FIGURE 1.

Exemplary plots of stride length values over the time course of the trial. Turns can be identified as gaps (dark gray overlay) in the timeseries as they are segmented and excluded by the implemented algorithm. Additional strides that were removed around turns in the alternative turn segmentation approach are marked with a lighter gray overlay. (A) Healthy subject with a turn that was not segmented by the software algorithm. Note the drastically shorter steps while turning, resulting in CoV beyond physiological range when calculated from all strides. (B) Patient with cerebellar ataxia and proper detection of all four performed turns. Note the pronounced fluctuation of stride length values from stride to stride, resulting in a CoV beyond physiological range. (C) Patient with idiopathic Parkinson’s disease and proper detection of all four performed turns. Note the decreased average stride length and the shorter strides right after the turns which may correspond to impaired step initiation.