The Detrimental Effect of Stroke on Motor Adaptation.

Small studies throughout: individual datasets had 2 to 36 people with stroke and 5 to 31 controls, and the pooled sample was 283 people with stroke. Effect sizes varied substantially between datasets (I2 = 60.4%), so studies are not all telling the same story. Most participants were years past their stroke (on average 4.88 years), so the results may not describe people in the early weeks or months after stroke. Stroke severity was not captured: the authors could not use Fugl-Meyer scores because studies reported impairment inconsistently, so it is unclear how mild or severe the participants were. The review was not pre-registered and screening was not blinded, though two authors screened independently and data were entered twice. The hemisphere comparison rested on only 5 datasets per group. Which arm was used for reaching was usually not controlled: of ten datasets looking at hemispheric contributions, only one had participants reach with the arm affected by their stroke, and most used the unaffected arm. Many included studies did not report whether participants were instructed to drop any deliberate aiming strategy during aftereffect trials, so it is unclear whether the aftereffect truly reflects subconscious learning. Only 21 datasets from 18 studies were available, and one dataset was unpublished.

The authors declared no potential conflicts of interest with respect to the research, authorship or publication of the article. The work was funded by the National Institutes of Health to Darcy S. Reisman (R01HD078330).

Do not treat the left hemisphere finding as established. The impairment in people with left hemisphere damage was large but did not reach statistical significance (P = .055), the comparison rested on only 5 datasets per group, and only one of the ten datasets had participants reach with the arm affected by their stroke.