University of Tennessee, Knoxville

TRACE: Tennessee Research and Creative

Exchange

Supervised Undergraduate Student Research

Chancellor’s Honors Program Projects and Creative Work

Fall 12-2000

An Introduction to Magnetic Resonance Imaging Physics

Douglas Wayne Adkisson
University of Tennessee - Knoxville

Follow this and additional works at: https://trace.tennessee.edu/utk_chanhonoproj

Recommended Citation
Adkisson, Douglas Wayne, "An Introduction to Magnetic Resonance Imaging Physics" (2000). Chancellor’s
Honors Program Projects.
https://trace.tennessee.edu/utk_chanhonoproj/360

This is brought to you for free and open access by the Supervised Undergraduate Student Research and Creative
Work at TRACE: Tennessee Research and Creative Exchange. It has been accepted for inclusion in Chancellor’s
Honors Program Projects by an authorized administrator of TRACE: Tennessee Research and Creative Exchange.
For more information, please contact [email protected].
 UNIVERSITY HONORS PROGRAM

SENIOR PROJECT • APPROVAL

N am e: ___\..).Qii.-~klS.._\L(~'...~_A4:~'LSj_O_Vl, __________ : ______________ _

College: J~~.9-~JA~------- Department:. __ M.N::S_______________

F ac u 1ty Men to r: ___ J2.Ll.._RY.b..erl__ b-~_JoAL ______________________ _
P ROJE C T TITLE: _A.\l.ly:ito-"1.~~i.\Ii~___\.y___ M\1~iIldLc-------------
--Rtj~~~~~_.l[~~~j--Pj:~~c~-----------------------------
-~--~-----------------------~--------------~--------------

that it is a project co
field.
Signed: ___ _
a
I have reviewed this completed senior honors thesis with this student and certify
ensurate with honors level undergraduate research in this

----~2~_________ , Faculty Mentor

Da te: --1i2jLi!-(1{2____ _
Comments (Optional):

()J~ 7)t»<-e - t/7 h~ d totf7/~

f l WtJ?:l A...pJ~or.e. tut1!U1 W/rf?, tfJO'j
~(S t(~ Ik- ~-/wJtJld /)~ urfJ/O(/~
/ J1. flu- #VK e I
 An Introduction to Magnetic
Resonance Imaging Physics

Douglas W. Adkisson
Faculty Mentor: Dr. Robert L. Toal
Acknowledgements

To Dr. Toal, many thanks for his patience and expertise

Abstract
Since its inception in the late 1970s, magnetic resonance imaging

(MRI) has been pushing the limits of diagnostic medicine. Diagnoses which

were made with great difficulty (or not made at all) using computed

tomography or x-ray imagings are now made with relative ease with MRI.

One of the most exciting aspects of MRI is the utility it has in so many

medical specialties. For example, MRI is currently being used to study the

functioning within the brains of Alzheimer's patients, take "snapshots" of

the human heart, and trace nerve tracts within the human brain. However,

the physical concepts upon which MRI technology was built seem to often be

misunderstood or ignored. An understanding of these physics is essential to

fully appreciating the images MRI provides us and capabilities MRI will offer

clinicians in the future. The purpose of this forum is to provide the reader

with a comprehensive overview of MRI physics in a practicat easily readable

manner.
Introduction
Magnetic resonance imaging magnets produce very large magnetic

fields. An MRI-induced magnetic field of 1.5 Tesla (1 Tesla (MKS units) =

10,000 Gauss (CGS units)) is 30,000 times greater than the strength of the

Earth's magnetic field (0.00005 Tesla) and 15 times greater than the

magnetic field that exists on the surface of common refrigerator magnets

(0.1 Tesla). During the relatively brief history of magnetic resonance imaging
(MRI), three different varieties of magnets have been utilized in obtaining

images. These magnets are classified as permanent, resistive, and

superconducting. However, while both permanent and resistive magnets have

their respective advantages, over 95'0 of the imagers today make use of
superconducting magnets (1).

Imaging Magnets
Permanent Magnets. Permanent MRI magnets consist of approximately one
hundred tons of bricklike solids made from a ceramic ferromagnetic alloy (1).

