When you run isxconcat-sess add -paired-diff. This will compute 1-2 for each subject. You can then run mri_glmfit. With that FSGD file, your contrast should be 1 not 1 0
doug


On 2/6/16 11:17 AM, Thục Trinh wrote:
Hello,

I'm working on group analysis in fMRI activation using FSFAST. I tend to test whether any significant changed in fMRI activation within group across two time points ( pre-training vs post). I ran first level already.  I still confuse about the concat data and the mri_glmfit command. Here are the command lines, I'm wondering if it runs exactly what I want it to do.

--  idlist

Input NCP140005_1 
Input NCP140005_2 
Input NCP140012_1 
Input NCP140012_2 
Input NCP140071_1 
Input NCP140071_2 
Input NCP140079_1 
Input NCP140079_2 
Input NCP140082_1 
Input NCP140082_2 
Input NCP140543_1 
Input NCP140543_2 
Input NCP140556_1 
Input NCP140556_2 
Input NCP140558_1 
Input NCP140558_2 

-- concat data
isxconcat-sess -sf idlist -analysis work.sm5.lh -contrast con-v-arrow -o group2

-- FSGD file

GroupDescriptorFile 1
Title Con-arrow Gr2
Class Defi
Input NCP140005 Defi
Input NCP140012 Defi
Input NCP140071 Defi
Input NCP140079 Defi
Input NCP140082 Defi
Input NCP140543 Defi
Input NCP140556 Defi
Input NCP140558 Defi

-- run glm analyisis

mri_glmfit --glmdir lh.paired.gr2-diff --y ces.nii  --fsgd con-arrow_group2.fsgd --C mean.mat

-- content in matrix contrast
 C = [ 1 0 ]

Thank you,
Trinh


_______________________________________________
Freesurfer mailing list
Freesurfer@nmr.mgh.harvard.edu
https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer