515 – Statistical Modeling for Sports Science and Applications
Identifying Individuals with Anterior Cruciate Ligament Injury Through Spectral and Linear Discriminant Analyzes
Kristin D. Morgan
University of Connecticut
Carolyn B. Morgan
Meck Limited, LLC
Heather Bush
University of Kentucky
Brian Noehren
University of Kentucky
In human movement biomechanics, healthy and injured populations are often differentiated based on discrete measures, such as peak knee flexion angles. Discrete measures do not capture the changing dynamics that can characterize altered joint motion in injured populations. To capture these changing dynamics, spectral and damping analyses will be employed. The results of these analyses will be used to develop a regression model to classify individuals into healthy and injured populations. This study will use fast Fourier transform and control theory techniques to obtain frequency, phase, damping and stability metrics to characterize gait in healthy individuals and post anterior cruciate ligament (ACL) reconstruction individuals. Then linear discriminant analysis will be used to identify the critical variables that best classify the healthy and injured populations. It will also yield a model that describes the relationship between these variables that are very critical to identifying individuals with adverse gait biomechanics. This analysis will allow us to potentially predict individuals with future injury risk and design more targeted ACL prevention and rehabilitation programs.