Dear FS developers and expert users,

I am preprocessing with freesurfer 6.0 recon-all a set of pre defaced subjects and the results are generally good, but the Talairach transform QA measure are terribly bad and exactly the same on all defaced:

#-----------------------------------------------------

TalAviQA: 0.00000 z-score: -222 WARNING: Talairach QA check failed! z-score of -222 is <= threshold of -9 #----------------------------------------------------

I have tried a few local subject of the set with face and the QA is acceptable (z-score=-2) and is the same if I pre-deface the subject with mri_deface in FS6.0 with the correct parameters. I don't know how the original datasets were defaced, but I suppose with mri_deface in an unknown older version. I need to process the defaced datasets, but I am warring about the recon-all output measure, in fact I found a thickness error of 28% between deface and face temporal lobe of the same subject. How do you suggest to proceed? Thanks in advance, cheers, Giuseppe