Save tracts in MNI space

[tuijnde...@gmail.com]

Dear Frank and DSI users,

Could you provide me with some more information on how DSI studio handles the transformation from tracts in native space to MNI space?
The tracts that I am trying to transform were not generated in DSI studio but in Matlab and uploaded to DSI studio. I did create a corresponding .fib file that contains the fa image in diffusion space.
During the calculation of the transformation ( affine, non-linear?), does it only use information stored in the .fib file, or does it use the vox to ras transformation stored in the trk header?
I noticed that after the transformation both the coordinates and the vox to ras field in the header are changed. 

Lastly, In the 3D view, the tracts overlay perfectly on the FA map stored in the fib file. Although when I create a tract density image in diffusion space and lets say open both of them in mricron, they do not overlay anymore.
I believe DSI studio handles everything in image space? So probably while it creates the file, I am missing some extra header information that is not necessary in DSI studio for correct rendering?

I appreciate your advice.

Thank you,
Emilie





 

[Fang-Cheng Yeh]

DSI Studio first did a linear registration between the subject's anisotropy map and a built-in template to get a linear transformation matrix. Then an SPM-like normalization will be conducted between linearly transformed subject anisotropy map and the built-in template to get the nonlinear transformation. The tracks will be converted to MNI space using both the linear transformation and nonlinear transformation.

The misalignment problem you met could be more complicated.  If possible, please send me the FIB file and the tracks file to fix the issue here.

Best regards,

Frank 

--
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+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Emilie Tuijnder]

Thank you Frank for the quick reply. I uploaded an example .fib and .trk file to the dropbox.

As I am working with stroke patients with pretty large lesions, I assume that the non-linear transformation does not work properly on the left side which is messing up the rendering in MNI space.

I will try and apply the linear transformation only to the .trk file and see if this solves my problem.

Best,

Emilie

--
You received this message because you are subscribed to a topic in the Google Groups "DSI Studio" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/dsi-studio/XpsAIJCxrVE/unsubscribe.
To unsubscribe from this group and all its topics, send an email to dsi-studio+unsubscribe@googlegroups.com.

[Fang-Cheng Yeh]

Hi Emilie,

    I loaded the fa map (exported from DSI Studio using [View][Save qa map..]) and the track density image in mricron, and they overlay nicely:

    Here you have all data in the native space, and thus DSI Studio does not use any transformation information in the trk file or in the fib file. You may see the header in the nifti file generated from DSI Studio, and the transformation matrix is a identity matrix with scaling only.

Best,

Frank

[Emilie Tuijnder]

Thank you Frank,  I was using the original fa file, which has a different transformation matrix in the header. So either the header of the tract density image needs to be adjusted or the fa.nii. That makes sense.

Last question, whenever you apply the linear transformation to the coordinates in the .trk files, do you just multiple the 3x3 with the coordinates to rotate them? what do you do with the linear translation?

I apologize for the many questions, this switching between coordinate systems and different programs has been a very frustrating problem to fix and there is very little documentation on how to do it correctly.

My best,

Emilie

[Fang-Cheng Yeh]

Hi Emilie,

    DSI Studio does not apply any transformation matrix here. It ignores the information in the trk file.

    If there are more questions, please feel free to let me know.

Best regards,

Frank

[Emilie Tuijnder]

Sorry, I meant whenever DSI studio saves the trk's in MNI space. I assume in a first step it multiplies the transformation matrix to the coordinates one way or another?

I have been trying to replicate this in matlab without success, as I need to take out the non linear transformation.

Best,

Emilie

[Fang-Cheng Yeh]

Saving the tracks in MNI space is more complicated. It follows the approach I mentioned in the previous email. DSI Studio calculates its own transformation matrix and nonlinear mapping parameters. This step cannot be replicated in MATLAB.

Best,

Frank