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

Area of Science:

• Neurology
• Rehabilitation Engineering
• Biomedical Engineering

Background:

• Neurological disorders like stroke significantly impair walking and balance.
• Robotic rehabilitation and exoskeletons show promise but have variable patient responses.
• Personalized treatment strategies are needed for effective post-stroke recovery.

Purpose of the Study:

• To cluster gait patterns in stroke patients.
• To gain insights into stroke pathology and patient response to rehabilitation.
• To inform the development and application of robotic rehabilitation devices.

Main Methods:

• Analysis of joint angle data from 45 sub-acute stroke patients.
• Application of Principal Component Analysis (PCA) for dimensionality reduction.
• K-means clustering algorithm to identify distinct gait patterns.

Main Results:

• Identification of eight unique gait pattern clusters.
• Clinical rearrangement of clusters into four distinct categories.
• Demonstration of heterogeneity in gait post-stroke.

Conclusions:

• Gait pattern clustering provides valuable insights into stroke patient pathology.
• Results support the personalization of robotic rehabilitation and exoskeleton selection.
• Future research should validate these clusters as clinical indicators for tailored treatments.