JMIR Publications

ABSTRACT

Background:

Depression is the top contributor to global disability, with prevalence estimates approaching 30% among college students. Early detection of depression enables timely intervention and reduces its physical and social consequences. Passive, device-based sensing further enables its detection at a low burden to the individual.

Objective:

We leverage an ensemble machine learning method (light gradient-boosting machine; LightGBM) to detect depression entirely through passive sensing.

Methods:

A diverse sample of undergraduate students (N = 28; Mage = 19.96, SDage= 1.23; 53.6% female; 46.4% Latine, 35.7% Asian, 14.3% non-Latine White, 3.6% other) participated in an intensive longitudinal study. Participants wore two devices (Oura ring [sleep, physiology], Samsung smartwatch [physiology, movement data]) and installed the AWARE software on their mobile devices, which collects passive sensing data such as screen time. Participants were derived from a randomized controlled trial of a positive psychology mhealth intervention. They completed a self-report measure of depressive symptoms administered weekly over a 19-22 week period.

Results:

The LightGBM model achieved an F1 score of 0.744 and a Cohen's kappa coefficient of 0.474, indicating moderate agreement between the predicted labels and the ground truth. The most predictive features of depression were sleep quality and missed mobile interactions.

Conclusions:

Findings suggest that data collected from passive sensing devices may provide real-time, low-cost insight into the detection of depressive symptoms in college students and may present an opportunity for future prevention and perhaps intervention. Clinical Trial: This study was not preregistered.