Identifying Gait Patterns in Sub-Acute Stroke Patients Based on Open Access Gait Kinematics.

Abstract

Neurological disorders, such as stroke, can affect the ability to walk and balance. Robotic rehabilitation assists in training walking and balance capabilities of patients with neurological disorders. However, not all participants are good responders when using exoskeletons. This study aims to cluster gait patterns in stroke patients to provide insights into the pathology of stroke patients. Joint angles of the affected lower limb of 45 sub-acute stroke patients from an open access database were clustered based on a principal component analyses, followed by a k-means cluster analysis. A total of eight gait pattern clusters were retrieved and clinically rearranged into four categories. The results can be used in the field of robotic devices as well as a more clinical setting. Future research should focus on validating and using the retrieved clusters as clinical indicators for selecting suitable treatments, such as robotic devices or exoskeletons, to personalize rehabilitation.