significant coherence

jan-mathijs schoffelen j.schoffelen at PSY.GLA.AC.UK
Mon May 18 22:13:05 CEST 2009


Dear Nina,

There is no fieldtrip function in the release allowing for the  
testing of significant coherence, if I understand correctly what you  
want. This would be to infer whether a given estimate of coherence is  
significantly different from zero. This means that you are not  
testing two conditions against each other (which is what  
stafun_indepsamplesZcoh is doing; on top of this: it does it for all  
combinations of channels, and not just for all channels versus a  
given external reference signal).
There are two ways in which you can assess the 'significance'. First  
would be something like a 'shift predictor'. In short, one could  
shuffle the order of the replications for the reference channel and  
recompute coherence between the reference and the rest. Provided the  
reference is not the same signal in each trial, doing this many times  
would give you an estimate of the 'bias' in coherence. Any observed  
coherence values in the upper 5% tail then would count as significant  
(this by the way only works for an external reference channel, and  
not when you use one of the MEG channels as a reference).
This shift predictor is perhaps something of an overkill, and you  
could use a parametric test instead. One way to do this, is to 'Z- 
transform' the coherence. I am used to using a particular formula  
which is described in one of our papers (Schoffelen, Science 2005  
(supplementary material)); the appropriate references can also be  
found therein.

I hope this helps,

Cheers,

Jan-Mathijs

On May 16, 2009, at 4:05 PM, Nina Kahlbrock wrote:

> Dear Fieldtrip users,
>
> I have a question concerning significant coherence. I saw that a  
> similar question was asked before on the mailing list. As I am new  
> to coherence, I did not quite understand the answer.
> So here is my question: I have computed coherence between different  
> MEG sensors and a photodiode. Now, I would like to compute  
> significance levels for these coherence values at different  
> frequencies. Is there a function in fieldtrip that I can use to  
> solve this task? I have found the function  
> ‘statfun_indepsamplesZcoh’. However, I am not sure whether this is  
> the right one to use.
>
> I would greatly appreciate your help!
>
> Thanks in advance.
> Nina
> ----------------------------------
>
> 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.
>
> http://listserv.surfnet.nl/archives/fieldtrip.html
>
> 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/neuroimaging/fieldtrip.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.science.ru.nl/pipermail/fieldtrip/attachments/20090518/3b19b97b/attachment-0002.html>


More information about the fieldtrip mailing list