MAGNETIC RESONANCE

IMAGING: The Basics

Dr Fatma Elawamy, Ph.D.

Department of Radiology
University Of Benghazi
 Outline
• Basics of Magnetism
• Generating MR
signals
(in 5 easy steps!)
• Hardware overview
• MR image
construction
• MR vs. CT
• Clinical examples
• Research topics
 Q: What is the origin of magnetism?
A: Moving charge (usually electrons)

Orbital Motion Current in a wire Spinning Charge

(for paired electrons, field cancels)

Magnetic phenomena were first observed ~1000 BC with the discovery of

lodestone (magnetite, Fe3O4) near the town of Magnesia in Asia Minor.
Lodestones attract iron particles and, when suspended from a string, tend to
align with earth’s magnetic field. This led to the invention of the compass.
 Q: What is magnetic susceptibility?
A: The degree to which a material becomes
magnetized when placed in a magnetic field

3 important
classes of
susceptibility
are:

Effect on external Strongly Weakly Weakly

magnetic field: enhanced enhanced opposed
 Magnetic Susceptibility
Diamagnetic: No unpaired electrons
Weak negative susceptibility
E.g. most tissue & organic materials, water

Paramagnetic: At least one unpaired electron

Weak positive susceptibility
E.g. Gd contrast agents, deoxyhemoglobin,
methemoglobin, molecular O2

Ferromagnetic: Cooperative alignment of magnetic moments

of neighboring atoms (property of bulk matter)
Strong positive susceptibility
E.g. alloys of iron, nickel, cobalt
 How is a Permanent Magnet Created?
An unmagnetized ferromagnetic sample
has multiple regions (domains) of
spontaneously aligned spins
Unmagnetized
Because they are randomly oriented, the
net magnetization cancels

When exposed to a strong external

magnetic field, domains will align
cooperatively, creating remanent
magnetization which persists after the
field is removed Hysteresis Curve

Permanent magnets have essentially all

domains in alignment, and exhibit
hysteresis when acted upon by an
Magnetized
external field
 What is Superparamagnetism?
Superparamagnetism is more closely related
to ferromagnetism than to paramagnetism
Superparamagnetism occurs in small single-
domain magnetic particles Single-domain particle

When exposed to a strong external magnetic field, the magnetic

domain will align with the field, but the particle does not retain any
net magnetic ‘memory’ after the field is removed
Examples are iron oxide contrast agents, ferritin and hemosiderin
Susceptibility Summary
 Basic Laws of Magnetism
• Magnetic fields exist as dipoles; in
contrast to electric charges, magnetic
monopoles do not exist (breaking a bar
magnet in half creates 2 bar magnets)
• Like poles repel; unlike poles attract
• The 2 poles of a magnet are always of
equal strength but opposite sign
Magnetic lines of force
• Magnetic field strength follows an inverse
square dependence on distance
• Permanent magnets can be created from
ferromagnetic objects by induction

Polarity of a permanent magnet

depends on polarity of external field
 Units of Magnetic Strength
The symbol for magnetic strength (flux density) is B
The SI unit of B is the tesla (T)
The CGS unit of B is the gauss (G)
1 T = 10,000 G

The earth’s magnetic field varies geographically, but

averages ~ 0.5 G or 0.05 mT
The axis of the earth’s magnetic field is tilted ~11°
from the axis of rotation
The earth’s principal dipolar field has been declining
~7% per century
 Electromagnetism
Historical Highlights:
1820: Oersted observes that an
electrical current deflects a compass
needle

DC generates a magnetic field

1831: Faraday discovers that a

changing magnetic field can generate
an electric current

Oscillating field creates AC

 Electromagnetism
Historical Highlights:
1834: Lenz predicts the direction of current induced by a
changing magnetic field: The current in the wire is such
that the induced magnetic field opposes the motion of the
external magnet
 Electronic vs. Nuclear Magnetism

