The Detrimental Effect of Stroke on Motor Adaptation

Abstract

BackgroundWhile it is evident that stroke impairs motor control, it remains unclear whether stroke impacts motor adaptation-the ability to flexibly modify movements in response to changes in the body and the environment. The mixed results in the literature may be due to differences in participants' brain lesions, sensorimotor tasks, or a combination of both.ObjectiveWe first sought to better understand the overall impact of stroke on motor adaptation and then to delineate the impact of lesion hemisphere and sensorimotor task on adaptation poststroke.MethodsFollowing the Preferred Reporting Items for Systematic reviews and Meta-Analyses guidelines, we conducted a systematic review and meta-analysis of 18 studies comparing individuals poststroke to neurotypical controls, with each group consisting of over 200 participants.ResultsWe found that stroke impairs motor adaptation (d = -0.63; 95% confidence interval [-1.02, -0.24]), and that the extent of this impairment did not differ across sensorimotor tasks but may vary with the lesioned hemisphere. Specifically, we observed greater evidence for impaired adaptation in individuals with left hemisphere lesions compared to those with right hemisphere lesions.ConclusionsThis review not only clarifies the detrimental effect of stroke on motor adaptation but also underscores the need for finer-grained studies to determine precisely how various sensorimotor learning mechanisms are impacted. The current findings may guide future mechanistic and applied research at the intersection of motor learning and neurorehabilitation.

Keywords: meta-analysis; motor adaptation; motor control; motor impairment; stroke; systematic review.

Figure 1.

Preferred Reporting Items for Systematic reviews and Meta-Analyses flow diagram describing the study inclusion process of the systematic review. We identified 18 studies, resulting in 21 total datasets, that fulfilled our eligibility criteria.

Figure 2.

Overview of datasets included in this meta-analysis. Schematic of motor adaptation tasks involving (A) reaching, (B) saccades, and (C) walking. (D) The number of patients (PT) and controls (CT), lesion side (shading indicates left, right, or both; where “both” means that some patients have left hemisphere lesions and others have right hemisphere lesions), and lesion location for each dataset. We also report the effector (eg, arms, legs, eyes), the limb used for the task (contralesional limb, ipsilesional limb, or both; where “both” means that some patients used the contralesional limb and others used the ipsilesional limb), and the outcome measure provided (late adaptation or aftereffect; denoted by the shaded region).^-,-

Figure 3.

Stroke impairs motor adaptation. Forest plot comparing the performance of individuals poststroke to neurotypical controls, where negative values indicate greater adaptation in controls (ie, impaired adaptation poststroke). The overall effect size is indicated by the blue vertical line. Each square represents a single dataset with its size indicating the weight assigned to that dataset in the random-effects model. Whiskers represent the 95% confidence intervals.

Figure 4.

Minimal impact of experimental tasks on motor adaptation poststroke. We assigned subgroups based on whether datasets used visuomotor rotation or split-belt walking tasks. The effect size for each subgroup (purple for visuomotor rotation and green for split-belt walking) is indicated by the vertical line. Each square represents a single dataset with its size indicating its weight and whiskers representing 95% confidence intervals.

Figure 5.

Impact of lesion hemisphere on motor adaptation poststroke. We assigned subgroups based on whether datasets compared motor adaptation between individuals with left hemisphere lesions or right hemisphere lesions against controls. The effect size for each subgroup (red for left hemisphere and orange for right hemisphere) is indicated by the vertical line. Each square represents a single dataset with its size indicating its weight and whiskers representing 95% confidence intervals.

Figure 6.

(A) Methods for isolating implicit recalibration. (B) Evidence suggesting which brain regions are involved in which learning processes.^,,,,, The solid arrows indicate studies that involve populations with stroke and open arrows indicate studies that involve populations with progressive neurodegenerative disorders. (C) Methods for isolating explicit re-aiming strategies.