Hi Helen,

So you tried the -template option with 3dfractionize, and
the -master option with 3dresample, I guess. Is that
correct?

Assuming it is, the reason for the shift is the same old
problem with SPM data, in that their format of choice does
not contain the orientation information (unless they store
it in a matlab-formatted .mat file).

So if you know the orientation and origins of your dataset,
then you can copy the ANALYZE dataset to AFNI format
(with 3dcopy), and use 3drefit on the AFNI dataset to set
the orientation and origins properly.

If there is a matlab file there, you could read it in with
matlab and get the transformation matrix out. Then you can
use 3dWarp to apply that transformation to the data.

---

In any case, now assuming the dataset is in AFNI format,
you should now be able to apply either of the commands,
'3dfractionize -template' or '3dresample -master' without
losing part of the brain.

Note that succeeding with the ANALYZE to AFNI steps is
not necessarily trivial.

- rick