Characterizing visual compensation for proprioceptive impairments during the subacute phase of stroke.

This study was designed for assessment, not to test any treatment, so it cannot show that vision-based training works. The authors state the data do not allow comment on robotic intervention versus standard care or recommendations for implementation. Participants received different amounts and types of rehabilitation, which may have influenced recovery and compensation. Oculomotor function was not measured with an instrumented assessment, so eye movement problems may have affected performance in the vision condition. The robot assessed shoulder and elbow position sense but not wrist or hand, so results may not apply to distal upper limb function. The matching task involved movement of the unaffected arm, although the authors say ipsilateral motor impairment was unlikely because unaffected arm and hand CMSA scores were at least 6. Control data were previously collected rather than concurrent, so the comparison with normal performance relies on an existing database.

Funding was reported from the Canadian Institutes of Health Research, Cumming School of Medicine, Alberta Innovates Health Solutions, Ontario Research Fund, and Heart and Stroke Foundation of Canada. No author conflicts of interest are stated in the supplied text.

Do not conclude that having patients look at their arm is a proven treatment for proprioceptive deficits. This was an assessment study, not an intervention trial, and at 26 weeks 22%, 13%, and 9% of participants were still in partial, absent, or maladaptive compensation groups for the three robotic measures.