Artificial Intelligence–Based Gait Analysis in Neurological Rehabilitation: A Systematic Review

Abstract

Background: Gait impairment is a common consequence of neurological disorders, including stroke, Parkinson’s disease, multiple sclerosis, and cerebral palsy. Accurate gait assessment is essential for identifying functional limitations, monitoring recovery, and optimizing rehabilitation interventions. Traditional gait assessment methods may be limited by their dependence on specialized equipment, clinical expertise, and controlled laboratory environments. Artificial intelligence (AI), including machine learning, deep learning, and computer vision, offers promising opportunities to improve gait analysis through automated movement assessment, pattern recognition, and data-driven clinical decision support.

Objective: This systematic review aims to examine the applications of AI-based gait analysis in neurological rehabilitation, focusing on computational approaches, data acquisition technologies, gait assessment outcomes, clinical applications, and challenges affecting clinical implementation.

Methods: The review will systematically identify and evaluate relevant studies using predefined eligibility criteria and a reproducible search strategy across appropriate biomedical and engineering databases. Eligible studies will be examined for neurological populations, gait data sources, AI techniques, assessment objectives, validation methods, and reported outcomes. Study selection and reporting will follow the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA 2020) guidelines. Risk of bias and methodological quality will be assessed using tools appropriate to the included study designs. Findings will be synthesized narratively, with quantitative synthesis considered when sufficient methodological and clinical comparability exists.

Conclusion: AI-based gait analysis has the potential to support more objective, scalable, and personalized approaches to neurological rehabilitation. However, its clinical usefulness depends on reliable measurement, representative datasets, transparent model development, external validation, and demonstrated relevance to patient-centered outcomes. This review will identify current evidence, methodological limitations, and research priorities to guide the responsible integration of AI-based gait assessment into neurological rehabilitation practice.