Dear Surferteam,
I have been trying to understand details on how the cortical thickness is calculated. I read in the article:
"we compute the closest vertex on the pial from each vertex on the white, and vice-versa, then set the thickness to the average of the two."
I have tried to carry out this calculation myself, using the generated surfaces lh.white and lh.pial and compare my results to the thickness generated by FS (lh.thickness). There is a one-to-one correspondence between the vertices of the white and pial meshes, which makes it quite straightforward:
e.g.: Vertex 100 on white surface has distance from pial surface = 2.5 mm, vertex 100 on pial surface has distance from white surface = 2.6 mm, average == thickness at vertex 100 = 2.55 mm.
However, the results that I am getting are different from lh.thickness. I expected discrepancies on the order of rounding errors, but half of the thickness values differ more than 0.001, with some differing as much as 2 mms! What am I missing?
Could you please elaborate a bit on how you define the cortical thickness?
Thank you very much, Miro Drahos
Hi Miro,
the closest vertex is not necessarily the "same" one on the pial/white surface. We actually have a newer and hopefully better version that we are working on, but it's not quite ready for distribution.
cheers Bruce On Fri, 30 Apr 2010, miro drahos wrote:
Dear Surferteam,
I have been trying to understand details on how the cortical thickness is calculated. I read in the article:
"we compute the closest vertex on the pial from each vertex on the white, and vice-versa, then set the thickness to the average of the two."
I have tried to carry out this calculation myself, using the generated surfaces lh.white and lh.pial and compare my results to the thickness generated by FS (lh.thickness). There is a one-to-one correspondence between the vertices of the white and pial meshes, which makes it quite straightforward:
e.g.: Vertex 100 on white surface has distance from pial surface = 2.5 mm, vertex 100 on pial surface has distance from white surface = 2.6 mm, average == thickness at vertex 100 = 2.55 mm.
However, the results that I am getting are different from lh.thickness. I expected discrepancies on the order of rounding errors, but half of the thickness values differ more than 0.001, with some differing as much as 2 mms! What am I missing?
Could you please elaborate a bit on how you define the cortical thickness?
Thank you very much, Miro Drahos
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
freesurfer@nmr.mgh.harvard.edu
-
Bruce Fischl -
miro drahos