Online Program

Return to main conference page
Friday, May 18
Bioinformatics/Biomedical
Fri, May 18, 10:00 AM - 10:45 AM
Regency Ballroom B
 

Data-Driven Statistical Methods for Detecting Gait Instability Using Physiological Signal Metrics (304643)

Carolyn Morgan, MECK Limited LLC 
*Kristin Morgan, University of Connecticut 
Brian Noehren, University of Kentucky 

Keywords: Gait, Stability, Logistic Regression, Discriminant Analysis

The detection of altered gait has proven useful in the early identification of movement abnormalities. While most studies investigate abnormal gait in injured populations, the detection of gait instability in healthy populations has been limited. The early detection of gait instability is critical for predicting and preventing injury, especially in sports. Current research in engineering biomechanics is directed at investigating physiological signals; such as, ground reaction force and joint kinematics to detect instability during walking. For this study, our walking protocol will involve inducing a destabilizing gait pattern in healthy individuals and employing techniques; such as, power spectral density, Poincare plots, sample entropy and wavelet analyses. Data-driven statistical methods such as logistic regression, discriminant analysis and multivariate linear regression will be used to identify altered movement strategies and classify stable and unstable gait. A preliminary investigation conducted on a study population will be discussed. The results of this study will be used to develop new methodologies for identifying non-invasive gait biometrics for early injury detection.