The Use of Kinematic Features in Evaluating Upper Limb Motor Function Learning Progress Based on Machine Learning.

Area of Science:

• Biomechanics
• Rehabilitation Science
• Machine Learning in Healthcare

Background:

• Current clinical assessments for rehabilitation lack sensitivity and continuity for tracking patient progress.
• Kinematic features offer a potential solution due to their sensitive and continuous nature in reflecting movement changes.
• Identifying optimal kinematic features for rehabilitation evaluation remains a challenge.

Purpose of the Study:

• To explore the application of kinematic features for classifying movement patterns and qualities in rehabilitation.
• To identify sensitive kinematic features that can monitor patient learning progress.
• To compare selected kinematic features with established clinical measures like normalized jerk.

Main Methods:

• Extracted 12 kinematic features from a 7-segment triangle motion pattern.
• Performed statistical analysis comparing kinematic features with normalized jerk.
• Developed two supervised machine learning models to classify movement patterns and qualities.
• Collected data from 14 participants using a single position sensor.

Main Results:

• Six kinematic features demonstrated sensitivity in reflecting changes during the rehabilitation experiment.
• All analyzed kinematic features contributed significantly to the classification tasks.
• Features derived from movement velocity proved most beneficial for classification, aligning with existing literature.

Conclusions:

• Kinematic features, especially velocity-based ones, are effective for classifying movement patterns and qualities in rehabilitation.
• This approach provides a more sensitive and continuous method for evaluating patient progress.
• Machine learning models utilizing kinematic features can enhance personalized and evidence-based rehabilitation interventions.