Hi
I have 2 groups, A and B, each group has 20 subjects.
And I have one psychological score for every subject.
Can I find out the brain regions with significantly different correlation to the psychological score between A and B?
For example, in area X, the "correlation between score and brain" is higher in group A than group B.....
I think the GLM design might be:
EV1=group A, EV2=group B, EV3= score
( for all subjects in group A, EV1=1, EV2=0.... and for group B, EV1=0, EV2=1)
And the contrast might be:
contrast 1( EV1=1 , EV2= -1 , EV3=1 )
contrast 2( EV1= -1, EV2=1, EV3=1 )
contrast 3( EV1=1, EV2=-1, EV3= -1 )
contrast 4( EV1= -1, EV2=1, EV3= -1 )
When I run randomise, I will use the "-D" function, so that I don't need to demean any values....
Is that okay?
Thanks a lot....
|