Preferred Name
Ian A. M. Barry
Creative Commons License
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License
ORCID
https://orcid.org/0009-0006-8408-1558
Date of Graduation
2026
Semester of Graduation
Spring
Degree Name
Master of Science (MS)
Department
Department of Biology
First Advisor
Lauren Sarringhaus
Second Advisor
Steven Cresawn
Third Advisor
Emily Naylor
Abstract
Chimpanzees (Pan troglodytes) are an endangered species of great ape whose survival is contingent on successful navigation of complex arboreal and terrestrial habitats. Our current understanding of chimpanzee locomotor performance is limited to captive studies. Captive environments do not necessarily reflect the entropy of wild environments; thus, captive studies might not illustrate how chimpanzees move in situ. Although large video databases exist that can provide insight into wild chimpanzee movement, prohibitively high time and labor investment are required to manually extract information from these databases. This study explores the application of a machine learning model to videos of wild and captive chimpanzees and demonstrates the successful creation and deployment of a model to track chimpanzee joint centers of rotation in wild and captive individuals. This study further demonstrates that viable joint angles at the hip, knee, and shoulder can be extracted from this model’s output and preliminarily tests for differences in locomotor performance between adult and adolescent males at Ngogo, Kibale National Park, Uganda. No differences were found between adult and adolescent locomotor performance in a sample from this population of wild chimpanzees, suggesting that it is acceptable to group these two age classes together in studies of chimpanzee locomotion. Finally, this study validates a method using linear equations in conjunction with joint angles derived from model output to estimate absolute speed of wild chimpanzees in the absence of a scale reference. Ultimately, the model provided by this study can be augmented to improve its performance on chimpanzee videos recorded in diverse habitats and diminish barriers to parsing kinematic information from large datasets.
