Our lab uses a multimodal approach – including passive sensing from smartphones and wearables, experience sampling, neuroimaging, laboratory-based experiments, and clinical interviewing – integrated with AI and machine learning methods to investigate the etiology and pathophysiology of depression in adolescents, as well as the mechanisms of change in psychotherapeutic and digital interventions for depression. Ongoing studies examine predictors and mechanisms of change in app-based mindfulness training, use AI to detect emotional distress and depression risk from smartphone and wearable data, and develop just-in-time interventions that deliver personalized support during moments of heightened need. We also investigate predictors of depression onset in at-risk youth.