Effectiveness of perception-interaction-enhanced rehabilitation in post-stroke recovery: A real-world propensity-matched cohort study.

BACKGROUND

Technology-assisted rehabilitation is increasingly used after stroke, but the real-world contribution of perception-interaction modalities remains unclear. This study evaluated their effectiveness and predictive factors in post-stroke motor recovery.

METHODS

We conducted a retrospective, single-center, propensity score-matched cohort study of adults (40-85 years) with mild to moderate upper-limb impairment admitted for inpatient rehabilitation (2020-2023). Patients received either conventional therapy alone or therapy supplemented with at least one perception-interaction modality-functional electrical stimulation (FES), virtual reality (VR), or electromyographic (EMG) biofeedback. Motor and functional outcomes were assessed at admission and discharge using the Fugl-Meyer Assessment (FMA) and Modified Barthel Index (MBI). Clinically meaningful improvement was defined as ΔFMA ≥8. Multivariable logistic regression identified independent predictors.

RESULTS

A total of 160 patients were analyzed (80 per group). The intervention group showed greater motor gains (median ΔFMA: 12.0 vs 7.0; P < 0.001) and functional independence (ΔMBI: 18.0 vs 11.0; P < 0.001). Responders were more frequent in the intervention group (70.0 % vs 45.0 %; P = 0.002). Multivariable analysis confirmed perception-interaction therapy as an independent predictor of motor recovery (OR = 2.21, 95 % CI: 1.15-4.24), along with adherence ≥80 % (OR = 1.84) and longer session duration (OR = 1.22 per 10 minutes). Subgroup analyses indicated superior outcomes with multimodal use and high adherence.

CONCLUSION

Perception-interaction-enhanced rehabilitation was independently associated with improved motor and functional outcomes after stroke. These findings support individualized, technology-integrated rehabilitation strategies, warranting further prospective validation.