3D ICC, a lot of outside brain sig voxels

AFNI version info (afni -ver): 22.0.13

Hello!

I have been using 3DICC to calculate some whole brain ICCs on some task based fMRI data and I am seeing a lot of voxels outside the brain (see attached image).
The olay is the ICCs themselves and the threshold is from the ICC F stat.
The voxel clusters look very spread out and grainy and thta whole outside brain map is contiguous.

Any help appreciated.
Maria

code text  # or delete if not needed

Hi Maria,

For most of the voxels outside the brain, the model likely failed to converge, resulting in unrealistic estimates at the parameter boundary. It would be better to simply mask out the area outside of the brain.

Gang Chen