Custom atlas -- THOMAS

[fhaos123]

I would like to visualize the THOMAS Atlas by following the instructions on this website: https://anniegbryant.github.io/subcortex_visualization/custom_segmentation.

The original labels of the THOMAS Atlas have been remapped to new labels (ranging from 1 to 10), with the Hb and MTT labels excluded. I ran the script volume_to_mesh_mz3.py and successfully generated the corresponding mz3 files; however, these mz3 files fail to render the 3D images correctly in Surf Ice.

However, the Melbourne_S1_subcortex.mz3 file — which is generated from Melbourne_S1_subcortex.nii.gz — can be successfully rendered without any issues.

code:

(subcortex_vis) PS C:\Users\78270\Desktop\Pattern_of_thalamus_atrophy\Viscode> python volume_to_mesh_mz3.py --input_volume Atlas-Thalamus_left_relabel_1to10.nii.gz --output_path ./ --out_file Atlas-Thalamus_space-MNI_hemi-left_label-AllNuclei_desc-MaxProb.mz3 --index_max 10 --colors my_colors.txt --delete_mz3
Processing indices from 1 to 10.
Running: niimath Atlas-Thalamus_left_relabel_1to10.nii.gz -thr 1 -uthr 1 -bin -s 1.2 -mesh -i 0.5 -r 0.5 -q b ./Atlas-Thalamus_left_relabel_1to10.nii_1
Saved mesh: ./Atlas-Thalamus_left_relabel_1to10.nii_1.mz3
Running: niimath Atlas-Thalamus_left_relabel_1to10.nii.gz -thr 2 -uthr 2 -bin -s 1.2 -mesh -i 0.5 -r 0.5 -q b ./Atlas-Thalamus_left_relabel_1to10.nii_2
Saved mesh: ./Atlas-Thalamus_left_relabel_1to10.nii_2.mz3
Running: niimath Atlas-Thalamus_left_relabel_1to10.nii.gz -thr 3 -uthr 3 -bin -s 1.2 -mesh -i 0.5 -r 0.5 -q b ./Atlas-Thalamus_left_relabel_1to10.nii_3
Saved mesh: ./Atlas-Thalamus_left_relabel_1to10.nii_3.mz3
Running: niimath Atlas-Thalamus_left_relabel_1to10.nii.gz -thr 4 -uthr 4 -bin -s 1.2 -mesh -i 0.5 -r 0.5 -q b ./Atlas-Thalamus_left_relabel_1to10.nii_4
Saved mesh: ./Atlas-Thalamus_left_relabel_1to10.nii_4.mz3
Running: niimath Atlas-Thalamus_left_relabel_1to10.nii.gz -thr 5 -uthr 5 -bin -s 1.2 -mesh -i 0.5 -r 0.5 -q b ./Atlas-Thalamus_left_relabel_1to10.nii_5
Saved mesh: ./Atlas-Thalamus_left_relabel_1to10.nii_5.mz3
Running: niimath Atlas-Thalamus_left_relabel_1to10.nii.gz -thr 6 -uthr 6 -bin -s 1.2 -mesh -i 0.5 -r 0.5 -q b ./Atlas-Thalamus_left_relabel_1to10.nii_6
Saved mesh: ./Atlas-Thalamus_left_relabel_1to10.nii_6.mz3
Running: niimath Atlas-Thalamus_left_relabel_1to10.nii.gz -thr 7 -uthr 7 -bin -s 1.2 -mesh -i 0.5 -r 0.5 -q b ./Atlas-Thalamus_left_relabel_1to10.nii_7
Saved mesh: ./Atlas-Thalamus_left_relabel_1to10.nii_7.mz3
Running: niimath Atlas-Thalamus_left_relabel_1to10.nii.gz -thr 8 -uthr 8 -bin -s 1.2 -mesh -i 0.5 -r 0.5 -q b ./Atlas-Thalamus_left_relabel_1to10.nii_8
Saved mesh: ./Atlas-Thalamus_left_relabel_1to10.nii_8.mz3
Running: niimath Atlas-Thalamus_left_relabel_1to10.nii.gz -thr 9 -uthr 9 -bin -s 1.2 -mesh -i 0.5 -r 0.5 -q b ./Atlas-Thalamus_left_relabel_1to10.nii_9
Saved mesh: ./Atlas-Thalamus_left_relabel_1to10.nii_9.mz3
Running: niimath Atlas-Thalamus_left_relabel_1to10.nii.gz -thr 10 -uthr 10 -bin -s 1.2 -mesh -i 0.5 -r 0.5 -q b ./Atlas-Thalamus_left_relabel_1to10.nii_10
Saved mesh: ./Atlas-Thalamus_left_relabel_1to10.nii_10.mz3
Combining individual .mz3 files into a single colored atlas...
Using input path: ./
Combined mesh saved as: Atlas-Thalamus_space-MNI_hemi-left_label-AllNuclei_desc-MaxProb.mz3
Deleted intermediate mesh files.

