Figure 5. Nice figure. It would be interesting to see step width distributions as well, comparing perturbed and unperturbed, as it would answer my question of whether people’s sideways dynamics were substantially affected as well. This is not absolutely necessary for the point that the authors wish to make, but could be a quick easy thing for the authors to generate from their data (especially given that they have already estimated the stride-length, step width is probably only a couple of lines of code!).
Reviewer 3 (Srinivasan):