Dear Kevin,
>I never heard back on the post below. Briefly: when I specify an explicit
>mask for a SPM99 PET analysis, does the software restrict its calculation of
>the global mean to voxels within the explicit mask, or does it calculate the
>global mean using all suprathreshold voxels whether in or out of the mask?
The global estimator is evaluated by spm_global. It does not refer to
any mask other than the one it creates for itself. If you want the
global estimator to be contrained to in-mask voxels you would have
to apply the mask to the data before design specification (remember
the global estimator enters the design matrix).
I hope this helps - Karl
|