Abstract Rumination is a transdiagnostic risk factor for developing depression and anxiety in adolescents. Rumination reflects compromised inhibitory control capability and is fundamentally a maladaptive psychological process in response to stressors. Understanding the neural correlates of rumination and inhibitory control enriches our knowledge on the neural and cognitive propensity for rumination, thereby informing the design of effective clinical programs for the prevention and treatment of adolescent mental health issues. To achieve the objectives of identifying neural features associated with trait rumination and inhibitory control capabilities, the proposed project will analyze the multimodal data from the parent R01 including anatomical MRI, resting state fMRI, the severity and temporal dynamics of trait rumination measured via Ecological Momentary Assessments (EMA), and performance profiles from a Sustained Attention to Response Task (SART). Voxelwise-based morphometry (VBM) will be used to obtain voxelwise estimate of gray matter density. Algorithms for calculating the Amplitudes of Low Frequency Fluctuation (ALFF) from the resting state fMRI data will be utilized to obtain voxelwise ALFF maps. Resting State Functional Connectivity (RSFC) analyses will be conducted using Regions of Interest (ROI) including subgenual prefrontal cortex, left amygdala and Posterior Cingulate Cortex (PCC). Voxelwise maps of gray matter density, ALFF and RSFC will be used for linear regression analyses with EMA and SART metrics as regressors to identify significant clusters after false discovery rate correction. Based on existing knowledge on the neurobiology of rumination and inhibitory control, we hypothesize trait rumination to be significantly associated with the gray matter density of the Dorsal Lateral Prefrontal Cortex (DLPFC), the ALFF values of the medial prefrontal cortex, amygdala, hippocampus and subgenual cingulate cortex, as well as the RSFC between the subgenual prefrontal cortex and Default Mode Network (DMN); we also hypothesize SART commission errors (incorrect response to “no-go” trials) and frequencies of task-unrelated thoughts during SART probe trials to be associated with the gray matter density at DLPFC and dorsal anterior cingulate cortex as well as the RSFC between DMN, frontal-parietal network and salience network. Finally, the MRI metrics of the significant clusters identified above will be entered into Machine Learning (ML) processes to predict depression severity as measured by the Center for Epidemiological Studies Depression (CESD) scale scores. The ML processes include data-driven feature selection and cross-validation steps to quantitatively evaluate the predictive power of these neural features.