Close
Home
Collections
Login
USC Login
Register
0
Selected
Invert selection
Deselect all
Deselect all
Click here to refresh results
Click here to refresh results
USC
/
Digital Library
/
University of Southern California Dissertations and Theses
/
Understanding how infants born full-term and preterm learn, move, and explore
(USC Thesis Other)
Understanding how infants born full-term and preterm learn, move, and explore
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
UNDERSTANDING HOW INFANTS BORN FULL-TERM AND PRETERM
LEARN, MOVE, AND EXPLORE
by
Jeongah Kim
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 2020
Copyright 2020 Jeongah Kim
ii
Dedication
To my family, especially my grandparents who see me from heaven
To all the babies and families who have deeply inspired this research
To Drs. Barbara Sargent and Linda Fetters who have encouraged me throughout this journey
iii
Acknowledgements
I am very grateful that I have the opportunity to highlight the following incredible people
who have supported me in achieving this dissertation work.
First, I would like to express my deepest gratitude to my dissertation committee. I will
never forget how Dr. Barbara Sargent and Dr. Linda Fetters, my co-advisors, welcomed me so
warmly when I first came to Los Angeles during my campus visit five years ago. They invited
me, a naïve international student, to amazing dinners together with the developmental group
people and made me feel to be a part of the team at USC. During the past five years, they have
always been responsive, responsible, supportive, attentive, and understanding and made me their
priority. It touched me greatly because I knew they have numerous other duties. Dr. Sargent and
Dr. Fetters have made me love my research and career as a scientist, which requires creativeness
and detail-oriented work at the same time. I will miss our weekly meetings and the moments
when three of us discussed my raw data. I cannot express how grateful I am to both, and not to
mention, Dr. Sargent and Dr. Fetters are the best leaders, scientists, and mentors in my life. I
respect them both so much and hope to someday be as excellent as they are.
To Dr. Nina Bradley, I am grateful for her critical advice and feedback. My conversations
with her have changed for me what it means to become an independent scientist, and she has
been an important model for me as I start my own journey as a researcher. In particular, she has
strongly advised me to become an active discussant. That was very challenging to me, but she
helped me overcome my personal and cultural barriers, practice speaking up and finally discover
a part of myself who enjoys discussions. I am also thankful for the readings I had with her on
neuropathology. That was my first time to dive deep into studying neural and physiological
iv
development. She taught me how to critically look at literature and obtain information out of it. I
hope I further develop my knowledge in this field and apply it to my future infant work to make
the basis stronger.
To Dr. Sandy Eckel, at USC Division of Biostatistics, I feel incredibly fortunate to have
worked with her. I took her course – PM511a Data Analysis, which helped me learn regression
models and use SAS for statistical analysis. I am deeply thankful for her efforts to fit me in her
hectic schedule. Every meeting with her was short but was most productive. Whenever I was
distressed about stats, she was always there with excellent solutions.
To Dr. Eileen Fowler, at UCLA Department of Orthopaedic Surgery, I would like to
thank for her pioneering work in defining and quantifying selective motor control in children
with cerebral palsy. Her research and methodology became a strong basis for my research. Her
research in children has given me big picture and foundational knowledge in regards to studying
infants. I appreciate her interest in my research and her thoughtful inputs.
I am also grateful to Sungwoo Park: our conversations about my final research project
deeply enriched it and resulted in his coding an essential Matlab program that made my work
possible, and to Hyeshin Park, Jeremy Welch, Yukikazu Hidaka, Younggeun Choi, Yongwan
Lim, Kate Havens, Nicolas Schweighofer, and Masayoshi Kubo who provided technical
resources for the development of the infant kick-activated mobile system, and to Anvitha
Shivakumar, Jessica Nguyen, Maggie Ridenhour, Manjima Sarkar, Sophia Zhou, Nicole Crisan,
Alisa Kokanoutranon, Aubrey Baker, Joshua Limlingan, Emily Puzo, Jamie Proffitt, Vivian Pae,
Stephanie Horwitz, Carlos Marroquin, Brandon Mamou, Andres Caballero, and Christine Lee
who assisted with data collection and analysis.
v
Special thanks to Dr. Lisa Guerra Vargas and staff members at the High-Risk Newborn
Clinic at Los Angeles County + USC Medical Center for allowing me to recruit participants and
welcoming me every Friday. I am also grateful for the parents and infants who participated in my
study. I owe my deepest gratitude to them for their time, support, and dedication to science.
It would have been impossible to accomplish this work without support from faculty,
staff, and colleagues in the USC Division of Biokinesiology and Physical Therapy. I want to
thank Dr. James Gordon for all his academic, emotional, and financial support. Very special
thanks to the faculty: Dr. Beth Smith, Dr. Carolee Winstein, Dr. Lucinda Baker, Dr. Rob Landel,
Dr. Beth Fisher, and Dr. Ryota Nishiyori, to the staff: Oshawa Smith, Troy Lord, Matthew
Sandusky, Chad Louie, Raj Singh, Janet Stevenson, Trihn Nguyen, Lydia Vazquez, and Tasha
Hsu, to the students, my dear colleagues: Steffi Shih, Wayne Deng, Marcelo Rosales, Ivan
Trujillo-Priego, Kyuwan Lee, Rini Varghese, Sara Almansouri, Bokkyu Kim, Helen Bacon,
Akira Nagamori, Andrew Hooyman, Alex Garbin, Aram Kim, Soo Yeon Sun, Victor Barradas,
Sujin Kim, and my sincere thanks to Kyungmi Park.
I would like to give my thanks to my family and people from the Los Angeles Power of
Praise Church who always trust me without reservation. To my mom and dad, I cannot express in
words what they have done for me. With their undying love, I have been encouraged, supported,
and protected. I know they have thought of me, prayed for me, and always been on my side. To
my lovely younger brother, Whanjoo, a precious memory was when he visited LA for one and a
half months during my second year. Our time together was joyful and reminded me that I am
never alone. I love and bless you all so much.
vi
To my husband, Kyungkeun, I am very fortunate to be married to him. He and I met at
USC at the perfect timing, and we have been able to share our campus lives, study passions, and
future dreams. Although a PhD life is known as a period of blood, sweat, and tears, the time I
spent here will remain as a beautiful memory because of him.
I am appreciative of the financial support from the USC Division of Biokinesiology and
Physical Therapy. I also received additional funding from the Southern California Clinical and
Translational Science Institute (SC CTSI) Voucher Program, the American Physical Therapy
Association (APTA) Academy of Pediatric Physical Therapy Mentored Grant to Jeongah Kim
and by the National Institutes of Health (NIH) Eunice Kennedy Shriver National Institute of
Child Health and Human Development (NICHD) under award number K12-HD055929 (PI:
Ottenbacher) to Barbara Sargent. The content is solely the responsibility of the authors and does
not necessarily represent the official views of the NIH.
vii
Table of Contents
Dedication ∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙ ii
Acknowledgements ∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙ iii
List of Tables ∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙ ix
List of Figures ∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙ x
Abstract ∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙ xiii
Chapter I: Introduction to the Dissertation ∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙ 1
Chapter II: Infants Born Full-term and Preterm Learn a Scaffolded Mobile Task ∙∙∙∙∙∙∙∙∙∙∙∙∙∙ 4
Abstract ∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙ 4
Introduction ∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙ 4
Methods ∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙ 7
Participants ∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙ 7
Experimental protocol ∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙ 7
Data reduction ∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙ 11
Statistical analysis ∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙ 13
Results ∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙ 14
Participants ∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙ 14
Learning ∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙ 17
Kick height ∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙ 19
Arousal ∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙ 22
Visual attention ∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙ 23
Discussion ∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙ 24
Chapter III: Infants Born Preterm and Infants Born Full-term Generate More Selective Leg
Joint Movement during the Scaffolded Mobile Task ∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙ 28
Abstract ∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙ 28
Introduction ∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙ 29
Methods ∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙ 31
Participants ∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙ 31
Experimental protocol ∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙ 32
Data reduction ∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙ 34
Statistical analysis ∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙ 35
Results ∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙ 36
Participants ∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙ 36
viii
Selective hip-knee movement∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙ 37
Discussion ∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙ 40
Clinical implications and future directions∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙ 42
Limitations ∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙ 43
Conclusions ∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙ 43
Chapter IV: Quantifying Exploratory Learning of Infants Born Full-term ∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙ 44
Abstract ∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙ 44
Introduction ∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙ 44
Methods ∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙ 48
Participants ∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙ 48
Experimental protocol ∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙ 48
Data reduction ∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙ 50
Statistical analysis ∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙ 52
Results ∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙ 53
Participants ∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙ 53
Final statistical models ∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙ 53
Variance ∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙ 54
Exploration volume ∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙ 57
Exploration region ∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙ 58
Discussion ∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙ 60
Conclusions ∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙ 62
Chapter V: Summary and Conclusions ∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙ 63
References ∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙ 66
ix
List of Tables
Table 2.1 Demographics of Participants 15
Table 2.2 Demographics of Individual Infants Born Preterm 16
Table 2.3 Mobile Active Time Ratio over Conditions, Mean (SE) 17
Table 2.4 Threshold Height (cm), Mean (SD) 21
Table 2.5 Number of Kicks Analyzed and Kick Height (cm) over Threshold
Conditions, Mean (SE)
21
Table 2.6 Alert Time (Percentage) over Conditions, Mean (SE) 22
Table 2.7 Mobile Looking Time (Percentage) over Conditions, Mean (SE) 23
Table 3.1
Demographics of Participants 36
Table 3.2 Hip-Knee Z-transformed Correlation Coefficient (ZCC) over Condition 37
Table 3.3 Hip-Knee Z-transformed Correlation Coefficient (ZCC) over Threshold
Interval
37
Table 3.4
Hip-Knee Relative Phase (RP) at Joint Reversal over Condition 40
Table 3.5 Hip-Knee Relative Phase (RP) at Joint Reversal over Threshold Interval 40
x
List of Figures
Figure 2.1 Infant set-up. One infrared light-emitting diode (circled) on each foot
activated the mobile. The dotted lines depict the three heights of the
virtual threshold.
6
Figure 2.2
Scaffolded mobile task timeline. On Day 1, the scaffolded mobile task
included a 2-min baseline condition followed by an 8-min mobile
condition, which used the lowest threshold height (mean baseline
kicking + 1 standard deviation; sd). On Day 2 and 3, the task included a
10-min mobile condition during which the threshold height was
systematically increased from the lowest (+1sd) to the middle (+1.5sd)
to the highest (+2sd).
9
Figure 2.3 Mobile active time (MAT) ratios in the infants born full-term (A) and
the infants born preterm (B) over 2-min intervals on Day 1, Day 2, and
Last Day; B = baseline condition; +1 = mean+1sd threshold condition;
+1.5 = mean+1.5sd threshold condition; +2 = mean+2sd threshold
condition; Error bars denote standard errors; asterisks denote adjusted p
< .05; The MAT ratios are plotted for each 2-min interval to show the
change in the time series, but for statistical analysis, the mean values of
the five 2-min intervals on Day 2 and Last Day, respectively, were
compared to the Day 1 baseline condition.
17
Figure 2.4 Kick height in the infants born full-term (A) and the infants born
preterm (B) over threshold conditions on Day 1, Day 2, and Last Day; B
= baseline condition; +1 = mean+1sd threshold condition; +1.5 =
mean+1.5sd threshold condition; +2 = mean+2sd threshold condition;
Error bars denote standard errors; asterisks denote p < .05; The kick
heights are plotted for each threshold condition to show the change in
the time series, but for statistical analysis, the kick heights during the
19
xi
+2sd threshold conditions on Day 2 and Last Day, respectively, were
compared to the Day 1 baseline condition.
Figure 4.1 Infant set-up. One marker (circled) on each foot activated the mobile.
The dotted lines depict the three heights (low, middle, high) of the
virtual threshold.
47
Figure 4.2 One infant’s right and left foot marker data captured for one minute.
(A) foot marker position data with red indicating the right foot and blue
indicating the left foot; (B) the position data wrapped by a convex hull;
(C) the convex hull. The gray plane depicts a threshold location (the
individually determined, low threshold height for this infant = 15.5 cm);
(D) the exploration region area. The darker area depicts region around
the threshold (z = 15.5 ± 1) and the less dark area depicts region above
the threshold (z ≥ 16.5).
51
Figure 4.3 Cross-correlation analysis result between the z- and x-variances. Shows
low correlation between the z- and x-variances.
54
Figure 4.4 The slope and 95% confidence interval of the z-variance in millimeters
(mm) across days. Individual 2-min estimates and standard errors are
indicated (B: baseline, L: low threshold, M: middle threshold, H: high
threshold).
55
Figure 4.5 Individual line graphs of the z-variance (mm) in 15 infants over 1-min
intervals across days.
55
Figure 4.6 The slope and 95% confidence interval of the x-variance in millimeters
(mm) across days. Individual 2-min estimates and standard errors are
indicated.
56
Figure 4.7 Individual line graphs of the x-variance (mm) of 15 infants over 1-min
intervals across days.
56
xii
Figure 4.8 The slope and 95% confidence interval of the exploration volume in
cubic centimeters (cm
3
) across days. Individual 1-min estimates and
standard errors are indicated (B: baseline, L: low threshold, M: middle
threshold, H: high threshold)
57
Figure 4.9 Individual line graphs of the exploration volume (cm
3
) in 15 infants
over 1-min intervals across days.
58
Figure 4.10 The slope and 95% confidence interval of the exploration region around
and above the threshold in percentage (%) across days. Individual 1-min
estimates and standard errors are indicated.
59
Figure 4.11 Individual line graphs of the exploration region around and above the
threshold (%) in 15 infants over 1-min intervals across days.
59
Figure 4.12 The exploration volume from one infant’s data (left: a pyramidal shape
during a 1-min interval on Day 1, right: an inverse pyramidal shape
during a 1-min interval on the learning day).
61
xiii
Abstract
This dissertation aims to gain a better understanding of how infants born full-term (FT)
and very preterm (PT) learn, move, and explore during a scaffolded mobile task that
systematically reinforces selective leg movements.
Children with spastic cerebral palsy have impaired selective leg joint movement which
may contribute to lifelong walking limitations. Selective joint movement is defined as the ability
to isolate the movement of one joint from the movement of the other joints within the limb.
Infants born PT are at increased risk of developing cerebral palsy. Infants born PT with brain
lesions, at highest risk for cerebral palsy, show impaired selective leg movements during
spontaneous kicking as early as one month of age. However, there is no evidence if infants born
PT have the capacity to produce more selective leg movements when a task reinforces them to
move their legs selectively.
To encourage selective hip-knee movements of infants born PT, prior to 32 weeks of
gestation, we scaffolded an infant kick-activated mobile task. Data were analyzed on 18 infants
born FT and 18 infants born PT who participated in the scaffolded mobile task for 2 to 3
consecutive days. This dissertation includes three studies, titled as: 1) Infants born FT and PT
learn a scaffolded mobile task, 2) Infants born PT and infants born FT generate more selective
leg joint movement during the scaffolded mobile task, and 3) Quantifying exploratory learning of
infants born FT.
These studies found that 1) both the FT and PT groups demonstrated learning of the
scaffolded mobile task, 2) the infants born FT and PT, who learned the scaffolded mobile task,
demonstrated more selective hip-knee joint movements while interacting with the mobile,
compared to their spontaneous kicking, and 3) the infants born FT, who learned the scaffolded
xiv
mobile task, increased their exploration of the task-specific space over the course of the task.
