Smart-phone based audio and video digital biomarkers observed in naturalistic settings predict current negative and positive symptom severity in schizophrenia

 

• Outpatients with schizophrenia participating in an open-label clinical trial completed a remote picture description task on their smartphone
• We leveraged over 1000+ longitudinal datapoints collected over the course of a year by applying generalized estimating equations (GEE), enabling maximal use of all available clinical trial datapoints
• Digital measurements of facial expression, head movement, and speech were associated with distinct profiles of negative and positive symptom severity
• These results provide further clinical validation of audio-video digital biomarkers and demonstrate the ability to predict symptom severity from light-weight active remote assessments