[Mrtrix-discussion] help with streamtrack

[Thijs Dhollander]

Hi again, :-)

One extra thing to add: if you really have no clue where to put the blue seeding regions and your only real goal is to find all possible tracks between both green regions, you might define the blue seeding region as being the full brain mask.  Many wrong tracks will be generated and discarded (so it might take quite a long time), but in the end you'll only get out the ones between both green regions.  This also causes the least possible "seeding-bias", as far as I can see.

Greetings,
Thijs

________________________________
From: mrtrix-discussion-bounces at public.nitrc.org [mailto:mrtrix-discussion-bounces at public.nitrc.org] On Behalf Of Thijs Dhollander
Sent: vrijdag 28 oktober 2011 11:23
To: 'mrtrix-discussion at public.nitrc.org'
Subject: RE: [Mrtrix-discussion] help with streamtrack

Hi Gabriel,

First, it's always helpfull to reply to the full mailing list (mrtrix-discussion at public.nitrc.org ; even though that is indeed not what happens by default if you click "reply" :-)): you'll get an answer faster and other people might also find the questions/answers usefull in the future.

About your question: what you're looking for is the "-stop" option (with no extra arguments).  If you include this one in your streamtrack command, all tracks will stop as soon as they enter any of your include regions: i.e. they will not track beyond your target ROI (but also not "within", only "up to", I suppose).

But... I guess this way your tracks will still also extent beyond your seeding region (away from the direction to the include region); something you might not want if you only want to see tracks "between" the start (seed) and end (include) region.  This is because, by default, tracking starts in 2 opposite directions from the seed.  To fix this, you could use the "-unidirectional" and "-initdirection" options; but then you also need to provide a good guess for the starting direction.
Apart from that, there's also an inherent bias in your choice of which of both regions will be the seed and which region will be the "include".

To fix all of this at once, and still incorporate your goal of "only-between-region" tracking, you might for example try a setup such as the one pictured in the (very professional ;-)) paint-drawing that is attached to this mail.  Suppose you want to track between both green regions, knowing that the tracks in between certainly pass through both blue regions.  You can also see this as tracking "between" (and also "through", but not "far beyond") both blue regions.  You could then streamtrack, providing both blue regions as seeding region, and both green regions as include regions, and also adding the "-stop" option.  Note that you need to define both green regions separately from each other (and put 2 -include options in your streamtrack command: this means the tracks have to hit both of them, i.e. one is not enough).  Only tracks such as the red ones (seeded from the red dots) then meet all criteria, and there is no "asymmetrical" bias: on a large scale, tracking can go both ways between your ROI's.

I hope this is somewhat clear! :-)

Greetings,
Thijs


Thijs Dhollander
PhD Student

thijs.dhollander at uz.kuleuven.ac.be
tel. +32 16 34 90 37
gsm. +32 475 36 44 27

Medical Image Computing (K.U.Leuven, ESAT/PSI, MIC)
Medical Imaging Research Center | UZ Gasthuisberg | Herestraat 49 - bus 7003 | B-3000 Leuven | http://www.medicalimagingcenter.be

From: Gabriel Gonzalez Escamilla [mailto:ggonesc at upo.es]
Sent: vrijdag 28 oktober 2011 10:04
To: Thijs Dhollander
Subject: RE: [Mrtrix-discussion] help with streamtrack

Hello Thijs,

First, so sorry for taking so long in answer, and thank you so much for the information, it was so clear and was exactly what I was wondering about it.

Now I would like to know about the calculation of the fibers that goes from one region to another, as I read in the manual and make my own trials, I put a seed region and a target region (options seed and include) but the include option has as condition that the tracks generated need to pass through the ROI (i.e. the tracks considered can extend beyond the target ROI), but I'm interested in only the part of the tracks which goes from one region and terminated to another one, is it possible to perform such analysis?

Bests,
Gabriel.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.nitrc.org/pipermail/mrtrix-discussion/attachments/20111028/030fa8ea/attachment-0001.html