Long-term impact of robot-assisted training on walking pattern reconstruction in chronic stroke survivors: insights...

To determine whether robot-assisted lower-limb training promotes superior walking pattern reconstruction in chronic stroke survivors, as quantified by multiparametric gait analysis and supported by concurrent improvements in motor, balance, and functional outcomes.

In this retrospective observational study, 80 patients with chronic stroke sequelae treated between January 2024 and January 2026 were allocated according to routine rehabilitation exposure to a control group (n = 45) or an observation group (n = 35); the primary endpoint was post-intervention gait remodeling, defined by change in gait speed and multiparametric spatiotemporal gait indices, while secondary endpoints included lower-limb motor recovery, spasticity, balance, functional independence, clinically meaningful responder status, and multivariable analysis of determinants of gait-speed change as the principal marker of walking pattern reconstruction.

After 10 weeks, the observation group showed significantly greater improvement than the control group in lower-limb motor function, including Fugl Meyer Assessment of the Lower Extremity total, proximal, and distal subscores, Timed Up and Go performance, and 10-m walking speed (all P < 0.001). Spasticity was also more effectively reduced, with lower Modified Ashworth Scale composite and segment-specific scores (P = 0.002 to < 0.001). In addition, balance, spatiotemporal gait parameters, and motor-related functional independence improved more markedly in the observation group (all P < 0.001 for key outcomes).

Robot-assisted lower-limb rehabilitation, when added to conventional therapy, significantly enhanced gait remodeling, motor recovery, spasticity reduction, balance performance, and locomotion-related independence in chronic stroke survivors, supporting its value as an effective strategy for promoting more physiological and clinically meaningful walking restoration after stroke.