Error-augmented bimanual therapy for stroke survivors.

Only 26 adults were analysed, so the trial is small and may not detect or generalise modest effects. The primary outcome showed no significant difference between groups, and the positive error-augmentation result came from a secondary analysis after removing an outlier. The apparent benefit appeared during the week after therapy ended, not immediately after treatment, which makes the timing difficult to interpret. The intervention lasted only two weeks and six sessions, shorter than a typical rehabilitation course. Participants had to have some proximal strength and were excluded for severe aphasia, neglect, visual problems, cognitive impairment, or other severe deficits, so results may not apply to more impaired stroke survivors. Both groups received the same amount of bimanual practice, so the comparison was between error augmentation and no error augmentation, not between bimanual practice and no practice.

Do not conclude that error-augmented bimanual robot practice is better than sham practice. The primary Fugl-Meyer comparison was null, and the apparent advantage appeared only after excluding one control participant and at one-week follow-up.