Brain Volume Correlations and Machine Learning Classification in Epilepsy Diagnosis

Epilepsy is a neurological disorder associated with complex structural and functional changes in the brain. Although volumetric brain measurements can present an indication of ictal characteristics, the complex nature and variability of volumetric differences in the epileptic brain pose challenges for consistent interpretation and generalization. In this regard, his study investigates the relationship between brain region volumes and epilepsy diagnosis, utilizing MRI structural data from individuals with diagnosed epileptic conditions and healthy controls. As such specific brain regions were measured, and correlation analysis was conducted along with statistical Mann-Whitney U testing. These analyses revealed significant differences in brain volumes between healthy participants and those with epilepsy, particularly in regions implicated in seizure generation and propagation. Subsequently, machine learning classifiers were employed to differentiate individuals based on these brain metrics, alongside with demographic and clinical details, with the Support Vector Machines (SVM) and K-Nearest Neighbors (KNN) classifiers demonstrating the highest classification accuracy. These findings underscore the potential of brain volume metrics as biomarkers for epilepsy and highlight the utility of machine learning models in navigating the complexities of epilepsy diagnosis.