up. All testing was completed in the Development of Infant Motor
Performance Laboratory (DIMPL), Division of Biokinesiology & Physical Therapy, University
of Southern California. Infants were undressed, placed supine on the central testing table, and
secured to the table using a 4-inch Velcro band placed across the trunk and around the table
(Figure 3.1). The midline position of the head was maintained using a horseshoe-shaped support
pillow surrounding the infant’s head. Throughout data collection, the experimenter and parent
were in visual and vocal contact with the infant to maintain the infant in an alert state.

Figure 3.1 Experimental set-up.

Kinematic data. Three-dimensional lower extremity time-position data were collected
at 100 Hz using an Optotrak Certus Motion Capture System (Northern Digital Inc., Waterloo,
ON, Canada) with two sensor banks which are sensitive to the position of infrared light emitting
diodes (IREDs). Each Optotrak sensor bank consisted of three position sensors connected in
series to a System Control Unit and a Dell Precision 690 computer.
INFANT LEG COORDINATION 38
The two Optotrak sensor banks were placed horizontally approximately 2.5 meters on
opposite sides of a central testing table. The root-mean-squared (RMS) error in the calibration of
the sensor banks was less than 0.3mm for each data collection session. A global coordinate
system was defined in relation to the central testing table with the x-axis parallel to the width of
the table, y-axis parallel to the length of the table, and z-axis perpendicular to the table.
Dynamic kicking trial. Rigid marker arrays (RMAs) with 4 embedded IREDs were
attached bilaterally to the foot, shank, thigh, and pelvis using Velcro straps. A small plastic array
with 2 embedded IREDs was placed on the sternum using a double-sided sticky EKG collar.
Leg action was collected for 2 or 3 trials of 5 minutes each.
Static leg calibration trial. After the dynamic kicking trials, a static calibration trial was
collected for each leg. This trial was necessary to define a local/segment coordinate system,
align it with the global coordinate system, and define the orientation of each body segment in
space. Ten individual IREDs were fixed bilaterally to the infant’s skin using double-sided sticky
EKG collars at the following locations: lateral midline of the trunk below the tenth rib, greater
trochanter of the hip, lateral knee joint line, ankle lateral malleolus, and distal end of the 5
th

metatarsal. The static calibration trial for each leg was collected by holding the infant’s lower
extremity in an extended, anatomical position for 5 seconds. All joint angles in this calibration
position were defined as zero degrees.
Video recording. Infants were video recorded with 3 video cameras that surrounded the
testing table (Basler Pylon IEEE1394 cameras using Streampix5 x64 edition multi-camera
software) with a right lateral, left lateral, and overhead view of the infant. Video data were
synchronized with the kinematic data.
INFANT LEG COORDINATION 39
Anthropometric data. Each infant was weighed on a digital electric scale (Health-o-
meter). The total length of the infant was measured and for both legs the following measures
were recorded: circumference at mid-segment of thigh, shank, and foot; width of the knee (at the
knee joint line), ankle (at the malleoli), and foot (at the metatarsal heads); and length of the thigh
(greater trochanter to knee joint line), shank (knee joint line to lateral malleolus), and foot
(medial malleolus to first metatarsophalangeal joint).
Data Reduction
Position data. Position data were converted into 3D coordinates with a direct linear
transformation algorithm using Optotrak system software. A custom Matlab (The Mathworks,
Inc., Natick, MA) program was used to: (1) interpolate missing position data (maximum of 20
consecutive frames) using a cubic spline, (2) filter position data using a fourth-order Butterworth
with a cut-off frequency of 5 Hz as determined from power spectrum analysis, (3) compute joint
angles of hip flexion/extension, hip abduction/adduction, hip external/internal rotation, knee
flexion/extension, ankle dorsiflexion/plantarflexion, ankle eversion/inversion, and (4) extract
kicks. A kick start was defined as the onset of a continuous leg movement for which: (a) the
infant’s foot moved for at least five consecutive frames (1 frame = 10ms), and (b) the hip joint
angle change exceeded 11.5° (2 radians) into flexion (Chen et al., 2002; Fetters et al., 2010;
Jensen et al., 1994; Schneider et al., 1990). The end of the kick was defined as the frame of peak
extension amplitude following a flexion movement (Chen et al., 2002; Fetters et al., 2010). All
kicks that met criteria were included in the kinematic analysis. Only kicks that did not contact
any object, including the non-kicking limb or the table, were included in the kinetic analysis.
Joint angles. Joint angles were computed for the hip, knee, and ankle using the method
of Söderkvist and Wedin (Soderkvist & Wedin, 1993) at the following time points: (a) kick
INFANT LEG COORDINATION 40
initiation, (b) peak hip flexion velocity, (c) peak hip flexion angle referred to as hip joint
reversal, (d) peak hip extension velocity, and (e) kick end (Figure 3.2). These were chosen
because they are points of either movement initiation or change in direction of movement where
there is potentially a change in forces to control the limb. Joint angles at these five time points
were used either as dependent measures or to compute additional measures.
Joint coordination. Joint coordination was defined through the analysis of joint angle
correlations and relative phase.
Joint angle correlations. Joint angle correlations were computed using Pearson
correlation coefficients (r) at zero lag between hip, knee, and ankle joint angle excursions for all
kicks extracted for each infant. For each kick, correlations were computed for joint angle
relations within a joint, i.e., hip flexion/extension-hip abduction/adduction, and interjoint
correlations were computed for motions between two joints, i.e. hip flexion/extension-knee
flexion/extension. All joint angle correlations were converted to Fisher Z scores to allow
comparison of correlations (r) among infants (Chen et al., 2002; Fetters et al., 2004; Fetters et al.,
2010; Jensen et al., 1994).
Relative phase. Relative phase describes the phase relations between two joint motions.
For each kick, joint angle data was time-normalized and continuous relative phase (CRP) was
computed from the angular position/velocity data for joint angle coordination within a joint, i.e.
hip flexion/extension-hip abduction/adduction, and joint angle coordination between two joints,
i.e. hip flexion/extension-knee flexion/extension after the method of van Emmerick and
Wagenaar (Kelso, Scholz, & Schoner, 1986; van Emmerick & Wagenaar, 1996). We then
analyzed results of the CRP computation at the five time points specified above under “Joint
Angles”. Values approaching zero indicate more in-phase coordination; values approaching
INFANT LEG COORDINATION 41

0 20 40 60 80 100 120 140
0
10
20
30
40
50
60
70
80
Angle (degrees)
Angle Data
Frames (100/sec)
2
1
3
5
4
1. kick initiation
2. peak hip flexion velocity
3. hip joint reversal
4. peak hip extension velocity
5. kick end
Hip
Knee
Ankle
Figure 3.2 Five time points in each kick.

INFANT LEG COORDINATION 42
±180° indicate more out-of-phase coordination. A positive value indicates that the proximal
segment is leading the distal segment in phase space; a negative value indicates that the distal
segment is leading the proximal segment. Since we were interested in the magnitude of out-of-
phase coordination, we analyzed the absolute value of the CRP at each discrete point.
Joint torques. NET and partitioned (MUS, GRA, MDT) joint torques were computed
using a Lagrangian approach and evaluated at the five discrete points indicated above. This
approach uses the biomechanic equation of motion, 3D kinematic data, and body-segment
inertial parameters to estimate the forces and torques that cause limb motion.
Kinematic data. Each lower extremity was modeled as three interconnected rigid links
(thigh, shank, foot) with frictionless joints at the hip, knee, and ankle. Joint velocities and
accelerations were determined numerically by differentiating the joint angles using finite
differences.
Body-segment inertial parameters. Anthropometric data were used to compute body-
segment inertial parameters, including estimates of the mass, center-of-mass, and moments of
inertia of each thigh, shank, and foot segment. The segmental mass and center-of-mass were
computed from equations modified for infants from Hatze’s anthropometric model for adults
(Schneider & Zernicke, 1992). The 3D moments of inertia of each thigh, shank, and foot
segment were computed from equations modified for infants from Jensen’s anthropometric
model for adults (Sun & Jensen, 1994).
Biomechanic equation of motion.
The biomechanic equation of motion
( )
̈ (
̇)
̇ + N( ) = T

(1)
INFANT LEG COORDINATION 43
was derived using the Lagrangian approach, where M ( ) is the mass or inertia matrix, C(
̇ )
represents the MDT torques, and N( ) the GRA torques. These terms were calculated using the
screw theory of spatial manipulations (Murray, Li, & Sastry, 1994). This allowed determining
the applied torques T generated by both active muscle contractions and passive elastic effects
from tendons and ligaments.
These terms were separated into torques that are directly responsible for motions at a
single joint and torques that arise from the mechanical effects of the linkage between the
different joints, following Galloway and Koshland (Galloway & Koshland, 2002). Torques in
the first group are proportional to the joint accelerations, where the factors are the diagonal
elements of the inertia matrix
NET = M
dia
( )
̈

. (2)
The second group consists of MDT torques, which comprise terms depending upon both joint
velocities and accelerations. The latter depend upon the off-diagonal elements in the inertia
matrix M
off

MDT = -M
off
( )
̈

- (
̇)
̇ . (3)
The remaining two groups
GRA = -N, MUS = T (4)
are GRA torques and applied torques and correspond directly to the terms in the equation of
motion. Note that for GRA and MDT the sign was changed to make these terms directly
comparable with MUS.
Relation between knee joint dynamics and changes in hip-knee joint coordination.
In order to understand the joint dynamics associated with changes in joint coordination, we focus
on the coordination of the hip and knee into flexion and extension. Hip flexion-extension was
INFANT LEG COORDINATION 44
chosen because it defined each kick, allowing for consistency of the motion analyzed across
subjects and trials. Knee flexion-extension was chosen because it did not occur within the hip
joint reducing the potential confounding effects of intrajoint motion, and it did not occur within
the ankle joint reducing the concern that the ankle would already be demonstrating less in-phase
coordination with the hip by 6-weeks of age.
We quantified the contributions of each partitioned knee (MUS, GRA, MDT) torque to
knee NET torque and computed two variables: percent interval and torque impulse (Sainburg &
Kalakanis, 2000; Tseng, Scholz, & Galloway, 2009). The percent interval was defined as the
percent of the total duration of the kick cycle that each partitioned knee torque contributed to the
knee NET torque. The torque impulse was defined as the magnitude of the contribution of each
partitioned knee torque to knee NET torque.
Percent interval. First, we analyzed the duration of the contribution of each partitioned
knee torque to the knee NET torque. For this analysis, we analyzed the sign of the partitioned
torques throughout the entire kick cycle. Intervals during which the sign of the knee MUS,
GRA, or MDT torques were the same or opposite to that of the knee NET torque were
considered to make a positive or a negative contribution, respectively, to the knee NET torque.
We determined the percentage of positive contribution intervals relative to the total kick cycle.
Torque impulse. Second, we analyzed the magnitude of the contribution of each
partitioned knee torque to the knee NET torque. For this analysis, we computed the positive or
negative torque impulse (torque * time) during intervals in which the knee MUS torque acted in
the same or opposite direction compared with the knee NET torque. Knee GRA and MDT
torque impulse were likewise computed as a contribution to knee NET torque. All positive and
INFANT LEG COORDINATION 45
negative impulses were summed for each torque component to yield a measure of total knee
MUS, GRA and MDT torque impulse.
Percent interval and torque impulse across ages. We analyzed the extent to which the
coordination of the kick, as defined by the Z-transformed correlation coefficient, was dependent
on the percent interval and torque impulse of the MUS and MDT torques. All dependent
variables were computed using Matlab (The Mathworks, Inc., Natick, MA).
Statistical Analysis
Mixed regression models were used for each dependent variable between ages (6-weeks,
15-weeks). Dependent variables included 15 joint correlations, 75 relative phase relations, 15
NET torques, 15 MUS torques, 15 GRA torques, 15 MDT torques, 3 contributions of partitioned
torque to NET torque in terms of percent interval, 3 contributions of partitioned torque to NET
torque in terms of torque impulse.
Linear regression was used to assess the extent to which the coordination of the kick, as
defined by the Z-transformed correlation coefficient, was dependent on the percent interval and
torque impulse of the MUS and MDT torques. The alpha level was set at 0.05 for overall F
values and adjusted using a Bonferroni correction for preplanned post hoc comparisons. SAS
(version 7.0, SAS Institute Inc.) was used for all statistical analysis.
Results
Joint Coordination
We first investigate whether all joint angle pairs change from a predominately in-phase
coordination pattern at 6-weeks to an increased variety of joint coordination patterns at 15-
weeks. We define coordination using joint angle correlation coefficients and relative phasing
between joint angle pairs.
INFANT LEG COORDINATION 46
Joint angle correlations. We define a predominately in-phase coordination pattern as a
positive correlation between joint angle pairs. We define an increase in the variety of joint
coordination patterns when the value of correlations between joint angle pairs approaches a zero
correlation. Joint correlation data are graphed in Figure 3.3 and included in Table 3.2. At 6-
weeks of age, paired joint angle correlations were all positive ranging from 0.07 to 1.63.
Generally, correlations which included hip external/internal rotation or ankle eversion/inversion
were positive (0.07-0.53); all other joint correlations were also positive but with higher
correlations (0.68-1.63). At 15-weeks of age, joint correlations ranged from -0.32 to 0.44, except
for hip external/internal rotation-hip abduction/adduction (1.00) and hip flexion/extension-knee
flexion/extension (0.60). All joint correlations moved toward a zero correlation between 6 and
15-weeks of age (11 joint correlations, adjusted p<0.002), except hip external/internal rotation-
hip abduction/adduction and external/internal rotation-hip flexion/extension which became more
positively correlated (adjusted p<0.05) and 2 other correlations associated with hip
external/internal rotation which did not change (adjusted p>0.05).
These data suggest that at 6-weeks of age joint angle pairs which include hip
external/internal rotation or ankle eversion/inversion have lower correlation coefficients than the
other joint angle pairs. From 6 to 15-weeks, the 10 joint correlation coefficients that include
motion in the sagittal (flexion/extension) and frontal (abduction/adduction, eversion/inversion)
planes consistently moved toward a zero correlation. The 5 joint angle pairs that include hip
external/internal rotation (transverse plane motion) generally do not demonstrate a consistent
change (2 of 5 joint correlations remained the same, 2 increased, and 1 decreased) from 6 to 15-
weeks of age.

INFANT LEG COORDINATION 47
A. 6-weeks B. 15-weeks

1. ↓ hip flexion / extension - knee flexion / extension *
2. ↓ hip flexion / extension - ankle dorsiflexion / plantarflexion*
3. ↓ knee flexion / extension - ankle dorsiflexion / plantarflexion*
4. ↓ hip flexion / extension - hip abduction / adduction*
5. ↓ hip abduction / adduction - knee flexion / extension*
6. ↑ hip external rotation / internal rotation - hip abduction / adduction*
7. ↓ hip abduction / adduction - ankle dorsiflexion / plantarflexion*
8. ↓ ankle eversion / inversion - ankle dorsiflexion / plantarflexion*
9. ↓ hip abduction / adduction - ankle eversion / inversion*
10. ↓ knee flexion / extension - ankle eversion / inversion*
11. ↓ hip flexion / extension - ankle eversion / inversion*
12. hip external / internal rotation - ankle eversion / inversion
13. hip external / internal rotation - ankle dorsiflexion / plantarflexion
14. ↓ hip external / internal rotation - knee flexion / extension*
15. ↑ hip external / internal rotation - hip flexion / extension*

Figure 3.3 Full-term infants: joint correlations. Mean joint correlations (Pearson’s r converted
to Fisher Z). Error bars represent standard errors. Figure A: 6-weeks n=2183 kicks. Figure B:
15-weeks n=2190 kicks.
* adjusted p < 0.05; ↑ and ↓ indicate direction of difference.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
- 2. 00
- 1. 20
- 0. 40
0. 40
1. 20
2. 00
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
- 2. 00
- 1. 20
- 0. 40
0. 40
1. 20
2. 00
INFANT LEG COORDINATION 48
Table 3.2

Full-Term Infants: Joint Correlations
Joint Angle Pairs

6 weeks
n=2183
M (SE)
15 weeks
n=2190
M (SE)
Adjusted

p
Correlation coefficients decreased
Hip flexion / extension
- knee flexion / extension
1.63 (0.10) 0.60 (0.10)* <0.002
Hip flexion / extension
- ankle dorsiflexion / plantarflexion
1.15 (0.08) 0.44 (0.08)* <0.002
Knee flexion / extension
- ankle dorsiflexion / plantarflexion
1.07 (0.10) 0.24 (0.10)* <0.002
Hip flexion / extension
- hip abduction / adduction
0.87 (0.09) 0.37 (0.09)* <0.002
Hip abduction / adduction
- knee flexion / extension
0.84 (0.05) 0.13 (0.05)* <0.002
Hip abduction / adduction
-ankle dorsiflexion / plantarflexion
0.68 (0.07) 0.16 (0.07)* <0.002
Ankle eversion / inversion
-ankle dorsiflexion / plantarflexion
0.53 (0.08) 0.27 (0.08)* <0.002
Ankle eversion / inversion
-hip abduction / adduction
0.41 (0.03) 0.12 (0.03)* <0.002
Ankle eversion / inversion
- knee flexion / extension
0.32 (0.06) -0.22 (0.06)* <0.002
Ankle eversion / inversion
- hip flexion / extension
0.31 (0.04) -0.16 (0.04)* <0.002
Hip external / internal rotation
- knee flexion / extension
0.09 (0.06) -0.32 (0.06)* <0.002
Correlation coefficients increased
Hip external / internal rotation
- hip abduction / adduction
0.83 (0.13) 1.00 (0.13)* <0.01
Hip external / internal rotation
- hip flexion / extension
0.07 (0.10) 0.20 (0.10)*

<0.05
No change in correlation coefficients
Hip external / internal rotation
- ankle eversion / inversion
0.28 (0.03) 0.18 (0.03)

0.06
Hip external / internal rotation
- ankle dorsiflexion / plantarflexion
0.26 (0.06) 0.20 (0.06) 0.90

Note. Joint correlations (Pearson’s r converted to Fisher Z). SE=standard error.
* = adjusted p < 0.05
INFANT LEG COORDINATION 49
Relative phase. We define a predominately in-phase coordination pattern as a value of
relative phase between joint angle pairs close to 0°. We describe an increase in the variety of
joint coordination patterns when the value of relative phase between joint angle pairs increases
toward 180°. Relative phase data are included in Table 3.3. Relative phase at each of the five
data points (kick initiation, peak hip flexion velocity, hip joint reversal, peak hip extension
velocity, and kick end) ranged from 19.8 to 112.8 at 6-weeks and 51.8 to 117.6 at 15-weeks.
Comparing relative phase relations at 6 and 15-weeks, 11 of 15 relative phase relations exhibited
a significant increase in at least 3 of the 5 data points (adjusted p<0.01). Four relative phase
relations, all which included hip external/internal rotation, demonstrated no change in at least 3
of the 5 data points (adjusted p>0.05).
Relative phase results were consistent with joint correlation results. The 11 joint
correlations that moved toward a zero correlation from 6-weeks to 15-weeks, were the same 11
relative phase relations that exhibited a significant increase in magnitude in at least 3 of the 5
data points. The 4 joint correlations involving hip external/internal rotation that either stayed the
same or increased from 6-weeks to 15-weeks, were the same 4 relative phase relations that
demonstrated no change in 3 or more of the 5 data points.
Joint Torques
We investigate the intersegmental dynamics of infant spontaneous kicks between 6 and
15-weeks of age by comparing NET joint torques and partitioned torques (GRA, MUS, MDT) at
five discrete data points within each kick (kick initiation, peak hip flexion velocity, hip joint
reversal, peak hip extension velocity, and kick end). NET joint torque and partitioned joint
torque data normalized to the mass of the leg of each individual participant are included in Table
3.4. At 15-weeks as compared to 6-weeks, kick initiation had greater hip NET torque
INFANT LEG COORDINATION 50
Table 3.3
Full-Term Infants: Relative Phase
Joint Angle Pair Kick
Initiation

M (SE)
Peak Flexion
Velocity

M (SE)
Hip Joint
Reversal

M (SE)
Peak
Extension
Velocity
M (SE)
Kick
End

M (SE)

Relative phase increases in at least 3 of 5 data points

hip flex/ext-knee flex/ext
6 weeks
15 weeks

28.3 (3.9)
67.6 (3.9)*

31.8 (3.6)
65.7 (3.6)*

21.6 (3.4)
62.4 (3.4)*

19.8 (4.1)
62.8 (4.0)*

32.3 (3.4)
70.0 (3.4)*
hip flex/ext-ankle DF/PF
6 weeks
15 weeks

47.2 (3.3)
75.8 (3.3)*

42.4 (2.1)
67.4 (2.1)*

34.9 (2.9)
65.1 (2.9)*

38.4 (3.8)
69.5 (3.8)*

55.8 (4.0)
79.4 (4.0)*
knee flex/ext-ankle DF/PF
6 weeks
15 weeks

47.2 (3.8)
80.7 (3.8)*

50.5 (3.4)
84.0 (3.4)*

40.3 (4.0)
81.8 (3.9)*

39.3 (4.3)
79.4 (4.3)*

50.9 (3.9)
82.0 (3.8)*
hip flex/ext-hip abd/add
6 weeks
15 weeks

47.1 (4.2)
72.4 (4.2)*

61.5 (3.7)
78.1 (3.7)*

60.6 (3.4)
77.8 (3.4)*

51.1 (3.5)
73.9 (3.5)*

51.0 (2.9)
73.7 (2.9)*
hip abd/add - knee flex/ext
6 weeks
15 weeks

47.8 (2.5)
82.9 (2.5)*

74.1 (3.0)
90.4 (3.0)*

61.4 (2.5)
86.1 (2.4)*

44.7 (2.4)
81.4 (2.3)*

46.2 (2.6)
81.0 (2.5)*
hip abd/add-ankle DF/PF
6 weeks
15 weeks

62.5 (3.0)
84.4 (3.0)*

67.8 (3.3)
85.0 (3.3)*

61.3 (2.8)
81.9 (2.8)*

54.5 (3.4)
81.2 (3.4)*

62.9 (2.9)
83.1 (2.8)*
ankle ev/inv-ankle DF/PF
6 weeks
15 weeks

69.6 (3.7)
77.7 (3.7)

69.5 (3.5)
78.0 (3.4)**

67.1 (3.4)
79.3 (3.4)*

67.1 (3.4)
81.2 (3.3)*

66.6 (3.5)
78.3 (3.4)*
ankle ev/inv-hip abd/add
6 weeks
15 weeks

77.0 (1.6)
87.8 (1.6)*

72.8 (2.4)
85.2 (2.3)*

69.1 (2.1)
82.7 (2.0)*

69.8 (1.5)
85.1 (1.5)*

74.3 (2.2)
83.5 (2.1)**
ankle ev/inv-knee flex/ext
6 weeks
15 weeks

74.5 (2.8)
100.8 (2.7)*

86.9 (3.0)
101.4 (2.9)*

77.0 (2.6)
101.8 (2.6)*

74.8 (2.3)
102.4 (2.3)*

73.3 (3.2)
97.7 (3.1)*
ankle ev/inv-hip flex/ext
6 weeks
15 weeks

77.1 (2.1)
95.7 (2.0)*

78.9 (1.9)
98.4 (1.9)*

77.6 (2.0)
100.6 (1.9)*

80.1 (2.4)
104.4 (2.4)*

82.2 (2.9)
95.0 (2.9)*
hip ER/IR - knee flex/ext
6 weeks
15 weeks

77.0 (3.5)
98.6 (3.4)*

112.8 (3.7)
117.6 (3.7)

96.5 (3.9)
108.9 (3.9)*

75.1 (3.7)
99.4 (3.7)*

68.8 (2.6)
94.1 (2.5)*

(continued)
INFANT LEG COORDINATION 51
Joint Angle Pair Kick
Initiation

M (SE)
Peak Flexion
Velocity

M (SE)
Hip Joint
Reversal

M (SE)
Peak
Extension
Velocity
M (SE)
Kick
End

M (SE)

Relative phase does not change consistently in at least 3 of 5 data points

hip ER/IR-hip abd/add
6 weeks
15 weeks

58.8 (5.2)
54.8 (5.1)

65.0 (5.3)
55.2 (5.2)*

52.4 (4.5)
51.8 (4.4)

50.0 (4.8)
52.6 (4.7)

51.0 (4.5)
52.5 (4.5)
hip ER/IR-ankle ev/inv
6 weeks
15 weeks

84.2 (2.3)
83.8 (2.2)

84.1 (2.4)
80.8 (2.3)

72.5 (1.8)
77.3 (1.7)

67.4 (1.6)
80.6 (1.5)*

74.6 (1.4)
82.4 (1.4)
hip ER/IR-ankle DF/PF
6 weeks
15 weeks

79.4 (3.2)
82.5 (3.2)

93.7 (4.2)
83.7 (4.1)**

81.3 (3.3)
78.7 (3.2)

68.3 (3.1)
76.4 (3.0)**

69.3 (2.3)
80.3 (2.2)*
hip ER/IR-hip flex/ext
6 weeks
15 weeks

79.4 (5.1)
75.9 (5.0)

106.7 (5.9)
91.3 (5.9)*

92.9 (4.6)
88.3 (4.5)

79.9 (4.9)
80.8 (4.9)

75.4 (4.8)
76.8 (4.8)

Note. Abd = abduction; add = adduction; DF=dorsiflexion; ER=external rotation; ev = eversion;
ext=extension; flex=flexion; IR = internal rotation; inv = inversion; PF=plantarflexion; SE = standard
error.
* adjusted p < 0.005
** adjusted p < 0.05

INFANT LEG COORDINATION 52
Table 3.4

Full-Term Infants: Normalized Torques

Kick
Initiation

M (SE)
Peak Flexion
Velocity

M (SE)
Hip Joint
Reversal

M (SE)
Peak
Extension
Velocity

M (SE)
Kick
End

M (SE)
Net Joint Torque (Nm/kg)
Hip 6 weeks
15 weeks
0.30 (0.05)
0.40 (0.05)↑
0.10 (0.01)
0.09 (0.01)
0.33 (0.03)
0.32 (0.03)
0.08 (0.01)
0.08 (0.01)
0.26 (0.03)
0.24 (0.03)
Knee 6 weeks
15 weeks
0.05 (0.01)
0.03 (0.01)↓
0.03 (0.00)
0.02 (0.00)
0.03 (0.00)
0.02 (0.00)↓
0.02 (0.00)
0.02 (0.00)
0.03 (0.01)
0.02 (0.01)
Ankle 6 weeks
15 weeks
0.00 (0.00)
0.00 (0.00)
0.00 (0.00)
0.00 (0.00)
0.00 (0.00)
0.00 (0.00)
0.00 (0.00)
0.00 (0.00)
0.00 (0.00)
0.00 (0.00)

Muscle Torque (Nm/kg)
Hip 6 weeks
15 weeks
1.25 (0.05)
1.28 (0.05)
0.75 (0.04)
0.74 (0.04)
0.34 (0.03)
0.39 (0.03)
0.68 (0.04)
0.68 (0.04)
1.10 (0.04)
1.01 (0.04)
Knee 6 weeks
15 weeks
0.32 (0.01)
0.25 (0.01)↓
0.12 (0.01)
0.11 (0.01)
0.15 (0.02)
0.12 (0.02)
0.10 (0.01)
0.10 (0.01)
0.25 (0.01)
0.18 (0.01)↓
Ankle 6 weeks
15 weeks
0.01 (0.00)
0.01 (0.00)
0.01 (0.00)
0.01 (0.00)
0.02 (0.00)
0.02 (0.00)↓
0.01 (0.00)
0.01 (0.00)↓
0.01 (0.00)
0.01 (0.00)

Gravitational Torque (Nm/kg)
Hip 6 weeks
15 weeks
0.89 (0.03)
0.86 (0.03)
0.75 (0.04)
0.73 (0.04)
0.48 (0.05)
0.53 (0.05)
0.69 (0.04)
0.66 (0.04)
0.81 (0.03)
0.74 (0.03)
Knee 6 weeks
15 weeks
0.19 (0.01)
0.13 (0.01)↓
0.12 (0.01)
0.10 (0.01)
0.09 (0.01)
0.10 (0.01)
0.10 (0.01)
0.10 (0.01)
0.16 (0.01)
0.11 (0.01)↓
Ankle 6 weeks
15 weeks
0.01 (0.00)
0.01 (0.00)
0.01 (0.00)
0.01 (0.00)
0.02 (0.00)
0.01 (0.00)↓
0.01 (0.00)
0.01 (0.00)↓
0.01 (0.00)
0.01 (0.00)

Motion-Dependent Torque (Nm/kg)
Hip 6 weeks
15 weeks
0.11 (0.01)
0.09 (0.01)
0.07 (0.01)
0.07 (0.01)
0.09 (0.01)
0.08 (0.01)
0.05 (0.01)
0.07 (0.01)
0.07 (0.01)
0.07 (0.01)
Knee 6 weeks
15 weeks
0.09 (0.01)
0.11 (0.01)
0.01 (0.00)
0.02 (0.00)
0.08 (0.01)
0.08 (0.01)
0.02 (0.00)
0.02 (0.00)
0.07 (0.01)
0.07 (0.01)
Ankle 6 weeks
15 weeks
0.01 (0.00)
0.01 (0.00)
0.00 (0.00)
0.00 (0.00)
0.00 (0.00)
0.01 (0.00)↑
0.00 (0.00)
0.00 (0.00)↑
0.01 (0.00)
0.01 (0.00)

Note. 6-weeks (n=544 kicks) and 15-weeks (n=431 kicks). SE=standard error.
↑ ↓ = adjusted p < 0.05 arrow indicates direction of 15 weeks as compared to 6 weeks

INFANT LEG COORDINATION 53
[F
(1,7)=
19.27, adjusted p=0.045], but no other differences between ages at other data points were
found in hip NET or partitioned torques. Knee NET torque at kick initiation [F
(1,7)=
31.45,
adjusted p=0.015] and hip joint reversal [F
(1,7)=
25.43, adjusted p=0.03] were lower at 15-weeks
as compared to 6-weeks, due to lower knee MUS and GRA torques. Ankle NET torques were
not different between ages, however, from hip joint reversal to kick end, ankle MDT torques
were higher and ankle MUS and GRA torques were lower at 15-weeks as compared to 6-weeks.
These data suggest that, unlike joint coordination, there is not an obvious developmental
change of net joint torque and partitioned torque profiles between 6 and 15-weeks of age.
Relation between Knee Joint Dynamics and Changes in Hip-Knee Joint Coordination
In order to investigate differences in joint dynamics which contribute to changes in joint
coordination, we focus on hip and knee flexion and extension. We quantified the contributions
of knee partitioned (MUS, GRA, and MDT) torques to knee NET torque in terms of duration
(percent interval) and magnitude (torque impulse) (Sainburg & Kalakanis, 2000; Tseng et al.,
2009). Knee torque data are included in Figures 3.4 and 3.5.
Percent interval and torque impulse at 6-weeks. At 6-weeks, the normalized mean
hip-knee flexion/extension correlation coefficient was 1.75 (SE=0.12). The knee MUS torque
contributed to knee NET torque for a mean of 74% of the kick cycle, whereas knee MDT torque
contributed 29% and GRA torque contributed 35%. The knee MUS impulse (0.03 Nm sec), on
average, was in the same direction as the knee NET torque, whereas the knee MDT (-0.01 NM
sec) and GRA (-0.02 Nm sec) impulses, on average, were in the opposite direction of the knee
NET torque.

INFANT LEG COORDINATION 54

Figure 3.4 Full-term infants: mean partitioned torque percent interval contribution to net
knee torque. Error bars represent standard errors.

Figure 3.5 Full-term infants: mean partitioned torque impulse contribution to net knee
torque. Error bars represent standard errors.

0
10
20
30
40
50
60
70
80
90
100
6 w eeks
15 w eeks
- 0. 20
- 0. 10
0. 00
0. 10
0. 20
0. 30
0. 40
( E- 1)
6 w eeks
15 w eeks
Percent Interval of Kick Cycle (%)
Muscle Motion-Dependent
Gravitational
Torque Impulse (Nm-s)
Muscle Motion-Dependent Gravitational
0.04
.
0.03
.
0.02
.
0.01
.
0.00
.
-0.01
.
-0.02
.
INFANT LEG COORDINATION 55
Percent interval and torque impulse at 15-weeks. At 15-weeks, the normalized mean
hip-knee flexion/extension correlation coefficient was 0.32 (SE=0.12); a significant decrease
from 6-weeks [F
(1,7)
= 261.17, adjusted p<0.001]. The knee MUS torque contributed to knee
NET torque for an average of 60% of the kick cycle, significantly less than at 6-weeks [F
(1,7)
=
81.44, adjusted p<0.001]. The knee MDT contributed 44% and GRA torque contributed 45%,
significantly more than at 6-weeks [F
(1,7)
= 80.09, adjusted p<0.001; F
(1,7)
= 54.41, adjusted
p<0.001]. The contribution of torque impulse was similar to 6-weeks; the knee MUS impulse
(0.03 Nm sec), on average, was in the same direction as the knee NET torque, whereas the knee
MDT (-0.01) and GRAV (-0.01) impulses, on average, were in the opposite direction of the knee
NET torque. The knee GRA impulse was significantly increased at 15-weeks as compared to 6-
weeks [F
(1,7)
= 36.31, adjusted p<0.002], which means that knee GRA impulse was, on average,
in the opposite direction of the knee NET torque at a lower magnitude. There was no difference
in the MDT [F
(1,7)
= 3.53, adjusted p=0.30] impulses, however knee MUS impulses trended
towards a decrease at 15-weeks as compared to 6-weeks [F
(1,7)
= 7.86, adjusted p=0.08], which
means that MUS impulse was in the same direction as the knee NET torque at a lower
magnitude.
Percent interval and torque impulse across ages.
We examined the extent to which the hip-knee coordination of the kicks, as assessed
using Z-transformed hip-knee correlation coefficients across ages, was dependent on the percent
interval and torque impulse of the knee MUS and MDT torques (Figure 3.6, Figure 3.7). For all
kicks between ages, the Z-transformed correlation coefficient was highly dependent on the
percent interval of the MUS and MDT torque contribution to the knee NET torque, accounting
INFANT LEG COORDINATION 56
A.

B.

Figure 3.6 Full-term infants: relation between kick coordination and percent interval across
ages. Figure A: knee muscle torque contribution to knee net torque. Figure B: knee motion
dependent torque contribution to knee net torque.

Z-transformed Correlation Coefficient
Percent interval knee motion dependent torque contribution to knee net torque
Z-transformed Correlation Coefficient
Percent interval knee muscle torque contribution to knee net torque
R
2
=0.32
R
2
=0.46
INFANT LEG COORDINATION 57
A.

B.

Figure 3.7 Full-term infants: relation between kick coordination and torque impulse across ages.
Figure A: knee muscle torque contribution to knee net torque. Figure B: knee motion dependent
torque contribution to knee net torque.

Knee muscle torque impulse contribution to knee net torque (Nm-sec)
Z-transformed Correlation Coefficient Z-transformed Correlation Coefficient
Knee motion dependent torque impulse contribution to knee net torque (Nm-sec)
R
2
=0.23
R
2
=0.23
INFANT LEG COORDINATION 58
for 32 and 46% of the variance respectively [F
1,973
=453.49, adjusted p<0.001; F
1,973
=828.26,
adjusted p<0.001]. For kicks with a high positive correlation, the MUS torques contributed to
NET torque for a relatively high proportion of the kick cycle, whereas the MDT torque
contributed for a relatively low proportion of the kick cycle. For the kicks with a high negative
correlation, the MUS torque contributed for a low proportion of the kick cycle, whereas the
MDT torque contributed for a relatively high proportion of the kick cycle.
For all kicks between ages, the Z-transformed correlation coefficient was moderately
dependent on the MUS and MDT torque impulse, with each accounting for 23% of the variance
[F
1,973
=284.06, p<0.0001; F
1,973
=192.61, p<0.0001]. For kicks with a high positive correlation,
the MUS torque impulse was in the same direction as the knee NET torque, whereas the knee
MDT torque impulse was in the opposite direction as the knee NET torque. For the kicks with a
high negative correlation, the MUS torque impulse was in the opposite direction as the knee NET
torque, whereas the knee MDT torque impulse was in the same direction as the knee NET torque.
Taken together, these data suggest that at 6-weeks knee MUS torque, which occurs in the
same direction as knee NET torque and contributes to knee NET torque during a greater duration
of the kick cycle, minimized the knee GRA and MDT torques and contributed to the in-phase hip
flexion/extension-knee flexion/extension coordination (expressed by the Z-transformed
correlation coefficient). By 15-weeks, knee MUS torques were generated during less of the kick
cycle and knee GRA and MDT torques were generated during more of the kick cycle. In
addition, knee MUS torques trended toward contributing to the knee NET torque at a lower
magnitude allowing for the emergence of a greater variety of coordination patterns between hip
flexion/extension and knee flexion/extension.
INFANT LEG COORDINATION 59
These results are consistent with the second hypothesis of our third objective. A greater
variety of joint coordination patterns between the hip and knee joint from 6 to 15-weeks of age is
associated with a decreased influence of knee MUS torques which allow passive knee GRA and
MDT to have a greater influence on the coordination of the kick.
Discussion
Our first objective was to characterize the change in coordination from 6 to 15-weeks of
age. We hypothesized that all joint angle pairs would change from a predominately in-phase
coordination pattern at 6-weeks to an increased variety of joint coordination patterns at 15-
weeks. The hypothesis was confirmed for the 11 joint angle pairs that included motion in the
sagittal (flexion/extension) and frontal (abduction/adduction, eversion/inversion) planes. This
supports previous studies which describe a change in the coordination patterns of spontaneous
kicks over the first months of life from a predominately in-phase coordination pattern of
synchronous flexion and extension of the hip, knee, and ankle joints to a greater variety of hip,
knee, and ankle coordination patterns (Fetters, et al., 2004; Jeng, et al., 2002; Piek, 1996; Thelen,
1985; Vaal, van Soest, Hopkins, et al., 2000; Thelen, 1985; Vaal, van Soest, Hopkins, et al.,
2000). Our findings contribute to this line of inquiry by adding that the coordination of joint
angle pairs which include hip abduction/adduction and ankle eversion/inversion also demonstrate
a greater variety of joint coordination patterns over the first months of life. Contrary to our
hypothesis, we found that joint angle pairs which include hip external/internal rotation
(transverse plane movement) generally did not demonstrate a consistent coordination change
from 6 to 15-weeks of age. Hip external/internal rotation may be more influenced by the
intrajoint motions of hip flexion/extension and abduction/adduction due to hip muscle activations
INFANT LEG COORDINATION 60
resulting in multiple actions at the hip, for example, the gluteus medius, piriformis, and sartorius
all abduct as well as externally rotate the hip.
Our second objective was to identify changes in intersegmental dynamics of infant
spontaneous kicks. We hypothesized that hip NET torques would increase from 6 to 15-weeks
of age as infants gained greater control of their hip motion and hip co-contraction decreased,
whereas knee and ankle NET torques would decrease as muscle co-activation decreased. Our
results suggest that, unlike joint coordination, there is not an obvious developmental change in
joint torques between 6 and 15-weeks of age. The lack of consistent relationship of joint torque
changes between 6 and 15-weeks could be due to differences in the coordination of the kicks
between ages which confound the torque analysis; essentially different kicking motions are being
analyzed. A more focused analysis which stratifies kicks based on kick coordination may be
needed to clarify changes in intersegmental changes across ages.
Our third objective was to describe the relation between changes in joint coordination and
intersegmental dynamics. We hypothesized that an increase in the variety of coordination
patterns between hip-knee flexion and extension would be associated with either: (1) an
increased influence of knee MUS torque to dampen the passive knee GRA and MDT torques
which are associated with more in-phase joint coordination, or (2) a decreased influence of knee
MUS torque which allow passive knee GRA and MDT torque to have a greater influence on the
coordination of the kick, contributing to a greater variety of hip-knee joint coordination patterns.
Our results support the second hypothesis that a greater variety of hip-knee flexion/extension
coordination is associated with a decrease in the duration and magnitude that the knee MUS
torque contributes to the knee NET torque and an increase in the duration and magnitude that the
MDT and GRA torques contribute to the knee NET torque.
INFANT LEG COORDINATION 61
This finding of a strong contribution of MUS torque to the in-phase character of young
infants’ kicks supports the interpretation of previous research using EMG analysis. Thelen and
colleagues suggested that the highly synchronous kicks of 2- and 4-week old infants were
initiated by phasic synchronous firing bursts of the rectus femoris, hamstring, tibialis anterior,
and gastrocnemius-soleus muscles (Thelen & Fisher, 1983b). The authors speculated that this
excessive antagonist co-contraction is characteristic of immature movement and the release of
joints from this obligatory synergism is essential to acquire new coordinative structures (Thelen,
1985; Thelen & Fisher, 1983b). We provide support for this speculation with our finding that as
a greater variety of hip-knee joint coordination emerges at 15-weeks of age, a decrease is
observed in the duration and magnitude that the knee MUS torque contributes to the knee NET
torque, whereas the converse occurs for the passive GRA and MDT torques. We suggest that
between 6 and 15-weeks of age the decrease in the duration and magnitude of the MUS torque
generated at the knee allows the knee passive GRA and MDT torques to have a greater influence
on the coordination of the kick, contributing to a greater variety of hip-knee joint coordination
patterns.
The emergence of a greater variety of joint coordination patterns supports the progress to
more complex movements (Fetters et al., 2010; Thelen & Fisher, 1983b), yet the mechanism
underlying this change in coordination remains an open area of inquiry. Our current results
support the hypothesis that the emergence of a greater variety of hip-knee joint coordination
patterns observed in the spontaneous kicking of young infants is due to a decreased influence of
knee MUS torque which allows for an increased influence of passive knee GRA and MDT
torques. We believe that a decrease in knee and ankle muscle co-activations allows the passive
torques to have a greater influence on the kicking coordination contributing to the greater variety
INFANT LEG COORDINATION 62
of joint coordination patterns. This hypothesis could be tested with the addition of EMG to our
3D kinematic and kinetic analysis. However, perhaps an even more critical next step is the
comparison of two infant leg actions: infant spontaneous kicking and infant goal-directed
kicking. From our current study, we believe that the increased variety of hip-knee joint
coordination patterns observed in the spontaneous kicking of 15-week old infants is associated
with an increased influence of passive knee GRA and MDT torques as compared to knee MUS
torque. However, it is unknown how the intersegmental dynamics change when infants are
actively trying to perform a task which requires more out-of-phase hip-knee joint coordination.
We hypothesize that there would be an increase in knee MUS torque to dampen the passive knee
GRA and MDT torques during the initial stage of learning. This is consistent with Bernstein’s
work in which excessive co-activation is used in the initial stages of learning to reduce the
number of independent joint variables to be controlled (Bernstein, 1967; Sporns & Edelman,
1993). However, as infants interact within task constraints, we expect that infants will decrease
muscle co-activation and exploit the passive GRA and MDT forces to achieve the out-of-phase
leg joint coordination required by the task, but in a more energy-efficient manner.

INFANT LEG COORDINATION 63
CHAPTER IV
CHANGES IN INFANT LEG COORDINATION:
INFLUENCE OF PREMATURITY

Abstract
The purpose of this study is to investigate the contribution of torque changes to the early
changes in leg joint coordination of preterm infants. We analyze preterm infant spontaneous
kicking actions within subjects and between 6 and 15-weeks of age using a three-dimensional
kinematics and kinetic approach and compared the results to kicking actions from full-term
infants. The kinematic analysis supported that preterm infants, similar to full-term infants,
exhibit an increased variety of joint coordinations from 6-weeks to 15-weeks of age. However,
some joint coordinations involving the ankle at 6-weeks of age were less in-phase in preterm
infants as compared to full-term infants. By 15-weeks of age, differences in joint coordination
between full-term and preterm infants had resolved. The kinetic analysis supported that preterm
infants demonstrated similar intersegmental dynamics as full-term infants; net joint torque and
partitioned joint torque data normalized to the mass of the leg of each individual subject
remained relatively consistent from 6 to 15-weeks of age. There were no significant differences
in intersegmental dynamics between full-term and preterm infants. Further analysis supported
that although preterm infants demonstrated significantly less in-phase coordination between the
hip and knee joint from 6 to 15-weeks of age, the contribution of knee muscle, gravitational, and
motion dependent torques to knee NET torque did not change as it did in the full-term group.
However, the preterm infants demonstrated a smaller change in hip-knee coordination from 6 to
INFANT LEG COORDINATION 64
15-weeks which may have been insufficient to document the relation between intersegmental
dynamics and joint coordination.
Introduction
Approximately 500,000 preterm (PT) births occur in the United States each year; this is
twelve percent of total births (Hamilton et al., 2011). PT infants are at high risk for white matter
damage (WMD), which may increase the risk of developing spastic cerebral palsy (CP), a severe
dysfunction of motor coordination (Himpens et al., 2008; Oskoui et al., 2013; Rosenbaum et al.,
2007). Even PT infants who do not sustain WMD are at high risk for impairments in motor
coordination, which increase in severity with decreasing gestational age and birth weight
(Aarnoudse-Moens et al., 2009; de Kieviet et al., 2009; Delobel-Ayoub et al., 2009; Larroque et
al., 2008; Pearsall-Jones et al., 2010). The prevalence of mild-moderate motor impairment in
school-aged children born PT, excluding children with CP, is estimated at 40.5% (Williams, Lee,
& Anderson, 2010).
Early therapeutic intervention is shown to improve motor skill acquisition of PT infants
at risk for impairments in motor coordination (Heathcock & Galloway, 2009; Heathcock et al.,
2008). The plasticity of the brain is considerable during the first months of life which
necessitates that intervention be initiated early, to optimize the neural circuitry which
implements motor coordination (de Graff-Peters & Hadders-Algra, 2006; Lobo et al., 2013;
Ulrich, 2010). The early identification of motor coordination impairments with the potential for
early intervention may be critical to optimize motor skill acquisition during the first months of
life.
Functional coordination patterns of the upper extremities are apparent very early in
development. Infants attempt to coordinate their upper limbs as toys and objects are within sight
INFANT LEG COORDINATION 65
and reach (Bhat et al., 2005; Bhat et al., 2007; Lee et al., 2008). The functional coordination
patterns of the lower extremities do not become apparent until months later when infants start
crawling and creeping. The joint coordination of spontaneous kicking, however, changes
significantly during the early months of life. The spontaneous kicks of newborn infants are
dominated by an in-phase intralimb coordination pattern of synchronous flexion or extension of
the hip, knee, and ankle joints. Within the first months of life, infants demonstrate a greater
variety of intralimb coordination patterns (Fetters et al., 2004; Fetters et al., 2010; Heriza, 1988a;
Jeng et al., 2002; Thelen & Fisher, 1983b; Vaal, van Soest, Hopkins, et al., 2000), which is
thought to support the progress to more functional coordination of the lower extremities
necessary for optimal motor skill acquisition (Fetters et al., 2010; Thelen & Fisher, 1983b).
Hence, the early identification of differences in spontaneous kicking between typically
developing full-term (FT) infants and PT infants at risk for motor coordination impairments may
be necessary to identify those PT infants in need of early therapeutic intervention to promote
more functional coordination patterns of the legs.
Early differences in the spontaneous kicking of FT infants, PT infants, and preterm
infants with white matter damage (PTWMD) have been documented using kinematic analysis.
At one-month corrected-age (CA) PT infants and FT infants demonstrate similar joint
coordination during spontaneous kicking (Fetters et al., 2004). However, in comparison to PT
and FT infants, PTWMD infants, at highest risk for CP, exhibit excessive in-phase joint
coordination (Fetters et al., 2004). Specifically, PTWMD infants tend to flex and extend their
hip, knee, and ankle in more synchrony than FT and PT infants, and demonstrate a decreased
variety of joint coordinations. The finding that PTWMD infants, at highest risk for CP,
demonstrate more in-phase joint coordination was expected since children with spastic CP also
INFANT LEG COORDINATION 66
demonstrate excessive in-phase joint coordination of the lower extremities, which affects their
functional mobility (Fowler & Goldberg, 2009; Ostensjo, Carlberg, & Vollestad, 2004; Sanger et
al., 2006).
By 5 months, FT, PT, and PTWMD infants all generate an increased variety of joint
coordinations (Fetters et al., 2010). However, PT infants in comparison to FT infants
demonstrate a greater variety of joint coordinations, and FT and PTWMD infants demonstrate
similar leg joint coordinations (Fetters et al., 2010). Since it would be expected that PTWMD
infants, at highest risk for CP, would continue to demonstrate more in-phase joint coordination,
this finding is surprising. Also, it appears that PT infants at 5 months are more “advanced” than
FT infants since they exhibit a greater variety of joint coordinations, yet PT infants are at high
risk for impairments in motor coordination. The authors suggest that the effect of practice might
explain both results as the PT and PTWMD infants were included in the research study not based
on their chronological age, but rather on their age “corrected” for their prematurity. Thus, the PT
and PTWMD infants had at least two extra months of additional exposure to the extra-uterine
environment which afforded more time to explore kicking dynamics. In addition, all of the
infants born PT in this study were in early intervention programs, although the extent of
intervention for lower extremity coordination was not determined. The results from this research
highlight that the leg joint coordination which emerges during infant spontaneous kicking is
complex requiring study of the processes which implement changes in joint coordination.
Differences in joint coordination between FT, PT, and PTWMD have focused on the
coordination among the hip, knee, and ankle joints into flexion and extension. Coordination
among joint angle pairs that include the rotational aspects of infant leg motion, such as hip
abduction/adduction, hip external/internal rotation, and ankle eversion/inversion, have not been
INFANT LEG COORDINATION 67
investigated. These rotational movements may provide a critical means of quantifying
differences in leg coordination between typically developing infants and PT infants at risk for
coordination impairments and may be necessary to study the processes which implement the
emergence of a greater variety of leg joint coordinations.
There are differences in the characteristics of spontaneous kicking between FT and PT
infants that suggest that the dynamics of this action may require more in-depth investigation.
Fetters, et al, 2010, found that at 5 months FT infants in comparison to PT infants move with
higher velocities and from a more extended position at the start of the movement. This may
indicate that FT infants in comparison to PT infants are generating more force during
spontaneous kicking, although the methods used in the study precluded an estimate of force
generation.
PT infants may also be managing the forces associated with kicking differently than FT
infants. Infants kick in a gravitational environment which imposes two specific passive torques
on the moving limb: a gravitational (GRA) torque related to gravity acting downward on the
limb and a motion-dependent (MDT) torque related to the mechanical interactions among the
moving interconnected segments of the limb. Even if PT infants demonstrate similar force
generating capabilities as FT infants, they may demonstrate different intersegmental dynamics,
in other words, they may demonstrate differences in the anticipation and control for the effects of
GRA and MDT torques that are generated as a result of their active muscle force generation
(Bernstein, 1967; Jensen et al., 1994; Schneider et al., 1990). Force generating capability and the
management of intersegmental dynamics may be critical factors in the emergence of a greater
variety of leg joint coordinations.
INFANT LEG COORDINATION 68
We do not as yet have an understanding of the dynamics which underlie the emergence of
a greater variety of leg joint coordinations. However, we do know that this is a typical process in
the development of FT infants and that the ability to combine joint action through combinations
of muscle actions is a hallmark of the development of functional movement. The kinematic
analysis of the leg coordination of FT and PT infants provides a description of the changes in
coordination which occur in the first months of life (Fetters et al., 2004; Fetters et al., 2010; Jeng
et al., 2002; Vaal, van Soest, Hopkins, et al., 2000), but a kinematic analysis alone is insufficient
to understand the processes which may support the emergence of a greater variety of joint
coordinations. Kinetic analyses, derived from the kinematics, can be used to quantify the net
(NET) torques that are generated at each joint during infant kicking actions and can further be
decomposed into the active muscle (MUS) torque and passive GRA and MDT torques which are
produced as a result of active muscle force generation (Bernstein, 1967; Jensen et al., 1994;
Schneider et al., 1990). A kinetic analysis may demonstrate how changes in joint coordination
are influenced by differences in NET torque (force-generating capacity) or differences in the way
that MUS torques are used to anticipate and control for the effects of GRA and MDT torques
(intersegmental dynamics) (Bernstein, 1967; Jensen et al., 1994; Schneider et al., 1990).
We studied kinetics derived from kinematics to investigate the contribution of torque
changes to early changes in infant leg joint coordination. In a previous study with FT infants
(Chapter III), we analyzed the coordination of the hip and knee into flexion and extension to
determine the relation between changes in joint coordination and intersegmental dynamics from
6 to 15-weeks of age. At 6-weeks the correlation coefficient (measure of in-phase coordination)
between the hip and knee was 1.75 (SE=0.12). This in-phase coordination was accompanied by
a long duration and high magnitude of knee MUS torques as compared to passive knee MDT and
INFANT LEG COORDINATION 69
GRA torques. At 15-weeks, the correlation coefficient had decreased to 0.32 (SE=0.12). This
less in-phase coordination was accompanied by a shorter duration and lower magnitude of knee
MUS torque as compared to passive knee MDT and GRA torques. We suggest that for FT
infants between 6 and 15-weeks of age the decreased influence of MUS torque generated at the
knee allows passive knee GRA and MDT torques to have a greater influence on the coordination
of the kick, contributing to a greater variety of hip-knee joint coordinations.
We now extend this research to PT infants in order to investigate the relation between
changes in joint coordination and intersegmental dynamics in a population of infants with known
differences in coordination from FT infants. This study has three objectives. The first objective
is to investigate changes in joint coordination between all joint angle pairs from 6 to 15-weeks of
age in PT infants and compare the results to FT infants. We hypothesize that, similar to FT
infants, the majority of joint angle pairs will change from a predominately in-phase coordination
pattern at 6 weeks to an increased variety of joint coordination patterns at 15-weeks. However,
at both 6-weeks and 15-weeks, PT infants, as compared to FT infants, will exhibit less in-phase
coordination between joint pairs. The second objective is to identify changes in intersegmental
dynamics of infant spontaneous kicks from 6 to 15-weeks of age in PT infants and compare the
results to FT infants. We hypothesize that between 6 and 15-weeks of age PT infants will
demonstrate similar intersegmental dynamics as the FT infants; NET joint torque and partitioned
joint torque data normalized to the mass of the leg of each individual participant will remain
relatively consistent from 6 to 15-weeks of age. The third objective is to describe the relation
between changes in coordination and changes in intersegmental dynamics through the analysis of
hip and knee flexion and extension coordination. We hypothesize that, similar to FT infants, a
greater variety of joint coordinations will be associated with a decreased influence of knee MUS
INFANT LEG COORDINATION 70
forces which allow passive knee GRA and MDT to have a greater influence on the coordination
of the kick, contributing to a greater variety of hip-knee joint coordinations.
This study is a unique contribution to the literature since it is the first time that a three-
dimensional (3D) kinematic and kinetic approach has been used to investigate joint coordination
and intersegmental dynamics in a population of infants with known differences in joint
coordination. An understanding of the similarities and differences in joint coordination and
intersegmental dynamics between FT and PT infants may further clarify the processes underlying
the early changes in joint coordination of young infants.
Method
Participants
Seven PT infants participated in this longitudinal study at both 6 and 15-weeks of age (±1
week). This is the developmental period during which a greater variety of leg joint coordinations
emerge and infants kick spontaneously in supine without attempts to roll over. Infants were
excluded from the study based on parent report if they were diagnosed with congenital
malformations, chromosomal abnormalities, prenatal drug exposure, orthopedic impairments,
and visual and hearing impairments. Participant characteristics are listed in Table 4.1. Parents
signed a consent form prior to participation in the study, and families received a small gift for
their participation. The Institutional Review Board at the University of Southern California
approved the study. The data from 10 FT infants from a previous study were used to compare
differences between PT and FT infant groups (Chapter III).
Procedure
Experimental set-up. All testing was completed in the Development of Infant Motor
Performance Laboratory (DIMPL), Division of Biokinesiology & Physical Therapy, University
INFANT LEG COORDINATION 71
Table 4.1