This dissertation research lays a scientific foundation for clinical practice. The studies of this
dissertation provide an example of how to structure and scaffold a task to support more selective
hip-knee movements of infants born PT at increased risk of developing cerebral palsy.
1
Chapter I: Introduction to the Dissertation
Children with spastic cerebral palsy have impaired selective leg joint movement which
may contribute to lifelong walking limitations (Farmer, 2003; Farmer, Pearce, & Stewart, 2008;
Fowler, Staudt, & Greenberg, 2010). Selective joint movement is defined as the ability to isolate
the movement of one joint from the movement of the other joints within the limb (e.g., hip flexes
while knee extends) (Fowler & Goldberg, 2009). Infants born very preterm (PT), prior to 32
weeks of gestation, are at increased risk of developing cerebral palsy; 6-15% develop cerebral
palsy (Himpens, Van den Broeck, Oostra, Calders, & Vanhaesebrouck, 2008; Pascal et al.,
2018). Infants born PT with brain lesions are at highest risk for developing cerebral palsy
(Gopagondanahalli et al., 2016; Hagberg, Hagberg, Beckung, & Uvebrant, 2001) and show
impaired selective leg movement during spontaneous kicking as early as one month of age
(Fetters, Chen, Jonsdottir, & Tronick, 2004). There is no evidence if infants born PT have the
capacity to produce more selective leg movements when a task reinforces them to move
selectively.
This dissertation aims to gain a better understanding of how infants born full-term (FT)
and PT learn, move, and explore during a task that systematically reinforces selective leg
movements. To encourage selective movements, we used an infant kick-activated mobile.
Previous research supports that infants born FT, who learned that their leg movements activated
the mobile, generated more selective leg movements (Sargent, Schweighofer, Kubo, & Fetters,
2014), but infants born PT who learned the task did not generate more selective leg movement
(Sargent, Kubo, & Fetters, 2018).
Here, we scaffolded a kick-activated mobile task to encourage infants born PT to
generate more selective leg movements. In the scaffolded mobile task, an infant lies supine under
2
a mobile, which rotates and plays music when the infant lifts the feet vertically above an
individually determined threshold height. On the first day, a low threshold height is used to
support learning of the association between the infant’s leg movement and the mobile activation.
On subsequent days, the threshold height is increased from low to middle to high in order to
encourage infants to kick high, thereby generating more selective hip-knee movements of hip
flexion with knee extension.
This dissertation includes three studies: Chapter II: Infants born full-term and preterm
learn a scaffolded mobile task; Chapter III: Infants born preterm and infants born full-term
generate more selective leg joint movement during the scaffolded mobile task; and Chapter IV:
Quantifying exploratory learning of infants born full-term.
Chapter II: Infants Born Full-term and Preterm Learn a Scaffolded Mobile Task.
This study has two aims: to assess (1) if each group of infants (FT, PT) learned the scaffolded
mobile task and (2) if each group changed the kick heights. We hypothesized that (1) both groups
of infants would learn the association between their leg actions and mobile activation, but the PT
group would require an extra day of practice, compared to the FT group, and (2) both groups of
infants would significantly increase the kick heights by the last day of participation, compared to
spontaneous kicking actions during the baseline period on the first day.
Chapter III: Infants Born Preterm and Infants Born Full-term Generate More
Selective Leg Joint Movement during the Scaffolded Mobile Task. This study assesses if
infants born FT and infants born PT who learned the scaffolded mobile task had more selective
hip-knee joint movement when interacting with the mobile, compared to their baseline
spontaneous kicking. We tested three hypotheses: (1) on the first day, infants born FT and infants
born PT would not demonstrate more selective hip-knee movement when interacting with the
3
mobile, compared to their baseline spontaneous kicking, (2) on the day that infants learned the
task (learning day), infants born FT would exhibit more selective hip-knee movement when
interacting with the mobile during the intervals using the low, middle, and high threshold
heights, compared to their baseline spontaneous kicking, and (3) on the learning day, infants
born PT would not exhibit more selective hip-knee movement during the interval using the low
threshold height, but would exhibit more selective hip-knee movement during the intervals using
the middle and high threshold heights, compared to their baseline spontaneous kicking.
Chapter IV: Quantifying Exploratory Learning of Infants Born Full-term. This
study quantifies how infants explore task space with their feet when learning to activate a mobile
during the scaffolded mobile task. We hypothesized that (1) on the first day, as infants explored
the low threshold height, the task-specific exploration variables (z-variance, volume, region)
would increase, (2) on the learning day, when the low, middle, and high threshold heights were
used, task-specific variables (z-variance, volume, region) would significantly increase, compared
to the first day when only the low threshold height was used, and (3) the non-task-specific
exploration variable (x-variance) would not change on either the first day or the learning day.
Chapter V: Summary and Conclusion. This chapter summarizes the findings, and
describes the limitations of this work and recommendations for future research.
4
Chapter II: Infants Born Full-term and Preterm Learn a Scaffolded Mobile Task
Abstract
Prior research supports that infants born preterm (PT), compared to full-term (FT), have
early differences in rate of learning and motor control that may hinder their ability to learn
challenging tasks. This study used an infant kick-activated mobile task with scaffolding to
reinforce progressively higher kicks. We found that the FT group (n=18) at 4 months learned the
association between their leg movements and mobile activation on the second day, but the PT
group (n=18) at 4 months corrected age learned the association on the third day. Both groups of
infants learned the challenging task of kicking high to activate the mobile. These findings
suggest that scaffolding may support infants born PT to independently learn challenging tasks.
Introduction
Infants born preterm (PT) are at increased risk for learning and motor disabilities (Lobo &
Galloway, 2013a; Van Hus, Potharst, Jeukens-Visser, Kok, & Van Wassenaer-Leemhuis, 2014;
Williams, Lee, & Anderson, 2009). As early as 3 to 4 months of age, infants born PT, compared
to full-term (FT), have differences in learning tasks (Gekoski, Fagen, & Pearlman, 1984; Haley,
Grunau, Oberlander, & Weinberg, 2008; Lobo & Galloway, 2013a; Sargent, Kubo, & Fetters,
2018). This difference in learning may be due to impairments commonly reported for infants
born PT, such as reduced rate of learning (Gekoski et al., 1984), decreased attention span (Rose,
Feldman, Jankowski, & Van Rossem, 2005), difficulty regulating arousal (Haley et al., 2008),
and decreased strength and motor control (Heathcock, Bhat, Lobo, & Galloway, 2005; Sargent,
Reimann, Kubo, & Fetters, 2017). When infants born PT first begin to have difficulty learning
tasks in early infancy, scaffolding the task may assist them to independently learn challenging
5
tasks and develop skills in a more age-appropriate manner. We define “scaffolding” as
modifying a task so that the infant is first reinforced for an easily attainable goal, then reinforced
for progressively more challenging goals, until the infant is finally reinforced for the most
challenging, but still achievable task goal.
A mobile, that is activated by leg movement, has been used to investigate learning and
memory in infants born FT and PT (Gekoski et al., 1984; Rovee-Collier, 1997; Rovee & Rovee,
1969). In this paradigm, 3 to 4-month-old infants are supine under an infant mobile and their leg
movements are reinforced with sound and movement of the mobile. Both infants born FT and
infants born PT associate their leg movement and mobile reinforcement. This association is
demonstrated by increased kicking frequency (Gekoski et al., 1984) or duration that the infant
moves the mobile (Sargent et al., 2018; Sargent et al., 2014). However, infants born PT have
difficulty learning the more challenging task of kicking only the reinforced leg. When only one
leg was reinforced by the mobile, infants born PT continue to kick both legs, whereas infants
born FT learned to kick only the reinforced leg (Heathcock et al., 2005).
Scaffolding may assist infants born PT to independently learn challenging tasks. In the
present study, as a challenging task, we used a mobile paradigm that requires infants to lift their
feet high to activate the mobile, thereby reinforcing more anti-gravity strength and control. We
believe that this task is challenging for infants born PT because they have decreased strength and
motor control compared to infants born FT (Ahmad et al., 2010; Dassios, Kaltsogianni, Krokidis,
Hickey, & Greenough, 2018; Dionisio, Santos, & Tudella, 2017; Sargent et al., 2017; Sargent et
al., 2018). We scaffolded the task for infants born PT by reinforcing three different heights of
kick to activate the mobile (Figure 2.1). We used the lowest height of kick on the first day to
support learning of the association between the infant’s leg movement and mobile activation. On
6
the second day, the task began with the lowest height that the infants had practiced on the first
day, and then the heights were systematically increased to the middle and highest height to
motivate the infants to kick higher. We asked participants who did not learn the task on the
second day to participate on a third day.
This study had two objectives. The first was to assess if each group of infants (FT, PT)
learned the scaffolded mobile task. We tested the hypothesis that both groups of infants would
learn the association between their leg actions and mobile activation, but the PT group, compared
to the FT group, would require an extra day of practice. The second objective was to assess if
each group of infants changed their kick heights. We tested the hypothesis that both groups of
infants would significantly increase their kick heights by their last day of participation, compared
to their spontaneous kicking actions during the baseline period on the first day.
Figure 2.1. Infant set-up. One infrared light-emitting diode (circled) on each foot activated the
mobile. The dotted lines depict the three heights of the virtual threshold.
7
Methods
Participants
Infants born FT and PT were recruited from the University of Southern California (USC)
and the Los Angeles County + USC Medical Center from July 2016 to August 2019. All
participants were screened for eligibility via telephone or email interview prior to participation.
Infants born FT were included if they were born at >37 weeks gestation, were developing
typically and without medical conditions per parent report. Infants born PT were included if they
were born at <32 weeks gestation. Infants were excluded if parents reported that the infant was
ill or had congenital malformations, chromosomal abnormalities, orthopedic impairments, or
visual or hearing deficits.
A sample size calculation was completed using G*Power 3.1 (Dusseldorf, Germany)
(Faul, Erdfelder, Lang, & Buchner, 2007) based on the learning outcome in previous studies
(Sargent et al., 2018; Sargent et al., 2014). An effect size f of 1.23 was obtained using an F-test,
analysis of variance for repeated measures, within-between interactions. With an alpha of .05 and
power of .95, a necessary sample size was estimated at 6 per group. We increased the sample
size to 18 per group to raise statistical power to detect potential within- and between-group
differences of our new outcome variable, kick height, which has not been measured in previous
mobile paradigms (Sargent et al., 2018; Sargent et al., 2014).
Experimental Protocol
The USC Health Sciences Institutional Review Board approved the study (#HS-17-
00119). Data collection took place in the Development of Infant Motor Performance Laboratory
at USC. Infants born FT and PT participated between the ages of 4 months 0 day to 4 months 15
days. Infants born PT participated based on their corrected ages, determined by subtracting the
8
number of days preterm from their chronological ages. Procedures were explained to the parents
and written informed consent was obtained before data collection.
Experimental setup. Each day, infants were undressed and placed supine on a testing
table with their feet under a mobile (Figure 2.1). Infants were secured to the table using a Velcro
band across the chest and their heads were placed on a horseshoe-shaped support pillow to
maintain a midline position. Custom, rigid arrays with imbedded 7-mm infrared light-emitting
diodes (IREDs) were attached to the sternum using double-sided adhesive flexible collar tape; to
the pelvis, thighs, and shanks using Velcro straps; and to each foot using a sock. Two Optotrak
Certus motion capture system sensor banks (Northern Digital Inc., Waterloo, ON, Canada) were
positioned on each side of the testing table to capture the position data of the IREDs at 100 Hz. A
video camera (Canon Inc., HD VIXIA HF R700 Camcorder) was positioned overhead to capture
the infants’ facial expressions and eye gaze. The parent sat on a chair next to the infant, but out
of the infant’s sight. The parent was instructed to stay quiet and not touch the infant. However, if
the infant became fussy or cried, the parent was instructed to say, “I am here” or “you are okay”
until the infant was soothed.
Scaffolded mobile task. Figure 2.2 is the timeline of the task. Infants participated in the
10-min scaffolded mobile task for two to three consecutive days depending on rate of learning.
On Day 1, the scaffolded mobile task included a 2-min baseline condition followed by an 8-min
mobile condition. During the 2-min baseline condition, each infant kicked spontaneously but the
mobile did not move. During baseline, the heights of the center IRED on the marker array on
each foot were collected and used to compute each participant’s individual kick height for the
virtual threshold.
9
Figure 2.2. Scaffolded mobile task timeline. On Day 1, the scaffolded mobile task included a 2-
min baseline condition followed by an 8-min mobile condition, which used the lowest threshold
height (mean baseline kicking + 1 standard deviation; sd). On Day 2 and 3, the task included a
10-min mobile condition during which the threshold height was systematically increased from
the lowest (+1sd) the middle (+1.5sd) to the highest (+2sd).
During the 8-min mobile condition, the musical mobile rotated only if the infant lifted
either foot IRED vertically to cross the virtual threshold, which was computed as the average of
the mean height of right foot IRED plus one standard deviation (sd) and the mean height of left
foot IRED plus 1 sd. The equation is as follows:
( 𝑚𝑚 𝑚𝑚𝑚𝑚𝑚𝑚 ℎ𝑚𝑚 𝑒𝑒𝑒𝑒 ℎ𝑡𝑡 𝑜𝑜𝑜𝑜 𝐼𝐼𝐼𝐼 𝐼𝐼𝐼𝐼 𝑜𝑜 𝑚𝑚 𝑟𝑟 𝑒𝑒𝑒𝑒 ℎ𝑡𝑡 𝑜𝑜𝑜𝑜𝑜𝑜𝑡𝑡 + 1𝑠𝑠𝑠𝑠 ) + (𝑚𝑚 𝑚𝑚𝑚𝑚𝑚𝑚 ℎ𝑚𝑚 𝑒𝑒𝑒𝑒 ℎ𝑡𝑡 𝑜𝑜𝑜𝑜 𝐼𝐼𝐼𝐼 𝐼𝐼𝐼𝐼 𝑜𝑜 𝑚𝑚 𝑙𝑙 𝑚𝑚𝑜𝑜𝑡𝑡 𝑜𝑜𝑜𝑜𝑜𝑜𝑡𝑡 + 1 𝑠𝑠 𝑠𝑠 )
2
If the computed threshold was less than 5 cm, the threshold was set at 5 cm to prevent the mobile
from activating without sufficient leg movement. The minimum 5 cm threshold was used for two
infants born FT and one infant born PT. If the computed threshold was greater than 20 cm, the
threshold was set at 20 cm to make it possible for infants to cross the threshold, considering that
the mean of the highest kick height during infants’ baseline spontaneous kicking was 22.5 cm.
The maximum 20 cm threshold was used for one infant born FT and one infant born PT. When
10
either foot crossed the threshold, the musical mobile rotated for the duration that the IRED was
above the threshold up to a maximum of 3 sec. After 3 sec, mobile reactivation required infants
to lower the foot below the threshold and again cross the threshold. This 3-sec rule was used to
encourage infants to move their legs versus holding their feet above the threshold.
