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

Only 59 of the 94 randomised participants finished the six weeks and went into the main analysis. The 35 who withdrew before week 6 left mainly for personal or scheduling reasons, but they still leave the trial well short of the planned group size and open to the possibility that the people who stayed were the ones most likely to improve. The sham group was not a do-nothing control. They played a one-muscle version of the same game, so this trial asked whether training several muscles at once is better than training one, not whether the device beats no training. The three-muscle group's advantage over sham appeared only in the per-protocol analysis, which drops people who did not train enough. In the pre-planned main analysis, no group beat sham. Everyone enrolled was at least six months past their stroke and almost all had severe arm impairment with little or no hand function, so the findings say nothing about people earlier after stroke or with milder or more hand-focused problems. People who could not see a laptop screen, who could not follow three-step instructions, or who had botulinum toxin in the arm in the previous three months were excluded, so the patients this applies to are the ones who can see, follow instructions and are not receiving spasticity injections. The device was bulky, needed leads clipped to electrodes that had to be replaced often, which some severely impaired users found hard going, and the results are group averages that can hide wide differences between individuals.

The authors declared no potential conflicts of interest in relation to the research, authorship or publication of the article. The paper states in its methods that the wireless surface EMG recording system used for the training was custom-built by Myomo, Inc.

Do not read the 6.8-second improvement as proof that home myoelectric training works. The main analysis of this trial found no training group better than sham; the three-muscle group only beat sham in the per-protocol analysis of people who completed enough training, and 35 of the 94 randomised participants dropped out before week 6, so the comparison that succeeded may be the one left with the most motivated and able patients.