After doing manual editing to the surface of a brain (brainmask.mgz, wm.mgz, and brain.finalsurfs.mgz) through freeview, I noticed that the volumes of some subcortical structures (specifically the hippocampus, amygdala, and accumbens) had changed, even though I did not edit aseg.mgz. Is that to be expected?

Most of the differences were minimal, a 5% change or less. However some of the structures changed by as much as 10 or 15%. I am preparing to do a second-level analysis of ROIs of the hippocampus and amygdala on a group of subjects. Given that cortical edits seem to influence subcortical structure volumes, would it be unreliable to run a second-level analysis with subcortical volumes recorded prior to cortical editing?

Thank you,

-Francesco

Francesco Siciliano, B.A.