Hello dear CAT12 & SPM Community,
I am relatively new to the MRI analysis and we have a rather complicated design with data taken from a drug interaction study. So maybe there is someone who could help me to understand:
We have 5 groups with different drugs (4 drugs, one placebo) and 3 time points for each group. I am about to do the VBM of the structural scans with CAT12 and will later go through the resting state scans.
The VBM the data has been preprocessed. For the group analysis I followed the instructions in the CAT12 Manual and created a flexible factorial 5x3 Design (With factors subject, group (5 levels) and time (3 levels) as well as TIV, gender, and age as covariates.
So now my questions:
1. Do you have experiences with the contrast-specification in this case?
can I test, for example, group 1 against all other groups? How do I do that? And how could I create an equivalent T-Contrast?
I tried an F-Contrast with:
[placebo]1/3(1,3) [group 2-4 each] -1/3(1,3) [covariates each] 0 [subjects placebo] 1/7(1,7) [subjects group2] 1/n2(1,n2) [subjects group3] 1/n3(1,n3) [subjects group4] 1/n4(1,n4) [subjects group5] 1/n5(1,n5)
(with n2-n5 each number of subjects within the concerning group)
The result is rather unsatisfying (there is one huge cluster over the whole glassbrain)
2. If I choose global scaling for TIV, do I delete it as a covariate?
Thank you so much in advance!
Eliska
|