Hi Anderson,
Thank you for your response to my question. I am still trying to navigate this, so thank you for your time and patience.
I looked at the page suggested by you. Instead then, I would do :
(hypothetical) model:
1 0 0 2 0 0
1 0 0 2 0 0
0 1 0 0 3 0
0 1 0 0 5 0
0 0 1 0 0 1
0 0 1 0 0 1
where my covariate values are not mean centred.
As I mentioned, I am running randomise on a difference image (scan 2 subtracted from scan 1). If I wanted to look at areas that show gray matter change that is correlated with the covariate, I'd simply enter a contrast like [ 0 0 0 1 0 0 ], for the first group, or [ 0 0 0 1 1 1 ] to look at changes that correlate across all groups. For regions that differ in correlation by group, I'd have this [ 0 0 0 1 -1 0 ].
Does this seem right?
Big thanks again.
Sam