Note: Descriptions are shown in the official language in which they were submitted.
~L2~
The field oE the invention is gyromagnetic resonance spectro-
scopy, and particularly, nuclear magnetic resonance (NMR) techniques
for measuring the properties of materials.
Gyromagnetic resonance spectroscopy is conducted to study
nuclei that have magnetic moments and electrons which are in
a paramagnetic state. The former is referred to in the art
as nuclear magnetic resonance (~MR), and the latter is referred
to as paramagnetic resonance (EPR) or electron spin resonance
(ESR). There are other forms of gyromagnetic spectroscopy that
are practiced less frequently, but are also included in the
field of this invention.
Any nucleus which possesses a magnetic moment attempts
to align itself with the direction of the magnetic field in
which it is located. In doing so, however, the nucleus precesses
around this direction at a characteristic angular frequency
(Larmour frequency) which is dependent on the strength of the
magnetic field ancl on the properties of the specific nuclear
species (the magnetogyric constant ~ of the nucleus).
When a substance such as human tissue is subjected to a
uniform magnetic field (polarizing field Bz) the individual
magnetic moments of the paramagnetic nuclei in the tissue attempt
to align with this field, but precess about i-t in random order
at their characteristic Larmour frequency. A net magnetic moment
Mz is produced in the direction of the polarizing field but
the randomly oriented components in the perpendicular plane
(x-y plane) cancel one another~ If, however, the substance,
or tissue, is irradiated with a magnetic field (excitation field
Bl) which is in the x-y plane and which is near the Larmour
frequency, the net aligned moment, Mz, can be rotated into the
x-y plane to produce a net transverse magnetic moment Ml which
is rotating in the x-y plane at the Larmour frequency. The
f~L2(~3~
degree to which the rotation of Mz into an Ml component is achieved,
and hence, the magnitude and the direction of the net magnetic
moment (M = Mo + Ml) depends primarlly on the length of time
of the applied excita~ion field Bl.
The practical value of this gyromagnetic phenomena resides
in the radio signal which is emitted after the excitation signal
Bl is terminated. When the excitation signal is removed, an
oscillating sine wave is induced in a receiving coil by the
rotating field produced by the transverse magnetic moment Ml.
The frequency of this signal is the Larmour frequency, and its
initial amplitude, Ao, is determined by the magnitude of Ml.
The amplitude A of the emission signal (in simple systems) decays
in an exponential fashion with time, t:
A = A~e~t/T2
The decay constant l/T2 is a characteristic of the process
and it provides valuable information about the substance under
study. The time constant T2 is referred to as the "spin-spin
relaxation" constant, or the "transverse relaxation" constant,
and it measures the rate at which the aligned precession of
the nuclei dephase after removal of the excitation signal Bl.
Other factors contribute to the amplitude of the free induc-
tion decay (FID) signal which is defined by the T2 spin-spin
relaxation process. One of these is referred to as the spin-lat-
tice relaxation process which is characterized by the time con-
stant Tl. This is also called the longitudinal relaxation process
as it describes the recovery of the net magnetic moment M to
its equilibrium value Mo along the axis of magnetic polarization
(Z). The Tl time constant is longer than T2, much longer in
most substances, and its independent measurement is the subject
of many gyromagnetic procedures.
The measurements described above are called "pulsed NMR
--2--
~za~3;~
measurements." They are divided into a period of excitation
and a period of emission. As will be discussed in more detail
below, this measurement cycle may be repeated many times to
accumulate different data during each cycle or to make the same
measurement at different locations in the subject. A varie-ty
of preparative excitation techniques are known which involve
the application of one or more excitation pulses of varying
duration. Such preparative excitation techniques are employed
to "sensitize" the subsequently observed free induction decay
signal (FID) to a particular phenomena. Some of these excitation
techniques are disclosed in U.S. Patent Nos. 4,339,716; 4,345,207;
4,021,726; 4,115,73n and 3,474,329.
Although NMR measurements are useful in many scientific
and engineering fields, their potential use in the field of
medicine is enormous. NMR measurements provide a contrast mecha-
nism which is quite different from x-rays, and this enables
differences between soft tissues to be observed with NMR which
are completely indiscernible with x-rays. In addition, physiolog-
ical differences can be observed with NMR measurements, whereas
x-rays are limited primarily to anatomical studies.
