MRI Reconstruction Principles

Hands-on MATLAB examples for learning and teaching core MRI reconstruction methods.

Why This Repository?

This project accompanies the talk "Principles of MRI Reconstruction", presented in the session "Demystifying MRI Reconstruction: Classical Foundations to AI Frontiers" at ISMRM Cape Town 2026.

The goal is simple: make classical reconstruction ideas concrete through compact, readable implementations you can run and inspect.

What You Will Find

This repository includes illustrative implementations of:

• SENSE
• CG-SENSE
• GRAPPA

These scripts are intentionally educational and lightweight, designed to clarify concepts rather than optimize production performance.

Repository Structure

• SENSE.m - basic SENSE reconstruction example.
• CG_SENSE.m - conjugate-gradient SENSE reconstruction workflow.
• GRAPPA.m - GRAPPA reconstruction example.
• data/data_MRI_recon_principles.mat - example dataset used by the scripts.
• toolbox/ - helper functions for Fourier transforms, coil matrix construction, and utilities.

Getting Started

1. Open the repository in MATLAB.
2. Ensure the toolbox/ folder is on your MATLAB path.
3. Load the provided data file in data/.
4. Run any of the main scripts:
   • SENSE.m
   • CG_SENSE.m
   • GRAPPA.m

Audience

This material is useful for:

• students learning parallel MRI reconstruction,
• researchers teaching fundamentals before advanced/AI methods,
• anyone who wants a compact reference for classic reconstruction pipelines.

Questions

If you have questions or suggestions, feel free to reach out.