Open AccessEpilepsy Detection Using Artificial Neural Networks
Author(s) -
Matheus Adler Soares Pinto,
Bruno Rocha Gomes,
João Pedro Moreno Vale,
André Luis Rolim de Castro Silva,
Joadson Teixeira Castro das Chagas,
Wellison Silva Santos,
Victor Alves,
Davi Nascimento,
Marta de Oliveira Barreiros
Publication year - 2020
Publication title -
deleted journal
Language(s) - English
DOI - 10.31686/ijier.vol8.iss4.2292
Subject(s) - epilepsy , electroencephalography , artificial neural network , computer science , artificial intelligence , multilayer perceptron , epileptic seizure , pattern recognition (psychology) , neuroscience , psychology
Epilepsy is a neurological disorder, where there is a cluster of brain cells that behave in a hyperexcitable manner, the individual can promote injuries, trauma or, in more severe cases, sudden death. Electroencephalogram (EEG) is the most used way to detect epileptic seizures. Therefore, more simplified methods of analysis of the EEG can help in the diagnosis and treatment of these individuals more quickly. In this study, we extracted pertinent EEG characteristics to assess the epileptic seizure period. We use Perceptron Multilayer artificial neural networks to classify the period of the crisis, obtaining a more efficient diagnosis. The multilayer neural network obtained an accuracy of 98%. Thus, the strategy of extracting characteristics and the architecture of the assigned network were sufficient for a rapid and accurate diagnosis of epilepsy.