Extracting of EEG signal features to identify the relationship between these features and gene variants of FAAH enzyme in epilepsy patients

Abstract

Epilepsy is a nervous system disorder that affects the lives of many people worldwide and is often diagnosed through an electroencephalogram (EEG). While drug responses vary among epilepsy patients, genetic factors may influence these differences. On the other hand, performing genetic tests on patients to detect the genotype of genes that are effective in the treatment and control of epilepsy is usually invasive, expensive and time-consuming. Estimating a patient's genotype using data such as information obtained from EEG can be considered a significant achievement. Considering the role of the fatty acid amide hydrolase (FAAH) gene in regulating the physiological function of the brain through the endocannabinoid system and its relationship with diseases such as epilepsy, it is thought that there can be a relationship between the FAAH gene and brain activities. Therefore, identifying genotypes that are associated with epileptic EEG changes is very important for clinical monitoring of epilepsy and control of brain oscillations. By pre-processing EEG signals and extracting features from time, frequency and time-frequency domains, this study investigates the classification of FAAH rs2295633 polymorphism genotypes through EEG signals using machine learning algorithms. In this regard, people carrying the "T" allele are classified in a class. In the classification section of CC and TT genotypes with the K-nearest neighbor (KNN) classifier with 92.36% accuracy (±3.83), in the classification section of CC and (CT, TT) genotypes of treatment-resistant epilepsy patients with the support vector machine (SVM) classification 100% accuracy and in the classification of CC and (CT, TT) genotypes based on EEG signals of epilepsy patients resistant to treatment, the accuracy of 83.63% (6.16±6.16) was achieved. These findings show the potential influence of FAAH rs2295633 polymorphism on brain activity and EEG patterns. However, more extensive multifactorial studies are necessary to establish a precise association between this polymorphism and epileptic EEG. Keywords: Epilepsy, EEG, Endocannabinoid system, FAAH rs2295633 polymorphism, Machine learning, Classification