Development of A Knowledge Graph-Based Cause Retrieval System for Post-Stroke Gait Analysis

Stroke often results in gait impairments, requiring detailed assessments to support effective rehabilitation. This study presents the development and preliminary evaluation of a knowledge graph-based cause retrieval system for post-stroke gait analysis. The event knowledge graph, constructed using insights from the Rancho Los Amigos framework and refined through expert feedback, represents the relationships between gait deviations and their underlying contributors. This knowledge graph is integrated into a system that includes an automated cause retrieval process, enabling clinicians to efficiently explore potential contributors to observed gait deviations. A case study and a user experience based on the Technology Acceptance Model (TAM) suggest that the system improves interpretability and supports clinical workflows, highlighting the potential of knowledge graphs to enhance AI-driven gait assessment tools for post-stroke rehabilitation.Clinical Relevance- This study presents a knowledge graph-based system to assist clinicians in identifying potential contributors to gait deviations in post-stroke patients. By organizing clinical knowledge in a structured format, the system supports cause identification and may aid in developing more targeted rehabilitation strategies, potentially improving clinical workflows and decision-making.