Dear Madam or Sir,
I encountered a problem when masking my image data in SPM5 in the 2nd
Level Analysis:
Consider a single voxel which contains regular values for some subjects
but zeros for the others. I realized that these zero-values are included
in the analysis, which of course change my results. My aim is to make
the calculations for this voxel (e.g., multilinear regression) only with
those subjects and values which are non-zero.
Unfortunately, setting these zero-values to NaN's and (implicitly)
masking them out does not help because then SPM does not include this
voxel at all (because this voxel contains NaN for at least one subject).
In other words, is it possible to create/include individual masks for
each subject?
I would be very happy about any comments or help on this topic :)
Kind regards,
Andreas Hahn
Functional Neuroimaging Group,
Medical University of Vienna,
Austria.
|