Electron magnetic moment -

Proton magnetic Neutron magnetic

moment +
moment

Spinning of the subatomic particles generates a

magnetic field, called a “magnetic moment” or
“magnetic dipole”
 Electronic vs. Nuclear Magnetism
Electronic Nuclear
The magnetic moment of the A net magnetic moment exists
electron is 658 times larger than for atoms with odd numbers of
that of the proton. Therefore, protons, neutrons, or both
electronic magnetism, when
When nucleons are paired,
present, tends to dominate
nuclear magnetism vanishes
magnetic behavior
Nuclear magnetism is the
Electronic magnetism is
basis of NMR spectroscopy
responsible for susceptibility
and MRI
effects (e.g. paramagnetism,
ferromagnetism), on which most
MR contrast agents are based
Electronic paramagnetism is the
basis of electron paramagnetic
resonance (EPR) spectroscopy
 QUIZ
Why does a neutron have a magnetic
moment?
Nucleons are
composed of charged “quarks” +
-
Proton + +
2 Up quarks (+2/3 esu) and 1 Down quark (-1/3 esu)
= +1 esu net charge

-
Neutron -
2 Down quarks (-1/3 esu) and 1 Up quark +
(+2/3 esu) = no net charge
 MRI Timeline
1946 MR phenomenon - Bloch & Purcell
1952 Nobel Prize - Bloch & Purcell
1950-1970 NMR developed as analytical tool
1971 T1 of tumors elevated above normal - Damadian
1972 Computed Tomography - Hounsfield
1973 Backprojection MRI - Lauterbur
1975 Fourier Imaging - Ernst
1977 Echo-planar imaging - Mansfield
1980 FT MRI demonstrated - Edelstein
1986 Gradient Echo Imaging
1986 NMR Microscope
1987 MR Angiography - Dumoulin
1991 Nobel Prize - Ernst
1992 BOLD effect (Functional MRI) - Ogawa
1994 Hyperpolarized 129Xe Imaging
2003 Nobel Prize - Lauterbur & Mansfield
 2003 Nobel Prize Controversy
In 2003, Raymond Damadian was passed over by the Nobel
Committee in its award for MRI development

One of several full-page ads published by Damadian in major

newspapers across the US to protest the Nobel decision
 The First MRI Units

1977: Damadian’s apparatus 1979: Aberdeen unit

 Recipe for MR Signal
1. Sample having magnetic nuclei (an odd
# of protons and/or an odd # of
neutrons) e.g. H-1
P-31
Na-23
C-13
F-19
etc.

Hydrogen nuclear spins

in random orientations
 Recipe for MR Signal

2. Static magnetic field B0

(preferably strong) to induce
alignment of nuclear spins

Stronger B0 field produces

stronger MR signals
 Recipe for MR Signal

3. B0 field also induces

spin precession

Precession Frequency
(Larmor Frequency)
f = (/2) B0 Top in Nuclear
gravitational moment in
field magnetic
field
(for protons, /2 = 42.576 MHz/tesla)
 Recipe for MR Signal
4. Apply EM energy at
Larmor frequency (on
resonance) to induce
energy absorption by
the nuclear moments. Plucking the violin string induces
the same note on the harp to
This energy lies in the radio- sound. This is resonance.
frequency (RF) band of the
electromagnetic spectrum.
 Recipe for MR Signal

5. Immediately after
excitation, the sample
briefly emits a MR
signal (also a radio-
wave)
 Hardware Overview
• Main magnet (B0 field)
• Gradient coils
• Radiofrequency (RF) coils
• Faraday cage
 Magnets: 3 Types
• Permanent (<0.3T)

• Resistive Electromagnet (< 0.6T)

• Superconducting
Electromagnet
(high field)
 Mobile MRI
Actively-shielded superconducting
magnets are best suited for mobile MRI
 Gradient Magnetic Field Coils
• Gradients are used to
select image planes
and to determine
where in space
signals originate
• Are essential for
imaging
• Produce audible
noise in MRI
 Gradient Magnetic Fields
• Gradients “tilt” the main magnetic field linearly in
space
• The magnitude of tilt is small (< 1% of B0)
 Radiofrequency (RF) Coils
• Used to excite nuclear
moments and detect
MR signals
• Can be transmit only,
receive only, or both
transmit/receive
• Volume coils yield
uniform image intensity
• Surface coils yield high
SNR near their surface
 RF Shielding

• To prevent
unwanted noise
from corrupting
the images, the
magnet is
protected by a
Faraday cage
 Creating MR Images

• Position patient
on motorized
table
• Attach RF coil
• Advance body
part of interest to
magnet center
 Creating MR Images
• One of hundreds
of possible pulse
sequences are
run to collect raw
image data

Timing diagram of a “spin echo” MRI sequence

 Creating MR Images
• Raw image data is
accumulated in a
2D array in
computer memory

