Using data processing to understand inconsistency in smartphone behavior among patients with serious mental illness: results of a digital phenotyping biomarker study

Abstract

BackgroundSmartphone-based digital phenotyping can serve as novel transdiagnostic markers of mental illness including circadian routines and anhedonia. In proposing transdiagnostic digital phenotypes for circadian routines and anhedonia in depression and bipolar disorder patients, this paper explores their derivation, comparison to naive models, and replicability across two different research sites/teams. Methods84 participants (bipolar disorders, depression, and controls) used the mindLAMP app for 12 weeks to capture digital phenotypes on their personal smartphones with the mindLAMP app. Active data captured included surveys about mood symptoms including anhedonia and passive data captured included device acceleration, geolocation, and screen state. A chronotype feature related to sleep and activity states was derived from this passive data. Within and between-participant models were created to assess how time-varying features collected through digital phenotyping could predict anhedonia scores for each week. ResultsWithin-person models outperformed between-person models in predicting anhedonia. Chronotype was the strongest predictor of weekly anhedonia scores as indicated by Shapley values. Shapley scores revealed that many of the time-varying predictors variables are significant but differ in their direction of action. DiscussionThis analysis reveals the meaningful but potentially misleading nature of digital phenotyping signals. Results suggest that each participant has a unique set of relationships between time-varying digital phenotype variables; therefore, it is challenging to predict trends between participants. Bayesian models, with appropriate population priors, may thus offer the next step for improving the potential of personalized digital phenotyping insights.