Hi Sophie,
the result image and the additional information (esp. the astronomical
number of resels)
makes one think that something has gone wrong with the preprocessing.
Sometimes, smoothness estimates can be completely off, of there are
outliers, i. e.
some corrupted image. So, best check the following points:
- all images smoothed in the same manner?
- all smw...img files looking ok if compared to eachother (or in small
groups of e. g. 4) in the check reg mode?
- your mask.img which you find in your result folder seems to extend the
normal brain edges... so SPM analyzes voxels
far ot that nobody is interested in. Best you need to define an absolute
threshold (round 0.1) when you set up the model.
So, it looks like a technical/preprocessing problem.
Comment on NaN: I also don't know the specific reason, BUT in a proper =
working analysis these values should appear (the statistical problem aside).
Let us know what the bug was here...
best regards,
Philipp
Max-Planck-Institute of Psychiatry
NMR Research Group
Kraepelinsr. 2-10
80804 Munich
Mail: [log in to unmask]
Phone: 0049-89-30622-413
|