Dwi2tensor on ROI image in .tck file format

1

Hi MrTrix3 experts!

I’m working on a project where I would like to find out certain DTI metrics (FA, MD, AD) for a selected roi- in this case the cerebellum.

At the moment I have a .tck file for the cerebellum, which I created using the connectome2tck command, and have converted it into a .mif file using tckmap. I have then used this .mif file (cerebellum_tracks.mif) as the input into the dwi2tensor command:
dwi2tensor cerebellum_tracks.mif cerebellum_dt.mif

However I get the following error : "error importing diffusion gradient table for image ‘cerebellum_tracks.mif’ ". How could I go about solving this, such that I can create an output diffusion tensor image for this ROI and subsequently run the tensor2metric command

Many thanks in advance

2

Dear Yuscah,

I think you want to do the following:

  1. From the diffusion-weighted images (dwi.mif), calculate the diffusion tensor (dt.mif) for the whole brain (brainmask.mif) :
dwi2tensor -mask brainmask.mif dwi.mif dt.mif
  1. From the diffusion tensors (dt.mif) of the whole brain, obtain FA (fa.mif), MD (md.mif) and AD (ad.mif) for the whole brain:
tensor2metric dt.mif -fa fa.mif -adc md.mif -ad ad.mif
  1. For the metric of your choice (e.g. fa.mif), obtain statistics within your ROI of choice (e.g. cerebellum_roi.mif):
mrstats fa.mif -mask cerebellum_roi.mif

Some important remarks:

  1. The file you obtained from a tractogram using tckmap (cerebellum_tracks.mif) is a visitation count map (containing integers) and not a binary ROI (containing just true or false). In order to turn cerebellum_tracks.mif into a binary image (which you can then use in step 3. above) you can use:
mrthreshold cerebellum_tracks.mif cerbellum_roi.mif
  1. Both mrstats and mrthreshold come with a lot of options, so if you want to tweak their behavior or specify outputs, please carefully read their respective manuals. You can obtain them by issuing the commands without any parameters or options, or by providing the -help option.
3

Hi Ben,

Thank you so much for your help and clarifying the point on the use of tckmap ! You have no idea how much you’ve helped :slight_smile:

Kind Regards,
Yuscah