Tensor2metric question

Dear MRtrix experts,

I have downloaded the preprocessed HCP young adults dti data, converted the data.nii.gz to dwi.mif and ran below to obtain FA, MD, AD, and RD images

dwi2tensor -force -mask nodif_brain_mask.nii.gz dwi.mif tensor.mif
tensor2metric -force -fa FA.nii.gz tensor.mif
tensor2metric -force -adc MD.nii.gz tensor.mif
tensor2metric -force -ad AD.nii.gz tensor.mif
tensor2metric -force -rd RD.nii.gz tensor.mif

FA

MD

AD

RD

FA image, however, looks a little different from what it normally looks like. Could you please advise if I have missed any steps or did anything incorrectly? Also, could you please let me know if MD, AD, and RD images look ok?

Best,
JW

Hi @JWon ,

From your code, I can’t decide whether there is a problem. The visualization results can be different because of a lot of things. You can check the view options in mrview to see if there is any difference, I think Intensity scaling may have some effects. Or maybe you should use other toolboxes(e.g. FSL) to generate metrics and directly check the voxel-vise values.

Still, some optimization can be applied.

First, from where I stand, the brain mask that HCP provided is not so good, so maybe you can generate a new mask for DWI image using MRtrix3.

Secendly, although HCP pipeline has preprocessed the data. It is suggested that mrdenoise and mrdegibbs should be used in the tutorial.

Lastly, the order of your command is a bit confusing, I prefer to type like this:

tensor2metric tensor.mif -fa FA.nii.gz -adc MD.nii.gz -ad AD.nii.gz -rd RD.nii.gz -force

 
Best,
Volcano

Hi Volcano,

As you suggested, I ran mrdenoise, mrdegibbs, and dwi2mask and the images look now fine.

Screen Shot 2022-11-04 at 9.23.22 AM

Thank you!

Best,
JW