Hi FSL users,
I'd rather compare mean of two group, with age as an effective regressor and
with education as an effective confounding regressor to be pre-eliminated
before the permutation testing begins. What contrast can work for separating
the effective regressor (demeaned age) from the effective confounding
regressor (demeaned education). New Glm GUI does not seem like different old
one (FSL3.*).
GroupA GroupB Age Education
A > B 1 -1 0? 0?
Thanks,
bsjeong
|