For most medical applications utilizing NMR, an imaging
technique must be employed to obtain gyromagnetic information
at specific locations in the subject. The foremost NMR imaging
technique is referred to as "zeugmatography" and was first pro-
posed by P. C. Lauterbur in a publication "Image Formation by
Induced Local Interactions: Examples Employin~ Nuclear MagneticResonance", Nature, Vol. 242, Mar. 16, 1973, pp. 190-191. Zeugma-
tography employs one or more additional magnetic fields which
have the same direction as the polarizing field Bo, but which
have a nonzero gradient. By varying the strength (G) of these
gradients, the net strength of the polarizing field Bo =
--3--
~Z~3Z~
Bz+GxX+GyY~G7Z at any location can be variedO As a result,
if the frequency response of the receiver is narrowed to respond
to a single frequency, WO~ then gyromagnetic phenomena, will
be observed only at a location where the net polarizing field
Bo is of the proper strength to satisfy the Larmour equation;
WO = y Bo: where WO is the Larmour frequency at that location.
By "linking" the resulting free induction signal FID with
the strengths of the gradients (G = Gx, Gy~ Gz) at the moment
the signal is generated, the NMR signal is "tagged", or "sensi-
tized", with position information. Such position sensitizingof the NMR signal enables an NMR image to be produced by a series
of measurements.
The series of free induction decay signals produced during
a scan of the subject are digitized and processed by a computer
lS to extract their various frequency components for display on
a screen. The most prevalent method involves the application
of a discrete Fourier transform to the digitized NMR signals.
Such transform may be in one or several variables as discussed
in "The Fourier Transform and Its Applications", by R. N. Brace-
wall, published in 1978 by McGraw-Hill. Computer programs for
performing such discrete Fourier transforms are well known,
as discussed in "Fourier Analysis of Time Series: An Introduction",
by P. Bloomfield, published in 1976 by Wiley. Two files of
digital data are produced by the Fourier transformation of the
time domain NMR signals. One file represents the "real" component
and the second file represents the "imaginary" component. As
discussed in U.S. Patent No. 4,070,611 it can be demonstrated
that the imaginary file is not required to reproduce an accurate
image of the NMR phenomena of interest, and it is common practice
to ignore this data.
The use of NMR to measure the flow of fluids in vessels
--4--
r h ~ d ~l
is well ~nown. A paper "The NMR Blood Flowmeter-Theory and
History" by J. H. Battocletti et al., published in Medical Physics,
Vol. 8, No. 4, July/August, 1981, describes the theory and history
of this effort. The techniques heretofore employed to measure
S flow require special NMR apparatus with coils arranged to magnet-
lze a sample of the fluid "upstream`' of the coils which are
employed to sense the FID signal. The physical distan~e between
this "tagging" coil and the sensing coil is known, and the level
of the FID signal provides velocity information in the direction
of fluid flow. In an article "NMR Rheotomography: Feasibility
and Clinical Potential", by J. P. Grant et al. and published
in Medical Physics, Vol. 9, No. 2, March/April 1982, imaging
techniques are employed to provide a flow intensity distribution
in a tubeO Such techniques are limited to measuring flow in
a known direction, and have been limited in practice to the
measurement of flow in inanimate objects or to the measurement
of blood flow in the arms and legs of animals.
The present invention relates to an NMR imaging apparatus,
and particularly, to a method and means for sensitizing the
NMR signals to provide not only the conventional NMR image,
but also to provide data from which a motion image can be con-
structed. In a gradient imaging NMR scanner the invention includes
the application of a motion sensitizing magnetic field gradien-t
(~) after the excitation portion of each measurement cycle and
prior to the emission portion of each cycle. The resulting
free induction decay signals which are produced by a series
of such measurements are processed by performing an inverse
Fourier transform to produce conventional image data mixed with
motion image data in the real and imaginary da-ta files. These
data files may be processed to produce spin~density images modu-
lated by conventional NMR phenomena, such as Tl or T2 relaxation,
_5_
or the motion data may be processed to produce an image of the
mo~ion alone, or the data files may be processed to produce
an image of conventional NM~ phenomena modulated by motion.
A general object of the invention is to measure the motion
of gyromagnetic material at any location within a subject. ~on-
ventional zeugmatographic scanners may easily be modified to
provide motion data along with image data. Such modifications
include the application of a motion sensitizing magnetic field
gradient F during the measurement cycle. The free induction
decay signal FID which is produced is "linked" to a position
and to the motion of the gyromagnetic material at that position.
The same processing employed on the image data alone can be
employed to construct a motion image.
Another object of the invention is to measure motion in
any direction at any position within the subject. A conventional
zeugmatographic scanner capable of exciting a gyromagnetic response
from a location within the subject is used to produce a motion
sensitized response. Motion sensitization is accomplished with
a magnetic field gradient F of alternating polarity which is
applied for a period 2T after the excitation portion of the
measurement cycle is completed. The direction of the field
(whose gradient is F) is the same as the polarizing field Bz,
but its strength is graduated in the x, y and z directions to
"tag" the flow data with a direction, as well as a magnitude
and position.
Another object of the invention is to provide motion image
data without significantly altering the NMR measurement cycle.
The motion sensitizing field gradient ~ can either be added
to each measurement cycle, or separate motion measurement cycles
can be inter]aced within standard NMR measurement cycles. In
either case the numerous preparative excitation techniques and
--6--
328;~:
emission measurement techniques known to the art can be carried
out with only minor modification. Independence of the motion
measurements is maintained by providing that the integral of
the field gradient F over the time period 2T is substantially
zero. Although the motion sensitizing field ~ is thus separate
and independent of the field gradient ~ used to position sensitize
the NMR signals, the two fields may be generated using the same
coils.
Fig. 1 is a schema~ic elevation view of an NMR scanner
which employs the present invention;
Figs. 2A-2C are graphic illustrations of gradient magnetic
fields produced in the scanner of Fig. l;
Figs. 3A~3C are perspective views of the gradient coils
which form part of the scanner of Fig. l;
Fig. 4 is an electrical block diagram of the control system
which forms part of the scanner of Fig. l;
Fig~ 5 is a graphic illustration of a typical conventional
measurement cycle performed by the scanner of Fig. l;
Fig. 6 is a graphic illustration oE a typical measurement
cycle performed according to the present invention;
Fig. 7 is a graphic illustration of a portion of an alter-
native measurement cycle according to the present invention;
and
Figs. 8A-8C are graphic illustrations of alternative forms
of the motion sensitizing field gradient ~ which may be employed
in the measurement cycles of Figs. 6 and 7.
Although the present invention may be easily implemented
in a variety of gyromagnetic scanner or NMR spectrometer struc-
tures, the preferred embGdiment of the invention employs a large
electromagnet to generate the polarizing field. Referring particu-
larly to Fig. 1, this polarizing magnet 1 is comprised of fou~
~A~
circular cylindrical segments 2-5 of sufficient size to receive
a table 6. A patient may be placed on the table 6 and any portion
of his body may be scanned by suitably positioning him with
respect to excitation coils 7. The polarizing magnet 1 produces
S a strong magnetic field Bz which is constant and homogeneous
within the space defined by the excitation colls 7. The excita-
tion coils 7 produce an excitation field Bl which is in the
transverse plane, perpendicular to the polarizing field Bz.
The excitation field Bl oscillates at a radio frequency WO and
it is applied as one or more pulses. The coils 7 are then switched
to a passive mode in which they operate as receivers for the
NMR signals produced in the patient's body.
Referring to Figs. 3A-3C, three sets of gradient field
coils are also formed around the table 6. A set of Z gradient
field coils lOa and lOb produce a magnetic field (Gz Z) which
is directed along the z axis of the machine, but which has a
strength that changes as a function of position along the z
axis. As shown in Fig. 2A, this field is additive to the polariz-
ing magnetic field Bz to provide a total field Bo which varies
in strength substantially linearly (i.e., Gz-Z) as a function
of Z position on the table 6.
Referring to Figs. 2B and 3B, a second set of gradient
field coils lla-lld produce a magnetic field (GX X) which is
directed along the z axis of the machine, but which has a strength
that changes as a function of position along the x axis. This
field is additive to the polarizing magnetic field Bz to provide
a total field Bo which varies in strength substantially linearly
as a function of x position on the table 6.
Referring to Figs. 2C and 3C, a third set of gradient field
3n coils 12a-12d produce a magnetic field (Gy-Y) which is directed
along the z axis of the machine, but which has a strength that
~328;~
changes as a function of position along the y axis. This field
is additive to the polari~ing magnetic field Bz to provide a
total field Bo which varies in strength substantially linearly
as a function of y position on the table 6.