Preterm Infants: Participant Characteristics

Participant
Number
Gender Gestational
Age

weeks
Birth
weight

grams
Age

days
Ponderal
Index

kg/m
3

Bayley III
Motor
Subtest

# Kicks for
Kinematic
Analysis
# Kicks for
Kinetics
Analysis

percentile rank
PT 1 M 28 885 48 24.4 533 20
112 22.9 79 205 22

PT 2 M 28 835 48 21.9 359 55
112 22.7 75 321 23

PT 3 F 34 1559 39 28.5 802 84
116 25.4 68 498 57

PT 4 F 34 1786 39 26.0 695 35
116 26.2 58 272 2

PT 5 M 29 1588 48 25.4 118 20
102 24.4 68 283 35

PT 6 F 33 2268 52 21.1 362 15
104 20.9 84 44 2

PT 7 F 34 2410 50 27.0 126 15
122 26.3 58 115 3
Note: F=female, M= male, PT = preterm
INFANT LEG COORDINATION 72
of Southern California. Infants were undressed, placed supine on the central testing table, and
secured to the table using a 4-inch Velcro band placed across the trunk and around the table. The
midline position of the head was maintained using a horseshoe-shaped support pillow
surrounding the infant’s head (Figure 3.1). Throughout data collection, the experimenter and
parent were in visual and vocal contact with the infant to maintain the infant in an alert state.
Kinematic data. Three-dimensional lower extremity time-position data were collected
at 100 Hz using an Optotrak Certus Motion Capture System (Northern Digital Inc., Waterloo,
ON, Canada) with two sensor banks which are sensitive to the position of infrared light emitting
diodes (IREDs). Each Optotrak sensor bank consisted of three position sensors connected in
series to a System Control Unit and a Dell Precision 690 computer.
The two Optotrak sensor banks were placed horizontally approximately 2.5 meters on
opposite sides of a central testing table. The root-mean-squared (RMS) error in the calibration of
the sensor banks was less than 0.3mm for each data collection session. A global coordinate
system was defined in relation to the central testing table with the x-axis parallel to the width of
the table, y-axis parallel to the length of the table, and z-axis perpendicular to the table.
Dynamic kicking trial. Rigid marker arrays (RMAs) with 4 embedded IREDs were
attached bilaterally to the foot, shank, thigh, and pelvis using Velcro straps. A small plastic array
with 2 embedded IREDs was placed on the sternum using a double-sided sticky EKG collar.
Leg action was collected for 2 or 3 trials of 5 minutes each.
Static leg calibration trial. After the dynamic kicking trials, a static calibration trial was
collected for each leg. This trial was necessary to define a local/segment coordinate system,
align it with the global coordinate system, and define the orientation of each body segment in
space. Ten individual IREDs were fixed bilaterally to the infant’s skin using double-sided sticky
INFANT LEG COORDINATION 73
EKG collars at the following locations: lateral midline of the trunk below the tenth rib, greater
trochanter of the hip, lateral knee joint line, ankle lateral malleolus, and distal end of the 5
th

metatarsal. The static calibration trial for each leg was collected by holding the infant’s lower
extremity in an extended, anatomical position for 5 seconds. All joint angles in this calibration
position were defined as zero degrees.
Video recording. Infants were video recorded with 3 video cameras that surrounded the
testing table (Basler Pylon IEEE1394 cameras using Streampix5 x64 edition multi-camera
software) with a right lateral, left lateral, and overhead view of the infant. Video data were
synchronized with the kinematic data.
Anthropometric data. Each infant was weighed on a digital electric scale (Health-o-
meter). The total length of the infant was measured and for both legs the following measures
were recorded: circumference at mid-segment of thigh, shank, and foot; width of the knee (at the
knee joint line), ankle (at the malleoli), and foot (at the metatarsal heads); and length of the thigh
(greater trochanter to knee joint line), shank (knee joint line to lateral malleolus), and foot
(medial malleolus to first metatarsophalangeal joint).
Data Reduction
Position data. Position data were converted into 3D coordinates with a direct linear
transformation algorithm using Optotrak system software. A custom Matlab (The Mathworks,
Inc., Natick, MA) program was used to: (1) interpolate missing position data (maximum of 20
consecutive frames) using a cubic spline, (2) filter position data using a fourth-order Butterworth
with a cut-off frequency of 5 Hz as determined from power spectrum analysis, (3) compute joint
angles of hip flexion/extension, hip abduction/adduction, hip external/internal rotation, knee
flexion/extension, ankle dorsiflexion/plantarflexion, ankle eversion/inversion, and (4) extract
INFANT LEG COORDINATION 74
kicks. A kick start was defined as the onset of a continuous leg movement for which: (a) the
infant’s foot moved for at least five consecutive frames (1 frame = 10ms), and (b) the hip joint
angle change exceeded 11.5° (2 radians) into flexion (Chen et al., 2002; Fetters et al., 2010;
Jensen et al., 1994; Schneider et al., 1990). The end of the kick was defined as the frame of peak
extension amplitude following a flexion movement (Chen et al., 2002; Fetters et al., 2010). All
kicks that met criteria were included in the kinematic analysis. Only kicks that did not contact
any object, including the non-kicking limb or the table, were included in the kinetic analysis.
Joint angles. Joint angle positions were computed for the hip, knee, and ankle using the
method of Söderkvist and Wedin (Soderkvist & Wedin, 1993) at the following time points: (a)
kick initiation, (b) peak hip flexion velocity, (c) peak hip flexion angle referred to as hip joint
reversal, (d) peak hip extension velocity, and (e) kick end (Figure 3.2). These were chosen
because they are points of either movement initiation or change in direction of movement where
there is potentially a change in forces to control the limb. Joint angles at these five time points
were used either as dependent measures or to compute additional measures.
Joint coordination. Joint coordination was defined through the analysis of joint angle
correlations and relative phase.
Joint angle correlations. Joint angle correlations were computed using Pearson
correlation coefficients (r) at zero lag between hip, knee, and ankle joint angle excursions for all
kicks extracted for each infant. For each kick, correlations were computed for motions within a
joint, i.e. hip flexion/extension-hip abduction/adduction, and interjoint correlations were
computed for motions between two joints, i.e. hip flexion/extension-knee flexion/extension. All
joint angle correlations were converted to Fisher Z scores to allow comparison of correlations (r)
among infants (Chen et al., 2002; Fetters et al., 2004; Fetters et al., 2010; Jensen et al., 1994).
INFANT LEG COORDINATION 75
Relative phase. Relative phase describes the phase relations between two motions. For
each kick, joint angle data was time-normalized and continuous relative phase (CRP) was
computed from the angular position/velocity data for motions within a joint, i.e. hip
flexion/extension-hip abduction/adduction, and motions between two joints, i.e. hip
flexion/extension-knee flexion/extension after the method of van Emmerick and Wagenaar
(Kelso et al., 1986; van Emmerick & Wagenaar, 1996). We then analyzed results of the CRP
computation at the five time points specified above under “Joint Angles”. Values approaching
zero indicate more in-phase coordination; values approaching ±180° indicate more out-of-phase
coordination. A positive value indicates that the proximal segment is leading the distal segment
in phase space; a negative value indicates that the distal segment is leading the proximal
segment. Since we were interested in the magnitude of out-of-phase coordination, we analyzed
the absolute value of the CRP at each discrete point.
Joint torques. NET and partitioned (MUS, GRA, MDT) joint torques were computed
using a Lagrangian approach and evaluated at the five discrete points indicated above. This
approach uses the biomechanic equation of motion, 3D kinematic data, and body-segment
inertial parameters to estimate the forces and torques that cause limb motion.
Kinematic data. Each lower extremity was modeled as three interconnected rigid links
(thigh, shank, foot) with frictionless joints at the hip, knee, and ankle. Joint velocities and
accelerations were determined numerically by differentiating the joint angles using finite
differences.
Body-segment inertial parameters. Anthropometric data were used to compute body-
segment inertial parameters, including estimates of the mass, center-of-mass, and moments of
inertia of each thigh, shank, and foot segment. The segmental mass and center-of-mass were
INFANT LEG COORDINATION 76
computed from equations modified for infants from Hatze’s anthropometric model for adults
(Schneider & Zernicke, 1992). The 3D moments of inertia of each thigh, shank, and foot
segment were computed from equations modified for infants from Jensen’s anthropometric
model for adults (Sun & Jensen, 1994).
Biomechanic equation of motion.
The biomechanic equation of motion
( )
̈ (
̇)
̇ + N( ) = T

(1)
was derived using the Lagrangian approach, where M ( ) is the mass or inertia matrix, C(
̇ )
represents the MDT torques, and N( ) the GRA torques. These terms were calculated using the
screw theory of spatial manipulations (Murray et al., 1994). This allowed determining the
applied torques T generated by both active muscle contractions and passive elastic effects from
tendons and ligaments.
These terms were separated into torques that are directly responsible for motions at a
single joint and torques that arise from the mechanical effects of the linkage between the
different joints, following Galloway and Koshland (Galloway & Koshland, 2002). Torques in
the first group are proportional to the joint accelerations, where the factors are the diagonal
elements of the inertia matrix
NET = M
dia
( )
̈

. (2)
The second group consists of MDT torques, which comprise terms depending upon both joint
velocities and accelerations. The latter depend upon the off-diagonal elements in the inertia
matrix M
off

MDT = -M
off
( )
̈

- (
̇)
̇ . (3)
The remaining two groups
INFANT LEG COORDINATION 77
GRA = -N, MUS = T (4)
are GRA torques and applied torques and correspond directly to the terms in the equation of
motion. Note that for GRA and MDT the sign was changed to make these terms directly
comparable with MUS.
Relation between knee joint dynamics and changes in hip-knee joint coordination.
In order to understand the joint dynamics associated with changes in joint coordination, we focus
on hip and knee flexion and extension. Hip flexion-extension was chosen because it defined
each kick, allowing for consistency of the motion analyzed across subjects and trials. Knee
flexion-extension was chosen because it did not occur within the hip joint reducing the potential
confounding effects of intrajoint motion, and it did not occur within the ankle joint reducing the
concern that the ankle may already be demonstrating less in-phase coordination with the hip by
6-weeks of age.
We quantified the contributions of each partitioned knee (MUS, GRA, MDT) torque to
knee NET torque and computed two variables: percent interval and torque impulse (Sainburg &
Kalakanis, 2000; Tseng et al., 2009). The percent interval was defined as the percent of the total
duration of the kick cycle that each partitioned knee torque contributed to the knee NET torque.
The torque impulse was defined as the magnitude of the contribution of each partitioned knee
torque to knee NET torque.
Percent interval. First, we analyzed the duration of the contribution of each partitioned
knee torque to the knee NET torque. For this analysis, we analyzed the sign of the partitioned
torques throughout the entire kick cycle. Intervals during which the sign of the knee MUS,
GRA, or MDT torques were the same or opposite to that of the knee NET torque were
INFANT LEG COORDINATION 78
considered to make a positive or a negative contribution, respectively, to the knee NET torque.
We determined the percentage of positive contribution intervals relative to the total kick cycle.
Torque impulse. Second, we analyzed the magnitude of the contribution of each
partitioned knee torque to the knee NET torque. For this analysis, we computed the positive or
negative torque impulse (torque * time) during intervals in which the knee MUS torque acted in
the same or opposite direction compared with the knee NET torque. Knee GRA and MDT
torque impulse were likewise computed as a contribution to knee NET torque. All positive and
negative impulses were summed for each torque component to yield a measure of total knee
MUS, GRA and MDT torque impulse.
Percent interval and torque impulse across ages. We analyzed the extent to which the
coordination of all kick across ages, as defined by the Z-transformed correlation coefficient, was
dependent on the percent interval and torque impulse of the MUS and MDT torques. All
dependent variables were computed using Matlab (The Mathworks, Inc., Natick, MA).
Statistical Analysis
Preterm Group
Mixed regression models were used for each dependent variable between ages (6-weeks,
15-weeks). Dependent variables included 15 joint correlations, 75 relative phase relations, 15
NET torques, 15 MUS torques, 15 MDT torques, 15 GRA torques, 3 contributions of partitioned
torque to NET torque in terms of percent interval, 3 contributions of partitioned torque to NET
torque in terms of torque impulse.
Linear regression was used to assess the extent to which the coordination of the kicks
across ages, as defined by the Z-transformed correlation coefficient, was dependent on the
percent interval and torque impulse of the MUS and MDT torques.
INFANT LEG COORDINATION 79
Differences between Preterm and Full-term Groups
To assess kinematic and kinetic differences between PT and FT infant groups, mixed
regression models with group (PT, FT) as the between-subject factor were used to test the
differences of each dependent variable listed above between ages (6-weeks, 15-weeks). All
statistical tests were completed using SAS (version 7.0, SAS Institute Inc.) with overall alpha
value at 0.05. Preplanned post-hoc comparisons were performed using a Bonferroni correction
to adjust for multiple comparisons.
Results
Joint Coordination
We first investigate whether all joint angle pairs change from a predominately in-phase
coordination pattern at 6-weeks to an increased variety of joint coordination patterns at 15-
weeks. We define joint coordination using joint angle correlation coefficients and relative
phasing between joint angle pairs.
Joint angle correlations.
Preterm group. Joint correlation data are graphed in Figure 4.1 and included in Table
4.2. At 6-weeks of age, 12 paired joint angle correlations were positive ranging from 0.01 to
1.12, and 3 paired joint angle correlations were negative ranging from -0.32 to -0.20. At 15-
weeks of age, joint correlations ranged from 0.05 to 0.81, except for hip external/internal
rotation-knee flexion/extension (-0.21). Comparing joint correlations at 6 and 15-weeks, 5 joint
correlations decreased toward a zero correlation (adjusted p<0.01), 2 joint correlations increased
toward a zero correlation (both included ankle eversion/inversion; adjusted p<0.01), 2 joint
correlations increased to a more positive correlation (both included hip external/internal rotation
with another hip intrajoint movement; adjusted p<0.01), and 6 joint correlations remained the
INFANT LEG COORDINATION 80
A. 6 weeks B. 15 weeks

1. ↓ hip flexion/extension-knee flexion/extension*
2. ↓ hip flexion/extension-hip abduction/adduction*
3. ↓ hip abduction/adduction-knee flexion/extension*
4. ↓ ankle eversion/inversion-ankle dorsiflexion/plantarflexion*
5. ankle dorsiflexion/plantarflexion-hip flexion/extension
6. ankle dorsiflexion/plantarflexion-hip abduction/adduction
7. ↓ hip external/internal rotation-ankle eversion/inversion*
8. ankle dorsiflexion/plantarflexion-knee flexion/extension
9. hip external/internal rotation-ankle dorsiflexion/plantarflexion
10. ↑ hip external/internal rotation-hip abduction/adduction*
11. ankle eversion/inversion-hip abduction/adduction
12. ↑ hip external/internal rotation-hip flexion/extension*
13. hip external/internal rotation- knee flexion/extension
14. ↑ ankle eversion/inversion-hip flexion/extension*
15. ↑ ankle eversion/inversion-knee flexion/extension*

Figure 4.1 Preterm infants: mean joint correlations (Pearson’s r converted to Fisher Z). Error
bars represent standard errors. Figure A. 6 weeks, N=2995 kicks. Figure B. 15 weeks, N=1738
kicks.
* adjusted p < 0.01; ↑ and ↓ indicate direction of difference.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
- 2. 00
- 1. 20
- 0. 40
0. 40
1. 20
2. 00
P ret erm
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
- 2. 00
- 1. 20
- 0. 40
0. 40
1. 20
2. 00
P ret erm
INFANT LEG COORDINATION 81
Table 4.2

Preterm Infants Compared to Full-Term Infants: Joint Correlations
A. Differences between full-term and preterm groups at 6-weeks and 15-weeks

Full-term
N=4373

M (SE)
Preterm
N= 4733

M (SE)
Between
Groups

adjusted p
Ankle dorsiflexion/plantarflexion
-hip flexion/extension
6 weeks
15 weeks
1.15 (0.08)
0.44 (0.08)↓
0.55 (0.09)*
0.69 (0.09)
<0.01
>0.05

Ankle dorsiflexion/plantarflexion
-knee flexion/extension
6 weeks
15 weeks
1.07 (0.10)
0.24 (0.10)↓
0.33 (0.12)*
0.51 (0.12)
<0.01
>0.05

Ankle eversion/inversion
-hip flexion/extension
6 weeks
15 weeks
0.31 (0.07)
-0.16 (0.07)↓
-0.21 (0.08)*
0.06 (0.08)↑
<0.01
>0.05

Ankle eversion/inversion
-hip abduction/adduction
6 weeks
15 weeks
0.41 (0.05)
0.12 (0.05)↓
0.12 (0.05)*
0.20 (0.06)
<0.05
>0.05

Ankle eversion/inversion
-knee flexion/extension
6 weeks
15 weeks
0.33 (0.08)
-0.22 (0.08)↓
-0.33 (0.09)*
0.04 (0.09)↑
<0.01
>0.05

(continued)
INFANT LEG COORDINATION 82
B. No differences between full-term and preterm groups at 6 weeks and 15 weeks

Full-term
N=4373

M (SE)
Preterm
N= 4733

M (SE)
Between
Groups

adjusted
p
Correlation coefficients in both groups increased in magnitude from 6 to 15 weeks
Hip external/internal rotation
-hip abduction/adduction
6 weeks
15 weeks
0.83 (0.10)
1.01 (0.10)↑
0.31 (0.12)
0.81 (0.12)↑
>0.05
>0.05

Hip external/internal rotation
-hip flexion/extension
6 weeks
15 weeks
0.07 (0.10)
0.20 (0.10)↑
0.01 (0.11)
0.31 (0.12)↑
>0.05
>0.05

Correlation coefficients in both groups decreased in magnitude from 6 to 15 weeks
Hip flexion/extension
-knee flexion/extension
6 weeks
15 weeks
1.63 (0.12)
0.60 (0.12)↓
1.12 (0.15)
0.67 (0.15)↓
>0.05
>0.05

Hip flexion/extension
-hip abduction/adduction
6 weeks
15 weeks
0.87 (0.11)
0.37 (0.11)↓
0.97 (0.13)
0.54 (0.13)↓
>0.05
>0.05

Hip abduction/adduction
-knee flexion/extension
6 weeks
15 weeks
0.84 (0.09)
0.13 (0.09)↓
0.82 (0.10)
0.22 (0.10)↓
>0.05
>0.05

Ankle eversion/inversion
-ankle dorsiflexion/plantarflexion
6 weeks
15 weeks
0.53 (0.07)
0.27 (0.07)↓
0.56 (0.09)
0.35 (0.09)↓
>0.05
>0.05

Correlation coefficients didn’t change in either group from 6 weeks to 15 weeks
Hip external/internal rotation
-ankle dorsiflexion/plantarflexion
6 weeks
15 weeks
0.26 (0.06)
0.20 (0.06)
0.32 (0.06)
0.28 (0.07)
>0.05
>0.05

Correlation coefficients decreased in one group, but not the other group from 6 to 15 weeks
Ankle dorsiflexion/plantarflexion
-hip abduction/adduction
6 weeks
15 weeks
0.68 (0.07)
0.16 (0.07)↓
0.55 (0.08)
0.39 (0.08)
>0.05
>0.05

Ankle eversion/inversion
-hip external/internal rotation

6 weeks
15 weeks
0.28 (0.03)
0.18 (0.03)
0.45 (0.04)
0.18 (0.04)↓
>0.05
>0.05

Hip external/internal rotation
-knee flexion/extension
6 weeks
15 weeks
0.09 (0.07)
-0.32 (0.07)↓
-0.20 (0.08)
-0.21 (0.08)
>0.05
>0.05
Note. Mean correlation coefficient (Pearson’s r converted to Fisher Z). SE=standard error.
* Between-group difference at the same age, adjusted p<0.05;
↑↓ Within-group difference between ages, adjusted p < 0.05; ↑ and ↓ indicate direction of
difference.

