Dear experts,
I am not too familiar with SPM. During my analysis I got problems
comparing contrasts of conditions between two different groups. I hope
you can help me with this problem. Here is the study-design:
3 groups (I,II,III)(total N=16) with different proficiency levels.
Each subject was tested in 4 different conditions (A,B,C,D).
Our designmatrix (spm.mat- file) includes all 16 subjects and all
conditions (from left to right: Group I, Group II, Group III).
First we are interested in contrasts between condition A-D, D-A, B-C,
C-B in each group and defined contrasts ([1 0 0 –1][-1 0 0 1][0 1 –1
0][0 –1 1 0]) within each group.
Secondly we are interested in differences (or interactions) between
group I and II in the contrasts (A-D, D-A, B-C, C-B) and between group
II and III in the contrasts (A-D, D-A, B-C, C-B).
The designmatrix:
(A-D Group I) – (A-D Group II)
Group (I) (II)
Subject (1 2 3....) (1 2 3....)
A-D [100-1 100-1 100-1 …] – [100-1 100-1 100-1 …]
= 100-1 100-1 100-1 … –1001 –1001 –1001 …
Now I can inclusive mask the contrast with (A-D Group I). Is this the
right
way?
but if I contrast
(A-D Group II)- (A-D Group I)
Group (I) (II)
Subject (1 2 3....) (1 2 3....)
A-D - [100-1 100-1 100-1 …] [100-1 100-1 100-1 …] ???
= –1001 –1001 –1001 … 100-1 100-1 100-1 … #???
I get the same contrasts as if I compare
(D-A Group I)- (D-A Group II)
Group (I) (II)
Subject (1 2 3....) (1 2 3....)
D-A [-1001 -1001 -1001 …] - [-1001 -1001 -1001 …] ???
= –1001 –1001 –1001 … 100-1 100-1 100-1 … #???
(#= same contrasts)
Or do I have to do a second-level analysis (comparison between contrasts
in groups using the *CON.IMG of each individual subject)?
Thanks in advance, Isabell.
|