[FieldTrip] Automatic artifact detection (no rejection)

[Stephen Politzer-Ahles]

I second Phil's suggestion. While I haven't done this myself, what i would
do if I were in this situation is something like the following:

1. Create a dataset with the shorter epochs
2. Run the artifact rejection
3. Save the list of good trials (trials that weren't rejected)
4. Throw out that dataset, and make a new dataset with the epoch length I
actually want
5. Directly use the list of good trials at whatever stage is relevant
(ft_timelockanalysis or whatever) to make sure you're only keeping the
trials that would have made it through artifact rejection

Best,
Steve



---
Stephen Politzer-Ahles
University of Oxford
Language and Brain Lab
Faculty of Linguistics, Phonetics & Philology
http://users.ox.ac.uk/~cpgl0080/


>
> Message: 4
> Date: Fri, 1 Jul 2016 14:10:09 +0200
> From: Philipp Taesler <p.taesler at uke.de>
> To: <fieldtrip at science.ru.nl>
> Subject: Re: [FieldTrip] Automatic artifact detection (no rejection)
> Message-ID: <15e80a9d-3eba-0c7b-cd77-af2e5f1d80c4 at uke.de>
> Content-Type: text/plain; charset="utf-8"
>
> Hi Robin,
>
> you are right, the comment in the function explicitly says:
>
> % Contrary to the other artifact detection functions, this function
> % will mark the whole trial as an artifact if the threshold is exceeded.
> % Furthermore, this function does not support artifact- or filterpadding.
>
> I didn't know that. Hmmm, then I guess your only option is to either
> visually review the marked trials and unmark the ones you want to keep.
> Another possibilty might be to use ft_redefinetrial to create a dataset
> with shorter trials, run the rejection on this data, and use the
> resulting artefact definitions for your original data. (Not sure if that
> is straightforward, since I never did this myself).
>
> Cheers,
> Phil
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.science.ru.nl/pipermail/fieldtrip/attachments/20160701/9ec194e0/attachment-0002.html>