Understanding Phase Contrast MR Angiography A Practical

Approach with MATLAB examples

Visit the link below to download the full version of this book:

https://medipdf.com/product/understanding-phase-contrast-mr-angiography-a-practi
cal-approach-with-matlab-examples/

Click Download Now

More information about this series at http://www.springer.com/series/10059
Joseph Suresh Paul Subha Gouri Raveendran
•

Understanding Phase
Contrast MR Angiography
A Practical Approach with MATLAB
Examples

123
Joseph Suresh Paul Subha Gouri Raveendran
Medical Image Computing Medical Image Computing
and Signal Processing Group and Signal Processing Group
Indian Institute of Information Technology Indian Institute of Information Technology
and Management-Kerala and Management-Kerala
Trivandrum Trivandrum
India India

Additional material to this book can be downloaded from http://extras.springer.com.

ISSN 2191-8112 ISSN 2191-8120 (electronic)

SpringerBriefs in Electrical and Computer Engineering
ISBN 978-3-319-25481-4 ISBN 978-3-319-25483-8 (eBook)
DOI 10.1007/978-3-319-25483-8

Library of Congress Control Number: 2015957051

© The Author(s) 2016

This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part
of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations,
recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission
or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar
methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publication does not imply, even in the absence of a specific statement, that such names are exempt from
the relevant protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this
book are believed to be true and accurate at the date of publication. Neither the publisher nor the
authors or the editors give a warranty, express or implied, with respect to the material contained herein or
for any errors or omissions that may have been made.

Printed on acid-free paper

This Springer imprint is published by SpringerNature

The registered company is Springer International Publishing AG Switzerland
Preface

This book discusses the basic concepts of MRI leading to PC-MRA. An intuitive
understanding of PC-MRA concepts is provided through simulation techniques
using the extended form of Bloch equation. For completeness, quantitative flow
measuring techniques and post-processing using statistical models are also included
as separate chapters. Implementation details of important techniques discussed in
this book are provided in the form of MATLAB codes.
PC-MRA is one of the non-contrast MRA techniques using the idea that blood
flow velocities can be encoded by phase. This was first developed by Paul R. Moran
in the early 1980s. Moran analyzed the phase effects on stationary and moving spins
subjected to a pair of bipolar gradients. A stationary spin subjected to such a
gradient pair will experience no net phase shift, but a moving spin will have a net
phase shift proportional to its velocity. Two spins flowing at the same speed but in
opposite directions will have equal but opposite phase shifts, and by measuring
changes in phase, the velocity can be computed.
PC-MRA is based on use of bipolar gradients that create phase shifts of moving
spins proportional to their velocities. The key applications include flow measure-
ments, Cine CSF flow studies, and venography. The degree of sensitivity to slow or
fast flows is determined by the amplitude, duration, and spacing of bipolar gradi-
ents, which is controlled by parameter VENC—velocity encoding. Simulation
experiments outlined in this book provide a sound understanding of the encoding
strategy and enable the reader to apply this knowledge in acquisition and
post-processing methods.

Acknowledgments

Some sections of this book are based on previous articles: “Computer Simulation of
Magnetic Resonance Angiography Imaging: Model Description and Validation,”
Plos one 9. (2014) and “In Silico Modeling of Magnetic Resonance Flow Imaging in
Complex Vascular Networks,” IEEE transactions on medical imaging, 33: 11 (2014).

v
Contents

1 Introduction to MR Imaging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1 Magnetic Resonance Imaging . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 MRI Physics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.2.1 Spin Physics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.2.2 RF Excitation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.2.3 Relaxation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.3 Signal Generation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
1.3.1 Spatial Encoding of MR Signal . . . . . . . . . . . . . . . . . . . . 9
1.3.2 2D Imaging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
1.3.3 Small Tip Angle Approximation . . . . . . . . . . . . . . . . . . . 13
1.4 Phase Contrast Imaging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2 Simulation Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
2.1 Bloch Equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
2.1.1 Solution of Bloch Equation. . . . . . . . . . . . . . . . . . . . . . . 21
2.1.2 Time Update Form of Bloch Equation . . . . . . . . . . . . . . . 23
2.2 Working Principle of MR Simulator. . . . . . . . . . . . . . . . . . . . . . 24
2.2.1 Imaging Parameters and K-Space Generation . . . . . . . . . . 26
2.3 Incorporation of T2* Effects in Gradient-Echo Imaging. . . . . . . . . 28
2.4 Incorporation of Susceptibility Effects . . . . . . . . . . . . . . . . . . . . 28
2.4.1 Susceptibility Artifacts . . . . . . . . . . . . . . . . . . . . . . . . . . 29
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
3 Working Principle of PC-MRA with MATLAB Examples . . . . . . . . 33
3.1 Gradient Echo Imaging. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
3.2 Velocity Encoding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
3.3 Effects of Flow on the Image . . . . . . . . . . . . . . . . . . . . . . . . . . 39
3.4 Phase Contrast Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40