INFANT LEG COORDINATION 83
same (5 included an ankle motion and 1 included hip external/internal rotation; adjusted p>0.05).
Differences between preterm and full-term groups. Between-group joint correlation
data are included in Table 4.2. When comparing the PT to the FT group, 10 joint correlations
were not significantly different between the two groups. The 5 correlations that were different
between the two groups included hip or knee motion paired with ankle eversion/inversion or
ankle dorsiflexion/plantarflexion. In all five cases, at 6-weeks of age the joint correlations of the
PT group were significantly less positively correlated than the correlations of the FT group
(adjusted p<0.03), but at 15-weeks of age the joint correlations were not significantly different
between the two groups (adjusted p>0.05).
These data suggest that similar to FT infants, PT infants generally exhibit a greater
variety of joint coordinations from 6 to 15-weeks of age. However, some joint angle pairs
involving the ankle at 6-weeks are more positively correlated in FT infants as compared to PT
infants. By 15-weeks, differences in joint coordination between FT and PT infants have
resolved.
Relative phase. We describe increasing out-of-phase coordination when the value of
relative phase between joint angles increases toward 180°.
Preterm group. Relative phase data are included in Table 4.3. Relative phase at each of
the five data points (kick initiation, peak hip flexion velocity, hip joint reversal, peak hip
extension velocity, and kick end) ranged from 40.0 to 125.1 at 6-weeks and 55.0 to 110.3 at 15-
weeks. Comparing relative phase relations at 6 and 15-weeks, 5 of 15 relative phase relations
exhibited a significant increase (more out-of-phase coordination) in at least 3 of the 5 data points
(adjusted p<0.05), 4 of 15 relative phase relations exhibited a significant decrease (more in-phase
coordination) in at least 3 of the 5 data points (adjusted p<0.05), and 5 relative phase relations
INFANT LEG COORDINATION 84
Table 4.3
Preterm Infants Compared to Full-Term Infants Relative Phase
A. Differences between full-term and preterm groups at 6 weeks and 15 weeks

Joint Angle
Pairs
Kick
Initiation
Peak
Flexion Velocity
Hip Joint
Reversal
Peak
Extension Velocity
Kick
End

FT
M (SE)
PT
M (SE)
FT
M (SE)
PT
M (SE)
FT
M (SE)
PT
M (SE)
FT
M (SE)
PT
M (SE)
FT
M (SE)
PT
M (SE)

ankle dorsiflexion/plantarflexion- hip flexion/extension
6 weeks 47.2 (3.4) 71.2 (3.9)* 42.2 (2.8) 63.4 (3.2)* 34.8 (3.7) 61.7 (4.3)* 38.4 (4.0) 63.9 (4.6) 55.8 (3.9) 74.3 (4.5)
15 weeks 75.8 (3.3)↑ 66.7 (3.9) 67.3 (2.8)↑ 56.3 (3.3)↓ 65.0 (3.6)↑ 56.6 (4.3) 69.4 (3.9)↑ 56.4 (4.7) 79.4 (3.8)↑ 67.7 (4.6)

ankle dorsiflexion/plantarflexion -knee flexion/extension
6 weeks 47.2 (3.8) 77.8 (4.4)* 50.4 (4.1) 82.0 (4.8)* 40.2 (4.4) 73.4 (5.1)* 39.3 (4.3) 71.2 (5.0)* 50.9 (4.0) 77.4 (4.7)
15 weeks 80.8 (3.8)↑ 71.4 (4.5) 83.9 (4.0)↑ 66.4 (4.8)↓ 81.7 (4.3)↑ 65.3 (5.2)↓ 79.5 (4.2) 67.7 (5.0)↓ 82.0 (4.0)↑ 72.6 (4.7)

ankle eversion/inversion-knee flexion/extension
6 weeks 74.3 (3.9) 102.6 (4.6)* 86.9 (3.6) 111.0 (4.2) 76.8 (3.6) 111.4 (4.2)* 74.4 (4.0) 106 (4.6)* 73.1 (4.1) 99.7 (4.8)
15 weeks 100.8 (3.9)↑ 91.4 (4.6)↓ 101.3 (3.6)↑ 89.0 (4.3)↓ 101.6 (3.6)↑ 88.5 (4.3)↓ 102.3 (3.9)↑ 89.3 (4.7)↓ 97.6 (4.1)↑ 88.1 (4.9)↓

ankle eversion/inversion-hip flexion/extension
6 weeks 77.0 (3.0) 98.9 (3.4)* 79.1 (3.5) 96.2 (4.1) 77.5 (4.0) 101.7 (4.7)* 80.1 (3.9) 107 (4.6)* 82.1 (3.6) 100.1 (4.2)
15 weeks

95.8 (3.0)↑

89.8 (3.5)↓

98.6 (3.5)↑

86.2 (4.1)↓

100.8 (4.0)↑

87.5 (4.7)↓

104.5 (3.9)↑

87.3 (4.6)↓

95.0 (3.6)↑

87.3 (4.3)↓

(continued)
INFANT LEG COORDINATION 85
B. No differences between full-term and preterm groups at 6 weeks and 15 weeks

Joint
Angle
Pairs

Kick
Initiation
Peak
Flexion Velocity
Hip Joint
Reversal
Peak
Extension Velocity
Kick
End
FT
M (SE)
PT
M (SE)
FT
M (SE)
PT
M (SE)
FT
M (SE)
PT
M (SE)
FT
M (SE)
PT
M (SE)
FT
M (SE)
PT
M (SE)

Both groups more out-of-phase from 6 to 15 weeks

hip flexion/extension-knee flexion/extension
6 weeks 28.2 (5.0) 44.4 (5.9) 31.7 (4.4) 47.6 (5.1) 21.5 (4.7) 39.9 (5.5) 19.7 (5.3) 37.5 (6.3) 32.2 (5.0) 43.4 (5.9)
15 weeks 67.6 (5.0)↑ 63.2 (6.0)↑ 65.7 (4.3)↑ 62.3 (5.2)↑ 62.3 (4.7)↑ 58.2 (5.6)↑ 62.8 (5.3)↑ 57.6 (6.3)↑ 69.9 (5.0)↑ 63.9 (6.0)↑

hip flexion/extension-hip abduction/adduction
6 weeks 47.2 (4.7) 49.9 (5.5) 61.6 (4.7) 52.5 (5.5) 60.76 (4.5) 53.3 (5.3) 51.2 (4.4) 49.0 (5.2) 51.2 (3.8) 52.9 (4.4)
15 weeks 72.4 (4.7)↑ 63.8 (5.6)↑ 78.1 (4.6)↑ 69.0 (5.5)↑ 77.8 (4.5)↑ 69.4 (5.3)↑ 73.9 (4.4)↑ 65.1 (5.2)↑ 73.7 (3.8)↑ 69.4 (4.5)↑

hip abduction/adduction-knee flexion/extension
6 weeks 47.5 (4.3) 54.1 (5.1) 74.2 (3.9) 65.4 (4.5) 61.6 (3.9) 56.4 (4.6) 44.9 (4.0) 46.5 (4.7) 45.9 (3.8) 52.2 (4.5)
15 weeks 82.8 (4.3)↑ 77.9 (5.1)↑ 90.4 (3.8)↑ 86.8 (4.6)↑ 86.1 (3.9)↑ 81.9 (4.6)↑ 81.5 (4.0)↑ 76.1 (4.7)↑ 80.9 (3.8)↑ 76.5 (4.5)↑

ankle eversion/inversion-ankle dorsiflexion/plantarflexion
6 weeks 69.5 (3.4) 64.8 (3.9) 69.6 (3.6) 62.8 (4.1) 67.1 (3.4) 66.0 (4.0) 67.1 (3.2) 68.5 (3.7) 66.5 (3.1) 61.6 (3.5)
15 weeks 77.6 (3.3)↑ 76.1 (4.0)↑ 77.9 (3.5)↑ 74.3 (4.2)↑ 79.3 (3.4)↑ 77.5 (4.0)↑ 81.1 (3.2)↑ 76.1 (3.8)↑ 78.1 (3.0)↑ 74.0 (3.6)↑

Neither group changed from 6 to 15 weeks

hip external/internal rotation-ankle dorsiflexion/plantarflexion
6 weeks 79.4 (2.9) 75.9 (3.2) 93.7 (3.7) 80.7 (4.2) 81.3 (3.0) 72.6 (3.5) 68.3 (3.1) 68.3 (3.6) 69.3 (2.3) 72.7 (2.5)
15 weeks 82.5 (2.8) 76.3 (3.3) 83.7 (3.6)↓ 81.7 (4.3) 78.7 (3.0) 77.3 (3.5) 76.4 (3.1)↑ 74.5 (3.7)↑ 80.3 (2.2)↑ 73.7 (2.6)

(continued)
INFANT LEG COORDINATION 86
Joint Angle
Pairs
Kick
Initiation
Peak
Flexion Velocity
Hip Joint
Reversal
Peak
Extension Velocity
Kick
End

FT
M (SE)
PT
M (SE)
FT
M (SE)
PT
M (SE)
FT
M (SE)
PT
M (SE)
FT
M (SE)
PT
M (SE)
FT
M (SE)
PT
M (SE)

One group changed, but the other did not change from 6 to 15 weeks

ankle dorsiflexion/plantarflexion-hip abduction/adduction
6 weeks 62.5 (2.9) 69.3 (3.3) 67.8 (3.2) 63.5 (3.6) 61.3 (3.0) 60.0 (3.4) 54.5 (3.5) 66.2 (4.0) 62.9 (2.8) 71.5 (3.1)
15weeks 84.4 (2.8)↑ 76.3 (3.4) 85.1 (3.1)↑ 72.8 (3.7)↑ 81.9 (3.0)↑ 70.8 (3.5)↑ 81.2 (3.4)↑ 70.3 (4.1) 83.1 (2.7)↑ 76.9 (3.2)

ankle eversion/inversion-hip abduction/adduction
6 weeks 77.2 (2.0) 87.9 (2.1) 72.8 (3.0) 78.1 (3.4) 69.2 (2.6) 82.9 (2.9) 69.6 (2.6) 92.0 (3.0)* 74.1 (2.6) 88.3 (2.9)
15 weeks 87.8 (1.9)↑ 86.8 (2.3) 85.1 (2.9)↑ 80.7 (3.5) 82.8 (2.5)↑ 78.9 (3.0) 85.1 (2.6)↑ 80.8 (3.1)↓ 83.4 (2.6)↑ 82.5 (3.0)

hip external/internal rotation-knee flexion/extension
6 weeks 77.0 (4.3) 85.9 (5.0) 112.8 (4.1) 125 (4.7) 96.5 (3.6) 117.0 (4.2) 75.1 (3.6) 94.5 (4.2) 68.8 (3.4) 79.6 (3.9)
15 weeks 98.5 (4.3)↑ 96.5 (5.1)↑ 117.6 (4.0) 110 (4.8)↓ 108.9 (3.6)↑ 100.8 (4.3)↓ 99.4 (3.6)↑ 92.5 (4.2) 94.0 (3.3)↑ 91.1 (4.0)↑

hip external/internal rotation-ankle eversion/inversion
6 weeks 84.2 (2.5) 78.2 (2.8) 84.1 (2.4) 66.1 (2.7)* 72.6 (2.3) 54.6 (2.5)* 67.4 (1.6) 63.7 (1.7) 74.1 (2.1) 75.5 (2.3)
15 weeks 83.9 (2.5) 82.2 (2.9) 80.8 (2.3) 80.9 (2.8)↑ 77.4 (2.2) 79.0 (2.6)↑ 80.6 (1.6)↑ 80.8 (1.8)↑ 82.3 (2.1)↑ 81.3 (2.4)

hip external/internal rotation-hip flexion/extension
6 weeks 79.4 (4.6) 79.6 (5.4) 106.7 (5.5) 109.2 (6.5) 92.9 (4.6) 101.4 (5.3) 79.9 (4.7) 87.4 (5.5) 75.4 (4.3) 78.9 (5.0)
15 weeks 75.8 (4.6) 75.3 (5.5) 91.3 (5.5)↓ 82.5 (6.5)↓ 88.2 (4.5) 78.8 (5.4)↓ 80.7 (4.7) 73.2 (5.6)↓ 76.7 (4.3) 76.8 (5.1)

hip external/internal rotation-hip abduction/adduction
6 weeks 58.7 (4.2) 72.7 (4.8) 65.0 (4.4) 87.4 (5.2) 52.3 (3.8) 80.3 (4.4)* 49.9 (3.9) 75.1 (4.5) 50.9 (3.7) 69.2 (4.3)
15 weeks 54.7 (4.1) 61.8 (4.9)↓ 55.2 (4.4)↓ 61.0 (5.2)↓ 51.8 (3.8) 55.0 (4.5)↓ 52.5 (3.8)↑ 55.6 (4.6)↓ 52.4 (3.7) 56.5 (4.4)↓

(continued)
INFANT LEG COORDINATION 87
Note. FT = full-term, PT = preterm, SE = standard error.
* Between-group difference at the same age, adjusted p<0.03
↑↓ Within-group difference between ages, adjusted p < 0.05; ↑ and ↓ indicate direction of difference.

INFANT LEG COORDINATION 88
demonstrated no consistent change in at least 3 of the 5 data points (adjusted p>0.05).
Relative phase results were consistent with joint correlation results. The 5 joint
correlations that decreased towards a zero correlation from 6-weeks to 15-weeks, were the 5
relative phase relations that exhibited a significant increase in at least 3 of the 5 data points
(more out-of-phase coordination). The 4 joint correlations that increased from 6-weeks to 15-
weeks, were the same 4 relative phase relations that increased in at least 3 of 5 data points (more
in-phase coordination). The 6 joint correlations that did not change from 6-weeks to 15-weeks,
were the same 6 relative phase relations that did not change consistently in 3 of the 5 data points.
Differences between preterm and full-term groups. Between-group relative phase
data are included in Table 4.3. When comparing the PT to the FT group, 11 relative phase
relations were not significantly different between the two groups in at least 3 of the 5 data points
(adjusted p>0.05). The 4 relative phase relations that were different between the two groups in at
least 3 of the 5 data points included hip flexion/extension or knee flexion/extension paired with
ankle dorsiflexion/plantarflexion or ankle eversion/inversion. In all four cases, at 6-weeks of age
the relative phase data points of the PT group were significantly increased (more out-of-phase) as
compared to the FT group (adjusted p<0.01), but at 15-weeks of age the relative phase data
points were not significantly different between the two groups (adjusted p>0.05).
Relative phase results were consistent with joint correlation results. The 10 joint
correlations that were not significantly different between FT and PT groups, were 10 of the 11
relative phase relations that were not significantly different between groups in at least 3 of the 5
data points. The 5 joint correlations that were significantly different at 6-weeks in the PT as
compared to the FT group, included 4 relative phase relations that were significantly different
between groups in at least 3 of the 5 data points and 1 relative phase relation that was not
INFANT LEG COORDINATION 89
significantly different. Analysis of both joint correlations and relative phase relations supported
that the 4 joint coordinations which include hip flexion/extension or knee flexion/extension
paired with ankle dorsiflexion/plantarflexion or ankle eversion/inversion were less in-phase in
the PT as compared to the FT group at 6-weeks, but there were no significant difference between
groups at 15-weeks.
Joint Torques
Preterm group. We investigate the intersegmental dynamics of infant spontaneous kicks
between 6 and 15-weeks of age by comparing NET joint torques and partitioned torques (GRA,
MUS, MDT) at five discrete points within each kick (kick initiation, peak hip flexion velocity,
hip joint reversal, peak hip extension velocity, and kick end). NET joint torque and partitioned
joint torque data normalized to the mass of the leg are included in Table 4.4. At 15-weeks as
compared to 6-weeks, kick initiation had greater hip NET torque [F
(1,6)
=25.86, adjusted p<0.03],
hip MUS torque [F
(1,6)
=25.60, adjusted p<0.03], and hip GRA torque [F
(1,6)
=39.78, adjusted
p<0.02]. In addition, greater hip MDT torque was found at peak extension velocity [F
(1,6)
=24.95,
adjusted p<0.05], but no other differences between ages at other data points were found in hip
NET or partitioned torques. Knee NET torque was not different at any data point between ages.
Ankle NET torque was lower at 15-weeks as compared to 6-weeks at peak extension velocity
[F
(1,6)=
47.32, adjusted p<0.02], hip joint reversal [F
(1,6)=
26.31, adjusted p<0.03], and peak
extension velocity [F
(1,6)=
24.46, adjusted p<0.04], but no other differences were found in ankle
NET or partitioned torques.
Differences between PT and FT groups. There were no statistically significant
differences between hip, knee, or ankle torques at each of the five discrete points between PT
and FT groups when adjusting for multiple comparisons.
INFANT LEG COORDINATION 90
Table 4.4
Preterm Infants: Normalized Torques

Kick
Initiation

M (SE)
Peak Flexion
Velocity

M (SE)
Hip Joint
Reversal

M (SE)
Peak
Extension
Velocity
M (SE)
Kick
End

M (SE)
Net Joint Torque (Nm/kg)
Hip 6 weeks
15 weeks
0.28 (0.03)
0.41 (0.03)↑
0.07 (0.01)
0.09 (0.01)
0.29 (0.03)
0.31 (0.03)
0.08 (0.01)
0.08 (0.01)
0.21 (0.02)
0.21 (0.02)
Knee 6 weeks
15 weeks
0.04 (0.00)
0.02 (0.01)
0.02 (0.00)
0.02 (0.00)
0.02 (0.00)
0.01 (0.00)
0.01 (0.00)
0.02 (0.00)
0.01 (0.00)
0.02 (0.00)
Ankle 6 weeks
15 weeks
0.00 (0.00)
0.00 (0.00)
0.00 (0.00)
0.00 (0.00)↓
0.00 (0.00)
0.00 (0.00)↓
0.00 (0.00)
0.00 (0.00)↓
0.00 (0.00)
0.00 (0.00)

Muscle Torque (Nm/kg)
Hip 6 weeks
15 weeks
1.09 (0.05)
1.26 (0.06)↑
0.73 (0.03)
0.76 (0.03)
0.36 (0.03)
0.41 (0.03)
0.71 (0.04)
0.73 (0.05)
0.97 (0.06)
0.95 (0.06)
Knee 6 weeks
15 weeks
0.25 (0.03)
0.24 (0.03)
0.11 (0.01)
0.08 (0.02)
0.08 (0.01)
0.09 (0.01)
0.11 (0.02)
0.10 (0.02)
0.20 (0.02)
0.16 (0.03)
Ankle 6 weeks
15 weeks
0.01 (0.00)
0.01 (0.00)
0.01 (0.00)
0.01 (0.00)
0.01 (0.00)
0.01 (0.00)
0.01 (0.00)
0.01 (0.00)
0.01 (0.00)
0.01 (0.00)

Gravitational Torque (Nm/kg)
Hip 6 weeks
15 weeks
0.80 (0.03)
0.88 (0.03)↑
0.73 (0.03)
0.77 (0.04)
0.58 (0.04)
0.61 (0.04)
0.70 (0.05)
0.70 (0.05)
0.77 (0.05)
0.75 (0.05)
Knee 6 weeks
15 weeks
0.15 (0.02)
0.14 (0.02)
0.12 (0.02)
0.10 (0.02)
0.07 (0.01)
0.07 (0.01)
0.11 (0.02)
0.09 (0.02)
0.13 (0.02)
0.11 (0.02)
Ankle 6 weeks
15 weeks
0.01 (0.00)
0.01 (0.00)
0.01 (0.00)
0.01 (0.00)
0.01 (0.00)
0.01 (0.00)
0.01 (0.00)
0.01 (0.00)
0.01 (0.00)
0.01 (0.00)

Motion-Dependent Torque (Nm/kg)
Hip 6 weeks
15 weeks
0.08 (0.01)
0.08 (0.01)
0.05 (0.00)
0.06 (0.01)
0.06 (0.01)
0.07 (0.01)
0.03 (0.00)
0.06 (0.01)↑
0.04 (0.00)
0.06 (0.01)
Knee 6 weeks
15 weeks
0.07 (0.01)
0.10 (0.01)
0.01 (0.00)
0.02 (0.00)
0.08 (0.01)
0.07 (0.01)
0.02 (0.00)
0.02 (0.00)
0.06 (0.01)
0.05 (0.01)
Ankle 6 weeks
15 weeks
0.01 (0.00)
0.01 (0.00)
0.00 (0.00)
0.00 (0.00)
0.01 (0.00)
0.00 (0.00)
0.00 (0.00)
0.00 (0.00)
0.00 (0.00)
0.00 (0.00)

Note. 6-weeks (n=262 kicks) and 15-weeks (n=126 kicks).
↑ ↓ = adjusted p < 0.05, arrow indicates direction of difference

INFANT LEG COORDINATION 91
In summary, within-group the PT group demonstrated similar intersegmental dynamics as
the FT group; NET joint torque and partitioned joint torque data normalized to the mass of the
leg of each individual participant remained relatively consistent from 6 to 15-weeks of age.
Between-groups, no significant differences were noted.
Relation between Knee Joint Dynamics and Changes in Hip-Knee Joint Coordination
Preterm group. In order to investigate differences in joint dynamics which contribute to
changes in joint coordination, we focus on hip and knee flexion and extension. We quantified
the contributions of knee partitioned (MUS, GRA, and MDT) torques to knee NET torque in
terms of duration (percent interval) and magnitude (torque impulse) (Sainburg & Kalakanis,
2000; Tseng et al., 2009). Knee torque data are included in Figures 4.2 and 4.3.
Percent interval and torque impulse at 6-weeks. At 6-weeks, the normalized mean hip
and knee flexion/extension correlation coefficient was 0.91 (SE=0.19). The knee MUS torque
contributed to knee NET torque for a mean of 61% of the kick cycle, whereas knee GRA torque
contributed 48% and MDT torque contributed 38%. The knee MUS impulse (0.01 Nm sec) was
in the same direction as the knee NET torque, whereas the knee GRA (-0.005 Nm sec) and MDT
(-0.003 NM sec) impulses were in the opposite direction of the knee NET torque.
Percent interval and torque impulse at 15-weeks. At 15-weeks, the normalized mean
hip and knee flexion/extension correlation coefficient was 0.09 (SE=0.21); a significant decrease
from 6-weeks [F
(1,6)
= 48.23, adjusted p<0.001]. The knee MUS torque contributed to knee
NET torque for an average of 57% of the kick cycle, whereas knee GRA torque contributed 51%
and MDT torque contributed 45%. The duration of the contribution of knee partitioned torques
to knee NET torque was not different across ages. The contribution of torque impulse was
similar to 6-weeks; the knee MUS impulse (0.01 Nm sec) was in the same direction as the knee
INFANT LEG COORDINATION 92

Figure 4.2 Preterm infants: mean partitioned torque percent interval contribution
to net knee torque. Error bars represent standard errors.

Figure 4.3 Preterm infants: mean partitioned torque impulse contribution to net
knee torque. Error bars represent standard errors.
0
10
20
30
40
50
60
70
80
90
100
6 w eeks
15 w eeks
- 0. 20
- 0. 10
0. 00
0. 10
0. 20
0. 30
0. 40
( E- 1)
6 w eeks
15 w eeks
Muscle Motion-Dependent Gravitational
Percent Interval of Kick Cycle (%) Torque Impulse (Nm-s)
Muscle
Motion-Dependent Gravitational
0.04
.
0.03
.
0.02
.
0.01
.
0.00
.
-0.01
.
-0.02
.
INFANT LEG COORDINATION 93
NET torque, whereas the knee GRAV (-0.004) and MDT (-0.003) impulses were in the opposite
direction as the knee NET torque. The contribution of partitioned torque impulse to knee NET
torque impulse was not different across ages.
In summary, at 15-weeks as compared to 6-weeks, hip flexion/extension-knee
flexion/extension correlation coefficients were significantly decreased [F
(1,5)
= 51.29, adjusted
p<0.001], however, there were no significant differences in the duration or magnitude of knee
partitioned torques to knee NET torques.
Percent interval and torque impulse across ages. We examined the extent to which
the hip-knee coordination of the kicks, as assessed using Z-transformed hip-knee correlation
coefficients across ages, was dependent on the percent interval and torque impulse of the knee
MUS and MDT torques (Figure 4.4, Figure 4.5). For all kicks between ages, the Z-transformed
correlation coefficient was moderately dependent on the percent interval of the MUS and MDT
torque contribution to the knee NET torque, accounting for 17 and 33% of the variance
respectively [F
1,386
=76.73, adjusted p<0.001; F
1,386
=190.87, adjusted p<0.001]. For kicks with a
high positive correlation, the MUS torques contributed to NET torque for a relatively high
proportion of the kick cycle, whereas the MDT torque contributed for a relatively low proportion
of the kick cycle. For the kicks with a high negative correlation, the MUS torque contributed for
a low proportion of the kick cycle, whereas the MDT torque contributed for a relatively high
proportion of the kick cycle.
For all kicks between ages, the Z-transformed correlation coefficient was moderately
dependent on the MDT torque impulse, accounting for 20% of the variance {F
1,386
=94.11,
adjusted p<0.001]. However, the Z-transformed correlation coefficient was minimally
dependent on the MUS torque impulse, accounting for only 9% of the variance [F
1,386
=37.70,
INFANT LEG COORDINATION 94

A.

B.

Figure 4.4 Preterm infants: relation between kick coordination and percent interval across ages.
Figure A: knee muscle torque contribution to knee net torque. Figure B: knee motion dependent
torque contribution to knee net torque.

Z-transformed Correlation Coefficient
Percent interval knee motion dependent torque contribution to knee net torque
Z-transformed Correlation Coefficient
Percent interval knee muscle torque contribution to knee net torque
R
2
=0.17
R
2
=0.33
INFANT LEG COORDINATION 95
A.

B.

Figure 4.5 Preterm infants: relation between kick coordination and torque impulse across ages.
Figure A: knee muscle torque contribution to knee net torque. Figure B: knee motion dependent
torque contribution to knee net torque.

