Localization of Epileptic Seizure Focus by Computerized Analysis of fMRI Recordings

TL;DR

This paper presents a machine learning approach using fMRI data to accurately localize epileptic seizure focus in pediatric patients by analyzing cortical interconnectivity patterns.

Contribution

It introduces a novel method combining mutual information analysis with a small, efficient MLP classifier for seizure focus localization.

Findings

High accuracy in classifying seizure focus location.

Effective reduction of model complexity to prevent overfitting.

Potential for clinical application in pediatric epilepsy diagnosis.

Abstract

By computerized analysis of cortical activity recorded via fMRI for pediatric epilepsy patients, we implement algorithmic localization of epileptic seizure focus within one of eight cortical lobes. Our innovative machine learning techniques involve intensive analysis of large matrices of mutual information coefficients between pairs of anatomically identified cortical regions. Drastic selection of pairs of regions with significant inter-connectivity provide efficient inputs for our Multi-Layer Perceptron (MLP) classifier. By imposing rigorous parameter parsimony to avoid over fitting we construct a small size MLP with very good percentages of successful classification.