Robust tracking of the upper limb for functional stroke assessment

Area of Science:

• Biomedical Engineering
• Rehabilitation Technology
• Movement Science

Background:

• Stroke survivors often experience upper limb motor impairments.
• Accurate kinematic assessment is crucial for evaluating motor function and recovery.
• Existing motion capture systems can be expensive and difficult to deploy in clinical or home settings.

Purpose of the Study:

• To develop and validate a robust 3-D parts-based (PB) tracking system for monitoring upper limb kinematics in stroke survivors.
• To assess the system's accuracy, robustness, and correlation with established functional measures.
• To provide a potentially more accessible tool for tracking post-stroke limb function during daily activities.

Main Methods:

• A probabilistic, parts-based (PB) model of the arm was developed, incorporating limb shape, color, and likely configurations.
• The system processed image sequences from multiple camera angles to track upper limb movement.
• Seven stroke survivors' functional arm motion was recorded concurrently using the PB system and a commercial motion capture device.

Main Results:

• The PB system demonstrated robustness to cluttered environments and temporary occlusions.
• Kinematic measurements showed high accuracy relative to a commercial motion capture system (e.g., elbow angle difference: 5.7°, hand velocity difference: 2.6 cm/s).
• Key kinematic metrics, such as standard deviation in shoulder velocity, significantly correlated with Fugl-Meyer Assessment (FMA) scores (r < -0.75).

Conclusions:

• The 3-D parts-based (PB) tracking system is a viable tool for accurately capturing upper limb kinematics in stroke survivors.
• The system's ability to correlate with functional assessments like the FMA suggests its utility in rehabilitation and monitoring.
• This technology offers a promising approach for objective, quantitative assessment of post-stroke motor recovery during functional tasks.