open-discussion
20210 Subscribers 
open-discussion > ParaToSPHARMMeshCLP amygdala values
Showing 1-3 of 3 posts
Feb 7, 2017 12:02 PM | Gozde Kizilates
ParaToSPHARMMeshCLP amygdala values
Hi,
I would like to analyze amygdala's shape via SPHARM. To analyze my data, spharmdegree and subdivlevel values are required. How can I found these values?
Thanks in advance,
Gozde
I would like to analyze amygdala's shape via SPHARM. To analyze my data, spharmdegree and subdivlevel values are required. How can I found these values?
Thanks in advance,
Gozde
Feb 7, 2017 01:02 PM | Beatriz Paniagua
RE: ParaToSPHARMMeshCLP amygdala values
Hi Gozde,
Thanks for your interest in SPHARM-PDM. For future inquiries, please use the SPHARM-PDM forum
A description on the spharmDegree and the subDivLevel can be found in the Styner 2006 paper.
However, I would like to point out that SPHARM-PDM alone has not been able to represent the shape of the amygdala properly. The amygdala is a very round shape with a lot of symmetry and that has caused the method to not find consistent correspondences across a population. Dr. Styner's group has developed a population-wide optimization to solve this sort of problems, that will be incorporated into SlicerSALT. SlicerSALT first release will not be ready until the end of the summer 2017, so I encourage you to contact Dr. Styner's about his study-wise research code.
Thank you!
Beatriz
Thanks for your interest in SPHARM-PDM. For future inquiries, please use the SPHARM-PDM forum
A description on the spharmDegree and the subDivLevel can be found in the Styner 2006 paper.
The two main parameters of this tool are the
maximal degree for the SPHARM computation and the subdivision level
for the icosahedron subdivision. In our experience, the SPHARM
maximal degree is should be chosen between 12 (hippocampus) to 15
(lateral ventricle, caudate) for brain structures. If the degree is
chosen too high the reconstructed SPHARM surface will often show
signs of voxelization. By using the '-paraOut' command line option,
the spherical icosahedron subdivision, as well as local phi and
theta attribute files for the quality control visualization with
KWMeshVisu are written out.
However, I would like to point out that SPHARM-PDM alone has not been able to represent the shape of the amygdala properly. The amygdala is a very round shape with a lot of symmetry and that has caused the method to not find consistent correspondences across a population. Dr. Styner's group has developed a population-wide optimization to solve this sort of problems, that will be incorporated into SlicerSALT. SlicerSALT first release will not be ready until the end of the summer 2017, so I encourage you to contact Dr. Styner's about his study-wise research code.
Thank you!
Beatriz
Feb 7, 2017 01:02 PM | Juan Felipe Orejuela
RE: ParaToSPHARMMeshCLP amygdala values
Regarding the answer of Mrs. Paniagua, maybe you can try with FIRST
(https://fsl.fmrib.ox.ac.uk/fsl/fslwiki/F...), you can do a group analysis as it is detailed there (they use
the hippocampus, but you can do it with any structure), and you
will get the vtk files of stats and p-values.
Regards,
Juan.
Regards,
Juan.