Multivariable closed-loop control of deep brain stimulation for Parkinson's disease

Area of Science:

• Neuroscience
• Biomedical Engineering
• Control Systems

Background:

• Current closed-loop deep brain stimulation (DBS) for Parkinson's disease (PD) typically controls a single biomarker, leading to suboptimal regulation for uncorrelated symptoms or varying biomarker-symptom relationships.
• The control of stimulation-induced side effects is often not addressed in existing DBS methods.

Purpose of the Study:

• To present a multivariable control architecture for selectively suppressing tremor or subthalamic nucleus beta band oscillations in PD.
• To modulate DBS pulse amplitude and duration to maintain amplitude below a threshold, avoiding stimulation of distal axons and associated side effects.
• To investigate a computational model of the PD motor network to simulate the proposed control architecture.

Main Methods:

• A multivariable control architecture was developed, modulating DBS pulse amplitude and duration.
• A supervisor controller selected between tremor or beta activity control based on detected muscle electromyographic (EMG) activity.
• A secondary controller limited pulse amplitude and modulated pulse duration to target smaller diameter axons, preventing side effects.

Main Results:

• The system achieved good control of both rest tremor and beta activity, with reduced power consumption compared to open-loop stimulation.
• The supervisor controller prevented over- or under-stimulation issues encountered with single-biomarker controllers.
• When DBS amplitude was limited, the secondary controller maintained stimulation efficacy by adjusting pulse duration, demonstrating dual-parameter control benefits.

Conclusions:

• Non-linear multivariable control offers targeted suppression of motor symptoms in PD patients.
• Dual-parameter control enables automatic regulation of DBS therapeutic dosage, preventing overstimulation and improving power efficiency.