Summary: Researchers at UC San Francisco have developed a closed-loop adaptive deep brain stimulation (aDBS) system that senses and responds to walking in real time. Unlike conventional continuous DBS, which delivers a constant pattern of stimulation, this implantable device reads neural signals linked to each left and right step and adjusts stimulation within fractions of a second, working in synchrony with the patient’s natural gait.
By embedding predictive algorithms into the implanted neurostimulator, the system monitors individualized electrical signatures from both sides of the body during each stride. Operating autonomously and without an external computer, the implant modifies its output on a step-by-step basis to support stable, symmetric walking and reduce falls.
Key facts
- Stride-by-stride pacemaker analogy: The device functions like a cardiac pacemaker for walking. Instead of pacing the heart, it continuously tracks and responds to the brain’s rhythm of gait, calculating stimulation bursts that align with each phase of movement.
- Limitations of continuous DBS: Although continuous DBS benefits tremor, rigidity, and slowness in Parkinson’s disease, it often fails to address gait impairment, freezing of gait, and falls because it cannot match the rapid timing demands of walking.
- Bilateral left–right neural mapping: The research team identified individualized neural signatures associated with lifting and planting the left and right foot. These step-specific patterns were embedded in the stimulator to drive millisecond-level adjustments.
- Laboratory and real-world validation: In-clinic testing showed improved gait symmetry and reduced variability. A subsequent blinded, multi-day crossover trial in participants’ homes demonstrated fewer falls while overall symptom control remained stable and no serious adverse events were reported.
- Dual cortical–subcortical recording array: Trial participants received both deep subcortical stimulation leads and research electrodes over movement-related cortical regions, enabling clear detection of intentional movement signals and simultaneous stimulation.
- Behaviorally driven neuromodulation: Whereas prior adaptive therapies typically respond to slow physiological changes, this approach ties stimulation directly to dynamic behavior, adjusting on millisecond timescales to support real-time actions.
- Potential for personalized therapies: Because the system demonstrates that implanted devices can sense and react to ongoing behavior, researchers anticipate extensions to conditions such as speech disorders, treatment-resistant depression, and cognitive decline.
Source: UCSF
Study overview
Published June 15 in Nature Medicine, the study reports the first implanted brain stimulator that detects neural signals linked to individual steps and automatically synchronizes stimulation to gait phase. The device recognizes unique cortical or subcortical patterns associated with intent to move the left or right leg, then delivers time-locked adjustments that promote more stable walking.
“Difficulty walking is one of the most disabling symptoms of Parkinson’s disease and one of the hardest to treat,” said Doris D. Wang, MD, PhD, associate professor of neurological surgery at UCSF and senior author. She emphasizes that walking is highly dynamic and requires precise bilateral timing; the new system recognizes those movement patterns and responds in real time so stimulation works with the patient’s movements.
How the adaptive system works
The personalized aDBS approach identifies neural biomarkers for left and right leg movement using both deep brain leads and cortical research electrodes. Those biomarkers are programmed into a bidirectional implanted neurostimulator so that it can sense gait phase and modulate stimulation without an external processor. The embedded algorithms operate within the sub-second timeline of stepping, enabling stimulation to align with each stride.
In-clinic testing documented improvements in step symmetry and reduced step-to-step variability—both indicators of a more efficient, stable gait. In a following blinded, multi-day crossover phase conducted in participants’ daily lives, the adaptive mode reduced falls while maintaining control of other Parkinsonian motor symptoms. Participants tolerated rapid stimulation adjustments without serious adverse events.
Clinical significance and future directions
Gait dysfunction is a major source of disability in Parkinson’s disease and is often inadequately treated by continuous DBS. This study shows that synchronizing neuromodulation to behavior—specifically to gait phase—can be feasible, safe, and potentially more effective at reducing falls. The results support moving toward larger, multi-center randomized trials to evaluate clinical efficacy at scale.
Beyond gait, the study demonstrates a broader principle: implanted devices can continuously sense neural activity and deliver precisely timed, personalized therapy only when it is needed. Researchers foresee rapid expansion of behavior-contingent neuromodulation to treat speech disturbances, mood disorders, and cognitive symptoms by aligning stimulation with relevant neural and behavioral markers.
Study details and contributors
This randomized, blinded crossover feasibility trial involved five participants with Parkinson’s disease who underwent pallidal DBS plus subdural paddle electrodes for research recordings. Personalized gait-phase biomarkers were identified in all participants and embedded into the neurostimulator. Acute testing showed improvements versus continuous DBS; a multi-day blinded crossover phase confirmed reduced falls and participant-specific gait gains. No adverse events were reported.
Key authors (UCSF): Kenneth H. Louie, PhD; Jannine P. Balakid, BS; Jessica E. Bath, DPT, PhD; Seongmi Song, PhD; Hamid Fekri Azgomi, PhD; Jacob H. Marks, BA; Philip A. Starr, MD, PhD; Doris D. Wang, MD, PhD. Additional author: Julia T. Choi, PhD (University of Florida).
Funding: Supported by the Michael J. Fox Foundation (Grant MNS135499A), UCSF Burroughs Wellcome Fund Career Award for Medical Scientist, NIH/NINDS (1R01NS130183), and UCSF Catalyst Grants. Funding was obtained by D.D.W.
Disclosures: D.D.W. consults for Medtronic, Boston Scientific, and Iota Bioscience and receives research support from Boston Scientific. P.A.S. receives support from Medtronic and Boston Scientific for fellowship education. K.H.L. is employed by Echo Neurotechnologies; this work was completed prior to that employment and the company had no role in study design, data collection, analysis, or publication decisions.
Key questions answered
A: Walking is a high-speed, bilateral behavior that needs millisecond-level coordination between brain, spinal cord, and muscles. Conventional DBS delivers a fixed, continuous stimulation pattern that cannot adapt to rapid, step-by-step changes, so it often cannot restore the precise timing needed for safe, fluent walking.
A: Research electrodes capture movement-related field potentials that form reproducible, step-specific neural signatures for left and right leg movements. The implant’s embedded algorithms recognize those signatures in real time and trigger phase-matched stimulation adjustments.
A: This is an early, promising clinical feasibility study. The trial demonstrated safety and real-world functioning in a small cohort, creating a rationale for larger, multi-center trials required for regulatory approval. Widespread availability will depend on the outcomes and timing of those subsequent studies.
Editorial notes
- Edited and reviewed by Neuroscience News editorial staff.
- Journal paper reviewed in full by the editors.
- Additional contextual reporting provided by editorial staff.
About this research news
Author: Brooke Thornton
Source: UCSF
Contact: Siyun Qin – Brooke Thornton
Image credit: Neuroscience News
Original research: Adaptive Deep Brain Stimulation for Dynamic Gait Control in Parkinson’s Disease: a randomized feasibility trial. DOI: 10.1038/s41591-026-04434-2. Open access.
Abstract
Adaptive Deep Brain Stimulation for Dynamic Gait Control in Parkinson’s Disease: a randomized feasibility trial
A randomized crossover study of five patients with Parkinson’s disease demonstrates that gait-synchronized adaptive deep brain stimulation is feasible and safe, and reduces falls compared with continuous stimulation. Personalized gait-phase biomarkers were identified from cortical or pallidal field potentials and embedded in a bidirectional neurostimulator. Acute in-clinic testing showed improvements in step variability and symmetry versus continuous DBS. A subsequent blinded, multi-day crossover phase in daily life reduced falls while preserving general motor symptom control. No serious adverse events occurred. These findings support the development of larger randomized trials to determine clinical efficacy. ClinicalTrial.gov registration: NCT04675398.