[Eeglablist] Error in running ICA

[Makoto Miyakoshi]

Dear Edwige,

This typically happens when you have problems in your time-series data.

   1. Check if you have channels with zero or 1000000 uV data.
   2. Check data stationarity in time domain.

These checks can be done by using trimOutlier() plugin.
https://sccn.ucsd.edu/wiki/TrimOutlier
I made this because there was no way to see all channels & all timepoints
in a single figure. Download it via EEGLAB plugin manager and use it.

Most likely you have problem in your data. If you don't see the problem in
using trimOutlier(), you should still go through all the time-series
browsing by eye to find the problem. Send me the screenshot of
trimOutlier() initial report plots if you don't see the problem. Edwige,
this problem can be fixed for sure.

Makoto

On Fri, Jul 13, 2018 at 2:51 AM E. Sijyeniyo <e.sijyeniyo at student.rug.nl>
wrote:

> Dear Makoto,
>
> Thank you for your answers!
>
> However, I think you might not have fully understood what my problem was:
>
> >*Good to discover the problem. Check what was wrong and confirm if ICA
> can be performed successfully.*
>
> I have tried to run ICA how you suggested I should do it, so removing the
> eyes first and not having to click on ‘…channel’ in the Run ICA pop-up
> screen. I have recorded what happens if I try to run ICA for this
> participant. In the video, you can see how it looks like:
>
> https://drive.google.com/file/d/16emh7TdllROu6s074nd6CacQKr5fMPto/view?usp=sharing
>  (you should watch until the end of the video, that’s where the error
> message pops up).
>
> Here you can see the problem, so as soon as I type in the following in*
> Commandline options : ‘extended’, 1, ‘pca’ , [54] and then click ok *, it
> starts lowering the learning rate and then at some point when it gets to
> step 1, it has difficulty going on with further steps and eventually this
> error message pops up. I have tried to re-process the continuous data
> again, the same way I have done all my other datasets, but for some
> reason, the error message only pops up with this particular dataset. So I
> can confirm that ICA cannot be performed successfully, why and how, I do
> not know because the preprocessing procedure is the same as for all my
> other participants where ICA was performed successfully.
>
> *>Yes, you should run ICA again to see if it works.*
> *If ICA takes too long time, you can specify the ICA iteration number by
> using 'maxsteps', 20 to stop at 20th iteration for example.*
>
> Ok, I will run ICA again for these participants even though ICA was
> successful while icasphere and nbchan were NOT the same…
>
> To summarise, as you can see in the video, the error pops up while ICA is
> being performed, so I do not know what happens for this particular
> participant.
>
> Thank you once more for your time!!
>
> Best,
> Edwige
>
> P.S. I have shared the Google drive link in a separate email, in case the
> link in this email does not work accordingly.
>
> Op 13 jul. 2018, om 02:48 heeft Makoto Miyakoshi <mmiyakoshi at ucsd.edu>
> het volgende geschreven:
>
> Good to discover the problem. Check what was wrong and confirm if ICA can
> be performed successfully.
>
>
>

-- 
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/20180713/e8443180/attachment.html>