large group_mask

AFNI Message Board
1

Hello experts

The group_mask+tlrc from the afni_proc.py script (below) is too large (pic attached: underlay anat_final, overlay group_mask). I would like to create a common group mask for the study for later group comparisons.

My questions: 1) Is there something I can do to get a better mask; 2) should I use the GM mask to create a common mask for group comparisons instead.
Thank you.

afni_proc.py -subj_id ${subj}
-script proc.$subj
-scr_overwrite
-out_dir $out_res/subj/func/rest.results \ -blocks despike tshift align tlrc volreg blur mask scale regress \ -copy_anat ../anatSS.{subj}.nii
-dsets ${subj}_task-rest_desc-AcUn_bold.nii.gz
-anat_has_skull no
-tcat_remove_first_trs 5
-align_opts_aea -ginormous_move -deoblique on -cost lpc+ZZ
-volreg_align_to MIN_OUTLIER
-volreg_align_e2a
-volreg_tlrc_warp -tlrc_base $tpath/btemplate \ -tlrc_NL_warp \ -tlrc_NL_warped_dsets \ ../anatQQ.{subj}.nii
…/anatQQ.{subj}.aff12.1D \ ../anatQQ.{subj}_WARP.nii
-volreg_warp_dxyz 3
-mask_apply group
-mask_segment_anat yes
-mask_segment_erode yes
-blur_size 8.0
-regress_anaticor_fast
-regress_ROI WMe CSFe
-regress_apply_mot_types demean deriv
-regress_bandpass 0.0067 0.25
-regress_est_blur_epits
-regress_est_blur_errts
-regress_run_clustsim no

group_mask.png
group_mask.png1048×374 98.6 KB

2

Hello,

I will have to alter that group mask creation, thanks. That ended up taking the union of 6 volumes…

In any case, that is not a great mask to use for group analysis, since it probably has too much coverage (even without the extra size). It might be better to take an intersection or 70% overlap of the mask_epi_anat datasets (or of the full_mask ones).

Thanks for letting us know,

  • rick
3

Thanks!

-Shankar

4

Hi Shankar,

We will build new binaries with a fix for the large mask_group result. They (presumably version AFNI_19.2.04) should be ready in by tomorrow morning.

Thanks again for bringing it up.

  • rick