(subcortex_vis) PS C:\Users\78270\Desktop\Pattern_of_thalamus_atrophy\Viscode> python volume_to_mesh_mz3.py --input_volume Atlas-Thalamus_right_relabel_1to10.nii.gz --output_path ./ --out_file Atlas-Thalamus_space-MNI_hemi-right_label-AllNuclei_desc-MaxProb.mz3 --index_max 10 --colors my_colors.txt --delete_mz3
Processing indices from 1 to 10.
Running: niimath Atlas-Thalamus_right_relabel_1to10.nii.gz -thr 1 -uthr 1 -bin -s 1.2 -mesh -i 0.5 -r 0.5 -q b ./Atlas-Thalamus_right_relabel_1to10.nii_1
Saved mesh: ./Atlas-Thalamus_right_relabel_1to10.nii_1.mz3
Running: niimath Atlas-Thalamus_right_relabel_1to10.nii.gz -thr 2 -uthr 2 -bin -s 1.2 -mesh -i 0.5 -r 0.5 -q b ./Atlas-Thalamus_right_relabel_1to10.nii_2
Saved mesh: ./Atlas-Thalamus_right_relabel_1to10.nii_2.mz3
Running: niimath Atlas-Thalamus_right_relabel_1to10.nii.gz -thr 3 -uthr 3 -bin -s 1.2 -mesh -i 0.5 -r 0.5 -q b ./Atlas-Thalamus_right_relabel_1to10.nii_3
Saved mesh: ./Atlas-Thalamus_right_relabel_1to10.nii_3.mz3
Running: niimath Atlas-Thalamus_right_relabel_1to10.nii.gz -thr 4 -uthr 4 -bin -s 1.2 -mesh -i 0.5 -r 0.5 -q b ./Atlas-Thalamus_right_relabel_1to10.nii_4
Saved mesh: ./Atlas-Thalamus_right_relabel_1to10.nii_4.mz3
Running: niimath Atlas-Thalamus_right_relabel_1to10.nii.gz -thr 5 -uthr 5 -bin -s 1.2 -mesh -i 0.5 -r 0.5 -q b ./Atlas-Thalamus_right_relabel_1to10.nii_5
Saved mesh: ./Atlas-Thalamus_right_relabel_1to10.nii_5.mz3
Running: niimath Atlas-Thalamus_right_relabel_1to10.nii.gz -thr 6 -uthr 6 -bin -s 1.2 -mesh -i 0.5 -r 0.5 -q b ./Atlas-Thalamus_right_relabel_1to10.nii_6
Saved mesh: ./Atlas-Thalamus_right_relabel_1to10.nii_6.mz3
Running: niimath Atlas-Thalamus_right_relabel_1to10.nii.gz -thr 7 -uthr 7 -bin -s 1.2 -mesh -i 0.5 -r 0.5 -q b ./Atlas-Thalamus_right_relabel_1to10.nii_7
Saved mesh: ./Atlas-Thalamus_right_relabel_1to10.nii_7.mz3
Running: niimath Atlas-Thalamus_right_relabel_1to10.nii.gz -thr 8 -uthr 8 -bin -s 1.2 -mesh -i 0.5 -r 0.5 -q b ./Atlas-Thalamus_right_relabel_1to10.nii_8
Saved mesh: ./Atlas-Thalamus_right_relabel_1to10.nii_8.mz3
Running: niimath Atlas-Thalamus_right_relabel_1to10.nii.gz -thr 9 -uthr 9 -bin -s 1.2 -mesh -i 0.5 -r 0.5 -q b ./Atlas-Thalamus_right_relabel_1to10.nii_9
Saved mesh: ./Atlas-Thalamus_right_relabel_1to10.nii_9.mz3
Running: niimath Atlas-Thalamus_right_relabel_1to10.nii.gz -thr 10 -uthr 10 -bin -s 1.2 -mesh -i 0.5 -r 0.5 -q b ./Atlas-Thalamus_right_relabel_1to10.nii_10
Saved mesh: ./Atlas-Thalamus_right_relabel_1to10.nii_10.mz3
Combining individual .mz3 files into a single colored atlas...
Using input path: ./
Combined mesh saved as: Atlas-Thalamus_space-MNI_hemi-right_label-AllNuclei_desc-MaxProb.mz3
Deleted intermediate mesh files.

