EEG Brain Signal Classification for Epileptic Seizure Disorder Detection

Sandeep Kumar Satapathy, Satchidananda Dehuri, Alok Kumar Jagadev, Shruti Mishra

• Publisher: Academic Press
• Year: 2019
• Print ISBN: 9780128174265
• eBook ISBN: 9780128174272
• Topic: Ciencias biológicas
• Subtopic: Biotecnología

EEG Brain Signal Classification for Epileptic Seizure Disorder Detection provides the knowledge necessary to classify EEG brain signals to detect epileptic seizures using machine learning techniques. Chapters present an overview of machine learning techniques and the tools available, discuss previous studies, present empirical studies on the performance of the NN and SVM classifiers, discuss RBF neural networks trained with an improved PSO algorithm for epilepsy identification, and cover ABC algorithm optimized RBFNN for classification of EEG signal. Final chapter present future developments in the field.This book is a valuable source for bioinformaticians, medical doctors and other members of the biomedical field who need the most recent and promising automated techniques for EEG classification.- Explores machine learning techniques that have been modified and validated for the purpose of EEG signal classification using Discrete Wavelet Transform for the identification of epileptic seizures- Encompasses machine learning techniques, providing an easily understood resource for both non-specialized readers and biomedical researchers- Provides a number of experimental analyses, with their results discussed and appropriately validated

Chapter 1

Introduction

Abstract

This chapter presents the basic introduction for the research work carried out in this book. It presents the introduction to electroencephalogram (EEG) signals and its characteristics and behaviors along with the methodologies used for recording of these signals. General and specific goals are briefly discussed in this chapter. It also introduces the motivation and objectives of the research work, along with different techniques used to support them. This chapter also gives a clear picture of the popularly used machine learning algorithms which support in fulfilling our objectives. Different feature extractions to extract the features from EEG signal are also discussed here.

Keywords

Electroencephalogram (EEG); Feature extraction; Discrete wavelet transform (DWT); Artificial neural network (ANN); Support vector machine (SVM)

Electroencephalogram or electroencephalography (EEG) is a trial performed on mental capacity to record the electrical activity in brain. The neural structure of the brain consists of several neurons in terms of lacs. These neurons communicate by colliding among themselves and passing information to each other. This collision leads to the generation of a very small amount of electricity. This is utterly different from the general electricity, which is very high in magnitude. This electric signal flow decides the behavior of a person. In a human brain the normal stream of electrical signal leads to a healthy person. And an abnormal electrical signal flow can pass to an unhealthy person. Hence, these signs can be recorded and analyzed to solve many neurological disorder diseases. The transcription of the electrical activity is essentially caused by putting electrodes on the scalp for 20–40 min, which evaluates the potential fluctuations in the brain [1]. The nerve cells in the psyche are the origin of electric charge, and then they exchange ions with the extracellular milieu. Ions of the same charge repel each other and in this manner they are forced out of the neurons when a number of ions are driven out at the same time they promote each other and form a way known as volume conduction [2]. When this wave reaches the electrode they push or force the ions along the air foil of the electrode which create potential difference and this voltage difference recorded over time gives EEG signals. The key motivation behind this research work is the rapid growth in volume of biological and clinical data or records. To extract knowledge from these data which can be served to be a clinical application, there are different data analysis difficulties which need to be overcome. Many analytical tools based on machine learning (ML) approaches have been invented to tackle with such challenging task of data analysis problems. Around 1% of the total population in the world are affected by a neurological disease called epilepsy. A careful analysis of these EEG signals can solve many neurological disorder diseases.

1.1 Problem Statement

Nowadays, the recording of EEG signal can be easily managed with the aid of various hardware and software techniques. By simply seeing at these signals with naked eyes one cannot make out any abnormality in the sign. Hence, the most important problem is to study these signals properly and extract the hidden features present inside. A neurological disease can occur in a human brain due to abnormal EEG flow. This abnormality should be properly analyzed to specify the pattern of this disease that can help with prediction of any such type of diseases in human brains. EEG recording generally leads to the collection of a huge amount of numerical information that consists of the state of electrical activity at different time. This recording is generally taken for 10–15 transactions. This duration is sufficient to understand the state of a human brain which leads to collection of huge quantities of data. By plotting this information graphically, we can conclude some behavior of the brain, though not completely. As a result, it is more significant to collect these data and pull information from this. In this research study, the main concentration is on the neurological disorder disease called epilepsy. The whole problem of this research work has been broadly classified into two groups. First, is feature extraction and analysis of a very nonstationary signal like EEG signal. Second, is the classification of EEG signals to detect epileptic seizures.

First, we have developed a well-defined and well-structured process for extracting the hidden features from a very transient and nonstationary signal like EEG signal. For this a signal transformation technique called as discrete wavelet transform (DWT) was used. To compare the significance of these extracted features, other features based on some mathematical computations were also extracted.

Second, we have developed a well-defined and most efficient classifier model that can identify and distinguish epileptic seizures from nonepileptic ones. For this, we have considered mostly ML-based classification techniques like ANN based classifiers, support vector machines (SVM), and evolutionary theory-based classifiers.

1.2 General and Specifi...

Table of contents

• Cover image
• Title page
• Table of Contents
• Copyright
• Preface
• Chapter 1: Introduction
• Chapter 2: Literature Survey
• Chapter 3: Empirical Study on the Performance of the Classifiers in EEG Classification
• Chapter 4: EEG Signal Classification Using RBF Neural Network Trained With Improved PSO Algorithm for Epilepsy Identification
• Chapter 5: ABC Optimized RBFNN for Classification of EEG Signal for Epileptic Seizure Identification
• Chapter 6: Conclusion and Future Research
• References
• Index