Hi Krista,
If 3dAutomask is clipping actual brain regions, you might
verify the contents of your datasets. One thing that jumps
out a bit is that your dataset is called "detrend+orig". Have
you removed the average and drift already? That would
cause some strange results, since 3dAutomask is looking
for that average.
What is the output of "3dinfo detrend+orig | head -20"?
How about the same command using mean+orig?
- rick