If the perturbations are only in the fore-aft direction, it is possible that the data set is insufficiently rich to infer the human walking control system; but it is also possible that the data set is rich enough due to sufficient coupling of the various degrees of freedom (fore-aft and sideways degrees of systems, to be a bit colloquial).
The new stride width box plot shows that we do get stride width variation in the perturbed walking but not as much a relative variation difference as stride length and frequency.
Reviewer 3 (Srinivasan):