Abstract:
|
Gait asymmetry is often observed in populations with varying degrees of altered neuromuscular control. Traditionally, gait asymmetry is evaluated using mean peak vertical ground reaction forces (vGRF), however, mean peak vGRF does not have the level of sensitivity needed to detect subtle changes in gait movement patterns. Biomechanically, it appears that time series generated from peak vGRF waveforms may be suitable to assess changes in abnormal gait patterns. Previously, autoregressive (AR) modeling has been used successfully to delineate between healthy and pathological gait during running; but has been little explored in walking. To expand upon this work, we will investigate the ability of AR modeling to both quantitatively and graphically delineate between healthy and abnormal gait patterns in healthy controls. We will use an asymmetric walking protocol to simulate abnormal gait patterns and fit the AR model to both the normal and asymmetric walking patterns. To address the scarcity of our limited dataset, classical bootstrapping techniques will be employed to increase the number of peak vGRF time series and improve our ability to construct viable and robust AR models.
|