Gait asymmetry and reduced gait speed (GS) are common after stroke. This systematic review and meta-analysis evaluated the effects of gait training with lower-limb robotic exoskeletons or exoskeleton-type devices on gait asymmetry and GS compared with conventional rehabilitation or non-robotic gait training. PubMed, Embase, Web of Science, the Cochrane Library, and Scopus were searched from inception to 21 June 2026. Randomized controlled trials (RCTs) reporting spatial gait asymmetry (SGA), temporal gait asymmetry (TGA), or GS were included. Standardized mean differences (SMDs; Hedges’ g) and 95% confidence intervals (CIs) were pooled using random-effects models. Eleven RCTs involving 532 randomized participants were included. Training with these devices reduced SGA (SMD = −0.68, 95% CI −1.13 to −0.22, p < 0.01) and improved GS (SMD = 0.46, 95% CI 0.12 to 0.81, p = 0.01), but did not significantly affect TGA (SMD = −0.85, 95% CI −1.90 to 0.20, p = 0.11). The certainty of evidence, assessed using the Grading of Recommendations Assessment, Development and Evaluation (GRADE) framework, was very low for all three outcomes. Gait training with lower-limb robotic exoskeletons or exoskeleton-type devices may reduce SGA and improve GS after stroke, whereas its effect on TGA remains uncertain.

Objective: To evaluate the effects of exercise-based interventions on self-reported sleep outcomes in patients with Parkinson’s disease (PD) and to assess the certainty of the available evidence. Methods: PubMed, Web of Science, Scopus, the Cochrane Central Register of Controlled Trials (CENTRAL), and China National Knowledge Infrastructure (CNKI) were searched from inception to 17 August 2026. The review protocol was registered with INPLASY (INPLASY202480029). Risk of bias was assessed using the Cochrane Risk of Bias tool (RoB 1). Random-effects meta-analyses were performed using Hedges’ g, with between-study variance estimated by restricted maximum likelihood. Exploratory subgroup analyses and univariable meta-regression were used to investigate heterogeneity, and sensitivity analyses were conducted to assess the robustness of the findings. Certainty of evidence was assessed using Grading of Recommendations Assessment, Development and Evaluation (GRADE). Results: Fifteen randomized controlled trials involving 696 participants were included. Exercise-based interventions were associated with improved self-reported sleep outcomes (standardized mean difference [SMD] = −0.75, 95% confidence interval [CI] −1.28 to −0.23; p = 0.0084), although between-study heterogeneity was substantial (I² = 82.4%). Instrument-stratified analyses showed no statistically significant differences across sleep questionnaires. Sensitivity analyses restricted to studies using the two most frequently reported sleep questionnaires and leave-one-out analyses yielded estimates consistent in direction with the primary analysis. Subgroup analyses showed no statistically significant differences across exercise intensity, session duration, weekly frequency, intervention period, or exercise type. None of the continuous moderators examined in univariable meta-regression was statistically significant. Conclusions: Exercise-based interventions may improve self-reported sleep outcomes in patients with PD, but the certainty of evidence is very low, and substantial unexplained heterogeneity limits the precision and clinical interpretability of the pooled estimate. Current evidence is insufficient to identify an optimal exercise prescription. Future adequately powered trials should incorporate standardized exercise reporting and objective sleep assessments.