On Day 2, the task was offered for 10-min total while the threshold height was
systematically increased. For the first 2-min, the threshold height was the mean height of the foot
IREDs during baseline from Day 1 spontaneous kicking plus 1sd (the lowest dotted line in
Figure 2.1), identical to the height from Day 1. For the following 4-min, the threshold height was
increased to the mean plus 1.5sd (Figure 2.1, the middle line), except for 2 infants born FT and
one infant born PT who used the 20 cm maximum threshold height. For the last 4-min, the height
of the threshold was the mean plus 2sd (Figure 2.1, the highest line), except for 4 infants born FT
and 3 born PT who used the 20 cm maximum threshold height.
All infants participated for two consecutive days. If an infant did not learn the task on
Day 2 based on our individual learning criterion (described in Data Reduction below), the
parents were asked to participate on the following day, Day 3, to give the infants more practice
time. The scaffolded mobile task on Day 3 was identical to Day 2.
Anthropometric and developmental measures. On Day 1, the infant’s medical history
and current health status were obtained from the parents. On Day 2, after the scaffolded mobile
task, participants were weighed on a digital electronic scale (Health-o-meter). The infant’s height
and lengths of the lower extremity (greater trochanter to lateral knee joint, lateral knee joint to
lateral malleolus) were measured using a standard measuring tape, and the widths of knee and
ankle were measured using a caliper. The lengths and widths of the lower extremity were
measured twice by an examiner with an intra-rater reliability of .99 (95% confidence interval, CI
11
.98-.99) estimated by the intraclass correlation coefficient (ICC), which was calculated with
SPSS ver. 25.0 (IBM Corp., Armonk, NY) based on a mean-rating (k=2), absolute-agreement, 2-
way mixed-effects model.
Participants were also evaluated using the motor subtest of the Bayley Scales of Infant
and Toddler Development, 3rd edition (BSITD-III) by one examiner with an intra-rater
reliability ICC of .89 (95% confidence interval, CI -1.35-.99) by comparing three infants’ scores
measured first at testing and second by evaluating recorded videos based on a mean-rating (k=2),
absolute-agreement, 2-way mixed-effects model. The BSITD-III is a norm-referenced test of
development with well-established psychometric properties and good test-retest reliability, r =
.67–.77, for ages 2 to 4 months (Bayley, 2005). Parents of infants born PT were contacted when
the child was 18 to 24 months corrected age to ask the age their child started walking and if their
child had been diagnosed with any medical conditions (e.g., cerebral palsy).
Data Reduction
Learning. Mobile active time (MAT) was defined as the amount of time that the infant
activated the mobile. MAT was measured for every 2-min interval. Since the mobile did not
activate during the baseline condition on Day 1, MAT during baseline was computed post hoc
using the coordinates of the foot IREDs that crossed the threshold, as if the mobile had activated.
Note that infants were excluded if the infant’s baseline MAT was less than 5% (6 sec) because
an extremely low baseline MAT may misidentify the infant as a learner. The dependent measure
for learning was the MAT ratio, which was computed as a ratio of the MAT of each condition
divided by the MAT of the baseline condition. Since three different thresholds were set to
activate the mobile, each threshold had a unique baseline MAT used to compute the MAT ratio.
12
Therefore, the MAT ratio during the baseline condition was always “1” because it was computed
as a ratio of baseline MAT over the baseline MAT.
Group learning. Learning was evaluated for Day 2 and Last Day, respectively, by
assessing if an average MAT ratio of the Day 2 (or Last Day) mobile condition was greater than
the MAT ratio of the Day 1 baseline condition. This learning evaluation is similar to previous
research (Haley et al., 2008; Sargent et al., 2018; Sargent et al., 2014).
Individual learning. Individual infants were identified as learning the task if the MAT
ratio during the entire Day 2 mobile condition was ≥1.5 (Sargent et al., 2018; Sargent et al.,
2014). The parents of infants who did not meet this criterion were asked for their infants to
participate on Day 3.
Kick height. The dependent measure for kick height was peak foot height, defined as the
vertical distance from the testing table to the highest foot marker position for each kick of either
leg. The initiation of a kick was defined as the onset of a continuous leg movement for which the
infant’s foot moved ≥5 consecutive frames (50 ms), and the hip or knee joint angle changed
>11.5° (0.2 radians) into either flexion or extension. The termination of a kick was defined as the
frame of peak flexion or extension amplitude following a movement in the opposite direction.
The definition of a kick is consistent with previous research (Fetters et al., 2004; Fetters, Sapir,
Chen, Kubo, & Tronick, 2010; Jensen, Schneider, Ulrich, Zernicke, & Thelen, 1994; Sargent et
al., 2018; Sargent et al., 2014). The peak foot height of all extracted kicks was computed using a
custom Matlab R2019b (The Mathworks, Inc., Natick, MA).
Video data. We analyzed infants’ arousal level and visual attention. Based on video
analysis, infants were excluded if they cried (Sargent et al., 2018; Sargent et al., 2014) or rolled
13
(Heathcock, Bhat, Lobo, & Galloway, 2004; Watanabe & Taga, 2009) more than two minutes on
any day during data collection.
Arousal. The dependent measure for arousal was alert time, defined as the percent of
time that the infant was alert. Video tapes were coded using Datavyu (Datavyu Team, 2014.
Databrary Project, New York University) by three evaluators blinded to group status with an
inter-rater reliability ICC of .99 (95% CI .98–.99) based on a single-rating, absolute-agreement,
2-way mixed-effects model. The descriptors of arousal consisted of sleeping, drowsy, alert,
fussy, and crying, consistent with previous research (Sargent et al., 2018; Sargent et al., 2014;
Thelen & Ulrich, 1991; Tiernan & Angulo-Barroso, 2008).
Visual attention. The dependent measure for visual attention was mobile looking time,
defined as the percent of time that the infant looked at the mobile during the task. Video tapes
were coded using Datavyu by the three evaluators blinded to group with an inter-rater reliability
ICC of .84 (95% CI .63–.95) based on a single-rating, absolute-agreement, 2-way mixed-effects
model. The descriptors of attention consisted of looking at the mobile and not looking at the
mobile consistent with previous research (Sargent et al., 2018; Sargent et al., 2014; Thelen &
Ulrich, 1991; Tiernan & Angulo-Barroso, 2008).
Statistical Analysis
Demographics. A Chi-square test and independent t-tests were used to test group
differences of demographic information using SPSS v.25.0 (IBM Corp., Armonk, NY).
MAT ratio. Mixed regression models using repeated measures were used to test
between-group (FT, PT) and within-group differences of MAT ratio between the Day 1 baseline
condition and the Day 2 and Last Day mobile condition, respectively. The interaction was also
analyzed between infant group (FT, PT) and condition (baseline, mobile).
14
Peak foot height. Mixed regression models using repeated measures were used to test
between-group (FT, PT) and within-group differences of peak foot height of the kicks between
the Day 1 baseline condition and the mean+2sd threshold condition on Day 2 and Last Day. The
interaction was analyzed between infant group (FT, PT) and threshold condition (baseline, +2sd).
Alert time and mobile looking time. Mixed regression models using repeated measures
were used to test between-group (FT, PT) and within-group differences of the percent of alert
time and mobile looking time between the Day 1 baseline condition and the Day 2 and Last Day
mobile condition. The interaction was analyzed between infant group (FT, PT) and condition
(baseline, mobile).
Mixed regression models were performed using SAS v.9.0 (SAS Institute Inc., Cary,
NC). For all mixed regression models, an autoregressive covariance structure was selected
because it was consistent with the study design and was the best fit based on Akaike Information
Criteria and Bayesian Information Criteria. Alpha level was set at .05 and Bonferroni corrections
were used to adjust for multiple comparisons.
Results
Participants
Sixty-three infants participated in the study (34 FT, 29 PT). Seventeen were excluded for
crying > 2 min (13 FT, 4 PT), 3 were excluded for rolling > 2 min (2 FT, 1 PT), 6 were excluded
by not returning to participate on Day 2 (1 FT, 5 PT), and 1 had a low Day 1 baseline MAT (1
PT). Thirty-six infants completed the study (18 FT, 18 PT). Demographics of participants are in
Tables 2.1 and 2.2. There were statistically significant differences between the two groups for
gestational age and birth weight (p < .001). There were no significant differences between the
two groups for age at time of testing (corrected for prematurity), ponderal index at testing, and
15
scores on the motor subset of the BSITD-III (p > .05). All infants scored >10th percentile on the
motor subtest of the BSITD-III, which is defined as an age-appropriate score (Bayley, 2005).
Table 2.1
Demographics of Participants
Characteristics, Mean (SD) Full-term (n = 18) Preterm (n = 18) p
Sex, N (%) .18
Female 7 (39%) 11 (61%)
Male 11 (61%) 7 (39%)
Gestational age, week 39.3 (1.0) 27.9 (2.2) <.001
Birth weight, kg 3.5 (0.3) 1.1 (0.4) <.001
Testing age, days CA 124.7 (5.1) 126.3 (4.8) .34
Ponderal index, kg/cm
3
2.45 (0.26) 2.44 (0.32) .91
BSITD-III motor, percentile 76.3 (17.4) 67.5 (18.8) .15
Note. CA, corrected age; BSITD-III motor, Bayley Scale of Infant/Toddler Development 3
rd
edition motor subtest. A Chi-square (χ
2
) test was used for sex. Independent t-tests were used
for other variables.
16
Table 2.2
Demographics of Individual Infants Born Preterm
Subject Sex
Birth
weight
(grams)
Gestational
age
(weeks)
Testing age
(days, CA)
BSITD-3
motor
(percentile)
Results of
brain
imaging
Individual
learning
1 F 1300 32.0 126 58 WNL
Not
learned
2 M 1130 27.9 122 79 WNL Learned
3 F 460 25.0 121 75 WNL
Not
learned
4 F 910 25.4 120 99 WNL Learned
5 M 960 26.0 122 58 WNL Learned
6 F 1330 30.9 132 27 WNL Learned
7 F 1380 30.1 120 79 Cystic PVL Learned
8 M 950 30.1 120 79 WNL
Not
learned
9 F 1470 30.1 133 88 WNL Learned
10 M 2030 31.0 133 68
Grade II
GMH
Learned
11 M 1065 27.3 131 58 WNL
Not
learned
12 F 890 28.0 127 42 WNL Learned
13 F 1100 27.0 128 58 WNL Learned
14 M 800 26.3 125 58
Grade II
IVH
(resolved)
Learned
15 F 900 26.0 129 75
Grade II
GMH
(resolved)
Learned
16 F 1000 26.0 129 84
Grade II
IVH
Learned
17 F 1100 28.0 123 88 WNL Learned
18 M 500 26.0 132 42 WNL
Not
learned
M 1070.8 28.0 126.3 67.5
SD 359.1 2.2 4.8 18.8
Note. BSITD-3, Bayley Scales of Infant and Toddler Development, 3
rd
edition: motor subtest
given at 4-months corrected age; CA, corrected age; DD, developmental delay; F, female;
GMH, germinal matrix hemorrhage; IVH, intraventricular hemorrhage; M, male; PVL,
periventricular leukomalacia; WNL, within normal limits. Medical history was recorded by
parental report based on the results from head ultrasound or magnetic resonance imaging.
17
Learning
Means and standard errors of the MAT ratios of infants born FT and PT are graphed in
Figure 2.3 and Table 2.3.
Figure 2.3. Mobile active time (MAT) ratios in the infants born full-term (A) and the infants
born preterm (B) over 2-min intervals on Day 1, Day 2, and Last Day; B = baseline condition; +1
= mean+1sd threshold condition; +1.5 = mean+1.5sd threshold condition; +2 = mean+2sd
threshold condition; Error bars denote standard errors; asterisks denote adjusted p < .05; The
MAT ratios are plotted for each 2-min interval to show the change in the time series, but for
statistical analysis, the mean values of the five 2-min intervals on Day 2 and Last Day,
respectively, were compared to the Day 1 baseline condition.
Table 2.3
Mobile Active Time Ratio over Conditions, Mean (SE)
Condition Full-term (n=18) Preterm (n=18) t (df)
Day 1 Baseline 1.00 1.00 -
Mobile 1.26 (0.13) 1.00 (0.13) 1.41 (34)
Day 2 Mobile 2.43 (0.22)*† 1.47 (0.22)† 3.39 (34)
Last Day Mobile 2.98 (0.32)* 2.76 (0.32)* 0.47 (34)
Note. *adjusted p < .05 within group. †adjusted p < .05 between groups.
18
Day 1 baseline vs. Day 2 mobile. The effect of condition on the MAT ratio varied by
group, F(1,34) = 5.75, p = .02. Among the infants born FT, the MAT ratio was higher during the
Day 2 mobile condition than during the Day 1 baseline condition (t = -4.89, adjusted p < .05).
Among the infants born PT, the MAT ratio was not significantly different between the Day 1
baseline condition and the Day 2 mobile condition (t = -1.50, adjusted p > .05). The MAT ratio
during the Day 2 mobile condition was higher in the infants born FT than in the infants born PT
(t = 3.39, adjusted p < .05). These results suggest that the infants born FT learned the scaffolded
mobile task on Day 2 while the infants born PT did not learn the task.
Individual data. Based on the individual learning criteria, 13 of 18 (72%) infants born FT
and 8 of 18 (44%) infants born PT learned the task on Day 2, which was analyzed as their last
day of participation. The parents of infants who did not learn (5 FT, 10 PT) were asked to
participate in a 10-min session on the following day (Day 3). Three infants (2 FT, 1 PT) did not
participate on Day 3 and Day 2 was analyzed as their last day of participation. Twelve infants (3
FT, 9 PT) participated on Day 3, which was analyzed as their last day of participation. On Day 3,
an additional seven infants (2 FT, 5 PT) met the learning criterion.
Day 1 baseline vs. Last day mobile. The main effect of condition on the MAT ratio was
significant, F(1,34) = 33.71, p < .0001. The MAT ratio was higher during the Last Day mobile
condition, compared to the Day 1 baseline condition, in both the infants born FT (t = -4.34,
adjusted p < .05) and PT (t = -3.87, adjusted p < .05). The main effect of group was not
significant, F(1,34) = 0.11, p = .74 and the interaction effect was not significant, F(1,34) = 0.11,
p = .74. These results suggest that both the infants born FT and PT learned the scaffolded mobile
task on their last day of participation.
19
Individual data. Based on the individual learning criteria, 15 of 18 (83%) infants born FT
and 13 of 18 (72%) infants born PT learned the task on Day 2 (13 FT, 8 PT) or Day 3 (2 FT, 5
PT).
Kick Height
Means and standard errors of the kick heights, the analyzed number of kicks, and means
and standard deviations of individually determined threshold heights of infants born FT and PT
are graphed in Figure 2.4 and in Tables 2.4 and 2.5.