The generation and control of the polarizing magnetic field
Bz and the field gradients Gx, Gy and Gz is well known in the
art and is employed in existing NMR scanners.
Referring particularly to Fig. 4, the control system for
the NMR scanner includes a set of four static power converters
15-18 which connect to an a.c. power source 19. he static
power converters 15-18 produce d.c. currents for the respective
coils 1, 10, 11 and 12 at levels determined by commands received
from a processor 20. The polarity, or direction, of the d.c.
currents produced for the gradient field coils 10-12 can also
be controlled. Thus, both the magnitude and the direction of
the gradient fields in the x, y and z direction can be switched
on command from the processor 20.
The excitation winding 7 is driven by a radio frequency
oscillator 21 when an electronic switch 22 is toggled to its
active position. The switch 22 is controlled by the processor
20 and when the switch 22 is toggled to its passive position,
the excitation winding 7 is coupled to the input of an amplifier
and phase-coherent detector circuit 23. The NMR si~nals in
the patient induce a voltage in the excitation winding 7 which
is amplified and demodulated in the circuit 23. The oscillator
21 provides a reference signal to the circuit 23 that enables
one phase-coherent detector therein to produce an in-phase,
or sine, free induction decay (FID) signal to an analog-to-digital
converter 24. A second phase-coherent detector produces an
orthogonal, or cosine, FID signal to an analog-to-digital con-
verter 25.
_g _
8Z
The free induction decay signals produced by the phase-coher-
ent detector 23 are digitized by the A~D converters 24 and 25.
The sample rate of this digitization is con~rolled by the processor
20, and the digital numbers which are produced by the A/D con-
verters 24 and 25 are input to the processor 20 and stored in
a memory 26. The processor 2G also stores values indicative
of the gradient field strengths at the moment the FID signals
are produced, and in this manrler, the FID signals are linked
to a specific position within the patient.
Referring particularly to Fig 5, a typical measurement
cycle for the NMR scanner in its imaging mode is illustrated.
Such measurement cycles are repeated many times during a single
scan, with the strengths of the field gradients Gx, Gy and Gz
being changed for each measurement to obtain the desired NMR
response from a series of points in the subject. In the example
cycle of Fig. 5 a first transverse excitation pulse 30 at the
desired Larmour frequency is applied and the field gradients
Gx, Gy and Gz are switched on at their desired levels. The
length of the excitation pulse 30 is selected to provide maximum
transverse magnetization (90) of the gyromagnetic nuclei, and
the resulting free induction decay signal 31 has an amplitude
Ao. The rate at which the FID signal 31 decays (as indicated
by dashed line 31') is a measure of the frequency distributions
of the gyromagnetic nuclei excited in the subject by the field
gradients (GXr Gy~ Gz).
To measure the T2 relaxation time within the same measurement
cycle, a second excitation pulse 32 is applied. This pulse
32 is at the same Larmour frequency, but it is twice as long
as the pulse 30, and phase shifted by 90, with the result that
the transverse magnetization is rotated 180. This "echo" pulse
stimulates the free induction decay signal 33 after the field
--10--
~LZ~;~Z~
gradients Gx, Gy and G~ are again applied. I~he peak value of
this FID siqnal 33 is less than the value Ao of the first FID
signal 31, and as indicated by dashed line 3~, it provides an
indication of the T2 relaxation time.
It should be apparent to those skilled in the art that
the NMR measurement cycle illustrated in Fig. 5 is but one of
many possi~le measurements that can be performed by the scanner
system of Fig. 4. With this particular cycle, a number of images
can be constructed which are of medical significance. Since
the measurement variables such as gradient field strengths,
and excitation pulse generation are under control of the processor
20, the NMR scanner system can be programmed to carry out any
number of different measurement cycles.
Referring again to Fig. 4, the digitized representations
of the FID signals generated during the complete scan are stored
in the memory 26 as two files Sl(t) and S2(t). Sl(t) is that
portion of the FID signal S(t) which is phase referenced to
the "cosine" phase of the transverse excitation signal produced
by oscillator 21, and S2(t) is the "sine" phase. Sl(t) and
S2(t) may be combined to form the complex signal,
S(t) = Sl(t) ~ iS2(t~. (1)
This may be written as the spectral transform:
S(t) = X~m(w)e dw, (~)
where:
w - w(Larmour) - Wrf~
and K is a constant electronic conversion factor. This signal
has been spacially modulated by the field gradient (G = Gxx
+ Gyy + Gzz) and is equivalent to the following:
S(t) =SM1(r)e i~G rtdr (3)
--11--
where: Ml = transverse magnetization
r = a position (x, y, z)
y = magnetogyric constant
Thls can be expressed in "q" space as:
S(q) = KJM1(r)e q dr (4)
where: "q" is a position in three-dimensional space
which is determined by the field gradient
G, q=~Gt.
