A five-patient feasibility study found that a form of deep brain stimulation programmed to adjust its output with each step of walking reduced falls and improved gait measures in people with Parkinson’s disease, compared with the standard approach of continuous stimulation.
Why this question mattered
Walking difficulties are among the most disabling features of Parkinson’s disease. They reduce independence, increase injury risk from falls, and tend to worsen as the disease progresses. Standard deep brain stimulation (DBS) electrical pulses delivered to specific brain regions via implanted electrodes can reduce tremor, stiffness, and slowness, but its benefits for gait are inconsistent and often diminish over years. Attempts to tune conventional DBS for walking, including lowering stimulation frequency or shifting electrode location, have produced mixed results, sometimes improving gait at the cost of worse tremor control.
What was already known
Researchers have known for some time that brain signals in the deep nuclei targeted by DBS particularly a structure called the globus pallidus internus (GPi) fluctuate in patterns that reflect a person’s physical state: sleep, medication levels, and movement all leave detectable signatures in local field potentials. A class of devices called adaptive DBS (aDBS) systems, which modulate stimulation in response to these shifting signals, has shown promise for controlling general motor symptoms. Most aDBS work to date has used slow-changing signals tied to disease state rather than moment-to-moment movement. Applying that logic to gait is more technically demanding: walking is cyclical and fast, with each stride lasting roughly a second, so useful control signals need to be identified and acted upon at sub-second timescales.
How the study was conducted
Each participant underwent surgery to implant bilateral electrodes in the GPi and subdural cortical electrode paddles. An investigational bidirectional neurostimulator the Medtronic Summit RC+S was connected to both the therapeutic DBS lead and the cortical paddle, allowing simultaneous recording and stimulation. Researchers recorded local field potentials and gait data during overground walking at self-selected speeds for at least 250 steps, both before stimulation was activated and after DBS parameters had been clinically optimized.
Using each participant’s own neural data, the team developed an aDBS therapy that modulated stimulation amplitude between 0.5 and 1.0 times the clinically optimized setting during contralateral leg swing. Control signals came from single-frequency band power at one recording site per hemisphere, selected through a data-driven search for features that could reliably distinguish the swing phase from other phases of the gait cycle.
For three of the five participants, a double-blind crossover experiment was then conducted, randomizing three conditions: continuous DBS, aDBS ramping up from half to full amplitude during contralateral leg swing, and aDBS ramping down from full to half amplitude during that phase. The multi-day crossover was carried out in participants’ daily lives, not only in the clinic.
What the results showed
Researchers successfully identified personalized gait-phase biomarkers from cortical or pallidal field potentials in all five participants and embedded them into the bidirectional neurostimulator. During acute in-clinic testing, aDBS improved step variability and step symmetry compared with continuous DBS.
In the blinded, multi-day portion of the study, participants experienced fewer falls while the adaptive system was active and maintained overall control of Parkinson’s symptoms. No serious adverse events occurred, and patients tolerated the rapid stimulation adjustments well.
The improvements in gait were patient-specific that is, the particular measures that improved differed across individuals which the authors noted as consistent with the personalized nature of the biomarker identification process.
What the researchers concluded
The authors concluded that the findings establish the feasibility of biomarker-driven, movement-synchronized neuromodulation and support the development of a larger randomized trial to determine clinical efficacy.
“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. “Walking is a highly dynamic behavior that requires precise timing across both sides of the body. We developed a system that can recognize those movement patterns and respond in real time, effectively allowing the stimulation to work with the patient as they move.”
First author Kenneth H. Louie, PhD, a UCSF postdoctoral scholar, noted: “The brain contains remarkably rich information about movement. We found that we could identify neural signatures linked to each step and use them to guide stimulation in real time.”
Limitations
The study enrolled five participants at a single center; only three completed the multi-day blinded crossover phase. The sample is too small to draw conclusions about clinical efficacy, and the authors explicitly frame this as a feasibility trial rather than an efficacy trial. All participants used an investigational device that is not commercially available. The subdural cortical electrode paddles implanted as part of this research protocol are not part of standard clinical DBS surgery, which adds procedural complexity. Because gait improvements were patient-specific and not uniform, the findings cannot be generalized to a typical DBS population. Longer-term follow-up data are not yet available. The researchers note that conflicts of interest exist: the senior author consults for Medtronic and Boston Scientific, and the first author is now employed by a neuromodulation company though that employment began after the study was completed.
Reference
Kenneth H. Louie, Jannine P. Balakid, Jessica E. Bath, Seongmi Song, Hamid Fekri Azgomi, Jacob H. Marks, Julia T. Choi, Philip A. Starr, Doris D. Wang. “Adaptive deep brain stimulation for dynamic gait control in Parkinson’s disease: a randomized feasibility trial.” Nature Medicine, June 15, 2026. DOI: 10.1038/s41591-026-04434-2
