EEG epileptic seizure detection using k-means clustering and marginal spectrum based on ensemble empirical mode decomposition

The detection of epileptic seizures is of primary interest for the diagnosis of patients with epilepsy. Epileptic seizure is a phenomenon of rhythmicity discharge for either a focal area or the entire brain and this individual behavior usually lasts from seconds to minutes. The unpredictable and rare occurrences of epileptic seizures make the automated detection of them highly recommended especially in long term EEG recordings. The present work proposes an automated method to detect the epileptic seizures by using an unsupervised method based on k-means clustering end Ensemble Empirical Decomposition (EEMD). EEG segments are obtained from a publicly available dataset and are classified in two categories “seizure” and “non-seizure”. Using EEMD the Marginal Spectrum (MS) of each one of the EEG segments is calculated. The MS is then divided into equal intervals and the averages of these intervals are used as input features for k-Means clustering. The evaluation results are very promising indicating overall accuracy 98% and is comparable with other related studies. An advantage of this method that no training data are used due to the unsupervised nature of k-Means clustering.