Research Frontiers | Paired Vagus Nerve Stimulation Sustains Stroke Rehabilitation Benefits for Two Years

（invasive vagus nerve stimulation）

Whether a person can once again hold a cup steadily or fasten a button often determines how independently they can live after a stroke. Once stroke becomes chronic, further recovery of arm and hand function is more difficult. Even when rehabilitation produces gains, patients and clinicians still want to know how long those gains will last. Invasive vagus nerve stimulation (invasive vagus nerve stimulation，iVNS) can deliver stimulation in synchrony with rehabilitation movements, but it requires surgical implantation and therefore carries practical barriers such as procedural burden and cost.

A 2026 follow-up study published in Neurology examined people with chronic ischemic stroke who had completed implanted paired stimulation therapy. The investigators tracked upper-extremity impairment, activity, participation, and quality of life to determine whether benefits persisted for 2 years; some participants were also assessed at 3 years. The central question was not whether treatment could produce a short-lived boost, but whether rehabilitation gains could endure in everyday life.

The vagus nerve contains both afferent and efferent fibers and supports two-way communication between the brain and multiple internal organs. In implanted vagus nerve stimulation, an electrode is surgically placed around the cervical vagus nerve. For stroke rehabilitation, brief bursts of stimulation are synchronized with training movements such as grasping, reaching, and extending the arm. Previous research suggests that this precise temporal pairing may recruit neuromodulatory systems and promote plastic changes in motor networks associated with the practiced movement. Paired VNS is therefore not a substitute for rehabilitation; it is intended to enhance how training shapes the nervous system. The present follow-up study assessed long-term clinical outcomes and did not directly test these neural mechanisms.

The key evidence comes from VNS-REHAB, a randomized, triple-blind, sham-controlled rehabilitation trial. It showed that implanted paired VNS combined with upper-extremity rehabilitation could improve motor function in people with chronic ischemic stroke, and a subsequent study reported that benefits were maintained for 1 year. Stroke, however, is a long-term condition. Whether early improvements on clinical scales translate into sustained arm use, participation, and quality of life - and whether they persist beyond 1 year - requires longer observation. The current study therefore conducted a post hoc analysis of 2-year data from the original trial and explored 3-year changes among participants with available data.

Conducted at 19 centers in the United States and the United Kingdom, the original trial enrolled 108 participants aged 22-80 years. All had a unilateral supratentorial ischemic stroke, an FMA-UE score of 20-50 out of 66, and some active wrist or finger movement. Every participant received an implanted device and was randomly assigned to active stimulation (n = 53) or sham stimulation (n = 55). Both groups completed 18 task-oriented outpatient sessions over 6 weeks, followed by 3 months of home training for 30 minutes per day.

After the blinded phase, the sham group crossed over to active stimulation, and both groups self-activated stimulation during home tasks. The analysis pooled all active-treatment participants against the original baseline; 49 completed the 2-year assessment. Linear mixed-effects models adjusted for age, sex, time since stroke, and side of paresis. Outcomes were the FMA-UE, the Wolf Motor Function Test-Functional Ability Scale (WMFT-FAS), and seven patient-reported measures of activity, participation, quality of life, and depression; Bonferroni correction accounted for multiple comparisons.

Figure 1. Study Timeline

After receiving active VNS, both groups entered long-term self-activated treatment and were assessed at 1 and 2 years; a subset was also assessed at 3 years.

Figure 2. Participant Flow Diagram

The original trial randomized 108 participants; 49 completed the 2-year assessment and were included in the pooled analysis.

At 2 years, the mean FMA-UE score among 49 participants had increased by 7.51 points from baseline (95% CI, 5.80-9.22; P < 0.001), and the mean WMFT-FAS score had increased by 0.63 points (95% CI, 0.50-0.75; P < 0.001). Both outcomes remained stable from year 1 to year 2. Based on the minimal clinically important difference, 37 participants (76%) achieved a clinically meaningful improvement on at least one measure.

The gains were not confined to laboratory-based scales. Of seven patient-reported outcomes, five improved significantly from baseline at 2 years: the Activities of Daily Living and Hand subscales of the Stroke Impact Scale, the Amount of Use and Quality of Movement scales of the Motor Activity Log, and the Stroke-Specific Quality of Life scale. The study found no significant association between changes in FMA-UE or WMFT scores and age, sex, time since stroke, or side of paresis.

Among the 16 participants who completed 3-year follow-up, the mean FMA-UE score remained 7.47 points above baseline (95% CI, 4.84-10.11; P < 0.001), and the mean WMFT-FAS score remained 0.69 points above baseline (95% CI, 0.50-0.89; P < 0.001). These subgroup findings indicate that improvements in upper-extremity motor function were still present at 3 years. However, the long-term follow-up phase had no concurrent control group and included only participants who completed follow-up. The observed long-term changes therefore cannot all be attributed to paired VNS on the basis of these data alone.

Figure 3. Long-Term Changes in Upper-Extremity Impairment and Activity

Upper panel: FMA-UE change; lower panel: WMFT change. Green points show pooled long-term results, and error bars show uncertainty.

This study primarily evaluated long-term clinical outcomes and did not include brain imaging, electrophysiology, or molecular measurements. It therefore cannot directly identify the neural circuits or biological mechanisms involved in maintaining treatment effects. Based on the principles of paired VNS, delivering stimulation as a rehabilitation movement occurs may enhance movement-related neural plasticity and help consolidate motor patterns formed through repeated practice. This interpretation comes mainly from previous research and was not directly tested in the present follow-up study.

The investigators placed greater emphasis on the possible contribution of long-term home training. Participants continued to self-activate stimulation outside the clinic and incorporated training into specific tasks such as holding a cup or picking up objects, potentially supporting renewed everyday use of the affected arm. Some participants achieved further gains meeting the minimal clinically important difference between years 1 and 2, while others reached that threshold for the first time. This pattern suggests that continued engagement in paired training may help preserve earlier improvements and may support further recovery in some individuals. However, this interpretation is based on follow-up trajectories and does not establish that long-term changes were caused entirely by stimulation or by any single mechanism. The study evaluated implanted paired VNS, and its findings cannot be directly extrapolated to transcutaneous auricular vagus nerve stimulation.

Future work could integrate implanted paired stimulation into a continuous stroke rehabilitation pathway linked to patient-prioritized goals such as dressing, eating, and writing. This would test whether scale improvements translate into greater independence at home and in the community. Device logs, wearable sensors, and telerehabilitation platforms could capture stimulation timing, training dose, and real-world arm use, helping refine individualized prescriptions and identify behaviors associated with sustained benefit. Direct comparisons and parameter studies, while clearly distinguishing implanted from transcutaneous stimulation, may support lower-burden neuromodulation and clarify its potential across stages of stroke and other motor disorders.

For clarity, Zenowell taVNS is a non-invasive form of vagus nerve stimulation and is distinct from the surgically implanted VNS technology evaluated in this study.

Kimberley, T. J., Vora, I., Cramer, S. C., Wolf, S. L., & Dawson, J. (2026). Two-year retention of benefits after paired vagus nerve stimulation in stroke: Follow-up of the VNS-REHAB randomized clinical trial. Neurology, 107(3), e218298.