[Freesurfer] Difference in sub-cortical volumes

Hi I noticed in my FreeSurfer runs, that there is a significant volume difference for sub-cortical structures obtained using the aseg.stat file and the mri_binaraize command. For example I got 14454mm3 for Thalamus-Proper (obtained by summing up the values for Left_Thalamus-Proper and Right-Thalamus-Proper) using aseg.stats file. For the same subject I got 15091mm3 using the command " mri_binarize --i aseg.mgz --match 10 --match 49 --o Thalamus-Proper_mask.nii.gz". Is this normal ? If so what is the most accurate method ? Thanks in advance, Indika Walimuni Research Associate University of Texas Health Science Center Houston, TX.