Print

Print
Hmm.  I just tried batching this on a per session basis, using the All-Session contrast to define the ROI.  I got the same xY.XYZmm values, but the eigenvariates are slightly different.

Here is the code from the job file (I edited out the filenames and commented):

matlabbatch{1}.spm.util.voi.spmmat = {'Directory/GLM_for_DCM/SPM.mat'};
matlabbatch{1}.spm.util.voi.adjust = 15; % Number of F-contrast for run 1
matlabbatch{1}.spm.util.voi.session = 1; % run 1?
matlabbatch{1}.spm.util.voi.name = 'Calcarine_batch';
matlabbatch{1}.spm.util.voi.roi{1}.spm.spmmat = {''};
matlabbatch{1}.spm.util.voi.roi{1}.spm.contrast = 14; %Number of T-Contrast of Interest for All Sessions (not run 1)
matlabbatch{1}.spm.util.voi.roi{1}.spm.conjunction = 1;
matlabbatch{1}.spm.util.voi.roi{1}.spm.threshdesc = 'none';
matlabbatch{1}.spm.util.voi.roi{1}.spm.thresh = 0.05;
matlabbatch{1}.spm.util.voi.roi{1}.spm.extent = 0;
matlabbatch{1}.spm.util.voi.roi{1}.spm.mask = struct('contrast', {}, 'thresh', {}, 'mtype', {});
matlabbatch{1}.spm.util.voi.roi{2}.mask.image = {'Directory/Intracalcarine.img,1'};
matlabbatch{1}.spm.util.voi.roi{2}.mask.threshold = 0.5;
matlabbatch{1}.spm.util.voi.roi{3}.sphere.centre = [0 0 0];
matlabbatch{1}.spm.util.voi.roi{3}.sphere.radius = 8;
matlabbatch{1}.spm.util.voi.roi{3}.sphere.move.global.spm = 1;
matlabbatch{1}.spm.util.voi.roi{3}.sphere.move.global.mask = 'i2';
matlabbatch{1}.spm.util.voi.expression = 'i1&i2&i3';



On Sun, Jun 14, 2015 at 8:50 PM, Rita Elena Loiotile <[log in to unmask]> wrote:
Hi,
Can someone please confirm whether my understanding of the GUI interface is correct? I would like to extract VOIs for multiple sessions.

I load Results for my contrast of interest - All Sessions (made using replicate & scale in the model setup) and threshold.
Press Eigenvariate
Adjust Data for Conditions of Interest- All Sessions
Which Session: Inf

It's my understanding that this is the preferred way to extract VOIs for DCM because it guarantees that the same voxels are used, across all runs, for a given subject (i.e. if you load the sessions separately, some voxels won't pass threshold for a subset of runs and will, therefore, be excluded from the Eigenvariate for only those runs).  

My question is: The way that I have setup the VOI extraction, does SPM understand that I mean to adjust run 1 for run 1's condition of interest, adjust run 2 for run 2's condition of interest, etc.?  (Or does SPM try to adjust run 1 for All Sessions-Condition of Interest... whatever that would mean?)

Also, I understand from the SPM archives, that adjusting for conditions of interest (i.e. an F-contrast of the regressors of interest) demeans the eigenvariate by all OTHER regressors.  Is this correct?

Thanks!