This type of magnet is comparatively inexpensive to produce, and its low

maintenance cost makes permanent MRI magnets a very cost-effective

alternative in smaller clinical settings. The two poles of the magnet are
situated above and below the patient, producing a vertically oriented main
magnetic field (along y-axis), and provide a spacious area for the patient
(see Figure 1). This "open" design is advantageous in instances where patients

suffer from claustrophobia; therefore, permanent magnets are often used in

"open" MRI scanners (10). However, this classification of magnet has its

share of shortcomings. Permanent magnets are only capable of producing

magnetic fields up to the order of 0.3 Tesla (T) and consequently are not

2
suitable for all diagnostic purposes. The fact that the magnetic fields are

fixed and that great difficulty must be taken in producing a uniform

magnetic field with these magnets are two other disadvantages associated

with permanent magnets (1).

Figure 1: MRI scanner coordinate system (see reference 2)

Resistive Magnets. Resistive MRI magnets (electromagnets), much like

permanent magnets, allow for a nonconfining patient area and form a

vertically oriented magnetic field (see Figure 2).

3
Figure 2: Open MRI scanner (see reference 10)

This main magnetic field is created by sending an electric current through

concentrically arranged coils of copper wire, which lie in a plane horizontal to

the patient. Magnetic fields up to 0.6 T are produced by the coils and exert
large forces on the remaining coils, which make it necessary to rigidly secure

each coil.
The magnitude of the magnetic field needed for MRI requires an

enormous amount of electrical current. This large power prerequisite can

prove to be very expensive! An added consequence of this power necessity

is the fact that the coils must be water-cooled. Although only a small amount

4
of resistance retards the current within the copper coils, the enormous

amount of electrical current causes the coils to become extremely hot.

Superconducting Magnets. In the early 1960s, a phenomenon was discovered

in which certain metal alloys acted as perfect conductors at temperatures

near absolute zero (0 degrees Kelvin). Gradually, materials have been

discovered that function as superconductors at higher and higher

temperatures. Magnets whose function depends upon this discovery are

referred to as superconducting magnets and represent over 95% of "the

magnets used clinically today. The magnetic field is produced by passing

current through miles of superconducting wire that is precisely wound into

several coils that ring the patient (10). Unlike permanent and

electromagnets, the magnetic field of superconducting magnets is oriented

parallel to patient, along the z-axis (see Figure 1). Recently, materials have

been identified that function as superconductors room temperature which

promise to have a significant influence on MRI in the near future (8).

The advantages of superconducting magnets should be rather obvious.

Since no resistance to electric conduction is present in superconducting

coils, heat dissipation is not an issue. An added benefit of the absence of
resistance is that once an electrical current is introduced into a
superconducting coil, it will remain indefinitely. The ability to produce larger
magnetic fields than either permanent or resistive magnets is also an

unmistakable advantage. Superconducting magnets are often used to produce

fields of the order of 1.5 T or 3.0 T for commercial use; however, larger

magnetic fields have been produced for the purpose of research (8). For

5
example, at the Ohio State University, a scanner is in use that can produce

B-fields up to 8.0 T in magnitude.

The main disadvantage of superconducting magnets in MRI scanners is

the fact that the coils must be maintained at a cryogenic temperature at all

times. In order to accomplish this task, manufacturers use a combination of

liquid nitrogen and helium chambers called a cyrostat (see Figure 3). The

nitrogen chamber acts as a buffer between the external environment and

the coils while the helium chamber actually performs the job of supercooling

the coils (1). Additional electricity is needed to operate the refrigeration

units to keep the cryogens at the proper temperature (8). Due to this

cyrostat design, many superconducting devices currently in use do not

provide the added patient area of permanent and resistive magnet scanners.

However, open MRI devices utilizing superconducting technology have begun

to surface (8).

Vacuum

------------r ,,,I \

Liquid I
Nitrogen
, ;

-----------_ .....,
\ I

1-4f--1m...J
~14~------3m------~~1

Figure 3: Nitrogen and helium chambers of cyrostat (see reference 1)

6
Magnetic Resonance Imaging Physics

Spin and Its Consequences. Subatomic particles such as protons and

neutrons are known to possess a characteristic spin of either +1/2 or -1/2.

