Multi-TE ASL Processing Scripts

This repository contains a set of Bash and Python scripts for post-processing multi–echo arterial spin labeling (Multi-TE ASL) MRI data, with a particular focus on blood–brain barrier (BBB) integrity measurements.
The pipeline implements a modified Multi-TE ASL signal model and voxelwise perfusion calibration, following published and validated methods.

---

Contents

1. Software Requirements
2. Modified Multi-TE ASL Model
3. Pipeline Overview & Scripts
4. Suggested Reading
5. References

---

1. Software Requirements

This pipeline is based on the FMRIB Software Library (FSL) and is intended to run on Linux systems (e.g., Ubuntu or CentOS).

Required software

• FSL (recommended: most recent stable version)
  Installation instructions:
  https://fsl.fmrib.ox.ac.uk/fsl/fslwiki/FslInstallation

• Fabber (from FSL)
  Fabber is a tool for fitting timeseries data to a forward model using a Bayesian approach. (Chappell et al., 2008).

• dcm2niix for DICOM to NIfTI conversion
  https://github.com/rordenlab/dcm2niix

• Python (for M0 fitting)

  • numpy
  • scipy
  • nibabel

Important
The modified Multi-TE ASL model used here is not yet part of the standard FSL distribution.
Model source files must be manually copied and fabber recompiled to enable fitting.

---

2. Modified Multi-TE ASL Model

This pipeline uses a modified version of the original Multi-TE ASL model (Gregori et al., 2013), optimized for BBB integrity analysis.

Key features:

• Joint modeling of perfusion and multi-echo decay
• Estimation of:
  • Perfusion (CBF)
  • Arterial transit time (ATT)
  • Exchange time (T_exch)
  • Intra-voxel transit time (ITT)

Details of the model modification, validation, and performance are described in:

Mahroo et al., 2021 — Robust Multi-TE ASL-Based Blood–Brain Barrier Integrity Measurements

The modified model must be compiled into Fabber before running the fitting step.

---

3. Pipeline Overview & Scripts

The scripts are numbered chronologically and must be executed in order.

1. 1_master

Master configuration script:

• Defines ASL acquisition parameters
• Sets directory structure
• Specifies model parameters used by Fabber
  Model parameters are literature-based and should not be modified unless explicitly justified.

---

2. 2_dcm2niix

• Creates directory structure
• Converts DICOM data to NIfTI using dcm2niix

---

3. 3_Structural_Preprocessing

• Structural preprocessing using fsl_anat
• Brain extraction
• Tissue segmentation
• Partial volume estimation (PVE)

Outputs are used for:

• Registration
• Partial-volume–aware ROI analysis

---

4. 4_Mask_ASL_Space

• Generates ASL brain mask
• Uses first echo of HAD-4 dataset as reference
• Two-step registration between ASL and T1 space

---

5. 5_Reg_M0_to_ASL_with_topup

• Registers M0 images to ASL reference
• Applies distortion correction using TOPUP if M0 data with opposing phase-encoding directions are available

---

6. 6_HAD_4_MultiTE_MotionCorrection_Decoding

• Motion correction
• Distortion correction
• Hadamard decoding for HAD-4 multi-TE dataset

Decoding matrix is provided and should be verified.
Current implementation assumes Walsh-ordered Hadamard encoding.

---

7. 7_HAD_8_SingleTE_MotionCorrection_Decoding

• Same preprocessing steps as above
• Applied to HAD-8 single-TE dataset

---

8. 8_Concatenating_Datasets

• Concatenates HAD-4 and HAD-8 datasets
• Produces final input for model fitting

---

9. 9_FabberCommand_Merged

• Calls Fabber with the modified Multi-TE ASL model
• Outputs voxelwise parameter maps:
  • mean_ftiss (perfusion)
  • mean_delttiss (ATT)
  • mean_T_exch (exchange time)
  • mean_ITT (intra-voxel trans_