Significance of coherence

Vladimir Litvak v.litvak at ION.UCL.AC.UK
Wed Jul 16 13:44:41 CEST 2008


Dear JM,

Is there any paper where the issue of shift-predictors for coherence
is properly looked at? I've also been using shift-predictors but I
started having serious doubts in the case of MAR-based measures. So I
thought that for coherence there is something better justified and the
distribution under the null is known at least under some assumptions.
But if you say that shift-predictor is state of the art, I'll take
your word for it since I know you've researched this subject quite
well.

Thanks,

Vladimir

On Wed, Jul 16, 2008 at 11:56 AM, jan-mathijs schoffelen
<j.schoffelen at psy.gla.ac.uk> wrote:
> Dear Vladimir,
>
> I usually use a shift predictor to test for this (computing coherence
> between signal A and a shuffled set of trials for signal B: provided both
> signals are not phase-locked to trial onset).
> However, this only works well when you are estimating for coherence between
> non-volume conducted activities.
> When coherence is estimated between two reconstructed dipoles, the amount of
> volume conduction is generally too high for this method to make sense.
> In other words, you will always get "significant" coherence in this case.
> The other way would be to use the traditional thresholding which is applied
> for example in Kilner et al 2001 J Neurosci. But this obviously suffers from
> the same problems as the
> shift predictor.
> Jackknifing could also be used with a t-test against 0, but again with the
> same problem.
>
> Yours,
>
> Jan-Mathijs
>
>
> On Jul 16, 2008, at 12:38 PM, Vladimir Litvak wrote:
>
>> Dear all,
>>
>> Could you point me to the present state of the art method to determine
>> significance of coherence? I'm not talking about comparison between
>> two conditions and not about activation to baseline comparison. Just
>> you have some stationary signals, compute coherence and want to say
>> whether it's real. I suspect there is no ideal solution, but what is
>> the most accepted presently? Can jackknife implemented in
>> freqdescriptives be used for that? Any other suggestions.
>>
>> Thanks,
>>
>> Vladimir
>>
>> ----------------------------------
>> 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.
>
>

----------------------------------
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.



More information about the fieldtrip mailing list