Hi
While performing double subtraction (between group and between conditions)
in SPM2, we have been encountering an error as follows:
Let’s say we have a face condition and a gender condition; there are two
groups, group A and group B.
We have 4 contrasts here:
Face vs. Gender (Group A > group B)
Face vs. Gender (Group B > Group A)
Gender vs. Face (Group A > group B)
Gender vs. Face (Group B > Group A)
What we are seeing is that the activation seen in the latter two contrasts
is a flip of the activation seen in the former. For example, Gender vs. Face
(A>B) activation will be identical to Face vs. Gender (B>A).
Is there a certain way that double subtraction needs to be performed so that
the results aren't ambiguous?
Thanks,
Rishi Deshpande
Autism Brain Imaging Research Laboratory
University of Alabama at Birmingham
|