coherence and phase spectrum

[Jasper Poort]

Hi all,

I have a few questions: 

first, in order to look at induced power rather than evoked power, one can subtract the power spectrum of the evoked potential (average over trials) from the power spectrum (average of power in single trials). Is this not the same as first subtracting the average over trials from every trial and then doing a single freqanalysis? If not, would it then mean you have to do a procedure like: do 1) freqanalysis on preprocessed data 2) timelockanalysis on preprocessed data, then freqanalysis 3) take (result of 1)-(result of 2)?

second, does a similar approach also exist for coherence? For example, in case of presenting a flash to a subject, different sensors will get common input that is time and phase locked, resulting in high coherence values, however for this no coupling between the sensors needs to exist. Is it possible to correct for this to get an "induced" coherence?

third, in principle the phase spectrum can give information additional to the power spectrum: in the case of a power spectrum, it can give the latency of the response if one fits a linear function to the phase spectrum. How can this be done in FieldTrip? My first approach would be:

cfg = [];
cfg.method = 'mtmfft';
cfg.output = 'fourier'; % complex spectrum
cfg.tapsmofrq = 4;
[freq]=freqanalysis(cfg,data) % data is preprocessed data

the power spectrum is then:
powspctrm=squeeze(mean(freq.fourierspctrm.*conj(freq.fourierspctrm)));

the phase spectrum is then simply unwrapping the angle of the complex spectrum and trying to fit a line on this?:
phsspctrm = squeeze(mean(unwrap(angle(freq.fourierspctrm))));

Besides practical advice I'd much appreciate references to any standard papers on this procedure.

related to this, how could one use the phase of the cross-spectrum to get an idea of the latency of covariance at some frequency?

Many thanks, best, Jasper

----------------------------------
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.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.science.ru.nl/pipermail/fieldtrip/attachments/20070904/2fca55c6/attachment-0001.html>