[Freesurfer] erroneuos segmentation labeling of skull outside of brain

Dear FreeSurfer community,

I am visually inspecting the quality of my recon -all output for a large dataset and have a rookie question.

In some instances, when skull stripping ended up leaving some of the skull, these voxels are associated with a "sgmtn label" (cerebral cortex). In other instances, although there are a small number of skull voxels that were not removed automatically, these voxels do not have an associated "sgmnt label" available in TkMedit Tools.

Is it the case that only the voxels for which there is a segmentation label are actually added to the total grey volume estimate?

Thus, voxels, which have not been automatically labelled would not impact volume computations?

Thank you kindly for your help.