Hello FSL experts,
I have been using randomise successfully for two-group comparisons, and I would like to add a nuisance regressor to my analysis. Specifically, I would like to correct for motion during the DTI scans that I used for tractography. I have endeavored to figure this out from the FSL website and posts to this list, but I haven't been able to be certain that my attempt is correct. I generated movement measures, mean-centered them (all-subject mean, not group mean), and added them to my design.mat as a third column.
So what was:
/NumWaves 2
/NumPoints 34
/PPheights 1 1
/Matrix
1 0
1 0
etc.
Is now:
/NumWaves 3
/NumPoints 34
/PPheights 1 1 1
/Matrix
1 0 -0.02133
1 0 -0.17644
etc.
I changed the original design.con file from:
/NumWaves 2
/NumContrasts 2
/PPheights 1 1
/Matrix
1 -1
-1 1
To this:
/NumWaves 3
/NumContrasts 2
/PPheights 1 1
/Matrix
1 -1 0
-1 1 0
Though these contrasts run in randomise without errors, I'm not certain I've set them up correctly. Does this contrast design treat the third column (my motion data) as a nuisance regressor as I hope, or have I just told it to ignore this column?
Thank you for the help!
- Jeff
|