International Journal of Computer Applications® (IJCA) (0975 – 8887) International Conference on Simulations in Computing Nexus, ICSCN-2014 1 Analysis of EEG Signal for the Detection of Brain Abnormalities M.Kalaivani PG Scholar Department of Computer Science and Engineering – PG National Engineering College Kovilpatti, Tamilnadu V.Kalaivani, Ph.D Associate Professor(SG) Department of Computer Science and Engineering – PG National Engineering College Kovilpatti, Tamilnadu V.Anusuya Devi Assistant Professor Department of Computer Science and Engineering – PG National Engineering College Kovilpatti, Tamilnadu ABSTRACT In the field of medical science, one of the major ongoing researches is the diagnosis of the abnormalities in brain. The Electroencephalogram (EEG) is a tool for measuring the brain activity which reflects the condition of the brain. EEG is very effective tool for understanding the complex behaviour of the brain. The aim of this study is to classify the EEG signal as normal or abnormal. It is proposed to develop an automated system for the classification of brain abnormalities. The proposed system includes pre-processing, feature extraction, feature selection and classification. In pre-processing the noises are removed. The discrete wavelet transform is used to decompose the EEG signal into sub-band signals. The feature extraction methods are used to extract the time domain and frequency domain features of the EEG signal. General Terms Methodology for Information in brain abnormality using EEG Signal. Keywords Electroencephalogram, brain diseases, wavelet transform, EEG waves, feature extraction 1. INTRODUCTION A disease is an abnormal condition that affects the body of an organism. Any deviation from the normal structure of a body part or organ is displayed by a characteristic set of symptoms and sign. Electroencephalogram is used for detecting the brain diseases. Electroencephalogram is the recording of electrical activity of the brain from scalp. It measures the voltage fluctuations resulting from ionic current flows within the neurons of the brain. Diagnostic applications generally focus on spectral content of EEG that is the type of neural oscillations that can be observed in EEG signals. EEG is painless and harmless. And it does not pass any electricity into your brain or body. The EEG signals are commonly decomposed into five EEG sub-bands: delta, theta, alpha, beta and gamma. Alpha waves are rhythmic and its frequency range is from 8 to 13 Hz. The amplitude of the alpha wave is low. Each region of the brain has the characteristic of alpha rhythm but mostly it is recorded from the occipital and parietal regions. It oscillates from adult in awake and relaxed state with eyes closed. Beta waves are irregular and its frequency range is greater than 13 Hz. The amplitude of the beta wave is very low. It is mostly recorded from temporal and frontal lobe. It oscillates from during the deep sleep, mental activity and is associated with remembering. Delta waves are rhythmic and its frequency range is 4 to 7 Hz. The amplitude of the delta wave is high. It oscillates from the children in sleep state, drowsy adult and emotional distress occipital lobe. Theta waves are slow and its frequency range is less than 3.5 Hz. The amplitude of the theta wave is low-medium. It oscillates from adult and normal sleep rhythm. Gamma waves are the fastest brainwave frequency and its frequency range is from 31 to 100 with the smallest amplitude. Figure 1 Normal EEG waves In the proposed work the EEG signals are given as input to the pre processing. From the pre processing the discrete wavelet transform are used to remove noises and the EEG signal are decomposed into five sub-band signals. The non linear parameters (time and frequency) were extracted from each of the six EEG signals (original EEG, delta, theta, alpha, beta and gamma). A genetic algorithm was used to extract the best features from the extracted time and frequency domain features. Then the classifier is used to classify the given EEG signal as normal or abnormal. 2. RELATED WORK Some literature survey has been focused for the pre- processing of EEG signals, Feature extraction, Feature selection and Classification methods. Siuly [1] has proposed a cross correlation based LS-SVM [4] [6] for improving the classification accuracy of EEG signals. Sabeti M [2] uses the discrete wavelet transform for preprocessing [4] [9] and genetic algorithm, which is used to select the best features from the extracted features. The two classifiers SVM [4] and LDA are used to classify the EEG signal abnormalities. Stevenson N J [3] has developed the automated grading system for EEG abnormality in neonates. Multiple linear discriminant classifier are used to classify the EEG abnormality in neonates with HIE. Marcus [5] has presented the time-frequency distributions of EEG signals. Here the SVM are used to classify the epilepsy from EEG signals.