Much like particle charge, spin of subatomic particles is an additive

property; therefore, by summing spin of individual particles, it is possible to

obtain a "net" nuclear spin. Luckily, spin values are not randomly distributed

among nuclear particles. In fact, for each pair of like particles (example:

proton-proton pair), each entity will possess a spin opposite to its

counterpart, and net spin of the pair will be zero. A simple deduction can be

made from this observation: net spin of a nucleus with an even atomic

number is zero. Thus, many common elements within the body, such as

oxygen, carbon, and calcium, possess an even number of nuclear particles and

are therefore useless for purposes of MRI (6). However, the most abundant

of all tissue components, hydrogen, has an atomic number of only one

(remember, a large percentage of the body is water and each water molecule

contains two hydrogen atoms) and therefore will possess a net spin equal to

the spin of that one proton. Note: Since a hydrogen nucleus contains only one

proton, the words "hydrogen nucleus" and "proton" will be used

interchangeably throughout this text.
Laws of quantum mechanics dictate that a magnetic field will be
produced by the spin of a charged particle (neutrons, with a charge of zero,
can therefore be ignored). The magnetic field direction is determined by a

rule known as the right-hand-rule. Taking your right hand, if you curl your

four fingers in the direction of spin and extend your thumb, the direction of

your thumb will correspond to the north pole of the magnetic field produced

7
by the particle. The direction of the north pole is often denoted by a vector

(see Figure 4) known as a "nuclear" magnetic moment (9).

South

I
f"".,.~
: : : '!'\: :
I I I I I
I 'I I I I
I 'I I I I
, I I
I I I
I I I
I I
I I
, I
I
I

Figure 4: Magnetic field produced by a spinning proton (see reference 10)

In soft tissue of patients, as in most materials, magnetic moments of

protons are oriented in a random manner. Since individual magnetic moments

of particles is also an additive property, random orientation essentially

cancels out the effect of each individual magnetic moment and results in a

net magnetization of zero within tissue (1). If however the patient is placed
I I

in a strong external magnetic field, these small magnetic moments will align

either parallel or anti-parallel to the external magnetic field (see Figure 5).
The nuclei aligned parallel to the external magnetic field are at a lower
energy level than those of the opposite orientation. Intuitively this should I

seem natural since the north pole of the particle s magnetic field is aligned I

with the south pole of the main magnetic field (just as with bar magnets,

opposites attract and Ii ke poles repel). Due to this advantageous energy

state, a slightly larger portion of the spinning nuclei (less than 1 in 106

additional protons) are aligned parallel to the external field (So) than against

it. When this occurs, the patient is said to possess a small net magnetization

8
(Mo) in the direction of the external magnetic field (see Figure 5). The

reason that some protons align anti-parallel is because of thermal molecular

interactions induced by molecular motions, rotations, and vibrations that

impact proton orientation while in the magnetic field (8).

~ t
••••• t --x
individual
nuclear 0'-.,

magnetic
moments

Figure 5: Net magnetization of patient in MRI scanner (see reference 1)

Nuclear Precession. Not only do small particles spin, but in a strong magnetic
field, they also have a tendency, much like a gyroscope, to wobble in a circle

about the axis of spin (9). This wobbling is known as precession. Thus, as it
applies to MRI, nuclei not only spin but also precess (wobble) about the

externdl magnetic field lines (see Figure 6).

9
 South

a b

- - - - - - - - ........
'\ ,
'\
, '\
,
'\
'\ I
'\
,.-
.I
'\

H. . apid spin '\. ,,

about slo\vJy '\ , /

prc<.-:cssing axis '\ I

.I

../

Figure 6: Precession of nuclei (see reference 10)

The frequency with which hydrogen nuclei precess about the external

magnetic field lines is known as the Larmor or resonant frequency. The

Larmor frequency of individual nuclei is directly dependent on strength of

the external magnetic field and a value known as the gyromagnetic constant.

Where m = Larmor frequency (MHz)

y = gyromagnetic constant (MHz/Tesla)

Bo =magnetic field strength (Tesla)

The gyromagnetic constant is a nucleus-specific property (hydrogen: y =

42.57 MHz/T). Therefore, for a given magnetic field strength, different

body elements will precess at different frequencies (7). Although hydrogen

nuclei all precess at the same frequency in a given external magnetic field,

they do not precess in phase with one another (random phase orientation),

10
For example, the resonant frequencies of a proton in a 0.5 T and 1.5 T B-

field are 21.285 MHz and 63.855 MHz, respectively.

