Superficial WM

Surfaces at various white matter depths (Laplace-field based)

See repository

Superficial White Matter generates surfaces at various white matter depths (default 1, 2, and 3 mm). The depths are calculated from the real-world voxel size, and each surface is produced by computing a Laplace field over white matter (cortex to subcortex plus ventricles) and shifting an existing white matter surface along that gradient, with stopping conditions set by geodesic distance travelled. The tool is also implemented in micapipe v0.2.3.

Usage

The code expects standard NIFTI orientation (run inputs through fslreorient2std). With a FreeSurfer/FastSurfer subject directory:

# 1. Calculate the Laplace field over white matter
python sWM/laplace_solver.py \
    ${aparc_aseg_nifti} \
    ${OUT}/${SUBJECT}_laplace-wm.nii.gz

# 2. Generate the depth surfaces for each hemisphere
for hemi in lh rh; do
    python sWM/surface_generator.py \
        "${OUT}/${SUBJECT}_hemi-${hemi}_label-white.surf.gii" \
        ${OUT}/${SUBJECT}_laplace-wm.nii.gz \
        ${OUT}/${SUBJECT}_hemi-${hemi}_label-sWF_depth- \
        1 2 3
done

Installation is git clone plus pip install superficial-white-matter/; a complete worked example is in the repository.

Jordan DeKraker, Ricardo Cruces, and Yu Hwang. “Superficial White Matter” Zenodo, 2024.