AUTHOR=Feng Guang , Chai Guohong , Zhang Jiaji , Song Tao , Shi Changcheng , Xu Jialin , Zuo Guokun TITLE=Effective unilateral/bilateral robot-assisted training for upper limb motor function rehabilitation: a cross-sectional study JOURNAL=Frontiers in Human Neuroscience VOLUME=Volume 19 - 2025 YEAR=2025 URL=https://www.frontiersin.org/journals/human-neuroscience/articles/10.3389/fnhum.2025.1571624 DOI=10.3389/fnhum.2025.1571624 ISSN=1662-5161 ABSTRACT=ObjectiveThe therapeutic effect of robot-assisted training is still indecisive due to the lack of patient-tailored protocols and dose-matched training strategies when compared to traditional treatment. The objective of this study was to investigate the optimal robot-assisted training strategies for the upper limb functional recovery in hemiparetic stroke patients.ApproachA bilateral upper limb rehabilitation robot was employed to execute unilateral and bilateral training. Eighteen able-bodied subjects were recruited to test the effective of robot-assisted training strategies before transferring them to stroke patients. We compared unilateral passive training (UPT), bilateral passive training (BPT), and unilateral active training (UAT) with various feedback types (visual, force, and visual-force, none). These trainings were performed on three kinds of virtually-guided (straight-line, circular, S-shaped) tasks. Tracking error (TE), interactive force (IF) and target muscle activation level were quantified to characterize the motion capability and active participation of subjects.Main resultsResults revealed that BPT-visual (0.63 ± 0.26) significantly increased muscle activation level when compared to those of BPT-none (0.45 ± 0.27) and UPT-visual (0.24 ± 0.05) (p < 0.01). UAT with single-modality feedback (visual/force) enabled higher TE (22.5 ± 3.40 mm) and active participation (0.78 ± 0.12) when compared with UAT with multi-modality (visual-force) feedback (TE: 6.6 ± 0.8 mm; activation level: 0.53 ± 0.13) (p < 0.01). The relatively complex circular and S-shaped tasks significantly enhanced the benefits of various training strategies.SignificanceThe current outcomes provide valuable guidelines for designing individualized robot-assisted training protocols, potentially promoting the clinical rehabilitation effect.