source coherence in one condition

[Eric Maris]

Dear Andrew,



> Thank you very much for the code and the answers. It helps to clarify
things alot.
> My understanding of coherence was that if the phase difference between two
signals
> is confined to within 2pi over some time interval, they are phase locked,
ie the
> temporal variations in the signals tend to follow each other. But that
this is
> regardless of their differences in amplitudes. When amplitude also
covaries then
> there is 'true' coherence. I thought that you could test for signifant
coherence
> between two signals.

I would forget about the influence of amplitude covariation on coherence (on
top of phase consistency). This is a side-issue for the problem that you are
addressing (testing for significant coherence in a single condition).

>
> For example in BESA you can do permutation tests on individual subjects
for
> 'significant' coherence between a reference dipole and the rest of the
dipoles in a
> model in one condition. I might be misunderstanding something though
becuase I
> was wondering if one could do this on a group level by importing the
coherence data
> from BESA into fieldtrip. I suppose this would be coherence relative to
some
> baseline, but I am not sure if that makes any sense. I don't know exactly
how BESA
> calculates coherence.
>
> So I guess H0 would be that there is no coherence between two dipoles that
is not
> due to chance in a particular condition - is that a possible null
hypothesis for
> coherence?

In my opinion, it does not make sense to test the null hypothesis that
coherence between two EEG/MEG-channels is equal to zero. The reason for this
is biophysical: due to volume conduction, the activity produced by a given
source will be picked up by all EEG/MEG-channels, and therefore it does not
make sense to test the null hypothesis that coherence is zero.

Testing coherence relative to baseline certainly makes sense, and is
possible in Fieldtrip, although it is not documented very well. (I remember
a work-around posted by Jan-Mathijs Schoffelen for testing coherence
differences in within-subjects designs.) However, you should keep in mind
that significant coherence-differences can also be produced by differences
in power. This is because coherence not only measures the phase consistency
between the signals at the source level but also the signal-to-noise ratio
(SNR) at the different sensors: For a fixed source-level phase consistency,
the larger the SNR, the larger the coherence.

Greetings,





dr. Eric Maris
NICI/Biological Psychology and
F.C. Donders Center for Cognitive NeuroImaging
University of Nijmegen
P.O. Box 9104
6500 HE Nijmegen
The Netherlands
T:+31 24 3612651 (NICI)
T:+31 24 3610754 (FCDC)
F:+31 24 3616066 (NICI)
E: maris at nici.ru.nl
MSc Cognitive Neuroscience :www.ru.nl/master/cns/


>
> Thank you again for all your help!
>
> andy
>
> ----- Original Message -----
> From: Robert Oostenveld <r.oostenveld at FCDONDERS.RU.NL>
> Date: Monday, May 7, 2007 3:12 pm
> Subject: Re: [FIELDTRIP] source coherence in one condition
> To: FIELDTRIP at NIC.SURFNET.NL
>
>
> > On 26 Apr 2007, at 0:35, Andrew Smart wrote:
> >
> >  > I would like to try to compute whether there is significant
> >  > coherence across
> >  > subjects but within one condition and between two dipoles. This code
> >  > generates graphs but I am unclear about what they mean:
> >
> >  Hi Andrew,
> >
> >  It is unclear to me what the comparison is in the statistical test
> >  that you want to perform. There should be a hypothesis like
> >  "coherence is larger in XXX than YYY", which rephrased would be H0:
> >
> >  the coherence is the same. Based on the H0, the probablility can be
> >
> >  computed and H0 can (perhaps) be rejected. You cannot say that
> >  "coherence is significant" just like that.
> >
> >  If the hypotheis is that coherence between one dipole pair is larger
> >
> >  than between the other dipole pair, then I am affraid that you cannot
> >
> >  test that using a permutation approach. With a permutation or
> >  randomizarion test you test the hypothesis that the data in two
> >  observed conditions is exchangeable. Simple differences in power in
> >
> >  one of the 4 sources underlying the two dipole pairs would already be
> >
> >  enough to reject the null-hypothesis that the data is exchangeable,
> >
> >  but those power differences do not have to mean that there is a
> >  coherence difference.
> >
> >  Robert
> >
> >  ----------------------------------
> >  The aim of this list is to facilitate the discussion between users of
> > the FieldTrip  toolbox, to share experiences and to discuss new ideas
> > for MEG and EEG analysis. See also http://www.ru.nl/fcdonders/fieldtrip.
> >
>
> ----------------------------------
> The aim of this list is to facilitate the discussion between users of the
FieldTrip
> toolbox, to share experiences and to discuss new ideas for MEG and EEG
analysis.
> See also http://listserv.surfnet.nl/archives/fieldtrip.html and
> http://www.ru.nl/fcdonders/fieldtrip.

----------------------------------
The aim of this list is to facilitate the discussion between users of the FieldTrip  toolbox, to share experiences and to discuss new ideas for MEG and EEG analysis. See also http://listserv.surfnet.nl/archives/fieldtrip.html and http://www.ru.nl/fcdonders/fieldtrip.