A Beginner's Guide to Bloch Equation Simulations of Magnetic Resonance Imaging Sequences

TL;DR

This paper provides a beginner-friendly guide to simulating magnetic resonance imaging sequences using Bloch equations, emphasizing conceptual understanding and practical implementation for novice users.

Contribution

It introduces accessible methods and example code for Bloch equation simulations, including analytical and numerical approaches, tailored for newcomers in MRI simulation.

Findings

Analytical solutions for no-RF pulse regimes

Efficient numerical simulation using Rodrigues' rotation formula

Practical code snippets for MR sequence simulation

Abstract

Nuclear magnetic resonance (NMR) concepts are rooted in quantum mechanics, but MR imaging principles are well described and more easily grasped using classical ideas and formalisms such as Larmor precession and the phenomenological Bloch equations. Many textbooks provide in-depth descriptions and derivations of the various concepts. Still, carrying out numerical Bloch equation simulations of the signal evolution can oftentimes supplement and enrich one's understanding. And though it may appear intimidating at first, performing these simulations is within the realm of every imager. The primary objective herein is to provide novice MR users with the necessary and basic conceptual, algorithmic and computational tools to confidently write their own simulator. A brief background of the idealized MR imaging process, its concepts and the pulse sequence diagram are first provided. Thereafter, two regimes of Bloch equation simulations are presented, the first which has no radio frequency (RF) pulses, and the second in which RF pulses are applied. For the first regime, analytical solutions are given, whereas for the second regime, an overview of the computationally efficient, but often overlooked, Rodrigues' rotation formula is given. Lastly, various simulation conditions of interest and example code snippets are given and discussed to help demonstrate how straightforward and easy performing MR simulations can be.