running proc.py without align block but using anaticor option?

AFNI Message Board
1

Dear afni experts,

Can I run afni_proc.py without the align block?
I would like to run afni_proc.py without the align block and do the alignment between the structural and functional image with ANTS afterwards as it seems to give me a better result.
However, I noticed that even with no align block, the anatomical image is used during the mask block for the WM, CSF segmentation, which is later used for anaticor.
Does it mean that my WM and CSF estimations will not be accurate and hence including the anaticor regressors will not give a good result as the anatomical image was not aligned with the functional?

I am using the following blocks: tshift despike volreg mask blur scale regress
And I am using anaticor and reml

Thank you very much in advance!
Carolin

2

Hi Carolin,

Yes, that is certainly true. If the anat/EPI alignment is not good, then it would not be appropriate to apply ANATICOR, or any other sort of tissue-based regression. If you are asking for ANATICOR and give afni_proc.py no mask to apply it with, the mask must come from the anat.

If you would rather use ANTS than align_epi_anat.py for the registration (which could likely be improved for this data, but that is a separate discussion), then it might be good to run ANTS first, and let afni_proc.py simply assumed they are registered (which is what you are doing now).

Were you to do this, maybe you could extract the MIN_OUTLIER volume from run 1 as a registration base, and then pass that back to afni_proc.py.

How does that seem?

  • rick
3

HI rick,

thank you so much for your quick response!
So you are suggesting to first use ants to register the anatomical image to the min outlier, and then running afni_proc.py with this registered anatomical image as anatomical dataset ?

Is there a short way to get the min_outlier without me having to run afni_proc.py twice? (one time to get the min_outlier and another time to run the analysis with the registered anatomical image).

Carolin

4

The MIN_OUTLIER from run 1 should be good enough, so do the same thing that is done in the proc script, e.g. in tcsh:

input: EPI_RUN1+orig
output: vr_base_myMIN_OUTLIER+orig

3dToutcount -automask -fraction -polort 3 -legendre EPI_RUN1+orig'[4..$]' > outcount.r01.1D
set min_outlier_index = `3dTstat -argmin -prefix - outcount.r01.1D\'`
set min_outlier_index = `ccalc -i $min_outlier_index+4`
3dbucket -prefix vr_base_myMIN_OUTLIER EPI_RUN1+orig"[$min_outlier_index]"
  • rick
5

Ok. Thanks a lot! :slight_smile:

Carolin

6

Oh, but remove the first few time points of input in the both EPI_r01 commands, to get past the pre-steady state.

I’ll modify the previous response to reflect this.

  • rick