On 13 Jan 2011, at 10:04, Chris Chapman wrote:
Hello FSLers,
I am wondering about using TBSS to compare diffusion in the inferior part
of the cingulum (retro-/infrasplenial and parahippocampal).
In "Smith et al. Tract based spatial statistics: Voxelwise analysis of
multi-subject diffusion data. Neuroimage. 2006;31:1487-1505", at the 3rd
paragraph starting on page 1494, the authors describe using a pre-defined
volume in which to search for the inferior cingulum skeleton, and searching
for the local max FA within a circular space instead of perpendicular
direction.
How can this be done using FSL/TBSS? I see that there is a "LowerCingulum"
mask in the "standard" folder, but I am not sure how to apply it.
The TBSS scripts/programs will do this for you.
Also, what if the images are being registered to the "best fit" brain of
the group, not the FMRIB58_FA standard brain? Will a new "LowerCingulum"
mask need to be drawn?
You can check this by loading in the standard space version of your template image (which TBSS will generate) to FSLView and also loading in the FSL cingulum mask - you can then compare the two. The mask is fairly liberal so it should probably be fine.
Can this method be applied to other tubular white-matter structures in
which there is large inter-subject variability? Ex. fimbria/fornix?
The reason we only generated this approach for the lower cingulum was that it is fairly isolated from other tracts and hence the method is more necessary and robust there than elsewhere. It is possible that the same could also be true of the fornix - however this would require unusually good quality / resolution data to make sense.
Cheers.
Thanks for any help,
Chris Chapman
University of Michigan Medical School
