A computer-aided diagnosis of multiple sclerosis based on mfVEP recordings

Area of Science:

• Neuroscience
• Medical Imaging
• Biomedical Engineering

Background:

• Multiple Sclerosis (MS) is a demyelinating disease of the central nervous system.
• Early diagnosis of MS is crucial for effective management and treatment.
• Current diagnostic methods can be invasive or time-consuming.

Purpose of the Study:

• To develop and evaluate a computer-aided diagnosis (CAD) system for identifying different stages of MS.
• To utilize multifocal visual-evoked potentials (mfVEPs) for objective MS assessment.
• To differentiate between various stages of MS, including Radiologically Isolated Syndrome (RIS), Clinically Isolated Syndrome (CIS), and definite MS.

Main Methods:

• Acquired mfVEP signals from 30 eyes with RIS, 62 eyes with CIS, 56 eyes with definite MS, and 44 control eyes.
• Developed feature vectors from mfVEP signal intensity, latency, and singular values.
• Implemented and compared a flat multiclass classifier (FMC) and a hierarchical classifier (HC) using the k-Nearest Neighbour (k-NN) algorithm for eye and subject classification.

Main Results:

• The hierarchical classifier (HC) outperformed the flat multiclass classifier (FMC) in eye classification (accuracy = 0.74, MCC = 0.68).
• The CAD system achieved high accuracy (0.95) and MCC (0.93) in classifying subjects with MS.
• Singular values of mfVEP signals provided significant discriminatory information beyond amplitude and latency.

Conclusions:

• The developed computer-aided diagnosis system using mfVEPs shows significant promise for diagnosing multiple sclerosis.
• The system can accurately differentiate between various stages of MS, including early forms.
• mfVEP singular values offer valuable insights into axonal loss and demyelination, aiding in disease staging.