A Surrogate Technique for Investigating Deterministic Dynamics in Discrete Human Movement

Abstract

Entropy is an effective tool for investigation of human movement variability. However, before applying entropy, it can be beneficial to employ analyses to confirm that observed data are not solely the result of stochastic processes. This can be achieved by contrasting observed data with that produced using surrogate methods. Unlike continuous movement, no appropriate method has been applied to discrete human movement. This article proposes a novel surrogate method for discrete movement data, outlining the processes for determining its critical values. The proposed technique reliably generated surrogates for discrete joint angle time series, destroying fine-scale dynamics of the observed signal, while maintaining macro structural characteristics. Comparison of entropy estimates indicated observed signals had greater regularity than surrogates and were not only the result of stochastic but also deterministic processes. The proposed surrogate method is both a valid and reliable technique to investigate determinism in other discrete human movement time series.

Characterizing and modeling the joint-level variability in human walking

“…To more fully investigate the effects of continuity constraints, further experiments with continuous walking are planned. In addition, non-parametric methods will be used to more fully characterize the variability across steps (Hausdorff et al, 1995; Taylor et al, 2016). …”