Automatic Preprocessing of EEG Signals in Long Time Scale

Electroencephalography (EEG) signals are highlyaffected by physiological artifacts. Establishing a robust andrepeatable EEG pre-processing is fundamental to overcomethis issue and be able to use fully EEG data especially in longtime scale experiments. In this work, starting from theIndependent Component Analysis (ICA) of the EEG data, acontrol feedback scheme aiming to manage the cleaning of theindependent component signals in an automatic way avoidingcut-bind solutions is presented, both with and without coregistrations.The method implemented combines different approachesbased on the residual artifact contents check, identification andcleaning. The results of this procedure are shown on a testdataset. This analysis tool is embedded as core module, in aplatform that can manage the automatic clearing of EEGrecordings for multiple-subjects studies.