Knee muscle torque impulse contribution to knee net torque (Nm-sec)
Z-transformed Correlation Coefficient Z-transformed Correlation Coefficient
Knee motion dependent torque impulse contribution to knee net torque (Nm-sec)
R
2
=0.09
R
2
=0.20
INFANT LEG COORDINATION 96
adjusted p<0.001]. For kicks with a high positive correlation, the MUS torque impulse was in
the same direction as the knee NET torque, whereas the knee MDT torque impulse was in the
opposite direction as the knee NET torque. For kicks with a high negative correlation, the MUS
torque impulse was in the opposite direction as the knee NET torque, whereas the knee MDT
torque impulse was in the same direction as the knee NET torque.
Differences between PT and FT groups.
Correlation coefficient differences between groups. Comparing PT and FT infant
groups, the interaction of AGE*GROUP was significant [F
(1,13)
=17.40, adjusted p=0.001]. The
hip flexion/extension-knee flexion/extension correlation coefficient was significantly decreased
in PT as compared to FT infants at 6-weeks (adjusted p<0.02), but not at 15-weeks (adjusted
p>0.40).
Percent interval differences between groups. Mean percent of the kick interval during
which partitioned torques assist knee NET torque between groups are included in Table 4.6. FT
infants at 15-weeks, compared to 6-weeks, exhibited a significant decrease in the duration that
the knee MUS torque contributed to the knee NET torque and a significant increase in the
duration that the passive knee GRA and MDT torques contributed to the knee NET torque. The
PT infants did not exhibit this difference.
Comparing PT and FT infant groups, the interaction of AGE*GROUP was significant for
the duration of MUS torque [F
(1,13)
=15.8, adjusted p<0.03], MDT torque [F
(1,13)
=8.63, adjusted
p=0.03], and GRA torque [F
(1,13)
=10.5, adjusted p<0.03] contribution to knee NET torque. At 6-
weeks, the PT group, as compared to the FT group, exhibited a decreased contribution of knee
MUS torque to knee NET torque (adjusted p<0.01) and an increased contribution of knee GRA

INFANT LEG COORDINATION 97
Table 4.6
Preterm Infants Compared to Full-Term Infants: Relation of Kick Coordination and Percent
Interval
Full-term
975 kicks

M (SE)
Preterm
388 kicks

M (SE)
Between-
groups

adjusted
p
Hip flexion/extension-
knee flexion/extension
correlation coefficient
6 weeks
15 weeks

1.75 (0.17)
0.31 (0.17)*
0.91 (0.19)**
0.09 (0.21)*
<0.02
0.43

Percent cycle that knee
MUS torque contributes to
knee NET torque
6 weeks
15 weeks

0.74 (0.02)
0.61 (0.02)*
0.61 (0.02)**
0.57 (0.03)
<0.01
0.36

Percent cycle that knee
GRA torque contributes to
knee NET torque
6 weeks
15 weeks

0.35 (0.02)
0.45 (0.02)*
0.48 (0.02)**
0.51 (0.03)
<0.01
0.27

Percent cycle that knee
MDT torque contributes to
knee NET torque
6 weeks
15 weeks

0.30 (0.02)
0.45 (0.02)*
0.38 (0.02)**
0.45 (0.03)
0.03
0.97

Note. SE = standard error.
* Within-group difference between ages, adjusted p<0.05
** Between-group difference at the same age, adjusted p<0.05

INFANT LEG COORDINATION 98
Table 4.7
Preterm Infants Compared to Full-Term Infants: Relation of Kick Coordination and Torque
Impulse
Full-term
975 kicks

M (SE)
Preterm
388 kicks

M (SE)
Between-
groups

adjusted
p
Hip flexion/extension-
knee flexion/extension
correlation coefficient
6 weeks
15 weeks

1.75 (0.17)
0.31 (0.17)*
0.91 (0.19)**
0.09 (0.21)*
<0.02
0.43

Knee MUS impulse
contribution to knee
NET torque
6 weeks
15 weeks

0.03 (0.002)
0.03 (0.002)
0.01 (0.003)**
0.01 (0.003)**
<0.001
<0.03

Knee GRA impulse
contribution to knee
NET torque
6 weeks
15 weeks

-0.015 (0.001)
-0.008 (0.001)*
-0.005 (0.001)**
-0.004 (0.002)
<0.001
0.16

Knee MDT impulse
contribution to knee
NET torque
6 weeks
15 weeks

-0.009 (0.001)
-0.007 (0.001)
-0.003 (0.001)**
-0.004 (0.001)
<0.001
0.12

Note. SE = standard error.
* Within-group difference between ages, adjusted p<0.02
** Between-group difference at the same age, adjusted p<0.02

torque (adjusted p<0.01) and knee MDT torque (adjusted p=0.03) contribution to knee NET
torque. At 15-weeks, no significant differences were noted between PT and FT groups (p>0.30).
Torque impulse differences between groups. Mean partitioned torque impulses between
groups are included in Table 4.7. FT infants at 15-weeks, compared to 6-weeks, exhibited a
significant difference in GRA impulse contribution to the knee NET torque. There were no
differences in MUS and MDT torque impulses across ages. The PT infants did not exhibit
differences in partitioned impulses between 6-weeks and 15-weeks.
INFANT LEG COORDINATION 99
Comparing PT and FT infant groups, the interaction of AGE*GROUP was significant for
the GRA impulse contribution to NET torque [F
(1,13)
=12.89, adjusted p<0.01], trended toward
significance for MUS impulse contribution to NET torque [F
(1,13)
=7.44, adjusted p=0.05], and
was not significant for MDT impulse contribution to knee NET torque [F
(1,13)
=1.86, adjusted
p=0.60]. At 6-weeks, the PT group, as compared to the FT group, exhibited a significant
difference in the knee MUS impulse (adjusted p<0.001), knee GRA impulse (p<0.001), and knee
MDT impulse (adjusted p<0.001) contribution to the knee NET torque. At 15-weeks, the
magnitude of the knee MUS impulse contribution to the knee NET torque was significantly
decreased in the PT group as compared to the FT group (adjusted p<0.03), but no other
significant differences were noted (adjusted p>0.05).
Summary. Taken together, these data suggest that at 6-weeks in PT infants, as compared
to FT infants, the knee MUS torque contributed to knee NET torque during a lesser duration of
the kick cycle whereas passive GRA and MDT torques contributed during a greater duration of
the kick cycle. In addition, at 6-weeks in PT infants, as compared to FT infants, the knee MUS
torque impulse and passive GRA and MDT torque impulses demonstrated a lower magnitude of
contribution to the knee NET torque. This distribution in duration and magnitude of the
portioned torque impulses to NET torque in PT infants is similar to the contributions of the
impulses in the FT infants at 15-weeks.
Discussion
We hypothesized that, similar to FT infants, the majority of joint angle pairs would
change from a predominately in-phase coordination pattern at 6-weeks to an increased variety of
joint coordination patterns at 15-weeks, however, at both 6-weeks and 15-weeks PT infants, as
compared to FT infants, would exhibit less in-phase coordination between joint pairs. Our
INFANT LEG COORDINATION 100
hypothesis was partially confirmed. PT infants demonstrated a change from a predominately in-
phase coordination to an increased variety of joint coordination patterns in 30% of joint angle
pairs versus 73% of joints angle pairs in FT infants. At 6-weeks FT infants as compared to PT
infants demonstrated more in-phase coordination in 4 of 15 joint angle pairs, all which included
hip or knee flexion/extension paired with ankle motion. This is consistent with previous research
that more out-of-phase joint coordination begins with the hip-ankle joint at 1 to 2 months of age
followed by the hip-knee and knee-ankle joints over the next few months (Fetters, et al., 2004;
Jeng, et al., 2002; Piek, 1996; Thelen, 1985; Vaal, van Soest, Hopkins, et al., 2000). We suggest
that at 6-weeks PT infants may exhibit less in-phase ankle joint coordination than FT infants due
to additional exposure to the extra-uterine environment. By 15-weeks of age, we found no
differences in joint coordination between FT and PT infants.
Our second hypothesis was also confirmed. Similar to FT infants, PT infants
demonstrated relatively consistent normalized NET joint torque and partitioned joint torque data
from 6 to 15-weeks of age. No significant between-group differences were noted in NET torque
between the FT and PT infant groups. To investigate differences in force generation capacity
between FT and PT infants, spontaneous kicking may not be the ideal movement to analyze since
there is no overload to the muscle. The kinetic analysis of a goal-directed action, such as foot
reaching, with the progressive addition of leg weights, may clarify differences in the maximum
force generation capacity of PT and FT infants from 3 to 6 months of age. In addition, it may
also clarify how joint coordination and intersegmental dynamics change as the capacity to
generate muscle forces becomes insufficient for task constraints.
Our third hypothesis was associated with the relation between changes in joint
coordination and changes in intersegmental dynamics through the analysis of hip
INFANT LEG COORDINATION 101
flexion/extension-knee flexion/extension coordination. We hypothesized that, similar to FT
infants, a greater variety of joint coordinations would be associated with a decreased influence of
knee MUS forces which would allow passive knee GRA and MDT to have a greater influence on
the coordination of the kick, contributing to a greater variety of hip-knee coordination patterns.
Although PT infants exhibited a greater variety of joint coordinations from 6-weeks to 15-weeks
of age, no differences were noted in the influence of knee MUS, GRA, and MDT torques on
knee NET torques. However, the change in coordination from 6-weeks to 15-weeks was smaller
in the PT group as compared to the FT group. A larger coordination change may be necessary to
document the relation between intersegmental dynamics and joint coordination. When we
examined the extent to which the hip-knee coordination of the kicks was dependent on the
percent interval and torque impulse of the knee MUS and MDT torques, we found a similar
relation as the FT infants; as kicks demonstrated a greater variety of joint coordinations the knee
MUS torque had less influence on the knee NET torque and MDT torques had more influence on
the knee NET torque. Future research could investigate the relation between intersegmental
dynamics and joint coordination in PT infants with WMD, at highest risk for differences in joint
coordination. Since changes in leg joint coordination which emerge during infant spontaneous
kicking are complex, a single subject analysis of the spontaneous kicking of a few PT infants
with WMD at multiple time points within the first six months of life may be necessary to
understand the relation between intersegmental dynamics and joint coordination.

INFANT LEG COORDINATION 102
CHAPTER V
INFANT EXPLORATORY LEARNING:
INFLUENCE ON LEG JOINT COORDINATION

Abstract
Infants born preterm with very low birth weight are at high risk for developing spastic cerebral
palsy, which is characterized by walking limitations due to a reduced ability to move the joints of
the legs in an out-of-phase pattern; e.g., flexing the hip while extending the knee. Previous
research demonstrates that full-term infants will exhibit more out-of-phase hip-knee joint
coordination when their leg actions are reinforced with activation of an overhead infant mobile,
however it is unknown whether preterm infants at risk for cerebral palsy will be able to exhibit
more out-of-phase joint coordination. The purpose of this study is to determine the ability of full-
term and preterm infants to: (1) learn through discovery, the contingency between leg action and
mobile activation, and (2) demonstrate a greater amount of out-of-phase hip-knee joint
coordination when leg actions are reinforced with mobile activation. Fourteen full-term infants
and six preterm infants participated at 3 - 4 months corrected age. Each infant participated in 2
sessions of mobile reinforcement on consecutive days. During each session, the infant was
positioned supine under an overhead infant mobile. Day 1 consisted of a 2-minute non-
reinforcement condition (spontaneous kicking) followed by a 6-minute reinforcement condition
(the infant mobile rotated and played music when the infant moved either foot vertically across a
virtual threshold). Day 2 consisted of a 2-minute non-reinforcement condition, 6-minute
reinforcement condition, and 2-minute non-reinforcement condition. The full-term group, but
not the preterm group, increased the percentage of mobile activation to meet performance criteria
INFANT LEG COORDINATION 103
the 2
nd
day. Neither group met learning criteria. However, both the full-term and preterm
groups contained infants that learned the contingency. Infants were separated into infants that
learned the contingency and infants that did not learn the contingency. Infants who learned the
contingency demonstrated a greater amount of out-of-phase hip-knee joint coordination during
the reinforcement condition on the second day as compared to spontaneous kicking during the
initial non-reinforcement condition on the first day. This coordination change was not
demonstrated by the group of infants that did not learn the contingency. These results indicate
that some full-term and preterm infants can demonstrate a greater amount of out-of-phase hip-
knee joint coordination when participating in a task in which their leg actions are reinforced with
mobile activation.
Introduction
Approximately a half million preterm (PT) births occur in the United States each year;
this is twelve percent of total births (Hamilton et al., 2011). PT infants are at increased risk for
cognitive, sensorimotor, and social-emotional problems which may persist from childhood into
young adulthood (Aarnoudse-Moens et al., 2009; de Kieviet et al., 2009; Delobel-Ayoub et al.,
2009; Larroque et al., 2008). Preterm infants are also at high risk for white matter damage,
which has been associated with the development of spastic cerebral palsy (CP), a disorder of
movement coordination (Himpens et al., 2008; Rosenbaum et al., 2007).
We have demonstrated that as early as one-month corrected-age (CA) PT infants with
white matter damage (PTWMD), in comparison to PT infants without damage and full-term (FT)
infants, exhibit atypical leg movements during spontaneous kicking which we quantified and
described as constrained by excessive in-phase intralimb joint coordination (Fetters et al., 2004).
Specifically, kicking of PTWMD infants is characterized by a preponderance of synchronous
INFANT LEG COORDINATION 104
flexion or synchronous extension of all intralimb joints. Although FT infants demonstrate some
intralimb joint synchronies, they have increasing out-of-phase intralimb joint coordination during
the first months of life. The difference among FT, PT, and PTWMD infants reduces by 5
months, but some PTWMD infants continue to have difficulty generating out-of-phase intralimb
joint coordination (Fetters et al., 2010). Children with spastic CP also demonstrate difficulty
generating out-of-phase intralimb joint coordination, which affects their functional mobility,
although the mechanism for this difficulty has not been identified (Fowler & Goldberg, 2009;
Ostensjo et al., 2004; Sanger et al., 2006). In this study, we further investigate this difficulty
with out-of-phase leg joint coordination in PT infants and seek to enhance out-of-phase patterns
as they are developing. Since we had not previously studied the performance of PT infants in
terms of changing their coordination patterns, we included PT infants regardless of WMD
determination.
Previous research supports that FT infants can learn to modify their spontaneous kicking
actions when interacting with an infant mobile that reinforces specific leg actions, such as
increasing the kicking rate of one leg (Rovee-Collier & Gekoski, 1979), crossing a specific knee
angle (Angulo-Kinzler et al., 2002; Tiernan & Angulo-Barroso, 2008), or extending the hip and
knee together (Angulo-Kinzler, 2001). Our first study using the mobile paradigm supported that
FT infants increase the frequency of kicking and exhibit an out-of-phase joint coordination
pattern of flexing their hip while extending their knee if this movement provided a more direct
means to kick a physical switch that activated an infant mobile (Chen et al., 2002). This went
beyond previous research in that infants demonstrated that they could learn a more out-of-phase
pattern of intralimb joint coordination in response to mobile reinforcement. However, one
limitation of our first experimental approach was that the intralimb joint coordination pattern that
INFANT LEG COORDINATION 105
was reinforced was constrained by the placement of the plate and was limited in the out-of-phase
leg actions that were promoted relative to the extensive movement repertoire that is available to
typically developing infants.
We now extend our use of a modified mobile paradigm to PT infants and to the study of
the promotion of out-of-phase joint coordination during infant exploratory learning. The
Perception-Action framework proposes that infants learn task-specific action through a discovery
learning process in which they demonstrate a wide range of exploratory actions to generate
information about possible outcomes of actions, then select actions which result in outcomes
with adaptive value (Gibson & Pick, 2000; Thelen & Smith, 1994). Therefore, we have
modified our infant-activated mobile paradigm to exploit this discovery learning process. Supine
infants activate our mobile by moving their legs vertically across a “virtual threshold” which is
individualized to each infant’s baseline spontaneous kicking action (Figure 5.1). What is unique
about this modification is that as infants spontaneously move their legs, they discover that their
leg actions activate the mobile. This better approximates the infant learning process and
potentially will contribute to our understanding of the learning process in order to better specify
treatment protocols for PT infants that have difficulty generating out-of-phase intralimb joint
coordination. Additionally, our new mobile paradigm, as compared to previous paradigms,
allows a wider range of leg coordination patterns to activate the mobile, yet an individualized
placement of the threshold for activation of the mobile promotes an out-of-phase intralimb joint
coordination pattern of flexing the hip while extending the knee.

INFANT LEG COORDINATION 106

Figure 5.1 Infant mobile task set-up. Note the infrared light-emitting diode (IRED) placed on
each foot (yellow circle) moving vertically to cross virtual threshold (red dashed line) and
activate mobile.

Previous research with PT infants and the mobile paradigm supports that age-corrected 3-
month old PT infants require more time, 2 versus 1 session, to learn the contingency between leg
action and mobile reinforcement (Gekoski et al., 1984). Therefore, we have extended our infant
mobile paradigm to 2 days in order to evaluate performance, memory, and learning differences
between FT and PT infants. The traditional mobile paradigm has 3 successive conditions: a non-
reinforcement baseline condition (kicking with no mobile reinforcement), a reinforcement
acquisition condition (kicking reinforced with mobile activation), and a non-reinforcement
extinction condition (kicking with no mobile reinforcement). Performance is assessed each day
to determine the extent to which each infant’s response rate (typically, kicking rate) during
mobile reinforcement exceeds the baseline rate (Gekoski et al., 1984; Rovee & Rovee, 1969).
Memory is assessed across days to determine the extent to which each infant’s response rate
INFANT LEG COORDINATION 107
during the baseline condition of the second day exceeds the baseline condition of the first day
(this is also called a retention test or a baseline ratio) (Gekoski et al., 1984). Learning is assessed
across days to determine the extent to which each infant’s response rate during mobile
reinforcement on the second day exceeds the baseline rate of the first day (Haley et al., 2006).
In summary, the purpose of this study is to (1) determine the ability of 3-month old FT
and PT infants to learn, through discovery, the contingency between leg action and mobile
activation using a virtual threshold, and (2) determine whether FT and PT infants who learn the
contingency exhibit more out-of-phase hip-knee joint coordination patterns when leg actions are
reinforced with mobile activation. We hypothesized that (1) the FT group would demonstrate
performance of the contingency on the first day and both memory and learning of the
contingency on the second day, and (2) the PT group would not demonstrate performance of the
contingency on the first day or learning and memory on the second day, but would demonstrate
performance on the second day. Since both FT and PT infant groups were expected to include
infants that learned the task, infants were separated into those that learned and did not learn the
task based on an individualized learning criteria. We hypothesized that (3) the FT and PT infants
who learn the contingency, as compared to infants who did not learn the contingency, would
exhibit more out-of-phase hip-knee joint coordination patterns when leg actions were reinforced
with mobile activation.
Method
Participants
Twenty FT infants and 7 PT infants participated in data collection at 3-months of age
(determined for PT infants by subtracting number of days premature from chronological age).
The FT infants were born ≥37 weeks gestational age (GA) without birth complications, and were
typically developing as per parent report and scores on the motor subtest of the Bayley Scales of
INFANT LEG COORDINATION 108
Infant and Toddler Development, 3
rd
Edition (Bayley, 2006). The PT infants were born ≤ 34
weeks GA. Infants were excluded from the study based on parent report if they were diagnosed
with congenital malformations, chromosomal abnormalities, prenatal drug exposure, orthopedic
impairments, visual deficits that would hinder viewing the infant mobile, or hearing deficits that
would hinder hearing the music from the infant mobile. In addition, infants were excluded for
excessive crying (crying >2 minutes) since crying infants are not likely participating in the task
(Heathcock et al., 2004). Since the standard practice of matching PT and FT infants for
conceptual age may not be an adequate maturational control (Gekoski et al., 1984), specifically
in regards to the ability to regulate arousal level (Haley et al., 2008), PT infants excluded from
the study at 3-months were asked to participate again at 4-months. Participant characteristics are
included in Table 5.1.
Parents signed a consent form prior to participation in the study, and families received a
small gift for their participation. The Institutional Review Board at the University of Southern
California approved the study.
Procedure
Experimental set-up. Each infant participated in 2 testing sessions conducted on
consecutive days at the Development of Infant Motor Performance Laboratory, University of
Southern California. During each session, infants were undressed, positioned supine on a table
under a conventional infant mobile (Fisher-Price, 2008), and secured to the table using a 4-inch
Velcro band placed across the trunk and around the table (Figure 5.2). The midline position of
the head was maintained using a horseshoe-shaped support pillow surrounding the infant’s head.
Video recording. Infants were video recorded with 3 video cameras that surrounded the
testing table (Basler Pylon IEEE1394 cameras using Streampix5 x64 edition multi-camera
INFANT LEG COORDINATION 109
Table 5.1
Full-Term and Preterm Infants: Participant Characteristics
Participant
Number
Gender Group Gestational
Age at Birth
(weeks)
Birth Weight

(grams)
Age at
Testing
(Days, CA)
Ponderal
Index
(kg/m
3
)
Bayley III
Motor Subtest
(Percentile Rank)
FT 1 F FT 37 3515 102 29.1 75
FT 2 F FT 39 3430 99 23.6 58
FT 3 F FT 39 3345 107 24.7 68
FT 4 F FT 39 3260 105 24.4 84
FT 5 M FT 39 3118 111 27.1 58
FT 6 F FT 39 3799 99 22.8 75
FT 7 M FT 39 3799 99 22.8 75
FT 8 M FT 38 3232 103 26.1 68
FT 9 M FT 39 3374 109 25.2 34
FT 10 F FT 40 3118 108 26.2 68
FT 11 M FT 38 3515 111 23.3 27
FT 12 M FT 39 3645 103 22.9 79
FT 13 F FT 40 3515 104 21.9 88
FT 14 M FT 40 3147 110 22.9 68

PT 1 M PT 28 885 104 25.3 84
PT 2 M PT 28 835 104 23 84
PT 3 M PT 29 1588 108 24 50
PT 4 F PT 34 2410 101 26.6 92

PT 1 (4 months) M PT 28 885 132 25.7 58
PT 2 (4 months) M PT 28 835 132 22.5 34
PT 3 (4 months) M PT 29 1588 136 28.6 50
PT 4 (4 months) M PT 24 765 147 22.3 21
PT 5 (4 months) M PT 24 695 147 23 12
PT 6 (4 months) F PT 34 2410 133 26.9 58
Note. CA=corrected-age, F = female, FT=full-term, M=male, PT=preterm
INFANT LEG COORDINATION 110

Figure 5.2 Infant mobile experimental set-up.

software) with a right lateral, left lateral, and overhead views of the infant. An additional
overhead video camera (Canon HG10) was used to record facial expressions including visual
attention.
Experimental protocol.
Day 1. A 2-min non-reinforcement condition (baseline) was followed by a 6-min
contingent mobile reinforcement condition (acquisition).
Day 2. A 2-min non-reinforcement condition (baseline) was followed by a 6-min
contingent mobile reinforcement condition (acquisition) and a 2-min non-reinforcement
condition (extinction). Figure 5.3.

INFANT LEG COORDINATION 111

Day 1 0 min 2 min 8 min
|__________|__________|__________|_________|
Baseline Acquisition

Day 2 0 min 2 min 8 min 10 min
|__________|__________|__________|_________|__________|
Baseline Acquisition Extinction

Figure 5.3 Infant mobile experimental protocol.

Baseline. During baseline, the infant was allowed to spontaneously move his legs, but
the mobile did not activate in response to the infant’s leg movements.
Acquisition. During acquisition, the mobile rotated and music played when the infrared
light-emitting diode (IRED) placed on either foot crossed a virtual threshold (Figure 5.1). The
threshold was placed perpendicular to the table and its height was individually determined for
each infant as one standard deviation (SD) above the average height of both feet during the day 1
baseline condition using the following equation:

Height of threshold (mm)=
( ) ( )

.

Mobile activation continued for as long as the foot was above the virtual threshold to a maximum
of 3 seconds. After 3 seconds, the mobile reactivated only if the infant moved the foot below the
virtual threshold, and then moved a foot vertically and again crossed the virtual threshold. This
“3 seconds rule” was added because during pilot testing infants would simply hold their feet
above the threshold to receive mobile reinforcement. Since we were interested in how infants
INFANT LEG COORDINATION 112
learned the contingency and whether interacting with the mobile would change hip-knee
coordination patterns, we added the 3 seconds rule to encourage leg exploratory movements
versus maintenance of a leg posture. The 6-min acquisition condition was divided into three, 2-
min intervals (Acq 1, Acq 2, Acq 3) for analyses of performance changes each day.
Extinction. During extinction, the infant was allowed to spontaneously move his legs,
but the mobile did not activate in response to the infant’s leg movements.
The parent and investigator maintained visual and verbal contact with the infant in order
to keep the infant in an alert state. Parents were instructed, “You can smile and talk to your baby
to keep him calm and happy, but don’t point out the mobile or encourage kicking by clapping or
making your voice excited.”
Kinematic data. Throughout the mobile paradigm, three-dimensional lower extremity
time-position data were collected at 100 Hz using an Optotrak Certus Motion Capture System
(Northern Digital Inc., Waterloo, ON, Canada) with two sensor banks. Each Optotrak sensor
bank consisted of three position sensors connected in series to a System Control Unit and a Dell
Precision 690 computer.
The two Optotrak sensor banks were placed horizontally approximately 2.5 meters on
opposite sides of the testing table (Figure 5.2). The root-mean-squared (RMS) error in the
calibration of the sensor banks was ≤ 0.3mm for each data collection session. A global
coordinate system was defined in relation to the testing table with the x-axis parallel to the width
of the table, y-axis parallel to the length of the table, and z-axis perpendicular to the table.
Rigid marker arrays with 4 embedded IREDs were attached bilaterally to the foot, shank,
thigh, and pelvis using Velcro straps. A small plastic piece with 2 embedded IREDs was placed
on the sternum using a double-sided sticky EKG collar. After the mobile paradigm, ten
INFANT LEG COORDINATION 113
individual IREDs were fixed bilaterally to the infant’s skin using double-sided sticky EKG
collars at the following locations: lateral midline of the trunk below the tenth rib, greater
trochanter of the hip, lateral knee joint line, ankle lateral malleolus, and distal end of the 5
th

metatarsal. Then, a static calibration trial was collected for each leg by holding the infant’s
lower extremity in an extended, anatomical position for 5 seconds. This trial was necessary to
define a local/segment coordinate system, align it with the global coordinate system, and define
the orientation of each body segment in space. All joint angles in this calibration position were
defined as zero degrees.
Anthropometric data. Each infant was weighed on a digital electric scale (Health-o-
meter). The total length of the infant was measured and for both legs the following measures
were recorded: circumference at mid-segment of thigh, shank, and foot; width of the knee (at the
knee joint line), ankle (at the malleoli), and foot (at the metatarsal heads); and length of the thigh
(greater trochanter to knee joint line), shank (knee joint line to lateral malleolus), and foot
(medial malleolus to first metatarsophalangeal joint).
Data Reduction
Performance, memory, and learning.
Dependent measure.
Reinforced leg action (RLA). During acquisition, RLA is equal to the duration of mobile
activation. During baseline and extinction since the mobile did not activate, we computed the
RLA post-hoc using the coordinates of the IRED on each foot that activated the mobile. We
computed the total duration of time one or both IREDs were above the virtual threshold
subtracting the duration of time in which one or both IREDs were above the virtual threshold for
INFANT LEG COORDINATION 114

Baseline Acquisition 1 Acquisition 2 Acquisition 3 Extinction
Figure 5.4 Raw data from one representative participant. Day 2 data from a 3-month old
preterm infant that demonstrated learning based on the individual learning criteria. Raw position
data of the z-coordinate of the infra-red light-emitting diode (IRED) placed on each foot of the
infant. Red line is IRED on right foot. Blue line is IRED on left foot. Thick black line is the
table. Dotted line is virtual threshold placed 14 cm above the table (5 ½ inches) as individually
determined for each infant based on the height of their kicking during baseline condition of the
first day. Note how the infant moves his feet during baseline when the mobile does not activate
and during the first 30 seconds of acquisition 1, then he consistently keeps both feet off the table
and moves his feet right around the threshold for the next 5½ minutes until the mobile no longer
activates during the extinction condition.