• A 2-dimensional
Fourier Transform
is done to generate
the final image
 Tissue Contrast Manipulation

Proton-density image T1 image T2 image

• MRI contrast is determined by the proton density and

relaxation times (T1 and T2) of tissues.
• By varying timing parameters of the imaging
sequence, tissue contrast can be significantly altered
 MRI vs. Computed Tomography

MRI CT
• Nonionizing radiation • Ionizing X-rays
• Direct multiplanar • Transverse plane only
imaging
 MRI vs.Computed Tomography

MRI CT
• Contrast depends on several • Contrast only related to
parameters electron density
• Soft tissues well differentiated • Poor soft tissue discrimination
 Clinical Images

Cervical spine: Liver and spleen:

Whole body MR:
Sagittal plane Transverse plane
Coronal plane
 Clinical Images

Knee
Brain astrocytoma
Cervical spine
Gadolinium Contrast Agent

Tumor only
clearly
visible with
contrast
agent
 MR: Beyond Anatomic Imaging

• Blood flow and angiography

• Cardiac dynamics
• Functional brain imaging
• In vivo spectroscopy
• Novel contrast agents
 MR Angiography
• Sequences can be designed to preferentially acquire
signal from moving spins (arterial or venous blood)
• May be unenhanced or Gd contrast-enhanced
• This technique can often replace risky X-ray angiography,
which requires injection of iodinated contrast agent

Carotids Heart & great vessels Cerebral venogram

 Coronary Artery Imaging

Both right and left coronary arteries are

visible in this heart image
 Cardiac Function
Cine

Regurgitant jets indicative of aortic insufficiency

 Functional Brain Imaging

Colored regions
denote areas of brain
“activated” in
response to a pain
stimulus to the right
hand
 MR Spectroscopy

By deliberately suppressing water, other hydrogen-containing metabolites

can be observed. In this head-injured patient, the affected hemisphere
demonstrates a lower concentration of N-acetyl aspartate (NAA).
 Spectroscopic Imaging

Constructing metabolite maps yields a visual indication of the

relative concentration of various chemicals in the brain
Ultra-small (USPIO) Contrast Agent
An investigational pulmonary contrast agent consisting of starch-coated
single USPIO crystals with diameters of 5 -7 nm has demonstrated
potential utility as a T1 agent
Images from a 3D breath-held exam:

Oblique Coronal Sagittal

Hyperpolarized Contrast Agents
It has been discovered that certain gases (e.g. 129Xe and 3He) can be
excited in the presence of rubidium vapor “pumped” by a laser to
achieve up to 25% spin polarization (parallel alignment with B0)
Pulmonary imaging using this technology is an active area of current
research

Image obtained during breath-hold Image obtained during inhalation

of 1 liter of 3He phase
 MR Horizons
• Faster (real-time) scanning
• Interventional procedures during MRI
• Further mapping of brain function
• Specialized hardware (e.g. dedicated
head or extremity imaging)
• High-temperature superconductors?
• ?
 MRI WWW Resources
A fairly comprehensive list of MR-related web
sites can be found at:
Ismrm.org
Sites are categorized by topic, e.g.
• Anatomy
• Safety
• Physics
• Artifacts
• Tutorials
• Software
• Etc.
 Outline
• Spins in a Magnetic Field
– Precession
– Frames of Reference
– Effect of RF Magnetic Field
• MR signal Generation
– FID
– Spin Echo (SE)
• Relaxation
– T2*
– T2
– T1
 Imaging Via Windows into Matter
The Electromagnetic (EM) Spectrum

• X-ray techniques employ

highly energetic (~1019 Hz)
ionizing radiation
• MR uses radiofrequency (RF)
waves (~107 Hz), i.e. 1012
times less energetic
• Because RF is nonionizing,
MR is a much safer modality
• Moreover, MR boasts
superior soft tissue contrast
than X-ray methods
• However, signal-to-noise
(SNR) is inherently low in MR
 Generating MR Signals
The Basic Requirements
• A sample containing magnetic nuclei (preferably very many of them)
• An external magnetic field B0 (preferably very strong and uniform)
• A source of radiowaves, tuned to the correct Larmor frequency (f0)
• Appropriate transmitter and receiver RF coils, also tuned to (f0)
• For imaging, a set of three orthogonal gradient coils, which are used
to deduce the origin of signals in space
 Magnetic Nuclei
The Dominance of Hydrogen
• Any atom with unpaired nucleons is
capable of generating a MR signal
• However, there are relatively few
magnetic nuclei of biological
significance
• Of these, only the hydrogen nucleus
(proton) exhibits both a relatively high
intrinsic signal and high concentration
in vivo
• Therefore, proton resonance currently
dominates both MR imaging and MR
spectroscopy

