Removal of cardiac and respiratory artifacts from EEG recordings under increased intracranial pressure

An automatic approach is presented to isolate and remove the electrocardiogram and respiration waveform artifacts from the electroencephalogram (EEG) recordings. The collected signals from the laboratory rabbit model of intracranial pressure increased are decomposed linearly into independent components by the extended-ICA. The artifactual components can be identified automatically by spectrum analysis. The approach is able to process a long period of data continuously with successive data segments, which can create artifact-reduced EEG signals. To evaluate the performance, the power spectra and the relative wavelet energy (RWE) were calculated, which show the approach can identify the artifacts correctly and suppress them strongly.