I have uploaded my relevant data as an attachment.

Any help will be greatly appreciated.

Atlas-Thalamus_left_relabel_1to10.nii.gz
Atlas-Thalamus_right_relabel_1to10.nii.gz
Atlas-Thalamus_space-MNI_hemi-left_label-AllNuclei_desc-MaxProb.nii.gz
Atlas-Thalamus_space-MNI_hemi-right_label-AllNuclei_desc-MaxProb.nii.gz

mz3.tar.gz

my_colors.txt

[anniegbryant]

Hey there, thank you for checking out the package and also for sharing this detailed summary of your experience trying to create and visualize the mesh for the THOMAS atlas.

Just wanted to let you know I've seen this and will try to figure this out (or raise an issue with the group that developed the mesh generation and visualization software, if not). I will follow up in this thread either way.

Cheers, Annie 😊

[anniegbryant]

Hi @fhaos123, thanks again for providing such a detailed overview of this issue, which I was indeed able to replicate. I've opened an issue with the group that developed the Python scripts used in this workflow (led by Professor Chris Rorden), and will loop you in when this issue is (hopefully) resolved. You can read more about this issue here: rordenlab/pythonScripts#1.

For the time being, I will note that a MATLAB alternative that is working for this atlas, and could be helpful to you if you have access to MATLAB. Either way, I've provided the successfully converted atlases from MATLAB below for your convenience; one limitation to this option is that I haven't figured out how to customize the color scheme in this pipeline yet. The repository for this MATLAB version can be found at https://github.com/neurolabusc/surfice_atlas/tree/master, and the specific MATLAB script that I used for this is nii_nii2atlas.m. It's very straightforward -- just pass in the NIFTI volume you want to convert to a mesh, e.g.:

nii_nii2atlas('Atlas-Thalamus_left_relabel_1to10.nii.gz')

This creates the file merge.mz3, which you should be able view in Surfice. Here's a screenshot of the left hemisphere mesh created with this MATLAB workflow:

Converted meshes:
📁THOMAS_left.mz3.gz
📁THOMAS_right.mz3.gz

I'll continue to try to debug this myself, but wanted to provide this update and the rendered meshes for now. I have also corresponded with the lead author on the THOMAS manuscript about how best to incorporate that atlas into this toolbox; if you're interested in contributing a vectorized version of this atlas for community use, feel free to let me know.

Cheers,
Annie 😊