[Eeglablist] How rejection small parts of epoch affects time-frequency analysis?

[Tarik S Bel-Bahar]

Sean, brief notes below.

*if epoching occurs, it's usually for cleaning purposes (to detect brief
artifact periods), or isolating times of interest for entry into ICA (e.g.,
specific epochs around a stimulus onset). ICA doesn't require epochs as it
mixes up time points anyway, as with eeglab it's looking for spatial
stabilities.

*within eeglab, you can generate an ICA solution (from properly cleaned
data, etc..) and then apply it to the unepoched data, and then remove bad
ICs.Just make sure you use data with the same number of channels, and
ideally from the "just before the epoching for cleaning" step.
the application of an ICA solution from one file to another is accessible
via the GUI if two files are loaded into memory, and of course, a script
can be generated by using eegh.

*I'm not sure what you mean by two ICAs, but if you mean running an ICA,
dropping ICs, and then running ICA again, I don't think this is recommended
nor required. If you see it in high-quality studies, however, you have a
precedent to cite for reviewers.

*Also
if you haven't had a chance to yet:
reviewing makotos pipeline suggestions and googleing eeglab list + your
topic  may be helpful.
take a quick look at ASR plugin for interesting alternative



On Wed, May 23, 2018 at 12:52 PM, Sean Gilmore <sean.gilmore at ryerson.ca>
wrote:

> I am currently preprocessing my data using ICA to prep the data for
> time-frequency analysis. My stimuli was administered for ~54 seconds and so
> I am defining each epoch as [1 54] seconds after each trigger. I intend to
> examine the time-frequency information contained across these epochs,
> specifically examining the power of the frequency spectrum.
>
> I have been reading about conducting two ICA in order to obtain clean ICs.
> To do this I have seen some people sub-dividing their epochs into  smaller
> epochs as to prevent the rejection of large amounts of data when rejecting
> artifactual epochs. I would like to implement this pipeline as it seems
> like a good way to clean the data without having to discard large amounts
> of data.
>
> My concern is with rejecting smaller sections of the larger epochs and how
> this will effect time-frequency decompositions. For example, If I ran a FFT
> on a epoch which had smaller sections removed (due to the presence of
> artifacts) would this effect the legitimacy of the results? Likewise, how
> would this effect my ability to obtain phase values from the epoch?
>
> Please let me know.
>
>
> Cheers
> -
> Sean Gilmore, M.A. candidate
> Department of Psychology
> Ryerson University
>
>
>
>
>
>
>
>
> _______________________________________________
> 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
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://sccn.ucsd.edu/pipermail/eeglablist/attachments/20180525/d4b8db89/attachment.html>