I have just changed the way that DEV version recon-all creates the aseg.stats file. Previously, it would report everything that it found in FreeSurferColorLUT.txt. This included a lot of stuff that would never be in the subcortical segmentation, and this made the aseg.stats file very messy.
The new version only reports on segmentations that are part of the subcortical segmentation process (ie, in mri_ca_label). Note that there may still be some segmentations reported in aseg.stats that are not found in a particular aseg.mgz because that segmentation was not found during labeling (eg, hypointensities).
The values in the table (ie, volumes and intensities) should be exactly the same.
This change only applies to dev (not to stable) and should be live on Tue.
doug