Got it, thanks!

There are some odd things here, which lead me to ask how
this dataset was created. Was is made at the scanner, or
by third party software? I don't have a coherent reply,
so I'll just itemize some observations for now...

1. When I run to3d, my result looks the same as your Old version.

to3d -time:zt 27 2 2000 alt-z -epan -prefix new 6*.dcm

2. If I apply the internal oblique tranfomation on the dataset
that was just created, it almost aligns with your New version,
but not quite.

3dWarp -deoblique -prefix new.obl new+orig

3. The brains all appear upside-down. So something is incorrect,
both in the basic xyz coordinates, and in the transformation.

I am confused why your New version behaves differently from mine,
since you appear to be running a current version of AFNI. Did you
compile from the archive yourself?

Where were the DICOM files created?

Daniel, do you have any thoughts about this?

- rick