Hi FSLers,
I ran a 3rd level feat analysis today and hit a weird snag -- the
output statistic images had a different dimensionality than the
input. In brief, I had normalized all the data to 1mm MNI space (182
x 218 x 182) and checked all the first and second level results,
which were fine. When I ran a paired t-test at the 3rd level, the
bg_image came out with the right dimensions but all the files in the
cope1.feat/stats directory had only 160 slices (instead of 182).
I can't figure out what's going up here. The
cope1.feat/filtered_func_data file has 182 slices and I manually
checked all the individual input files and they do too. But every
one of the files in the stats/ dir has only 160 slices.
Am I missing something obvious? I'm wondering if something weird
happened and will re-run the analysis to see if this replicates (I hope not...)
Thanks.
- Joe
|