Optimization

Neural engineering treatments, particularly neuromodulation treatments, are still largely trial-and-error based treatments, involving numerous adjustable parameters that significantly influence patient-related outcomes. Inter-subject variability requires treatments that are both personalized and effective for broader patient cohorts, adding strain to medical device manufacturers and healthcare systems. Computational modeling thus emerges as a more efficient way to simulate, test, refine, and personalize neural engineering treatments.

The vast high-dimensional parameter space of neural engineering treatments makes current computer-based approaches to therapy design resource-intensive. To address this, we are developing efficient computational tools, such as surrogate models and activation function-based axon models, while adapting our frameworks for high-performance computing. Additionally, we are integrating optimization techniques to effectively navigate this complex parameter space.