Artificial Intelligence in the Neurorehabilitation of Patients with Neurological Disabilities

Abstract

Artificial intelligence has established itself as an emerging tool with significant potential to transform neurorehabilitation for patients with neurological disabilities. Its application enables the integration of vast amounts of clinical, functional, and digital data to optimize assessment, estimate prognoses, tailor interventions, and enhance therapeutic monitoring. To critically analyze the available evidence, a narrative review of literature from biomedical and rehabilitation databases was conducted, prioritizing systematic reviews, scoping reviews, and clinical studies related to artificial intelligence, neurorehabilitation, and neurological disabilities published between 2020 and 2026. Findings revealed significant applications in stroke, spinal cord injury, Parkinson's disease, cognitive impairment, and other neurological disorders, particularly in the areas of robotics, telerehabilitation, adaptive systems, and decision-support tools. However, methodological limitations, outcome heterogeneity, insufficient external validation, and ethical challenges regarding transparency, privacy, and algorithmic bias persist. Overall, artificial intelligence represents a promising strategy for advancing toward more precise, personalized, and efficient rehabilitation, although its clinical implementation requires robust evidence, multicenter study designs, and well-defined regulatory frameworks.