IDENTIFYING RELATIONSHIPS BETWEEN REINFORCEMENT LEARNING AND MOOD IN DEPRESSION - PROJECT SUMMARY DEPRESSION IS THE LEADING CAUSE OF DISABILITY WORLDWIDE, BUT AN UNDERSTANDING OF ITS UNDERLYING PROCESSES HAS LIMITED THE DEVELOPMENT OF IMPROVED TREATMENT. REWARD LEARNING RESEARCH IDENTIFIES DISRUPTIONS IN VALUATION AND LEARNING ABOUT REWARDS AS CENTRAL TO DEPRESSION. COMPUTATIONAL PSYCHIATRY STUDIES USING REINFORCEMENT LEARNING (RL) MODELS TO TEST HYPOTHESES ABOUT REWARD LEARNING HAVE POSITIONED RL PARAMETERS AS PROMISING THERAPEUTIC TARGETS FOR DEPRESSION. HOWEVER, A CRITICAL GAP REMAINS IN UNDERSTANDING HOW RL CONTRIBUTES TO DEPRESSION. THIS F31 PROJECT AIMS TO IDENTIFY SPECIFIC, MODIFIABLE RL PROCESSES UNDERLYING LOW MOOD IN DEPRESSION. EMERGING EVIDENCE SUGGESTS RECIPROCAL INFLUENCES BETWEEN RL AND MOOD, SUCH THAT MOOD REFLECTS THE MOMENTUM OF REWARD AVAILABILITY, AND MOOD CHANGES INFLUENCE RL EFFICIENCY; THESE DATA OFFER A TESTABLE PATHWAY LINKING RL TO DEPRESSION. THIS PROJECT WILL ADAPT AN RL-ONLY COMPUTATIONAL MODEL WITH A MOOD UPDATING PARAMETER TO CREATE AN RL-MOOD MODEL. DATA WILL COME FROM SPONSOR DR. PEARL CHIU'S R01MH127773, WHICH INCLUDES PARTICIPANTS ASSIGNED TO AN ACTIVE (N=100) OR CONTROL ARM (N=100) OF A GUIDED REWARD LEARNING PARADIGM DESIGNED TO EVOKE CHANGES IN LEARNING PARAMETERS. AIM 1 HYPOTHESIZES THAT THE RL-MOOD MODEL WILL BETTER FIT BASELINE DATA AND PREDICT SYMPTOMS THAN AN RL-ONLY MODEL, SUPPORTING INTERACTIONS BETWEEN DISRUPTED RL AND MOOD CONTRIBUTING TO DEPRESSION. AIM 2 BUILDS ON THE APPLICANT'S MASTER'S THESIS, WHICH SHOWED THAT RL CHANGES CORRELATED WITH SYMPTOM REDUCTIONS FOLLOWING THE GUIDED REWARD LEARNING PARADIGM, TO TEST THE HYPOTHESIS THAT INCREASING RL- MOOD MODEL PARAMETERS MEDIATES THE RELATIONSHIP BETWEEN RL CHANGES AND SYMPTOM CHANGES. SUCCESSFUL COMPLETION OF THIS PROJECT HAS SIGNIFICANT LONG-TERM IMPLICATIONS, INCLUDING ADVANCING UNDERSTANDING OF LOW MOOD IN DEPRESSION, INFORMING BEHAVIORAL LEARNING INTERVENTIONS, AND ENABLING PERSONALIZED TREATMENT. THE TRAINING PLAN INCLUDES WEEKLY MENTORSHIP MEETINGS WITH SPONSOR DR. CHIU, AN EXPERT IN COMPUTATIONAL PSYCHIATRY, AND MONTHLY MEETINGS WITH BIOSTATISTICIAN MENTOR DR. ALEXANDRA HANLON. IN ADDITION, THE APPLICANT WILL ATTEND AND PRESENT AT NATIONAL AND INTERNATIONAL CONFERENCES, ATTEND TARGETED SEMINARS AND WORKSHOPS, AND PURSUE FURTHER CLINICAL TRAINING. CONDUCTING THIS RESEARCH IN DR. CHIU'S LAB AT VIRGINIA TECH, A TOP-TIER RESEARCH INSTITUTION, THE APPLICANT WILL HAVE ACCESS TO STATE-OF-THE-ART RESOURCES, INCLUDING A DEDICATED COMPUTING CLUSTER, A COLLABORATIVE INTERDISCIPLINARY RESEARCH COMMUNITY, AND EXTENSIVE STATISTICAL AND PROFESSIONAL DEVELOPMENT SUPPORT. THESE ACTIVITIES WILL FULFILL THE APPLICANT'S TRAINING GOALS OF ENHANCING EXPERTISE IN LONGITUDINAL ANALYSES, COMPUTATIONAL MODELING, DEPRESSION, AND SCIENTIFIC WRITING WHILE POSITIONING HER TO ACHIEVE HER LONG-TERM GOAL OF AN ACADEMIC CAREER IN COMPUTATIONAL PSYCHIATRY, FOCUSED ON MECHANISTICALLY-INFORMED INTERVENTIONS FOR MOOD DISORDERS.