# MRzero -- Fully automated discovery of MRI sequences using supervised learning

arXiv: https://arxiv.org/abs/2002.04265 · HTML: https://mlsanity.com/paper/2002.04265/

## Key facts

- Cited by 10 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 11 Feb 2020.

## Built on by (most cited first, 10 of 10)

- [Physics-driven Synthetic Data Learning for Biomedical Magnetic Resonance](https://mlsanity.com/paper/2203.11178/) (2022)
- [Known Operator Learning and Hybrid Machine Learning in Medical Imaging -- A Review of the Past, the Present, and the Future](https://mlsanity.com/paper/2108.04543/) (2021)
- [Multi-Parameter Molecular MRI Quantification using Physics-Informed Self-Supervised Learning](https://mlsanity.com/paper/2411.06447/) (2024)
- [MRpro: open framework for model-based, learned, and quantitative MR imaging](https://mlsanity.com/paper/2507.23129/) (2025)
- [Joint MR sequence optimization beats pure neural network approaches for spin-echo MRI super-resolution](https://mlsanity.com/paper/2305.07524/) (2023)
- [Artificial Intelligence for Neuro MRI Acquisition: A Review](https://mlsanity.com/paper/2406.05982/) (2024)
- [Decoding the human brain tissue response to radiofrequency excitation using a biophysical-model-free deep MRI on a chip framework](https://mlsanity.com/paper/2408.08376/) (2024)
- [BMapEst: Estimation of Brain Tissue Probability Maps using a Differentiable MRI Simulator](https://mlsanity.com/paper/2404.14739/) (2024)
- [Controlling sharpness, SNR and SAR for 3D FSE at 7T by end-to-end learning](https://mlsanity.com/paper/2409.20251/) (2024)
- [Sequence Search: Automated Sequence Design using Neural Architecture Search](https://mlsanity.com/paper/2604.14788/) (2026)

## Abstract (arXiv)

Purpose: A supervised learning framework is proposed to automatically generate MR sequences and corresponding reconstruction based on the target contrast of interest. Combined with a flexible, task-driven cost function this allows for an efficient exploration of novel MR sequence strategies. Methods: The scanning and reconstruction process is simulated end-to-end in terms of RF events, gradient moment events in x and y, and delay times, acting on the input model spin system given in terms of proton density, T1 and T2, and DeltaB0. As a proof of concept, we use both conventional MR images and T1 maps as targets and optimize from scratch using the loss defined by data fidelity, SAR penalty, and scan time. Results: In a first attempt, textitMRzero learns gradient and RF events from zero, and is able to generate a target image produced by a conventional gradient echo sequence. Using a neural network within the reconstruction module allows arbitrary targets to be learned successfully. Experiments could be translated to image acquisition at the real system (3T Siemens, PRISMA) and could be verified in the measurements of phantoms and a human brain textitin vivo. Conclusions: Automated MR sequence generation is possible based on differentiable Bloch equation simulations and a supervised learning approach.