Identifying Person-Specific Drivers of Adolescent Depression via Idiographic Network Modeling of Active and Passive Smartphone Data - PROJECT SUMMARY/ABSTRACT
Adolescents experience escalating risk for developing clinical depression, which can lead to lifelong morbidity
and mortality. The neural, physical, cognitive, and socioemotional changes that may contribute to this risk also
signal an opportunity for high impact intervention. Unfortunately, psychotherapy trials demonstrate modest
effects on youth depression. To improve long-term outcomes for adolescents, this study will identify person-
specific drivers of adolescent depression that can guide treatment personalization. Prior research with
depressed or anxious adults demonstrates the existence of such drivers—symptoms and related processes
that are influential (i.e., predict change in other symptoms), modifiable, and exhibit individual differences.
Personalized selection and sequencing of cognitive behavioral therapy (CBT) modules to target these drivers
early in adults have produced larger treatment effects compared to a historical benchmark. Identifying person-
specific drivers during adolescence could inform treatments that account for both developmental and individual
differences to shift the trajectory of depression onset and maintenance. Investigating person-specific drivers
usually involves intensive surveying of self-reported experience via smartphone-based ecological momentary
assessment (EMA). Emerging evidence suggests that smartphones can also monitor mood through passive
sensing of depression-related behaviors with minimal response burden. However, nearly all such studies have
been conducted with adults, despite near universal smartphone ownership among adolescents in the US.
Thus, this study will leverage depressed adolescents' everyday smartphone use to assess the validity of
mobile sensing against established ambulatory methods (i.e., EMA and actigraphy) to identify person-specific
drivers of adolescent depression. Fifty adolescents (12–18 years old) with elevated depressive symptoms will
participate in 30 days of: a) smartphone-based EMA of depressive symptoms, processes, and affect (4x/day),
sleep diary (1x/day); (b) mobile sensing of mobility, physical activity, sleep, natural language use in typed
interpersonal communication, screen-on time and call frequency/duration; and (c) wrist actigraphy of physical
activity and sleep. Adolescents and caregivers will complete diagnostic interviews and other measures (e.g.,
developmental, clinical, Research Domain Criteria) at baseline, as well as user feedback interviews at follow-
up. To address study aims: 1) idiographic, within-subject networks of EMA symptoms will be modeled to
identify each adolescent's drivers; 2) correlations among EMA, mobile sensor, and actigraph measures of
sleep, physical, and social activity; and machine learning prediction of core depressive symptoms (self-
reported mood and anhedonia) will be used to assess the validity of mobile sensing for identifying person-
specific drivers; 3) between-subject baseline characteristics will be explored as predictors of person-specific
drivers. These results will inform future development of a scalable, low-burden smartphone-based tool that can
guide personalized treatment decisions for depressed adolescents, with potential public health impact.
