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

2025-04-01

AI Summary

To address low control accuracy, slow adaptation, and pronounced side effects in closed-loop deep brain stimulation (CLDBS) for Parkinson’s disease, this work proposes a novel deep learning–based nonlinear model predictive control (MPC) paradigm. Methodologically, we introduce a differential input-convex neural network (ICNN)-structured multi-step beta-band oscillation predictor—uniquely integrating neural dynamics modeling with online optimization efficiency—and develop a data-driven, generalizable closed-loop control framework that overcomes the limitations of conventional proportional/integral controllers. Validated on both simulated and real patient neurophysiological data, our approach reduces tracking error and stimulation energy consumption by over 20%, significantly enhancing control precision and delaying the onset of stimulation tolerance. To the best of our knowledge, this is the first deep learning solution for CLDBS that jointly achieves high-fidelity neural signal prediction and real-time MPC optimization.

Technology Category

• Intelligent Robots: Learning & Optimization for ROB
• Cognitive Modeling & Cognitive Systems: Neural Spike Coding
• Machine Learning: Online Learning & Bandits

Application Category

• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

We present a nonlinear data-driven Model Predictive Control (MPC) algorithm for deep brain stimulation (DBS) for the treatment of Parkinson's disease (PD). Although DBS is typically implemented in open-loop, closed-loop DBS (CLDBS) uses the amplitude of neural oscillations in specific frequency bands (e.g. beta 13-30 Hz) as a feedback signal, resulting in improved treatment outcomes with reduced side effects and slower rates of patient habituation to stimulation. To date, CLDBS has only been implemented in vivo with simple control algorithms, such as proportional or proportional-integral control. Our approach employs a multi-step predictor based on differences of input-convex neural networks to model the future evolution of beta oscillations. The use of a multi-step predictor enhances prediction accuracy over the optimization horizon and simplifies online computation. In tests using a simulated model of beta-band activity response and data from PD patients, we achieve reductions of more than 20% in both tracking error and control activity in comparison with existing CLDBS algorithms. The proposed control strategy provides a generalizable data-driven technique that can be applied to the treatment of PD and other diseases targeted by CLDBS, as well as to other neuromodulation techniques.

Problem

• Develops MPC for closed-loop deep brain stimulation in Parkinson's.
• Improves beta oscillation prediction accuracy using neural networks.
• Reduces tracking error and control activity by 20%.

Innovation

• Nonlinear data-driven MPC for DBS in PD
• Multi-step predictor with convex neural networks
• 20% better tracking and control than existing