vii
viii Contents

3.5 Quantitative Flow Image Analysis . . . . . . . . . . . . . . . . . . . . . . . 42

3.5.1 Two-Point Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
3.5.2 Simple Four Point Method . . . . . . . . . . . . . . . . . . . . . . . 43
3.5.3 Balanced Four Point Method . . . . . . . . . . . . . . . . . . . . . 44
3.5.4 Processing of Multi-channel PC-MRA . . . . . . . . . . . . . . . 45
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
4 Numerical Simulation of PC-MRA . . . . . . . . . . . . . . . . . . . . . . . . . 49
4.1 Flow Phantom Model. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
4.1.1 Masking Function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
4.2 Simulation of Magnetization Transport . . . . . . . . . . . . . . . . . . . . 54
4.2.1 Lattice Boltzmann Method (LBM). . . . . . . . . . . . . . . . . . 55
4.3 Simulation of MRI Signal Generation Using LBM
and Bloch Equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
4.3.1 Integration of LBM and Blochequation Simulation . . . . . . 59
4.4 MRA Simulation Using Particle Trajectory Models . . . . . . . . . . . 61
4.5 Bloch Flow Equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
5 Modeling of PC-MRA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
5.1 An Overview of PC-MRA Modeling . . . . . . . . . . . . . . . . . . . . . 71
5.1.1 Partial Volume Effect. . . . . . . . . . . . . . . . . . . . . . . . . . . 72
5.2 Global Segmentation of Speed Images . . . . . . . . . . . . . . . . . . . . 75
5.3 Initial Estimation of Mixture Parameters . . . . . . . . . . . . . . . . . . . 80
5.3.1 Iterated EM Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . 81
5.3.2 Segmentation Using Local Phase Coherence . . . . . . . . . . . 83
5.3.3 Segmentation Using MRF Formulation . . . . . . . . . . . . . . 83
5.4 Vascular Tree Construction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
5.4.1 Skeletonization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88

Appendix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
Chapter 1
Introduction to MR Imaging

Abstract This chapter provides a quick introduction to the basics of MRI,

adequate for understanding of the terminology and concepts used in Phase-Contrast
MR Angiography (PC-MRA). Chapter begins with a description of signal gener-
ation, followed by an explanation for the need of spatial encoding necessary for
derivation of the Fourier imaging approximation. A brief introduction to the
magnetization dynamics using Bloch equation, with extension to small tip angle
approximation is also provided. The chapter concludes with an introduction to the
basic principle of flow measurement using PC-MRA, and the need for additional
flow encoding gradients.

Keywords Spin physics
RF excitation
Bloch equation
Spatial encoding

PC-MRA

1.1 Magnetic Resonance Imaging

Magnetic Resonance Imaging (MRI) is a non-invasive method of mapping the

tissue in terms of its spatial density of protons. Since the proton density varies with
the type of tissue, imaging contrast is achieved to differentiate tissue types based on
variations in their proton density. The imaging contrast can be easily maneuvered
by utilizing the dependence of received signal intensity on the physical parameters
of proton density, longitudinal relaxation time (T1), and transverse relaxation time
(T2). For example, chemical and structural changes of tumors over time directly
affect the signal intensity on MR images, providing information about their age.
Unlike many other medical imaging modalities, the contrast in an MR image is
strongly dependent upon the way the image is acquired. By adding Radio
Frequency (RF), or gradient pulses, and by careful choice of timings, it is possible
to highlight different components in the object being imaged.
The MR Image is constructed by placing the patient inside a large magnet, which
induces a relatively strong external magnetic field (B0). This causes the nuclei of

