Experimental Study on Detection of Epilepsy

Author Name(s): Prof. Bahubali Shiragapur, Prof. Nishikant Surwade, Aishwarya Nair, Shefali Singh, Priya Choudhary
Author Email: bahu2001@gmail.com

Abstract

Epilepsy is a neurological disease which can be referred as a disorder of the central nervous system, considered by the loss of consciousness and seizures. Epileptic patients are subjected to epileptic seizures that cause abnormal excessive brain activity. This condition affects approximately 5% of the world’s population. Ideally, it is possible to detect and predict epilepsy. For detection, we can consider computer electroencephalograph, smart watches etc., whereas for prediction we can use machine learning artificial intelligence, phase correlation. Our objective is to survey different approaches to Epileptic seizure detection. In our work, we have pre-processed the EEG signal of both normal and Epileptic signal by using a Chebyshev filter. Later, we have used Wavelet analysis to decompose the filtered signal into five sub-levels. However, we have considered only the lower sublevel for further processing. The features are extracted using wavelet decomposition for further analysis.

Introduction

Epilepsy is a neurological disorder causing abnormal excessive brain activity. This condition affects approximately 5% of the world’s population. It is the fourth most common neurological disorder and it affects people of all ages. According to the World Health Organization (WHO), it estimates that epilepsy affects 50 million people worldwide [1]. The seizure is called as a single episode whereas multiple episodes of seizure are called epilepsy, which is a disorder. Epilepsy often follows many medical disorders such as traumas, and drug or alcohol intoxication [1]. They are also associated with brain tumours, abscesses, congenital malformation, stroke; which is mainly of two types: One is the Ischemic i.e. obstruction of blood vessels and the other one is Haemorrhage i.e. rupture of blood vessels. Epilepsy begins before 20 years of age in more than 75% of patients. Epilepsy is not synonymous with mental retardation or illness; it is not associated with an intellectual level. The cause for epilepsy in most of the cases is unknown (clinical expert’s opinion). Epilepsy that occurs because of other issues may be preventable. Seizures can be controlled with 20 types of medication, depending on the type of seizures. Inexpensive options are often available. It is not necessary that all cases of epilepsy are lifelong, in some cases it can be cured, if proper medication is being taken [2]. Out of the four lobes of the brain i.e. Frontal, Temporal, Occipital, and Parietal lobes, if the disorder is in Frontal, Occipital, Parietal then the person can be treated with the help of proper medications. But, if it is affected on temporal lobe then the person must undergo surgery because it is the most common indication for surgery. Before that, the person shouldn’t have any other diseases such as high blood pressure, diabetes etc., before undergoing surgery. Seizures beginning at the frontal lobe area of the brain are usually shorter than those that start at temporal lobe area [2].

Conclusion

In this paper we have studied and expressed the cause of epileptic seizure with the help of clinical experts. For experimentation analysis we have extracted the features using DWT followed by Chebyshev filter type-I. The various statistical features are extracted using Wavelet transform, are useful to train ANN for classification. Furthermore, Classification and detection of Epileptic Seizure will be covered in part-II article. We request reader to please refer part-II article for clear understanding of detection and prediction using ANN.

217 total views, 4 views today

Download PDF File

About the author: admin