This code is for calulating and visualising the volumes of different brain tissues, specifically using SPM 12 grey matter (c1), white matter (c2), and cerebrospinal fluid (CSF, c3). It fetches pre-segmented anatomical images, calculates volumes, and showcases interactive visualisations. Note that I wrote it for Gannet Toolbox segmentation outputs (which uses SPM).
nibabel numpy os nilearn
-
Define Image Paths: Sets the paths for the original anatomical T1-weighted image and its corresponding segmented images (grey matter, white matter, and CSF).
-
Compute Volume from Mask: A function to calculate the volume of non-zero voxels from a mask. The volume calculation is based on the voxel dimensions of the image.
-
Volume Calculations: Utilises the above function to compute volumes of grey matter, white matter, and CSF. Additionally, it calculates the total brain volume excluding and including CSF.
-
Volume Conversion to mL: For ease of understanding, computed volumes (originally in mm^3) are converted to mL.
-
Results Display: Outputs the computed volumes in a structured format.
-
Interactive Visualisation:
- Displays the anatomical T1-weighted image in an interactive grayscale mode within Jupyter Lab.
- Visualises the grey and white matter segmentation overlayed on the anatomical image. These visualisations are saved as HTML files and can be viewed directly in a browser.
-
Modify Base Directory: Ensure you set
base_dirto the path where your anatomical and segmented images are stored. -
Run the Script: Execute the Python script in a Jupyter Lab environment or a similar platform that supports interactive visualisation. If not using Jupyter Lab, you might want to comment out or modify the inline visualisation sections.
-
View Results: Check the printed results for volume calculations and open the generated HTML files ("sub-013_GreyMatter.html" and "sub-013_WhiteMatter.html") in a browser for visual representation.
- The code assumes that the segmented images and the anatomical T1-weighted image are named in a specific format. If your filenames differ, ensure you update the paths in the code accordingly.
- The visualisations use a threshold of "90%". You can adjust this value if needed.