ok, i found a solution for my problem below, please don't bother.
i needed to adjust the variable 'defaults.mask.thresh' in the
spm_defaults.m file to something feasible for our data. interestingly
one does not seem to be able to do this on the fly, i.e., kind of
applying a volatile parameter that stays just for a single batch run.
bye, pisti
On Mon, 18 Feb 2008, Istvan Akos Morocz wrote:
> dear SPM gurus,
>
> why do mask.img files after a 1st level fixed effect analysis of
> several spatially normalized fMRI studies sets turn out to be
> NON-identical, even after inclusion of an explicit mask file in the
> variable :
>
> jobs{1}.stats{1}.fmri_spec.mask{1} = 'myMask.img' ;
>
> moreover, during the 2nd level RFX analysis, i experience a similar
> problem: specifying a mask file in variable complex :
>
> jobs{1}.stats{1}.factorial_design.masking.tm.tm_none = [] ;
> jobs{1}.stats{1}.factorial_design.masking.im = 0 ;
> jobs{1}.stats{1}.factorial_design.masking.em{1} = 'myMask.img' ;
>
> doesn't have much an effect on the outcome of the resulting 2nd level
> mask.img file. while computing with hundreds of 1st level contrasts,
> somewhere down the line one or more strong outliers spoil and ruin the
> entire analysis.
>
> i need to add on here that undesired voxels are, as epxected, excluded
> with the supplied mask at the 1st level, but additional voxels within
> the brain disappear as well although the explicit mask specifically
> marks them with a value of '1'.
>
> what am i missing here? thanks for any input, bye,
>
> pisti
>
-----
www.pistikem.org
[log in to unmask]
|