Using smart phone mobility traces for the diagnosis of depressive and manic episodes in bipolar patients

Agnes Grünerbl, Venet Osmani, Gernot Bahle, Jose C. Carrasco, Stefan Öhler, Oscar Mayora, Chrisitan Haring, Paul Lukowicz

In: Proceedings of the 5th ACM Augmented Human International Conference. Augmented Human International Conference (AH) 5th March 7-9 Kobe Japan ACM 3/2014.


In this paper we demonstrate how smart phone sensors, specifically inertial sensors and GPS traces, can be used as an objective “measurement device” for aiding psychiatric diagnosis. In a trial with 12 bipolar disorder patients conducted over a total (summed over all patients) of over 1000 days (on average 12 weeks per patient) we have achieved state change detection with a precision/recall of 96%/94% and state recognition accuracy of 80%. The paper describes the data collection, which was conducted as a medical trial in a real life every day environment in a rural area, outlines the recognition methods, and discusses the results.

Deutsches Forschungszentrum für Künstliche Intelligenz
German Research Center for Artificial Intelligence