Migraine classification by machine learning with functional near-infrared spectroscopy during the mental arithmetic...

Area of Science:

• Neuroscience
• Medical Imaging
• Machine Learning

Background:

• Migraine is a complex neurovascular disorder with subjective diagnostic criteria.
• Current diagnostic methods for migraine and medication-overuse headache lack objectivity.
• A need exists for an objective system to aid in accurate migraine diagnosis.

Purpose of the Study:

• To investigate the feasibility of using functional near-infrared spectroscopy (fNIRS) for objective migraine classification.
• To develop a machine learning model for differentiating between healthy controls, chronic migraine, and medication-overuse headache patients.

Main Methods:

• Functional near-infrared spectroscopy (fNIRS) was employed to measure prefrontal cortex (PFC) hemoglobin changes.
• Participants included healthy controls (HC), chronic migraine (CM) patients, and medication-overuse headache (MOH) patients.
• A mental arithmetic task (MAT) was used to elicit brain activity responses.

Main Results:

• The fNIRS-based machine learning model achieved 100% sensitivity and 75% specificity for chronic migraine (CM).
• The model demonstrated 75% sensitivity and 100% specificity for medication-overuse headache (MOH).
• These results indicate high classification accuracy for distinguishing between HC, CM, and MOH groups.

Conclusions:

• Functional near-infrared spectroscopy (fNIRS) combined with machine learning provides a feasible approach for objective migraine classification.
• This objective method can assist clinicians in making more accurate diagnoses of migraine and related disorders.
• The study highlights the potential of neuroimaging techniques in improving headache diagnosis.