Movement-responsive deep brain stimulation for Parkinson's disease using a remotely optimized neural decoder

Area of Science:

• Neuroscience
• Biomedical Engineering
• Neurological Disorders

Background:

• Deep brain stimulation (DBS) is a common treatment for advanced Parkinson's disease.
• Conventional DBS (cDBS) uses fixed stimulation parameters, which may not address dynamic patient needs.
• Adaptive DBS (aDBS) offers a promising alternative by adjusting stimulation based on real-time physiological or behavioral states.

Purpose of the Study:

• To develop and evaluate a novel adaptive deep brain stimulation (aDBS) algorithm for Parkinson's disease.
• To mitigate movement slowness by delivering stimulation increases during movement using decoded motor signals.
• To assess the efficacy of movement-responsive aDBS compared to conventional DBS and a control condition.

Main Methods:

• An aDBS algorithm was designed to increase stimulation during movement based on decoded brain signals.
• The algorithm's performance was compared against an inverted control and conventional DBS (cDBS).
• A machine learning pipeline was developed for remote optimization of aDBS parameters in a home setting.

Main Results:

• The movement-responsive aDBS algorithm improved dominant hand movement speed and participant-reported therapeutic efficacy compared to the control.
• Typing speed increased and dyskinesia decreased with aDBS compared to cDBS.
• Proof of principle for remote, machine learning-assisted optimization of aDBS parameters was demonstrated.

Conclusions:

• Movement-responsive aDBS shows potential as a therapeutic strategy for Parkinson's disease, specifically targeting motor symptoms like slowness.
• This approach allows for dynamic alignment of therapy with patient-specific needs, potentially improving outcomes.
• Machine learning-assisted programming can simplify the optimization of aDBS, facilitating its clinical translation and scalability.