Hello.
I’m just after confirmation that this is correct please.
I have three patient groups, scanned twice. I’ve run VBM and created a difference image (scan 2 minus scan 1), as per the suggested route on the mailing list. I want to run randomise to see if there are any changes in GMD that correlate with a behavioural measure. I’m interested to know whether changes in GMD and the correlation with this behavioural measure differ between groups.
If I had just 6 subjects (2 in each group), would I simply enter one column per group, and enter the demeaned (within group) behavioural score? e.g.,
model:
2 0 0
1 0 0
0 3 0
0 2 0
0 0 1
0 0 2
contrasts:
groupA: 1 0 0
groupB: 0 1 0
groupC: 0 0 1
A>B: 1 -1 0
..
I'm also going to add in age and sex as nuisance variables but have left this out for simplicity here.
Thanks.
Sam.
|