ERPLAB Studio: Artifact Correction and Rejection - ucdavis/erplab GitHub Wiki

ERPLAB Studio provides three types of methods for minimizing artifacts in the data-

1. ICA-based artifact correction. This method uses EEGLAB's independent component analysis routines to estimate and remove the artifactual signals.
2. Artifact rejection in continuous EEG. This method deletes periods of artifactual data from the EEGset. You can determine which periods to delete by visual inspection or by means of an automated algorithm. This is mainly used to delete periods of "crazy noise" prior to ICA (which makes ICA work better). not to eliminate ordinary artifacts.
3. Artifact detection in epoched EEG. This method uses either visual inspection or automated algorithms to "mark" epochs containing artifacts. The marked epochs still remain in the dataset, but they are excluded when averaged ERPs are created (unless you tell the averaging routine to include them). If an artifact is confined to a single channel, it is possible to instead interpolate the voltages for the channel in epochs containing artifacts using the Interpolate marked epochs option in the Interpolate Channels panel.

For details, click the links above for each of these methods.

For a conceptual overview of artifact correction and rejection, see Chapter 6 in An Introduction to the Event-Related Potential Technique, 2nd Edition.

For a detailed description of how to implement these concepts in real data, see Chapters 8 and 9 in Applied Event-Related Potential Data Analysis.

The ERPLAB Studio Tutorial also contains brief demonstrations of artifact correction and artifact detection.