[Eeglablist] Artifact detection and ICA on data with multiple epoch types

[Mikołaj Magnuski]

Dear Becky,

good to hear of your success. :)
Yes, I was thinking about eeg_epoch2continuous and I advise you to stick
with it instead of trying epoch_linearize (I wote it once just for some
tests with ICA, its definitely less reliable than an eeglab function). The
other function I mentioned (for cutting continuous data into 'fake' epochs
for further ICA analysis) - 'onesecepoch' works ok and has some nice
options, but you should be fine with you current pipeline.
Concerning your question, removing ICA components should be done on the
data you want to remove them from (epoched data in your case). This is
because apart from changing the ica fields you also remove the signal that
given component contributes to the data - and you want to remove it from
the data you will be working on.
Good luck!
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://sccn.ucsd.edu/pipermail/eeglablist/attachments/20130509/6b46dc65/attachment.html>