Fig 1.
Definition of step onsets and resulting temporal performance measures.
a). Schematic representation of step onsets (red circles) extracted from the vertical heel marker trajectory (blue line) during the stepping task. Onsets are defined by the time point when the vertical axis coordinate of the captured heel marker rose above a predefined threshold height above the floor surface. Onsets were extracted from the avatar and participants in the same way. b). Example schematic showing temporal performance measures based on the step onset times of the avatar and participant. Asynchrony (A) is measured as the time difference between the participant’s and corresponding avatar step onsets. Inter-step intervals for the avatar (IAv) and participant (IP) are calculated as the time between a participant’s/avatar’s own step onsets. Step T, refers to the avatar step onset that was perturbed (shown here as a shortened interval by shifting the onset earlier). In this example, the participant onsets prior to the perturbation are before the avatars (denoted by a negative asynchrony).
Fig 2.
Mean Inter-step intervals (ISI) of participant steps (filled bar) compared to the target Avatar ISI (lighter shaded bars) pre-perturbation for all conditions.
Error bars show standard error of the mean (SEM). * represents significance at p<.05, ** is p<.01.
Fig 3.
Asynchrony between participant step onsets and corresponding Avatar steps pre-perturbation.
a). Mean Asynchrony between participant step onsets and corresponding Avatar steps, for all conditions. This asynchrony reflects baseline performance. Negative asynchronies indicate the participants are, on average, stepping ahead of the Avatar cue. Error bars show SEM. b). Mean standard deviation of asynchrony for all conditions. Error bars show SEM. * represents significance at p<.05, ** is p<.01.
Fig 4.
Mean relative asynchronies before and after the perturbation (vertical grey bar) at time T for Visual-Only cues.
Dotted horizontal line shows the zero relative asynchrony measure, to which participants were expected to correct towards following the perturbation. Separate plots are shown for Fast, shortened (a) and lengthened (b) intervals, and Slow, shortened (c) and lengthened (d) intervals. Error bars show SEM.
Fig 5.
Mean relative asynchronies before and after the perturbation (vertical grey bar) at time T for Auditory-Visual cues.
Dotted horizontal line shows the zero relative asynchrony measure, to which participants were expected to correct towards following the perturbation. Separate plots are shown for Fast, shortened (a) and lengthened (b) intervals, and Slow, shortened (c) and lengthened (d) intervals. Error bars show SEM.
Fig 6.
Mean correction gain as calculated by a linear phase correction model fitted to the asynchrony data of each participant.
Higher values indicate participants are correcting their movements to a bigger proportion of the preceding step asynchrony between themselves and the Avatar. Correction gains are shown for Fast versus Slow and Visual-Only (VO) versus Auditory-Visual (AV) conditions. Error bars show SEM. * represents significance at p<.05, ** is p<.01.
Fig 7.
View of the avatar seen by participants when completing the stepping task.
Fig 8.
Example of asynchronies before and after phase wrapping removed.
a). Asynchronies observed for a fast tempo trial when matching participant onsets to the nearest Avatar step onsets. Wrapping of the asynchrony can be seen between the intervals of +/-0.4 seconds which does not correspond to an attempt to regain synchrony. b). Asynchronies for the same trial but with phase wrapping removed and steps assigned to ensure asynchronies are continuous. This is a severe example of the drift observed in some fast tempo visual-only trials to demonstrate the unwrapping procedure.