Dear Philipp,
there one hidden default in SPM which unfortunately can't yet be set
via the batch system: defaults.mask.thresh. If you set this to zero,
then the estimation should run on all voxels within the explicit mask.
The reason for this threshold is to exclude non-brain voxels from
analysis based on their mean intensity (brain tissue usually has
higher intensity than non-brain regions in the images). You can either
change your spm_defaults.m, or run
global defaults
defaults.mask.thresh=0
in the MATLAB command window.
Volkmar
Quoting "Philipp G. Saemann" <[log in to unmask]>:
> Hello,
>
> when defining an explicit mask in SPM5 (fmri model, job manager) that is a
> byte image of the same resolution with binary values, SPM5 kindly ignores
> this... The estimation process runs, and a different mask (with parts
> missing that should be included) is produced.
>
> Does anybody have an explanation or idea on what the problem is?
>
> (I should add that the signal in all fMRI raw images is definitely different
> from zweo and also not NaN values)
>
> Thanks for any help here,
> Philipp
>
>
|