ABSTRACT Epilepsy is a devastating neurological illness that affects over 50 million people worldwide. Approximately one-third of patients do not respond to anti-seizure medication (ASM) and require additional diagnostic work-up, including consideration for surgery. Structural neuroimaging plays a pivotal role in the diagnostic evaluation of epilepsy, identifying visible lesions in many patients that co-localize with the seizure focus. However, up to 40% of patients have normal-appearing MRIs and this number is growing. As a result, there is increased interest in identifying subtle brain network abnormalities that could help to delineate the epileptogenic network and aid in the prediction of treatment response (i.e., response to ASMs and surgical outcomes). Unfortunately, methods for reliably identifying which patients will be drug-responsive versus drug- resistant, and which patients will achieve successful versus unsuccessful surgical outcomes are lacking. A major barrier to progress in this field has been obtaining quantitative imaging, including structural MRI (sMRI) and diffusion-weighted imaging (dMRI), clinical, and genetic data on large, geographically diverse samples of patients in whom different treatment outcomes can be evaluated. In the past, sample sizes have been insufficient to detect subtle, but reliable, brain abnormalities in patients with focal or generalized epilepsies that are genuinely associated with epilepsy and not with vicissitudes related to small or geographically restricted samples. A new, large-scale data initiative, ENIGMA4-Epilepsy, coupled with technological advancements that enable improved data harmonization are now lifting these barriers and allowing us to combine multi-site sMRI/dMRI, clinical, genetic data to predict important clinical outcomes, and making the results generalizable to a global epilepsy community. In this grant, we will leverage data collected through ENIGMA-Epilepsy—a consortium of 24 epilepsy centers from 14 countries (more than 2,250 patient and 1,727 healthy control sMRI/dMRI datasets) and the Human Epilepsy Project (HEP). We will include new network models (i.e., individualized connectomes) and polygenic risk scores (PRS) to test whether a combination of imaging, clinical, and genetic risk can accurately predict two clinical outcomes: drug-resistance and post-operative seizure outcome. Our scientific premise is that MRI-based assessment of whole-brain network properties, in combination with clinical data and PRS derived from genetic data, are able to predict (i) drug response in recently diagnosed epilepsy cases and (ii) postsurgical outcomes in individuals with drug-resistant epilepsy. This R01 addresses NIH's call for more reproducible studies by introducing a highly-powered design capable of capturing variability across patients with diverse clinical characteristics and treatment outcomes. This grant is also directly aligned with NINDS's 2020 Epilepsy Benchmarks (IIIB), which encou...