Every molecule of water has 2

magnetic nuclei
 Nuclear Magnetism: The Proton
The spin angular momentum (µ)
of a charged proton creates a
dipolar magnetic field

The magnetic dipole may be

simplified by using a single
vector representation of the    h I
magnetic moment

In the absence of an external

magnetic field, nuclear moments
are randomly
oriented; therefore the net
magnetization is zero

Random orientation of magnetic moments

Effect of External Magnetic Field
When placed in an external magnetic field (B0), nuclear
moments will
• Instantly begin precessing
• Gradually come into alignment with the B0 field vector

Precession about B0 Alignment with B0

Why Do Nuclear Moments Precess?
The cross product of the magnetization vector M with
the external magnetic field B0 creates a torque
oriented orthogonal to both vectors

The net result is a clockwise precession

of M about B0 with angular frequency ω0
 Effect of External Magnetic Field
Spin Precession
The precession or “wobble” of a nuclear magnetic moment
about the B direction is exactly analogous to that of a spinning
0

top under the influence of a gravitational field

A magnetic moment (M) precesses A top precesses in the earth’s

about B0 with angular frequency ω0 gravitational field
 Effect of External Magnetic Field
Spin Precession – Larmor Equation
The frequency of precession about B is determined by the
0

product of the external field strength and the nuclear

gyromagnetic ratio

f0  B0
2
Where f0 = Larmor frequency (MHz)
γ/2π = 42.58 MHz/T for protons
B0 = external field strength (T)

The precessional frequency is of fundamental importance because

• It is the frequency of radiowaves needed to induce MR excitation
• The precessional frequency instantly responds to changes in the
local magnetic field (e.g. gradient fields or changes in orientation
or proximity of other nearby paramagnets or magnetic nuclei)
 Effect of External Magnetic Field
Spin Alignment

Under the influence of B0, the

magnetic moments assume a
nonrandom alignment parallel or
antiparallel with the B0 vector,
creating a net magnetic moment

The rate of approach to alignment

is governed by the T1 relaxation
time (a radiationless process)
The parallel orientation is in a slightly lower (thermodynamically
favored) energy state, so that a small excess of spins acquire a
parallel orientation

This population difference is essential to producing MR signals

(Equal numbers of parallel and antiparallel spins would cancel each other out)
 Why Don’t All Spins Align With B0?
Because even in a strong B0 field, the ambient thermal energy
is ~ 1 million times higher than the energy difference between
the parallel and antiparallel spin states

The ratio Nparallel/Nantiparallel is predicted by Boltzmann to be:

N parallel
 e E kT  eB0 kT

N antiparallel
Note that since γ, h and k are constants, only B0 and T
(temperature) can be varied to improve the spin population ratio

In vivo, only a higher B0 helps; at 3T and assuming body

temperature (310K), for every million hydrogen spins aligned
antiparallel, there will be one million plus 20 aligned parallel
In “hyperpolarized” gas contrast agents ( He & Xe), optical pumping with
3 129

a laser is used to vastly increase the spin population ratio

 Effect of External Magnetic Field
Spin Alignment – Net Magnetization

Despite unfavorable spin

populations, the huge number of
spins per voxel usually leaves a
sufficient net magnetization (M) for
signal to be detected

The net M vector is oriented

exactly along B0

After several T1 time intervals, the net magnetization reaches a

maximum amplitude, called the equilibrium magnetization (M0)

Now we are in a position to manipulate the net magnetization in

order to elicit a coherent response
Effect of External Magnetic Field
Summary
To summarize, the purpose of the external field B0 is to
induce the nuclear spins to align cooperatively and thus
create an equilibrium or net magnetization M0

Note that simply placing a sample in a magnetic field does

not produce any signal

However, rotation about B0 (precession) of a coherent net

magnetization vector would emit an electromagnetic signal
(at least briefly)

Therefore, we would like to push M out of alignment with B0

somehow

The precession of the individual spin vectors comprising

the net magnetization provides a means to do exactly that
 Frames of Reference
Definition of Mz and Mxy
Before proceeding, it is convenient to establish a Cartesian
(x, y, z) coordinate system to describe nuclear magnetic
behavior

