Dear FS's experts,
I have tried the recon-all process on one subject with both FS v5.3 and v6_beta.
Although subcortical structures seems to be better segmented in v6_beta, I find that v6_beta over-segmented some other structures as cerebellum, hippocampus, ...
I will attached in the FileDrop my T1.mgz, and aparc.a2009s+aseg.mgz files for both freesurfer versions.