APS
2026 APS Annual Convention
Advancing Predictive Models of Mental Health Outcomes Using Intensive Longitudinal Data
Intensive longitudinal data, gathered through frequent assessments over time, offer opportunities to improve prediction of mental health outcomes. This symposium covers symptom dynamics predicting anxiety disorder onset, baseline predictors of cognitive therapy response, time series foundation models for digital phenotyping, and validating longitudinal prediction models for clinical use.
Chairs & Discussants
- Dawson HaddoxChair
University of Arizona - Eiko FriedDiscussant
Leiden University
Presentations
- Investigating the Role of Symptom Activation Patterns As a Disorder-Generating Mechanism Predicting the Onset of Anxiety DisordersOmid Ebrahimi
- For Whom Is Cognitive Therapy for Anxiety Most Efficacious? a Machine Learning Model.Matt Southward, Douglas Terrill, Madeline Kushner, Shannon Sauer-Zavala
- Benchmarking Time Series Foundation Models for Digital Phenotyping Using Zero-Shot and Fine-Tuning ApproachesDawson Haddox
- Just-In-Time or Just between People? Tackling the Obstacles of Validating Prediction Models That Use Longitudinal DataNicholas Jacobson, Anna Langener
