Classification and Assessment of the Patelar Reflex Response through Biomechanical Measures

Area of Science:

• Biomedical Engineering
• Neurology
• Machine Learning

Background:

• The patellar reflex test is a common diagnostic tool, but its evaluation is typically manual and subjective.
• Objective measurement and classification of the patellar reflex can improve diagnostic accuracy and consistency.

Purpose of the Study:

• To develop a device for controlled elicitation and objective measurement of the patellar reflex.
• To classify patellar reflex responses using machine learning based on biomechanical data.

Main Methods:

• A novel device with a controlled pendulum hammer was used to elicit the patellar reflex.
• An inertial measurement unit and tapping sensor recorded leg angle and velocity from 106 volunteers.
• Machine learning classifiers were employed to differentiate reflex responses based on NINDS scale criteria.

Main Results:

• The system successfully recorded and processed patellar reflex data, including angle and angular velocity.
• Domain-specific features were extracted from the biomechanical signals.
• The proposed method achieved accurate classification of patellar reflexes into NINDS scale groups.

Conclusions:

• Automated, objective classification of the patellar reflex is feasible using biomechanical measurements.
• The developed device and machine learning approach offer a promising tool for neurological assessment.
• This technology has the potential to standardize and enhance the clinical evaluation of the patellar reflex.