Combining Transformations with afni_proc.py

AFNI Message Board
1

Dear AFNI experts,

I am relatively new to AFNI and afni_proc.py, and am trying to use proc.py to combine transformations. More specifically, I am trying to take our slice-timing corrected dataset and simultaneously transform the subject’s functional data into standard space, using the same parameters from the previous transformation while also maintaining the resolution of the functional data, and perform the second volume registration.

This is my afni_proc.py script so far:

afni_proc.py -subj_id 20180328
-script 20180328.tcsh
-out_dir Results
-dsets /home/user/Test/Datasets/20161028_0406_f1_f2_v1_am_r1_r2_st.nii
-blocks align tlrc volreg
-copy_anat /home/user/Test/Datasets/anat.nii
-tcat_remove_first_trs 0
-anat_has_skull yes
-align_opts_aea -Allineate_opts -warp shift_rotate
-align_opts_aea -giant_move
-tlrc_opts_at -apar /home/user/Test/Datasets/anat_tl.nii
-tlrc_opts_at -dx 2.5 -dy 2.5 -dz 2.4
-volreg_align_to first
-volreg_align_e2a
-volreg_tlrc_warp
-bash

After executing the generated script, I receive this error:

cat_matvec anat_ns+tlrc::WARP_DATA -I
** ERROR: THD_read_dvecmat: can’t find attribute WARP_DATA in dataset anat_ns+tlrc
** FATAL ERROR: Can’t read matrix from ‘anat_ns+tlrc::WARP_DATA’

Also, when looking at the generated script, it looks as though -giant_move isn’t executed at all. Any guidance would be greatly appreciated!

2

I suspect having multiple -align_opts_aea is why you are not seeing the -giant_move option executed. I believe that you could use a combination of --tlrc_anat, -volreg_tlrc_adwarp, and -volreg_warp_dxyz to perform what you want in afni_proc.py.

If you’re willing to use the nonlinear transform, you can use auto_warp.py once and then apply that transform to all of your afni_proc.py calls using the -tlrc_NL_warped_dsets option.

That said, perhaps you could clarify how you are warping the anatomic to standard space (are you using @auto_tlrc?) and what the eventual goal is. Why not include all of your runs in the same afni_proc.py call?