Biomedical Engineering Reference
In-Depth Information
use, consists of pulsed RF excitation followed by the
detection of the resultant preprecession signal. Hence,
rather than being simultaneous, in this scheme excitation
and detection are performed sequentially.
A major milestone was the discovery of the chemical
shift by Warren Proctor, F. C. Yu, and W. C. Dickinson.
They found that in ammonium nitrate, two nitrogen-14
resonances could be observed, which they ascribe to the
different chemical environments to which the nitrogen
nucleus is exposed in the nitrate and ammonium ions.
Similar findings were later made by others for nuclei such
as fluorine, phosphorus, and hydrogen. These observa-
tions constitute the basis of modern NMR spectroscopy.
A few years later, as the magnetic homogeneity that de-
termines the frequency resolution achievable in NMR
was improved further, another type of fine structure was
discovered. This structure, which is due to spin-spin
coupling, is fundamental to modern high-resolution
spectroscopy, and together with the chemical shift
provides the basic ingredients for molecular structure
determination. Today, NMR is the preeminent method
for determining the structures of biomolecules with
molecular masses up to 100,000 Da.
The development of pulse Fourier transform by
Richard Ernst and Weston Anderson was of great im-
portance for NMR. This alternative mode of signal cre-
ation, detection, and processing led to an unprecedented
enhancement in per-unit-time detection sensitivity
compared with continuous wave excitation techniques. If
N
channels are used simultaneously in an experiment,
then, provided the dominant source of noise is not the
excitation, the sensitivity increases by a factor of
N
1/2
.
Ernst and Anderson demonstrated that one can affect
broadband excitation by exciting the nuclear spins with
short RF pulses of a single carrier frequency.
A new breakthrough was added to NMR technology in
1973 when Paul Lauterbur at the State University of New
York at Stony Brook first proposed generating spatial maps
of spin distributions by what he called ''NMR zeugma-
tography.'' The key to this method was the idea of super-
imposing magnetic field gradients onto the main magnetic
field to make the resonance frequency a function of the
spatial origin of the signal. In the presence of a magnetic
field gradient the frequency domain signal is the equivalent
of a projection of the object onto the gradient axis. By
rotating the gradient in small angular increments, one
obtains a series of projections from which an image can be
reconstructed using back projection techniques.
In and around 1980 whole body experimental NMR
scanners were in operation, and by 1981 clinicians began
to explore the clinical potential of MRI. NMR has several
advantages over x-ray computerized tomography (CT).
First, it was noninvasivedthat is, it did not require ion-
izing radiation or the injection of contrast material.
Second, it provided intrinsic contrast far superior to that
of x-ray CT. Some of this early work showed that MRI was
uniquely sensitive to diseases of the white matter of the
brain, such as multiple sclerosis. Further, the contrast
could be controlled to a significant extent by the nature
and timing relationships of the RF pulses. Third, MRI was
truly multiplanar and even three dimensional (3D)
that
is, it could provide images in other than the traditional
transverse plane without the subject having to be reposi-
tioned. This property turned out to be of great value over
x-ray CT for the study of the brain and other organs.
In 1975 Ernst introduced a new class of NMR exper-
iments now known as two-dimensional (2D). NMR, and
should be regarded as the parent of modern NMR tech-
niques. One can understand the principle by reference to
Figure 6.1-1
. Suppose the spins in a point object of spin
density r(
x
,
y
) are excited by an RF pulse in the presence
of a magnetic field gradient of magnitude
G
y
, which we
call a phase-encoding gradient. These spins will resonate
at a relative frequency u
¼
g
B
(
y
), where
B
(
y
)
¼ G
y
Y
.
If the gradient is active for a period
t
y
, the phase at the
end of the gradient period is f
y ¼
g
G
y
yt
y
. Let us then
step the time
t
y
in equal increments, as applied by
Figure 6.1-1b
. We readily notice that the phase at the end
of the gradient period varies cyclically with time. At time
t ¼ t
y
, the gradient
G
y
is turned off and an orthogonal
d
(a)
Radio
-
frequency pulse
Gy
Gx
t
y
t
x
(b)
Phase Encoding
Grad
ient Gy
t
y
Phase
Magnetization
Figure 6.1-1 Fourier zeugmatography principle.
