This collaborative project uses digital phenotyping methods to measure emotion, mood, and functional outcomes within the context of Prechter Longitudinal Study of Bipolar Disorder. It integrates three technologies – our proprietary PRIORI (Predicting Individual Outcomes for Rapid Intervention) app, an ecological momentary assessment (EMA) protocol, and a Fitbit wearable device. The PRIORI app was developed by Dr. Emily Mower Provost in Engineering and Computer Sciences and Dr. Melvin McInnis, and uses machine algorithms to detect emotion activation and valence from speech collected passively on participants smartphones. Initial validation was completed on clinical interview telephone calls and has been expanded to capture ambient speech and noise as participants go about their normal daily lives. In our new deployment, the EMA protocol that will assess participants self-perceptions of their emotional valence and arousal and integrate with Fitbit devices that will measure physiological arousal and regulation through activity counts, heart rate, and sleep and circadian patterns. We will use anomaly detection methods to identify whether changes in passively detected speech and activity are associated with self-reported emotional valence and arousal, and whether this differs as a function of context (e.g., socialization, location).