Characterizing the Complexity of Human Unipedal Balance

We investigate the data obtained as a person stands one-legged on a force plate to identify underlying pathologies concerning a participant’s balance process. In this pilot study, we utilize information-theoretic tools to quantify the degree of complexity in a signal, allowing subjects to be classified based on their complexity profile. To determine which tools are most appropriate for characterizing human balance, we filtered the medial-lateral jerk signal to reduce the influence of machine noise. Moreover, our preliminary analysis suggests that such signals may have a deterministic character along with temporally evolving statistics. The Lempel-Ziv Complexity was chosen due to its ubiquity in characterizing coarse-grained signals that have qualities of randomness but may have an underlying deterministic mechanism. After decomposing the signal into its constituent modes using multiresolution analysis, we observed that our complexity measure exhibits signs of regular behavior among the modes when applied to overlapping time-series windows. This suggests that the Lempel-Ziv Complexity can indicate the existence of latent sources of determinism in human balance, but more trials need to be conducted to get a full classification. Further developments of this project can provide health professionals with a measure they can use to diagnose problems non-invasively and affordably.