Decoding of Ankle Joint Movements in Stroke Patients Using Surface Electromyography.

Area of Science:

• Biomedical Engineering
• Rehabilitation Science
• Neuroscience

Background:

• Stroke, a cerebrovascular disease, often leads to motor impairments, particularly in lower limbs.
• Current rehabilitation methods for lower limb motor recovery are limited, especially for home-based, patient-assisted devices.
• Surface electromyography (sEMG) offers a potential non-invasive method for controlling rehabilitative devices.

Purpose of the Study:

• To investigate the feasibility of using surface electromyography (sEMG) for controlling a home-based lower limb rehabilitation device for stroke survivors.
• To decode ankle joint movements from sEMG signals in stroke patients.
• To assess the correlation between sEMG-based movement classification accuracy and the level of motor impairment.

Main Methods:

• Collected sEMG data from three channels from 11 stroke patients during ankle joint movements.
• Extracted Hudgins time-domain features from sEMG signals.
• Classified movements using linear discriminant analysis (LDA) and artificial neural network (ANN), and quantified motor impairment using the Fugl-Meyer Assessment (FMA) scale.

Main Results:

• Average movement classification accuracies were 63.86% ± 4.3% for LDA and 67.1% ± 7.9% for ANN.
• Significant differences in classification performance were observed across different movements for both LDA (p < 0.001) and ANN (p = 0.014).
• Moderately positive correlations were found between FMA scores and classification accuracies (ρ = 0.75 for LDA, ρ = 0.55 for ANN).

Conclusions:

• sEMG signal decoding shows potential for developing home-based lower limb rehabilitation systems for stroke patients.
• The correlation between classification accuracy and motor impairment suggests that sEMG-based systems can be tailored to individual patient needs.
• This technology could lead to customized, accessible, and effective physical therapy for improving functional lower limb recovery post-stroke.