Dear Group,
I have a question concerning the correspondence between con*.img files and
SPM*.img files:
I was quite surprised to find that con*img and SPM*img files do not
correspond voxel by voxel
in terms of what was "masked out" in each. In other words, there were
non-zero/non NaN voxels
in an SPM*img that were NaN's in the corresponding con*img files. How could
this be if SPMt*img
files are simply con*img files divided by an appropriate error image?
Thanks,
Karen Andersen, MD
Columbia University
|