Figure 1

Schematic representation of the five experimental sessions. A) Normalization procedure before the experiment. First, the dynamic balance control capabilities of each subject were quantified to determine the amplitude and velocity of the moving box (Calibration). Then, subject rested for 5-min. To avoid possible learning effect during the main experiment, subjects performed 40 trials followed by a 5-min rest period (Practice session). Then, baseline reaction times were collected (Baseline RT). Finally, subjects walked on the treadmill to determine maximum treadmill velocity of each subject (Familiarization). B) Main experiment. The first two blocks are without fatigue (No fatigue 1 and No fatigue 2). There was a 2 minutes rest period between both blocks of no fatigue. The next three blocks composed the fatigue condition (Fatigue 1, Fatigue 2 and Fatigue 3). The gray areas before each of the three block of fatigue correspond to the fast walking periods. After the last block of no fatigue (No fatigue 2), subjects started the first fast walking session (gray area). When they could not keep the pace with the treadmill, they started tracking the moving box for 10 30-s trials. They repeated this procedure twice. C) Within trial. Before each trial, a two seconds period of data acquisition served to align the centre of the moving box with the average centre of pressure calculated during this period. This 2-s period was not included in the centre of pressure data analysis. Each trial lasted 30 seconds.