Interindividual differences in digital phenotypes of major depressive disorder: A passive sensing study using...

Area of Science:

• Digital Health
• Psychiatry
• Computational Social Science

Background:

• Major Depressive Disorder (MDD) exhibits significant heterogeneity in symptom presentation.
• Previous subtype identification relied on subjective self-reported symptoms.
• Objective data from digital devices offer potential for novel MDD phenotyping.

Purpose of the Study:

• To identify latent profiles of MDD using digital biomarkers from smartphones and wearables.
• To explore associations between identified digital phenotypes and MDD severity and functioning.

Main Methods:

• Collected passive sensing data from smartphones and Garmin smartwatches in 297 individuals with MDD.
• Utilized digital biomarkers including sleep patterns, physical activity, screen time, social engagement, and heart rate variability.
• Employed latent profile analysis to identify distinct digital phenotypes.

Main Results:

• A two-profile solution emerged: 'average' (85.7%) and 'deficient sleep, low HRV, low social engagement' (14.3%).
• No significant difference in MDD symptom severity between profiles.
• Profile 2 exhibited trends towards lower social and occupational functioning, though not statistically significant after correction.

Conclusions:

• Digital phenotypes of MDD can be identified using wearable and smartphone data.
• Sleep dysregulation, low heart rate variability, and reduced social engagement may indicate functional impairments.
• Further research with diverse samples and additional biomarkers is needed to validate these digital phenotypes.