OBJECTIVES/GOALS: This study will collect multimodal and longitudinal data in adults with obsessive-compulsive disorder and healthy controls. A mixed effects random forest machine learning approach will be taken to develop a model that can predict individualized longitudinal OCD symptom burden. METHODS/STUDY POPULATION: Baseline resting state functional MRI (rsfMRI) and measures of symptom burden will be collected in adults with OCD and healthy controls. Longitudinal measures of behavior and physiology–such as heart rate, activity, and sleep metrics - will be collected using Fitbit Charge 5 tracker. Daily assessments of symptom burden and functional status will be collected through a smartphone app. Individuals with OCD will start pharmacotherapy during the study period and all participants will be followed for a total of 10 weeks. Repeat rsfMRI imaging will occur at study conclusion. Data will be analyzed using a mixed effects random forest machine learning algorithm with assessment of model performance. RESULTS/ANTICIPATED RESULTS: Prior studies of symptom severity in psychiatric illness and affect in non-clinical populations have found longitudinal features - such as lexical and acoustic measures, participant context, heart rate, and sleep metrics–that were predictive of these states over time. It is anticipated that the present study will extend these results to individuals with OCD and identify physiologic and behavioral features that track personalized symptom burden longitudinally in this patient population. A model able to predict when symptoms are elevated could allow for provision of additional treatment or interventions targeted to times of high symptom burden. DISCUSSION/SIGNIFICANCE: This study will be the first to collect and analyze longitudinal measures of behavior, symptoms, and physiology in patients with OCD with a goal of predicting symptom burden. Identification of elevated symptom burden would allow for implementation of just-in-time treatment, during these periods.