I have noticed a discrepancy in the volumes generated by aparcstats2table and mri_segstats using aparc+aseg. Which is correct?

I am using version 5.3.