A Population Analysis Approach for the analysis of Depression Episodes using mobility data

Depression is one of the common mental illnesses that is impacting millions of people around the world. Although there are numerous clinical diagnostic procedures available, the majority of them are predominantly based on self-reporting of patients and/or observational assessment by clinicians. Recent studies show that depression could alter the motor skills and lessen the overall motor complexity of the affected person as compared to a healthy individual. This opens the door for the possibility of developing a new approach to diagnose depression by collecting the mobility data of individuals and looking for distinguishing characteristics associated with their mobility patterns. With the current availability of a wide range of wearable devices and the increasing advancement of wireless technology, this possibility appears to be very attractive. The main objective of this study is to develop a new approach that takes advantage of data collected from wearable devices and analyzes mobility patterns for a given group of subjects. We propose a novel population analysis approach using correlation networks that compares mobility parameters of the population and identifies subgroups that exhibit similar motor complexity. The proposed approach involves a two-step process. In the first step, we construct a correlation network modeled by graphs by using their mobility data. In the second step, we employ a clustering algorithm to discover groups with similar mobility profiles from the constructed network. Our results demonstrate that depressed patients are grouped into a separate cluster. Furthermore, we conducted enrichment analysis to identify similar and distinguishable properties associated with each cluster