Skip to main content

LUP Student Papers

LUND UNIVERSITY LIBRARIES

Maxwell term correction at low magnetic fields and strong gradients

Olsson, Viktor (2025) MSFT02 20252
Medical Physics Programme
Abstract
Diffusion magnetic resonance imaging (MRI) is a technique that utilizes the random thermal motion of water molecules to probe the microstructure of the tissue. Because it has excellent soft tissue contrast and high sensitivity to changes in the microstructure, it has established itself as a powerful tool to detect and quantify both normal development and pathological processes. Diffusion encoding is commonly achieved using an identical pair of pulsed gradients placed on either side of the refocusing pulse. Although this sequence design is not time-efficient, it exploits a symmetry that implicitly ensures negligible concomitant gradient effects and thus has negligible effect on signal accuracy. This experimental design is by far the... (More)
Diffusion magnetic resonance imaging (MRI) is a technique that utilizes the random thermal motion of water molecules to probe the microstructure of the tissue. Because it has excellent soft tissue contrast and high sensitivity to changes in the microstructure, it has established itself as a powerful tool to detect and quantify both normal development and pathological processes. Diffusion encoding is commonly achieved using an identical pair of pulsed gradients placed on either side of the refocusing pulse. Although this sequence design is not time-efficient, it exploits a symmetry that implicitly ensures negligible concomitant gradient effects and thus has negligible effect on signal accuracy. This experimental design is by far the dominant method for conventional diffusion-weighted imaging. To further boost the encoding efficiency and signal-to-noise-ratio, an alternative is to employ asymmetric gradient waveforms. The trade-off is that asymmetric waveforms do not have an implicit control for concomitant gradient effects and are therefore uniquely susceptible to unwanted spin dephasing and erroneous loss of signal. This problem has been addressed through numerical optimization of the gradient waveform wherein the so-called ‘Maxwell index’ is constrained to ensure that dephasing due to concomitant gradients is negligible. The approach, known as Maxwell compensation, has been successful at conventional clinical MRI hardware, however, the recent uptake of portable ultra-low-field MRI systems and ultra-strong gradient systems have more pronounced concomitant gradient effects, and the extended breadth of MRI hardware poses a challenge to the fundamental assumptions of current gradient waveform design.

The primary challenge is rooted in the fact that current mitigation of concomitant gradient effects assume that the compensation only needs to consider the leading order effects which are approximately proportional to the square of the gradient amplitude and the inverse of the magnetic field strength (g2/B0). Therefore, the purpose of this work is to (i) demonstrate that corrections for leading order effects are not sufficient when considering a wide range of MRI hardware configurations, (ii) develop a novel gradient waveform design that encompasses arbitrary order corrections (gp/B0p-1), and (iii) implement this design in an open-source numerical optimization framework. To achieve higher-order Maxwell compensation, the Maxwell index was extended to account for an arbitrary order of concomitant gradient terms for any given magnetic field strength and spin species. The resulting gradient waveforms were compared through numerical simulations to both lower-order Maxwell compensated waveforms and symmetric waveforms, in terms of signal accuracy and encoding efficiency for different b-tensor shapes across a wide range of MRI system configurations.

The proposed theory of higher order Maxwell-compensated gradient waveform design enables asymmetric diffusion encoding with negligible effects of higher order concomitant gradient fields. We have through simulations demonstrated that this is particularly important in ultra-low-field MRI or when using high-performance gradient systems. Additionally, the proposed waveform design entails only a small cost in terms of encoding efficiency. (Less)
Popular Abstract (Swedish)
Magnetresonanstomografi (MRI från engelskans, magnetic resonance imaging) är en bildtagningsteknik som låter oss avbilda kroppens morfologi och funktion, helt utan invasiva eller smärtsamma ingrepp. En av de funktionella metoderna är så kallat diffusionsviktad MRI (dMRI). Denna metod är känslig för slumpmässiga rörelsen hos de vattenmolekyler som finns i vävnaden. Eftersom rörelserna påverkas av sin omgivning på mikrometernivå så kan vi använda dMRI för att avbilda mikrostrukturen indirekt. Exempelvis kan dMRI användas för att studera riktningen av nervbanor i hjärnan, vilket inte är möjligt med andra bildgivande metoder. För att göra dMRI känsligt för diffusionsprocessen så ställs höga krav på magnetkamerans hårdvara—straka... (More)
Magnetresonanstomografi (MRI från engelskans, magnetic resonance imaging) är en bildtagningsteknik som låter oss avbilda kroppens morfologi och funktion, helt utan invasiva eller smärtsamma ingrepp. En av de funktionella metoderna är så kallat diffusionsviktad MRI (dMRI). Denna metod är känslig för slumpmässiga rörelsen hos de vattenmolekyler som finns i vävnaden. Eftersom rörelserna påverkas av sin omgivning på mikrometernivå så kan vi använda dMRI för att avbilda mikrostrukturen indirekt. Exempelvis kan dMRI användas för att studera riktningen av nervbanor i hjärnan, vilket inte är möjligt med andra bildgivande metoder. För att göra dMRI känsligt för diffusionsprocessen så ställs höga krav på magnetkamerans hårdvara—straka magnetfältsgradienter krävs för att mäta diffusionen med hög noggrannhet.

