B-Spline Modeling of Inertial Measurements for Evaluating Stroke Rehabilitation Effectiveness

Area of Science:

• Biomedical Engineering
• Rehabilitation Science
• Machine Learning in Healthcare

Background:

• Upper-limb paralysis post-stroke necessitates ongoing rehabilitation.
• Current evaluation methods rely on subjective scales or traditional assessments.
• Inertial Measurement Units (IMUs) offer objective data but present challenges in direct analysis.

Purpose of the Study:

• To develop an objective method for evaluating hand function and stability in stroke patients.
• To utilize B-splines for processing and analyzing IMU trajectory data.
• To enhance the effectiveness of rehabilitation by providing precise patient evaluations.

Main Methods:

• Employed B-splines to estimate and model IMU trajectory data.
• Utilized machine learning classifiers for data analysis.
• Developed mathematical indices to quantify hand function and stability.
• Validated the method using existing IMU data from a 2018 upper-limb rehabilitation study.

Main Results:

• Features extracted from B-spline trajectories achieved high accuracy in classifying individuals.
• The proposed mathematical indices effectively differentiated between patient groups.
• Demonstrated the feasibility and accuracy of the B-spline approach for IMU data.

Conclusions:

• The B-spline-based method provides a more objective and effective evaluation of upper-limb rehabilitation compared to conventional methods.
• This approach can aid in precise treatment adjustments for stroke survivors.
• Highlights the potential of advanced data analysis techniques in personalized rehabilitation.