Can we get the angle statistics of fiber bundle by mrtrix？

1

I’m studying the fiber bundles of muscles，I want to extract statistics of muscle angle,
Is itpossible with MRtrix3?

If so how can I do it?

Your help is highly appreciated,

Best,
Wang

1 Like
2

Welcome Wang!

This is a fairly open-ended question, so we would need further information before being able to provide tangible advice. Certainly diffusion models (diffusion tensor, CSD) provide “orientation”-based information. But the concept of an “angle” necessitates the definition of a reference orientation, the angle from which some other orientation can be measured. So, for instance, do you want to estimate the muscle fibre angle in each image voxel with respect to the main magnetic field, or with respect to some reference point within the muscle? Or are you in fact interested not in an absolute angle between two orientations, but in the curvature of the fibres?

Rob

3

I want to estimate the muscle fibre angle in each image voxel with respect to the main magnetic field.Thank you for your reply.

4

Hi Wang,

In MRtrix3, all orientation-based information is defined with respect to the scanner’s axes. So the third spatial axis—the “z-axis”—is the magnetic field orientation. As such, once you have estimated a fibre orientation in each voxel, you can obtain the angle relative to the z-axis using the relevant trigonometry.

If you have a 4D image with 3 volumes, where those volumes correspond to the X-Y-Z components of a unit 3-vector, the third volume contains the Z-component of that vector. So you can extract just that component, and then convert to an angle (in radians), using:
mrconvert directions.mif -coord 3 0 -axes 0,1,2 - | mrcalc - -acos angles.mif
You could also use mrcalc to scale the result into degrees if you wish.

The source of that orientation information could come from e.g. the tensor model:
dwi2tensor dwi.mif - | tensor2metric - -vector directions.mif -modulate none
Or in a couple of ways using the results of spherical deconvolution:

sh2peaks FOD.mif directions.mif -num 1
---------
fod2fixel FOD.mif fixels/
fixel2peaks fixels/ directions.mif -number 1

Hope that’s enough to get you experimenting
Rob

5

Thank you very much.

6

HI,I am a green hand and I have some questions for you.
Can I use mrstats to get specific values?

7

Sincerely rsmith
Nice to meet you
I’m interested in calculating the white matter fiber orientation with respect to the B0 field.
however, we acquired the DTI image with the slice direction not aligned with the direction of the scanner bore (the B0 field) , with oblique slice direction.

in this case, we calculated the largest eigenvector with the commands below
dwi2tensor dwi.mif dt.mif
tensor2metric dt.mif -modulate none -vec vec.mif
mrconvert vec.mif -coord 3 2 z-component.mif
mrcalc z-component.mif -acos angle.mif

should I correct the direction of the largest eigenvector due to the oblique scan?

If you have some time, please let me know this.

thank you…

1 Like
8

One issue with this is that sh2peaks generates values between -1.5 and 1.5 so if you do -acos you get quite a few NaN values.
Could you write out how we calculate the angle (in degrees) between the z-axis and peak orientation from the sh2peaks output?