[Eeglablist] EEG Coherence combination

[Makoto Miyakoshi]

Dear Tahani,

Check out SIFT by Tim Mullen
http://sccn.ucsd.edu/wiki/SIFT

When you compute coherence measures, it generates EEG.Conn.... you can
simply average them across all channels. FYI, averaging across column gives
you information outflow, and across raw gives you information inflow.

Makoto


2013/5/29 T. Almabruk <t.almabruk at yahoo.com>

> Hi List
> I have a question about combining coherence measurement for multi
> electrodes. I mean If I calculate the coherence between each pair in a
> dataset of 40 electrodes. Is there any way to combine theses measurements
> in one measure I need a way better that averaging. or any idea that help me
> to have just one measurement instead of this big number of measurements
>
> Cheers,
> Tahani
>
>    *From:* Makoto Miyakoshi <mmiyakoshi at ucsd.edu>
> *To:* T. Almabruk <t.almabruk at yahoo.com>
> *Sent:* Thursday, April 25, 2013 3:40 AM
> *Subject:* Re: [Eeglablist] EEG Coherence
>
> Dear Tahani,
>
> Ok I see. Most likely, the difference between the two results are due to
> the difference of parameters you used. Unfortunately I'm not familiar nor
> interested in pop_newcrossf() especially at the scalp channel level-because
> we have SIFT- I won't compare the details in these two methods.
>
> So I have no answer to your question technically, but starategically I can
> recommend that you use SIFT instead. It addresses many theoretical problems
> you run into when performing coherency analysis. Google SIFT Tim Mullen. He
> is one of our genius guys.
>
> Makoto
>
>
> 2013/4/19 T. Almabruk <t.almabruk at yahoo.com>
>
> Dear  Makoto
>
> Thanks alot for your reply. I will try to explain my problem so that may
> give you an idea about what i am trying to do. Simply i try to make a
> comparsion between calculating cross-coherence between two channals using
> first pop-newcrossf() and second by Mscoher.
> I followed the next steps:
> 1- I filtered the signals
> 2- extract the epoches
> 3-run ICA
> 4- i used pop_newcrossf() to calculate the cross-coherence between channel
> CP1(17) and CP2(18) in the first case as can be seen in the attached
> figures.
>
> http://www.flickr.com/photos/94417587@N05/
>
>  and then repeated the same steps with channel T7(11) and channel T8 (15).
>
> on the othere hand,
>  i calculated the cross-coherence between the same channals by using
> MSCOHER matlab function, i used the data after extracting epoches but
> without runing ICA.
>
> fs=128;
> % sampling frequency
> nfft=256;
> % width of subseries (# of samples)
>  ww=256;
> % width of hanning window applied to each subseries
> over=0;
> s_1=EEG.data(11,:);
> s_2=EEG.data(15,:);
>  q=mscohere(s_1,s_2,hanning(ww),over,nfft,fs);
>  figure
>  plot(q(1:50),
> 'b');
> ylabel(
> 'Coherence');
> xlabel(
> 'Frequences');
>
> *I can see a difference between the results in both cases i mean the
> coherence between the channals by using newcross() and that by using
> Mscoher!!!!*
> **
> *I am sure that i still missunderstand something*
> *and i'm willing to get help from you*
> **
> *Cheers,*
> *Tahani*
>
>   *From:* Makoto Miyakoshi <mmiyakoshi at ucsd.edu>
> *To:* T. Almabruk <t.almabruk at yahoo.com>
> *Cc:* "eeglablist at sccn.ucsd.edu" <eeglablist at sccn.ucsd.edu>
> *Sent:* Tuesday, April 2, 2013 8:36 AM
> *Subject:* Re: [Eeglablist] EEG Coherence
>
>  Dear Tahani,
>
> Looks like your channel 1 and 2 are very much alike. Did you choose two
> ICs? Did you choose subspace of alpha or mu that are similar to each other?
>
> If there were coherence of near 1 between two ICs that do not belong to
> the same subspace (and I guess that's the case here), something should be
> wrong because they are far from independent. Why don't you check the setup
> and tell us more detail about the comparison?
>
> Makoto
>
>
>
>
> 2013/3/27 T. Almabruk <t.almabruk at yahoo.com>
>
>
>
>    Hi List
>
> I asked a question before about measuring EEG coherence by EEGLAB. I got
> benifits from the answer i got  but still have something wrong. I applied
> the function to the same dataset provided with EEGLAB tutorial. i filtered
> the data first and then extract the epoches (153). After that i run ICA and
> ploted component cross-coherence between channel 4 and channel 9.
> As you can see fig(1) was the result and i tried to zoom it in (fig 2).
> I wanted to compare the result with that can be done by matlab function
> (Mcoher) and fig(3) represents the results i got.
> My question is that, there is a big difference in the coherence's
> amplitude between the two functions where using EEGLAB the amplitude is
> less that 0.1 where with Mcohere as you can see is close to 1.
>
> I'm sure i miss undeerstand something because i'm new with this area and
> hope i can find some help from you .
>
> Cheers,
> Tahani
>
>
> http://www.flickr.com/photos/94417587@N05/
>
>
>
> _______________________________________________
> 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
> Swartz Center for Computational Neuroscience
> Institute for Neural Computation, University of California San Diego
>
>
>
>
>
> --
> Makoto Miyakoshi
> Swartz Center for Computational Neuroscience
> Institute for Neural Computation, University of California San Diego
>
>
>


-- 
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/20130530/2693eff7/attachment.html>