Monitoring Changes In Daily Actigraphy Patterns Of Free-Living Patients

The current development of actigraphs integrated in small and discrete devices allows noninvasive recording of patient activity over several days or even months. The information obtained by these devices allows the analysis of daily activity patterns and therefore may be a useful tool to monitor the status of out-patients in diseases such as major depression. However, the full exploitation of this information requires automated systems that reduce the inherent complexity of these data and facilitate their interpretation by the clinicians. Thus in this paper a Daily Activity Monitoring System (DAMS) is presented based on functional data analysis algorithms for signal alignment and non-linear dimensionality reduction techniques based on manifolds. The DAMS allows robust processing of actigraphy data, and visual detection of changes or anomalies in routine activity.