Novel feature for identification of focal EEG signals with k-Means and fuzzy c-means algorithms

In this paper, a new method for automatic identification of focal electroencephalogram (EEG) signals is proposed. Detection of focal EEG signals locates the epileptogenic area which is an important task for successful surgery. The proposed method is based on empirical mode decomposition (EMD) that uses the ratio of amplitude modulation bandwidth (B_AM) and frequency modulation bandwidth (B_FM), as a feature for identification of focal EEG signals. The feature average bandwidths ratio (AvgB_ratio) extracted from analytic intrinsic mode functions (IMFs) is set to input in k-Means and fuzzy c-mean (FCM) unsupervised learning. Statistical test Kruskal-Wallis shows the effective discrimination ability of the feature. The experimental results shows that proposed method is precisely proficient to classify focal and non-focal EEG signals using single narrow frequency band. A comparative analysis of both unsupervised learning techniques is performed by elapsed time, time complexity, and accuracy.