Hi Adam,
what does it look like on the original surface? Is it a region of high gaussian curvature (i.e. curvature in 2 dimensions at the same time)? It may be that the distortion is unavoidable if you want to preserve metric properties to the extent that you can. Does this cause a problem for any of the FreeSurfer stuff, or is only the interaction with Suma? Ziad: is the problem that you are undersampling that area becasue of the dense grid? Or is there actually overlap in the triangles? That is, is it negative definite there?
Bruce
On Fri, 11 Nov 2005, Thomas, Adam (NIH/NIMH) wrote:
Florent,
I'm using MapIcosahedron from the SUMA package to do the spherical averaging and the distortions are visually quite striking in an area that is fairly important in this study. Here's a screen shot of the smoothwm surface after recreating it with an icosahedron mesh:
http://fmrif.nimh.nih.gov/~adamt/ridges.jpg
Note the ridges along the gyrus. There is also a spike coming out of the sulcus nearby that is a little hard to see.
Ziad (the developer for SUMA) also thought some sort of regional smoothing might be the best solution. If that's the consensus, we'll pursue that option.
-Adam
Adam Thomas adamt@nih.gov Functional MRI Facility, NIMH/NIH/DHHS 10 Center Dr, Room B1D708A Bethesda MD. 20892-1148 Phone:301-402-6351 Fax: 301-402-1370
-----Original Message----- From: florent segonne [mailto:fsegonne@csail.mit.edu] Sent: Thursday, November 10, 2005 4:51 PM To: Thomas, Adam (NIH/NIMH) Cc: freesurfer@nmr.mgh.harvard.edu; fischl@nmr.mgh.harvard.edu Subject: Re: [Freesurfer] problems with spheres
Dear Thomas,
Looking at the file rh.sphere.asc, it seems that the problem comes from the spherical registration that generates a spherical surface with, as you said, small triangles slightly overlapping.
However, these distortions should be extremely small. If you really need
to get rid of them, one hacky solution would be to detect these (small) regions and to smooth them untill the overlapping edges(faces) disappear. Since these regions are small, this should not generate large metric distortions.
But maybe Bruce has some other ideas.
Florent
Florent Segonne PhD Candidate Stata Center 32-D430 CSAIL MIT 1 617 253 2986 http://people.csail.mit.edu/~fsegonne
On Thu, 10 Nov 2005, Thomas, Adam (NIH/NIMH) wrote:
I recently noticed that some of my subject's spherical surfaces
appears
to have several small, pyramid-like points on them. One example is
shown
here in tksurfer:
http://fmrif.nimh.nih.gov/~adamt/sphere.jpg
Looking at these points in more detail in suma shows that several triangles are folder on top of each other so that all the vertices are on the spheres surface, but some triangles are entirely overlapped by others.
The spherical surface is here:
http://fmrif.nimh.nih.gov/~adamt/rh.sphere.asc (11MB)
And this is a close up of near node 141619 in suma:
http://fmrif.nimh.nih.gov/~adamt/sphere_suma_zoom.jpg
These distortions are causing us problems with subject averaging. Is there something we can do to fix or avoid them?
Thanks, -Adam
Adam Thomas adamt@nih.gov Functional MRI Facility, NIMH/NIH/DHHS 10 Center Dr, Room B1D708A Bethesda MD. 20892-1148 Phone:301-402-6351 Fax: 301-402-1370
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer