Predicting Epilepsy and Deep Learning

Mandana Sadat Ghafourian*; Mohammad Teshnehlab; Mohammad Hassan Moradi; Farveh Daneshvarfard

Epilepsy is one of the most common nervous diseases, which occurs unconsciously and unpredictably due to the brain transient disorders. In this paper, deep neural networks is used to predict epilepsy attacks thorough simultaneous use of EEG signals and Heart Rate Variability (HRV) analysis on a public database containing 8 patients. Deep neural networks is a type of neural architectures, which has more than one hidden layer and capable of better generalization in comparison with conventional neural networks. In this work, eight features are extracted from HRV signals in time and frequency domains...