Gyri parcellation for 3.1 graph version

Questions about BrainVisa usage and installation

Moderators: denghien, riviere

Post Reply
MSh
Posts: 40
Joined: Tue Nov 04, 2008 4:42 pm

Gyri parcellation for 3.1 graph version

Post by MSh »

Hi

I was wondering if I should use the gyri model within 3.0 folder (called gyri.gyr) for parcellation of the graphs labelled under 3.1 version or something else? (It seems to be working properly though)

Thanks
User avatar
riviere
Site Admin
Posts: 1361
Joined: Tue Jan 06, 2004 12:21 pm
Location: CEA NeuroSpin, Saint Aubin, France
Contact:

Re: Gyri parcellation for 3.1 graph version

Post by riviere »

Hi,

(sorry for the late answer, the email notifications of the forum seem to be broken)

Yes, you can use the gyri model 3.0 for gyral parcellation on 3.1 graphs, they are compatible enough.

Denis
joost
Posts: 86
Joined: Tue Feb 05, 2008 1:56 pm

Re: Gyri parcellation for 3.1 graph version

Post by joost »

Hi,

i'm using the latest BV version (4.1).

i'm doing Gyral Parcellation, choosing 'Cortical Ribbon'. the process produces a parcellation as on the attached image. my question: as can be seen on the attached image, the parcellations are not adjacent, i.e. there's an area between each of the the parcellations that is not labeled. it seems to be systematic. i would like the parcellations to be adjacent, i.e. all vertices to be labeled. can this be done? i have also attached the log file for this process.

merci,
-joost
Attachments
log.txt
(67.04 KiB) Downloaded 348 times
right_lateral.png
right_lateral.png (177.4 KiB) Viewed 7896 times
User avatar
Jean-Francois Mangin
Posts: 337
Joined: Mon Mar 01, 2004 10:24 am
Location: Neurospin, CEA, France
Contact:

Re: Gyri parcellation for 3.1 graph version

Post by Jean-Francois Mangin »

Are you sure some nodes are not labelled? It could be that the triangles linking nodes with different labels are white?
User avatar
riviere
Site Admin
Posts: 1361
Joined: Tue Jan 06, 2004 12:21 pm
Location: CEA NeuroSpin, Saint Aubin, France
Contact:

Re: Gyri parcellation for 3.1 graph version

Post by riviere »

Yes that's it: all vertices are labelled, but as the visualization shows extracted sub-meshes, the triangles between vertices belonging to different regions have been removed. But it's only a visualization artifact.
Denis
joost
Posts: 86
Joined: Tue Feb 05, 2008 1:56 pm

Re: Gyri parcellation for 3.1 graph version

Post by joost »

Hi,

thanks. can i undo the visualization artefact, i.e. get all the vertices labeled within Anatomist?

merci,
-joost
User avatar
riviere
Site Admin
Posts: 1361
Joined: Tue Jan 06, 2004 12:21 pm
Location: CEA NeuroSpin, Saint Aubin, France
Contact:

Re: Gyri parcellation for 3.1 graph version

Post by riviere »

Hi,
More or less: we have developed another way to extract sub-meshes which keeps triangles which contain two vertices with the same label (affecting the triangle to this label). There are still little holes at the junction of 3 regions, and the boundary between regions, although with no hole, will probably look like sawteeth.
However this other method has been coded after the gyral parcellation process has been written, so the process in its current state does not offer the possibility to call the new sub-mesh algorithm. We could modify it, but as it's called in compiled code, it would require a recompilation, and thus would be more difficult to patch in your current brainvisa binary release. I'll have a look if I find a minute, but I don't promise.
Denis
User avatar
riviere
Site Admin
Posts: 1361
Joined: Tue Jan 06, 2004 12:21 pm
Location: CEA NeuroSpin, Saint Aubin, France
Contact:

Re: Gyri parcellation for 3.1 graph version

Post by riviere »

Alternately, I think you also have the gyral parcellation map as a texture that you can use on the white matter mesh. The colors would probably not be the same as the graph representation (and depends on the colormap - use the "RGB interpolation mode for the texture in anatomist), but would contain no hole.
joost
Posts: 86
Joined: Tue Feb 05, 2008 1:56 pm

Re: Gyri parcellation for 3.1 graph version

Post by joost »

Hi,

thanks. i guess it's also possible outside BrainVisa (e.g. Matlab), i'll try that route first.

merci,
-joost
Post Reply