[Eeglablist] Doubt about provided data (first analysis)

[Makoto Miyakoshi]

Dear Juan,

Amplitude level of -110 dB/Hz^2 is like 100000 times smaller than my data.
Something is wrong.

Please follow these steps for diagnosis.
0. make sure you are using the latest EEGLAB
1. import your data straight from your raw data
2. copy and paste these lines

figure; bar(std(EEG.data,0,2));

This is to see standard variation for all channels. In the plot, the unit
is microvolt.

figure; plot(EEG.data(1,:))'

This is to see time series of one channel. Here 1 is channel number. Change
it to other channels to see the difference. In the plot, the unit is
microvolt.

3. apply 1 Hz (0.5 is fine) high-pass for all channels and run the above
test 2 again.

What do you see?

Mikolaj's initial guess is right: if the data were DC recorded, the data
can contain huge baseline drift (several thausands) that can completely
mess up spectopo plots.

Makoto

2013/3/15 Juan Carlos Haro Vicente <jharovicente at gmail.com>

> Dear EEGLAB list;
>
> This is my first post, I am a student from Spain who is doing a thesis
> about signal processing using EEGLAB. First of all, I want to say
> thank you for all the info and help that I found here. It has been
> provided to me data from different subjects in an experiment about the
> process of different languages. Once I export this data to EEGLAB I
> have 66 channels of information which look like EEG signals. I extract
> my epochs and here, it comes the problems. When I represent the
> channel spectra and maps it shows me next picture:
>
> http://www.flickr.com/photos/94067049@N03/8558863353/
>
> Which, as I think, it is very differents from what I have to expect.
> If I see a channel, I obtain what I think is normal:
>
> http://www.flickr.com/photos/94067049@N03/8559971178/
>
> And the channel ERP scalp map shows me a signal without sense:
>
> http://www.flickr.com/photos/94067049@N03/8558863377/
>
> After seeing this, I continued with ICA analysis, once I did, I
> obtained next result for spectra maps:
>
> http://www.flickr.com/photos/94067049@N03/8559971126/
>
> Again, far away from what I expected. If I represent components ERP in an
> array:
>
> http://www.flickr.com/photos/94067049@N03/8558863349/
>
> where we can see what it looks like ERPs, but very contaminated by noise.
>
> After say my situation, I would like to ask you if you have any idea
> about what it is the problem with the data. I trust your experience
> with the software and good management of these situations
>
> Thank you very much
> Regards
>
> Juan Carlos Haro
> _______________________________________________
> 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
JSPS Postdoctral Fellow for Research Abroad
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/20130318/30182825/attachment.html>