Hi AFNI experts,
I am running into an issue for applying some transforms. I have a T1 brain which I have rigidly aligned to my EPI scan (tx_1.1D) with the cmass flag in 3dAllineate. I have that same T1 brain also rigidly aligned to the MNI template for AC-PC alignment (tx_2.1D) using -cmass. I use cat_matvec to combine the transforms as follows:

cat_matvec tx_1.1D -I > tx_1_inv.1D
cat_matvec tx_1_inv.1D tx_2.1D > tx_EPItoACPC.1D


If I apply this transform using 3dAllineate to the T1 image (aligned to EPI after tx_1) it looks correct for APCP alignment with MNI template:
3dAllineate -base tpl.nii -source T1toEPI.nii -1Dmatrix_apply tx_EPItoACPC.1D

When I apply the same transform to the EPI scan it looks fairly shifted in the saggital plane by 15mm or so and also below the template by several millimeters.

1) Does the matrix coming from 3dAllineate take into account the initial cmass operation?
2) Could any of this be related to the fact that my T1 is 0.8mm iso while the EPI resolution is 2mm iso?

I also tried a few things with 3dNwarpApply

3dNwarpApply -nwarp 'IDENT(EPI.nii) tx_EPItoACPC.1D'
3dNwarpApply -nwarp 'IDENT(EPI.nii) tx_1_inv.1D IDENT(T1.brain.nii) tx_2.1D'

and a few other combinations but no success.


If I run it in two stages it works fine:

3dAllineate -1Dmatrix_apply tx_1_inv.1D -source EPI.nii -base T1.brain.nii -prefix temp1.nii
3dAllineate -1Dmatrix_apply tx_2.1D -source temp1.nii -base tpl.nii -prefix EPItotpl.nii



Any help for how to concatenate the warps properly would be appreciated.

Thanks,
Ajay