<ul><li>Researchers have developed a data-driven Model Predictive Control (MPC) algorithm for deep brain stimulation (DBS) in the treatment of Parkinson's disease (PD).</li><li>Closed-loop DBS (CLDBS) utilizes neural oscillations as a feedback signal, resulting in improved treatment outcomes and reduced side effects compared to open-loop DBS.</li><li>The proposed algorithm uses a multi-step predictor based on input-convex neural networks to model the future evolution of beta oscillations, improving prediction accuracy and simplifying online computation.</li><li>Through simulations and tests with PD patients, the algorithm achieved significant reductions in tracking error and control activity compared to existing CLDBS algorithms, offering a potential advancement in DBS treatment for PD and other diseases.</li></ul>

Deep Learning Model Predictive Control for Deep Brain Stimulation in Parkinson's Disease

Discover more