Is Implicit Motor Learning Preserved after Stroke? A Systematic Review with Meta-Analysis

• Published in: PloS one 2016-12, Vol.11 (12), p.e0166376-e0166376
• Main Authors: Kal, E, Winters, M, van der Kamp, J, Houdijk, H, Groet, E, van Bennekom, C, Scherder, E
• Format: Article
• Language: English
• Subjects: Aged; Analysis; Asymmetry; Automation; Bias; Biology and Life Sciences; Brain research; Controlled vocabularies; Feasibility studies; Fitness equipment; Human performance; Humans; Learning; Learning - physiology; Learning strategies; Medical imaging; Medicine and Health Sciences; Memory; Meta-analysis; Middle Aged; Motor skill learning; Motor Skills - physiology; Neuropsychology; Online databases; Patients; Physical Sciences; R&D; Reaction Time; Reaction time task; Recovery of Function; Rehabilitation; Research & development; Research and Analysis Methods; Social Sciences; Stroke; Stroke - physiopathology; Stroke - psychology; Stroke Rehabilitation; Systematic review; Walking
• ISSN: 1932-6203; 1932-6203
• DOI: 10.1371/journal.pone.0166376

Many stroke patients experience difficulty with performing dual-tasks. A promising intervention to target this issue is implicit motor learning, as it should enhance patients' automaticity of movement. Yet, although it is often thought that implicit motor learning is preserved post-stroke, evidence for this claim has not been systematically analysed yet. Therefore, we systematically reviewed whether implicit motor learning is preserved post-stroke, and whether patients benefit more from implicit than from explicit motor learning. We comprehensively searched conventional (MEDLINE, Cochrane, Embase, PEDro, PsycINFO) and grey literature databases (BIOSIS, Web of Science, OpenGrey, British Library, trial registries) for relevant reports. Two independent reviewers screened reports, extracted data, and performed a risk of bias assessment. Overall, we included 20 out of the 2177 identified reports that allow for a succinct evaluation of implicit motor learning. Of these, only 1 study investigated learning on a relatively complex, whole-body (balance board) task. All 19 other studies concerned variants of the serial-reaction time paradigm, with most of these focusing on learning with the unaffected hand (N = 13) rather than the affected hand or both hands (both: N = 4). Four of the 20 studies compared explicit and implicit motor learning post-stroke. Meta-analyses suggest that patients with stroke can learn implicitly with their unaffected side (mean difference (MD) = 69 ms, 95% CI[45.1, 92.9], p < .00001), but not with their affected side (standardized MD = -.11, 95% CI[-.45, .25], p = .56). Finally, implicit motor learning seemed equally effective as explicit motor learning post-stroke (SMD = -.54, 95% CI[-1.37, .29], p = .20). However, overall, the high risk of bias, small samples, and limited clinical relevance of most studies make it impossible to draw reliable conclusions regarding the effect of implicit motor learning strategies post-stroke. High quality studies with larger samples are warranted to test implicit motor learning in clinically relevant contexts.