© The Author(s) 2016 1

J. Suresh Paul and S. Gouri Raveendran, Understanding Phase Contrast
MR Angiography, SpringerBriefs in Electrical and Computer Engineering,
DOI 10.1007/978-3-319-25483-8_1
2 1 Introduction to MR Imaging

many atoms in the body, including Hydrogen, to align them with the magnetic field.
With application of RF signal, energy is released from the area being imaged, in the
form of a time-varying voltage. The detector demodulates the RF components, and
stores the discrete samples in a Fourier space, known as k-space. The MR image is
then obtained by performing an inverse Fourier transform on the acquired k-space.
Using slice-selection gradients, the imaging plane can be optimized for the
anatomic area being studied. Flow-sensitive pulse sequences and MR Angiography
(MRA) yield data about blood flow, as well as displaying the vascular anatomy. In
conventional MR imaging, the spins are considered to be stationary throughout the
imaging process. However, this does not hold true in the case of flow and vascular
imaging. This is particularly applicable to the situation of imaging a slice containing
blood vessels. A critical problem with flow imaging is that the excited spins in a
vessel can flow out of the slice by the time readout is performed. Since the
unexcited spins have flown in during readout, the image formed will not have
contributions from MR signals originated from the vessel. Consequently, mea-
surements of flow would require some form of spatial encoding that is flow sen-
sitive. This is done by applying a magnetic field gradient along the direction in
which flow is to be measured. A large enough gradient amplitude can dephase the
stationary as well as moving spins depending on their position along the gradient.
The gradient when reversed, will completely rephase only the stationary spins.
Spins that have moved will not be completely rephased. For uniform flow, the
phase difference and flow are directly related to the time delay between the forward
and reverse gradients.
Spins that are moving in the same direction as a magnetic field gradient develop
a phase shift that is proportional to the spin velocity. This is the basis of
Phase-Contrast MRA (PC-MRA). In PC-MRA pulse sequence, the spin velocities
are encoded using bipolar gradients (two gradients with equal magnitude but
opposite direction). Following the gradient application, stationary spins do not
undergo a net change in phase. Due to varying spatial position, the moving spins
will experience a different magnitude of the second gradient compared to the first.
As a result, the moving spins attain a net phase shift during readout. The resultant
phase information can be used directly to determine the spin velocity.
Understanding the key physical concepts in MRI is a prerequisite for analyzing
the theory and post processing methods used in PC-MRA. The successive chapters
in this book are organized to orient the reader towards systematically building up
the knowledge base needed to understand the technical aspects through a series of
methods which outline the interface between flow and image formation process.
This chapter provides the theoretical prelude to understand the key MRI concepts
from a technical perspective. Chapter 2 highlights the fundamental approach to
generate synthetic MR images from data consisting of relevant physical parameters,
and geometry using known MR sequences. Chapter 3 describes the theory of
PC-MRA together with details of methods used to derive flow information from
raw data. Chapter 4 enables the reader to extend the image generation ideas
1.1 Magnetic Resonance Imaging 3

presented in Chap. 2 for application to PC-MRA image generation. Chapter 5

summarizes the statistical methods for post-processing the flow information derived
from PC-MRA. Links are provided in each chapter, for access to relevant
MATLAB codes for simulation, pre-processing, PC-MRA image generation, and
statistical analysis.