Each measurement cycle thus produces a line sampling in
q-space, and the data files Sl(t) and S2(t) repr~sent a set
of such line samplings. An image Im~r) can be reconstructed
from this data by performing a numerical discrete Fourier inver-
slon to the desired yeometry:
Im(r) = K~H(q)S(q)e2 iq rdq (5)
15 Where: H(q) is the apodlzing function
associated with the digitizing
process.
When the Fourier inversion is performed according to equation
(5) by the processor 20, two data files are created, I(r) and
iJ~r), where:
Im(r) = I(r) ~ iJ(r) (6)
It is well known in the art that the image data in the
file I(r) may be output to a display device, such as the CRT
27 in Fig. 4, to produce an image. Such an image may represent
primarily the density of the excited gyromagnetic nuclei ("spin-den-
sity") or the image may be modulated by Tl or T2 factors to
provide improved contrast of the anatomical or physiological
phenomenon. The "imaginary" data file iJ(r) returns a null-value
-12-
3~
when the system is properly tuned, and it is usually discarded
in prior NMR scanner systems.
Although spin-density, Tl and T2 images provide useful
information of an anatomical nature, the present invention enables
a motion image to be produced. The flow of fluids in a human
subject is a most important phenomena, and its measure and imaging
provides diagnostic medicine with invaluable information for
functional assessment and physiological status. Although the
"motion-zeugmatographic" imaging method and system of the present
invention may be employed to image acceleration, jerk, etc.,
its primary value to medicine is believed to be in the production
of velocity images.
Referring particularly to Fig. 6, the present invention
may be implemented as part of a conventional NMR measurement
cycle. After a first free induction decay signal 40 is received
and digitized in the standard manner described above, a motion
sensitizing field gradient,
F = Fx~+ Fyy+ Fzz
is applied to the subject. This motion sensitizing field gradient
may be generated with the gradient field coils 10-12 (Fig. 3),
and it is characterized by thf? ~dCt that i L a 1 t:~?~ lt::-'S :i.ll E)~ t-.i ty
such that i~s ~ t ~ i 'S l ' Cl tl . l I t~ ro ( ~ V ~ ? l .i t ' - t i ~ )C? .?;'- i o~ ¦
2'['.
¦F( t)dt = O
l~? t. i,? l.U t'. i. V~` V~:l L. U~? 'i vf tll~? coor~tinilt~? con~ponelts Fx, Fy~
F'~ ctetermine the clilection in wl~ich the su?bscquent NMR signal
41 is motion sensitizecl.
The me?asurement cycle illustrated in Fig. 6 may be repeated
many -tlmf?s to motion sensitizf? a series oE NMR signals 41 in
many d:ir~?ctions. 'L'hfa s~t oE motion sensitizf?d data whicn results
32~12
from this series of "F cycles" is stored, and the gradient fields
Gx, G~ and Gz are then changed to position sensitize the next
series of F cycles to a different location in the subject. The
process is continued with a series of motion sensitized measure-
ments being made at each location in the scan.
It is a requirement of the present invention that the motionsensitizing field gradient F be applied after the application
of an excitation field which produces a transverse magnetic
~oment Ml. Furthermore, motion sensitization must occur prior
to the emission of the FID signal which it is to sensitize.
In the example measurement cycle of Fig. 6, excitation pulse
42 produces the required transverse magnetic moment Ml. The
flow sensitizing field gradient F is applied after the first
free induction signal 40 is produced, and hence the data which
is collected from the FID 40 is not motion sensitized. The
FID signal 41 on the other hand, is produced by "echo" excitation
pulse 43 after the motion sensitizing field gradient F has been
applied. It contains motion information. In this example the
echo excitation pulse 43 does not produce any additional trans-
verse magnetic moment Ml.