Figure 2.4. Kick height in the infants born full-term (A) and the infants born preterm (B) over
threshold conditions on Day 1, Day 2, and Last Day; B = baseline condition; +1 = mean+1sd
threshold condition; +1.5 = mean+1.5sd threshold condition; +2 = mean+2sd threshold
condition; Error bars denote standard errors; asterisks denote p < .05; The kick heights are
plotted for each threshold condition to show the change in the time series, but for statistical
analysis, the kick heights during the +2sd threshold conditions on Day 2 and Last Day,
respectively, were compared to the Day 1 baseline condition.
Day 1 baseline vs. Day 2 threshold. The effect of threshold condition on the kick height
varied by group, F(1,34) = 18.86, p < .001. During the baseline condition, the kick height was
20
not significantly different between the infants born FT (mean = 11.0 cm) and PT (mean = 11.9
cm) (t = -1.60, adjusted p > .05). Among the infants born FT, the kick height was higher during
the mean+2sd threshold condition on Day 2 (mean = 14.7 cm), compared to the Day 1 baseline
condition (t = -6.93, adjusted p < .05). However, among the infants born PT, the kick height was
not significantly different between the Day 1 baseline condition and the mean+2sd threshold
condition on Day 2 (mean = 12.5 cm) (t = -0.95, adjusted p > .05). During the Day 2 threshold
condition, the kick height was higher for the infants born FT than for the infants born PT (t =
5.22, adjusted p < .05). These results can be interpreted as the infants born FT, but not PT,
increased their kick heights during the Day 2 highest threshold condition, compared to the Day 1
baseline condition.
Day 1 baseline vs. Last day threshold. The effect of threshold condition on the kick
height varied by group, F(1,34) = 11.97, p < .01. The kick height was higher during the
mean+2sd threshold condition on the Last Day, compared to the Day 1 baseline condition, in
both the infants born FT (mean = 15.0 cm, t = -9.47, adjusted p < .05) and the infants born PT
(mean = 13.4 cm, t = -7.46, adjusted p < .05). During the Last Day threshold condition, the kick
height was higher in the infants born FT than in the infants born PT (t = 4.05, adjusted p < .05).
These results can be interpreted as both the infants born FT and PT increased their kick heights
during the Last Day highest threshold condition, compared to the Day 1 baseline condition, but
the kick height was higher in the infants born FT than that in the infants born PT.
21
Table 2.5
Number of Kicks Analyzed and Kick Height (cm) over Threshold Conditions, Mean (SE)
Threshold
condition
Full-term (n=18) Preterm (n=18)
Number of
kicks analyzed
Kick height
Number of
kicks analyzed
Kick height
Day 1 Baseline 821 11.0 (0.5) 793 11.9 (0.5)
Mean + 1sd 2,506 12.7 (0.3) 3,791 11.5 (0.2)
Day 2 Mean + 1sd 603 13.9 (0.5) 667 12.0 (0.5)
Mean + 1.5sd 1,475 14.5 (0.3) 1,768 12.6 (0.3)
Mean + 2sd 1,754 14.7 (0.3) 2,218 12.5 (0.3)
Last Day Mean + 1sd 621 13.7 (0.5) 783 14.0 (0.5)
Mean + 1.5sd 1,420 15.0 (0.4) 2,155 14.2 (0.3)
Mean + 2sd 1,657 15.0 (0.3) 2,489 13.4 (0.3)
Total number
of kicks
10,857
Total number
of kicks
14,664
Table 2.4
Threshold Height (cm), Mean (SD)
Threshold
condition
Full-term (n=18) Preterm (n=18)
Mean+1sd 10.2 (4.9) 9.3 (4.5)
Mean+1.5sd 11.9 (5.1) 11.0 (4.8)
Mean+2sd 13.3 (5.3) 12.5 (4.9)
22
Arousal
Means and standard errors of the alert times of infants born FT and PT are in Table 2.6.
Table 2.6
Alert Time (percentage) over Conditions, Mean (SE)
Condition Full-term (n=18) Preterm (n=18) t (df)
Day 1 Baseline 99.7 (1.4) 99.6 (1.4) 0.04 (34)
Mobile 94.5 (1.4) 96.7 (1.4) -1.15 (34)
Day 2 Mobile 88.5 (2.8)* 93.3 (2.8) -1.19 (34)
Last Day Mobile 93.2 (2.9) 91.2 (2.9) 0.48 (34)
Note. *adjusted p < .05 within group. No statistically significant group difference.
Day 1 baseline vs. Day 2 mobile. The main effect of condition on the alert time was
significant, F(1,34) = 9.40, p < .01. Among the infants born FT, the alert time was shorter during
the Day 2 mobile condition, compared to the Day 1 baseline condition (t = 2.77, adjusted p <
.05) while among the infants born PT, the alert time was not significantly different between the
Day 1 baseline condition and the Day 2 mobile condition (t = 1.56, adjusted p > .05). However,
the main effect of group was not significant, F(1,34) = 0.78, p = .38, and the interaction effect
was also not significant, F(1,34) = 0.68, p = .42. These results can be interpreted as the infants
born FT were less alert while participating in the task on Day 2, compared to Day 1, while the
infants born PT did not change their alertness.
The mean alert time during the Day 1 baseline condition was 99.7% in the infants born
FT and 99.6% in the infants born PT. The mean and the range of 2-min intervals during the Day
2 mobile condition were 88.5% (81.2 - 96.4%) in the infants born FT and 93.3% (85.7 - 99.5%)
in the infants born PT.
23
Day 1 baseline vs. Last day mobile. The main effect of condition on the alert time was
significant, F(1,34) = 6.61, p = .01. However, pairwise comparisons show that the alert time
during the Day 1 baseline condition was not significantly different from that during the Last Day
mobile condition in either the infants born FT (t = 1.59, adjusted p > .05) or the infants born PT
(t = 2.05 adjusted p > .05). The main effect of group was not significant, F(1,34) = 0.12, p = .73,
and the interaction effect was also not significant, F(1,34) = 0.11, p = .75. These results can be
interpreted as the alert time remained unchanged in both the infants born FT and the infants born
PT while participating in the task on the Last Day.
The mean mobile looking time and the range of 2-min intervals during the Last Day
mobile condition were 93.2% (89.2 - 95.0%) in the infants born FT and 91.2% (85.2 - 99.2%) in
the infants born PT.
Visual Attention
Means and standard errors of the mobile looking times of infants born FT and PT are in
Table 2.7.
Table 2.7
Mobile Looking Time (Percentage) over Conditions, Mean (SE)
Condition Full-Term (n=18) Preterm (n=18) t (df)
Day 1 Baseline 74.0 (4.4) 73.8 (4.4) 0.03 (34)
Mobile 80.9 (4.4) 84.2 (4.4)* -0.52 (34)
Day 2 Mobile 78.3 (4.1) 87.3 (4.1)* -1.57 (34)
Last Day Mobile 78.5 (4.6) 83.0 (4.6) -0.70 (34)
Note. *adjusted p < .05 within group. No statistically significant group difference.
24
Day 1 baseline vs. Day 2 mobile. The main effect of condition on the mobile looking
time was significant, F(1,34) = 7.76, p < .01. Among the infants born FT, the mobile looking
time was not significantly different between the Day 1 baseline condition and the Day 2 mobile
condition (t = -0.95, adjusted p > .05) while among the infants born PT, the mobile looking time
was longer during the Day 2 mobile condition, compared to the Day 1 baseline condition (t = -
2.99, adjusted p < .05). The main effect of group was not significant, F(1,34) = 0.85, p = .36, and
the interaction effect was also not significant, F(1,34) = 2.07, p = .16. These results can be
interpreted as the infants born PT looked at the mobile longer while participating in the task on
Day 2 than on Day 1, while infants born FT did not change their mobile looking time.
The mean mobile looking time during the Day 1 baseline condition was 74.0% in the
infants born FT and 73.8% in the infants born PT. The mean and the range of 2-min intervals
during the Day 2 mobile condition were 78.3% (75.4 - 81.6%) in the infants born FT and 87.3%
(86.4 - 88.3%) in the infants born PT.
Day 1 baseline vs. Last day mobile. There was no significant main effect of condition,
F(1,34) = 4.32, p = .05, or group, F(1,34) = 0.15, p = .70, or the interaction effect, F(1,34) =
0.51, p = .48. These results can be interpreted as both the infants born FT and PT did not change
their visual attention while participating in the task on the Last Day.
The mean mobile looking time and the range of 2-min intervals during the Last Day
mobile condition were 78.5% (75.9 – 82.4%) in the infants born FT and 83.0% (78.2 - 87.6%) in
the infants born PT.
Discussion
This study investigated if infants born FT and infants born PT at 4 months of age could
learn a task that required progressively higher kicks to activate a mobile. As hypothesized, both
25
the FT and PT groups learned the scaffolded mobile task. The FT group learned the task on Day
2, and the PT group learned the task on Day 3 after an extra day of practice. This is consistent
with previous research that reported slower learning rates for infants born PT compared to
infants born FT during the mobile paradigm (Gekoski et al., 1984; Haley et al., 2008). However,
research also supports that infants born PT and infants born FT have similar learning rates
(Haley, Weinberg, & Grunau, 2006), and that infants born PT did not learn the mobile task even
when provided additional time (Heathcock et al., 2004, 2005). We attribute the conflicting
findings to differences in task demands. When a task requires movements that are within the
infants’ preferred repertoire (e.g., increasing the frequency of kicks), infants born PT may
achieve learning of the task at a similar rate as infants born FT (Haley et al., 2006). However, if
the task is challenging and requires the infants to use movements that are not within their
preferred repertoire (e.g., kicking only a reinforced leg), infants born PT may not demonstrate
learning (Heathcock et al., 2004, 2005). Our task required infants born PT to use movements that
are not within their preferred repertoire particularly during the highest threshold condition, but
we speculate that the scaffolding assisted the infants to learn the task after an extra day of
practice.
We also investigated if infants born FT and PT would kick their legs higher to activate
the mobile. As hypothesized, both FT and PT groups increased their kick heights during the
highest threshold for mobile reinforcement. Infants born PT have known differences in strength
and motor control that may make anti-gravity movements challenging (Dassios et al., 2018;
Sargent et al., 2017). Participation in scaffolded tasks, such as this one, has potential as an early
intervention to both support learning of motor tasks and motivate more antigravity leg
movements to develop strength and motor control.
26
We assessed if infants were participating in the task by analyzing arousal level and
mobile looking time. Both FT and PT groups were alert for over 80% of the task and looked at
the mobile for over 70% of the task. This is consistent with previous research that documented
alertness of infants born FT and PT as ranging from 80-100% of the mobile task (Haley et al.,
2008; Sargent et al., 2018; Sargent et al., 2014). The result is also consistent with a previous
study that reported the mobile looking time as ranging from 58-85% of the mobile task in infants
born PT (Sargent et al., 2018). The mobile looking time of infants born FT in our task ranged
from 75-84%, which is lower than the 94-99% of mobile looking time reported in Sargent et al.
(2014). This inconsistency may be because the mobile in Sargent et al. (2014) was located
directly over the heads of the 3-month-old infants, attracting their attention to the mobile. In the
current study, the mobile was located over-the-legs of the 4-month-old infants. In addition,
infants at 4 months, who have better ability to control their head, may explore their surroundings
rather than attend to the mobile, especially if the mobile is not located directly overhead (Lima-
Alvarez, Tudella, van der Kamp, & Savelsbergh, 2014).
Our results suggest that a scaffolded environment may support independent learning of
motor tasks for infants born PT. Scaffolding, as a metaphor in development, refers to providing
additional support, which is usually offered by parents or other knowledgeable adults, to enable
infants or children to accomplish a task that challenges their current capabilities (Stone, 1998;
Wood, Bruner, & Ross, 1976). However, little is known about scaffolding a task itself or
designing tasks differently for infants born PT who may have learning and motor impairments.
Here, we define scaffolding of a task as systematically increasing the level of demand to
accomplish a final goal. As an example, Angulo-Kinzler (2001) provided a shaping
reinforcement schedule during the mobile paradigm for infants born FT who did not produce a
27
sufficient number of hip extension paired with knee extension kicks to activate the mobile
(Angulo-Kinzler, 2001). First, hip extension or knee extension kicks were reinforced, then hip
extension paired with knee extension kicks were reinforced. Scaffolding the task may improve
performance and learning for infants born PT at high risk for learning and motor impairments.
This study has limitations. First, we did not have a ‘non-scaffolded’ control group, so
conclusions cannot be drawn about learning and increased kick heights without scaffolding.
However, we speculate that the use of scaffolding helped the infants born PT to learn this
challenging task. Second, we did not investigate differences in kick height, arousal, and attention
between learners (n=28) and non-learners (n=8) due to the small sample size of non-learners.
Stratifying infants into learner and non-learner subgroups may have provided additional data
about the factors that contributed to the learning process.
In summary, our findings support previous research that infants born PT and infants born
FT have the ability to learn the mobile contingency with an appropriate task demand. This study
further contributes to our knowledge of the ability of infants born PT to adapt to challenging task
demands and generate higher anti-gravity kicks with a scaffolded task. The present findings
emphasize the importance of scaffolding task environments for educational and clinical practice
for infants born PT at risk for learning and motor impairments. These infants may require more
systematic scaffolding based on their individual learning rate to support their ability to
independently perform motor tasks. Future studies on task scaffolding may provide insight into
novel designs of scaffolded tasks that could be applicable to specific educational and clinical
practice.
28
Chapter III: Infants Born Preterm and Infants Born Full-term Generate More
Selective Leg Joint Movement during the Scaffolded Mobile Task
Abstract
Aim: To investigate whether infants born very preterm (PT) at increased risk for cerebral palsy
generate more selective hip-knee joint movement (e.g., hip flexes while knee extends) while
participating in a scaffolded, motor learning task.
Method: In this cross-sectional study, infants born PT and infants born full-term (FT) at 4
months corrected age participated in a mobile task for 2-3 consecutive days. We used a
scaffolded mobile task that required infants to raise their legs vertically over a virtual threshold.
Three threshold heights (low, middle, high) were used to test if the middle and high heights
encourage infants to move more selectively.
Results: Fifteen infants born FT (5 females, 10 males; mean age [SD] 4 months 3.7 days [4.8
days]) and 13 infants born PT (9 females, 4 males; 4 months 6.4 days [4.7 days]) learned the
task. The infants born FT showed more selective hip-knee movement on their learning day at
each of the three threshold heights, compared to their baseline spontaneous kicking. The infants
born PT showed more selective hip-knee movement on their learning day, but only when the
middle and high thresholds were used.
Interpretation: The scaffolded task effectively encouraged infants born PT to generate more
selective hip-knee joint movement.
Conclusion: Infants born PT, at increased risk of cerebral palsy, can generate more selective leg
movement when a structured task systematically reinforces the specific movement.
29
Introduction
Cerebral palsy is a disorder of posture and movement caused by damage to the
developing brain (Rosenbaum et al., 2007). Children with spastic cerebral palsy have impaired
selective leg joint movement which contributes to lifelong walking limitations (Farmer, 2003;
Farmer et al., 2008; Fowler et al., 2010). Selective joint movement is defined as the ability to
isolate the movement of one joint from the movement of the other joints within the limb (e.g.,
extending the knee while flexing the hip) (Chen, Fetters, Holt, & Saltzman, 2002; Fetters et al.,
2004; Fetters et al., 2010; Fowler & Goldberg, 2009; Sargent et al., 2018; Sargent et al., 2014).
