[FieldTrip] help with channel combination

jan-mathijs schoffelen jan.schoffelen at donders.ru.nl
Sat Feb 18 10:24:02 CET 2012

Hi Karl,

This is perfectly fine behavior.

ft_connectivityanalysis either outputs a 'chan_chan_freq' matrix, or a 'chancmb_freq' matrix. The former will be the output when the input freq-structure contains fourier-spectra, and the cfg lacks a channelcmb field. Ft_connectivityanalysis then computes connectivity between all channel pairs, which can conveniently be stored in a square matrix 'chan_chan' for each frequency. If the freq-structure contains cross-spectra (and power-spectra), or if you specify a cfg.channelcmb in combination with fourierspectra in the input, you will get the connectivity data to be 'chancmb_freq', i.e. each row in the matrix contains a spectrum. The reason for this is that it is a priori unlikely that the channelcmb-array defines a set of channel pairs that can be represented in a square all-to-all matrix. If it is your intention to end up with a square matrix, you may want to consider first doing a subselection of channels, i.e. computing your fourierspectra only on the subset of channels you want to use in a later step, and then call ft_connectivityanalysis without a channelcmb field in the cfg.

See for some additional information about the different representation of frequency domain data: http://fieldtrip.fcdonders.nl/faq/in_what_way_can_frequency_domain_data_be_represented_in_fieldtrip



On Feb 18, 2012, at 2:16 AM, Karl Doron wrote:

> Hello,
> I'm trying to run ft_connectivityanalysis on a reduced set of MEG channels, i.e., left and right posterior channels. Because neighboring channels are highly correlated, I'd like to exclude them and look only at "inter-hemispheric" connections.
> I'm having some troubling getting ft_channelcombination to produce this set of channels. I have 43 channels per hemisphere giving n(n-1)/2 connections. I made an NX2 matrix using repmat.m and passed it as cfg.channelcmb; however the output of ft_connectivityanalysis gave a vector for 'cohspctrm' rather than a chan_chan_freq matrix.
> Thanks for any suggestions.
> Karl Doron
> UC Santa Barbara
> _______________________________________________
> fieldtrip mailing list
> fieldtrip at donders.ru.nl
> http://mailman.science.ru.nl/mailman/listinfo/fieldtrip

Jan-Mathijs Schoffelen, MD PhD 

Donders Institute for Brain, Cognition and Behaviour, 
Centre for Cognitive Neuroimaging,
Radboud University Nijmegen, The Netherlands

Max Planck Institute for Psycholinguistics,
Nijmegen, The Netherlands

J.Schoffelen at donders.ru.nl
Telephone: +31-24-3614793

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.science.ru.nl/pipermail/fieldtrip/attachments/20120218/aef50280/attachment-0001.html>

More information about the fieldtrip mailing list