Wearable Myoelectric Interface for Neurorehabilitation (MINT) to Recover Arm Activity After Stroke: A Randomized Controlled Trial

Background: Abnormal muscle co-activation impairs arm function after stroke. Home-based myoelectric interface for neurorehabilitation (MINT) was developed to reduce abnormal co-activation. Objective: Evaluate MINT feasibility and effects on arm activity in a single-blind, randomized, sham-controlled trial. Methods: Moderately-to-severely impaired chronic stroke survivors were randomized to 1 of 3 MINT groups (who played customized games requiring independent activation of 2 or 3 abnormally co-activating muscles) or a sham control group (using 1 muscle). Participants trained 90 minutes/day, 6 days/week, switching upper-arm muscle sets every 2 to 3 weeks. Feasibility was assessed using training time, repetitions, and motivation. The primary efficacy outcome was the change in Wolf Motor Function Test (WMFT) from baseline to week 6 in each MINT group compared to sham. Secondary outcomes included WMFT at week 10, WMFT subscores, Fugl–Meyer, Modified Ashworth, Motor Activity Log, reaching kinematics, co-activation, and lesion location effects. Results: Fifty-nine participants completed training (mean: 86 ± 21 minutes/day, 315 ± 85 repetitions/day). The 3-muscle group improved WMFT by 6.8 seconds at week 6 (95% confidence interval [2.1-11.6 seconds], P = .001); other groups did not. Combined MINT groups improved by 4.1 seconds ([1.4-6.8 seconds], P = .0008). In per-protocol analysis, the 3-muscle group improved significantly more than sham on WMFT (7.5 seconds; [0.1-14.9 seconds], P = .046), shoulder/elbow subscores, and kinematics. At week 10, all MINT groups improved significantly, unlike sham. Reduced co-activation correlated significantly with improved arm function and movement. Stroke involving the internal capsule negatively predicted MINT response. Other secondary outcomes showed no clinically important changes. Discussion: MINT conditioning is feasible, reduces co-activation, and improves arm movement and activity. Clinical Trial Registration: ClinicalTrials.gov (NCT03401762).