Electrophysiological modeling in generalized epilepsy using surface EEG and anatomical brain structures

Abstract

Deep brain structures involve significantly in the pathology of brain diseases such as epilepsy, Alzheimer, and Parkinson. Physiological brain modeling has become an emerging approach to investigate the coupling dynamics of the brain activity of these diseases. We propose a method using the surface EEG signals integrated with the anatomical individual brain to build the electrophysiological model of the epileptic patient’s brain. The EEG-driven model is used to investigate the deep brain activities of 23 patients diagnosed with generalized epilepsy from CHB-MIT Scalp EEG Database. Significant changes in the electrical activities in hippocampus, accumbens, and amygdala provide us insights into the dynamics of active brain regions during epilepsy. All of these brain regions show the significant energy variation defined by 5 features (Mean, Max, Min, Standard deviation, Power spectral density) with the p-value, <0.05 in both pre-ictal versus ictal and ictal versus post-ictal. Such result shows the potential of using EEG as a tool to capture the deep brain activity of epilepsy and other diseases that alter deep brain structures. The proposed model may be used to enhance the sensitivity of detecting and predicting epilepsy, detect the progression of the brain lesion, and support the decision-making for a brain medical intervention.