yes, I agree that a new atlas might need a different scaling. And I also agree that the talairach_with_skull.lta should be adjusted relative to a volume with a skull on it, otherwise the eTIV wouldn't be right if there's atrophy. On Fri, 27 Oct 2006, Joao Pereira wrote:
Dear all,
I am posting the questions I made to Bruce without going through the list along with his reply. **I would still like to be sure about question 1)***, as it is central to my conclusions. If anyone has any other ideas, please let me know.
There are still a couple of questions in my mind I would really like to understand concerning this TIV estimation though:
- I first run FS without any visual inspection (-autorecon1 and 2) to get
the TIV estimate in the aseg.stats file. I then compared those values to
the
ones obtained *after* adjusting the affine transf (in some cases using the brain.mgz file, others manually) and the values changed. But is this the correct procedure to adjust the TIV output, or am I looking at an
artifact?
- Is it possible for me to build my own template for TIV estimation
purposes? If yes, which file should be replaced? Thank you again for your help (btw, is there any chance I can obtain the eTIV value without running -autorecon1 and 2 all the way?)
Reply:
(1) We have very little experience with the eTIV, but I guess the answer is yes, that would be the way to do it, and (2) the only way to build your own atlas would be to use the aseg.mgz from a bunch of subjects and mris_ca_train.
Although 2) is possible in a practical way (ie, I can build my own atlas), I don't believe it would be valid for TIV estimation with FS, due to the bootstrap method you used to calculate the "scale factor". Am I right?
Finally, could someone please explain me the point of trying to adjust the talairach output by using the T1.mgz and brain.mgz files? I know this works, but I don't understand why, especially with the brain.mgz: the reference atlas has skull... shouldn't using a skull stripped scan confound the process?
Thank you all. Hope to hear from you soon.
Joao
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer