I was searching the FS list archives and came across this now very old question. For the record, the DKT40 classifier in the FreeSurfer distribution is the same as the one described in the paper. The DK and DKT atlases are very similar. The goal for constructing the DKT atlas was to create a cortical classifier that would reliably label the major cortical gyri. We modified some of the DK region definitions to rely on boundaries well suited to FreeSurfer’s classifier algorithm, such as sulcal fundi that are approximated by surface depth and curvature. The boundary definitions were also made less ambiguous, making it easier to apply them in the training dataset more consistently.

Jason Tourville

On Thu, Jun 15, 2017 at 5:15 AM Gabor Perlaki <petzinger.gabor@gmail.com> wrote:

Dear all,

I've found an article comparing DK and DKT atlas based automatic segmentations to manual labels (Klein and Tourville, Front Neurosci. 2012; 6:171, https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3514540/). It seems that using DKT40 works better than DK in most of the cortical regions (based on Dice).
Does anybody know that the DKTatlas related results created by Freesurfer (e.g. aparc.DKTatlas+aseg.mgz, lh.aparc.DKTatlas.stats) are based on the same DKT40 atlas used in the article and whether the DKT40 atlas related results were also created by Freesurfer in the article by Klein and Tourville? Are the Freesurfer created results in lh.aparc.DKTatlas.stats generally more reliable than the results in lh.aparc.stats?

--
Gabor Perlaki
research associate
Diagnostic Center of Pécs
H-7623 Pécs, Rét str. 2.
Tel.: 0036-30-2084367
E-mail: petzinger.gabor@gmail.com

_______________________________________________
Freesurfer mailing list
Freesurfer@nmr.mgh.harvard.edu
https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer

The information in this e-mail is intended only for the person to whom it is
addressed. If you believe this e-mail was sent to you in error and the e-mail
contains patient information, please contact the Partners Compliance HelpLine at
http://www.partners.org/complianceline . If the e-mail was sent to you in error
but does not contain patient information, please contact the sender and properly
dispose of the e-mail.