Separation of Sources From Single-Channel EEG Signals Using Independent Component Analysis

The electroencephalogram (EEG) signals are often mixed with several sources such as electrooculogram and electromyogram signals. Independent component analysis (ICA) is often used to separate the sources from the multichannel EEG signals. Recently, the use of portable EEG devices has gained much attention due to their low power consumption and the ability to record the EEG signals in home environment. However, these systems are equipped with single or few EEG channels, and hence the direct application of ICA is not possible. In this paper, we proposed an efficient technique to separate the sources from single-channel EEG signals by combining the singular spectrum analysis (SSA) and ICA techniques. In this technique, the single-channel EEG data are first decomposed into multivariate data using SSA. Later, ICA is applied on the multivariate data to extract the source signals. In order to validate the performance of the proposed technique, we carried out the simulation on synthetic and real life EEG signals. In addition, we have also studied the performance of the proposed technique for detecting the seizures. The performance measures of a seizure detection classifier such as receiver operating curve, true positive rate, and false positive rate are obtained and found that the proposed technique exhibits better performance compared with the existing methods.