Impaired selective leg movement results in excessively coupled hip and knee flexion and
extension during the gait of children with spastic cerebral palsy, contributing to short stride
length, slow walking speed, and excessive energy consumption (Farmer, 2003; Fowler &
Goldberg, 2009).
Infants born very preterm (PT), prior to 32 weeks of gestation, are at increased risk of
developing cerebral palsy; 6-15% develop cerebral palsy (Himpens et al., 2008; Pascal et al.,
2018). Infants born PT with brain lesions are at highest risk for developing cerebral palsy
(Gopagondanahalli et al., 2016; Hagberg et al., 2001), and show impaired selective leg
movement during spontaneous kicking as early as one month of age, corrected for prematurity
(Fetters et al., 2004). We do not know if infants born PT with brain lesions, at highest risk of
cerebral palsy, have the capacity to produce more selective leg movement when a task reinforces
them to move more selectively. As a first step, we developed a task to encourage selective leg
movement in infants born PT without brain lesions, at increased risk for cerebral palsy.
One promising way to encourage selective leg movement is the mobile task (Sargent et
al., 2018; Sargent et al., 2014). In the mobile task, infants are supine under an infant mobile that
30
rotates and plays music only when the infants lift their legs vertically above an individually
determined threshold height. Infants born full-term (FT) who learned the association between
their leg movements and the mobile activation demonstrated more selective hip-knee
coordination when interacting with the mobile, compared to their baseline spontaneous kicking
(Sargent et al., 2014). However, infants born PT who learned the association did not change their
hip-knee coordination throughout the task (Sargent et al., 2018). Infants born PT may need
further support during the task, or scaffolding, to produce more selective hip-knee movement.
Here, we use the term “scaffolding” as modifying a task so that the infant is first reinforced for
an easily attainable goal, then reinforced for progressively more challenging goals, until the
infant is finally reinforced for the most challenging, but still achievable task goal.
We developed a scaffolded mobile task to motivate selective hip-knee movement for
infants born PT (refer to Chapter II). This task was designed to encourage infants to kick high by
using three different heights of virtual threshold (low, middle, high; three dotted lines in Figure
2.1). The infant kick-activated mobile plays music and rotates when the threshold is crossed by a
specific foot marker on each of the infant’s feet (a solid circle in Figure 2.1). On the first day,
only the low threshold was used to support infants to learn the association between their leg
movements and mobile activation. On the second and third days, we systematically increased the
threshold from the low to the middle to the high height to encourage infants to kick higher to
activate the mobile.
We found that 4-month-old infants born FT and infants born PT both demonstrated
learning of the scaffolded mobile task and increased kick heights when the high threshold was
used. We hypothesize that this task using high thresholds may motivate infants born PT to
generate more selective hip-knee movement by requiring them to kick high by extending the
31
knee while flexing the hip. To test this hypothesis, in the present study, we investigated the
kinematic data of kicks in the infants born FT and PT who learned the scaffolded mobile task to
determine if they demonstrate more selective hip-knee movement.
The objective of this study was to assess if infants born FT and infants born PT who
learned the scaffolded mobile task had more selective hip-knee joint movement when interacting
with the mobile, compared to their baseline spontaneous kicking. We tested three hypotheses, 1)
on the first day of participation, infants born FT and infants born PT would not demonstrate
more selective hip-knee movement when interacting with the mobile, compared to their baseline
spontaneous kicking, consistent with previous research (Sargent et al., 2018; Sargent et al.,
2014), 2) on the day that infants born FT learned the task, they would exhibit more selective hip-
knee movement when interacting with the mobile during the intervals using the low, middle, and
high threshold heights, compared to their baseline spontaneous kicking, consistent with previous
research (Sargent et al., 2014), and 3) on the day that infants born PT learned the task, they
would not exhibit more selective hip-knee movement during the interval using the low threshold
height, consistent with previous research (Sargent et al., 2018), but would exhibit more selective
hip-knee movement during the intervals using the middle and high threshold heights, compared
to their baseline spontaneous kicking.
Method
Participants
This study was approved by the University of Southern California (USC) Health Sciences
Institutional Review Board (#HS-17-00119). Refer to Chapter II for details of protocol. Infants
born PT (<32 weeks gestational age) and infants born FT (>37 weeks gestational age) were
recruited from USC and the Los Angeles County + USC Medical Center. Infants participated
32
between the ages of 4 months 0 day to 4 months 15 days. For infants born PT, corrected age was
used. Infants were excluded based on parent report if they were ill or had congenital
malformations, chromosomal abnormalities, orthopedic impairments, or severe vision or hearing
impairments that would hinder their ability to see or hear the mobile. In this analysis, data were
excluded if infants did not learn the scaffolded mobile task.
Experimental Protocol
Data were collected in the Development of Infant Motor Performance Laboratory at
USC. Study procedures were explained to the parents upon arrival and written informed consent
was obtained before data collection. Throughout data collection, the parent sat next to the infant,
out of the infant’s sight, and was instructed to not touch the infant and stay quiet, but use
soothing words (e.g., “you are okay”) if the infant fussed or cried.
Scaffolded mobile task. All infants participated in the 10-min scaffolded mobile task for
2 to 3 consecutive days. Each day, infants were supine on a testing table with their feet under an
infant kick-activated mobile.
On Day 1, the 10-min scaffolded mobile task consisted of a 2-min baseline condition
followed by an 8-min mobile condition. During the baseline condition, the mobile did not move
but each infant kicked spontaneously. The vertical heights of the center marker (infrared light-
emitting diode; IRED) on each foot were collected to compute each infant’s individual virtual
threshold. This was calculated as the average of the mean height of the right foot IRED plus one
standard deviation (sd) and the mean height of the left foot IRED plus 1 sd; refer to equation in
Chapter II. During the mobile condition, the musical mobile rotated only if the infant lifted either
foot IRED vertically higher than the virtual threshold. Once either foot crossed the threshold, the
musical mobile rotated for the duration the IRED stayed above the threshold up to a maximum of
33
3 sec. To reactivate the mobile after 3 sec, infants were required to lower the foot below the
threshold and then cross the threshold again. This 3-sec setting was used to reinforce infants to
continue moving their legs rather than holding their feet above the threshold.
On Day 2, the task consisted of a 10-min mobile condition in which the threshold height
was systematically increased. For the first 2-min (+1sd threshold interval), the threshold was
consistent with the mobile condition on Day 1 (mean plus 1sd baseline foot height). For the
following 4-min (+1.5sd threshold interval), the threshold was increased to the mean baseline
foot height plus 1.5sd. For the last 4-min (+2sd threshold interval), the threshold was increased
to the mean baseline foot height plus 2sd. Individual infants were identified as learners if the
duration that the infant activated the mobile during the entire Day 2 mobile condition was ≥1.5
times the duration that the infant would have activated the mobile during the Day 1 baseline
assuming the mobile could be activated. The parents of infants that did not meet this individual
learning criterion were asked to participate in the identical task on the next day (Day 3) to
provide the infants more practice time.
Video data. During the scaffolded mobile task, a video camera (Canon Inc., VIXIA HF
R700 Camcorder) was positioned overhead to record infants’ facial expressions and eye gaze.
Kinematic data. During the scaffolded mobile task, the three-dimensional position data
of 7-mm IREDs were collected using an Optotrak Certus Motion Capture System (Northern
Digital Inc., Waterloo, ON, Canada). Custom, rigid marker arrays with four embedded IREDs
were attached to the infants’ pelvis, thigh, shank, and foot bilaterally using Velcro straps. A
custom marker array with two embedded IREDs was attached to the infants’ sternum using
double-sided adhesive collar tape.
34
The scaffolded mobile task was followed by a 5-sec static calibration trial for each leg,
which defined the hip, knee and ankle angles at 0° (Sargent et al., 2014). During this trial, an
additional 5 individual IREDs were attached to the infants’ skin bilaterally at the lateral midline
of the trunk below the 10
th
rib, greater trochanter of the hip, lateral knee joint line, ankle lateral
malleolus, and distal end of the 5
th
metatarsal using double-sided collar tape. During the static
trial, the infants’ straight leg positions were held by an experimenter for 5 seconds.
Anthropometric and developmental measures. On Day 2, after the scaffolded mobile
task, an experimenter measured infants’ weights, heights, lengths of the lower extremity (greater
trochanter to lateral knee joint, lateral knee joint to lateral malleolus), and widths of knee and
ankle. Infants were also evaluated using the motor subtests of the Bayley Scales of Infant and
Toddler Development, 3rd Edition (BSITD-III) (Bayley, 2006). Refer to Chapter II for details.
Data Reduction
A custom Matlab R2019b (The Mathworks, Inc., Natick, MA) was used to process the
position data of the IREDs as follows: (a) interpolate missing position data up to 20 consecutive
frames using a cubic spline, (b) extract segments longer than 25 frames without missing position
data of all IREDs, (c) filter position data using a fourth-order Butterworth with a cut-off
frequency of 5 Hz, (d) compute hip and knee joint angles of flexion and extension, and (e)
extract kicks (Sargent et al., 2018; Sargent et al., 2014). A kick onset was defined as a change
into hip or knee flexion/extension angle of >11.5°, and the termination of a kick was defined as
the peak flexion or extension amplitude following a movement in the opposite direction (Fetters
et al., 2004; Fetters et al., 2010; Jensen et al., 1994; Sargent et al., 2018; Sargent et al., 2014).
For each extracted kick, selective movement was quantified using Pearson correlation
coefficients (r) between hip and knee joint angles (Hip-Knee CC). The Hip-Knee CC values
35
were converted to Fisher Z scores for statistical comparisons among infants (Hip-Knee ZCC).
The mean Hip-Knee ZCC of all kicks was computed for each infant. A more negative Hip-Knee
ZCC was interpreted as more selective movement, and a more positive Hip-Knee ZCC was
interpreted as less selective movement (Chen et al., 2002; Fetters et al., 2010; Sargent et al.,
2018; Sargent et al., 2014).
To further investigate selective movement, relative phase was also measured between hip
and knee joint angles (Hip-Knee RP) at joint reversal for each extracted kick. The absolute value
of RP was computed to analyze the magnitude of selective movement. The mean Hip-Knee RP
of all kicks was computed for each infant. A Hip-Knee RP approaching 0˚ was interpreted as less
selective movement, and a Hip-Knee RP approaching 180˚ was interpreted as more selective
movement.
Statistical Analysis
Demographic data. Independent t-tests and Chi-square tests were performed to test
group differences of demographic data using SPSS v.25.0 (IBM Corp., Armonk, NY).
Kinematic data. For each Hip-Knee ZCC and RP, mixed regression models using
repeated measures were performed using SAS v.9.0 (SAS Institute Inc., Cary, NC). An
autoregressive covariance structure for all mixed regression models was chosen for our repeated-
measure design. Post-hoc comparisons were performed using Bonferroni correction to adjust for
multiple comparisons. The between- and within-group differences of Hip-Knee ZCC and RP,
respectively, were tested: 1) between the 2-min baseline and the entire 8-min mobile conditions
on Day 1, and 2) between the 2-min baseline condition on Day 1 and the entire 10-min mobile
condition on the day that infants learned the task (Day 2 or 3; learning day). The interactions
between group (FT, PT) and condition (baseline, mobile) were also tested. The Bonferroni
36
adjusted alpha was set at .013 (.05/4 comparisons). The within-group differences of Hip-Knee
ZCC and RP were further tested between the baseline condition on Day 1 and each of the
threshold intervals (+1sd, +1.5sd, +2sd) on the learning day. The Bonferroni adjusted alpha was
set at .013 (.05/4 comparisons).
Results
Participants
Fifteen infants born FT and 13 infants born PT learned the scaffolded mobile task on
either Day 2 (13 FT, 8 PT) or Day 3 (2 FT, 5 PT), which was defined as the learning day (Refer
to Chapter II for details). Demographics of the infants who learned the task (15 FT, 13 PT) are in
Table 3.1. There were statistically significant differences between the infants born FT and the
infants born PT for gestational age and birth weight. The infants born FT were born at a mean of
39.0 weeks gestation with a mean birth weight of 3.5 kg. The infants born PT were born at a
mean of 27.9 weeks gestation with a mean birth weight of 1.2 kg. There were no statistically
significant differences between the two groups for sex, age at testing, ponderal index at testing,
and scores on the motor subset of the BSITD-III at testing.
Table 3.1
Demographics of Participants
Characteristics, Mean (SD) Full-term learner (n=15) Preterm learner (n=13) p
Sex, N (%) .06
Female 5 (33%) 9 (69%)
Male 10 (67%) 4 (31%)
Gestational age, week 39.0 (0.9) 27.9 (2.1) < .001
Birth weight, kg 3.5 (0.3) 1.2 (0.3) < .001
Testing age, days corrected age 123.7 (4.8) 126.4 (4.7) .15
Ponderal index, g/cm
3
2.45 (0.28) 2.54 (0.29) .42
BSITD-III motor, percentile 77.9 (17.3) 69.5 (20.3) .25
37
Note. BSITD-III motor, Bayley Scale of Infant/Toddler Development 3
rd
edition motor subtest.
A Chi-square (χ
2
) test was used for sex. Independent t-tests were used for other variables.
Selective Hip-Knee Movement
Means and standard errors of the Hip-Knee ZCC and the number of kicks analyzed in
infants born FT and PT on Day 1 and Learning day are in Table 3.2 for each condition and in
Table 3.3 for each threshold interval. Refer to Tables 3.4 and 3.5 for details on the Hip-Knee RP
results.
Table 3.3
Hip-Knee Z-transformed Correlation Coefficient (ZCC) over Threshold Interval
Threshold
interval
Full-term
learner
(n=15)
Number of
analyzed
kicks
Preterm
learner
(n=13)
Number of
analyzed
kicks
Day 1 Baseline 0.67 (0.07) 673 0.60 (0.07) 528
Mean + 1sd 0.26 (0.04)* 2,230 0.69 (0.03) 2,548
Learning Day Mean + 1sd 0.04 (0.07)* 526 0.45 (0.09) 483
Mean + 1.5sd 0.15 (0.05)* 1,232 0.24 (0.05)* 1,431
Mean + 2sd 0.26 (0.04)* 1,472 0.29 (0.04)* 1,890
Note. *adjusted p <. 05. Each day and group separately.
Table 3.2
Hip-Knee Z-transformed Correlation Coefficient (ZCC) over Condition
Condition
Full-term
learner
(n=15)
Analyzed
kick
number
Preterm
learner
(n=13)
Analyzed
kick
number
t
(between-
group)
Adjusted
p
Day 1 Baseline
0.56
(0.07)
673
0.60
(0.07)
528 -0.40 > .05
Mobile
0.24
(0.03)*
2,230
0.69
(0.03)
2,548 -9.42 < .05
Learning
Day
Mobile
0.13
(0.03)*
3,265
0.29
(0.03)*
3,804 -3.64 < .05
Note. *adjusted p <. 05. Each day separately.
