Automatic artifacts correction: improving on-line EEG analysis

In this paper an automatic ocular artifacts management procedure for EEG analysis on-line is proposed, composed of a detection algorithm followed by a correction which is based on canonical correlation analysis (CCA). The accuracy of the whole method is tested on simulated signals and its capability of recovering the original signals is shown to be comparable with non-automatic ‘gold standard’ procedure (independent component analysis - ICA). The method is implemented to be suitable for fast EEG processing to improve on-line signal interpretation. An example on real data is also provided.