[FieldTrip] applying ica rejection to differently epoched data

Russell G Port russgport at gmail.com
Thu Oct 2 20:10:58 CEST 2014


Dear All,

Sorry for this repetitive email. I realize one I have asked this question
before, I couldn't find the email before now, so I was not sure if it were
the exact same question. Please ignore this email discussion.

for those who were wondering here is the previous answer

This is the preferred way to do it. Read in all of your data in 1- or
2-second epochs at arbitrary points (just one right after another).
Run ICA on these epochs and save the components that you get (call
this 'comp'). Then make a note of the components that you want to
remove (usually blink, eye movement, and heart). Now read in your data
time-locked to your triggers and use the previously-defined components
to clean up the data (use the cfg to specify which components you want
to remove).

Again thank is due to Dr. Schurger, and my apologizes

On Tue, Sep 30, 2014 at 10:08 AM, Russell G Port <russgport at gmail.com>
wrote:

> Hi All,
>
> I have a question, which I hope someone can help with. Currently I read in
> my dataset (*.ds) via standard fieldtrip commands, epoching 1 second bins
> for the length of the data. After proper removal of all jump and muscle
> artifact, I apply ICA to the resample data (now 300Hz before it was
> 1200Hz). This is just like fieldtrip's website suggests. Then I verify
> which components I want to remove via topoplots and coherence in the time
> domain with the ECG channels. So I now have the list of components that I
> think are artifact and want removed. Can these components be applied for
>  rejection onto the same dataset, but differently parsed into epochs, now
> trials being centered 1 second bins around a trigger? I am trying to see if
> this works, because I hope if I include more data in the ICA analysis (all
> data in the dataset) I will get a better components since there are more
> trials to train the data on, rather than if I based trials around the
> trigger. I guess what I want to know is if I can get better estimates of my
> artifacts via using a larger dataset, and then apply then to the same
> dataset just differently epoched?
>
> Best,
> Russ Port
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.science.ru.nl/pipermail/fieldtrip/attachments/20141002/14484e49/attachment.html>


More information about the fieldtrip mailing list