[Eeglablist] basic question about artifact rejection using ICA

[Makoto Miyakoshi]

Dear Vivian,

(1) Is there a good how-to document for deciding which components to reject
based on the features of the ICA (e.g., distribution of activity over the
scalp, time, trials, power spectrum, etc.)?

Don't discard any IC by hand if you will use STUDY; it will automatically
threshold your ICs based on 1. residual variance of 15% (by default) and 2.
outside-brain IC rejections. Otherwise, you have to learn yourself which
ICs are good in terms of scalp topography, spectra, ERP and ERP images, etc.

(2) I have questions about 6 components (here's a
link<https://www.dropbox.com/s/yikdd2fpfwx2ziu/forEEGLABlistserv.pdf>
to
a document with their graphs). Any thoughts about whether these components
should be kept or rejected, and whym would be much appreciated!

They are all good in term of ERP and ERP images, but interestingly bad in
scalp topography patterns. Two suggestions. 1. Go back to continuous data
and apply 1-Hz high-pass filter 2. Check channel locations again. Is it
rotated by 90 degrees by any chance? Did you press 'optimize center' in the
channel location edit window on the top right?

Makoto


2014-03-24 14:39 GMT-07:00 Vivian Zayas <vz29 at cornell.edu>:

>  Hi,
>
> I'm using ICA for artifact rejection and have questions about whether to
> keep or reject a few components.
>
>  (1) Is there a good how-to document for deciding which components to
> reject based on the features of the ICA (e.g., distribution of activity
> over the scalp, time, trials, power spectrum, etc.)?
>
>  (2) I have questions about 6 components (here's a link<https://www.dropbox.com/s/yikdd2fpfwx2ziu/forEEGLABlistserv.pdf> to
> a document with their graphs). Any thoughts about whether these components
> should be kept or rejected, and whym would be much appreciated!
>
>  Best,
> V
>    ___________________________
> Vivian Zayas, Ph.D.
> Associate Professor
> Cornell University
> Department of Psychology
> 238 Uris Hall
> Ithaca, NY 14853-7601
> Email: vz29 at Cornell.edu
> Phone: 607-254-6332
> Fax: 607-255-8433
> Lab website: http://people.psych.cornell.edu/~pac_lab
> Faculty website: http://comp9.psych.cornell.edu/people/Faculty/vz29.htm
>
>
>
>
> _______________________________________________
> Eeglablist page: http://sccn.ucsd.edu/eeglab/eeglabmail.html
> To unsubscribe, send an empty email to
> eeglablist-unsubscribe at sccn.ucsd.edu
> For digest mode, send an email with the subject "set digest mime" to
> eeglablist-request at sccn.ucsd.edu
>



-- 
Makoto Miyakoshi
Swartz Center for Computational Neuroscience
Institute for Neural Computation, University of California San Diego
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://sccn.ucsd.edu/pipermail/eeglablist/attachments/20140401/898a3f04/attachment.html>