38
Day 1. The effect of condition on the Hip-Knee ZCC varied by infant group on Day 1,
F(1, 26) = 14.37, p < .001. There was no group difference during the baseline condition (t = -
0.40, adjusted p > .05). Among the infants born FT, the Hip-Knee ZCC was decreased during the
mobile condition, compared to the baseline condition (t = 4.32, adjusted p < .05) while it was not
significantly different in the infants born PT (t = -1.19, adjusted p > .05). The between-group
difference was significant during the mobile condition (t = -9.42, adjusted p < .05). These results
suggest that the infants born FT generated more selective hip-knee movement on the first day
when interacting with the mobile, compared to when kicking spontaneously, but the infants born
PT did not. Hip-Knee RP results show a similar trend, which is more selective movement during
the mobile condition for the infants born FT, but not PT, although the results were not
statistically significant.
Learning day. The main effect of condition on the Hip-Knee ZCC was significant on the
learning day, F(1, 26) = 40.34, p < .0001. The Hip-Knee ZCC was decreased during the mobile
condition, compared to the Day 1 baseline condition, in both the infants born FT (t = 5.53,
adjusted p < .05) and the infants born PT (t = 3.55, adjusted p < .05). However, the main effect
of group was not significant, F(1, 26) = 2.85, p = .10, and the interaction effect was also not
significant, F(1, 26) = 1.21, p = .28. These results suggest that both the infants born FT and the
infants born PT generated more selective hip-knee movement when interacting with the mobile
on the day that they learned the task, compared to when kicking spontaneously. Hip-Knee RP
results show a similar trend, more selective movement during the mobile condition for both
infant groups, although the results were only statistically significant for the infants born PT.
Threshold intervals in infants born FT. The effect of threshold interval on the Hip-Knee
ZCC was significant on the learning day in the infants born FT, F(3, 42) = 14.14, p < .0001. The
39
Hip-Knee ZCC was decreased during all three threshold intervals; the low (mean+1sd; t = 5.44,
adjusted p < .05), middle (mean+1.5sd; t = 5.95, adjusted p < .05), and high (mean+2sd; t = 4.39,
adjusted p < .05), compared to the Day 1 baseline, respectively. These results suggest that the
infants born FT generated more selective hip-knee movement on the day that they learned the
task at each of the three threshold intervals. Hip-Knee RP results show a similar trend, more
selective movement during all three threshold intervals, but the significant difference was found
only during the mean+1.5sd threshold interval.
Threshold intervals in infants born PT. The effect of threshold interval on the Hip-Knee
ZCC was significant on the learning day in the infants born PT, F(3, 36) = 5.20, p < .01. The
Hip-Knee ZCC was decreased during the middle (t = 3.58, adjusted p < .05) and high (t = 3.23,
adjusted p < .05) threshold intervals, but not during the low threshold interval (t = 1.19, adjusted
p > .05). These results suggest that the infants born PT generated more selective movement only
when the middle and high thresholds were used, but not when the low threshold was used. Hip-
Knee RP results are consistent with ZCC results.
40
Table 3.5
Hip-Knee Relative Phase (RP) over Threshold Interval
Threshold
interval
Full-term
learner
(n=15)
Number of
analyzed
kicks
Preterm
learner
(n=13)
Number of
analyzed
kicks
Day 1 Baseline 116.4 (2.4) 673 111.4 (3.0) 528
Mean + 1sd 119.7 (1.3) 2,230 116.1 (1.4) 2,548
Learning Day Mean + 1sd 125.2 (2.7) 526 117.5 (3.1) 483
Mean + 1.5sd 126.0 (1.8)* 1,232 121.5 (1.8)* 1,431
Mean + 2sd 118.5 (1.7) 1,472 127.5 (1.6)* 1,890
Note. *adjusted p <. 05. Each day and group separately.
Table 3.4
Hip-Knee Relative Phase (RP) at Joint Reversal over Condition
Condition
Full-term
learner
(n=15)
Analyzed
kick
number
Preterm
learner
(n=13)
Analyzed
kick
number
t
(between-
group)
Adjusted
p
Day 1 Baseline
116.3
(2.6)
673
111.3
(2.9)
528 1.28 > .05
Mobile
119.7
(1.4)
2,230
116.1
(1.3)
2,548 -1.51 > .05
Learning
Day
Mobile
122.5
(1.2)
3,265
124.0
(1.1)*
3,804 -0.95 > .05
Note. *adjusted p <. 05. Each day separately.
41
Discussion
The current study investigated if 4-month-old infants who learned the scaffolded mobile
task demonstrated more selective hip-knee joint movement. Infants born FT, contrary to our
hypothesis, demonstrated more selective hip-knee movement on the first day during the mobile
condition that offered the low threshold height, compared to their baseline spontaneous kicking.
This finding is different from a previous finding from a similar mobile task of 3-month-old
infants born FT in which the infants did not demonstrate more selective hip-knee movement on
the first day (Sargent et al., 2014). We speculate that the infants born FT at 4 months in the
current study were able to better disassociate their joint movement than at a younger age (E.
Thelen, 1985). Infants born PT at 4 months corrected age, as we hypothesized, did not exhibit
more selective hip-knee movement on the first day during the mobile condition, compared to
their baseline spontaneous kicking. This is consistent with a previous finding from a similar
mobile task of infants born PT at 4 months corrected age (Sargent et al., 2018).
On the day of learning, as hypothesized, the infants born FT demonstrated more selective
hip-knee movement during the scaffolded mobile condition, compared to their baseline
spontaneous kicking. This finding is similar to that of Sargent et al. (2014) in that 3-month-old
infants born FT who learned the mobile task demonstrated more selective hip-knee movement,
compared to their spontaneous kicking. The authors speculated that the infants born FT changed
their coordination to activate the mobile more efficiently by holding their hip flexed while
bending and straightening their knee joints to keep the feet right around the threshold. This
previous study, however, only used the low threshold (mean+1sd) throughout the task, whereas
the current study added the middle and high thresholds (mean+1.5sd and +2sd) on the second
and third day. Hence, we further analyzed associations between the three thresholds and selective
42
movement. Our results show that the infants born FT, as hypothesized, exhibited more selective
joint movement at each of the three threshold heights. This finding may imply that infants born
FT can generate more selective hip-knee movement even without task scaffolding.
As hypothesized, the infants born PT also exhibited more selective hip-knee movement
on the day of learning, compared to their baseline spontaneous kicking. This result highlights the
positive effect of task scaffolding because the previous study with a similar mobile task using
only the low threshold (mean+1sd) found that 4-month-old infants born PT who learned the task
did not change their hip-knee coordination (Sargent et al., 2018). In the current study, the infants
born PT also did not change their coordination using the low threshold, but they generated more
selective hip-knee movement when the middle and high thresholds (mean+1.5sd and +2sd) were
used. We speculate that this may indicate that the scaffolded mobile task effectively reinforced
more selective hip-knee movement in infants born PT.
Clinical Implications and Future Directions
Scaffolding may provide a means to support the early learning and motor control of
infants born PT, at increased risk for cerebral palsy. The current study provides an example of
how to systematically scaffold a task for infants born PT to generate more selective hip-knee
movement. Specifically, we scaffolded the task to reinforce progressively higher kicks to
encourage infants born PT to extend the knee while flexing the hip, thereby generating more
selective movement. Due to their early differences in motor control (Dionisio et al., 2017;
Sargent et al., 2017), infants born PT may require systematic scaffolding of tasks to generate
age-appropriate movement.
This study also provides a scientific foundation to develop a task to support more
selective movement of infants with brain lesions, at highest risk for cerebral palsy. Scaffolding
43
for infants with brain lesions may include extra days of practice to learn the task and customized
heights and timing of thresholds based on the infants’ real-time performance of the task (Pulido
et al., 2017). For example, if the infant demonstrated more selective movement at a height but
then demonstrated less selective movement at a higher height, the threshold could be adjusted a
little lower to support the infant’s ability to generate more selective movement. Further study is
required to develop this task for infants with brain lesions, at highest risk for cerebral palsy.
Limitations
The main limitation of this study is the lack of a ‘non-scaffolded’ control group. Since we
did not have the control group, conclusions cannot be drawn about coordination change without
scaffolding. However, based on the previous results in Sargent et al. (2018), which found that 4-
month-old infants born PT did not change their movement when they participated in a similar
mobile task using only the low threshold, it seems that the middle and high thresholds used in the
current study reinforced the infants born PT to generate more selective hip-knee movement.
Further study including a control group is required to confirm the effect of scaffolding.
Conclusions
In this study, the infants born FT and the infants born PT who learned the scaffolded
mobile task demonstrated more selective hip-knee movement. This finding provides an insight
into the effect of task scaffolding on encouraging selective leg movement in infants born PT.
44
Chapter IV: Quantifying Exploratory Learning of Infants Born Full-term
Abstract
Background: Exploration is considered essential to infant learning, but few studies have
quantified infant exploratory learning during a task.
Objective: The purpose of this study was to quantify how infants explored task space with their
feet while learning to activate an infant kick-activated mobile.
Design: Repeated measures.
Methods: Data were analyzed from fifteen 4-month-old infants who participated in a 10-min
scaffolded mobile task on 2-3 consecutive days. All infants had learned that their vertical leg
movements above a systematically increased threshold height activated the mobile. Three
variables were analyzed: variance of vertical and horizontal foot positions, exploration volume,
and exploration region around and above the threshold height that activated the mobile.
Results: The infants who learned the task increased their task-specific exploration (vertical
variance of foot positions, exploration volume, and exploration region) across days. However,
the non-task-specific exploration (horizontal variance of foot positions) did not change.
Limitations: This study is limited by the lack of a non-learner control group.
Conclusions: Quantifying infants’ task exploration may provide critical insights into how
learning emerges in early infancy and, in turn, enable researchers to more systematically
describe, interpret, and support learning.
Introduction
Exploration is considered essential to infant learning (Adolph, Eppler, & Gibson, 1993;
Gibson & Pick, 2000; Sargent et al., 2014). Infants, through their movements, explore the
45
people, objects, and spaces around them to learn the relation between their movement and the
environment. Infants may then exploit their movements to produce an expected response from
the environment. While exploration is considered fundamental to learning, few studies have
quantified infant exploratory learning. Quantifying exploration may provide critical insights into
how learning emerges in early infancy and, in turn, enable researchers to more systematically
describe, interpret, and support learning.
Lobo and Galloway (2013) quantified object exploration behaviorally in infants from 2 to
5 months of age. At 2 months, the infants explored only their own bodies (e.g., touching,
looking). Once they started reaching, the infants moved beyond themselves and began to explore
objects (e.g., touching the object). After this, they began to combine their behaviors (e.g.,
touching the object while looking at it). Goldfield et al. (1993) examined the exploration of 8-
month-old infants as they bounced in a jolly jumper. They found that the infants’ bouncing
amplitudes and frequencies were sporadic and irregular at first, but over the course of 6 weeks,
the infants began to exploit specific forces and timing to generate energetically efficient and
stable, successive bounces.
Sargent et al. (2014) quantified task exploration in 3-month-old infants as they played with
an infant kick-activated mobile for two days. The infants were positioned under the mobile and it
rotated and played music when the infants moved their feet vertically above a specific height.
The infants who learned the task demonstrated increased variance of their foot heights in the
vertical, task-specific direction, whereas infants who did not learn the task demonstrated no
change in the variance of their vertical foot heights. The results support that exploration of the
task-specific space may be important for learning.
46
We designed a scaffolded infant kick-activated mobile task to encourage infants to
continually explore their environment with their feet (Chapter II). Specifically, we were
interested in further quantifying exploratory learning as infants explored the spatial properties
(e.g., threshold heights) of their task environment. Quantifying exploration during the scaffolded
mobile task may provide an understanding of how infants explore the space with their feet as
they learn the association between their foot movement and the mobile activation.
In the scaffolded mobile task, an infant lies supine under a mobile for 10 min on 2 to 3
consecutive days (Figure 4.1). When the infant lifts the legs vertically above an individually
determined threshold height, the mobile rotates and plays music for up to 3 seconds. On the first
day (Day 1), a low threshold height is used to support learning of the association between the
movement of the infants’ feet and the movement of the musical mobile. On subsequent days, the
threshold height is increased from low to middle to high. As a group, 4-month-old infants
learned the task, shown by increased duration that the infants activated the mobile on Day 2 or
Day 3 (learning day) (refer to Chapter II for details). Data also confirmed that the infants
increased the peak heights of their kicks in the vertical, task-specific direction.
This current study quantifies how infants who learned the scaffolded mobile task explored
the task space with their feet using four variables: z-variance, x-variance, exploration volume,
and exploration region. First, we analyzed the z-variance of the foot position in the vertical, task-
specific direction, consistent with Sargent et al. (2014), to assess if infants explored in the task-
specific direction across days. Second, we analyzed the x-variance of the foot position in the
horizontal, non-task-specific direction to assess if the infants did not change their exploration of
the non-task-specific direction. Third, we analyzed the entire volume of space in which the
infants moved their feet to assess if the exploration volume increased across days. Fourth, we
47
analyzed the exploration region by computing the number of the foot position data located
around and above the threshold (task-specific space), to assess if the infants increased their
exploration of the task-specific space across days.
Figure 4.1. Infant set-up. One marker (circled) on each foot activated the mobile. The dotted
lines depict the three heights (low, middle, high) of the virtual threshold.
Aim and hypotheses
The purpose of this study was to quantify 4-month-old infants’ exploration space of their
feet while learning to activate a mobile during the scaffolded mobile task. We assessed four
variables: z-variance (vertical, task-specific), x-variance (horizontal, non-task-specific),
exploration volume, and exploration region around and above the threshold. We hypothesized
X
Z
Y
48
that the task-specific exploration variables (z-variance, volume, region) would increase on Day 1
as infants explored the low threshold height. We further hypothesized that, on their learning day
(Day 2 or 3) when the low, middle, and high threshold heights were used, the slope of the task-
specific variables (z-variance, volume, region) would be significantly increased, compared to the
slope of the variables on Day 1 when only the low threshold height was used. Last, we
hypothesized that the non-task-specific exploration variable (x-variance) would not change on
either Day 1 or the learning day.
Methods
Participants
Infants born full-term (>37 weeks of gestation) were recruited from the University of
Southern California (USC) and the Los Angeles County + USC Medical Center. Infants
participated in the study between the ages of 4 months 0 day to 4 months 15 days. Infants were
excluded if they were ill or had congenital malformations, chromosomal abnormalities,
orthopedic impairments, or visual/hearing deficits. Refer to Chapter III for details of the study
participants.
Experimental protocol
The study was approved by the USC Health Sciences Institutional Review Board (#HS-
17-00119) and conducted in the Development of Infant Motor Performance Laboratory at USC.
Refer to Chapter II for details of the protocol. Procedures were explained to the parents and
written informed consent was obtained prior to data collection.
