How Lausance 2008 ROIs are generated?

101 views
Skip to first unread message

Qingyang Li

unread,
Apr 26, 2012, 2:33:37 PM4/26/12
to cmtk-...@googlegroups.com
Dear CMTI-users,

 Hagmann et al.'s 2008 paper documented that the ROIs are generated by *first* generating 66 neuroanatomical regions, and then further subdividing them into super voxels that are 1.5cm^2 to get 998 ROIs.
While in CMP if we choose the Lausanne 2008 scheme for parcellation, we ended up with 1016 ROIs. Do you know how these are generated? are they still based on the same 66 regions? 

Thanks!

Yang

Alessandra Griffa

unread,
Apr 26, 2012, 2:41:37 PM4/26/12
to cmtk-...@googlegroups.com
Hi Yang,

you can refer to this paper for more details:

L. Cammoun et al.
Mapping the human connectome at multiple scales with diffusion spectrum MRI
Computational Neuroscience Volume 203, Issue 2, 30 January 2012, Pages 386–397


Alessandra

Qingyang Li

unread,
Apr 26, 2012, 2:53:52 PM4/26/12
to cmtk-...@googlegroups.com
Thank you very much for fast response.
so ,1016 ROIs are 998 cortical ROIs + 18 sub-cortical ROIs.

Yang
------
Qingyang Li

Balance and focus.

Qingyang Li

unread,
Apr 26, 2012, 6:29:12 PM4/26/12
to cmtk-...@googlegroups.com
Hi Alessandra,

From the Cammoun paper I read that there are 5 scales of ROIs generated. If there an easy way to generate graphs based on all of them in CMP? Thanks

Yang 
------
Qingyang Li

Balance and focus.

Alessandra Griffa

unread,
May 1, 2012, 5:39:58 PM5/1/12
to cmtk-...@googlegroups.com
Hi Yang,

yes, you have all the information you need in the 5 gpickle files saved in project/subject/timepoint/CMP/fibers/matrices.
The files contain the connectivity information in graph format already. Use the NetworkX (http://networkx.lanl.gov/) Python library to read the gpickle files and have access to the graphs.

Hope this can help
Alessandra

Qingyang Li

unread,
May 18, 2012, 3:10:10 PM5/18/12
to cmtk-users
Thank you very much for the answer!

Yang

On May 1, 5:39 pm, Alessandra Griffa <alessandrag...@gmail.com> wrote:
> Hi Yang,
>
> yes, you have all the information you need in the 5 gpickle files saved in *
> project*/*subject*/*timepoint*/CMP/fibers/matrices.
Reply all
Reply to author
Forward
0 new messages