Supporting everyday movement in chronic hemiparesis

Electrical stimulation of the cervical spinal cord improved arm strength and motor function in people with chronic stroke. The finding comes from a seven-person feasibility trial conducted by the University of Pittsburgh, Carnegie Mellon University, and collaborators. The key effect was immediate assistance when stimulation was on. This should be distinguished from evidence of a cure for paralysis or regeneration of damaged neural circuits. The paper was published in Nature Medicine on June 4, 2026.

Hemiparesis affecting the arm and hand after stroke involves more than reduced strength: it can also cause difficulty with fine movements, spasticity, and unintended simultaneous contraction of multiple muscles. High-dose motor rehabilitation can help, but the amount of training available in routine care is often insufficient. The researchers examined whether cervical epidural spinal cord stimulation (SCS) could serve as a neuroprosthetic aid during everyday movements.

Mean strength rose 32% in stimulation ON/OFF comparisons

The study was a single-arm, open-label trial without a control group. The seven participants who completed it were 2–10 years post-stroke, with baseline upper-extremity Fugl-Meyer Assessment (FMA) motor scores of 15–35. The FMA measures motor impairment of the arm and hand; the maximum upper-extremity motor score is 66.

The researchers implanted two electrode leads, each with eight contacts, in the cervical epidural space of each participant and connected them to an external stimulator. The electrodes were removed after 29 days. Participants underwent 19 assessments over four weeks, without a prescribed motor rehabilitation program. They performed a total of 8.6 hours of motor activity during assessments, including 5.5 hours with stimulation on.

SCS aims to assist voluntary movement by stimulating sensory afferent nerve fibers that connect to motor neurons controlling the upper limbs. The researchers adjusted electrode contacts and stimulation intensity and frequency while monitoring muscle responses and participants’ movements. Selective stimulation was important because activating inappropriate muscles could interfere with movement.

Within the same assessment sessions, maximum voluntary contraction (MVC) strength averaged 32% higher with stimulation on than off across movements at multiple joints. This does not represent a recovery rate relative to baseline or a percentage improvement in all daily functions. Reaching trajectories also became straighter and smoother. However, assistance with hand opening and individual finger movements was observed only in three participants who had residual finger movement or detectable motor-evoked potentials.

At four weeks, FMA motor scores were an average of 5.14 points higher than before implantation with stimulation off and 6.57 points higher with it on. The change with stimulation off reflects improvement observed over time in the off condition; the change with stimulation on also includes the immediate assistive effect. In addition, every participant had a clinically meaningful reduction in spasticity in at least one muscle group at four weeks with stimulation off, compared with baseline.

Continued use and safety require further evaluation

There were no serious adverse events, but 14 mild adverse events were recorded. One participant experienced shortness of breath when stimulation was delivered at 5 mA near C3–C4; it resolved immediately after stimulation stopped. The researchers suggested unintended activation of the phrenic nerve, which contributes to diaphragm movement, as a possible explanation and restricted stimulation conditions at higher cervical levels. Because the approach requires surgical electrode implantation, long-term safety needs separate evaluation.

One month after electrode removal, FMA and upper-limb function assessment scores were lower than at the four-week stimulation-on assessment. This suggests that continued use may be needed to maintain the maximum effect. The observation that participants with better-preserved sensory function responded more favorably is exploratory.

The small sample, lack of a control group, unblinded participants, and short assessment period make it difficult to establish the size and durability of the effect. Assessors did not know the stimulation condition, but participants could identify it from the sensations caused by stimulation. In particular, improvement observed over time with stimulation off cannot be attributed to SCS alone. The next step is to assess real-world arm use in larger controlled trials and long-term implant studies.

The study design and results are available in the Nature Medicine paper.