Upper-limb functional assessment after stroke using mirror contraction: A pilot study.

Area of Science:

• Biomedical Engineering
• Neurorehabilitation
• Clinical Assessment

Background:

• Current stroke clinical assessments often lack quantitative feedback, relying on subjective rating scales.
• Biomedical signals, such as surface electromyography (sEMG), offer potential for objective patient evaluation.
• Assessing bilateral limb function is crucial for understanding stroke's impact and recovery progress.

Purpose of the Study:

• To develop a unified assessment framework for post-stroke individuals using sEMG bias from bilateral limbs.
• To quantitatively evaluate stroke severity and qualitatively recognize stroke patients.
• To explore the potential for automated stroke rehabilitation through objective signal analysis.

Main Methods:

• Recorded sEMG signals from six channels on each arm of eleven healthy subjects and six stroke patients during four specific movements.
• Employed machine learning algorithms including Linear Discriminant Analysis (LDA), Random Forest (RF), and Support Vector Machine (SVM) for patient recognition.
• Developed a Bilateral Bias Diagnosis Algorithm (BBDA) using a similarity index (SI) to quantitatively assess stroke severity.

Main Results:

• Significant differences in sEMG feature bias between the bilateral arms of stroke patients and healthy individuals were observed.
• RF and SVM algorithms achieved higher recognition accuracy (0.92 ± 0.12 and 0.93 ± 0.12, respectively) compared to LDA (0.84 ± 0.20) in distinguishing stroke patients.
• A strong positive correlation (r = 0.93) was found between the SI derived from sEMG and the Fugl-Meyer clinical score.

Conclusions:

• sEMG-based bilateral bias analysis provides a quantitative complement to traditional qualitative stroke assessments.
• Machine learning algorithms, particularly RF and SVM, are effective in identifying stroke patients based on sEMG data.
• The developed framework and its correlation with clinical scores suggest a pathway towards automated stroke rehabilitation, reducing reliance on therapists.