[Mrtrix-discussion] CSD

[Robert Smith]

Hi Ralf,


1. A command you definitely want to have a look at in the new MRtrix
release is fod2metric.
It performs FOD segmentation as outlined in Appendix 2 of this paper
<http://www.sciencedirect.com/science/article/pii/S1053811912011615>, and
can output various statistics on the results (see the help file of the
command for full details).
It's also a neat way to familiarise yourself with concepts relating to how
to deal with multiple fibres per voxel in a context other than
tractography; for instance, the terminology
<https://github.com/jdtournier/mrtrix3/wiki/Dixels-and-Fixels> you may use
in such a context, how to store such data
<https://github.com/jdtournier/mrtrix3/wiki/MRtrix-Sparse-image-formats-(.msf-.msh)>,
and how to  perform
<http://www.sciencedirect.com/science/article/pii/S1053811911012092>
statistical <http://www.ismrm.org/13/session83.htm>  comparisons
<http://www.ismrm.org/14/program_files/Session57.htm>...


2. Regarding the image overlay, it's important to understand that the
viewer does not perform an explicit image registration step between images
when you load one of them using the overlay tool. If you want the images to
overlap, then most likely some explicit image registration step will first
be required. If however we assume that such a registration step has been
performed, the second important concept to understand is that two images
can be aligned with one another in *real space*, even if they do not
possess the same image transform or voxel sizes.

So let's imagine that you've loaded your MPRAGE in mrview; you can think of
the 'camera' in the viewer being a 2D plane that is 'snapped' to whatever
alignment the voxel grid of that MPrage happens to have. You then load your
DWI using the overlay tool. Now the DWI is drawn over the top of the
MPRAGE, based on the position of the 'camera', and the image transform of
the DWI. The fact that the DWI doesn't have the same image transform and
voxel size as the MPRAGE (i.e. the image data have not bee *re-gridded*)
doesn't matter; it's drawn over the top of the image you're looking at
using interpolation. This argument still holds even if the 'camera' is not
locked to the underlying MPRAGE (i.e. you've turned off the 'lock to image
axes' option), in which case the 'camera' defines a plane through space,
and *both images* are drawn upon this plane using interpolation.

Now imagine that you have *not* performed an image registration step first,
and the subject moved significantly between the two acquisitions. Now when
you overlay the DWI, not only do the image transforms and voxel sizes
differ, but the actual *position of the brain in real space* also differs,
i.e. the image information from the two images doesn't overlap, no matter
how you look at them.

The distinction between image registration, image re-gridding /
re-sampling, and image overlay, is very important to ensure that such
questions are interpreted correctly.

Best regards
Rob


--

*Robert Smith, Ph.D*
Research Officer, 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 Tue, Jun 17, 2014 at 6:27 PM, Luetzkendorf, Ralf <
Ralf.Luetzkendorf at med.ovgu.de> wrote:

> Hi Donald,
>
> thanks for the quick answer.
>
> 1. What I meant with 1. was, if I have one ODF in my image, can I get some
> values for that ODF, like:
> - how many directions it has and how strong
> - angle
> - vector
> or something. So that I have something for a statistics or to compare to
> subjects in that voxel or the same subject with different measurment
> parameters or different MR Scanners.
>
> Im not sure if "find_SH_peaks" does that. I will test it.
>
> 2. If I open MPRage with MRview and overlay DWI and they have not the same
> resolution. Does it register the DWI to the MPRage or the MPrage to the DWI?
>
> Thanks for the help and sorry for the inexplicit questions,
>
> Ralf
>
>
> +++++++++++++++++++++++++++++++++++++++++++++++
> Ralf Lützkendorf
> Department for Biometry and Medical Informatics
> University of Magdeburg, Germany
> phone: +49 391 67 13546
> fax: +49 391 67 13536
>
> ________________________________
> Von: Donald Tournier [jdtournier at gmail.com]
> Gesendet: Montag, 16. Juni 2014 19:29
> An: Lützkendorf, Ralf
> Cc: mrtrix mailinglist
> Betreff: Re: [Mrtrix-discussion] CSD
>
>
> Hi Ralf,
>
> I'm assuming you're talking about the stable version of MRtrix (the 0.2.x
> branch)? If so:
>
> > 1. Is there a way to have a selected ODF in values, angles or something?
> Would be interesting to see or compare!
> >
>
> I'm not sure what you mean by that exactly. You can use find_SH_peaks to
> get the vectors corresponding to the fODF peaks (scaled by their
> amplitude), maybe this is what you're after?
>
> > 2. What does MRViewer do if you overlay DWI over anatomie? Is it
> interpolating the DWI or the Anatomie or something in between?
>
> Both images are interpolated, unless you  explicitly turn it off AND 'lock
> to image axes' is on (the default), in which case the main image is not
> interpolated. The overlay image is always interpolated, but if both images
> align exactly, then the interpolation has negligible effect.
>
> > 3. Is there a way to do the TDI for the Gray matter? I set the thresgold
> for the FA Map very low, but its not enough I guess!
>
> There is no FA threshold with streamtrack unless you use the DT_STREAM
> algorithm, which isn't recommended for TDI (actually, it isn't recommended
> under any circumstances). If you set the cutoff sufficiently low, you
> should be able to track through these regions, although how good the
> results are will depend very much on how good your data are, these regions
> are really noisy at best... So not too sure why you would have trouble with
> that, unless 'setting the FA threshold' is a different operation from what
> I think it is... ?
>
> Cheers,
> Donald a
>
> >
> >
> > Thanks for the help,
> >
> > Ralf
> >
> >
> >
> > +++++++++++++++++++++++++++++++++++++++++++++++
> > Ralf Lützkendorf
> > Department for Biometry and Medical Informatics
> > University of Magdeburg, Germany
> > phone: +49 391 67 13546
> > fax: +49 391 67 13536
> > _______________________________________________
> > 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
> _______________________________________________
> 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/20140618/b7138211/attachment.html>