Overview
- The study published in Nature Medicine on Monday reports that AI-driven, activity-dependent DBS produced measurable improvements in walking and fewer falls in early tests.
- Researchers trained neural decoders on data from about 40 patients to detect locomotor states from brain activity and used those signals to guide stimulation changes within seconds.
- A UCSF team implanted personalized neural signatures into an approved neurostimulator and tested the system in five people, using blinded multi-day real-life testing that showed better gait symmetry and lower step variability when adaptive stimulation was active.
- No serious adverse events were reported and participants tolerated rapid stimulation shifts, but investigators stress the findings are preliminary and call for larger, longer trials to confirm durability, safety, and broader effectiveness.
- The work builds on decades of DBS treatment for Parkinson’s and, with industry collaboration including Medtronic, points toward future implants that could respond dynamically to movement, speech, mood, or cognition.