Energy can be transferred to nuclei by electromagnetic waves. In

clinical settings, these electromagnetic waves are known as RF (radio

frequency) pulses. If the B-field strength within the MRI device is known,

then the resonant frequency of the hydrogen nuclei within the patient can

be easily calculated. The passage through the patient of an RF pulse

possessing this Larmor frequency will cause hydrogen nuclei within the

patient to absorb a photon of energy and experience resonance (10). This

energy absorption allows those nuclei at a lower energy state to IIflip" and

jump to a higher energy state (align anti-parallel to the external B-field)

while at the same instance causing the hydrogen nuclei to precess in phase

with one another (see Figure 8).

tu SOll[h

l • • ;. • ..
, ,
~
•
,,
.. ~

.. . .
I , I •
: : : INl
I • I t I
,,
1
,, ••
•
,
1

,, , .
r I. • l 1
•
I
•
•I 5~pilt-llip :
J\l\flf\ ___ ~..,.~
I
I
___ I
I -"'" ...1'\:
r
I
-VUV- I' •
I
I '·;~I/: •
I

.:·
Lulllor-fr. . . ..:llH:JlCV: ~
,
I
I
I •• •
(".f.} photon ' 1:
, .
. •
I
I
I
•
I
I
,
I
I
I

, .
~ I , • t
, I I ,
I • ~
, I
, ,
, I I
, I I I I I
r

a
[~ . North [-~~-~~~~~:~~1~~·· ."]
I ligh cncTb'")' statt'

Figure 8: Resonance in hydrogen nuclei (see reference 10)

However since the electromagnetic wave is sent as a pulse, the excited

I

state of the hydrogen nuclei is only momentary. The nuclei will soon fade

12
back (relax) into their original orientation parallel to the external magnetic

field and lose the phase coherence gained by the RF pulse.

Essential Components of MRI Scanners (see Figure 9)

(;.-adicnt coils
'-. t coil

Gr,ldicnt driver:<-

C:Olllput\"I- systt.~Jl)
& display

Figure 9: Essentials within an MRI scanner (see reference 10)

Main magnet. The main magnet is the component of the scanner that is
responsible for producing a strong main magnetic field. As mentioned earlier,

these magnets can be classified into three groups: permanent, resistive

(electromagnets), and superconducting magnets.

Shim Coils. Homogeneity of the main external B-field is of utmost

importance in MRI scanners. Variation within this B-field will result in the

degradation of image quality (10). For example, a 1.0 T magnetic field should

vary no more than ±50 JJ T or ±50 ppm (parts per million). Sufficient

homogeneity can be achieved by the use of shim coils. Located just inside

the main magnet, each coil possesses its own power supply and is capable of

13
adjusting its current and polarity in such a way that the homogeneity of the

main magnetic field is maximized.

Gradient coils. Located just inside the shim coils are three independent coils
known as gradient coils (one for each of the x, y, and z-axes). These coils

are used in choosing location of the imaging plane within the patient (see

Slice mechanics for more details). These fields are intermittently switched
on and off throughout a scan and are responsible for the muffled, booming

sounds that can be heard by the patient (10).

RF/Receiving coils. RF/receiving coils serve a dual purpose in MRI scanners.

Their primary task is to produce the RF (radio frequency) pulses that cause

protons to absorb photons of energy and therefore result in proton spin

flips. However, RFIreceiving coils also detect the weak RF signals that are

given off by the relaxing hydrogen nuclei. The coils then amplify the signal

and send it to a computer for analysis (1).

RF/receiving coils are available in a variety of different shapes, with

each shape lending itself to a particular type of imaging (10). For example, an
RF/receiving coil known as a saddle coil is used primarily for neuroimaging

(see Figure la, top left corner).

14
 ......... -- ----
........- ....-... r.............
I~

FIG. 7-1. Various coils used in vertical field magnets.

Figure 10: Various RF/receiving coils (see reference 4)

Computer system. The control of gradient coil and RF pulse initiation is

under computer control. A computer system is also responsible for storing

and analyzing resonant data, carrying out reconstruction calculations, and

producing images for display (10).

Magnetic Resonance Imaging Parameters

Computed tomography (CT) and x-ray imaging both use a parameter known as

linear x-ray attenuation coefficient to form images. However, the formation

of magnetic resonance images is dependent upon three factors: spin density,

T1 relaxation time, and T2 relaxation time (7). This accounts for the

superior contrast resolution of MRI versus CT.

15
Spin density. Spin density is a measure of the concentration of hydrogen
nuclei within a particular region of interest. The greater the number of

hydrogen atoms present, the stronger (brighter) the signal received from

the patient. Tissues such as muscle and fat, which are rich in hydrogen

atoms, will appear much brighter than materials with little or no hydrogen

atom concentration. This is the reason that bone and air appear black in

MRI. Since spin density is directly proportional to the number of hydrogen

nuclei (protons), it is also known as proton or hydrogen density (1).

Tl relaxation time. When an RF pulse with a frequency corresponding to the

Larmor frequency of hydrogen irradiates a patient, some of the magnetic

moments (spins) aligned parallel to the strong external magnetic field will

absorb energy and flip to an orientation anti-parallel to the external field

(see Figure 8). The number of magnetic moments that absorb energy is

dependent on the strength and duration of the RF pulse. Therefore,

manipulating the RF pulse can control the number of spins that are flipped

(7).

As mentioned earlier, the application of an external magnetic field

results in a small surplus of magnetic moments parallel to the external B-
field. This surplus of spins yields a net magnetization in the direction of this
B-field (see Figure 5); in other words, the patient becomes "magnetized."
Let's imagine for a moment that an RF pulse has a strength and duration such

that all of the spins in this surplus flip to an anti-parallel orientation. As a

result of this pulse, the net magnetization of the patient would also flip
0
against the field (see Figure 11). This type of pulse is referred to as a 180

pulse since the net magnetization of the patient is reversed (1).

16
 Let us now consider an RF pulse that flips the magnetic moments of

half of the "surplus" hydrogen nuclei. Since this pulse would result in only

half of the spins being flipped, the net magnetic moment of the patient

would rotate only halfway to the anti-parallel position. In other words, the

net magnetization would lie in a plane perpendicular to the plane of the

external magnetic field. This type of pulse is referred to as a 90° pulse

since the net magnetic moment rotates through a right angle (1) .
•. _!..

Bo

~t··
a 900 pulse4t t
x
~tttt
a 180°pulse tt

Figure 11: 90 and 180 degree pulses (see reference 1)

Once an RF pulse is terminated, the energized protons begin losing

this absorbed energy and fade back (referred to as relaxation) to their

original (parallel) position. The released energy is given off in the form of an

RF signal (9). A receiving coil positioned parallel to the axis of the external

B-field detects these RF signals radiating from these protons. These

17
electromagnetic waves induce a voltage in the coil which is analyzed and used

to form magnetic resonance images.

The signal released by the protons is at its strongest when the entire

net magnetic moment of the patient is perpendicular (transverse) to the

external B-field. As this net magnetization regresses back to a parallel

orientation, its transverse component and the signal detected by the

receiving coil becomes smaller. This loss of energy and subsequent

realignment with the external B-field is known as T1 relaxation (10). T1

relaxation (regrowth of net magnetization) can be defined by the function:

t
l-e Tl

where t is time and T1 is T1 relaxation time. When t is equal to T1, this

equation reduces to 1 - e- 1, which has a value of 0.63212. In physical terms,

TI relaxation time is the timed required for the net magnetization of the
patient along the z-axis to regain 63.212'0 of its original value (see Figure

12).

Tl Relaxation

"2
~
.... s:::: 0.8
'51 .!:!
s:::: ....
.s
1; t
2 0.6
0.4
1- Tl Relaxation 1
..s:: '='
~ =E 0.2
I 0
o 50 100 150

Time (ms)

Figure 12: Tl relaxation with Tl relaxation time =37ms

18
 The return of net magnetization in the longitudinal plane does not give

off a signal. This means that T1 cannot be measured directly. Rather,

successive RF pulses are used to flip the nuclei from lower to a higher

energy state. The difference in the transverse Signal given off by the nuclei

following each RF pulse is proportional to the number of atoms returning to

equilibrium (relaxation). This then becomes a measure of T1 (8).

