MRI acquisition and reconstruction cookbook: recipes for reproducibility, served with real-world flavour

This repository contains the code to reproduce the results of the article MRI acquisition and reconstruction cookbook: recipes for reproducibility, served with real-world flavour.

Tamir JI, Blumenthal M, Wang J, Oved T, Shimron E, Zaiss M. MRI acquisition and reconstruction cookbook: recipes for reproducibility, served with real-world flavour. MAGMA. 2025 Mar 6. doi: 10.1007/s10334-025-01236-4. Epub ahead of print. PMID: 40048131. https://doi.org/10.1007/s10334-025-01236-4

Requirements

This repository has various requirements to run the experiments. Key requirements are:

• Fig 3/4:
  • pypulseq (version 1.4.2)
  • MRzeroCore (version 0.3.12)
  • A conda environment file is provided in environment.yml to install all required packages.
• Fig 6/7:
  • bart (version 0.9.00) c.f. https://github.com/mrirecon/bart
  • view (version 0.3.00) c.f. https://github.com/mrirecon/view
  • For interaction of BART and python (Fig 7), the BART_TOOLBOX_PATH environment variable must be set to the path of the BART toolbox.

For composing of figures to final PDFs, inkscape is required.

The code was tested on Debian 12.

Data

k-Space data to reproduce the reconstructions (Figures 6 and 7) is hosted on Zenodo . The data can be downloaded with the script in Data/download_data.sh.

Reproducing

All experiments can be reproduced by running

export BART_TOOLBOX_PATH=/path/to/bart/
./run_all.sh

How to cite

Tamir, J.I., Blumenthal, M., Wang, J. et al. MRI acquisition and reconstruction cookbook: recipes for reproducibility, served with real-world flavour. Magn Reson Mater Phy (2025). https://doi.org/10.1007/s10334-025-01236-4