Epilepsy classification using discriminant analysis and implementation with space time trellis coded MIMO-OFDM system for telemedicine applications

Abstract

Epilepsy is one of the major neurological disorders occurring in the brain which causes sudden and recurrent seizures. These seizures cause loss of motor control in the human brain thereby bringing dangerous situations in the life of an epileptic patient. The patients tend to develop anxiety problems throughout their life because the seizures occur without any preliminary warnings. Electroencephalogram (EEG) is used to measure the recordings of the human brain and it helps greatly as a tool for analyzing, recognizing, diagnosing and classifying the epilepsy from EEG signals. Since the recordings are pretty huge, it is quite hectic to process the entire data and therefore certain dimensionality reduction techniques are required to reduce the dimensions of the data. In this paper, the dimensions of the data is reduced with the help of Linear Discriminant Analysis (LDA) and then the dimensionally reduced data is transmitted with the help of Space Time Trellis Coded Multiple Input Multiple Output Orthogonal Frequency Division Multiplexing (STTC MIMO-OFDM) System. At the receiver side, the Expectation Maximization Based Gaussian Mixture Model (EM-GMM) is engaged to classify the epilepsy from EEG signals. Also the Bit Error Rate (BER) is analyzed at the receiver side along with the classification accuracy. The performance measures analyzed here are Specificity, Sensitivity, Time Delay, Quality Values, Performance Index and Accuracy.