[Mrtrix-discussion] tracking the fornix

[Patai, Zita]

Dear Luis,

That is great, thank you so much for your ROI placement strategy, and the references.
I will try your suggestions:)

Thanks again!
z

From: lconcha at gmail.com [mailto:lconcha at gmail.com] On Behalf Of Luis Concha
Sent: 13 May 2013 18:45
To: Patai, Zita; mrtrix-discussion at www.nitrc.org
Subject: Re: [Mrtrix-discussion] tracking the fornix

Hello Zita.

Well, this question is close to my heart...
Your general description for tracking the fornix sounds reasonable and most of the times it should give you accurate tracking. However, I find that placing the seed at the level of the crus or body of the fornix gives much better results than placing it at the level of the mammillary bodies (MB). If you do it this way, bidirectional tracking is suggested, and a couple of -include regions can be placed at the level of the MB and the hippocampus. I normally put an -exclude region between the MB and the hippocampus on a para-sagittal slice, to avoid getting anterior commissure tracks contaminating my fornix. I repeat the procedure for the opposite side.
While I did this approach using deterministic tracking a while ago, I find that these ROIs work just as well on probabilistic tractography. More details can be found here:

http://www.ncbi.nlm.nih.gov/pubmed/16219832
http://www.ncbi.nlm.nih.gov/pubmed/15562425

Finally, as the fornix is embedded in CSF, tracking errors occur more often due to partial volume effect than in any other bundle if some sort of CSF elimination is not used. It is still possible to track the fornix in the presence of CSF signal, of course, but it tends to be a  lot easier from CSF-nulled data.
Best of luck with the curviest of the tracks!

Luis



Dr. Luis Concha
Instituto de Neurobiología
Laboratorio C-13
UNAM, Campus Juriquilla
Boulervard Juriquilla 3001
Juriquilla, Querétaro.
C.P. 76230
México
Tel (442) 2 38 10 54
Fax (442) 2 38 10 46
www.inb.unam.mx<http://www.inb.unam.mx>

On Mon, May 13, 2013 at 12:30 PM, Patai, Zita <e.patai at ucl.ac.uk<mailto:e.patai at ucl.ac.uk>> wrote:
Dear MRTrixers

I am in the process of tracking the fornix in a large group of subjects. I mention large group because I am trying to find a way to track the fornix, without having to place to many post-hoc exclusion ROIs.
Currently I am using deterministic CSD (as recommended by Metzler-Baddeley et al, 2011/12), and I am placing a seed voxel in the mammillary bodies , and an inclusion see in the anterior hippocampus (there are additional parameters, please see example below).
What I find is that this works great for some people, but others do not look great, and some don't show up with anything. However, this latter group does indeed have a fornix (thankfully) when I use probabilistic tracking, but with many other fibres of non-interest.

The obvious question arises: am I doing this wrong, or do fornices really differ enough to warrant such a difference in appearance with this particular method?
And if I using a suboptimal procedure, does that mean that it is ok to do many post-hoc exclusion masks? I was hoping for a straightforward method that required few user input parameters as this may carry a large bias especially when I will be tracking my patient group...

Thank you!
zita


my code:
streamtrack SD_STREAM CSD4.mif -seed 1.81,3.76,-12.39,10 -initdirection 0,0,1 -include -20.88,-4.32,-13.39,10 -unidirectional -curvature 0.5 -mask mask.mif fornix_seedmb10_trghipp10_superiordir_unidir_curv0.5_csd_det.tck
_________________________
Eva Zita Patai
Postdoctoral Research Associate
UCL, Institute of Child Health
e.patai at ucl.ac.uk<mailto:e.patai at ucl.ac.uk>
020 7905 2730


_______________________________________________
Mrtrix-discussion mailing list
Mrtrix-discussion at www.nitrc.org<mailto:Mrtrix-discussion at www.nitrc.org>
http://www.nitrc.org/mailman/listinfo/mrtrix-discussion

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.nitrc.org/pipermail/mrtrix-discussion/attachments/20130513/794ef68b/attachment-0001.html