Each tissue has its own characteristic T1 relaxation time due to the

unique manner by which it is held together and the unique magnetic

environments these tissues produce. Since this type of relaxation occurs in

the plane of the external B-field, T1 relaxation time is also known as

longitudinal relaxation time. An additional moniker for this parameter, spin-

lattice relaxation time, is due to the interaction of released energy with the

lattice framework of the tissue (1).

T2 relaxation time. In addition to flipping the magnetic moments of protons,

properly tuned RF pulses also cause protons to precess in phase with one

another. Therefore, for a 90° pulse, the net magnetization can be imagined

to be lying in the transverse plane and precessing about the axis of the
external magnetic field. However, when the RF pulse is terminated, the
interaction of individual magnetic moments (fields) causes some protons to
precess faster, some slower. This process is known as dephasing (7).
As the protons fall further and further out of phase, the transverse

component of the patient's net magnetic moment becomes smaller and

smaller. Eventually, the individual magnetic moments will dephase completely

and the transverse component will be nonexistent, as it was before the

introduction of the RF pulse (see Figure 13).

19
 M
a b

Figure 13: Dephasing of proton spins (see reference 10)

This process of dephasing and subsequent shrinking of the transverse

component of the net magnetic moment is known as T2 relaxation. Since this

decay of phase coherence results from the interaction of individual spins

and occurs in a plane transverse to the external B-field, this parameter is

also known as spin-spin or transverse relaxation time (1).

Transverse relaxation with respect to time can be represented by the

equation

t
e T2

where t is time and T2 is T2 relaxation time (see Figure 14). When t is equal

to T2, the equation reduces to e-1, which has a value of 0.36788. In physical

terms, T2 relaxation time is the time required for the transverse net

magnetization of the patient to fall by 36.788c:to or to a value that is

63.212c:to of its original value.

20
 T2 Relaxation

1
~ ~
~ ~ 0.8
~ :::J
> 0 Ol
~ s:: .~ 06
a 0 '"
'- .+= ..2
f- a 0..
~ .~ ~
.

0.4
1- Tl Relaxation 1
°:to
'" s::
~ ~ 0.2
-l~

o
o 50 100 150

Time (ms)

Figure 14: T2 relaxation with T2 relaxation time =37 ms

Just as with T1 relaxation times, T2 relaxation times differ from

tissue to tissue due to their different magnetic environments. In fact, T1

and T2 relaxation times can differ by 20i'0 to 30i'0 between various soft

tissue types. This superior contrast resolution is one major advantage that

MRI has over both computed tomography (CT) and x-ray imaging (1). X-ray

imaging uses linear x-ray attenuation coefficient to form its images.

However, this imaging parameter differs by only approximately 1i'o in soft

tissues. These x-ray images are also degraded by scatter radiation. CT

imaging also uses linear x-ray attenuation coefficient as its imaging

parameter. CT does however eliminate scatter radiation degradation, which

provides greater contrast resolution than x-ray imaging.

Slice Mechanics

Slice selection. One distinctive advantage that magnetic resonance imaging

has over other imaging modalities is its ability to perform multiplanar

21
imaging without repositioning the patient (10). This method of slice selection

involves the use of a group of coils oriented along the x, y, and z-axes of the

patient known as gradient coils (see Figure 1). Initiation of a gradient coil

forms a magnetic field that can be superimposed onto the main B-field. This

superposition of B-fields creates a gradient magnetic field in the direction

of the gradient coil's orientation that varies linearly from one extreme of

the patient to the other (see Figure 15).

Gradient Field in Z-Direction

1='
......,
1.56
~ 1.54
::J

~ 1.52
-0
""i) 1.5 Gradient Field I
tL
~ 1.48
~
§, 1.46
a
:E 1.44

Distance Along Z-Axis

Figure 15: Gradient magnetic field along z-axis with main field = 1.5 T

For example, suppose for a moment that the gradient coil oriented along the
z-axis was initiated. The combination of the resulting B-field with the main

magnetic field produced a gradient field with values of 1.3 T and 1.5 T at the

lower and upper extremities, respectively, of the patient. Since this gradient

field is known to vary linearly along the z-axis, the value of the field at any

point along the z-axis can be calculated.