greater than a 3 seconds interval. This replicated the “3 seconds rule” of mobile activation. Raw
data from one representative subject is included in Figure 5.4 to illustrate the methods.
Definition of terms. Performance, memory, and learning were assessed using group and
individual methods as is consistent with previous literature (Chen et al., 2002; Haley et al., 2006;
Ohr & Fagen, 1991; Rovee-Collier & Gekoski, 1979).
Performance. Performance of each group was measured each day by determining
whether the percent of RLA during any one of the three, 2-min acquisition intervals significantly
Height above table
INFANT LEG COORDINATION 115
exceeded the 2-min baseline condition for the same day (Angulo-Kinzler et al., 2002; Chen et al.,
2002; Ohr & Fagen, 1991; Rovee-Collier & Gekoski, 1979).
Memory. Memory of each group was defined statistically by determining whether the
percent of RLA during the baseline condition day 2 significantly exceeded the percent of RLA
during the baseline condition day 1 (Angulo-Kinzler & Horn, 2001; Gekoski et al., 1984; Haley
et al., 2006).
Learning. Learning of each group was measured statistically by determining whether the
percent of RLA during the entire 6-min acquisition condition day 2 exceeded the percent of RLA
during the baseline condition day 1 (Haley et al., 2006).
Individual learning for each infant was measured by determining whether the percent of
RLA during the entire 6-min acquisition condition of day 2 was equal to or greater than 1.5 times
the baseline condition of day 1.
Change in performance. Change in performance for each group was measured
statistically by determining whether the percent of RLA during the entire 6-min acquisition
condition day 2 exceeded the percent of RLA during the entire 6-min acquisition condition day
1.
Video recording. Video tapes were coded for arousal and attention by an evaluator
blinded to group status. The arousal scale is described as: drowsy = 1, alert and inactive = 2,
alert and active = 3, fussy = 4, and crying = 5. The attention scale is described as: 0=not looking
at the mobile, 1=looking at the mobile. This is consistent with previous infant research (Thelen
& Ulrich, 1991; Tiernan & Angulo-Barroso, 2008).
Kinematic data reduction. Kinematic data were used to determine threshold crossings
as well as to compute dependent measures of hip-knee joint coordination. Position data were
INFANT LEG COORDINATION 116
converted into 3D coordinates with a direct linear transformation algorithm using Optotrak
system software. A custom Matlab (The Mathworks, Inc., Natick, MA) program was used to: (1)
interpolate missing position data (maximum of 20 consecutive frames) using a cubic spline, (2)
filter position data using a fourth-order Butterworth with a cut-off frequency of 5 Hz as
determined from power spectrum analysis, (3) compute joint angles of hip flexion/extension, hip
abduction/adduction, hip external/internal rotation, knee flexion/extension, ankle
dorsiflexion/plantarflexion, ankle eversion/inversion, and (4) extract kicks. A kick start was
defined as the onset of a continuous leg movement for which: (a) the infant’s foot moved at least
five consecutive frames (1 frame = 10ms), and (b) the hip and/or knee joint angle change
exceeded 11.5° (2 radians) in either flexion or extension (Chen et al., 2002; Fetters et al., 2010;
Jensen et al., 1994; Schneider et al., 1990). The end of the kick was defined as the frame of peak
extension amplitude following a flexion movement or peak flexion amplitude following an
extension movement (Chen et al., 2002; Fetters et al., 2010).
Joint angles. Joint angles were computed for the hip, knee, and ankle using the method
of Söderkvist and Wedin (Soderkvist & Wedin, 1993) at the following time points: (a) kick
initiation, (b) peak hip flexion velocity, (c) peak hip flexion angle referred to as hip joint
reversal, (d) peak hip extension velocity, and (e) kick end (Figure 3.2). These were chosen
because they are points of either movement initiation or change in direction of movement where
there is potentially a change in forces to control the limb. Joint angles at these five time points
were used either as dependent measures or to compute additional measures.
Kinematic dependent measure.
Hip-knee joint coordination. Joint coordination was defined through the analysis of
joint angle correlations and relative phase.
INFANT LEG COORDINATION 117
Hip-knee correlation coefficient. Hip and knee flexion and extension joint correlations
were computed using Pearson correlation coefficients (r) at zero lag between hip and knee joint
angle excursions for all kicks extracted for each infant. Hip-knee joint angle correlations were
converted to Fisher Z scores to allow comparison of correlations (r) among infants (Chen et al.,
2002; Fetters et al., 2004; Fetters et al., 2010; Jensen et al., 1994).
Relative phase. Relative phase describes the phase relations between two joint motions.
For each kick, joint angle data was time-normalized and continuous relative phase (CRP) was
computed for hip and knee flexion and extension from the angular position/velocity data after the
method of van Emmerick and Wagenaar (Kelso et al., 1986; van Emmerick & Wagenaar, 1996).
We then analyzed results of the CRP computation at the five time points specified above under
“Joint Angles”. Values approaching zero indicate more in-phase coordination; values
approaching 180° indicate more out-of-phase coordination. A positive value indicates that the
hip is leading the knee in phase space; a negative value indicates that the knee is leading the hip.
Since we were interested in the magnitude of out-of-phase coordination, we analyzed the
absolute value of the relative phase at each discrete point.
Statistical Analysis
Performance, Memory, Learning of Full-Term and Preterm Groups
Due to unequal sample sizes, performance, memory, and learning were assessed
separately for the FT group, PT group at 3 months, and PT group at 4 months. Performance was
assessed each day (Day 1, Day 2) using mixed regression models to test the differences of
percent of RLA in any 2-min interval in the baseline, acquisition, and extinction conditions
(INTERVAL). To assess memory, learning, and performance change across days, mixed
INFANT LEG COORDINATION 118
regression models were used to test the differences of percent of RLA among baseline,
acquisition, and extinction conditions across days (CONDITION).
Coordination Changes of Learners and Non-Learners
Based on the individual learning criteria, FT and PT infants were separated into those that
met criteria (Learners) and did not meet criteria (Non-Learners). To assess coordination changes
each day (Day 1, Day 2), mixed regression models with group (Learners, Non-Learners) as the
between-subject factor were used to test the differences of coordination (hip-knee correlation
coefficient, relative phase at each of the 5 discrete points) in any 2-min interval in the baseline,
acquisition, and extinction conditions (INTERVAL). To assess memory, learning, and
coordination change across days, mixed regression models with group (Learners, Non-Learners)
as the between-subject factor were used to test the differences of coordination (hip-knee
correlation coefficient, relative phase at each of the 5 discrete points) among baseline,
acquisition, and extinction conditions across days (CONDITION).
Arousal and Attention Differences of Learners and Non-Learners
To assess arousal and attention across days, mixed regression models with group
(Learners, Non-Learners) as the between-subject factor were used to test the differences of
arousal and attention among baseline, acquisition, and extinction conditions across days
(CONDITION). All statistical tests were completed using SAS (version 7.0, SAS Institute Inc.)
with overall alpha value at 0.05. Preplanned post-hoc comparisons were performed using a
Bonferroni correction to adjust for multiple comparisons.

INFANT LEG COORDINATION 119
Results
Performance, Memory, Learning of Full-Term and Preterm Groups
Participants. Twenty FT and 7 PT infants completed data collection at 3 months of age.
Two FT infants were excluded due to equipment difficulties. One FT infant was excluded due to
illness. Three FT infants and 3 PT infants were excluded due to excessive crying. Six PT infants
returned and participated at 4 months of age. Therefore, the study includes 14 FT infants at 3
months, 4 PT infants at 3 months, and 6 PT infants at 4 months.
Full-term.
Performance. Percent of RLA during each interval of day 1 and day 2 are graphed in
Figure 5.5. The main effect of interval was not significant for day 1 [F
(3,39)
=0.92, p=0.44], but
was significant for day 2 [F
(4,52)
=2.58, p=0.048]. Specifically, as compared to baseline of day 2,

Base Acq1 Acq2 Acq3 Base Acq1 Acq2 Acq3 Ext

Figure 5.5 Full-term infants: mean percent reinforced leg action by condition. N=14.
Error bars are standard error. Base=baseline, Acq=acquisition, Ext=extinction.
Base Acq 1 Acq 2 Acq 3
0
10
20
30
40
50
60
Base Acq 1 Acq 2 Acq 3
0
10
20
30
40
50
60
Base Acq 1 Acq 2 Acq 3 Ext
0
10
20
30
40
50
60
FT
p=0.01
p=0.02
p=0.01
% reinforced leg action (RLA)
INFANT LEG COORDINATION 120
the following intervals were significantly increased: acquisition 2 (p=0.012), acquisition 3
(p=0.015), and extinction (p=0.01). This can be interpreted as the FT group demonstrated
improved performance on day 2, but not on day 1.
Memory, learning, change in performance. Percent of RLA across days is graphed in
Figure. 5.6. The main effect of condition was significantly different [F
(4,52)
=3.26, p=0.019]. The
baseline condition of day 1 was not significantly different from the baseline condition of day 2
(p=0.41, memory) or the acquisition condition of day 2 (p=0.31, learning). However, when
controlling for multiple comparisons, the acquisition condition of day 2 trended toward a
significant increase as compared to the acquisition condition of day 1 (p=0.02, change in
performance). This can be interpreted as the FT group did not remember or learn the
contingency across days, but they did improve their performance on day 2 as compared to day 1.

Day 1 Day 2

Base Acq Base Acq Ext

Figure 5.6 Full-term infants: mean percent reinforced leg action by condition. N=14.
Error bars are standard error. Base=baseline, Acq=acquisition, Ext=extinction.
Base Acq 1 Acq 2 Acq 3
0
10
20
30
40
50
60
Base Acq
0
10
20
30
40
50
60
FT
Base Acq Ext
0
10
20
30
40
50
60
FT
% reinforced leg action (RLA)
p=0.02
INFANT LEG COORDINATION 121

Preterm group at 3 months.
Performance. Percent of RLA during each interval of day 1 and day 2 is graphed in
Figure 5.7. A box-and-whisker plot was used to depict the data because the data was not
normally distributed. Due to the small sample size (n=4), performance, memory, learning, and
change in performance were not statistically analyzed.

Day 1 Day 2

Base Acq1 Acq2 Acq3 Base Acq1 Acq2 Acq3 Ext
Figure 5.7 Preterm infants at 3 months: mean percent reinforced leg action by interval.
N=4. Base=baseline, Acq=acquisition, Ext=extinction.

Preterm group at 4 months.
Performance. Although the preterm group was small (n=6), the data was normally
distributed. Percent of RLA during each interval of day 1 and day 2 is graphed in Figure 5.8.
The main effect of interval was not significant on day 1 [F
(3,15)=
2.04, p=0.15] or day 2
[F
(4,20)
=0.50, p=0.73]. This can be interpreted as the PT group at 4 months did not demonstrate a
change in performance day 1 or day 2.
Base Acq 1 Acq 2 Acq 3
0
10
20
30
40
50
60
70
Base Acq 1 Acq 2 Acq 3
0
10
20
30
40
50
60
70
Base Acq 1 Acq 2 Acq 3 Ext
0
10
20
30
40
50
60
70
% reinforced leg action (RLA)
INFANT LEG COORDINATION 122
Memory, learning, change in performance. The main effect of condition was not
significantly different over the course of the two days [F
(4,20)
=1.03, p=0.42]. This can be
interpreted as the PT group at 4 months did not remember or learn the contingency across days,
and did not improve their performance on day 2 as compared to day 1.

Day 1 Day 2

Base Acq1 Acq2 Acq3 Base Acq1 Acq2 Acq3 Ext

Figure 5.8 Preterm infants at 4 months: mean percent reinforced leg
action by interval. N=6. Error bars are standard errors.

Summary of performance and learning results. The FT group at 3 months did not
demonstrate increased performance on day 1 and did not remember or learn the contingency
across days, but they did demonstrate increased performance on day 2. The PT group at 3
months was too small to analyze statistically. The PT group at 4 months did not demonstrate a
performance change either day and did not remember or learn the contingency across days.
Coordination Changes of Learners and Non-Learners
Participants. Since we were interested in determining whether FT and PT infants who
learned the contingency exhibited more out-of-phase hip-knee joint coordination when leg
Base Acq 1 Acq 2 Acq 3
0
10
20
30
40
50
60
Base Acq 1 Acq 2 Acq 3
0
10
20
30
40
50
60
Base Acq 1 Acq 2 Acq 3 Ext
0
10
20
30
40
50
60
PT
% reinforced leg action (RLA)
INFANT LEG COORDINATION 123
actions were reinforced with mobile activation, we separated all 20 infants into infants who
learned the contingency (Learners) and infants who did not learn the contingency (Non-
Learners). The infants were separated based on the individual learning criteria defined as
percent of RLA during the acquisition condition of day 2 equal to or greater than 1.5 times the
baseline condition of day 1. Note that there were 4 PT infants that participated in the study both
at 3 and 4-months. We separated these infants into Learner and Non-Learner groups based on
when they first met learning criteria, or if they did not meet learning criteria when they first
participated in the task. Therefore, the Learner group consisted of 8 infants: 5 FT infants at 3
months, 1 PT infant at 3 months, and 2 PT infants at 4 months. The Non-Learner group
consisted of 12 infants: 9 FT infants at 3 months, 2 PT infants at 3 months, and 1 PT infant at 4
months.
Hip-knee joint coordination. The dependent variables from the kinematic data were
computed from the 13,295 leg movements that met the criteria of a kick. Table 5.2 includes the
number of kicks analyzed for each group in each condition.

Table 5.2
Learners versus Non-Learners. Number of Kicks Analyzed During each Condition
Day 1 Day 2 Total
Baseline Acquisition Baseline Acquisition Extinction
Learners 455 1298 670 1887 745 5055
Non-Learners 902 2398 973 2822 1145 8240
Total 1357 3696 1643 4709 1890 13295
Note. Learners=8 infants, Non-Learners = 12 infants.
INFANT LEG COORDINATION 124

Day1 Day 2

Base Acq1 Acq2 Acq3 Base Acq1 Acq2 Acq3 Ext
Figure 5.9 Learners versus non-learners: mean Fisher Z correlation
coefficients of hip-knee pair by interval. Error bars are standard error.
* = within-group difference compared to baseline Day 2, p < 0.001
** = between-group difference, p < 0.002

Correlation coefficients.
Performance. Least squared means and standard error of the Fisher Z hip-knee
correlation coefficient across intervals each day is graphed in Figure 5.9. The interaction of
INTERVAL*GROUP was significant for Day 1 [F
(3,54)
=7.97, p=0.0002] and Day 2 [F
(4,72)
=8.26,
p<0.0001]. On day 1, between-group and within-group preplanned comparisons were non-
significant when adjusting for multiple comparisons. On day 2, between-group hip-knee
correlation coefficients during the third acquisition interval were significantly decreased
in the Learner group as compared to the Non-Learner group (p=0.002). Within-group, the
Learner group demonstrated a significant decrease in hip-knee correlation coefficients in all three
acquisition intervals as compared to baseline (p<0.001, performance). Within-group, the Non-
Base Acq 1 Acq 2 Acq 3
0. 00
0. 30
0. 60
0. 90
1. 20
1. 50
Base Acq 1 Acq 2 Acq 3 Ext
0. 00
0. 30
0. 60
0. 90
1. 20
1. 50
Learners
Non-l earners
Fisher Z hip-knee correlation coefficient
* * *
**
INFANT LEG COORDINATION 125
Learner group did not demonstrate a significant change between baseline and acquisition
intervals (p>0.40, performance).
This can be interpreted as the Learner group demonstrated a lower hip-knee correlation
coefficient when interacting with the mobile on day 2, but not day 1. The Non-Learner group
did not demonstrate a change in hip-knee coordination either day.
Memory, learning, change in task performance. Least squared means and standard
error of the Fisher Z hip-knee correlation coefficient across conditions each day is graphed in
Figure 5.10. Statistical results confirmed an interaction effect of CONDITION*GROUP for the
hip-knee pair [F
(4, 72)
= 53.4, p<0.0001]. Between-groups, the only difference between the
Learners and Non-Learners was during the day 2 acquisition condition with the Learners
demonstrating a significant decrease in hip-knee correlation coefficients in comparison to the
Non-Learners (p<0.002). Within-group, the Learners did not demonstrate a significant decrease
in hip-knee correlation coefficients between day 2 baseline and day 1 baseline (p=0.12,
memory), but did demonstrated a significant decrease in hip-knee correlation coefficients
between day 2 acquisition and day 1 baseline (p<0.0001, learning) and between day 2 acquisition
and day 1 acquisition (p<0.0001, change in performance). Within-group, the Non-Learners
demonstrated a significant increase in hip-knee correlation coefficients between day 2 baseline
and day 1 baseline (p=0.0002, memory), day 2 acquisition and day 1 baseline (p<0.0001,
learning), and day 2 acquisition and day 1 acquisition (p<0.0001, change in performance).

INFANT LEG COORDINATION 126

Base Acq Base Acq Ext

Figure 5.10 Learners versus non-learners: mean Fisher Z correlation coefficients
of hip-knee pair by condition. Error bars are standard error.
* = Within-group difference compared to baseline Day 2, p < 0.001.
** = Between-group difference, p < 0.002.

This can be interpreted as the Learner group demonstrated a decreased hip-knee
correlation coefficient when interacting with the mobile on day 2 as compared to both the
baseline and acquisition conditions of day 1. Conversely, the Non-Learner group demonstrated
an increased hip-knee correlation coefficient when interacting with the mobile on day 2 as
compared to both the baseline and acquisition conditions of day 1.
Relative Phase.
Performance. Least squared means and standard error of relative phase at each of the
five discrete points across intervals each day is included in Table 5.3. On day 1 at each discrete
point, between-group and within-group preplanned comparisons of hip-knee relative phase were
Base Acq Ext
0. 00
0. 30
0. 60
0. 90
1. 20
1. 50
Learners
Non-Learners
Base Acq
0. 00
0. 30
0. 60
0. 90
1. 20
1. 50
Learners
Non-l earners
Base Acq Ext
0. 00
0. 30
0. 60
0. 90
1. 20
1. 50
Learners
Non-Learners
Fisher Z hip-knee correlation coefficient
**
p<0.0001
p<0.0001
INFANT LEG COORDINATION 127
generally non-significant when adjusting for multiple comparisons. On day 2 between-group, the
Learner group, as compared to the Non-Learner group, demonstrated significantly higher values
of hip-knee relative phase during acquisition intervals (more out-of-phase hip-knee joint
coordination). Within-group, the Learner group demonstrated these higher values as compared
to baseline, the hip-knee relative phase at each of the five discrete points during the acquisition
intervals was generally increased as compared to baseline. Within-group, the Non-Learner group
demonstrated no change in hip-knee relative phase between baseline and acquisition intervals.
This is consistent with the results from the correlation coefficients that the Learner group
demonstrated a greater amount of out-of-phase hip-knee joint coordination when interacting with
the mobile on day 2, but not on day 1 and that the Non-Learner group did not demonstrate a
change in hip-knee joint coordination either day. It is also consistent with the between-group
comparison in which the Learner group demonstrated more out-of-phase hip-knee joint
coordination during the third acquisition interval of day 2.
Memory, learning, change in task performance. Least squared means and standard
error of hip-knee relative phase at each of the five discrete points across conditions each day is
included in Table 5.4. Between-groups, the only difference between the Learners and Non-
Learners was during the day 2 acquisition condition with the Learners demonstrating
significantly higher values of hip-knee relative phase during all 5 discrete points (p<0.001, more
out-of-phase hip-knee joint coordination). Within-group, the Learners did not demonstrated
INFANT LEG COORDINATION 128
Table 5.3

Learners versus Non-Learners: Relative Phase of Hip-Knee Pair by Interval

a. Day 1
Kick
initiation
Mean (SE)
Peak velocity
1
st
half of kick
Mean (SE)
Hip joint
reversal
Mean (SE)
Peak velocity
2
nd
half of kick
Mean (SE)
Kick
end
Mean (SE)
Day 1 Baseline
Learners
Non-Learners

63.0 (9.2)
60.5 (7.5)

55.4 (9.2)
51.3 (7.4)

55.5 (10.2)
52.4 (8.3)

55.6 (10.5)
51.1 (8.5)

61.3 (8.2)
58.3 (6.7)
Day 1 Acquisition 1
Learners
Non-Learners

56.1 (9.3)
62.3 (7.5)

50.6 (9.2)
57.8 (7.5)*

48.6 (10.3)
55.0 (8.3)

48.0 (10.6)
54.3 (8.6)

60.0 (8.3)
62.9 (6.7)
Day 1 Acquisition 2
Learners
Non-Learners

68.1 (9.3)
62.6 (7.5)

62.7 (9.2)
56.3 (7.5)

62.0 (10.2)
53.1 (8.3)

62.8 (10.6)
53.5 (8.6)

68.8 (8.3)
61.5 (6.7)
Day 1 Acquisition 3
Learners
Non-Learners

64.3 (9.2)
56.3 (7.5)

59.2 (9.1)
50.1 (7.4)

60.1 (10.1)
48.6 (8.3)

59.5 (10.5)
49.0 (8.6)

66.8 (8.2)
58.4 (6.7)

b. Day 2
Kick
initiation
Means (SE)
Peak velocity
1
st
half kick
Means (SE)
Hip joint
reversal
Means (SE)
Peak velocity
2
nd
half kick
Means (SE)
Kick
end
Means (SE)
Day 2 Baseline
Learners
Non-Learners

66.7 (6.5)
47.3 (5.3)

64.6 (7.4)
46.0 (6.1)

63.8 (7.7)
42.6 (6.3)

63.0 (7.6)
41.0 (6.2)

67.3 (5.9)
51.2 (4.9)
Day 2 Acquisition 1
Learners
Non-Learners

80.2 (6.6)**
49.2 (5.4)***

72.7 (7.5)**
45.7 (6.1)

71.8 (7.8)**
42.7 (6.3)***

72.1 (7.7)**
38.9 (6.2)***

74.7 (6.0)
51.6 (4.9)***
Day 2 Acquisition 2
Learners
Non-Learners

74.0 (6.5)
49.9 (5.3)***

70.7 (7.4)
45.9 (6.1)

71.8 (7.7)**
42.7 (6.3)***

71.7 (7.6)**
41.7 (6.2)***

75.2 (5.9)
51.7 (4.9)***
Day 2 Acquisition 3
Learners
Non-Learners

79.8 (6.5)**
50.9 (5.3)***

73.5 (7.4)**
46.1 (6.1)***

73.1 (7.8)**
41.3 (6.3)***

72.6 (7.6)**
38.7 (6.2)***

73.4 (6.0)
47.4 (4.9)***
Day 2 Extinction
Learners
Non-Learners

67.7 (6.5)
50.1 (5.3)

62.7 (7.4)
49.6 (6.1)

64.1 (7.7)
46.6 (6.3)

64.2 (7.6)
44.8 (6.2)

68.5 (5.9)
52.0 (4.8)
Note. SE=standard error.
* = Significantly different from Non-Learner group baseline day 1, p <0.006
* * = Significantly different from Learner group baseline day 2, p <0.006
*** = Significantly different from Learner group, p <0.006

INFANT LEG COORDINATION 129
Table 5.4
Learners versus Non-Learners: Relative Phase of Hip-Knee Pair by Condition
Kick
Initiation

M (SE)
Peak velocity
1
st
half of kick

M (SE)
Hip joint
Reversal

M (SE)
Peak velocity
2
nd
half of kick

M (SE)
Kick
End

M (SE)
Day 1 Baseline
Learners
Non-Learners

64.4 (6.7)*
60.3 (5.4)**

57.1 (6.8)*
52.6 (5.5)**

57.1 (7.5)*
53.2 (6.1)**

57.7 (7.6)*
51.8 (6.1)**

62.5 (6.0)*
58.3 (4.8)**
Day 1 Acquisition
Learners
Non-Learners

64.1 (6.4)*
60.0 (5.2)**

58.7 (6.6)*
55.6 (5.4)**

58.3 (7.3)*
52.7 (5.9)**

58.4 (7.4)*
52.7 (6.0)**

66.3 (5.6)*
61.0 (4.6)**
Day 2 Baseline
Learners
Non-Learners

65.9 (6.6)
44.7 (5.4)

63.6 (6.7)
42.6 (5.5)

62.7 (7.4)
39.3 (6.1)

61.9 (7.5)
37.8 (6.1)

66.7 (5.8)
48.6 (4.8)
Day 2 Acquisition
Learners
Non-Learners

76.3 (6.4)***
47.6 (5.2)

70.5 (6.6)***
42.3 (5.4)

70.5 (7.3)***
38.7 (5.9)

70.3 (7.3)***
36.6 (6.0)

73.2 (5.6)***
47.5 (4.6)
Day 2 Extinction
Learners
Non-Learners

65.6 (6.6)
48.1 (5.3)

60.5 (6.7)
46.7 (5.5)

61.7 (7.4)
43.9 (6.0)

61.7 (7.5)
42.3 (6.1)

66.7 (5.8)
49.8 (4.7)

Note. SE=standard error.
* = Significantly different from Learner acquisition day 2, p <0.001
** = Significantly different from Non-Learner acquisition day 2, p <0.001
*** = Significantly different from Non-Learner group acquisition day 2, p <0.001

higher values of hip-knee relative phase at each of the 5 discrete points between day 2 baseline
and day 1 baseline (p>0.01, memory), but did demonstrate significant differences between day 2
acquisition and day 1 baseline (p<0.001, learning) as well as between day 2 acquisition and day 1
acquisition (p<0.001, change in performance). Within-group, the Non-Learners demonstrated a
significant decrease in relative phase (more in-phase hip-knee joint coordination) between day 2
baseline and day 1 baseline (p<0.001, memory), day 2 acquisition and day 1 baseline (p<0.001,
learning), and day 2 acquisition and day 1 acquisition (p<0.001, change in performance). This
can be interpreted as the Learner group, but not the Non-Learner group, demonstrated a greater
INFANT LEG COORDINATION 130
amount of out-of-phase hip-knee joint coordination patterns when interacting with the mobile on
day 2 as compared to baseline leg action of day 1.
This is consistent with the results from the correlation coefficients that the Learner group
demonstrated a greater amount of out-of-phase hip-knee joint coordination patterns when
interacting with the mobile on day 2 as compared to both the baseline and acquisition conditions
of day 1. Conversely, the Non-Learner group demonstrated a greater amount of in-phase hip-
knee joint coordination patterns when interacting with the mobile on day 2 as compared to both
the baseline and acquisition conditions of day 1.
Relation between Learning and Coordination Changes
To depict the relation of learning and coordination changes, we graphed the difference in
percent RLA from the Day 2 acquisition condition to the Day 1 baseline condition (our
individual learning criteria) and difference in hip-joint coordination from the Day 2 acquisition
condition to the Day 1 baseline condition in Figure 5.12. Six of 8 infants in the Learner group
decreased their hip-knee correlation coefficient, including all 3 of the PT infants, whereas only 3
of 12 infants in the Non-Learner group decreased their hip-knee correlation coefficient.
Arousal and Attention Differences between Learners and Non-Learners
Arousal. The interaction of CONDITION*GROUP [F
(4,72)
=0.65, p=0.63] and the main
effect of GROUP [F
(1,18)
=0.30, p=0.59] were not significant. The main effect of CONDITION
was significant [F
(4,72)
=8.21, p<0.0001]. Infants in both groups were classified as alert (a score
of 2 or 3 on the arousal scale) for 94 – 98 percent (SE 4 percent) of the experimental time except
during the extinction condition of day 2 which was 74 percent (SE 4 percent), significantly
different from all other conditions (p<0.0001). This can be interpreted as there was no
INFANT LEG COORDINATION 131

Figure 5.11 Scatterplot of difference in reinforced leg action (RLA)
between day 2 acquisition condition and day 1 baseline and difference
in hip-knee correlation coefficient between day 2 acquisition condition
and day 1 baseline.

difference between the Learner and Non-Learner groups in terms of their arousal during the
mobile paradigm, however, both groups of infants were significantly more fussy or crying during
the extinction condition as compared to baseline and acquisition conditions.
Attention. In both groups, the mean estimates of the percent of time during each
condition in which the infant was looking at the mobile ranged from 91 – 97 percent (SE 2
percent). The interaction of CONDITION*GROUP [F
(4,72)
=2.15, p=0.08], the main effect of
GROUP [F
(1,18)
=1.26, p=0.28], and the main effect of CONDITION [F
(4,72)
=2.24, p=0.07], were
not significantly different. This can be interpreted as there was no difference between the
Learner and Non-Learner groups in terms of their attention during the mobile paradigm.

