A Review of Electroencephalogram Signal as Clinical Decision Support System

Systematic Reviews in Pharmacy,2018,10,1,49-54.
Published:December 2018
Type:Review Article

A Review of Electroencephalogram Signal as Clinical Decision Support System

Usha Govindarajan1, Narasimhan Kumaravelu2

1Department of Electronics and Communication Engineering, SRC, SASTRA Deemed University, Kumbakonam 612001, Tamilnadu, INDIA.

2School of Electrical and Electronics Engineering, SASTRA Deemed University, Thanjavur 613401, Tamilnadu, INDIA.

Abstract:

Computer Aided Diagnosis (CAD) systems have been helping us in various fields including the medical diagnosis of various diseases where the CAD generated output is taken as complementary to the doctor’s view. EEG signals which pick up the electrical activity of the brain are fed into a CAD system to further analyze and confirm the underlying abnormality. Neurological disorders affect a person and can make one’s day to day activities difficult which mainly deals with the brain’s electrical activity. Autism and Epilepsy are two of the major neurological disorders which can be analyzed through CAD systems. Alzheimer’s disease can also be diagnosed with the help of a CAD system. Signal processing is used to take the input EEG signal and then process or decompose it to help us in analyzing. Few of the many processing techniques include FFT, DCT, DWT etc., Four features are extracted from a processed signal like Shannon entropy, Band power, Standard deviation and Largest Lyapunov Exponent which enhance the study of the signal. The feature vector is formed which is then fed to different classifiers whose main purpose is to classify as a normal or an abnormal signal. Some of the classifiers used KNN, ANN, SVM and LDA in which KNN provides the best help in recognizing the abnormality in a short period of time. In this paper we study the usage of few techniques, features, classifiers and give emphasis to the advantages of using DWT, Shannon entropy and KNN which helps us get an efficient and better understanding of the abnormality.

Keywords:ANN, CAD, EEG, KNN, LDA, LLE, SVM