1.2 MRI Physics

1.2.1 Spin Physics

MRI uses large magnetic fields to magnetize the proton spins of human tissue. By
changing the RF pulses and gradients, different contrasts can be obtained which
enhance different aspects of the MR image. The way in which the RF pulses and
gradients are altered to create an image is called a sequence.
All protons are spinning, and therefore, creates a tiny magnetic field, referred to
as a “magnetic moment”. The Pauli’s exclusion principle [1] demands that no two
subatomic particles in the same atom exist simultaneously in the same state. If the
nucleus has an even number of protons, each proton will be accompanied by
another of exactly the opposite spin, and the two magnetic moments will cancel
each other.
Nuclear particles do not act as classical particles. A classical particle spinning in
a magnetic field will, according to Maxwell’s equation, radiate electromagnetic
energy. These nuclear particles, even though they possess a magnetic moment
aligned with the external field, do not emit, until stimulated by an RF pulse
described next. Thus, only if a nucleus has an odd number of protons, will it possess
a net magnetic moment. Placing this nucleus in a strong, constant magnetic field
will cause it to tend to align with the field.
Understanding that we are only using an analogy, we will continue to refer to the
quantum-mechanical property [2, 3] as “spin,” and say that the nucleus now spins in
such a way that the magnetic moment aligns with the external field. For simplifi-
cation, the spins are often looked upon as spin packets which represent all the spins
in a certain volume.
When there is magnetic dipole (or a current loop, constituted by orbiting elec-
tron) in applied magnetic field, there are various things which can happen: the
magnetic dipole may align in same direction, opposite direction or make some
angle with the field. When it makes some angle, it may draw some energy from the
field and precess about the applied field with some precession frequency. Now spin
angular momentum of a particle also acts as a dipole which can precess indepen-
dently. Usually Larmor precession term is reserved for dipole moment. Now we
must consider the rate of spin. Once the external field has been applied, the spins
precess with Larmor frequency
4 1 Introduction to MR Imaging

x0 ¼ cB0 ð1:1Þ

Under the presence of magnetic field along z direction, the spin packets precess
at a single frequency while they are not in phase. The resultant net magnetic field is
pointed in z direction. Net magnetization M is the vector sum of nuclear magnetic
moments μi produced by the spin packets. This is expressed as

X
Ns
M¼ ln ð1:2Þ
n¼1

1.2.2 RF Excitation

The signals generated due to static magnetic field contain frequencies in the RF
range, and are not useful for derivation of tissue-related information. Hence the
received signals are first amplitude demodulated to extract the desired information.
Considering the mathematical description of signals, this is equivalent to referring
the spin dynamics in a rotating coordinate system, rotating at the RF Larmor
frequency [4, 5]. The signal representation in the rotating frame of reference will
contain only the intermediate frequencies in the kHz range. Consider the coordinate
system for observation of the magnetization to be rotating at a frequency ω0. The
coordinate system is rotated about the z-axis in the same direction that the local
magnetization M rotates about B1. The “laboratory” frame of reference is the usual
frame of reference with coordinates (x, y, z). The “rotating” frame of reference has
coordinates (x′, y′, z′). For a rotating frame, the coordinate axes are transformed
using

i0 ¼ i cosðx0 tÞ  j sinðx0 tÞ
j0 ¼ i cosðx0 tÞ þ j sinðx0 tÞ ð1:3Þ
0
k ¼k

Consider an RF field B1(perpendicular to static field) applied for a short duration,

to rotate the spin away from the initial B0 alignment. The RF field is produced by a
coil system separate from the static field source, called transmit field. When B = B0
k, the field (Beff) in the rotating frame will be effectively zero. Such rotation leaves
the classical moment precessing at an angle around the original static field. The x-y
components of the magnetization in rotating frame is given by

Mxy:rot ðtÞ ¼ Mxy ðtÞ expðix0 tÞ ð1:4Þ

With the magnetization vector referenced in the static frame represented by

M = [Mx, My, Mz] and that in the rotating frame by Mrot = [Mx,rot, My,rot, Mz,rot],
1.2 MRI Physics 5

Mxy
rot ðtÞ ¼ Mx;rot þ iMy;rot

ð1:5Þ
¼ Mxy expðix0 tÞ

The longitudinal magnetization remains same in the rotating frame. For an RF

field B1 with frequency ω1 applied along the x-direction,

B1 ¼ B1 expðjx1 tÞ ð1:6Þ

Since ω1 = ω0 at resonance, the total applied B-field can be represented as

2 3
B1 cosðx0 tÞ
B ¼ BðtÞ ¼ 4 B1 sinðx0 tÞ 5 ð1:7Þ
B0

In the rotating frame, the rate of change of magnetization is

dMrot
¼ Mrot  cB1eff ð1:8Þ
dt

where B1eff ¼ B1 i0 is the effective magnetic field in the rotating frame. Substituting
for B1eff, the matrix version of (1.8) will be
2 3
0 0 0
dMrot
¼ Mrot 4 0 0 cB1 5 ð1:9Þ
dt
0 cB1 0

