AI-DBS: Applying Personalised Deep Brain Stimulation using ‘Neuronal Fingerprints’
This project brings together a new public-private partnership between Amsterdam UMC, Medtronic, and Runelabs to improve deep brain sJmulaJon (DBS) for people with Parkinson’s disease. The goal is to develop a smarter, adapJve form of DBS that automaJcally adjusts sJmulaJon based on real-Jme brain acJvity. This innovaJon could significantly reduce side effects, improve symptom control, and lessen the burden on both paJents and healthcare providers.
Parkinson’s disease affects over 50,000 people in the Netherlands and is expected to double in prevalence by 2040. DBS is a proven and effecJve treatment for paJents with advanced disease, but the current programming process is Jme-intensive, requires frequent hospital visits, and sJll relies on fixed sJmulaJon se^ngs. These fixed se^ngs do not account for symptom fluctuaJons throughout the day, which limits treatment efficacy and increases the risk of side effects. As healthcare costs and workforce shortages conJnue to rise, innovaJons that increase efficiency and autonomy for paJents are urgently needed.
Our approach uses a novel DBS system that can both sJmulate and record neuronal acJvity in the brain. One hundred Parkinson’s paJents will receive this system. Over the first six months, brain acJvity will be recorded and linked to symptoms measured through smartwatch movement data, clinical assessments, and paJent diaries. Using arJficial intelligence, we will idenJfy individual “neuronal fingerprints” of symptom severity. These paVerns will then be used to guide real-Jme, personalized DBS in a later phase of the study.
The key deliverables of this project include validated paJent-specific brain acJvity profiles, a potenJal algorithm for adapJve DBS, and clinical data comparing convenJonal and intelligent DBS. This project lays the foundaJon for a future where DBS therapy is more personalized, less dependent on hospital visits, and supports greater autonomy and quality of life for paJents.
