Date of Award

Summer 8-2020

Document Type


Degree Name

Doctor of Philosophy (PhD)


Health Services Research


Kinesiology and Rehabilitation

Committee Director

Daniel M. Russell

Committee Member

Steven Morrison

Committee Member

Eric J. Schussler

Committee Member

Mariana Szklo-Coxe


The purpose of this work was to study the relationships between age, measures of gait variability, and locomotor adaptability. Measures of gait variability are used to identify maladapted locomotor behavior, motor disease, and risk of falls. The first aim was to determine the relationships between age and measures of gait variability. Thirty-four participants (23-71 years old) walked on a treadmill for 6 minutes at their preferred speed. Variability of stride times and lengths was computed via linear measures (standard deviation and coefficient of variation) and nonlinear measures (sample entropy and detrended fluctuation analysis). Movement trajectory variability of the dominant knee angle, and vertical and medial-lateral positions of the pelvis were quantified using nonlinear measures (correlation dimension and local dynamic stability). The results showed little association of age and variability measures. Additional analyses revealed that preferred gait speed was a better predictor of gait variability measures, suggesting that variations in gait variability are driven more by preferred gait speed than age. The second aim of this dissertation was to investigate the relationships between measures of gait variability. While the relationships between measures of gait variability have received little investigation, many have been suggested to quantify the same underlying component of locomotion, the ability of an individual to adapt. A principal component analysis was performed to examine if measures of variability were related to one or more underlying constructs of gait variability. Four independent constructs of gait variability were identified, indicating there is no single construct underlying gait variability and different variability measures can be associated with the same constructs. The final aim was to determine if measures of variability quantify the ability of an individual to adapt to a novel split-belt gait adaptation task, where two treadmill belts were set at different speeds. The findings showed no significant association between measures of gait variability from the preferred walking trial and adaptability performance. To conclude, gait variability is more speed-related than age-related, measures of gait variability quantify at least four separate components of gait, and gait variability measures are relatively unrelated to the adaptability performance of an individual.