Devin Schoen Institute: University of California San Francisco Date: December 5, 2024 Title: Advancing predictive models for deep brain stimulation outcomes with multi-contrast MRI and hypergraph techniques Abstract: Deep brain stimulation (DBS) is a well-established therapy for treating Parkinson’s disease motor symptoms, but patient outcomes remain highly variable. To better account for this variance, there is a critical need for improved prognostic tools that integrate clinical metrics and imaging data already collected in routine care. This talk highlights the use of MRI biomarkers, fractal dimension from T1-weighted imaging, combined with clinical data through hypergraph neural networks. These techniques enhance our ability to predict DBS outcomes and demonstrate adaptability to varying data availability. Future directions include incorporating additional MRI contrasts and refining hypergraph-based methods to better support clinical decision-making and optimize personalized treatment planning.
