Segment GMWMI corresponding to cortical/subcortical segmentation

1

Hello Mrtrix Experts,

  1. I have used “labelconvert” and “labelsgmfix” commands to obtain nodes_fixSGM.mif in the diffusion space.
  2. I have also used “5tt2gmwmi” command to obtain gmwmi_b0.mif in the diffusion space.

However, i dont see a perfect/complete overlay of the gmwmi_b0.mif over the cortical/subcortical segmentations from nodes_fixSGM.mif. Is it possible to segment the gmwmi_b0.mif based on the cortical/subcortical segmentations from nodes_fixSGM.mif? I am interested in looking at the interaction between the grey matter cortical/subcortical measures and white matter measures from their corresponding interface.

BW
Jay

2

Hi Jay,
What algorithm was used to generate the 5TT image?

Cheers,
Arkiev

3

Hi Arkiev,

I used the 5ttgen fsl command to generate the 5TT image.

BW
Jay

4

Hi Jay - do you have any screenshots to share?

The issue may stem from the gmwmi_b0 image allowing for partial volume effects, whereas the output from labelsgmfix only allows for integers.

It might be possible to generate a boundary image from labelsgmfix output by:

  1. extract node of interest (using mrcalc and -eq)
  2. use maskfilter (with erode) to create a temporary mask
  3. compute mrcalc <extracted_node> <temporary_mask> -sub <extracted_node_boundary>
  4. repeat steps 1-3 for other nodes of interest
  5. add extracted boundary images together to create a single image containing boundary images for all nodes of interest.

Hope this helps!

Cheers,
Arkiev