[Mrtrix-discussion] Exclude slices from calculation

[Thijs Dhollander]

Hi Dorian,

As far as I know, excluding specific slices of specific volumes is
currently not possible for CSD analysis.  It is available for tensor
fitting (and subsequent MD or FA maps, etc..) though:
http://www.brain.org.au/software/mrtrix/commands/dwi2tensor.html .  If you
got whole volumes that are unusable due to serious artifacts, you can of
course throw them out beforehand.  If it's single slices that are affected
for all volumes, these could be excluded from the mask (but I guess this
scenario doesn't make a lot of sense in practice).

I'm not fully sure on all the options offered by ExploreDTI, but as far as
I know, the (or one of the) probabilistic CSD tracking algorithm(s) in
ExploreDTI is based on bootstrapping the data, yet following the
("deterministic") maxima of the subsequently obtained FODs for each
bootstrap realisation.  I.e., the probabilistic distribution sampled is a
distribution of the FOD maxima due to bootstrapping.  More details can be
found in Ben's paper:
http://onlinelibrary.wiley.com/doi/10.1002/hbm.21032/abstract .  The
current probabilistic algorithm in MRtrix, on the other hand, directly
samples the FOD itself.  More information on this one can be found in the
following paper:
http://onlinelibrary.wiley.com/doi/10.1002/ima.22005/abstract .
Do note, however, that there's a lot more details/settings/choices/... that
can affect the appearance of the final tractogram.  An overview of some of
these things is given in the MRtrix paper, but there's also others.  I'm
not sure, but I think ExploreDTI seeds tracks in the center of voxels for
some or all of its tractography algorithms, for example; while MRtrix seeds
randomly (uniformly) at any coordinate in any seeding regions.  This simple
difference can by itself already yield a different "qualitative" feeling of
the outcome.  Even a different rendering/visualisation can give a different
impression...!  So, in the end, it all depends on how dissimilar your
results look: is it just a qualitative thing, or are the distributions of
tracks really different, or do your tracks even segment different
structures all together?  Depending on what you want to achieve with the
final tractograms, some of these differences might not be all that
important, while others might be crucial.

Hope this helps!

Cheers,
Thijs

*Thijs Dhollander
thijs.dhollander at gmail.com
Tel. +32 475 36 44 27*
Medical Image Computing (MIC), ESAT-PSI, Department of Electrical
Engineering, KU Leuven


On Tue, Sep 24, 2013 at 12:00 AM, Dorian P. <alb.net at gmail.com> wrote:

> Dear all,
>
> This is my first post, congratulations on the software. A piece of cake to
> use on Win7, quick to use and no crashes.
>
> My question has to do with bad slices. Is there a way to exclude specific
> slices from analysis? My scanner is misbehaving at times, and I have slices
> completely collapsed (I guess the subject moved too).
>
> I noticed some people use ExploreDTI to preprocess, can it be used to
> "delete" bad slices too?
>
>
> One more question since we are in the topic. Does ExploreDTI have the same
> CSD/probabilistic algorithm of MRtrix? It seem to have the CSD option but
>  results don't look similar to MRtrix.
>
> Thank you.
> Dorian
> TJU
>
> _______________________________________________
> 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/20130924/e2c539e3/attachment.html