Cycle variance: proposing a novel ensemble-based approach to assess the gait rhythmicity on the MIT-Skywalker

Abstract

A novel ensemble-based approach of cycle variance is proposed for the assessment of gait rhythmicity of people affected by stroke or cerebral palsy. Three impaired participants (two with cerebral palsy and one with a stroke) completed a one-month rehabilitation training program utilizing a novel piece of robotic hardware, the MIT-Skywalker. The MIT-Skywalker trained patients in 3 modes (rhythmic, discrete, and balance mode). Before training, the cycle variance of the three participants across 4 different variables was observed to be significantly higher than that of a speed-matched healthy person. Of the 12 variables measured across the three participants, 11 showed statistically significant decreases after the one-month long trial. For the single stroke participant, cycle variance was measured before and after each trial, showing statistically significant decreases in cycle variance after rhythmic training but not during discrete or balance sessions. We speculate that cycle variance might serve as a diagnostic tool to assess gait rhythmicity.