Tracking mental health disorder symptomology using a smartphone app

A bipolar disorder targeted smartphone dataset comprising 291 individuals over several months was analyzed, with all information collected via an app which monitors smartphone interactions such as typing dynamics and accelerometer motion.
Authors concluded that smartphone accelerometer data can successfully be used to extract the necessary information to track and predict fluctuations in mental health disorder symptomology outside of the traditional clinical setting.
Results also suggested that accelerometer data improves the predictive performance of machine learning models.