Hello,
I am comparing my results for 3dAutomask using an older version of AFNI (2.56b of 3/12/04) and a recent upgrade to version AFNI_2006_12_22_0933. I am having trouble getting identical results (though they usually don't differ by a whole lot, maybe 1 voxel or so, especially if I use the "-nograd" option on the new version). I can't figure out what is going on here. The only other thing that has changed, aside from the version, is that I am now working on a server that uses little-endian data rather than big-endian data (I don't know the details of this, it's just what the techno-experts tell me). Any ideas on what might be causing the results to differ?
Thanks,
Karin