Machine learning for exploring neurophysiological functionality in multiple sclerosis based on trigeminal and hand...

Area of Science:

• Neuroscience
• Clinical Neurology
• Biomedical Engineering

Background:

• Brainstem dysfunction is a significant factor in Multiple Sclerosis (MS) progression and disability.
• Blink reflexes, like the Trigeminal Blink Reflex (TBR), assess brainstem function but may lack sensitivity for early detection.
• The Hand Blink Reflex (HBR) offers a novel method to evaluate the peripersonal space representation, potentially revealing subtle brainstem alterations.

Purpose of the Study:

• To investigate the utility of Machine Learning (ML) algorithms in conjunction with TBR and HBR for detecting brainstem dysfunction in MS.
• To assess if HBR provides additional clinical information beyond TBR for MS diagnosis.
• To evaluate the potential of ML-powered neurophysiological measurements for early MS detection.

Main Methods:

• Recording of Hand Blink Reflex (HBR) and Trigeminal Blink Reflex (TBR) in individuals with Relapsing-Remitting MS and healthy controls.
• Feature extraction from HBR and TBR recordings.
• Training AdaBoost classifiers using extracted features for binary classification (MS vs. Controls).

Main Results:

• Both TBR and HBR, when analyzed with ML, achieved classification accuracy comparable to or exceeding that of clinicians.
• The HBR, in particular, demonstrated potential for providing supplementary diagnostic information.
• ML techniques effectively identified individuals with MS based on neurophysiological reflex data.

Conclusions:

• Machine learning offers a powerful tool to enhance the clinical interpretation of neurophysiological reflexes in MS.
• HBR shows promise as a novel biomarker for assessing brainstem integrity and aiding early MS diagnosis.
• Integrating ML with HBR and TBR could significantly improve the early detection and management of brainstem dysfunction in Multiple Sclerosis.