Effect of Role of Robotic Assistance on Upper-limb Sensorimotor Recovery: A Systematic Review and Meta-Analysis

BackgroundRobot-assisted training is increasingly being adopted to deliver high-intensity, repetitive, and task-specific treatments with real-time feedback to augment recovery. However, the causal role of robotic assistance in sensorimotor recovery, an underlying core mechanistic feature, remains unclear. ObjectiveThis study systematically reviewed randomized controlled trials (RCTs) on upper-limb rehabilitation that isolated the effect of robotic assistance by comparing similar interventions differing only in presence, level, or mode of robotic assistance. MethodThree electronic bibliographic databases-PubMed, Scopus and Google Scholar were searched. Twelve RCTs were included in the qualitative synthesis, and data from eight studies (all involving stroke survivors) were meta-analysed with the Fugl-Meyer Assessment score (measure of impairment) as the primary outcome. The main analysis was done to assess the overall effect of robotic assistance, regardless of disease stage, assistance level and targeted limb area, whereas subgroup analyses were conducted to evaluate the impact of robot-assisted therapy on chronic stroke survivors and to compare therapies focused on the shoulder-elbow versus the hand (wrist/fingers). ResultsThe main analyses yielded a significant positive effect size (0.462) favouring robotic assistance in impairment recovery, however with a significant heterogeneity between the studies. Subgroup analyses suggested potential benefits in chronic phase (moderate significant effect size 0.588), and similar therapy gain for shoulder-elbow and hand-specific training. ConclusionLimited available RCTs with small sample sizes and moderate heterogeneity prevented any firm conclusion to be drawn, and highlighted the need for systematic, controlled studies with large sample sizes to better understand the direct role (mechanistic feature) of robotic assistance in sensorimotor recovery.