22
 You may be wondering how a linearly varying gradient field would be

helpful in slice selection. If you recall, the resonant (Larmor) frequency of

hydrogen nuclei is dependent upon the gyromagnetic constant of hydrogen

and the magnitude of the magnetic field to which a hydrogen nucleus is

exposed. Therefore, since the magnetic field at each point along the patient

is slightly different, if the B-field magnitude at the plane of interest is

known, the appropriate RF pulse can easily be obtained that will excite only

those protons within that plane (2). Referring back to the previous example,

imagine that physicians desire to image a plane exactly halfway between the

lower and upper extremities of the patient. Using a B-field of 1.4 T in the

Larmor equation, the frequency needed to induce resonance in that plane

could be found. The type of image in the example would be known as a

transverse (trans-axial) image. Images obtained from a gradient field along

the x-axis are known as sagittal images while those obtained from gradients

along the y-axis are known as dorsal (coronal) images. It is also possible to

obtain oblique images by using combinations of the x, y, and z-gradients (see

Figure 16).

23
Figure 16: (top left) transverse image; (top right) coronal image; (bottom) oblique image; sagittal images and slice
plane can be seen in bottom left corner of each image (see reference 4)

Slice thickness. The thickness of the slices is determined by the range of

frequencies included within the RF pulse (this has not been mentioned

earlier, but it is possible to send pulses containing a band of frequencies).

For example, an RF pulse containing all frequencies from 64 MHz to 64.5

MHz would excite all protons with resonant frequencies within that range.

24
Slice thickness can also be influenced by the severity of the slope of the

gradient field: the steeper the slope, the thinner the slice (2).

Image reconstruction. The formation and timing of gradient fields and RF

pulses is under computer control (see Figure 9). A computer also has the

responsibility of analyzing data from the receiving coil, carrying out

reconstruction calculations, and producing the resulting images for viewing.

The mathematical procedure that is utilized by the computer in the

reconstruction calculations is known as the Fourier transform (7). A two-

dimensional (2-D) Fourier transform is used in the analysis of slices formed

by a single-frequency RF pulse (a two-dimensional plane). In order to

determine the fraction of the signal that originated from each portion of

the plane, the slice must be divided into tiny squares, similar to the divisions

present on graphing paper. These tiny squares are known as pixels and this

process is known as spatial localization (2). Spatial localization is achieved

with the gradient coils within the scanner.

Once the slice has been selected by the slice-select gradient, the

remaining two gradient coils are used to divide the slice into strips and then
into squares. The remaining two gradient coils are referred to as the
frequency-encoding gradient and the phase-encoding gradient (3). In order
to grasp a better understanding of this process, let us again refer back to
the example of the transverse slice. In this particular case, the gradient coil

along the z-axis would be the slice-select gradient, and the x and y-axis

gradients would be the frequency-encoding and phase-encoding gradients,

respectively (with the x-axis gradient dividing the plane into vertical strips

and the y-axis gradient dividing it into vertical strips). Once the slice has

25
been spatially encoded, the Fourier transform (in very simplified terms) can

convert the data in each pixel into an intensity or gray-scale value. The

compilation of these intensity values, which range from white (highest

intensity) to black (lowest intensity), result in a viewable image.

A three-dimensional (3-D) Fourier transform is used in the analysis of

slices formed through the application of an RF pulse containing a range of

frequencies (2). The plane of interest is possesses a finite thickness and can

be considered as a volume element. Appropriately, the divisions into which

the volume element is separated are known as voxels. As one might expect,

the greater the number of divisions within the slice, the more precise the

image reconstruction will be for each point in space and the greater the

spatial resolution in the resulting image (this is applicable for the 2-D case

as well).

Hazards of MRI Scanning

Numerous studies have been performed on the effect of strong magnetic

fields on biological tissue, and no evidence has ever been found to suggest

that MRI scanning is directly harmful to the patient. Further, researchers

are extremely confident that no long-term effects will be found by
subsequent studies (1). There are, however, indirect dangers associated with
the fields produced during MRI scanning.

