Hi,

I would like to do a Group Analysis for my FA data.

For all my subjects (controls + patients) I have all the fa-masked.ANAT+CVS-to-avg35.mgz (the resampled registered image to CVS space).

I am not sure if my steps for the Group Analysis are the right one:

1) Concatenate FA from individuals into one file

 mri_concat */fa-masked.ANAT+CVS-to-avg35.mgz --o GroupAnalysis-fa-masked.ANAT+CVS-to-avg35.mgz

2) Make sure the order agrees with the fsgd:

 ls -l */fa-masked.ANAT+CVS-to-avg35.mgz

3) Create average of the input images for visualization:
    mri_average */fa-masked.ANAT+CVS-to-avg35.mgz GroupAnalysis.Average-fa-masked.ANAT+CVS-to-avg35.mgz

4) mri_glmfit --y GroupAnalysis-fa-masked.ANAT+CVS-to-avg35.mgz \
      --fsgd Group_Analysis.fsgd dods --C contrast.mtx --glmdir Group_Analysis.glmdir

To view uncorrected results:

tkmedit fsaverage orig.mgz -aparc+aseg -overlay /media/1Tra/Antonella_2011/DTI_FS/dtrecon/Group_Analysis.glmdir/contrast/sig.mgh -fminmax 2 3

where -fminmax 2 3 : Voxel-wise threshold of 2 (p<.01), color saturation at 3 (p<.001)


To correct for multiple comparisons run:


5) mri_glmfit-sim \
  --glmdir Group_Analysis.glmdir \
  --grf 3 neg \
  --cwpvalthresh .0166 \
  --overwrite


Is 3 (p<.001) neg as well as .0166 (.05/3) the best voxel-wise as well as cluster wise values to be tried?


To view corrected results:
tkmedit fsaverage orig.mgz \
  -ov /media/1Tra/Antonella_2011/DTI_FS/dtrecon/group_analysis.glmdir/contrast/grf.th3.pos.sig.cluster.mgh \
  -seg /media/1Tra/Antonella_2011/DTI_FS/dtrecon/group_analysis.glmdir/contrast/grf.th3.pos.sig.ocn.anat.mgh  \
       /media/1Tra/Antonella_2011/DTI_FS/dtrecon/group_analysis.glmdir/contrast/grf.th3.pos.sig.ocn.lut \
  -fminmax 1.3 5

with only positive values

Are this steps right?

When I run the registration over the cvs space which file is the equivalent file of the fa-tal.nii.reg from the TAL space?


MANY THANKS!

Antonella