Objective: Exercise is widely recognized as an effective non-pharmacological intervention to maintain health in older adults. With advances in artificial intelligence (AI), AI-assisted exercise has emerged as a novel rehabilitation approach, yet its comparative effectiveness against traditional and software-assisted programs remains unclear. This study aimed to evaluate and rank the relative effectiveness of these interventions on multiple physical and psychological outcomes using a network meta-analysis (NMA). Methods: Following the PRISMA-NMA guidelines, we systematically searched PubMed, Embase, Cochrane Library, Web of Science, and Scopus up to June 2025. Eligible studies were randomized controlled trials (RCTs) involving adults ≥ 60 years comparing AI-assisted, software-assisted, and conventional upper/lower limb rehabilitation. Six outcomes were analyzed: gait, balance, range of motion (ROM), muscle strength, cognitive function, and quality of life (QOL). Stata 17.0 was used to conduct the NMA, calculating the standardized mean differences (SMDs) and SUCRA rankings, with assessments of heterogeneity and risk of bias. Results: Seventy RCTs with 808 participants were included. All active interventions outperformed the placebo. AI-assisted programs showed the strongest effects on gait (SMD = 1.33) and balance (SMD = 0.76), while software-assisted interventions ranked highest for ROM (SMD = 0.69) and QOL (SMD = 1.06). Both AI and software interventions improved cognition and muscle strength. Heterogeneity was low (I2 ≤ 38.5%). Subgroup analysis indicated that AI-based methods were superior to traditional rehabilitation, although differences among novel interventions were not statistically significant. Conclusions: AI-assisted exercise is highly effective for gait and balance, while software-assisted approaches excel in ROM and QOL. These interventions hold promise for community and home-based rehabilitation. Future studies should investigate integrated “AI + traditional” models and incorporate biomechanical and neurophysiological indicators to optimize personalized care.
Stroke-induced motor impairment necessitates objective and quantitative assessment tools for rehabilitation planning. In this study, a gesture-specific framework based on high-density surface electromyography (HD-sEMG) was developed to characterize neuromuscular dysfunction using eight macroscopic features and two microscopic motor unit decomposition features. HD-sEMG recordings were collected from stroke patients (n = 11; affected and unaffected sides) and healthy controls (n = 8; dominant side) during seven standardized hand gestures. Feature-level comparisons revealed hierarchical abnormalities, with the affected side showing significantly reduced activation/coordination relative to healthy controls, while the unaffected side exhibited intermediate deviations. For each gesture, dedicated K-nearest neighbors (KNN) models were constructed for clinical validation. For Brunnstrom stage classification, wrist extension yielded the best performance, achieving 92.08% accuracy and effectively discriminating severe (Stage 4), moderate (Stage 5), and mild (Stage 6) impairment as well as healthy controls. For fine motor recovery prediction, the thumb–index–middle finger pinch provided the optimal regression performance, predicting Upper Extremity Fugl–Meyer Assessment (UE-FMA) scores with R = 0.86 and RMSE = 3.24. These results indicate that gesture selection should be aligned with the clinical endpoint: wrist extension is most informative for gross recovery staging, whereas pinch gestures better capture fine motor control. Overall, the proposed HD-sEMG framework provides an objective approach for monitoring post-stroke recovery and supporting personalized rehabilitation assessment.