Using the initial condition Mrot(0) = m0k, solution of (1.8) yields

2 3
0
Mrot ðtÞ ¼ 4 m0 sinðcB1 tÞ 5 ð1:10Þ
m0 cosðcB1 tÞ

Since B1eff is applied along the x0 axis, Mrot will precess around x0 in the z-y0
plane. In the rotating frame, the trajectory of the magnetization vector at any given
time point during excitation is shown in Fig. 1.1.
Net magnetization M can be flipped to an angle α by applying magnetic field B1
for a time period T such that

a ¼ cB1 T: ð1:11Þ

If the magnetization is to be flipped into the transverse (x-y) plane, the B1 field is
applied for a period of time and then removed. For a flip angle of α = π/2,
6 1 Introduction to MR Imaging

Fig. 1.1 The trajectory of

M in the presence of B0 and
B1 in the reference frame

p
T¼ : ð1:12Þ
2cB1

As noted from (1.8) the Bloch equation governs the precession of a nuclear
magnetic moment M(x) = ((Mx(x), My(y), Mz(z))T about a time varying external
magnetic field ω(t) = (ωx(t), (ωy(t), (ωz(t))T. In the presence of a gradient field,
Bloch equation relates gradient fields and RF magnetic fields to the time derivative
of magnetization. In a frame rotating at Larmor frequency, the Bloch equation is
2 3
0 GðtÞ
r B1;y ðtÞ
dM ðr; tÞ
¼ c4 GðtÞ
r 0 B1;x ðtÞ 5Mðr; tÞ ð1:13Þ
dt B ðt Þ B1;x ðtÞ 0
1;y

where γ is the gyromagnetic ratio, gradient fields G(t) = ((Gx(t), Gy(t), Gz(t))T, and
the RF magnetic field B1 in (1.6) becomes a function of t with components
B1,x(t) and B1,y(t) respectively. Here, the RF fields are assumed to be spatially
uniform and (1.13) can be used to determine the magnetization state following RF
excitation provided, an initial magnetization state, RF pulse and gradient wave-
forms are given.

1.2.3 Relaxation

The transverse component of the precessing magnetization Mxy can be detected by

placing coils in the x-y plane. The rotating Mxy induces electromotive force in the
coil. The magnetization vector (M) continuously precesses in the x-y plane under
the absence of relaxation effects. But the relaxation mechanisms will cause
dephasing of transverse magnetization, and a build-up of the longitudinal
1.2 MRI Physics 7

magnetization. The resulting process is called Free Induction Decay (FID). The
relaxation process is divided into two independent processes which happens
simultaneously. T1 relaxation describes change in magnetisation vector in
z-direction and T2 relaxation describes changes in the x-y plane [5].
T1 relaxation applies to protons in the volume following the 90° excitation pulse.
During relaxation process, the spin will transfer the energy which are gained from
RF pulse to the environment. Larger molecules exhibit slower relaxation and vice
versa. T1 values are determined from the relation between Larmor frequency and
natural frequency of the molecules. The values are small when both frequencies are
closer. The process of T1 relaxation is mathematically represented using exponential
growth, described using
 
Mz ðtÞ ¼ M0 1  et=T1 ð1:14Þ

The second relaxation process, the transverse relaxation, has its origin in the
interaction of the spins with each other and the magnetic properties of their envi-
ronment. The magnetic moments of neighboring spins trigger magnetic field fluc-
tuations. Through these field inhomogeneities, the Larmor frequencies of spins tend
to differ by small amounts, so that the spins lose their phase coherence. The
dephasing of spins results in decrease of the transversal magnetization. The duration
of this process is described by the time constant T2, following which the transverse
magnetization Mxy(t) has decreased to about 37 % of its initial value. The rate of
dephasing is different for each tissue. Fat tissue will dephase quickly, while water
will dephase much slower. The transverse relaxation is mathematically represented
using

Mxy ¼ Mxy ð0Þet=T2 ð1:15Þ

In reality, the dephasing is even more pronounced due to external field inho-
mogeneities, leading to the effective transversal relaxation time T*2 (T*2 < T2).
Mathematically, this effective relaxation rate is given by

1 1 1
 ¼ þ 0 ð1:16Þ
T2 T2 T2

where T2 is related to non-reversible spin-spin relaxation and T20 refers to the

reversible relaxation due to static field inhomogeneities. The two constants T1 and
T2 are tissue-specific, where usually T1 > T2. These relaxation times can be used in
MR imaging to generate different tissue contrasts in the images.
8 1 Introduction to MR Imaging

