I am doing cortical thickness and surface area analyses and had a few questions regarding editing the borders.

1) Is it reasonable to edit every (big and small) error that I find? Basically, how much a difference would it make in terms of the surface area and cortical thickness values if I did not edit the smaller errors? 

2) If the skull hasn't been completely stripped, but is not designated as either white or gray matter, do I need to adjust the thresholding for extracting the brain and re-extract? The only situation in which I feel this would be a problem is when doing whole brain analyses.

3) Would it make a difference to the data if I initially ran recon-all using freesurfer version 5.0 and then, after editing the segmentations/labels/skullstripping, reran autorecon 2,or 3 using freesurfer version 5.1? 

If anyone could answer any or all of these questions that would be great!