Ny teknisk utveckling av magnetkamerans hårdvara har nyligen förbättrat både tillgängligheten till undersökningar och kontrasten mellan olika vävnadstyper. Detta med introduktionen av portabla magnetkameror med ultra-låga magnetfält och kameror med högpresterande gradientsystem. Denna utveckling i hårdvara introducerar även utmaningar med experimentdesign relaterade till så kallade ’Maxwell-termer’ eller ’medföljande gradienter’. Då vi manipulerar magnetfältet (applicerar gradienter) för att skapa bilder samt erhålla diffusionskontrasten, så bidrar Maxwell-termerna till oönskade avvikelser. Beroende på hur experimentet genomförs kan dessa effekter vara helt osynliga eller leda till att vissa områden i bilden uppvisar felaktig signal. Därför kan Maxwell-termer ha en negativ påverkan på diagnostiska mätningar, då detta kan leda till felaktig tolkning av bilderna.

I detta arbete demonstrerar vi utbredningen av felen som orsakas av Maxwell-termer samt kartlägger deras relation till MR-hårdvaran. Vi fastställer att nuvarande metoder inte är tillräckliga för att säkerställa noggranna mätningar, utan kan orsaka total signalförlust. Vi presenterar en lösning på detta problem som etablerar en ny experimentdesign som är okänslig för Maxwell-termer, och anpassad för den ökande räckvidden för nya typer av MRI-hårdvara. Denna lösning sker genom numerisk optimering av experimentdesignen för att garantera ackurat mätning. (Less)
Please use this url to cite or link to this publication:
author
Olsson, Viktor
supervisor
organization
course
MSFT02 20252
year
type
H2 - Master's Degree (Two Years)
subject
language
English
id
9207783
date added to LUP
2025-07-02 10:05:31
date last changed
2025-07-02 10:05:31
@misc{9207783,
  abstract     = {{Diffusion magnetic resonance imaging (MRI) is a technique that utilizes the random thermal motion of water molecules to probe the microstructure of the tissue. Because it has excellent soft tissue contrast and high sensitivity to changes in the microstructure, it has established itself as a powerful tool to detect and quantify both normal development and pathological processes. Diffusion encoding is commonly achieved using an identical pair of pulsed gradients placed on either side of the refocusing pulse. Although this sequence design is not time-efficient, it exploits a symmetry that implicitly ensures negligible concomitant gradient effects and thus has negligible effect on signal accuracy. This experimental design is by far the dominant method for conventional diffusion-weighted imaging. To further boost the encoding efficiency and signal-to-noise-ratio, an alternative is to employ asymmetric gradient waveforms. The trade-off is that asymmetric waveforms do not have an implicit control for concomitant gradient effects and are therefore uniquely susceptible to unwanted spin dephasing and erroneous loss of signal. This problem has been addressed through numerical optimization of the gradient waveform wherein the so-called ‘Maxwell index’ is constrained to ensure that dephasing due to concomitant gradients is negligible. The approach, known as Maxwell compensation, has been successful at conventional clinical MRI hardware, however, the recent uptake of portable ultra-low-field MRI systems and ultra-strong gradient systems have more pronounced concomitant gradient effects, and the extended breadth of MRI hardware poses a challenge to the fundamental assumptions of current gradient waveform design. 

The primary challenge is rooted in the fact that current mitigation of concomitant gradient effects assume that the compensation only needs to consider the leading order effects which are approximately proportional to the square of the gradient amplitude and the inverse of the magnetic field strength (g2/B0). Therefore, the purpose of this work is to (i) demonstrate that corrections for leading order effects are not sufficient when considering a wide range of MRI hardware configurations, (ii) develop a novel gradient waveform design that encompasses arbitrary order corrections (gp/B0p-1), and (iii) implement this design in an open-source numerical optimization framework. To achieve higher-order Maxwell compensation, the Maxwell index was extended to account for an arbitrary order of concomitant gradient terms for any given magnetic field strength and spin species. The resulting gradient waveforms were compared through numerical simulations to both lower-order Maxwell compensated waveforms and symmetric waveforms, in terms of signal accuracy and encoding efficiency for different b-tensor shapes across a wide range of MRI system configurations.

The proposed theory of higher order Maxwell-compensated gradient waveform design enables asymmetric diffusion encoding with negligible effects of higher order concomitant gradient fields. We have through simulations demonstrated that this is particularly important in ultra-low-field MRI or when using high-performance gradient systems. Additionally, the proposed waveform design entails only a small cost in terms of encoding efficiency.}},
  author       = {{Olsson, Viktor}},
  language     = {{eng}},
  note         = {{Student Paper}},
  title        = {{Maxwell term correction at low magnetic fields and strong gradients}},
  year         = {{2025}},
}