fiber direction from DTI model

[Nian Wang]

Hi Frank,

I'm going to compare the fiber orientation angular deviations at different experimental conditions using DTI model. I'm wondering whether I can the fiber direction from DSI studio. I notice we can "save the fiber directions" from "Export" in the GUI, but I'm not sure how to calculate the fiber direction using this .nii file. 

Or do you have any other suggestions?

Thanks,

Nian

[Frank Yeh]

The most straightforward way is to load FIB file in MATLAB, and you get all the fiber orientations.

http://dsi-studio.labsolver.org/Manual/export-data-to-matlab#TOC-FIB-file

--
You received this message because you are subscribed to the Google Groups "DSI Studio" group.
To unsubscribe from this group and stop receiving emails from it, send an email to dsi-studio+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/dsi-studio/0343c145-4547-473e-b1c4-3ca8ca1328c4o%40googlegroups.com.

[Surendra Maharjan]

Hi,

I read FIB data in MATLAB. It shows like this. However, the dsi-studio webpage shows like this.

The index variable is missing. Would you please tell me which one is the index variable? In any case, are rd, rd1 and rd2 the index variables?

Best Regards,

Suren

[Frank Yeh]

You may check out dir0 at http://dsi-studio.labsolver.org/Manual/export-data-to-matlab

Frank

To view this discussion on the web visit https://groups.google.com/d/msgid/dsi-studio/0dbfc75a-ef07-43ab-a3b8-9e4d1ac2b94cn%40googlegroups.com.

[Frank Yeh]

Sorry I am not sure if it is correct.

Perhaps you may try a sample data that is much smaller to see if your pipeline works correctly.

Frank

On Wed, Mar 24, 2021 at 10:14 AM Surendra Maharjan <suren...@gmail.com> wrote:

Hello Frank,

Thank you very much for your kind response. I am afraid to let you know that this is my first time doing this analysis. I converted dir0 into nifti file with [dimension,3] (please refer to above screenshot for dimension). I read nifti file in MRtrix3. I would like to know if this dir0 image is normal or not-normal.

Suren

To view this discussion on the web visit https://groups.google.com/d/msgid/dsi-studio/35a48916-bf00-4aba-aa3e-fd6f949f0a16n%40googlegroups.com.

[wangn...@gmail.com]

Hi Frank,

Suren is working with me on this project. I did check the DSI studio documentary about the dir0. 

I found the dir0 after reshape ( matlab script below) is as same as the "XXXX_fiber.nii.gz" I exported from the GUI. 

 dir0 = reshape(dir0,[3 dimension]);

We got a 3*x*y*z matrix for each voxel, which means each voxel has a  (x,y,z)  in the 3D coordinate system. For my limited understanding, this is just a point in the coordinate system, not a vector. I'm wondering how could we know the fiber direction from this (x,y,z) value? 

I just uploaded the .fib file and XXX_fiber.nii.gz to the dropbox for your review. Let me know if you have any suggestions. 

Many thanks,

Nian

[Frank Yeh]

Hi Nian,

    Hope I understood your question. The fiber orientation is the (x,y,z) you get from dir0.

Best,

Frank

To view this discussion on the web visit https://groups.google.com/d/msgid/dsi-studio/8d6deeb7-0f84-4c11-8642-69a4df0795e6n%40googlegroups.com.

[Nian Wang]

Hi Frank,

The "XXX_fiber.nii.gz" in the dropbox is from the GUI. 

I did check the dir0, after I run dir0 = reshape(dir0,[3 dimension]) and dir0 = permute(dir0, [2 3 4 1])  in matlab, it's the same as the "XXX.fiber.nii.gz" exported from GUI.

The size of dir0 is: 82575360 =256*420*256*3; dimension =[256 420 256].

Hope it's clear?

Thanks,

Nian

To view this discussion on the web visit https://groups.google.com/d/msgid/dsi-studio/CAG_QrtJPtb2GdYGUOJu9i_h%2Bp4vJhFw0wj72aRgc73-JRy7fOw%40mail.gmail.com.

[Frank Yeh]

256*420*256*3 -> the last dimension of the "3" is the fiber orientations  (dx,dy,dz)

To view this discussion on the web visit https://groups.google.com/d/msgid/dsi-studio/CAPnc0V2byYa1634J8ve75v4_rP%3DtFMijxy4M3KyVT0bONdODyw%40mail.gmail.com.

[wangn...@gmail.com]

Thanks, Frank. If I have a voxel (red dot, image below) has the (dx,dy,dz) of (0.4, 0.6, 0.1). How could we know the fiber orientation/direction? Do we need a vector (like the purple stick) to know the orientation/direction? 

Nian

[Frank Yeh]

Yes, the  256*420*256 are the coordinates at x,y,z, and the remaining 3 are the fiber orientations in unique vector.

To view this discussion on the web visit https://groups.google.com/d/msgid/dsi-studio/13eeb720-7f4e-463e-b145-cf5b7c5bbef4n%40googlegroups.com.

[Nian Wang]

The unique vector according to [0 0 0]? Otherwise, it's a point, not a vector.

Thanks,

Nian

[Frank Yeh]

It seems that I misunderstood something here.

To clarify, if the 256*420*256*3 matrix is multiplied by 255, then it will become the DTI's directional color encoding map.

Does this help tell what it is?

To view this discussion on the web visit https://groups.google.com/d/msgid/dsi-studio/76c083f3-a1d6-44f9-b27d-e9ac78bc00f8n%40googlegroups.com.

[Nian Wang]

OK, I finally got it. Thank you so much!

Nian