Analysis and classification of EEG signals using spectral analysis and recurrent neural networks

This study proposes a three stages technique for automatic detection of epileptic seizure in EEG signals. In practical application of pattern recognition, there are often diverse features extracted from raw data which needs to be recognized. Proposed method is based on time series signal, spectral analysis and recurrent neural networks (RNNs). Decision making was performed in three stages:(i)feature extraction using Welch method power spectrum density estimation (PSD) (ii)dimensionality reduction using statistics over extracted features and time series signal samples (iii)EEG classification using recurrent neural networks. This study shows that Welch method power spectrum density estimation is an appropriate feature which well represents EEG signals. We achieved higher classification accuracy (specificity, sensitivity, classification accuracy) in comparison with other researches to classify EEG signals exactly 100% in this study.