By convention, B0 is always aligned with z

Magnetization along z (Mz) = “longitudinal”
Magnetization in the x-y plane (Mxy) = “transverse”

Only Mxy produces

detectable MR signal

Definition of Longitudinal and Transverse Magnetization

 Frames of Reference
Lab Frame vs. Rotating Frame
Although a stationary laboratory frame of reference seems
simpler, it is often more convenient to adopt a frame of
reference that rotates at the Larmor frequency

Note that in the rotating frame, precession stops

Rotating frame
denoted by x’ and y’

Laboratory Frame Rotating Frame

 Effect of RF (B1) Magnetic Field
On Resonance vs. Off Resonance
Radiowaves, like all EM radiation, have a magnetic component (a
rotating B1 magnetic field)
However, in the rotating frame, B1 on resonance appears stationary

Rotating Frame

RF that satisfies the Larmor condition is efficiently absorbed by

the nuclear spins
RF of incorrect frequency (off resonance) has no effect
 Effect of RF Magnetic Field
Lab Frame vs. Rotating Frame

In the lab frame, RF on resonance In the rotating frame, M

induces a complex perturbation of the undergoes simple
net magnetization vector rotation about B1
 Effect of RF Magnetic Field
Lab Frame vs. Rotating Frame
When RF is applied, the M vector begins to precess from its
initial orientation along z toward the x-y plane

In the lab frame, RF on resonance In the rotating frame, the

induces a complex perturbation of M; motion of M is much simpler
the vector tip traces out a “beehive”
 Effect of RF Magnetic Field
Flip Angle
The flip angle is calculated as the rate of angular precession
about B1 times the duration of the RF pulse

RF is applied to the Which is normally With a certain amplitude

sample as a pulse drawn as a rectangle B1 and duration t

The rate of angular    B1 Note that B1 << B0, so

precession is relatively slow
precession about B1

Flip angle    t   B1 t
A 180° pulse, for example, might have twice the amplitude (or
twice the duration) of a 90° pulse
 EXAMPLE


If a 1 ms square pulse induces a 90° flip angle
(π/2 radians), what is the amplitude of B1?
B1 
t

 2
B1    5.86 T
 t (2.68 E8s T )(1E  3s )
1 1

Note that the B1 field is much weaker than B0

 Effect of RF Magnetic Field
Flip Angle
Although a 90° flip angle yields the largest possible signal, often
smaller flip angles are used to avoid depleting (saturating) Mz

30° pulse 75° pulse

90° pulse 180° (Inversion)

pulse
 Mxy After the RF Pulse
When the RF is turned off, the M vector immediately precesses
about B0 in the x-y plane. A radiowave signal (at the Larmor
frequency) is emitted:

In the lab frame, Mxy shrinks In the rotating frame, loss of Mxy occurs due
toward zero as it precesses to fanning out (loss of phase coherence) of
the individual spin vectors
 Transmission & Reception of RF
RF Coils (Antennae)
RF energy is introduced into the patient using a transmitter coil
(usually a volume coil) and signal is detected by a receiver coil
(either a volume or a surface coil)

B0

B1

Birdcage transmit/receive head coil Butterfly surface coil for spine imaging
 Effect of RF Magnetic Field
Summary
The purpose of the radiofrequency field B1 is to drive the
net magnetization vector away from its initial alignment
along z (the B0 field direction)

The B1 field is introduced by pulsing RF through tuned coils

(antennae) surrounding the body part of interest

Immediately after the RF is turned off, M will again precess

about B0, but now the individual spin vectors are in phase

Rotation (precession) of the in-phase Mxy magnetization

emits an electromagnetic signal at the Larmor frequency (at
least briefly)

The signal is detected by a tuned coil and amplified

 MR Signals
FID and Spin Echo
The two most commonly employed MR signals are the Free
Induction Decay (FID) and the Spin Echo (SE)
The FID is used for gradient echo imaging and the spin
echo is used for spin echo imaging

The FID appears

The Spin Echo appears
immediately following a
following 2 RF pulses:
RF pulse
90° and 180°
 MR Signals
What Other Signals Are Observed in MR?

