Spike Detection from EEG Signals with Aid of Morphological Filters and Particle Swarm Optimization (PSO)

In order to detect the abnormality in brain signals, it is essential to study the behavior of spikes in Electroencephalogram (EEG). Normally, recorded EEG signals contain large amount of artifacts like spikes whose detection is a technically challenged one. Morphological filters are generally used to separate these spikes from the recorded EEG signal. In existing techniques, the Gaussian function is used in morphological filter to find out the optimal structuring element. Using this function, the accurate optimal structuring element cannot be found. Here, it is proposed an optimization technique along with a spike detection method using morphological filter. In this method, initially the noise within EEG signals is removed by the wavelet technique and the resultant preprocessed EEG signals are given to the spike detection process. In the proposed method, Particle Swarm Optimization (PSO) is used for the computation of optimal structuring elements in the Morphological filter used for the spike detection. After the computation, an amplitude threshold should be set to detect the occurrence of individual spikes. Hence, the spikes can be detected more effectively by achieving more number of correctly detected spikes rather than the conventional spike detection methods. KeywordsElectroencephalogram (EEG), Morphological Filter, Particle Swarm Optimization (PSO),Spike Detection, Wavelet.