AFNI_proc.py: Resting-State Analysis: .errts. defective result

AFNI Message Board
1

Dear all,

I am currently stuck on the problem that my resting-state analysis with afni_proc.py results in a flawed outcome. This problem only occurs for my resting-state runs (for all subjects) and not for the task-based runs. The results of the latter are fine.

The problem is that when I open my errts file (e.g. “errts.Subject1_Rest.anaticor+tlrc”) there is activation all around the brain (once I increase the threshold slightly) but not in the brain. Please notice the screenshot that I added, it makes the problem all clear.

Please let me know what you think. Where do you think could the problem be located? In what direction should I look?

By the way: A friend of mine reminded me that I have to use the errts for a resting-state analysis as underlay in the AFNI Gui. I did that. Even if I choose AnatQQ or AnatSS as underlay (and then errts as overlay) the problem remains the same for all kinds of combinations. So I think the problem is not that I simply don’t use the Gui accordingly, but the problem has to be somewhere else.

This is my resting-state analysis script:

Resting-State Preprocessing

for subject in Subject1 Subject2
do

Set directories

directory_subjects=/users/philipp/desktop/fmri/dataset/subjects/$subject
directory_subjects_general=/users/philipp/desktop/fmri/dataset/subjects
directory_fMRIruns=$directory_subjects/raw
directory_anatomical=$directory_subjects/raw
directory_SSwarper=$directory_subjects/SSwarper

Run SSwarper

mkdir $directory_subjects/SSwarper
@SSwarper
-input $directory_anatomical/3dto3d+orig
-base MNI152_2009_template_SSW.nii.gz
-subid $subject
-odir $directory_subjects/SSwarper

cd $directory_subjects

AFNI_proc.py

afni_proc.py
-subj_id ${subject}_Rest
-out_dir $directory_subjects/Preprocessing_RestingState
-dsets
$directory_fMRIruns/restto3d+orig
-blocks despike tshift align tlrc volreg mask blur scale regress
-copy_anat $directory_SSwarper/anatSS.$subject.nii
-anat_has_skull no
-tcat_remove_first_trs 4
-align_opts_aea -cost nmi
-volreg_align_e2a
-volreg_align_to MIN_OUTLIER
-volreg_tlrc_warp -tlrc_base MNI152_2009_template_SSW.nii.gz
-tlrc_NL_warp
-tlrc_NL_warped_dsets
$directory_subjects/SSwarper/anatQQ.$subject.nii
$directory_subjects/SSwarper/anatQQ.$subject.aff12.1D
directory_subjects/SSwarper/anatQQ.{subject}_WARP.nii
-volreg_post_vr_allin yes
-volreg_pvra_base_index MIN_OUTLIER
-mask_segment_anat yes
-mask_segment_erode yes
-regress_anaticor
-regress_ROI WMe
-regress_apply_mot_types demean deriv
-regress_motion_per_run
-regress_censor_motion 0.2
-regress_censor_outliers 0.05
-regress_bandpass 0.02 0.2
-blur_size 6.0
-regress_est_blur_epits
-regress_est_blur_errts
-html_review_style pythonic
-execute

done

111.png
111.png1628×1375 148 KB

1 Like
2

Hello,

This is expected, and is because of the ‘scale’ processing block (which notably does not affect correlations, nor does the censoring).

The voxels inside the brain have noise that may be somewhat similar to what is outside the brain. However, they also have far higher means. So when all voxels are scaled, the fluctuations outside the brain are heavily amplified when compared with fluctuations inside the brain.

  • rick
1 Like
3

Dear Rick,

does this mean that I do not have to use scaling for a standard resting-state analysis?

Thanks,

Philipp

4

Hi Philipp,

Yes. Scaling is often left in because it may be beneficial for other subsequent steps, but it has no effect on a single run correlation (it does have a very small effect when using multiple runs).

You can test it out and see, just to feel more confident (since you already have scaled results, and might be thinking of making unscaled ones, too).

  • rick
1 Like
5

Dear Rick,

thanks again for your help. I really appreciate that you help me here so often. And I am giving everything to learn AFNI.

Have a good weekend,

Philipp