NE-Motion: Visual Analysis of Stroke Patients Using Motion Sensor Networks

Area of Science:

• Biomedical Engineering
• Rehabilitation Science
• Data Science

Background:

• Stroke survivors often experience significant upper extremity (UE) functional deficits, necessitating effective rehabilitation.
• Current UE assessment methods, like the Fugl-Meyer Assessment, may not fully capture compensatory movements.
• Accurate assessment is crucial for tailoring and evaluating rehabilitation strategies.

Purpose of the Study:

• To develop and validate NE-Motion, a graph learning-based visualization tool for analyzing UE motion in stroke patients.
• To enhance the identification of abnormal movement patterns and compensatory strategies during UE tasks.
• To provide a more detailed assessment of UE impairment compared to traditional clinical tests.

Main Methods:

• Utilized a graph learning approach to process time-series motion capture data from sensors worn by patients.
• Developed NE-Motion, a visualization tool designed for analyzing Network Environment for Motion Capture Data Analysis.
• Collaborated with domain experts to ensure clinical relevance and identify key phenomena.

Main Results:

• NE-Motion effectively visualizes motion data, enabling the identification of abnormalities in stroke patients' movement patterns.
• The tool successfully uncovered compensatory movements that bypass impaired limb function.
• Demonstrated differences in movement patterns between stroke survivors and healthy individuals.

Conclusions:

• NE-Motion offers a powerful new approach for analyzing UE function in stroke rehabilitation.
• The visualization tool aids in understanding complex movement patterns and compensatory strategies.
• This method has the potential to improve the assessment of rehabilitation efficacy and patient outcomes.