Figure 1

Permutation entropy method. Permutation entropy (PE) was calculated for both unaveraged hindlimb trajectories and cycle averaged trajectories for each joint angle. Permutation entropy quantifies the probability that a signal will remain similar from one segment to the next. Changes in the direction of the signal (positive to negative slope, for instance), indicate increases in complexity, while a steadily decreasing slope would indicate less complexity. A unitless number 0–1 describes the results, with higher values having higher complexity. A: The angle trajectories are segmented into 3 point motifs. B: The motifs are then classified in to one of 6 possible categories (1–6). PE is then calculated using the standard Shannon uncertainty formula.