[Mristudio-users] Tracking of the cingulum bundle

[Igor Yakushev]

Dear Susumu, dear Experts,

I’m trying to establish an algorithm, which would allow tracking of the
cingulum bundle (CB) with a good reproducibility.
I have started with the superior part of CB. First, I place two ROIs (2
x option OR): one at the level of the splenium of the corpus callosum
(CC) (Wakana et al., 2007 NI) and one at the border between the anterior
1/3 and posterior 2/3 of the CC body, please see ppt slides. According
to Wakana et al. the second ROI should be placed at the level of the
genu, but in this case no tracks are found between the ROIs. Relatively 
large ROIs are used in order to ensure
initial inclusion of all voxels related to CC according to a color-coded
map. Then, I place two smaller ROIs (2 x option AND) at the same
locations. Fibers, whose direction roughly deviates from the direction
of the CB, are excluded with the option NOT.
1) Any critique, improvements to this schema, please?
Finally, I get a fiber track, which is substantially smaller than the
extension of (green) voxels related to CC, please see the slides. 2) How
should/may this be interpreted? Is this related to the procedure of the
ROI placement?

Back to tracking.. I’d like to continue tracking of CB going to its
anterior or descending part. I start again with the option OR. Then one
needs AND, but AND refers to all ROIs within the image, right?
Accordingly, placement of one more ROI (option AND) e.g. at the level of
the pons results in disappearance of the previously reconstructed
portion of CB. 3) Possible solutions? Should one just save the
previously reconstructed track? When all portions are tracked, it must
be possible to save all portions as one track, isn’t it?  The aim is to
get statistics from the whole CB, not only from its portions.

Many thanks!
Igor
-------------- next part --------------
A non-text attachment was scrubbed...
Name: CB_superior_part_IY.ppt
Type: application/vnd.ms-powerpoint
Size: 1093120 bytes
Desc: not available
Url : http://lists.mristudio.org/pipermail/mristudio-users/attachments/20090309/f2a4cc6b/attachment-0001.ppt