Understanding Heterogeneity Across PTSD, MDD and Anxiety By Leveraging Large-scale Multimodal Neuroimaging Datasets

Project summary Psychiatric comorbidities are common in individuals with posttraumatic stress disorder (PTSD), major depressive disorder (MDD) and anxiety disorders (ANXs) such as generalized anxiety disorder (GAD), social anxiety disorder (SAD). Patients with multiple comorbidities often experience more symptom severity, higher risk of suicidal behavior, greater impairments in neurocognitive functioning, and poorer treatment outcomes. Diagnostic heterogeneity and high comorbidity among these mental disorders significantly complicating progress in clarifying underlying neural mechanisms of mental disorders, and efforts to advance personalized medicine. To address this issue and respond to NIH’s call for precision medicine, our proposal aims to utilize large clinical and imaging datasets, along with explainable deep learning models, to identify neural biomarkers and biotypes underlying PTSD, MDD and ANXs, and to assess their utility of clinical outcomes. We will first apply a top-down Research Domain Criteria (RDoC) approach to identify distinct biomarkers. In addition, we will utilize a bottom-up data-driven computational framework to identify shared transdiagnostic biotypes to tackle heterogeneity across PTSD, MDD and ANXs. We will leverage multimodal neuroimaging data including structural magnetic resonance imaging (MRI) and resting-state functional MRI (rs-fMRI) from four globally representative lifespan datasets: the ENIGMA-PTSD, the ENIGMA-MDD, the ENIGMA-Anxiety, and UK Biobank (UKBB) datasets, to test the generalizability of the deep learning models, and the utility of the identified biomarkers (from aim 1) and the data-driven transdiagnostic biotypes (from aim 2) in predicting treatment outcomes (aim 3). The results of this study may enhance our understanding of unique and common neural profiles across PTSD, MDD, and ANXs, and help future treatment development by providing initial guidance of selecting best treatments that works the best for individual patient.