Decoding and Restoring Movement in Parkinson’s Disease

## Stimulation with and without feedback **Deep brain stimulation**, or **DBS**, delivers electrical pulses through implanted leads to modulate circuit activity. Conventional DBS can improve motor symptoms in carefully selected people with Parkinson’s disease. It does not replace lost neurons or establish a cure. Conventional continuous DBS is generally **open loop**: a clinician sets parameters, and stimulation does not automatically follow moment-to-moment brain activity or movement. “Continuous” does not mean that settings can never be changed; it means the delivered stimulation is not being continuously adjusted by the measured feedback signal. **Adaptive DBS**, or **closed-loop DBS**, adds a repeated cycle: **sense → analyze → decide → stimulate → sense again** The input might be a neural biomarker, a movement measurement or a combination. The controller must distinguish a meaningful change from noise or recording artifact. Feedback does not by itself make a system safer or more effective; calibration and comparison with appropriate controls remain essential. ## A beta-burst-driven gait study Wilkins, Bronte-Stewart and colleagues (2025) tested an investigational beta-burst-driven controller in seven people with Parkinson’s disease. Implanted STN signals streamed to a computer. If an ongoing beta burst exceeded a participant-specific duration threshold, stimulation ramped upward; shorter bursts led to a downward ramp. Output remained within an individually calibrated therapeutic window, with a minimum stimulation level that still provided benefit. This describes the study’s method, not instructions for programming anyone’s device. Researchers had to calibrate the signal, threshold, stimulation limits and tolerable changes separately for each participant. Changes in stimulation can distort sensing, creating another reason to interpret feedback carefully. ## What the comparison supports Double-blind testing included stimulation off, adaptive DBS, continuous DBS and randomly adapting stimulation. Participants performed harnessed stepping in place, a turning-and-barrier course and other motor tests. The random condition changed output within a similar window without following the neural biomarker, helping distinguish feedback from simply varying stimulation. Adaptive stimulation was feasible and tolerated during testing. Compared with stimulation off, freezing time decreased and leg swing speed increased. Benefits were similar to continuous DBS; random stimulation also performed similarly at the group level. The study therefore did not establish superiority of adaptive over continuous DBS. The test involved a small group, hours in a laboratory and a computer-in-the-loop research system. Freezing was not eliminated for everyone. Longer studies outside the laboratory are needed to test persistence and generalization. The translational step is that a person’s measured brain signal can guide stimulation during movement. That is a useful result without turning feasibility into universal clinical efficacy. ## Source trail See `sources/bs2-2.md` and the claim audit, especially Wilkins et al. (2025), DOI: 10.1093/braincomms/fcaf266.