Overview
- A Nature Medicine paper published Monday, June 15, 2026, reports that researchers in Lausanne and teams at UCSF developed adaptive deep brain stimulation that responds to real-time locomotor signals.
- The system uses machine-learning neural decoders trained on data from about 40 patients to detect walking states and adjust stimulation within seconds.
- In blinded, multi-day real-life testing at UCSF with five previously implanted patients, adaptive stimulation improved gait symmetry, cut step-to-step variability, and was associated with fewer falls.
- Investigators embedded movement signatures, including left‑ and right‑leg phase signals, directly into implanted neurostimulators and worked with Medtronic to adapt existing device hardware for automatic, on‑device adjustments.
- No serious adverse events were reported in these small studies, but researchers and clinicians say larger, longer trials are needed to confirm benefits, assess long‑term safety, and determine which patients would gain the most.