George,

Seeing the data sure clears up the questions, thanks. :)
The problem is that your new origin is basically at zero,
but that is not where you want it. When you say slice 0
is Anterior, was it 0.43 mm anterior? That would just be
because AFNI doesn't know where the origin is from the
ANALYZE file, so it uses zero. Nothing is reversed.

Just to make sure, you should use 3dcopy to create a new dataset.
Let me suppose that you execute the following commands:

3dAFNItoANALYZE -orient LPI aa.lpi t2_2+orig
3dcopy aa.lpi.hdr t2.lpi

Now "3dinfo t2.aa" should show one-sided extents, meaning
that 0,0,0 is at the outer corner of your dataset.

What you can do now is adjust the origin to match the LPI
origin from your original dataset. Dataset t2_2+orig had LPI
extents at 103.672 L, 77.204 P, -90.027 I (inferior is negative,
left and posterior are positive, in the AFNI world). So now you
can use 3drefit to set this as your new origin:

3drefit -xorigin_raw 103.672 -yorigin_raw 77.204 -zorigin_raw -90.027 t2.lpi+orig

Hopefully that will work out for you.

- rick