Dear FSL group members,
I have read the previous posts on how to compute this analysis, but I would like to double check if I am setting up the Glm design correctly before I run the randomise.. I am a complete novice, therefore please don't be very shocked if I made very obvious to you mistakes..:-)
What I have is two groups (patients-controls) and what I would like to see first is whether there is a correlation between age and my FA, so I would know if I would enter age later on as a covariate. I have set up the model like this in the Glm GUI:
Group EV1 EV2 EV3
Input1 1 1 0 18
Input2 1 1 0 23
Input3 1 1 0 16
Input4 2 0 1 13
Input5 2 0 1 20
and my contrasts
EV1 EV2 EV3
1 0 0
0 1 0
And I would then run the command :
randomise -i all_FA_skeletonised -o tbss2 -m mean_FA_skeleton_mask -d newtest.mat -t newtest.con -n 5000 --T2 -V
I am using the -D option, therefore I don't have to demean my EV, correct?
Do I also put -V? And why? (cannot find what -V stands for from the randomise help, but I have seen it used..)
And then I would like to see the association between my FA and my neuropsych scores, controlling for age, so I would add an EV 4 with my attention scores. What I would like to ask is how to set up my contrasts in this one.. I am a bit confused on how to specify in the model that age is my covariate and the attention scores are a variable of interest.
I would greatly appreciate your help, thank you in advance.
Maria Tziraki
|