[Eeglablist] Doubt about provided data (first analysis)

[Mikołaj Magnuski]

Ok, sorry, so filtering wasn't the problem.
(I had a similar problem some time ago, and
highpass - about 0.1 or 0.5 helped, but as I
think about it - my spectra were very different from
yours - without highpass filtering you usually get
funny saw-like ups and downs in the spectrum).

I assume you have performed baseline subtraction
(either subtracting mean of a channel from that channel
for all channels or the same but with respect to epochs -
taking the mean from the whole epoch OR epoch
baseline [for example prestimulus period]).

How does your signal look like in eeglab scroll-plot?
(plot --> channel data (scroll) )

(BTW when trying to remove AC it is much better to
use CleanLine than notch-filtering; CleanLine works
really well and does not distort your signal spectrum
or phase)

Pozdrawiam,
Mikołaj Magnuski


2013/3/18 Juan Carlos Haro Vicente <jharovicente at gmail.com>

> Dear Mikołaj;
>
> I have been trying different things. For example, If i do a band pass
> filter between 1-35Hz:
>
> http://www.flickr.com/photos/94067049@N03/8567343259/
>
> Later I'll see this spectra maps:
>
> http://www.flickr.com/photos/94067049@N03/8567343223
>
> As you can see, its not a good result and the filter doesnt works as I
> wanted (I have readen that, it is maybe because I have applied both
> limits to the filter at same time and, sometines, it doesnt work
> correctly). So, I have tried with two low pass filters with higher
> edge at 35 and 50Hz because I remember that my data is yet filtered
> for DC component:
>
> http://www.flickr.com/photos/94067049@N03/8567343253
>
> http://www.flickr.com/photos/94067049@N03/8567343247
>
> And the spectra maps:
>
> http://www.flickr.com/photos/94067049@N03/8567343217
>
> http://www.flickr.com/photos/94067049@N03/8568441508
>
> This seems to have more sense, but I continue without undertanding
> what it is that components with 0 magnitude.
>
> Taking 50Hz filtered signal and doing ICA analysis, I obtain:
>
> http://www.flickr.com/photos/94067049@N03/8568441526
>
> Where you can see clearer than before the ERPs
>
> Thanks for your help, I am a little bit lost and this email list is
> helping me really!
>
> Regards
>
> 2013/3/17 Mikołaj Magnuski <imponderabilion at gmail.com>:
> > Dear Juan,
> >
> > it looks to me as if you didn't high-pass filter the signal.
> > Try filtering (tools>filter the data>basic FIR filter (new) )
> > and see if it looks any better.
> > Good luck!
> >
> > Mikolaj Magnuski
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://sccn.ucsd.edu/pipermail/eeglablist/attachments/20130318/78cbfd3f/attachment.html>