[FieldTrip] granger causality on beamformer data

[Tyler Grummett]

I then do the following:


% look up which virtual channels correspond to particular areas
            % of the brain
            cfg = [];
            cfg.atlas = afni;
            cfg.inputcoord = 'mni';
            cfg.maskparameter = 'inside';
            labels = ft_volumelookup( cfg, source);

            % how many sources found in grey matter
            [tmp ind] = sort(labels.count,1,'descend');
            sel = find(tmp);
            for j = 1:length(sel)
                found_areas{j,1} = [num2str(labels.count(ind(j))) ': ' labels.name{ind(j)}];
            end​

However I dont know how to find out what sources are the 'found_areas', after that I dont know how to cluster the source in a particular area.

Tyler



*************************

Tyler Grummett ( BBSc, BSc(Hons I))
PhD Candidate
Brain Signals Laboratory
Flinders University
Rm 5A301
Ext 66124
________________________________
From: fieldtrip-bounces at science.ru.nl <fieldtrip-bounces at science.ru.nl> on behalf of Tyler Grummett <tyler.grummett at flinders.edu.au>
Sent: Wednesday, 28 May 2014 3:53 PM
To: FieldTrip discussion list
Subject: Re: [FieldTrip] granger causality on beamformer data


Hello Julian,


I think I got a few steps further into what I am trying to do, except I dont know whether I have done the correct thing or not, could you check it?


            % interpolate sources
            mri = ft_read_mri('Subject01.mri');
            mri = ft_volumereslice([], mri);

            % read in atlas from fieldtrip template
            afni = ft_read_atlas( fullfile( matlabrootpath, 'Matlab/fieldtrip/template/atlas/afni/TTatlas+tlrc.HEAD'));

            % construct grid that lies only in grey matter
            cfg = [];
            cfg.mri = mri;
            cfg.grid.warpmni = 'yes';
            cfg.grid = afni;
            grid = ft_prepare_sourcemodel( cfg);

            % Source Analysis: without contrasting condition
            cfg = [];
            cfg.channel = 'EEG';
            cfg.method = 'lcmv';
            cfg.grid = grid;
            cfg.vol = vol;
            cfg.keepfilter = 'yes';
            source = ft_sourceanalysis( cfg, timelock);

​

Tyler


*************************

Tyler Grummett ( BBSc, BSc(Hons I))
PhD Candidate
Brain Signals Laboratory
Flinders University
Rm 5A301
Ext 66124
________________________________
From: fieldtrip-bounces at science.ru.nl <fieldtrip-bounces at science.ru.nl> on behalf of Tyler Grummett <tyler.grummett at flinders.edu.au>
Sent: Wednesday, 28 May 2014 10:36 AM
To: FieldTrip discussion list
Subject: Re: [FieldTrip] granger causality on beamformer data


​Hey Julian,


Having trouble making sense of that link. Am I correct in saying that I should be downloading the brede toolbox? because it is taking a long time, plus the functions in the brede toolbox dont make a lot of sense.


Would you use ft_prepare_sourcemodel instead of ft_prepare_leadfield? or would you run in before running it?


Regards,


Tyler


*************************

Tyler Grummett ( BBSc, BSc(Hons I))
PhD Candidate
Brain Signals Laboratory
Flinders University
Rm 5A301
Ext 66124
________________________________
From: fieldtrip-bounces at science.ru.nl <fieldtrip-bounces at science.ru.nl> on behalf of Julian Keil <julian.keil at gmail.com>
Sent: Wednesday, 28 May 2014 1:55 AM
To: FieldTrip discussion list
Subject: Re: [FieldTrip] granger causality on beamformer data

Hi Tyler,

I can't comment on the usefulness of directionality between 1400 sources, but keep in mind, that you would have to compute something in the range of 1400*1400
connections, so I hope you have a fast computer.

As for the regions, in case you want to use anatomically defined regions, you an either use the atlases (atlanti? atlae?) that come with fieldtrip or generate a mask from this website: http://neuro.imm.dtu.dk/services/jerne/ninf/voi.html

The general idea is to build a grid with a gridpoint per voxel of your MRI using ft_prepare_sourcemodel. Then you can check which of your virtual channels is closest to the voxel-gridpoints and thus select the virtual channels that are inside your ROI.

In the first case, you can use ft_volumelookup to find the voxels corresponding to your ROI. In the latter case you can just use the mask and check which voxels are 1 (= inside your ROI).

I hope that helps, if you have specific questions, feel free to ask.

Best,

Julian


On Tue, May 27, 2014 at 7:44 AM, Tyler Grummett <tyler.grummett at flinders.edu.au<mailto:tyler.grummett at flinders.edu.au>> wrote:

​Hello fieldtrippers,


I was just wondering whether it would be sensible to do granger causality on all 1400 virtual channels, as calculated using beamformer.


Or should you do a PCA reduction of some description beforehand.


I was also wondering how to create regions of interest. Some of my colleagues think that we should use some kind of spatial ICA technique.


Im open to all suggestions.


Tyler


*************************

Tyler Grummett ( BBSc, BSc(Hons I))
PhD Candidate
Brain Signals Laboratory
Flinders University
Rm 5A301
Ext 66124

_______________________________________________
fieldtrip mailing list
fieldtrip at donders.ru.nl<mailto:fieldtrip at donders.ru.nl>
http://mailman.science.ru.nl/mailman/listinfo/fieldtrip

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.science.ru.nl/pipermail/fieldtrip/attachments/20140528/a5836f30/attachment.html>