To change fs_default

,
1

Dear community:

This is more than an error a general question. I hope it can be solved. My question is how can I perform a structural connectome analysis with more detail of the structures in the cerebellum than with fs_default (84x84 matrix) parcelation. As in this file there is only cerebellar cortex left and right but no lobes. I need to be able to report more specific results if possible.

Thanks for the help in advance.

1 Like
2

Hello @andraderenew ,

FreeSurfer doesn’t provide cerebellar parcellations by default, so you’ll have to use another method.

I just did a quick search and you could achieve it with this tool.

Hope this helps,
Nick

3

let me try to understand this that you are proposing. Is it to parcellate the image with this program and then what steps should I follow. Where am I at in the structural connectome processing?
I have to run this step (see commands below) where I substitute aparc+aseg.mgz and fs_default.txt I guess with something else and carry on the analysis.

labelconvert sub-CON02_recon/mri/aparc+aseg.mgz $FREESURFER_HOME/FreeSurferColorLUT.txt /usr/local/mrtrix3/share/mrtrix3/labelconvert/fs_default.txt sub-CON02_parcels.mif

4

Hello,

Yes this would consist of:

  1. Removing the cerebellar cortex labels from your existing whole brain parcellation (I usually create a mask for the existing value, invert it and multiply by the original image to make that value 0)
  2. Giving the new cerebellar parcels unique values and then adding them into the whole brain (minus cerebellar cortex) parcellation
  3. Adding the cerebellar labels into the FreeSurferColorLUT.txt - along with their current unique values in the first column
  4. Taking out the rows for the cerebellar cortex from the fs_default.txt and reordering, and adding the new cerebellar labels at the end

Then you can come back to this labelconvert step with the new inputs and continue from there…

All the best,
Nick

1 Like