Hi AFNI folks,

I've got two anatomical datasets I'm aligning using the align_epi_anat python script. All is
fine with the alignment. However, I've also got a set of ROIs that need to be aligned to the
same target anatomical. This ROI dataset is an anatomical mask with integer values shoved
into each voxel that's part of an ROI (5 for V1, 10 for V2, for example). So, I thought I'd be
able to simply mark this ROI dataset as a -child_dset in the align_epi_anat command and
all would be fine. However, after alignment, my nice homogeneous integer values in the
mask which mark each of my ROIs have been altered significantly. It looks as though the
edges of the ROIs have been smoothed/averaged with the surround such that the integer
values have changed at the edge, but are preserved towards the center of each ROI. Any
idea how I can avoid this? I'd like the output to maintain the integer values from before the
alignment.

Here's the command I ran:

align_epi_anat.py -dset1 AX_anat_IC1_TAL.nii -dset2 AXT_b0_image.nii -child_dset1 AXT_TALbrain_roisoverlay.nii -dset1to2
-Allineate_opts '-allcost' -suffix _DSI

Thanks for your help,
Adam Greenberg