Gait Analysis for Post-Stroke Hemiparetic Patient by Multi-Features Fusion Method

Area of Science:

• Biomedical Engineering
• Neurology
• Rehabilitation Science

Background:

• Walking is essential for daily activities, but neurological diseases like stroke impair gait.
• Previous research focused on temporal gait parameters post-stroke, neglecting symmetry, regularity, and stability.

Purpose of the Study:

• To investigate gait symmetry, regularity, and stability in post-stroke hemiparetic gait.
• To develop and evaluate novel gait features for distinguishing stroke survivors from healthy individuals.

Main Methods:

• Utilized dynamic time warping (DTW) for gait symmetry, sample entropy for regularity, and empirical mode decomposition for stability.
• Collected gait data from 15 healthy controls and 15 post-stroke survivors.
• Employed Mann-Whitney test and four classifiers (including k-nearest-neighbor) for analysis.

Main Results:

• The novel gait features significantly differentiated hemiparetic patients from healthy controls (p < 0.05).
• The k-nearest-neighbor classifier achieved a maximum area under the curve of 0.94.
• The proposed features demonstrated high potential for automated gait analysis.

Conclusions:

• Gait symmetry, regularity, and stability are valuable metrics for characterizing hemiparetic gait.
• These features can effectively distinguish stroke survivors, supporting clinical applications.
• The findings pave the way for advanced, automated gait analysis systems in rehabilitation.