Quantitative Assessment of Upper Limb Multi-Modal Feature Fusion Under Task-Oriented Movement.

Area of Science:

• Rehabilitation Medicine
• Biomedical Engineering
• Neuroscience

Background:

• Traditional post-stroke motor function assessment relies on clinical experience, lacking quantitative analysis.
• Objective assessment of upper limb motor status in hemiplegic patients is crucial for effective rehabilitation.

Purpose of the Study:

• To propose and validate a novel, objective assessment method for upper limb motor function in post-stroke hemiplegic patients.
• To develop a multi-modal feature fusion framework integrating kinematic and muscle synergy data for comprehensive motor function analysis.

Main Methods:

• Extracted features from multi-modal upper limb movement data (kinematics and muscle synergy).
• Quantified kinematic and muscle synergy using singular value decomposition (SVD) and similarity metric index.
• Developed an aggregated classifier and a probability-based functional scoring mechanism using multi-modal feature fusion.

Main Results:

• Integrating synergy analysis improved classification accuracy by over 2.3%.
• Multi-modal feature fusion achieved 90.75% classification accuracy.
• Assessment results showed significant correlation with standard clinical trial scores (r=-0.81, p=0.0147).

Conclusions:

• The proposed multi-modal feature fusion method provides an objective and quantitative assessment of upper limb motor function post-stroke.
• This novel framework demonstrates feasibility for clinical application in assessing hemiplegic patients, enhancing rehabilitation strategies.