1.3 Signal Generation

The received signal s(t) is composed of contributions from all spins possessing
transverse magnetization in the imaged volume. This is mathematically expressed
using
Z Z Z Z
sðt Þ ¼ Mxy ðr; tÞdV ¼ Mxy ðx; y; z; tÞdxdydz ð1:17Þ
vol z y x

In the presence of transverse relaxation, the received signal equation becomes

Z
sðtÞ ¼ Mxy ðx; y; z; tÞet=T2 ðx;y;zÞ dV ð1:18Þ
vol

In any MR imaging sequence, the maximum signal intensity is obtained at

t = TE (echo time). This is achieved by signal refocusing using either spin or
gradient echo formation (see Chap. 2 for further information). Either forms of
refocusing is performed in the presence of external gradient fields, described in later
sections. The signal received in the absence of refocusing is referred to as free
induction signal. In quadrature reception, the FID signals are acquired in the in-
phase (I) and quadrature (Q) coils upon application of a 90° RF pulse in the
phase-encode direction as shown in Fig. 1.2.
Imaging application often requires selection of a particular slice in which the
applied gradient fields cause changes in spin precession. Slice selection is
accomplished using combination of an RF excitation pulse and a slice selection
gradient pulse. For sagittal and coronal sections, the slices are selected in the x and
y directions respectively. By convention, the laboratory frames in the scanner are
set with the x-direction from Left ear-to-Right ear (LR), and the y-direction from
Anterior-to-Posterior (AP). Axial scanning requires slice selection gradients to be

Fig. 1.2 After RF-pulse application, the magnetization M is tilted by the flip angle α from the
equilibrium along B0 direction and precesses with the Larmor frequency about the z-axis, parallel
to B0. Mz is the longitudinal and Mxy is the transverse component of magnetization
1.3 Signal Generation 9

applied in the z-direction, which is same as direction of the main field B0. Assuming
that a selective slice excitation is used for axial scan, the dependence of Mxy on z
can be eliminated. Since the net magnetization Mxy(r, 0) at t = 0, is directly pro-
portional to the spin density ρ(r), the signal contribution from a voxel at position r
can be represented as
 
sðx; yÞ ¼ qðx; yÞf ðvÞ 1  eTR=T1 eTE=T2 ð1:19Þ

where f(v) represents signal modulation arising from fluid flow, and TR and TE
denote the repetition time and echo time respectively for any MR sequence. In
(1.19), the time independence of the local signal (on the LHS), as opposed to (1.18),
is obtained by choosing the maximum signal at the refocusing time. Since TR and TE
are image sequencing parameters, the above relationship illustrates how their values
can be manipulated to provide different types of imaging contrast. In the absence of
external field gradients, the time-independent form of received signal is the sum total
of s(x, y) from all locations in the imaging Field Of View (FOV). In the
time-dependent form, this is same as s(t) in (1.18). The imaging problem consists of
inverting the signal equation for computing the proton density as a function of
location in the imaging plane. Even with discrete samples of signals at different time
points, it is impossible to invert (1.18) for computation of ρ(r). It is for the sake of
enabling the inversion, that the locally generated MR signals are spatially encoded
using externally applied magnetic field gradients. The mathematical nature of
encoding signals, and their role in obtaining a Fourier approximation [6–8] for
imaging, are discussed further in the succeeding sections.

1.3.1 Spatial Encoding of MR Signal

By varying the magnetic field spatially, the spins at each location will experience
different magnetic field strengths and precess with different Larmor frequencies.
This is achieved by adding magnetic gradients to the B1 field. Addition of a positive
gradient (Fig. 1.3) in the z-direction results in increasing precession frequency
with z. When applying an RF-pulse, only spins in a slice with matching precession
frequency will flip down. This is called slice selection, and it is possible to vary the
thickness of a slice by changing the bandwidth of the RF-pulse [5, 8].
Consider a matrix of spins (spin packets) representing a slice. Following
application of an RF pulse, the received signal will be
ZZ
sðt Þ ¼ qðx; yÞ expðt=T2 Þ expðix0 tÞdxdy ð1:20Þ