PROJECT ABSTRACT The millions of brain MRIs acquired each year in clinical settings are a vastly underutilized scientific resource. Meanwhile, brain MRIs collected in research studies often struggle to recruit cohorts fully representative of populations of interest, particularly in the case of rare neurodevelopmental disorders. One clear opportunity for discovery science are pediatric MRIs where gross pathology has been ruled out in children with headache, concern for head trauma, or neurodevelopmental symptoms such as delayed milestones, autism, or psychosis. One reason why brain morphology phenotypes derived from clinical brain MRIs have been underutilized is due to the expected technical differences across sites and scanners, which pose methodological and statistical challenges. However, our study team recently developed brain growth charts based on one of the largest aggregated neuroimaging datasets to date, which provide an unprecedented model for brain MRI developmental imaging phenotypes -- including quantitative measures of brain volumes of cortical and subcortical brain regions – to be accurately benchmarked against population norms while controlling for technical differences. In preliminary investigations, we have now further developed this approach to benchmark individual patients' brain anatomy against hospital norms – so-called “clinical controls” – to provide high-quality imaging metrics to combine with data on genetics, environmental exposures, and neuropsychiatric outcomes. Using advanced, fully-reproducible image processing, deep learning, and statistical modelling, we will develop a clinical radiomics reference that allows precise quantification of where an individual clinically-acquired brain MRI lies in reference to expected developmental variation, using 20,000 clinical MRIs without gross pathology obtained by Children's Hospital of Philadelphia care teams (Aim 1). By integrating records of clinical MRIs with large-scale CHOP biobanks, we will identify developmental imaging phenotypes associated with high genetic risk for neuropsychiatric and neurodevelopmental disorders (Aim 2). By integrating secondary analysis of clinically-acquired scans with prospective behavioral phenotyping and neuroimaging, we will identify developmental imaging phenotypes associated with psychiatric vulnerability in individual patients (Aim 3). Throughout, we will generate practically useful brain chart resources that will facilitate analysis of clinically- acquired brain scans by other researchers (Aim 4). By leveraging clinically-acquired brain MRIs, this proposal harnesses untapped neurogenomic information to investigate neurodevelopment trajectories in high-risk youth, capitalizing on the PI and assembled team's expertise in psychiatric and developmental brain imaging, imaging-genetics and neuroinformatics. The project will provide the necessary infrastructure to support future HIPAA-compliant multi-site collaborative projects across health syste...