Neuromorphic Chip Cuts Deep Brain Stimulation Power by 75% for Parkinson's

Parkinson's disease (PD) affects millions worldwide, causing severe motor symptoms. Traditional deep brain stimulation (DBS) delivers constant, open-loop pulses, wasting energy and failing to adapt to symptom fluctuations. Adaptive DBS (aDBS) addresses this by using physiological biomarkers to adjust stimulation in real time, but most implementations remain software-based with limited circuit-level progress. Now, researchers from a team including Md Abu Bakr Siddique, Jakub Orłowski, Yan Zhang, and Hongyu An have introduced the Silicon Leaky Integrate-and-Fire Deep Brain Stimulation (SiLIF-DBS) controller—a neuromorphic silicon neuron stimulator built with standard CMOS technology. This chip is designed for low-power, implantable aDBS, using beta-band subthalamic local field potentials (STN-LFPs) as its control biomarker, specifically the average rectified value (Beta ARV) to drive closed-loop stimulation.

The SiLIF-DBS controller was validated within a Parkinsonian cortico-basal ganglia computational model. Key results: it suppresses pathological beta activity while consuming only 25% of the power required by open-loop stimulation, with a suppression efficiency of 5.85%/μW. This dramatic power reduction is critical for implantable devices where battery life and heat dissipation are major constraints. The chip's neuromorphic architecture mimics biological neurons to efficiently compute stimulation parameters, offering a hardware foundation for next-generation aDBS systems. While still in the modeling stage, the work demonstrates a viable path toward truly intelligent, low-power neural prosthetics that could significantly improve quality of life for Parkinson's patients.