- 45 - 36 - 27 - 18 -9 0 9 18 27 36 45
-2
-1
0
1
2
Lear ner s
N on- Lear ner s
Change in hip-knee correlation coefficient
Change in percent of reinforced leg action (RLA)

Learners
Non-Learners
INFANT LEG COORDINATION 132
Discussion
We hypothesized that the FT group would demonstrate performance of the contingency
on the first day and both memory and learning of the contingency on the second day. Contrary
to our hypothesis, the FT infants as a group demonstrated improved performance on the second
day, not the first day as reported in previous research. In addition, they did not demonstrate
either memory or learning of the contingency. There are two possible explanations for these
results. First, in our mobile paradigm, the infants independently discover the contingency as
their exploratory leg actions activate the mobile. In previous paradigms, FT infants were often
shown that the mobile moved either through the investigator passively guiding the leg (Chen et
al., 2002), a shaping reinforcement schedule (Angulo-Kinzler, 2001), or activation of the mobile
by the investigator at the beginning of the acquisition condition (Angulo-Kinzler et al., 2002).
Our research is grounded within a Perception-Action framework. Thus we allowed infants to
discover the contingency without external assistance. We believe that our experiment more
closely approximates the infant learning environment although it may have resulted in infants
requiring more time to demonstrate performance of the contingency. This raises the question for
intervention programs as to the benefit of demonstration versus self-discovery. If the immediate
task is to be learned, then demonstration of a contingency or aspects of the contingency (e.g. that
the mobile can move) may be the most rapid method for task learning. It is not clear that this
type of intervention would generalize to other tasks. However, discovery learning such as
provided in our mobile paradigm, may provide an opportunity for learning a process that could
be generalized to the learning of other tasks. This could be empirically addressed with an
experimental protocol that first offers a contingency task to be learned. One group of infants
could be shown or guided through the contingency while a second group was left to discover the
INFANT LEG COORDINATION 133
contingency. After the initial contingency is introduced and learned additional contingency tasks
are then tested. We hypothesize that the time needed to learn subsequent contingency tasks
would be reduced for the group “learning to learn” through discovery actions in comparison to
the group who is guided through the task. Adolph and colleagues have conducted a series of
studies in which infants were exposed to variable and novel challenges of balance and
locomotion, such as descending slopes, spanning gaps, and crossing bridges (Adolph, 2008).
Overall, they found that when infants first acquired a new posture, such as walking, they did not
take into account the limits of their abilities relative to risky environmental features, for example
they would attempt to walk down steep slopes (Adolph, 1997). Over weeks of walking
experience, their responses became more adaptive and they would attempt safe slopes and refuse
steep slopes. In fact, infants with walking experience could even respond adaptively to the novel
experience of descending a slope wearing either a lead-weighted or feather-weighted shoulder
pack. Experienced walkers demonstrated that they generalized their knowledge of slopes by
immediately walking down relatively steep slopes with the feather-weighted shoulder packs, but
refusing to walk down the same slope with the heavier, lead-weighted shoulder packs
(Garciaguirre, Adolph, & Strout, 2007). This demonstrates that they were able to generalize
their knowledge of walking down slopes to a novel task utilizing shoulder packs of varying
weight. They had “learned to learn.”
Another possible explanation why FT infants did not perform our task the first day is that
our paradigm requires more specified control of leg action than is typically required in the
mobile paradigm. In previous paradigms, in order to demonstrate performance of the
contingency infants simply needed to increase the frequency of leg actions that were within their
preferred movement repertoire of in-phase intralimb coordination, such as increasing kicking rate
INFANT LEG COORDINATION 134
(Heathcock et al., 2004; Rovee-Collier & Gekoski, 1979), demonstrating flexion or extension of
the knee (Angulo-Kinzler et al., 2002; Tiernan & Angulo-Barroso, 2008) or demonstrating in-
phase flexion or extension of the hip and knee joints (Angulo-Kinzler, 2001). Our mobile
paradigm required an out-of-phase hip-knee joint coordination pattern, which is supported by the
fact that 75 percent of infants who demonstrated learning, as compared to 25 percent of infants
who didn’t demonstrate learning, increased the amount of out-of-phase hip-knee coordination
when interacting with the mobile on the second day. This was supported in our previous
research with FT infants who changed their coordination patterns based on task demands, but
only after passive guidance through the contingency was provided (Chen et al., 2002). This
change in coordination pattern may have been difficult for the FT infants to independently
discover in one, 6-minute testing session.
Our second hypothesis was that the PT group would not demonstrate performance of the
contingency on the first day or learning and memory on the second day, but would demonstrate
performance on the second day. We did not have a sufficient sample size of 3-month PT infants
to statistically analyze for group performance, memory, and learning changes. However based
on individual learning criteria, 1 of 4 PT infants at 3-months learned the contingency. This
demonstrates that PT infants can learn the task. A notable finding was the difference in
proportion of 3-month old FT and PT infants excluded from the study due to excessive crying,
43 percent in the PT group compared to 18 percent in the FT group. Difficulty regulating arousal
level has been documented in PT infants during the mobile paradigm and has been associated
with slower rates of contingency learning (Haley et al., 2008; Haley et al., 2006). Specifically,
PT infants, as compared to FT infants, looked less at the mobile, demonstrated higher basal heart
rates, showed greater increases in heart rate responses to the contingency, and demonstrated
INFANT LEG COORDINATION 135
slower rates of contingency learning (Haley et al., 2008). The researchers suggested that PT
infants began the paradigm in a heightened state of arousal, as supported by the higher basal
heart rates, then became over-aroused during the mobile paradigm, exhibiting greater increases
in heart rate responses to the contingency, and looked away from the mobile in an attempt to
regulate their arousal (Haley et al., 2008). They speculated that this strategy to regulate arousal
contributed to the slower rate of learning of the PT group and may be one factor underlying the
learning difficulties of PT infants. Our results support the conclusion that PT infants at 3
months, as compared to FT infants, have more difficulty regulating their arousal level to
participate in the mobile task without excessive crying, but our sample size was insufficient to
investigate whether arousal and attention differences between FT and PT infants contributed to
their differences in task performance.
Due to the number of PT infants excluded from the study due to excessive crying, we
chose to test the PT infants again at 4 months of age. At 4 months, 6 of 6 PT infants participated
in the task without excessive crying, versus 4 of 7 PT infants at 3-months. In addition, 3 of 6
infants at 4-months learned the contingency, versus 1 of 4 infants at 3-months. These results
again raise the question regarding the standard practice of matching PT and FT infants in
research studies based on the criterion of conceptual age (Gekoski et al., 1984). It is notable that
2 of the 3 PT infants that demonstrated excessive crying in our task were our earliest PT infants,
born at 24 weeks gestation. Perhaps micro-preemies require more age adjustment than PT
infants born at later gestational ages. Matching for maturity of state control might be one
method to adjust that applies to all gestational ages.
Our third hypothesis was that FT and PT infants who learned the contingency, as
compared to infants who did not learn the contingency, would exhibit more out-of-phase hip-
INFANT LEG COORDINATION 136
knee joint coordination when leg actions were reinforced with mobile activation. The hypothesis
was confirmed using both an analysis of correlation coefficients and relative phase. Infants who
learned the task demonstrated an increase in out-of-phase hip-knee joint coordination during the
reinforcement condition on the second day compared to both the baseline condition and the
reinforcement condition on the first day. Infants who learned the task included 5 FT infants at 3
months, 1 PT infant at 3 months, and 3 PT infants at 4 months (1 PT infant demonstrated
learning at both 3-months and 4-months, so we only analyzed his 3-month data). In terms of
individual infants, 6 of 8 infants in the Learner group decreased their hip-knee correlation
coefficient, including all 3 of the PT infants, whereas only 3 of 12 infants in the Non-Learner
group decreased their hip-knee correlation coefficient. This supports that FT and PT infants as
young as 3 months of age can learn the task and demonstrate more out-of-phase hip-knee joint
coordination when interacting with the mobile.
Of note is that the hip-knee joint coordination of the Learner group became more in-
phase during the initial 2 minutes of interacting with the mobile on the first day, and then the
coordination became more out-of-phase until it reached significance during the second day of
interacting with the mobile. This increase in in-phase hip-knee joint coordination was also
observed in our previous work (Chen et al., 2002). Four-month old FT infants demonstrated an
overall increase of in-phase hip-knee joint coordination when interacting with the mobile,
although the kicks that contacted the plate and activated the mobile were more out-of-phase. We
speculate that mobile reinforcement might enhance an infant’s state of arousal (Rovee-Collier &
Gekoski, 1979) which could increase the kicking rate (Heriza, 1991; Thelen et al., 1981); a high
kick rate has been associated with a more in-phase kicking pattern (Heriza, 1991; Thelen, 1985).
However, all of the 3-month old FT infants in this study that learned the task decreased the
INFANT LEG COORDINATION 137
number of times they crossed the threshold, and subsequently their duration of mobile activation,
by greater than 50% during the first 2 minutes of interacting with the mobile during the first day
of acquisition. The majority of FT Non-Learners also demonstrated a decrease but not as large
as the FT Learners. These results suggest that when mobile reinforcement began the 3-month
old FT infants actually decreased their kicking rate, or more precisely, the height of their kicking.
We too noticed an overall decrease in infant movement when the mobile first began to activate.
Perhaps in the infants’ effort to determine whether their movement was associated with mobile
activation, they decreased the amount of exploratory leg movement and reverted to a more in-
phase kicking pattern. The use of more in-phase joint coordination has been proposed as a
means of reducing the number of independent joint variables to be controlled during the initial
stages of learning (Bernstein, 1967; Sporns & Edelman, 1993). As the learner interacts within
the task constraints, a more out-of-phase movement pattern emerges that more closely
corresponds with the requirements of the task. An essential aspect of infant contingency learning
may be to decrease overall movement to observe the environmental stimuli that is occurring and
then to explore movement options to determine which movements could be exploited to improve
task performance. Specifically, in the case of our mobile paradigm, although the mobile could
be activated using an in-phase strategy of flexing and extending the hip and knee together, a
more efficient strategy may have been an out-of-phase strategy of maintaining the hip flexed and
flexing and extending the knee. This out-of-phase strategy may have been more difficult for the
Non-Learner group to generate, which constrained their ability to demonstrate learning of the
contingency.
One clinical implication of this work is that, as early as 3 months of age, PT infants at
increased risk for cerebral palsy and other movement disorders may be able to generate a greater
INFANT LEG COORDINATION 138
amount of out-of-phase hip-knee joint coordination when participating in contingent learning
paradigms. Results from this research support that these interventions would need to be
provided over the course of days in environments and at times in which the infant is best able to
regulate arousal levels.

INFANT LEG COORDINATION 139
CHAPTER VI
SUMMARY AND GENERAL DISCUSSION

Summary and Discussion
The purpose of this dissertation is to gain a better understanding of the contribution of
torque changes to the early changes in leg joint coordination of typically-developing full-term
infants, clarify differences between the leg joint coordination and torques of preterm and full-
term infants, and determine whether preterm infants have the potential for a greater amount of
out-of-phase leg joint coordination when interacting with an infant-activated mobile that
reinforces out-of-phase leg joint coordination. Three studies were conducted to answer these
questions. The first study was a longitudinal research study which investigated the contribution
of torque changes to the early changes in leg joint coordination of typically-developing full-term
infants from 6 to 15-weeks of age. The second study was a longitudinal research study which
investigated the contribution of torque changes to the early changes in leg joint coordination of
preterm infants from 6 to 15-weeks of age. The final study investigated the ability of 3 to 4-
month old FT and PT infants to demonstrate a greater amount of out-of-phase hip-knee joint
coordination when participating in an innovative learning task using an infant-activated mobile.
Following is a discussion of each research study highlighting novel findings, implications for
clinical practice, and opportunities for future research.
Changes in Infant Leg Coordination: Insights from a 3D Kinematic and Kinetics Approach
Novel findings. In this study, for the first time, standard three-dimensional (3D)
kinematics were used to compute complete 3D kinematic data of infant spontaneous kicks from 6
to 15-weeks of age. This is a significant contribution to infant research because coordination
INFANT LEG COORDINATION 140
between joint angle pairs which include the rotational aspects of infant leg motion, such as hip
abduction/adduction, hip external/internal rotation, ankle eversion/inversion, had not yet been
investigated due to limitations in the method used to collect and analyze position data from
infant spontaneous kicks. We found that in full-term infants from 6 to 15-weeks of age, 11 of 15
joint angle pairs demonstrated a change from an in-phase intralimb coordination pattern at 6-
weeks to an increased variety of intralimb joint coordination patterns at 15-weeks. The only
joint angle pairs which did not demonstrate a greater variety of joint coordinations were those
which included hip external/internal rotation. We suggest that hip external/internal rotation may
be more influenced by the intrajoint motions of hip flexion/extension and abduction/adduction
due to hip muscle activations resulting in multiple actions at the hip.
Also in this study, for the first time, Lagrangian mechanics were used to determine the
3D kinetics of infant spontaneous kicks from 6 to 15-weeks of age. We found that, unlike joint
coordination, there is not an obvious developmental change in joint torques between 6 and 15-
weeks of age. We suggest that the lack of consistent relationship of joint torque changes
between 6 and 15-weeks could be due to differences in the coordination of the kicks between
ages which confound the torque analysis; essentially different kicking motions are being
analyzed.
The most innovative aspect of this study is the combination of 3D kinematic and kinetic
analyses to investigate the relation between intersegmental dynamics and hip-knee joint
coordination in FT infants from 6 to 15-weeks of age. We found that at 6-weeks the correlation
coefficient between the hip and knee was 1.75 (SE=0.12). This in-phase coordination was
accompanied by a long duration and high magnitude of knee muscle (MUS) torque as compared
to the passive knee motion-dependent (MDT) and gravitational (GRA) torques. At 15-weeks,
INFANT LEG COORDINATION 141
the correlation coefficient had decreased to 0.32 (SE=0.12). This less in-phase coordination was
accompanied by a shorter duration and lower magnitude of knee MUS torque as compared to
passive knee MDT and GRA torques. We suggest that for FT infants between 6 and 15-weeks of
age the decreased influence of MUS torque generated at the knee allowed the passive knee
gravitational and motion dependent torques to have a greater influence on the coordination of the
kick, contributing to a greater variety of hip-knee joint coordinations.
Opportunities for future research. In order to more fully understand the relation
between changes in joint coordination and changes in intersegmental dynamics, two research
directions are proposed. From our current study we believe that a decrease in muscle co-
activation between the hip and the knee and ankle allows the passive torques to have a greater
influence on the kicking coordination contributing to a greater variety of hip-knee joint
coordinations. This hypothesis could be tested with the addition of EMG to our 3D kinematic
and kinetic analysis.
We also believe that the greater variety of hip-knee joint coordinations observed in the
spontaneous kicking of 15-week old infants is associated with an increased influence of passive
knee GRA and MDT torques as compared to knee MUS torque. However, it is unknown how
the intersegmental dynamics change when infants are actively trying to perform a task which
requires more out-of-phase hip-knee joint coordination. We hypothesize that knee MUS torque
would increase to dampen the passive knee GRA and MDT torques during the initial stage of
learning. This is consistent with Bernstein’s work in which excessive co-activation is used in the
initial stages of learning and as a result reduce the number of independent joint variables to be
controlled (Bernstein, 1967; Sporns & Edelman, 1993). However, as infants interact within task
constraints, we expect that infants will decrease muscle co-activation and exploit the passive
INFANT LEG COORDINATION 142
GRA and MDT forces to achieve the out-of-phase leg joint coordination required by the task in a
more energy-efficient manner.
Changes in Infant Leg Coordination: Influence of Prematurity
Novel findings. This study is a unique contribution to the literature since it is the first
time that 3D kinematic and kinetics have been used to investigate the contribution of torque
changes to the early changes in leg joint coordination in a population of infants with known
differences in joint coordination from typically-developing FT infants. We found that PT infants
demonstrated less in-phase coordination in 30% of joint angle pairs versus 73% of joint angle
pairs in FT infants. At 6-weeks PT infants as compared to FT infants demonstrated less in-phase
coordination in 4 of 15 joint angle pairs, all which included hip or knee flexion/extension paired
with ankle eversion/inversion or plantarflexion/dorsiflexion. We suggest that at 6-weeks PT
infants may exhibit a greater variety of joint coordinations than FT infants due to additional
exposure to the extra-uterine environment. By 15-weeks of age, we found no differences in joint
coordination between FT and PT infants.
Similar to FT infants, PT infants demonstrated relatively consistent normalized net (NET)
joint torque and partitioned joint torque data from 6 to 15-weeks of age. We also found that
although PT infants exhibited a greater variety of hip-knee coordinations from 6-weeks to 15-
weeks of age, no differences were noted in the influence of knee MUS, GRA, and MDT torques
on knee NET torques. However, the PT group as compared to the FT group exhibited a smaller
change in coordination from 6-weeks to 15-weeks which may not have been sufficient to
document the relation between intersegmental dynamics and joint coordination.
Opportunities for future research. In order to more fully understand the relation
between changes in joint coordination and changes in intersegmental dynamics in high-risk
populations, two research directions are proposed. The kinetic analysis of a goal-directed action,
INFANT LEG COORDINATION 143
such as foot reaching, with the progressive addition of leg weights, may clarify differences in the
maximum force generation capacity of PT and FT infants from 3 to 6 months of age. It may also
clarify how joint coordination and intersegmental dynamics change as the capacity to generate
muscle forces becomes insufficient for task constraints.
Future research could also investigate the relation between intersegmental dynamics and
joint coordination in PT infants with WMD, at highest risk for differences in joint coordination.
Since changes in leg joint coordination which emerge during infant spontaneous kicking are
complex, a single subject analysis of the spontaneous kicking of a few PT infants with severe
WMD at multiple time points within the first six months of life may be necessary to understand
the relation between intersegmental dynamics and joint coordination.
Infant Exploratory Learning: Influence on Leg Coordination
Novel findings. Our infant-activated mobile paradigm was based on the Perception-
Action framework which proposes that infants learn task-specific action through a discovery
learning process in which infants demonstrate a wide range of exploratory actions to generate
information about possible outcomes of actions, then select actions which result in outcomes
with adaptive value (Gibson & Pick, 2000; Thelen & Smith, 1994). Our infant-activated mobile
is unique in that it exploits this discovery learning process. Infants independently discover the
contingency between mobile activation and their leg action as they spontaneously move their
legs. In previous paradigms, infants were often shown that the mobile moved either through the
investigator passively guiding the leg (Chen et al., 2002), a shaping reinforcement schedule
(Angulo-Kinzler, 2001), or activation of the mobile by the investigator at the beginning of the
acquisition condition (Angulo-Kinzler et al., 2002). We believe that our experiment more
INFANT LEG COORDINATION 144
closely approximates the infant learning environment although it may have resulted in infants
requiring more time to demonstrate performance of the contingency.
Our novel finding was that infants who learned the contingency between their leg action
and mobile activation demonstrated an increase in out-of-phase hip-knee joint coordination when
interacting with the mobile on the second day versus kicking spontaneously during the baseline
condition on the first day. Infants who learned the task included 5 FT infants at 3 months, 1 PT
infant at 3 months, and 3 PT infants at 4 months (1 PT infant demonstrated learning at both 3-
months and 4-months). This supports that FT and PT infants as young as 3 months of age can
demonstrate more out-of-phase hip-knee joint coordination when interacting with an infant-
activated mobile.
Another notable finding was the difference in proportion of 3-month old FT and PT
infants excluded from the study due to excessive crying, 43% in the PT group compared to 18%
in the FT group. Difficulty regulating arousal level has been documented in PT infants during
the mobile paradigm and has been associated with slower rates of contingency learning (Haley et
al., 2008; Haley et al., 2006). Due to the number of PT infants that were excluded from the study
due to excessive crying, we chose to test the PT infants again at 4 months of age. At 4 months, 6
of 6 PT infants participated in the task without excessive crying versus 4 of 7 PT infants at 3-
months. In addition, 3 of 6 PT infants at 4-months learned the contingency, versus 1 of 4 PT
infants at 3-months. These results again raise the question regarding the standard practice of
matching PT and FT infants in research studies based on the criterion of conceptual age (Gekoski
et al., 1984). It is notable that 2 of the 3 PT infants that demonstrated excessive crying in our
task were our earliest PT infants, born at 24 weeks gestation. Perhaps micro-preemies require
INFANT LEG COORDINATION 145
more age adjustment than PT infants born at later gestational ages. Matching for maturity of
state control might be one method to adjust that applies to all gestational ages.
Implications for clinical practice. One clinical implication of this work is that, as early
as 3 months of age, PT infants at increased risk for cerebral palsy (CP) and other movement
dysfunctions may be able to generate a greater amount of out-of-phase hip-knee joint
coordination when participating in contingent learning paradigms. Results from this research
support that these interventions would need to be provided over the course of days in
environments and at times in which infants are best able to regulate their arousal levels.
Opportunities for future research. Eight infants in our study, including one PT infant
at 3 months, demonstrated that they had learned the contingency between leg action and mobile
activation after participating in the mobile paradigm for 6-minutes on 2 consecutive days. It is
surprising that young infants can learn this after such a brief exposure to the contingency. We
collected the data from this study in a way that will allow us to investigate the learning process
of our Learner group in terms of managing the variance and intersegmental dynamics when
attempting to increase activation of the mobile. In our study, the infants activated the mobile by
moving either foot in the vertical direction across a virtual threshold, which was placed
perpendicular to the table (Figure 5.1). Attached to the foot of each infant was an infra-red
emitting diode (IRED) that acted as a trigger for mobile activation, but also allowed us to
consistently track the motion of the IRED on each foot and dissociate the motion into the z-
direction (task-specific action, perpendicular to the table) and the x-direction (non-task-specific
action, parallel to the table). If infants had learned the contingency between leg action and
mobile activation, it would be expected that infants would increase the variance in the z-direction
INFANT LEG COORDINATION 146
(task-specific action) to activate the mobile and decrease variance in the x-dimension (non-task-
specific action).
The data from this study was also collected such that intersegmental torques can be
computed. From our first study, we believe that the out-of-phase hip-knee joint coordination
observed in the spontaneous kicking of 15-week old infants is associated with an increased
influence of passive knee GRA and MDT torques as compared to knee MUS torque. However, it
is unknown how the intersegmental dynamics change when infants are actively trying to perform
a task which requires more out-of-phase hip-knee joint coordination. We could investigate this
question through the comparison of the intersegmental torques from the kicks that did not contact
the table or the other leg during the acquisition condition of the second day (task-specific action)
compared to the baseline condition of the first day (spontaneous kicking). We expect during the
acquisition condition of the second day (task-specific action), as compared to the baseline
condition of the first day (spontaneous kicking), the influence of the knee MUS torque on the
knee NET torque will be increased and the influence of the knee GRA and MDT torques on the
knee NET torque will be decreased.
In order to more fully investigate whether preterm infants with white matter damage
(PTWMD), at highest risk for cerebral palsy (CP), have the potential for a greater amount of out-
of-phase joint coordination, an in-home system may need to be developed. We were only able to
recruit one PTWMD infant for our study. PTWMD infants may have a more complicated
medical course than PT infants without white matter damage, which may make travel to a
research lab impractical, especially for consecutive days. The in-home system we envision will
incorporate wireless motion sensors to monitor the infant’s leg coordination patterns, an e-tablet
to set the reinforcement schedule for specific coordination patterns of the legs, and connection to
INFANT LEG COORDINATION 147
an infant-activated mobile. This system could be implemented in the home for several minutes
each day for a few weeks to determine if prolonged exposure to the contingency encourages
changes in leg joint coordination patterns. The strength of the in-home system is that it
specifically targets out-of-phase leg joint coordination, a primary contributor to the movement
dysfunction of children with CP and an impairment which has been shown to significantly limit
walking abilities. Another strength of the in-home system is that it can be used for multiple
sessions in the home environment at times when the infant is best able to regulate his arousal
levels. In addition, the technology could be expanded to activate any number of electronic toys
to reinforce specific limb movements of infants with a wide range of diagnoses, such as
hemiplegic cerebral palsy, Erb’s palsy, Down syndrome, and myelomeningocele.
Summary
This dissertation has increased our understanding of the contribution of torque changes to
the early changes in leg joint coordination of typically-developing full-term infants, highlighted
the differences between leg joint coordinations and torques of preterm and full-term infants, and
confirmed that PT infants have the potential for a greater amount of out-of-phase leg joint
coordination when interacting with an infant-activated mobile that reinforces out-of-phase leg
joint coordination. The results from this dissertation will be an essential step toward developing
more relevant and successful screening, intervention, and prevention programs for preterm
infants at risk for impairments in motor coordination.

