Passive smartphone sensors for detecting psychopathology

Importance Smartphone sensors can continuously and unobtrusively collect clinically relevant behavioral data, allowing for more precise symptom monitoring in clinical and research settings. However, progress in identifying unique behavioral markers of psychopathology from smartphone sensors has been stalled by research on diagnostic categories that are heterogenous and have many non-specific symptoms.Objective Determine which domains of psychopathology are detectable with smartphone sensors and identify passively sensed markers for (1) general impairment (the p-factor) and (2) specific transdiagnostic domains.Design Cross-sectional study completed in 2023 including a baseline survey and 15 days of smartphone monitoring.Setting Participants were recruited from the community via a clinical research registry.Participants Volunteer sample selected for mental health treatment status (51% currently in outpatient treatment, 29% with treatment history).Main outcomes and measures Transdiagnostic psychopathology dimensions of internalizing, detachment, disinhibition, antagonism, thought disorder, somatoform, and the p-factor; 27 behavior markers derived from GPS, accelerometer, motion, call logs, screen on/off, and battery status.Results Among the 557 participants included in the study, 82% were female and the mean age was 30.7 (SD=8.8). Multiple correlation (R) showed the domain most strongly related to sensed behavior was detachment (R=.42) followed by somatoform (R=.41), internalizing (R=.37), disinhibition (R=.35), antagonism (R=.33), and thought disorder (R=.28). Each psychopathology domain had significant, bivariate associations with 4-10 smartphone sensor variables. After adjusting for shared variance between psychopathology dimensions, all domains except thought disorder retained significant, incremental associations with sensor variables, reflecting unique behavioral signatures (range of standardized s=|.11|-|.24|The p-factor was associated with lower mobility, more time at home, later bedtime, and less phone charge (s=|.12|–|.24|).Conclusions and relevance In this cross-sectional study, we identified behavioral markers for domains encompassing most major forms of psychopathology using smartphone sensor data. In addition to establishing the breadth of psychopathology that was detectable, findings show smartphone sensors assessed markers that distinguish domains of dysfunction. Results showing many behavioral markers reflected non-specific psychopathology reinforces the need for dimensional, transdiagnostic models to maximize the potential of mobile sensing technology. Findings from this study can advance research on day-to-day maintenance mechanisms of psychopathology and inform development of symptom monitoring tools.