Whole-tract extraction

In this approach an atlas map is warped into a participant and a scalar map is averaged within it. It is suitable for descriptive summaries, replication checks and visualization. It is not recommended as a primary endpoint because spatial variation along the tract is not retained.

Procedure

The 50% binary map (_thr50) provides a conservative core; the probabilistic map (_overlap_prop, both in the atlas package) supports probability-weighted extraction. The map is warped to T1 space with the Step 1 transforms, using nearest-neighbour interpolation for binary maps and linear interpolation followed by thresholding for probabilistic maps.

ATLAS_PROB="$ATLAS_DIR/vta_posterior_hpc/left_vta_posterior_hpc_mni152_1mm_overlap_prop.nii.gz"
antsApplyTransforms -d 3 -i "$ATLAS_PROB" -r "$T1" -o prob_in_T1.nii.gz \
  -t mni2t1_1Warp.nii.gz -t mni2t1_0GenericAffine.mat -n Linear
fslmaths prob_in_T1.nii.gz -thr 0.50 -bin prob50_in_T1_bin.nii.gz

The mask is then moved to the scalar map's grid. When the T1 and the scalar map are header-aligned and differ only in grid, the mask is resampled (flirt -in mask -ref scalar -applyxfm -usesqform -interp nearestneighbour); otherwise the T1 → diffusion transform is applied. Voxel count, volume and the mean and standard deviation of the scalar within the mask are then extracted.

fslstats mask_on_scalar_grid.nii.gz -V                   # voxel count, mm3
fslstats "$NDI" -k mask_on_scalar_grid.nii.gz -M -S       # mean, SD

Reporting

For multi-shell data with a fitted neurite orientation dispersion and density imaging (NODDI) model, the neurite density index (NDI) is generally preferred to fractional anisotropy (FA) as a whole-tract summary. It remains a model-derived quantity sensitive to partial volume, registration and mask boundaries, and should be reported with covariates for age, sex, motion, intracranial volume and whole-white-matter NDI and described as a summary measure.