That is basically expected. I created a slightly liberal mask so as to not exclude any regions. Having said that, you should check the talairach.xfm transform of your subjects. Sometimes the scaling can be a little off.

On 08/25/2016 04:40 PM, Christopher J Markiewicz wrote:

Hi all,

This is the subcortical mask for a subject's functional data overlaid on top of $FREESURFER/average/mni305.cor.mgz: https://imgur.com/XQPwGVT

The mask was produced with

preproc-sess -s $SESS -fsd bold -nostc -per-session -fwhm 5 -mni305

Is the spread beyond the anatomical boundaries an artifact of the smoothing done, and thus we can simply apply ROI labels in MNI305 to the functional data, or does this suggest something has gone wrong in our preprocessing?

Thanks,