Besides the FID and the Spin Echo, several other signal
types have been given identifying names:

Hahn Echo: The echo signal resulting from two RF pulses

of arbitrary flip angle (e.g. 30° – 30°)

Stimulated Echo: The echo signal resulting from a train of

three RF pulses

Gradient Echo: Simply an FID that has been dephased and

then rephased by an applied bipolar gradient magnetic field
 MR Signals
The FID (Free Induction Decay)
The FID is so named because it is a RF-induced signal that decays
without (free of) the B1 field remaining on

The FID decays due to T2 relaxation (spin-spin interaction) plus

additional dephasing influences (mainly B0 inhomogeneity); the
overall decay rate is dictated by T2*

Mxy = M0 e-t/T2*

After a 90° pulse, the receiver (lab

The exponential decay of the FID
frame) detects precessing Mxy as
follows the time constant T2*
an oscillating voltage
 MR Relaxation: T2
Why Does the FID Decay?
Individual spin vectors Causes of T2* Dephasing Causes of T2 Dephasing
get out of phase (i.e. Spin-spin interactions Spin-spin interactions
lose phase coherence) Magnetic field
inhomogeneities
Magnetic susceptibility

T2* is always
shorter than
Exponential decay T2
constant defined by
time to reach 37% of
initial signal
 MR Relaxation: T2
What is the Difference Between T2 and T2*?
T2 is the rate of decay of transverse magnetization (Mxy) in a
perfectly uniform external magnetic field

T2 is very sensitive to tiny changes in magnetic field produced by

neighboring magnetic atoms, hence the name spin-spin relaxation

Spin-spin interactions induce small local magnetic field variations, which alter the
precessional frequency and hence cause spins to get out of phase
 MR Relaxation: T2 *

What is the Difference Between T2 and T2*?

T2* is the observed rate of decay of transverse magnetization (Mxy) in a
real (imperfect) external magnetic field

The magnet inhomogeneity plus dephasing produced by susceptibility

effects combine to accelerate the natural T2 process

Theoretical T2 decay
curve

Actual T2* decay curve

 MR Signals
The Spin Echo
The spin echo arises following a combination of two (usually 90° and
180°) RF pulses; echoes can be refocused by additional 180° pulses

The main reason for using the SE is to eliminate the effects of B0

inhomogeneity and bulk susceptibility

FID and 3 spin echoes produced by a train of RF pulses

The echo amplitudes are determined only by the sample T2 and not by T2*
 MR Signals
How Does the Spin Echo Overcome B0 nonuniformity?
The spin echo uses the 180° pulses to “reverse time” such that
static field inhomogeneity that accelerates spin dephasing before
the inversion pulse also causes compensatory rephasing after the
inversion pulse

In the rotating frame, diverging vectors before the 180° pulse become
convergent after the pulse; the spins rephase at the echo time (TE)
 MR Signals
Spin Rephasing: The Race Track Analogy
Similar to vehicles of differing speeds, the individual spin vectors
precess at different rates
The 180° pulse is akin to a directive to “go back to the starting line”

Regardless of how far out of phase the vectors get during TE/2, unless B 0
changes they will converge at the echo time TE
The 180° pulse only cancels B0 nonuniformity; spin-spin dephasing still occurs
 MR Relaxation: T1
The Return To Equilibrium
T1 is the rate of growth or recovery of longitudinal magnetization (Mz)
following a perturbing RF pulse

T1 is usually significantly longer than T2, because efficient T1 relaxation

requires fluctuating magnetic fields at or near the Larmor frequency

Mz = M0(1 – e-t/T1)
During one T1 interval,
the signal recovers
T1 is usually longer than T2
63% of M0
 Tissue Relaxation Times
Three important tissue properties that determine MR contrast are
proton density, T1 and T2

T1 and T2 vary considerably among different tissues:

Note that T1 values increase somewhat with B0 field strength

T2, however, is generally independent of B0
 MR Relaxation
Summary
T2* is the observed rate of decay of transverse magnetization (Mxy)
in a real external magnetic field

The FID follows T2* relaxation: Mxy = M0 e-t/T2*

T2 is the rate of decay of transverse magnetization (Mxy) in a

perfectly uniform external magnetic field

SE amplitude is determined by T2: Mxy = M0 e-t/T2

T1 is the rate of recovery of longitudinal magnetization (Mz)

following a perturbing RF pulse

T1 relaxation is not directly observable; Mz = M0 (1 – e-t/T1)

T2 * < T 2 ≤ T 1