Motor control complexity estimation using gait measures in individuals post-stroke

Motor control impairments post-stroke significantly limit walking ability, with residual gait impairments often persisting despite rehabilitation efforts. Integrating motor control-based assessments in post-stroke gait evaluations is essential for monitoring the underlying causes of the limited functionality and enhancing recovery outcomes. This study aimed to develop motor control-based post-stroke gait evaluation techniques using common gait measures to inform and guide rehabilitation decisions. Subject-specific, forward-dynamic simulations of eight individuals with post-stroke gait undergoing a 12-weeks FastFES gait retraining program were created pre- and post-treatment to determine muscle activation patterns for muscle module analysis. The motor control complexity index was defined by the variance not accounted for by one module (VNAF₁) as a summary measure of the analysis. Twenty-eight gait measures were investigated, and the relevant measures were selected using feature selection methods and fed into a multiple linear regression model to estimate the motor control complexity index. The motor control complexity of 182 gait cycles were quantified (0.164 ± 0.047). No strong relationship (quantified using Pearson correlation coefficients) was found between gait measures and the motor control complexity index. However, a combination of four gait measures from the paretic side (maximum hip abduction and knee flexion angle during swing, knee range of motion, and maximum paretic ankle power) explained most of the variation (R² = 0.66) in motor control complexity.