Scaffolded mobile task. Infants participated in a 10-min scaffolded mobile task for 2 to
3 consecutive days. Each day, infants were undressed and placed supine on a testing table with
their feet under an infant kick-activated mobile (Figure 4.1). The parent sat next to the infant and
49
was instructed to stay quiet and out of the infant’s sight throughout the task. If the infant became
fussy or cried, the parent tried to soothe the infant using words like “I am here” without touching
the infant.
Day 1. In the first 2-min baseline condition, the mobile did not move, but each infant
kicked spontaneously. During this baseline condition, the vertical position data of the center
marker on each foot were collected to determine the height of an individual virtual threshold. In
the following 8-min mobile condition, the musical mobile rotated only if the infant lifted either
foot vertically higher than the virtual threshold. The threshold was calculated as the average of
the mean height of the right foot marker plus one standard deviation (sd) and the mean height of
the left foot marker plus one sd (Refer to the equation in Chapter II). A minimum threshold was
set at 5 cm and a maximum threshold was set at 20 cm. Once the center marker on either foot
crossed the threshold, the mobile played music and rotated for the duration the foot stayed above
the threshold to a maximum of 3 sec. To re-activate the mobile after 3 sec, infants needed to
lower the foot below the threshold and then raise the foot to cross the threshold again. This 3-sec
activation time was set to prevent infants from holding their feet above the threshold and stop
moving their legs.
Day 2-3 (Learning Day). The 10-min task consisted of a mobile condition during which
the threshold height was systematically increased from low to middle to high. During the first 2-
min, low threshold interval, the threshold was consistent with the mobile condition on Day 1.
During the next 4-min, middle threshold interval, the threshold was increased to the mean
baseline foot height plus 1.5sd. During the last 4-min, high threshold interval, the threshold was
increased to the mean baseline height plus 2sd. Either Day 2 or 3 was each infant’s last day of
participation depending on whether they learned the task. Learning was identified if the duration
50
that the infant activated the mobile during the entire Day 2 mobile condition was greater than 1.5
times the duration that the infant would have activated the mobile during the Day 1 baseline
condition. If infants did not meet this individual learning criterion on Day 2, the infants
participated in the identical task on Day 3 and assessed again if they met the criterion.
Kinematic data. During the scaffolded mobile task, the infants’ leg movements were
captured using an Optotrak Certus Motion Capture System (Northern Digital Inc., Waterloo, ON,
Canada). Custom rigid arrays with four embedded infrared light-emitting diode markers were
attached to the infants’ pelvis, thigh, shank, and foot bilaterally using Velcro straps. A custom
array with two embedded markers was attached to the infants’ sternum position using collar tape.
The collected position data of the center marker on each foot, which activated the mobile, were
analyzed to investigate exploration (Figure 4.2.a).
Video recording. An overhead video camera (Canon Inc., VIXIA HF R700 camcorder)
recorded the infant during the scaffolded mobile task.
Anthropometric and developmental measures. After the scaffolded mobile task on
Day 2, an experimenter measured infants’ weights, heights, lengths of the lower extremity, and
widths of knee and ankle. The infants were also assessed using the motor subtest of the Bayley
Scales of Infant and Toddler Development, 3rd edition (BSITD-III) (Bayley, 2006). Refer to
Chapter II for details.
Data Reduction
A custom Matlab R2019b (Mathworks, Inc., Natick, MA) was used to interpolate missing
position data up to 20 consecutive frames using a cubic spline and to compute the following
dependent variables.
51
Figure 4.2. One infant’s right and left foot marker data captured for one minute. (A) foot marker
position data with red indicating the right foot and blue indicating the left foot; (B) the position
data wrapped by a convex hull; (C) the convex hull. The gray plane depicts a threshold location
(the individually determined, low threshold height for this infant = 15.5 cm); (D) the exploration
region area. The darker area depicts region around the threshold (z = 15.5 ± 1) and the less dark
area depicts region above the threshold (z ≥ 16.5).
A B
C D
52
Z- and x-variances. From the interpolated data, kicks were extracted. Refer to Chapter II
for details. The z- and x-variances of the position data of the center marker on each foot rigid
marker array were computed for each 2-min interval for all extracted kicks and were expressed
in centimeters. This calculation is consistent with previous research (Sargent et al., 2014).
Exploration volume. The smallest convex hull, the minimum polyhedron that wraps the
given position data of the marker on each foot, was constructed for each 1-min interval (Figure
4.2.b-c). The volume of the convex hull was computed for each minute and was expressed in
cubic centimeters.
Exploration region. The three-dimensional space was partitioned into three cross
sections; around the threshold (the vertical range of threshold ± 1 cm), and above and below this
section (Figure 4.2.d). The number of the foot marker points around and above the threshold was
counted for each 1-min interval and was divided by the total number of the foot marker points
collected during each 1-min interval. Since there was no threshold during baseline on Day 1, the
first 2 minutes were excluded from this analysis. The equation is as follows:
𝐼𝐼 𝐸𝐸𝐸𝐸 𝑙𝑙𝑜𝑜 𝑟𝑟𝑚𝑚𝑡𝑡𝑒𝑒𝑜𝑜 𝑚𝑚 𝑟𝑟 𝑚𝑚 𝑒𝑒𝑒𝑒𝑜𝑜 𝑚𝑚 (%) =
Num b er of f oot m ark er p o in t s around (± 1 cm) and a b ov e t he t h re s h old
Tot al num b er of f oot m ark er p o in t s
Statistical Analysis
We used linear mixed models with an autoregressive covariance structure to estimate
changes in each of the four dependent measures (z-variance, x-variance, exploration volume,
exploration region) over the time intervals on Day 1 and the learning day. Each model included a
subject-specific random intercept and fixed effects for time interval, day condition, and its
interaction (time interval*day condition). This interaction term was added to test if the slopes of
53
the exploration variables were different between baseline (Day 1; minutes 1-2), Day 1
acquisition (minutes 3-10), and the learning day (minutes 1-10). For the exploration region,
baseline was not included in the analysis as mentioned above. If there was no significant
interaction, the model was re-fit without the interaction term.
We additionally performed cross-correlation analysis to identify a possible interaction
between the z-variance and x-variance. For the exploration volume, we tested the interaction
effect between time interval and infants’ leg length to test whether the association between the
exploration volume and time interval was modified by the infants’ leg length. All statistical
analyses were performed using SAS v.9.0 (SAS Institute Inc., Cary, NC) and the alpha level was
set at .05.
Results
Participants
Data on 15 infants born full-term (5 girls, 10 boys) who learned the scaffolded mobile
task were analyzed. The learning day was the second day for 13 infants and the third day for 2
infants. The infants were born at a mean of 39 (±0.9) weeks gestation with a mean birth weight
of 3.5 (±0.3) kg and tested at 123.7 (±4.8) days of age with a mean ponderal index of 2.45
(±0.28) kg/cm
3
and BSITD-III score of 77.9 (±17.3) percentile. Refer to Chapter III for details.
Final statistical models
Since there was no statistically significant interaction between time interval and day
condition for all dependent variables, the interaction term (time interval*day condition) was
removed from each linear mixed model and the main effects of time interval were analyzed.
54
Variance
There was a low cross-correlation between the z- and x-variances (Figure 4.3). Thus, we
considered these two variables as independent measures. From Day 1 to the learning day, the
slope of the z-variance was 4.4 mm/min (95% confidence interval, CI: 1.5, 7.3, p = .003) (Figure
4.4; Visualization of individual infant time series data is in Figure 4.5). The slope of the x-
variance was 1.2 mm/min, but not statistically significant (p = .08) (Figure 4.6; Visualization of
individual infant time series data is in Figure 4.7).
Figure 4.3. Cross-correlation analysis result between the z- and x-variances. Shows low
correlation between the z- and x-variances.
55
Figure 4.4. The slope and 95% confidence interval of the z-variance in millimeters (mm) across
days. Individual 2-min estimates and standard errors are indicated (B: baseline, L: low threshold,
M: middle threshold, H: high threshold).
Figure 4.5. Individual line graphs of the z-variance (mm) in 15 infants over 1-min intervals
across days.
56
Figure 4.6. The slope and 95% confidence interval of the x-variance in millimeters (mm) across
days. Individual 2-min estimates and standard errors are indicated.
Figure 4.7. Individual line graphs of the x-variance (mm) of 15 infants over 1-min intervals
across days.
57
Exploration volume
There was no significant interaction of infants’ leg length on the association between the
exploration volume and time interval, thus, we used the raw volume in cubic centimeters for the
analysis. From Day 1 to the learning day, the slope of the exploration volume was 253.86
cm
3
/min (95% CI: 190.55, 317.16, p < .0001) (Figure 4.8; Visualization of individual infant time
series data is in Figure 4.9).
Figure 4.8. The slope and 95% confidence interval of the exploration volume in cubic
centimeters (cm
3
) across days. Individual 1-min estimates and standard errors are indicated (B:
baseline, L: low threshold, M: middle threshold, H: high threshold).
58
Figure 4.9. Individual line graphs of the exploration volume (cm
3
) in 15 infants over 1-min
intervals across days.
Exploration region
From Day 1 after baseline to the learning day, the slope of the exploration region was
0.75%/min (95% CI: 0.33, 1.17, p < .001) (Figure 4.10; Visualization of individual infant time
series data is in Figure 4.11).
59
Figure 4.10. The slope and 95% confidence interval of the exploration region around and above
the threshold in percentage (%) across days. Individual 1-min estimates and standard errors are
indicated.
Figure 4.11. Individual line graphs of the exploration region around and above the threshold (%)
in 15 infants over 1-min intervals across days.
60
Discussion
This study investigated the exploration of task space of 4-month-old infants who learned
the scaffolded mobile task. We hypothesized that the task-specific exploration variables (z-
variance, exploration volume, exploration region) would increase on Day 1 and the slopes of the
variables would increase on the learning day as infants explore the higher thresholds. Consistent
with our hypothesis, the task-specific exploration variables increased within both Day 1 and the
learning day. However, the slopes of the variables did not increase on the learning day. As
hypothesized, the slope of the non-task-specific variable (x-variance) did not change within or
across days.
Our finding of an increase in the vertical, task-specific z-variance was consistent with the
results from Sargent et al. (2014) who found that the infants who learned the task increased their
z-variance, whereas infants who did not learn the task had no change in z-variance. Our current
study adds that infants who learned the task did not change their non-task-specific horizontal
exploration, x-variance. The data from these two studies support that infants while learning
increase their exploration of the task-specific space but do not change their exploration of the
non-task-specific space. Further research is necessary to investigate exploration in infants who
did not learn the task to associate learning and early changes in the z- and x-variances.
We found a linear increase in the 3-dimensional volume of space explored by the infant.
Although no study has used volumetric measures to quantify infant exploratory learning,
volumetric measures have been used in other research to track moving objects (Demšar &
Virrantaus, 2010), detect tumor size (Nguyen, Jain, & Sabata, 2011), characterize the motion of
professional dancers (Ajili, Mallem, & Didier, 2019) and describe infant orofacial movements
(Green & Wilson, 2006). In our study, an incidental finding from analyzing volume graphs of
61
individual infants was that, for several infants, the shape of volume changed over the task from a
pyramid to an inverse pyramidal shape (Refer to example in Figure 4.12). We interpret this as the
infants moved their exploration from primarily exploring the supporting surface to primarily
exploring above the surface and around the threshold. Further research is needed to investigate
not only the size of the exploration volume, but also the shape of the volume to describe and
interpret the volume in relation to task space.
Figure 4.12. The exploration volume from one infant’s data (left: a pyramidal shape during a 1-
min interval on Day 1, right: an inverse pyramidal shape during a 1-min interval on the learning
day).
To investigate if infants spent more time in the reinforced, task-specific space, we
calculated the exploration region around and above the threshold. We found that there was a
linear increase in the task-specific region although the rate of increase was low (<1%). Infants
kept their feet around and above the threshold (the task-specific space) approximately 23% to
62
46% of the time over the course of two days. Considering that the threshold height changed on
the learning day, the small increase in the exploration region may indicate that the infants were
able to immediately explore the new threshold heights, thereby maintaining their exploration
region around and above the threshold. This ability to adapt to change has been supported in the
literature. For example, Thelen et al. (1987) increased the weight of infants’ legs while kicking.
When a weight was added to one leg, supine 6-week-olds began kicking their unweighted legs
more, thereby maintaining the same overall kicking frequency as when there was no weight on
either leg (Esther Thelen, Skala, & Kelso, 1987). Similarly, Rachwani et al. (2017) changed the
slope of a sitting surface. Experienced 6- to 8-month-olds quickly adjusted their sitting posture to
cope with changes to the slope of the surface (Rachwani, Soska, & Adolph, 2017).
The main limitation of this study is the lack of a ‘non-learner’ control group. We did not
compare the infants who learned the task with a group of infants who did not learn the task due
to the small sample size of non-learners (n=3). The addition of a non-learner group would allow
an analysis of how exploration is associated with learning.
Conclusions
Our findings add to the literature supporting infants’ spatial exploration while learning a
task. By quantifying infants’ exploration during the scaffolded mobile task, we discovered that
infants who learned the task explored the space systematically over the course of two days. This
study has implications for researchers and clinicians describing, interpreting, and supporting
infant exploratory learning.
63
Chapter V: Summary and Conclusions
This dissertation investigated how 4-month-old infants born full-term (FT) and infants
born preterm (PT) learn, move, and explore during the scaffolded task that systematically
reinforced selective leg movements.
Chapter II: Infants Born Full-term and Preterm Learn a Scaffolded Mobile Task.
Data were collected from 18 infants born FT and 18 infants born PT at 4 months of age,
corrected for prematurity. First, this study found that both the FT and PT groups demonstrated
learning of the scaffolded mobile task that required progressively higher kicks of the infants. As
expected from previous findings, the PT group showed a slower learning rate. The FT group
demonstrated learning on the second day of participation, whereas the PT group demonstrated
learning after an extra day of participation was given for some infants. This finding shows that
the scaffolded mobile task was still within the PT infants’ repertoire although it was not within
their preferred movement boundary. Second, this study found that both the FT and PT groups
significantly increased their kick heights during the high threshold condition. Although
conclusions cannot be drawn without having a non-scaffolded control group, the finding of this
study informs the capability of infants born PT, who have known differences in strength and
motor control, of producing challenging, anti-gravity movements.
This study contributes to understanding the importance of task scaffolding and infants’
capability of adapting to challenging task demands. The findings emphasize the importance of
scaffolding task environments for educational and clinical practice for infants born PT at risk for
learning and motor impairments.
64
Chapter III (Infants Born Preterm and Infants Born Full-term Generate More Selective
Leg Joint Movement during the Scaffolded Mobile Task)
In this study, data were analyzed on 15 infants born FT and 13 infants born PT who
learned the scaffolded mobile task by the second or third day of participation. This study found
that both the FT and PT learner groups demonstrated more selective hip-knee joint movements
when interacting with the mobile, compared to their spontaneous kicking. The FT group showed
more selective hip-knee movement from the first day of participation and throughout the low,
middle, and high threshold intervals. The PT group, on the other hand, did not show more
selective hip-knee movements during the low threshold intervals, but only during the middle and
high threshold intervals. It is speculated that there was a positive effect of task scaffolding using
the higher heights for the PT group.
