[Freesurfer] freesurfer segmentation: perirhinal Roi

Dear Freesurfer experts,

We would like to use volume and cortical thickness measures obtained for the perirhinal cortex in Freesurfer 5. When checking the region that has been segmented (using lh/rh.perirhinal.label), we noticed that the ROI does not seem to cover the entire cortical thickness of the region. On the coronal slice shown in example (attached file), one can see that there are still a few voxels of gray matter between the external border of the ROI and the external boundary of the cortex. How reliable is this segmentation? Is there a reason why the ROI appears so thin?

Many thanks in advance,

Best regards,