Hi Rick,
I am not sure if this also explains the issue I had regarding 3dAllineate/afni_proc.py/align_epi_anat. I would register the T1 to the EPI (far apart) and so during the initial alignment of centers the temporary file looks good. The centers matched the EPI and the T1 was centered in the window. However, no matter what option I tried (including supplying align centers with -pre_matrix) when the final output was displayed the T1 always appeared in the bottom left corner of the grid. I know that the original RS would be in that location so I understand that it didnt have to zpad in order to align the two objects, but I could never get the T1 window to be centered on that position.
1) Is the bug you are updating the binaries related to this at all?
2) Separate question is when I use 3dAllineate to apply my transformation matrix from T1 to EPI alignment to the freesurfer aseg file using nearest neighborbor interpolation, I am still getting voxels that are not integer numbers. Is there a better way to apply the trasnform which will preserve the values or would I need to use 3d calc and a floor/ceiling sort of fix?
Thanks,
Ajay