This study shows an example of how to systematically scaffold a task for infants born PT
to generate more selective hip-knee movement. Specifically, the scaffolded task reinforced
progressively higher kicks to encourage infants born PT to extend the knee while flexing the hip,
thereby generating more selective hip-knee movement. By doing so, this study has implications
for clinicians in terms of how to develop a task and support selective movements of infants with
brain lesions, at highest risk for cerebral palsy.
Chapter IV (Quantifying Exploratory Learning of Infants Born Full-term)
This study investigated how infants born FT explore task space while learning the
scaffolded mobile task. This is the first study to quantify spatial exploration of infants during a
task. Their exploratory foot movements were quantified using variance, volume, and region. The
FT learner group increased task-specific exploration (the vertical variance, exploration volume,
65
exploration region around and above the threshold) during the task, whereas non-task-specific
exploration (the horizontal variance) did not change.
This research is a first step toward quantifying infants’ task exploration. This study has
implications for researchers and clinicians describing, interpreting, and supporting infant
exploratory learning.
Overall, these dissertation studies contribute to the body of knowledge about how infants
born FT and PT learn a scaffolded task, move their bodies in response to reinforcement, and
explore the task space. Our data inform the capability of infants born PT to generate more
selective hip-knee movements when participating in a scaffolded task. The findings of this
dissertation may provide the basis for researchers and clinicians to structure tasks that can
provide necessary support for infants born PT at increased risk of developing cerebral palsy.
66
References
Adolph, K. E., Eppler, M. A., & Gibson, E. J. (1993). Crawling versus walking infants'
perception of affordances for locomotion over sloping surfaces. Child Development,
64(4), 1158-1174.
Ahmad, I., Nemet, D., Eliakim, A., Koeppel, R., Grochow, D., Coussens, M., . . . Waffarn, F.
(2010). Body composition and its components in preterm and term newborns: A cross-
sectional, multimodal investigation. American Journal of Human Biology, 22(1), 69-75.
https://doi.org/10.1002/ajhb.20955
Angulo-Kinzler, R. M. (2001). Exploration and selection of intralimb coordination patterns in 3-
month-old infants. Journal of Motor Behavior, 33(4), 363.
https://doi.org/10.1080/00222890109601920
Bayley, N. (2005). Manual for the Bayley scales of infant and toddler development (3
rd
edition).
San Antonio: Psychological Corporation.
Bayley, N. (2006). Bayley Scales of Infant and Toddler Development: Technical manual (3
rd
edition). San Antonio, TX: Harcourt Assessment.
Chen, Y-P., Fetters, L., Holt, K. G., & Saltzman, E. (2002). Making the mobile move:
Constraining task and environment. Infant Behavior & Development, 25(2), 195-220.
https://doi.org/10.1016/S0163-6383(02)00121-2
Dassios, T., Kaltsogianni, O., Krokidis, M., Hickey, A., & Greenough, A. (2018). Deltoid muscle
morphometry as an index of impaired skeletal muscularity in neonatal intensive care.
67
European Journal of Pediatrics, 177(4), 507-512. https://doi.org/10.1007/s00431-018-
3090-5
Demšar, U., & Virrantaus, K. (2010). Space–time density of trajectories: Exploring spatio-
temporal patterns in movement data. International Journal of Geographical Information
Science, 24(10), 1527-1542.
Dionisio, J., Santos, G. L. D., & Tudella, E. (2017). Influence of additional ankle weights on
kinematic variables of late preterm infants aged 3-4 months. Journal of Motor Behavior,
49(3), 306-311. https://doi.org/10.1080/00222895.2016.1204264
Farmer, S. E. (2003). Key factors in the development of lower limb co-ordination: Implications
for the acquisition of walking in children with cerebral palsy. Disability and
Rehabilitation, 25(14), 807-816.
Farmer, S.E., Pearce, G., & Stewart, C. (2008). Developing a technique to measure intra-limb
coordination in gait: Applicable to children with cerebral palsy. Gait & Posture, 28(2),
217-221. https://doi.org/10.1016/j.gaitpost.2007.12.005
Faul, F., Erdfelder, E., Lang, A.-G., & Buchner, A. (2007). G* Power 3: A flexible statistical
power analysis program for the social, behavioral, and biomedical sciences. Behavior
Research Methods, 39(2), 175-191.
Fetters, L., Chen, Y.-p., Jonsdottir, J., & Tronick, E. Z. (2004). Kicking coordination captures
differences between full-term and premature infants with white matter disorder. Human
Movement Science, 22(6), 729-748. http://dx.doi.org/10.1016/j.humov.2004.02.001
68
Fetters, L., Sapir, I., Chen, Y. P., Kubo, M., & Tronick, E. (2010). Spontaneous kicking in full-
term and preterm infants with and without white matter disorder. Developmental
Psychobiology, 52(6), 524-536. http://dx.doi.org/10.1002/dev.20455
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. http://dx.doi.org/10.1016/j.gaitpost.2008.07.007
Fowler, E. G., Staudt, L. A., & Greenberg, M. B. (2010). Lower-extremity selective voluntary
motor control in patients with spastic cerebral palsy: Increased distal motor impairment.
Developmental Medicine & Child Neurology, 52(3), 264-269.
http://dx.doi.org/10.1111/j.1469-8749.2009.03586.x
Gekoski, M. J., Fagen, J. W., & Pearlman, M. A. (1984). Early learning and memory in the
preterm infant. Infant Behavior and Development, 7(3), 267-276.
http://dx.doi.org/10.1016/S0163-6383(84)80042-9
Gibson, E. J., & Pick, A. D. (2000). An ecological approach to perceptual learning and
development: Oxford University Press, USA.
Goldfield, E. C., Kay, B. A., & Warren, W. H., Jr. (1993). Infant bouncing: The assembly and
tuning of action systems. Child Development, 64(4), 1128-1142.
Gopagondanahalli, K. R., Li, J., Fahey, M. C., Hunt, R. W., Jenkin, G., Miller, S. L., &
Malhotra, A. (2016). Preterm hypoxic–ischemic encephalopathy. Frontiers in Pediatrics,
4(114). http://dx.doi.org/10.3389/fped.2016.00114
69
Green, J. R., & Wilson, E. M. (2006). Spontaneous facial motility in infancy: A 3D kinematic
analysis. Developmental Psychobiology, 48(1), 16-28.
http://dx.doi.org/10.1002/dev.20112
Hagberg, B., Hagberg, G., Beckung, E., & Uvebrant, P. (2001). Changing panorama of cerebral
palsy in Sweden. VIII. Prevalence and origin in the birth year period 1991-94. Acta
Paediatrica, 90(3), 271-277.
Haley, D. W., Grunau, R. E., Oberlander, T. F., & Weinberg, J. (2008). Contingency learning
and reactivity in preterm and full-term infants at 3 months. Infancy, 13(6), 570-595.
http://dx.doi.org/10.1080/15250000802458682
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.
http://dx.doi.org/10.1016/j.psyneuen.2005.06.007
Heathcock, J. C., Bhat, A. N., Lobo, M. A., & Galloway, J. C. (2004). The performance of
infants born preterm and full-term in the mobile paradigm: Learning and memory.
Physical Therapy, 84(9), 808-821. http://dx.doi.org/10.1093/ptj/84.9.808
Heathcock, J. C., Bhat, A. N., Lobo, M. A., & 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(1), 8-18.
http://dx.doi.org/10.1093/ptj/85.1.8
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:
70
A meta-analytic review. Developmental Medicine & Child Neurology, 50(5), 334-40.
http://dx.doi.org/10.1111/j.1469-8749.2008.02047.x
Jensen, J. L., Schneider, K., Ulrich, B. D., Zernicke, R. F., & Thelen, E. (1994). Adaptive
dynamics of the leg movement patterns of human infants: I. The effects of posture on
spontaneous kicking. Journal of Motor Behavior, 26(4), 303-312.
http://dx.doi.org/10.1080/00222895.1994.9941686
Lima-Alvarez, C. D., Tudella, E., van der Kamp, J., & Savelsbergh, G. J. (2014). Early
development of head movements between birth and 4 months of age: A longitudinal
study. Journal of Motor Behavior, 46(6), 415-422.
http://dx.doi.org/10.1080/00222895.2014.929562
Lobo, M. A., & Galloway, J. C. (2013a). Assessment and stability of early learning abilities in
preterm and full-term infants across the first two years of life. Research in Developmental
Disabilities, 34(5), 1721-1730. http://dx.doi.org/10.1016/j.ridd.2013.02.010
Lobo, M. A., & Galloway, J. C. (2013b). The onset of reaching significantly impacts how infants
explore both objects and their bodies. Infant Behavior and Development, 36(1), 14-24.
http://dx.doi.org/10.1016/j.infbeh.2012.09.003
Pascal, A., Govaert, P., Oostra, A., Naulaers, G., Ortibus, E., & Van den Broeck, C. (2018).
Neurodevelopmental outcome in very preterm and very-low-birthweight infants born
over the past decade: A meta-analytic review. Developmental Medicine & Child
Neurology, 60(4), 342-355. http://dx.doi.org/10.1111/dmcn.13675
71
Pulido, J. C., González, J. C., Suárez-Mejías, C., Bandera, A., Bustos, P., & Fernández, F.
(2017). Evaluating the child–robot interaction of the NAOTherapist Platform in pediatric
rehabilitation. International Journal of Social Robotics, 9(3), 343-358.
http://dx.doi.org/10.1007/s12369-017-0402-2
Rachwani, J., Soska, K. C., & Adolph, K. E. (2017). Behavioral flexibility in learning to sit.
Developmental Psychobiology, 59(8), 937-948. http://dx.doi.org/10.1002/dev.21571
Rose, S. A., Feldman, J. F., Jankowski, J. J., & Van Rossem, R. (2005). Pathways from
prematurity and infant abilities to later cognition. Child Development, 76(6), 1172-1184.
http://dx.doi.org/10.1111/j.1467-8624.2005.00843.x
Rosenbaum, P., Paneth, N., Leviton, A., Goldstein, M., Bax, M., Damiano, D., . . . Jacobsson, B.
(2007). A report: The definition and classification of cerebral palsy April 2006.
Developmental Medicine & Child Neurology Supplement, 109, 8-14.
Rovee-Collier, C. (1997). Dissociations in infant memory: Rethinking the development of
implicit and explicit memory. Psychological Review, 104(3), 467-498.
http://dx.doi.org/10.1037/0033-295X.104.3.467
Rovee, C. K., & Rovee, D. T. (1969). Conjugate reinforcement of infant exploratory behavior.
Journal of Experimental Child Psychology, 8(1), 33-39. http://dx.doi.org/10.1016/0022-
0965(69)90025-3
Sargent, B., Kubo, M., & Fetters, L. (2018). Infant discovery learning and lower extremity
coordination: Influence of prematurity. Physical & Occupational Therapy in Pediatrics,
1-16. http://dx.doi.org/10.1080/01942638.2017.1357065
72
Sargent, B., Reimann, H., Kubo, M., & Fetters, L. (2017). Infant intralimb coordination and
torque production: Influence of prematurity. Infant Behavior & Development, 49, 129-
140. http://dx.doi.org/10.1016/j.infbeh.2017.08.009
Sargent, B., Schweighofer, N., Kubo, M., & Fetters, L. (2014). Infant exploratory learning:
Influence on leg joint coordination. PLoS One, 9(3), e91500.
http://dx.doi.org/10.1371/journal.pone.0091500
Stone, C. A. (1998). The metaphor of scaffolding: Its utility for the field of learning disabilities.
Journal of Learning Disabilities, 31(4), 344-364.
http://dx.doi.org/10.1177/002221949803100404
Thelen, E. (1985). Developmental origins of motor coordination: Leg movements in human
infants. Developmental Psychobiology, 18(1), 1-22.
http://dx.doi.org/10.1002/dev.420180102
Thelen, E., Skala, K. D., & Kelso, J. S. (1987). The dynamic nature of early coordination:
Evidence from bilateral leg movements in young infants. Developmental Psychobiology,
23(2), 179-186. http://dx.doi.org/10.1037/0012-1649.23.2.179
Thelen, E., & Ulrich, B. D. (1991). Hidden skills: A dynamic systems analysis of treadmill
stepping during the first year. Monographs of the Society for Research in Child
Development, 56(1), 1-104. http://dx.doi.org/10.2307/1166099
Tiernan, C. W., & Angulo-Barroso, R. M. (2008). Constrained motor-perceptual task in infancy:
Effects of sensory modality. Journal of Motor Behavior, 40(2), 133-142.
http://dx.doi.org/10.3200/jmbr.40.2.133-142
73
Van Hus, J. W., Potharst, E. S., Jeukens-Visser, M., Kok, J. H., & Van Wassenaer-Leemhuis, A.
G. (2014). Motor impairment in very preterm-born children: Links with other
developmental deficits at 5 years of age. Developmental Medicine & Child Neurology,
56(6), 587-594. http://dx.doi.org/10.1111/dmcn.12295
Watanabe, H., & Taga, G. (2009). Flexibility in infant actions during arm- and leg-based
learning in a mobile paradigm. Infant Behavior and Development, 32(1), 79-90.
http://dx.doi.org/10.1016/j.infbeh.2008.10.003
Williams, J., Lee, K. J., & Anderson, P. J. (2009). 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. http://dx.doi.org/10.1111/j.1469-
8749.2009.03544.x
Wood, D., Bruner, J. S., & Ross, G. (1976). The role of tutoring in problem solving. Journal of
Child Psychology and Psychiatry, 17(2), 89-100. http://dx.doi.org/10.1111/j.1469-
7610.1976.tb00381.x
Abstract (if available)
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
Learning acceleration control of leg movements during infancy
Asset Metadata
Creator
Kim, Jeong Ah (Jeongah)
(author)
Core Title
Understanding how infants born full-term and preterm learn, move, and explore
School
School of Dentistry
Degree
Doctor of Philosophy
Degree Program
Biokinesiology
Publication Date
04/14/2020
Defense Date
03/06/2020
Publisher
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
exploration,Infants,Learning,OAI-PMH Harvest,prematurity,scaffolding,selective movement
Language
English
Contributor
Electronically uploaded by the author
(provenance)
Advisor
Sargent, Barbara (
committee chair
), Bradley, Nina (
committee member
), Fetters, Linda (
committee member
), Eckel, Sandrah Proctor (
committee member
), Fowler, Eileen (
committee member
)
Creator Email
jeongah.jane.kim@gmail.com,jeongah@usc.edu
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-c89-281010
Unique identifier
UC11673960
Identifier
etd-KimJeongAh-8260.pdf (filename),usctheses-c89-281010 (legacy record id)
Legacy Identifier
etd-KimJeongAh-8260.pdf
Dmrecord
281010
Document Type
Dissertation
Rights
Kim, Jeong Ah (Jeongah)
Type
texts
Source
University of Southern California
(contributing entity),
University of Southern California Dissertations and Theses
(collection)
Access Conditions
The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...
Repository Name
University of Southern California Digital Library
Repository Location
USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA
Tags
exploration
prematurity
scaffolding
selective movement