Mobile Phone Sensor Correlates of Depressive Symptom Severity in Daily-Life Behavior: An Exploratory Study

TLDR

Depression is a common, burdensome disorder that is frequently undetected and untreated. This study aimed to investigate whether mobile phone GPS and usage sensor data could detect daily‑life behavioral markers indicative of depressive symptom severity. Forty community adults carried a phone with a data‑collection app for two weeks; 28 provided sufficient data, and participants completed a PHQ‑9 survey at baseline while GPS and usage features were extracted for analysis. GPS metrics such as circadian movement, normalized entropy, and location variance, along with phone usage duration and frequency, were strongly correlated with PHQ‑9 scores, and a classifier using normalized entropy achieved 86.5 % accuracy with a regression model error of 23.5 %, indicating that mobile sensors can serve as behavioral markers of depression.

Abstract

Depression is a common, burdensome, often recurring mental health disorder that frequently goes undetected and untreated. Mobile phones are ubiquitous and have an increasingly large complement of sensors that can potentially be useful in monitoring behavioral patterns that might be indicative of depressive symptoms.The objective of this study was to explore the detection of daily-life behavioral markers using mobile phone global positioning systems (GPS) and usage sensors, and their use in identifying depressive symptom severity.A total of 40 adult participants were recruited from the general community to carry a mobile phone with a sensor data acquisition app (Purple Robot) for 2 weeks. Of these participants, 28 had sufficient sensor data received to conduct analysis. At the beginning of the 2-week period, participants completed a self-reported depression survey (PHQ-9). Behavioral features were developed and extracted from GPS location and phone usage data.A number of features from GPS data were related to depressive symptom severity, including circadian movement (regularity in 24-hour rhythm; r=-.63, P=.005), normalized entropy (mobility between favorite locations; r=-.58, P=.012), and location variance (GPS mobility independent of location; r=-.58, P=.012). Phone usage features, usage duration, and usage frequency were also correlated (r=.54, P=.011, and r=.52, P=.015, respectively). Using the normalized entropy feature and a classifier that distinguished participants with depressive symptoms (PHQ-9 score ≥5) from those without (PHQ-9 score <5), we achieved an accuracy of 86.5%. Furthermore, a regression model that used the same feature to estimate the participants' PHQ-9 scores obtained an average error of 23.5%.Features extracted from mobile phone sensor data, including GPS and phone usage, provided behavioral markers that were strongly related to depressive symptom severity. While these findings must be replicated in a larger study among participants with confirmed clinical symptoms, they suggest that phone sensors offer numerous clinical opportunities, including continuous monitoring of at-risk populations with little patient burden and interventions that can provide just-in-time outreach.