1^ST ORDER MRI SIMS

First-order Spatial Encoding Simulations for Improved Accuracy in the Presence of Strong $\mathrm{B}_0$ and Gradient Field Variations

[Paper] [Github] [BibTeX]

Overview

This repository contains code to perform 1^st-order spatial encoding MRI simulations, developed especially for highly variable and inhomogeneous main magnetic (B₀) and gradient fields (BG_xyz). This simulation code is a starting point for exploring how to capture the encoding effects of strong field variations accurately and efficiently to enable the assessment of distortions, signal dropout, and foldover artifacts.

The following figure determines an example scenario in which these simulations can be used for improved accuracy:

For more details, please refer to the paper.

Installation

The code requires julia >= v"1.9.0"

Install 1^ST ORDER MRI SIMS: clone the repository locally and install with

git clone https://github.com/radhikatibrewala/1stOrderMRISims.git

Usage

The repository is centered around the MainSlurm.jl file. The following breaks down the basic structure:

• Project.toml: Contains the packages needed for the project. Each entry specifies the package name and its version requirement
  
• Initialize.jl: Contains Pkg.instantiate() which reads the Project.toml file and automatically creates a project environment with the specified dependencies and versions. Pkg.instantiate() will trigger package installation when run for the first time
  
• MainSlurm.jl: Main file to run the simulation
  
• src: Contains the source code for the project, including all relevant functions that are called upon in MainSlurm.jl
  
• examples: Contains the example quadratic field configuation, example input image and a bash file gen_kspace.sh that calls upon MainSlurm.jl
  
• RunTests.jl: Compiles the code and runs tests on a gradient echo and spin echo sequence using files from tests
  
• tests: Contains test scripts called upon in RunTests.jl
  
• FinalImageSim: Final file to be run once MainSlurm.jl has finished running, to create the image from the simulated k-space
  

Hardware Requirements

This is a very computationally heavy code and has been tested on High Performance Computing Systems. Julia uses threads to speed up the calculation by splitting it on multiple cores. To run the example gen_kspace.sh file, you will need:

• A computer with at least 256GB of RAM
  
• A multi-core CPU
  

License

1^ST ORDER MRI SIMS is MIT licensed, as found in LICENSE file

Cite

If you use the 1^ST ORDER MRI SIMS code in your research, please use the following BibTeX entry.