Multi-timescale neuromodulation strategy for closed-loop deep brain stimulation in Parkinson's disease

Area of Science:

• Computational neuroscience
• Neuromodulation
• Biomedical engineering

Background:

• Beta oscillations are key indicators in Parkinson's disease.
• Current closed-loop deep brain stimulation (DBS) faces challenges with neural dynamics.
• Tuning DBS requires accounting for multi-timescale neural variations.

Purpose of the Study:

• To develop a closed-loop DBS strategy for modulating beta oscillation amplitude across different timescales.
• To incorporate inherent variations of the basal ganglia-thalamus-cortical circuit into DBS control.
• To ensure the strategy respects clinical constraints on stimulation parameters.

Main Methods:

• A dynamic mean-field model of the basal ganglia-thalamus-cortical circuit was created.
• A dynamic target model was designed to represent multi-timescale beta power variations (milliseconds to minutes).
• A proportional-integral-differential (PID) controller with a dynamic target was implemented for closed-loop DBS, considering stimulation amplitude bounds.

Main Results:

• The model accurately simulated medication rhythms and captured beta power dynamics across timescales.
• The proposed closed-loop DBS strategy achieved smoother stimulation compared to constant-target approaches.
• Beta power successfully tracked the dynamic target under various modulation strengths and clinical constraints.

Conclusions:

• This study presents a novel closed-loop DBS approach for precise, multi-timescale beta power modulation.
• The method effectively addresses the complexities of neural dynamics and clinical limitations in Parkinson's disease treatment.
• The findings offer a promising advancement for optimizing DBS therapy efficacy and safety.