Eeg Denoising Independent Component Analysis Method And Its Application

Different biomedical signals derive from different sources. The sources of EEG, ECG and EOG are respectively brain, heart and eyes. Presently, the brain science research shows that multi-channel EEG data collected from various scalp electrodes placed according to a certain criterion contains rich physiological phenomena and much more dynamic information. The work of EEG source segregation, identification, and localization are very important and difficult in brain science research and cognition science research.Moreover, in clinic, the EEG interpretation and analysis is very essential to diagnosis and treatment of some brain diseases. However, the electrodes on scalp are near located, so the EEG signals collected from them are highly correlated. EEG signals of each channel contain great information about brain activity, but are often contaminated with artifacts and noise, for instance, EOG, ECG, EMG, and line noise. It poses a challenge to researchers of this field who try to extract the useful information reflecting brain state from EEG recordings. Recent efforts to sources identification and artifacts segregation have focused mostly on performing spatial segregation and localization of source activity. Independent Component Analysis(ICA) is one of these methods.As a new array processing technique, ICA is an effective means to resolve the BSS problem. In artifacts removal of EEG data, considering the fact that artifacts and EEG activity have respective independent sources, we use ICA technique to perform the task. Based on the expound and analysis of ICA theory and algorithm, we apply ICA to the removal of different kinds of artifacts from EEG recordings, and the experiments results show that it is a promising method. Compared with the traditional methods of artifacts elimination, ICA, a kind of spatial filter, does not be restricted by the case of spectrum overlapping, and have a good reservation of useful detail signals. In addition, the inverse weight matrix of ICA can be used to reflect the topographic structure of different independent sources of EEG, and therefore have great physiological significance.