Abstract
Epilepsy is the most unpredictable and recurrent disease among neurological diseases. Early detection of epileptic seizures can play a critical role in providing timely treatment to patients especially when a patient is in a remote area. This paper uses deep learning framework to detect epilepsy in the Electroencephalography (EEG) signal. The dataset used is publicly available and has a recording of three kinds of EEG signals: pre-ictal, inter-ictal (seizure-free epileptic) and ictal (epileptic with seizure). The proposed Long Short-Term Memory (LSTM) classifier classifies these three kinds of signals with up to 95% accuracy. For binary classification such as detection of inter-ictal or ictal only, its accuracy increases to 98%. The EEG signal is modelled as wide sense non-stationary random signal. Hurst Exponent and Auto-regressive Moving Average (ARMA) features are extracted from each signal. In this work, two different configurations of LSTM architecture: single-layered memory units and double-layered memory units are also modelled. After standardising the features, double-layered LSTM approach gives the highest accuracy in comparison to previously used Support Vector Machine (SVM) classifier and proved to be computationally efficient at Graphics Processing Unit (GPU).
Highlights
Epilepsy is a harmful disease which affects millions of people around the world
We compared our results with the work of Gupta [2] which used a 10-fold Support Vector Machine (SVM) classification technique to predict the ictal EEG signals
We modelled the EEG signal as Brownian movement of brain neurons and Hurst Exponent with Autoregressive Moving Average (ARMA) features used to train the Long Short-Term Memory (LSTM) classifier
Summary
Epilepsy is a harmful disease which affects millions of people around the world. It is a brain disease which occurs due to a chronic neurological disorder of neurons producing an abnormal signal. Machine learning approaches are used to predict the epileptic seizure which includes EEG signal acquisition, pre-processing, feature extraction and the classification of the seizure states. Threshold-based techniques are used for the prediction of seizure where the analysis is targeting a high/low change in the values of some features during the pre-ictal stage. The Adaptive Neural Network (ANN) and SVM are both widely used by researchers [5]–[8] In these schemes, the EEG signal data are classified by features which are extracted from the recorded EEG data signal. The pre-processing and feature extraction from an EEG signal plays an important role in improving true positive rates and prediction time. This is done to avoid the biasing in LSTM training towards any particular data class. 4) the experimenting is done for cases like pre-ictal vs ictal, pre-ictal vs inter-ictal, inter-ictal vs ictal, pre-ictal vs inter-ictal vs ictal
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
