Dear SPMers,
My design is a 2x2 and my 1st level design matrix looks like this
A1B1 A1B2 A2B1 A2B2
I ran contrasts like [1 0 0 0] at 1st level and ended up with 4 different con*.img for each subject.
Then, at the 2nd level I used the flexible factorial model and got a main effect of condition A at a particular coordinate (FWE-corrected).
I tried to replicate this result in SPSS by extracting the con value for that coordinate from the 4 con*img files created during the 1st level.
However, the output of SPSS doesn't match the output of SPM. When I look at the main effect of A (by collapsing across B) the t-value is much lower in SPSS than in SPM.
Is it somewhat incorrect to calculate con values in SPSS in this way?
Thank you in advance,
AJ
|