erode fs_ap_wm.nii.gz from suma_make_spec_FS?

AFNI Message Board
1

Hello Afni experts,

I have a quick question here:
Do you recommend to erode fs_ap_wm.nii.gz and fs_ap_latvent.nii.gz before using it as masks for tissue based regression? Freesurfer does not do a perfect job with outlining the white matter and I want to make sure I have no GM voxels inside these masks.
(I constructed those masks usingsuma_make_spec_FS and would like to run aCompCor to derive Principal components from wm and csf)

Thank you very much in advance!
Carolin

2

Hi Carolyn,

Yes, some erosion is important, to avoid regressing out gray matter signals. Note that particularly due to EPI distortion, imperfections in registration can lead to regressing out something inappropriate. So sufficient erosion is important.

Consider eroding one or 2 levels, depending on the voxel size. You can also pass the mask to afni_proc.py and tell it to erode (-anat_follower_erode) the mask. However, you might feel more comfortable doing the erosion yourself, to be sure the resulting mask seems reasonable, and then not having afni_proc.py do erosion at all.

  • rick