Revolutionizing Parkinson’s Treatment: Neuromorphic Stimulation for Energy-Saving Precision.

Parkinson’s Disease (PD) affects millions worldwide, bringing tremors, stiffness, and movement challenges that disrupt daily life. Managing this condition remains a global health puzzle, as standard treatments like medications often lose effectiveness over time, leading to higher doses and side effects such as depression or impaired movement. Deep Brain Stimulation (DBS) offers hope by delivering electrical pulses through electrodes implanted in the brain to ease symptoms. However, traditional DBS—known as open-loop DBS—has drawbacks. It uses rigid stimulation patterns that don’t adjust to a patient’s changing needs, drains batteries quickly (requiring frequent surgical interventions for device replacement), and can cause unwanted side effects like mood changes. Closed-Loop DBS (CL-DBS) steps in as a smarter alternative, adapting stimulation in real time by monitoring brain signals, such as beta oscillations in the subthalamic nucleus (STN)—a brain area tied to PD symptoms. Yet, its high power demands make it tough to use in small, implantable devices. This research, led by Biswas et al. (2025) and published in IEEE, introduces a brain-inspired, energy-saving approach using Leaky Integrate-and-Fire (LIF) controllers to make CL-DBS practical, promising better symptom relief and fewer surgeries for PD patients.

The Neuromorphic Solution: Leaky Integrate-and-Fire (LIF) Controllers

This study reveals a brain-inspired CL-DBS system that uses Leaky Integrate-and-Fire (LIF) controllers to adjust stimulation dynamically. Two types of these controllers are introduced. The on-off LIF controller and the dual-threshold LIF controller. These act like smart dials, continually tweaking electrical pulses by tracking beta oscillations, a brain signal linked to PD tremors to adapt to symptom severity. The on-off controller switches stimulation on or off as needed, while the dual-threshold controller uses two levels for finer adjustments, increasing or decreasing the pulse strength depending on whether the brain signal crosses an upper or lower limit. While designed for future hardware integration, the LIF controllers were initially implemented and tested through simulations on computational models, leveraging software to validate the conceptual approach. Inspired by how neurons process information, this neuromorphic design saves energy and handles tasks efficiently, mimicking the brain’s natural, low-power parallel processing—perfect for the tight energy constraints of implanted devices. This fits into a broader landscape of neuromorphic computing in healthcare, including epilepsy monitoring and prosthetic control, where brain-like systems are gaining traction for their efficiency and adaptability. Other players, like MintNeuro, are developing tiny, low-power chips to shrink brain implants, while Starfish Neuroscience, co-founded by Gabe Newell, is working on a 2x4mm, 1.1 mW brain chip for PD, showing a competitive field pushing for less invasive, efficient solutions.

Key Contributions & Results

The standout achievement here is energy efficiency. Convventional open-loop DBS systems consume around 50 milliwatts due to constant stimulation, but the on-off LIF controller drops this to 40.5 milliwatts (a 19% reduction), while the dual-threshold LIF controller cuts it to 22 milliwatts (a 56% reduction)—a major leap that could mean fewer battery replacements and surgeries. Treatment effectiveness also improves. The on-off LIF boosts “suppression efficiency” by 4.7%, and the dual LIF by 6.77%. Suppression efficiency measures how well the system calms the abnormal brain activity causing PD symptoms per unit of power—essentially, it delivers more relief with less energy. The study also addresses a data gap by creating a new PD dataset, modeling neural activity and beta oscillations in the STN, using computational models to simulate local field potentials (LFP)—combined electrical signals from groups of neurons. This dataset, made publicly available through the Parkinson’s Electrophysiological Signal Dataset (PESD) at https://dandiarchive.org/dandiset/001333/0.250304.0304, includes raw brain signals from both PD patients and healthy individuals, empowering researchers to refine neuromorphic approaches. Performance was tested with metrics like Mean Squared Error (MSE)—how close the system hits the target brain signal (11% for on-off, 30% for dual vs. 100% without DBS), power consumption, and suppression efficiency, showing clear gains. For the neuromorphic community, the system’s efficiency, achieving around 500 synaptic operations per joule, sets a promising benchmark for energy-efficient spike processing, indicating its high computational power for minimal energy—a key metric for advancing brain-inspired tech.

Broader Impact & Future Outlook

This research makes CL-DBS a more viable option for PD treatment by tackling its energy challenges, laying the foundation for devices that work longer and more effectively inside the body. It opens the door to personalized treatments, where stimulation adapts to each patient’s unique brain signals, and responsive therapies that adjust as symptoms change over time. With edge AI, the system could learn these patterns on-device, processing data locally without sending it to external servers, enhancing privacy and speed. Integration with digital health ecosystems—via smartphone apps for remote monitoring or data syncing—could further transform care, fitting into the growing digital health revolution. Beyond PD, this approach could extend to treating epilepsy, depression, or other conditions needing brain stimulation, leveraging its adaptive, low-power design. For people with PD, this could translate to fewer surgeries, better symptom management, and an improved quality of life. However, scaling to clinical use faces hurdles: in the U.S., regulatory approvals from the FDA require extensive safety and efficacy trials, potentially taking 5-10 years. In Europe, the European Medicines Agency (EMA) imposes similar rigorous standards, often requiring clinical trials for high-risk devices like neuromodulators, adding to the timeline. Manufacturing challenges include miniaturizing neuromorphic chips and ensuring reliable, biocompatible materials, which could add 2-3 years. This research fundamentally advances the feasibility of CL-DBS by offering unprecedented energy efficiency and adaptive precision. While clinical adoption faces regulatory and manufacturing hurdles, the groundwork for tailored neuromodulation therapies is now firmly established. The public release of the Parkinson’s Electrophysiological Signal Dataset (PESD) serves as an open invitation to the global neuromorphic and medical research communities to collaborate, innovate, and accelerate the development of these transformative treatments for Parkinson’s Disease.

A Call for Collaboration: The Open Dataset and Benchmarks

The Parkinson’s Electrophysiological Signal Dataset (PESD) at https://dandiarchive.org/dandiset/001333/0.250304.0304 fosters a collaborative research environment, enabling the neuromorphic community to refine LIF-based systems, test new applications, or improve energy efficiency. With spike rates of 100 Hz and 500 synaptic operations per joule, these benchmarks provide a foundation for comparing with other neuromorphic platforms, sparking further advancements in medical device technology.

Citation

[1] A. Biswas, M. Akhtaruzzaman, H. An, “Energy-Efficient Neuromorphic Closed-Loop Modulation System for Parkinson’s Disease,” 2025 26th International Symposium on Quality Electronic Design (ISQED), IEEE, 2025, doi: 10.1109/ISQED65160.2025.11014351.

Related Articles

PHP Code Snippets Powered By : XYZScripts.com
★

Track This Company

Create a free account to add companies to your watchlist and get notified when patents, funding, or key developments change.