[Mrtrix-discussion] RE: mask ROI

[Robert Smith]

Zita

Thijs is entirely correct in his description of the difference between the
two cases.

The reason you get a 'better' tract with the second approach is because all
streamlines that leave the mask are terminated, and most likely the vast
majority of those streamlines will be retained. With the first option,
whether or not a streamline is kept or discarded will depend on whether it
terminates just before the exclusion mask, or just reaches the exclusion
mask. This is placing a large emphasis on the appropriate termination of
streamlines, which is notoriously difficult using only the diffusion
information.

For instance, consider a situation where you have used the brain parenchyma
as the tracking mask, and the area outside the brain as an exclusion mask.
In one grey matter voxel with a supra-threshold FOD, the streamlines may
cross the voxel, enter the exclusion region, and be rejected; while in the
voxel adjacent to it with a sub-threshold FOD, the streamlines are
terminated in that voxel, and therefore retained (as they do not reach the
exclusion mask). This will result in a poorly reconstructed tract, as many
of its components will be missing.

The ill-posed problem of streamline termination is precisely why I
implemented the Anatomically-Constrained
Tractography<http://www.sciencedirect.com/science/article/pii/S1053811912005824>
framework.
If you have image data for performing EPI susceptibility distortion
correction, you'll be able to use this in the next major MRtrix release.

Remember: an exclusion mask results in the complete rejection of any
streamlines that enter it. If you want to 'exclude' streamlines from
traversing a particular area, i.e. terminate any streamlines that attempt
to enter that area (but still keep those streamlines), omit that area from
your propagation mask.

Rob

--

*Robert Smith*
Post-Doctoral Researcher, Imaging Division

The Florey Institute of Neuroscience and Mental Health
Melbourne Brain Centre - Austin Campus
245 Burgundy Street
Heidelberg Vic 3084
Ph: +61 3 9035 7128
Fax: +61 3 9035 7301
www.florey.edu.au


On Thu, May 23, 2013 at 6:15 AM, Patai, Zita <e.patai at ucl.ac.uk> wrote:

>  Dear Thijs****
>
> ** **
>
> Thank you for the quick response!****
>
> ** **
>
> I understand what you have said, and it makes sense. Do you recommend
> either option as being more ‘anatomically’ correct? ****
>
> What is interesting is that with option b, I got a better tract, even
> though as you say more streamlines may have stayed in there…****
>
> ** **
>
> Thanks again!****
>
> z****
>
> ** **
>
> *From:* mrtrix-discussion-bounces at www.nitrc.org [mailto:
> mrtrix-discussion-bounces at www.nitrc.org] *On Behalf Of *Thijs Dhollander
> *Sent:* 22 May 2013 20:57
> *To:* mrtrix-discussion at www.nitrc.org
> *Subject:* [Mrtrix-discussion] RE: mask ROI****
>
> ** **
>
> Hi Zita,****
>
> ** **
>
> There is in fact a difference between both scenarios: tracks which enter
> an exclusion ROI are entirely discarded, while tracks which leave a mask
> are just stopped (but not discarded if they meet the other requirements,
> e.g. the minimum track length, at that point).****
>
> Qualitatively, in the final distribution of accepted tracks, you are thus
> likely to see (a lot) less tracks ending “on” (i.e. very near) the boundary
> of the exclusion ROIs in scenario (a), as opposed to scenario (b).  Also,
> scenario (a) will take longer to obtain an equal amount of accepted tracks,
> as opposed to scenario (b).****
>
> ** **
>
> Cheers,****
>
> Thijs****
>
> ** **
>
> *Thijs Dhollander*
> PhD Student
>
> thijs.dhollander at uzleuven.be
> tel. +32 16 34 90 37
> gsm. +32 475 36 44 27*
>
> **Medical Image Computing (KU Leuven, ESAT/PSI, MIC)
> *Medical Imaging Research Center | UZ Gasthuisberg | Herestraat 49 - bus
> 7003 | B-3000 Leuven | http://www.medicalimagingcenter.be****
>
> ** **
>
> *From:* mrtrix-discussion-bounces at public.nitrc.org [
> mailto:mrtrix-discussion-bounces at public.nitrc.org<mrtrix-discussion-bounces at public.nitrc.org>]
> *On Behalf Of *Patai, Zita
> *Sent:* Wednesday, May 22, 2013 21:39
> *To:* mrtrix-discussion at www.nitrc.org
> *Subject:* [Mrtrix-discussion] mask ROI****
>
> ** **
>
> Dear MrTrixers****
>
> ** **
>
> Just checking that there is no difference in the ‘quality’ or ‘number of
> streamlines’ generated if I do****
>
> **a)      **a mask of whole brain but include many exclusion ROIs****
>
> or****
>
> **b)      **a mask ROI which is smaller than the whole brain and is
> already excluding those regions which I did as –exclude above****
>
> ?****
>
> ** **
>
> Many thanks!****
>
> zita****
>
> _________________________****
>
> Eva Zita Patai****
>
> Postdoctoral Research Associate****
>
> UCL, Institute of Child Health****
>
> e.patai at ucl.ac.uk****
>
> 020 7905 2730****
>
> ** **
>
> _______________________________________________
> Mrtrix-discussion mailing list
> 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/20130524/af531554/attachment.html