Ocular artifact correction with 21 EEG channels (+ VEOG and HEOG)

✉️ Mailing List Archive (read-only)
1

Hello,

In our lab, we usually reject epochs that are contaminated by any artifact.
I have a dataset where I would like to analyze the ERN (time-locked to the
response) . Problem: participants generally blink just right after the
response, and most of my epochs are contaminated by blinks. I am thus
looking for an efficient artifact correction method.
My dataset features 21 EEG channels [subset of the International 10/20
System sites: Fz, Cz, Pz, F3, F4, C3, C4, P3, P4, PO3, PO4, T3,
T4, T5, T6, O1, and O2 and the nonstandard sites OL (halfway
between O1 and T5) and OR (halfway between O2 and T6)] + VEOG (bipolar) and
HEOG (bipolar)]. Given this small number of electrodes, I don't think it is
appropriate to use ICA. Any suggestion?

Mathieu
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.nmr.mgh.harvard.edu/pipermail/mne_analysis/attachments/20170804/32165961/attachment.html

2

You can generate SSP projectors centered around the blink events, and use
those to project out the blink artifacts. Use
mne.preprocessing.create_eog_epochs and mne.compute_proj_epochs

3

If only a small number of channels are available, it may be good to regress out the EOG instead. See:
https://github.com/mne-tools/mne-sandbox/blob/master/examples/preprocessing/plot_eog_regression.py

Marijn.