Hello,
I am using SPM5 to do a "2nd-level" two-group t-test on a bunch of
aligned scans and I am concerned if I'm choosing the ANCOVA normalization
options consistently or not. Here are the settings I used:
======== ANCOVA - related selections in the GUI,
under "Factorial design specification" ====
Design:
Two-sample t-test
ANCOVA - yes
Covariates:
age
Centering - As implied by ANCOVA
sex
Centering - As implied by ANCOVA
Global normalisation:
Normalisation -> ANCOVA
=======================================
Did I miss anything or should I change some of these choices?
Also, if I was after significant regions where group A values EXCEED
group B's, and the design matrix had the 1's for group A on the first
column and those of group B on the second (the other covariates and
the ANCOVA columns being unimportant here), the correct contrast
for the 1-sided t-test done by SPM would be (1 -1 0 0 ...), right?
(Note: I am not looking for different values here, but precisely
for higher values in group A than in group B)
Thank you very much,
Silviu Podariu
UNMC, Omaha
|