INFANT LEG COORDINATION 148
REFERENCES
Aarnoudse-Moens, C., Weisglas-Kuperus, N., van Goudoever, J., & Oosterlaan, J. (2009). Meta-
analysis of neurobehavioral outcomes in very preterm and/or very low birth weight
children. Pediatrics, 124, 717-728.
Adolph, K. E. (1997). Learning in the development of infant locomotion. Monographs of the
Society for Research in Child Development, 62(3).
Adolph, K. E. (2008). Learning to move. Current Directions in Psychological Science, 17(3),
213-218.
Adolph, K. E., & Anthony, A. (2000). Walking infants adapt locomotion to changing body
dimensions. Journal of Experimental Psychology: Human Perception and Performance,
26(3), 1148-1166.
Angulo-Kinzler, R. (2001). Exploration and selection of intralimb coordination patterns in 3-
month-old infants. Journal of Motor Behavior, 33, 363-376.
Angulo-Kinzler, R., & Horn, C. (2001). Selection and memory of a lower limb motor-perceptual
task in 3-month-old infants. Infant Behavior and Development, 24, 239-257.
Angulo-Kinzler, R., Ulrich, B. D., & Thelen, E. (2002). Three-month-old infants can select
specific leg motor solutions. Motor Control, 6, 52-68.
Bax, M., Tydeman, C., & Flodmark, O. (2006). Clinical and MRI correlates of cerebral palsy:
the European cerebral palsy study. Journal of the American Medical Association, 296,
1602-1608.
Bayley, N. (1969). Manual for the Bayley scales of infant development. New York, NY:
Psychological Corporation.
Bayley, N. (2006). Manual for the Bayley scales of infant and toddler development, 3rd edition.
San Antonio: Psychological Corporation.
INFANT LEG COORDINATION 149
Beaino, G., Khoshnood, B., Kaminski, M., Pierrat, V., Marret, S., Matis, J., . . . Ancel, P.-Y.
(2010). Predictors of cerebral palsy in very preterm infants: The EPIPAGE prospective
population-based cohort study. Developmental Medicine & Child Neurology, 52, e119-
e125.
Bernstein, N. (1967). The coordination and regulation of movement. London: England:
Pergamon Press.
Bhat, A., Heathcock, J., & Galloway, J. C. (2005). Toy-oriented changes in hand and joint
kinematics during the emergence of purposeful reaching. Infant Behavior and
Development, 28(4), 445-465.
Bhat, A., Lee, H., & Galloway, J. C. (2007). Toy-oriented changes in early arm movements II--
Joint kinematics. Infant Behavior and Development, 30, 307-324.
Chapman, D. (2002). Context effects on the spontaneous leg movements of infants with spina
bifida. Pediatric Physical Therapy, 14, 62-73.
Chen, Y., & Fetters, L. (2002). A comparison of the leg coordination patterns of preterm and
fullterm infants: A meta-analysis. R.O.C. (Taiwan) Journal of Physical Therapy, 27, 303-
313.
Chen, Y., Fetters, L., Holt, K., & Saltzman, E. (2002). Making the mobile move: Constraining
task and environment. Infant Behavior and Development, 25, 195-220.
de Graff-Peters, V., & Hadders-Algra, M. (2006). Ontogeny of the human central nervous
system: What is happening when? Early Human Development, 82, 257-266.
de Kieviet, J., Piek, J. P., Aarnoudse-Moens, C., & Oosterlaan, J. (2009). Motor development in
very preterm and very low-birth-weight children from birth to adolescence: A meta-
analysis. Journal of the American Medical Association, 302, 2235-2242.
INFANT LEG COORDINATION 150
Delobel-Ayoub, M., Arnaud, C., White-Koning, M., Casper, C., Pierrat, V., Garel, M., . . . for
the, E. S. G. (2009). Behavioral problems and cognitive performance at 5 years of age
after very preterm birth: The EPIPAGE study. Pediatrics, 123(6), 1485-1492.
Einspieler, C., Prechtl, H., Bos, A. F., Ferrari, F., & Cioni, G. (2004). Prechtl's method on the
qualitative assessment of general movement of general movements in preterm, term, and
young infants. New York: Mac Keith Press.
Fetters, L., Chen, Y., Jonsdottir, J., & Tronick, E. (2004). Kicking coordination captures
differences between full-term and premature infants with white matter disorder. Human
Movement Science, 22, 729-748.
Fetters, L., & Ellis, T. (2006). A perception-action framework for physical therapy for persons
with neurological dysfunction: Use of therapeutic affordance and unitless ratio. Journal
of Neurologic Physical Therapy, 30, 142-147.
Fetters, L., Sapir, I., Chen, Y., Kubo, M., & Tronick, E. (2010). Spontaneous kicking in full-term
and preterm infants with and without white matter disorder. Developmental
Psychobiology, 52, 524-536.
Forssberg, H. (1999). Neural control of human motor development. Current Opinion in
Neurobiology, 9(6), 676-682.
Fowler, E. G., & Goldberg, E. J. (2009). The effect of lower extremity selective voluntary motor
control on interjoint coordination during gait in children with spastic diplegic cerebral
palsy. Gait & Posture, 29(1), 102-107.
Galloway, J. C., & Koshland, G. F. (2002). General coordination of shoulder, elbow, and wrist
dynamics during mulitjoint arm movements. Experimental Brain Research, 142(2), 163-
180.
INFANT LEG COORDINATION 151
Galloway, J. C., & Thelen, E. (2004). Feet first: object exploration in young infants. Infant
Behavior and Development, 27, 107-112.
Garciaguirre, J., Adolph, K. E., & Strout, P. E. (2007). Baby carriage: Infants walking with
loads. Child Development, 78, 664-680.
Gekoski, M. J., Fagen, J. W., & Pearlman, M. A. (1984). Early learning and memory in the
preterm infant. Infant Behavior and Development, 7, 267-276.
Gesell, A., Thompson, H., & Amatruda, C. (1934). Infant behavior: Its genesis and growth. New
York, NY: McGraw-Hill.
Gibson, E. J. (1988). Exploratory behavior in the development of perceiving, acting, and the
acquiring of knowledge. Annual Review of Psychology, 39, 1-41.
Gibson, E. J., & Pick, A. D. (2000). An ecological approach to perception, learning and
development. New York, NY: Oxford University Press.
Grillner, S. (1975). Locomotion in vertebrates: Central mechanisms and reflex interaction.
Physiological Review, 55, 247-304.
Haley, D. W., Grunau, R. E., Oberlander, T., & Weinberg, J. (2008). Contingency learning and
reactivity in preterm and full-term infants at 3 months. Infancy, 13, 570-595.
Haley, D. W., Weinberg, J., & Grunau, R. E. (2006). Cortisol, contingency learning, and
memory in preterm and full-term infants. Psychoneuroendocrinology, 31(1), 108-117.
Hamilton, B., Martin, J., & Ventura, S. (2011). Births: Preliminary data for 2010. National Vital
Statistics Reports. (Vol. 60). Hyattsville, MD: National Center for Health Statistics.
Han, T. R., Bang, M. S., Lim, J. Y., Yoon, B. H., & Kim, I. O. (2002). Risk factors of cerebral
palsy in preterm infants. American Journal of Physical Medicine, 81(4), 297-303.
INFANT LEG COORDINATION 152
Harris, S. R., & Heriza, C. B. (1987). Measuring infant movement: Clinical and technological
assessment techniques. Physical Therapy, 67, 1877-1880.
Heathcock, J., Bhat, A., Lobo, M., & Galloway, J. C. (2004). The performance of infants born
preterm and full-term in the mobile paradigm: Learning and memory. Physical Therapy,
84, 808-821.
Heathcock, J., Bhat, A., Lobo, M., & Galloway, J. C. (2005). The relative kicking frequency of
infants born full-term and preterm during learning and short-term and long-term memory
periods of the mobile paradigm. Physical Therapy, 85, 8-18.
Heathcock, J., & Galloway, J. C. (2009). Exploring objects with feet advances movement in
infants born preterm: A randomized controlled trial. Physical Therapy, 89(10), 1027-
1038.
Heathcock, J., Lobo, M., & Galloway, J. C. (2008). Movement training advances the emergence
of reaching in infants born at less than 33 weeks of gestational age: A randomized
clinical trial. Physical Therapy, 88(3), 310-322.
Heriza, C. B. (1988a). Comparison of leg movements in preterm infants at term with healthy
full-term infants. Physical Therapy, 68, 1687-1693.
Heriza, C. B. (1988b). Organization of leg movements in preterm infants. Physical Therapy, 68,
1340-1346.
Heriza, C. B. (1991). Implications of a dynamical systems approach to understanding infant
kicking behavior. Physical Therapy, 71, 222-236.
Himpens, E., Van den Broeck, C., Oostra, A., Calders, P., & Vanhaesebrouck, P. (2008).
Prevalence, type, distribution, and severity of cerebral palsy in relation to gestational age:
A meta-analytic review. Developmental Medicine and Child Neurology, 50, 334-340.
INFANT LEG COORDINATION 153
Jeng, S., Chen, L., Tsou, K., Chen, W., & Luo, H. (2004). Relationship between spontaneous
kicking and age of walking attainment in preterm infants with very low birth weight and
full-term infants. Physical Therapy, 84, 159-172.
Jeng, S., Chen, L., & Yau, K. (2002). Kinematic analysis of kicking movements in preterm
infants with very low birth weight and full-term infants. Physical Therapy, 82, 148-159.
Jensen, J. L., Schneider, K., Ulrich, B. D., & Zernicke, R. F. (1994). Adaptive dynamics of the
leg movement patterns of human infants: I. The effects of posture on spontaneous
kicking. Journal of Motor Behavior, 26, 303-312.
Kelso, J. A., Scholz, J. P., & Schoner, G. (1986). Nonequilibrium phase transitions in
coordinated biological motion: critical fluctuations. Phys. Lett. A, 134, 8-12.
Larroque, B., Ancel, P.-Y., Marret, S., Marchand, L., André, M., Arnaud, C., . . . Kaminski, M.
(2008). Neurodevelopmental disabilities and special care of 5-year-old children born
before 33 weeks of gestation (the EPIPAGE study): A longitudinal cohort study. The
Lancet, 371(9615), 813-820.
Lee, H., Bhat, A., Scholz, J., & Galloway, J. C. (2008). Toy-oriented changes during early arm
movements: IV: Shoulder-elbow coordination. Infant Behavior and Development, 31(3),
447-469.
Lloyd, M., Burghardt, A., Ulrich, D. A., & Angulo-Barroso, R. (2010). Physical activity and
walking onset in infants with Down syndrome. Adapted Physical Activity Quarterly, 27,
1-16.
Lobo, M., Harbourne, R., Dusing, S. C., & McCoy, S. (2013). Grounding early intervention:
Physical therapy cannot just be about motor skills anymore. Physical Therapy, 93(1), 94-
103.
INFANT LEG COORDINATION 154
McGraw, M. B. (1945). The neuromuscular maturation of the human infant. New York, NY:
Columbia University Press.
McKay, S. M., & Angulo-Barroso, R. (2006). Longitudinal assessment of leg motor activity and
sleep patterns in infants with and without Down syndrome. Infant Behavior and
Development, 29, 153-168.
Meinecke, L., Breitbach-Faller, N., Bartz, C., Damen, R., Rau, G., & Disselhorst-Klug, C.
(2006). Movement analysis in the early detection of newborns at risk for developing
spasticity due to infantile cerebral palsy. Human Movement Science, 25(2), 125-144.
Murray, R. M., Li, Z., & Sastry, S. S. (1994). A mathematical introduction to robotic
manipulation. Boca Raton, Florida: CRC Press.
Newell, K. (Ed.). (1986). Constraints on the development of coordination. Dordrecht: Martinus
Nijhoff Publishers.
Ohr, P., & Fagen, J. (1991). Conditioning and long-term memory in three-month-old infants with
Down syndrome. American Journal on Mental Retardation, 96(2), 151-162.
Oskoui, M., Coutinho, F., Dykeman, J., Jetté, N., & Pringsheim, T. (2013). An update on the
prevalence of cerebral palsy: A systematic review and meta-analysis. Developmental
Medicine and Child Neurology, 1-11.
Ostensjo, S., Carlberg, E., & Vollestad, N. (2004). Motor impairments in young children with
cerebral palsy: Relationship to gross motor function and everyday activities.
Developmental Medicine & Child Neurology, 46, 580-589.
Pearsall-Jones, J. G., Piek, J. P., & Levy, F. (2010). Developmental coordination disorder and
cerebral palsy: Categories or a continuum? Human Movement Science, 29(5), 787-798.
INFANT LEG COORDINATION 155
Piek, J. P. (1996). A quantitative analysis of spontaneous kicking in two-month-old infants.
Human Movement Science, 15, 707-726.
Rademacher, N., Black, D., & Ulrich, B. D. (2008). Early spontaneous leg movements in infants
born with and without myelomeningocele. Pediatric Physical Therapy, 20, 137-145.
Rosenbaum, P., Paneth, N., Leviton, A., Goldstein, M., Bax, M., Damiano, D., . . . Jacobsson, B.
(2007). A report: The definition and classification of cerebral palsy. Developmental
Medicine and Child Neurology, Suppliment, 8-14.
Rovee-Collier, C. K., & Gekoski, M. J. (1979). The economics of infancy: A review of conjugate
reinforcement. In H. W. Reese & L. P. Lipsitt (Eds.), Advances in child development and
behavior (Vol. 13, pp. 195-255). New York, NY: Academic.
Rovee, C. K., & Rovee, D. T. (1969). Conjugate reinforcement of infant exploratory behavior.
Journal of Experimental Child Psychology, 8(1), 33-39.
Sainburg, R., & Kalakanis, D. (2000). Differences in control of limb dynamics during dominant
and nondominant arm reaching. Journal of Neurophysiology, 83, 2661-2675.
Sanger, T., Chen, D., Delgado, M., Gaebler-Spira, D., Hallett, M., & Mink, J. (2006). Definition
and classification of negative motor signs in childhood. Pediatrics, 118, 2159-2167.
Schneider, K., & Zernicke, R. F. (1992). Mass, center of mass and moment of inertia estimates
for infant limb segments. Journal of Biomechanics, 25(2), 145-148.
Schneider, K., Zernicke, R. F., Ulrich, B. D., Jensen, J. L., & Thelen, E. (1990). Understanding
movement control in infants through the analysis of limb intersegmental dynamics.
Journal of Motor Behavior, 22, 493-520.
Scholz, J. P. (1990). Dynamic pattern theory--Some implications for therapeutics. Physical
Therapy, 70, 827-843.
INFANT LEG COORDINATION 156
Soderkvist, I., & Wedin, P. (1993). Determining the movements of the skeleton using well-
configured markers. Journal of Biomechanics, 26(12), 1473-1477.
Sporns, O., & Edelman, G. (1993). Solving Bernstein's problem: A proposal for the development
of coordinated movement by selection. Child Development, 64, 960-981.
Stephens, B., & Vohr, B. (2009). Neurodevelopmental outcome of the premature infant.
Pediatric Clinics of North America, 56, 631-646.
Sun, H., & Jensen, R. (1994). Body segment growth during infancy. Journal of Biomechanics,
27(3), 265-275.
Thelen, E. (1985). Developmental origins of motor coordination: Leg movements in human
infants. Developmental Psychobiology, 18, 1-22.
Thelen, E. (1994). Three-month-old infants can learn task-specific patterns of interlimb
coordination. Psychological Science, 5, 280-285.
Thelen, E., Bradshaw, G., & Ward, J. A. (1981). Spontaneous kicking in month-old infants:
Manifestation of a human central locomotor program. Behavioral and Neural Biology,
32, 45-53.
Thelen, E., & Fisher, D. (1982). Newborn stepping: An explanation for a "disappearing" reflex.
Developmental Psychology, 18, 760-775.
Thelen, E., & Fisher, D. (1983a). From spontaneous to instrumental behavior: Kinematic
analysis of movement changes during very early learning. Child Development, 54, 129-
140.
Thelen, E., & Fisher, D. (1983b). The organization of spontaneous leg movement in newborn
infants. Journal of Motor Behavior, 15, 353-377.
INFANT LEG COORDINATION 157
Thelen, E., Ridley-Johnson, R., & Fisher, D. (1983). Shifting patterns of bilateral coordination
and lateral dominence in the leg movements of young infants. Developmental
Psychobiology, 16, 29-46.
Thelen, E., Skala, K., & Kelso, J. (1987). The dynamic nature of early coordination: Evidence
from bilateral leg movements in young infants. Developmental Psychobiology, 23, 179-
186.
Thelen, E., & Smith, L. B. (1994). A dynamic systems approach to the development of cognition
and action. Cambridge, MA: MIT Press.
Thelen, E., & Ulrich, B. D. (1991). Hidden skills: A dynamical system analysis of treadmill
stepping in the first year Monographs of the Society for Reseach in Child Development,
223.
Tiernan, C., & Angulo-Barroso, R. (2008). Constrained motor-perceptual task in infancy: Effects
of sensory modality. Journal of Motor Behavior, 40, 133-142.
Tseng, Y., Scholz, J. P., & Galloway, J. C. (2009). the organization of intralimb and interlimb
synergies in response to different joint dynamics. Experimental Brain Research, 193,
239-254.
Ulrich, B. D. (2010). Opportunities for early intervention based on theory, basic neuroscience,
and clinical science. Physical Therapy, 90(12), 1868-1880.
Ulrich, B. D., & Ulrich, D. A. (1995). Spontaneous leg movements of infants with Down
syndrome and nondisabled infants. Child Development, 66, 1844-1855.
Ulrich, B. D., Ulrich, D. A., & Angulo-Kinzler, R. (1997). Sensitivity of infants with and without
Down syndrome to intrinsic dynamics. Research Quarterly for Exercise and Sport, 68(1),
10-19.
INFANT LEG COORDINATION 158
Vaal, J., van Soest, A. J., & Hopkins, B. (2000). Spontaneous kicking behavior in infants: age-
related effects of unilateral weighting. Developmental Psychobiology, 36, 111-122.
Vaal, J., van Soest, A. J., Hopkins, B., & Sie, L. T. L. (2002). Spontaneous leg movements in
infants with and without periventricular leukomalacia: effects of unilateral weighting.
Behavioural Brain Research, 129, 83-92.
Vaal, J., van Soest, A. J., Hopkins, B., Sie, L. T. L., & van der Knapp, M. S. (2000).
Development of spontaneous leg movements in infants with and without periventricular
leukomalacia. Experimental Brain Research, 135, 94-105.
van Emmerick, R., & Wagenaar, R. (1996). Effects of walking velocity on relative phase
dynamics in the trunk in human walking. Journal of Biomechanics, 29, 1175-1184.
Volpe, J. J. (2009). Brain injury in premature infants: A complex amalgam of destructive and
developmental disturbances. The Lancet Neurology, 8, 110-124.
Watanabe, H., & Taga, G. (2006). General to specific development of movement patterns and
memory for contingency between actions and events in young infants. Infant Behavior
and Development, 29, 402-422.
Watanabe, H., & Taga, G. (2009). Flexibility in infant actions during arm- and leg-based
learning in a mobile paradigm. Infant Behavior and Development, 32, 79-90.
Williams, J., Lee, K. J., & Anderson, P. J. (2010). Prevalence of motor-skill impairment in
preterm children who do not develop cerebral palsy: A systematic review. Developmental
Medicine & Child Neurology, 52(3), 232-237.
Yang, J. F., Lam, T., Pang, M. Y. C., Lamont, E., Musselman, K., & Seinen, E. (2004). Infant
stepping: A window to the behaviour of the human pattern generator for walking.
Canadian Journal of Physiology & Pharmacology, 662-674.

Abstract (if available)

Abstract Infants born preterm with very low birth weight are at increased risk for developing spastic cerebral palsy, which is characterized by walking limitations due to a reduced ability to move the joints of the leg in an out-of-phase pattern; e.g., flexing the hip while extending the knee. In typical development, the spontaneous kicks of newborn infants are dominated by an in-phase coordination pattern of synchronous flexion or extension of the hip, knee, and ankle joints. Within the first months of life, infants demonstrate a greater variety of intralimb joint coordination patterns which is thought to support the progress to more functional coordination patterns of the lower extremities necessary for independent mobility. PT infants also demonstrate this developmental change from in-phase to a greater variety of joint coordination patterns, however, some PT infants exhibit prolonged in-phase joint coordination which places them at increased risk of developing cerebral palsy. The purpose of this dissertation is to gain a better understanding of the contribution of torque changes to the early changes in leg joint coordination of typically-developing full-term infants, clarify the differences between leg joint coordinations and torques of preterm and full-term infants, and determine whether preterm infants have the potential for a greater amount of out-of-phase leg joint coordination when interacting with an infant-activated mobile that reinforces out-of-phase leg joint coordination. ❧ Three studies were conducted to answer these questions. The first study investigated the contribution of torque changes to the early changes in leg joint coordination of typically-developing full-term infants. We analyzed kicking actions within 10 full-term infants and between 6 and 15-weeks of age using a three-dimensional kinematics and kinetics approach. We found that although 73% of joint angle pairs demonstrated a change from an in-phase intralimb coordination pattern at 6-weeks to an increased variety of intralimb joint coordination patterns at 15-weeks, there was not an obvious developmental change in net joint torques or intersegmental dynamics. Further analysis supported that a greater variety of hip-knee joint coordinations from 6 to 15-weeks of age was associated with a decreased influence of knee muscle torques which allowed passive knee gravitational and motion dependent torques to have a greater influence on the coordination of the kick, contributing to a greater variety of hip-knee joint coordinations. ❧ The second study investigated the contribution of torque changes to the early changes in leg joint coordination of preterm infants. From 6 to 15 weeks, preterm infants demonstrated less in-phase coordination in 30% of joint angle pairs, versus 73% of joint angle pairs in full-term infants. At 6 weeks preterm infants as compared to full-term infants demonstrated less in-phase coordination in 4 of 15 joint angle pairs, but this difference resolved at 15 weeks. Similar to full-term infants, there was not an obvious developmental change in net joint torques or intersegmental dynamics from 6 to 15 weeks. Although PT infants exhibited a greater variety of hip-knee coordinations from 6 to 15 weeks of age, unlike full-term infants, no differences were noted in the intersegmental torques. However, the PT infants demonstrated a smaller change in hip-knee coordination from 6 to 15 weeks which may have been insufficient to document the relation between intersegmental dynamics and joint coordination. ❧ The final study investigated whether preterm infants could generate a greater amount of out-of-phase leg joint coordination when participating in an innovative learning task which utilized an infant-activated mobile. Fourteen full-term infants and 6 preterm infants participated at 3 - 4 months corrected age. Each infant participated in 2 sessions of mobile reinforcement on consecutive days. During each session, the infant was positioned supine under an overhead infant mobile. Day 1 consisted of a 2-minute non-reinforcement condition (spontaneous kicking) followed by a 6-minute reinforcement condition (the infant mobile rotated and played music when the infant moved either foot vertically across a virtual threshold). Day 2 consisted of a 2-minute non-reinforcement condition, 6-minute reinforcement condition, and 2-minute non-reinforcement condition. The full-term group, but not the preterm group, increased the percentage of mobile activation to meet performance criteria the 2nd day. Neither group met learning criteria. However, both the full-term and preterm groups included infants that learned the contingency between their leg action and mobile activation. Infants were separated into infants that learned the contingency and infants that did not learn the contingency. Infants who learned the contingency demonstrated a greater amount of out-of-phase hip-knee joint coordination during the reinforcement condition on the second day as compared to spontaneous kicking during the initial non-reinforcement condition on the first day. This coordination change was not demonstrated by the group of infants that did not learn the contingency. These results indicate that some full-term and preterm infants can demonstrate a greater amount of out-of-phase hip-knee joint coordination when participating in a task in which their leg actions are reinforced with mobile activation. ❧ The dissertation concludes with a discussion of each research study highlighting novel findings, implications for clinical practice, and opportunities for future research. The results from this dissertation will be an essential step toward developing more relevant and successful screening, intervention, and prevention programs for preterm infants at risk for impairments in motor coordination.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Understanding how infants born full-term and preterm learn, move, and explore

PDF

Learning acceleration control of leg movements during infancy

PDF

Learning self-agency during a contingency paradigm in infants at commnity and elevated risk of autism

PDF

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

PDF

The biomechanical changes associated with the transition from shod to barefoot running

PDF

Locomotor skill learning in virtual reality in healthy adults and people with Parkinson disease

PDF

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

PDF

Corticomotor excitability of gluteus maximus: influence on hip extensor strength and hip mechanics

PDF

Development of a movement performance assessment to predict ACL re-injury

PDF

The representation, learning, and control of dexterous motor skills in humans and humanoid robots

PDF

Factors influencing patellofemoral joint stress following anterior cruciate ligament reconstruction: implications for early onset osteoarthritis

PDF

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

Asset Metadata

Creator Sargent, Barbara A. (author)

Core Title Infant exploratory learning: influence on leg coordination

School School of Dentistry

Degree Doctor of Philosophy

Degree Program Biokinesiology

Degree Conferral Date 2013-05

Publication Date 04/30/2013

Defense Date 03/14/2013

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

Tag Infant,kicking,kinematics,kinetics,Learning,OAI-PMH Harvest

Format theses (aat)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Fetters, Linda (committee chair), Azen, Stanley P. (committee member), Bradley, Nina S. (committee member), Scholz, John P. (committee member), Schweighofer, Nicolas (committee member)

Creator Email bsargent@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-oUC11293734

Unique identifier UC11293734

Identifier etd-SargentBar-1640.pdf (filename)

Legacy Identifier etd-SargentBar-1640

Dmrecord 249911

Document Type Dissertation

Format theses (aat)

Rights Sargent, Barbara A.

Internet Media Type application/pdf

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 author, as the original true and official version of the work, but does not grant the reader permission to use the work if the desired use is covered by copyright. It is the author, as rights holder, who must provide use permission if such use is covered by copyright.

Repository Name University of Southern California Digital Library

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

Repository Email cisadmin@lib.usc.edu