run_T2_bianca.sh — White Matter Lesion Detection¶

Location: BRC_structural_pipeline/scripts/run_T2_bianca.sh
Authors: Ali-Reza Mohammadi-Nejad, Stamatios N Sotiropoulos
Last updated: 16 Mar 2021

Note

This script is called automatically by run_T2_preprocessing.sh when --regtype is 2 or 3. It does not need to be run manually.

Purpose¶

Applies FSL BIANCA (Brain Intensity AbNormality Classification Algorithm) to detect white matter lesions on T2-FLAIR images. BIANCA is a supervised k-nearest-neighbour classifier. The BRC pipeline uses a pre-trained classifier stored in BRC_GLOBAL_DIR/templates/bianca_class_data.

Pre-flight Validation¶

Before running BIANCA, the script verifies that all six required files are present. If any file is missing, it exits with a descriptive error:

Problem running Bianca. File <path> is missing

Required files:

File	Location
`T1_unbiased_brain.nii.gz`	T1 temp folder
`T1_unbiased.nii.gz`	T1 temp folder
`T2_unbiased.nii.gz`	T2 working directory
`T1_to_MNI_nonlin_coeff_inv.nii.gz`	T1 registration temp folder
`T1_to_MNI_linear.mat`	T1 registration temp folder
`T1_brain_pve_0.nii.gz`	T1 FAST folder

Processing Steps¶

1. Create Inclusion Mask¶

make_bianca_mask \
    T1_unbiased.nii.gz \
    T1_brain_pve_0.nii.gz \
    T1_to_MNI_nonlin_coeff_inv.nii.gz

Generates a white matter / periventricular inclusion mask (T1_unbiased_bianca_mask) to remove grey matter from BIANCA results.

2. Generate Configuration File¶

echo "T1_brain.nii.gz T2_unbiased.nii.gz T1_to_MNI_linear.mat" \
    > conf_file.txt

Single-line config listing the T1, T2-FLAIR, and T1→MNI affine for BIANCA's feature extraction.

3. Run BIANCA¶

bianca \
    --singlefile=conf_file.txt \
    --querysubjectnum=1 \
    --brainmaskfeaturenum=1 \
    --loadclassifierdata=bianca_class_data \
    --matfeaturenum=3 \
    --featuresubset=1,2 \
    -o bianca_mask

Features used: T2-FLAIR intensity (featuresubset=1) and T1 intensity (featuresubset=2).

4. Apply Inclusion Mask and Threshold¶

fslmaths bianca_mask \
    -mul T1_unbiased_bianca_mask.nii.gz \
    -thr 0.8 \
    -bin \
    final_mask

Probability values ≥ 0.8 within the WM inclusion mask are kept and binarised.

5. Extract Lesion Volume¶

fslstats final_mask -V | awk '{print $1}' > volume.txt

Total voxel count written to volume.txt.

Parameters¶

Parameter	Type	Description
`--workingdir`	Path	T2 temp folder (provides `T2_unbiased.nii.gz`)
`--tempt1folder`	Path	T1 temp folder (provides `T1_unbiased` files)
`--fastfolder`	Path	FAST folder (provides `T1_brain_pve_0` for inclusion mask)
`--biancatempfolder`	Path	BIANCA working directory for all outputs
`--regtempt1folder`	Path	T1 registration temp folder (MNI warp and affine)
`--logfile`	Path	Log file path

Key Outputs¶

File	Description
`BIANCA/bianca_mask.nii.gz`	Raw BIANCA probability map (0–1 per voxel)
`BIANCA/final_mask.nii.gz`	Binary lesion mask (threshold 0.8, inclusion mask applied)
`BIANCA/volume.txt`	Total lesion volume in voxels
`BIANCA/T1_unbiased_bianca_mask.nii.gz`	WM inclusion mask
`BIANCA/T1_unbiased_ventmask.nii.gz`	Ventricular mask (from `make_bianca_mask`)
`BIANCA/conf_file.txt`	BIANCA input configuration file