Cardiac pacemakers. Cardiac pacemakers use a small electromagnet in their

operation and exposure to a strong magnetic field can cause the pacemaker

to fail. Therefore, those individuals with cardiac pacemakers should refrain

from being near MRI scanners and should never be imaged (1).

26
Other internal metallic objects. Metallic objects within the body such as
aneurysm clips, surgical pins, shrapnel, artificial heart valves, etc. often pose

no threat to the patient (10). It is, however, advisable for one to inform

their physician of any possible source of metallic material within their body

to ensure no possible side effects will occur.

Personal Items. Items such as credit cards or wristwatches should not be

brought into contact with the strong magnetic fields of MRI scanners.

Credit cards contain magnetic strips and can be deactivated by a strong B-

field. Tiny moving parts within a mechanical watch can become magnetized

and cause the watch to malfunction (I),

Metallic objects. The main magnetic field of a MRI scanner can extent huge
forces on magnetic materials made of steel, iron, or other ferromagnetic

(easi Iy magneti zed) materials. Pens have been known to be sucked right out

of the pockets of researchers. Personnel must also be careful when

maintenance work is being performed near a scanner. For example, one story

surfaced at a prominent medical school in which a maintenance worker

veered a bit too close to a superconducting MRI scanner while vacuuming.

The vacuum was sucked into the core of the magnet, and the magnet had to

be quenched and the cooling mechanism deactivated in order to remove the

vacuum cleaner.

Fringe fields. An added problem with resistive magnets is the fact that the
magnetic fields produced extend beyond the MRI scanner (see Figure 17).

27
This excess field is known as a "fringe" field. Fringe fields affect the

performance of many types of equipment, most notably computers, electron

microscope, and cathode ray tubes, and must be accounted for in the
hospital or lab setting.

Figure 17: Fringe field due to MRI scanner (see Reference 1)

Conclusion

The discussion thus far has entailed the very basics upon which magnetic
resonance imaging is built as well as some practical aspects within MRI.
Other subjects, such as RF pulse sequences and the weighting of images to

emphasize one parameter over another, is most definitely worthy of

consideration but are not within the focus of this forum. MRI has made an

incredible impact on the field of diagnostic medicine, and with alternate

techniques such as magnetic resonance angiography becoming popular in

28
clinical settings, it promises to become the standard upon which all other
diagnostic tools are measured. A thorough understanding of the basic
concepts of MRI is essential to fully appreciating the images produced by
MRI and the capabilities that MRI will provide to medical professionals in
the future.

29
References
1.) Bushong SC. Radiologic Science for Technologists: Physics, Biology,

and Protection. St. Louis: Mosby, 1993.

2.) Friedman BR, et al. Principles of MRI. New York: McGraw-Hili:

1989.

3.) Miyabayashi T, Smith M" Tsuruno Y., Comparison of Fast Spin-Echo

and Conventional Spin-Echo Magnetic Resonance Spinal Imaging

Techniques in Four Normal Dogs. Veterinary Radiology and

Ultrasound2000; 41: 4: 308-312.

4.) Rothschild PA, Rothschild DR. Open MRI. Philadelphia: Lippincott

Williams & Wilkins, 2000.

5.) Seguin B, et al. Use of magnetic resonance angiography for

diagnosis of portosystemic shunts in dogs. Veterinary Radiology &

Ultrasound1999; 40: 3: 251-258.

6.) Shores A. Magnetic resonance imaging. The Veterinary Clinics of

North America Small Animal Practice 1993; 23: 2: 437-459.

7.) Tidwell AS. Advanced Imaging Concepts: A Pictorial Glossary of CT

and MRI Technology. Clinical Techniques in Small Animal Practice

1999; 14: 2: 65-111.
8.) Toal RL. Private conversations. 2000.
9.) Widmer WR, et al. Principles of magnetic resonance imaging and
application to the stifle joints of dogs. Veterinary Medicine Today

1991; 198: 11: 1914-1922.

10.) Wolbarst AB. Looking Within: How X-Ray, CT, MRI, Ultrasound,

and Other Medical Images Are Created. Berkeley: University of

California Press, 1999.

30