Another possible measurement cycle whlich produces motionsensitized data is illustrated in Fig. 7. In this cycle a 9~
excitation pulse 45 is applied to produce maximum transverse
magnetic moment and the position gradient field G is later switched
on to produce the free induction decay signal 46. At time TA
during the generation of the FID 46, a motion sensitizing field
gradient F is applied until time TB. The FID signal 46' (generated
after TB) is motion sensitized.
The direction of the motion which is measured is determined
by the direction of the field gradient F. The measurement sensi-
tivity is determined by a number of factors, including the strength
-14-
31.~;~3Z~3f~
(Fo) of the Eield gradient F and its duration (2T). If the
systematic phase errors produced by the system are denoted by
"e", then the minimum velocity which can reliably be measured
is as follows:
Vmin ~ e[2~yFOT ] (7)
where e is in radians.
For example, if the motion oE hydrogen nuclei is measured
with a system having a phase resolution error of e=.l radian,
then the following conditions are typical:
y ~ 4.6 X 103 EIz/gauss
T = 10 msec.
Fo = .5 gauss/cm.
Vmin ~ .06 cm./second
It should be apparent that the measurement process can
be shortened and simplified considerably if fluid flow in only
direction is imaged. For example, if only the motion sensitizing
field gradient Fz is employed, an average velocity image of
fluid flow along the z axis is generated. In such case, only
one flow sensitized measurement is required at each "G" position
of the scan.
The flow sensitized FID signal is both position and motion
sensitized. As indicated above in equation (3), the digitized
FID signal S(t) which is stored in the memory 26 is linked to
position by the gradient field ~. Similarly the motion sensitized
FID signal S~t) is linked to the velocity of the spin-density
at this same position by the motion field gradient F:
S(t) = ¦~Ml(r,v)e2~ G-rt+F-V(~)2]d ~8)
If a six-dimensional discrete Fourier inversion is performed
on this stored data file, an image ~(r,v) can be constructed
-15-
3~
on the CR~ 27 which displays spin-density (Po) distributed accord-
ing to the proportions of that density possessing particular
velocities in the direction F of the gyromagnetic nuclei at
that location.
~ ~(r,v) = K~ H(q~H(f)S(q f)e2~i(q r~f-v)d df (9)
where: H(q) and H(f) are apodizing functions associ.ated with
digitizing the FID, and "f" is a position in three-dimen-
sional Fourier-volocity space which is determined
by the velocity field gradient Fl where f~ F(T2).
This six-dimensional image, ~(r,v), is the most general
and ambitious direct image of true flow velocities since lt
enables many points in "f" space to be measured by sensitizing
a series of FID signals with velocity gradients F having different
directions and different magnitudes. The technique can be consider-
ably simplified if the image is modulated by a single velocity
gradient F at each point. This considerably shortens the data
collection portion of the process since it requires only one
F cycle for each ~ cycle, but it returns only the average velocity
of flow rather than a complete velocity distribution scale,
or profile.
There is an endless variety of modifications and simplifica-
tions by which the motion-~eugmatographic phase-modulation method
of the present invention can be applied to studies of practical
importance. Chemical shift distributions or T1 and T2 spectro-
scopy may be added and similarly "interlaced" in the data-collection
cycle. The B1 excitation field may be modulated to suppress
or isolate contributions to the T2-process, and as will be described
in more detail below, variations in the shape of the motion
sensitizing field gradient ~ are possible.
As indicated above (equation (5)), the Fourier inversion
-16-
2~
performed on a conventional ~eugmatographic NMR scanner returns
data in a "real." file I(r) from whi.ch an image of spin-density
PO(r) can be produced on the CRT 27. As indicated above by
equation (9), when the Fourier inversion is performed on motion
sensitized NMR data, the real file I(r) is returned with data
which enables an image of spin-density PO(r) modulated by velocity
V(r) to be produced on the CRT 27.
It is another aspect of the present invention that when
the Fourier inversion of flow sensiti~ed NMR data is performed,
the "imaginary" data file iJ(r) returns information from which
images of particular medical value can be produced. More specif-
ically, if the NMR data is flow sensitized in a single direction
(i.e., one ~ cycle per ~ cycle), and if the magnitude (Fo) and
duration (2T) ~f the flow sensitizing fi.eld gradient field are
kept small such that:
ei2~T FO-V~l ~ i2~yT2Fo-~7 (lo)
the real file I(r) returns conventional image data Po(r). How-
ever, the imaginary file iJ(r) now returns image data:
J(r) = (2~yT F0)-[pO(r)(V)] (11)
The values in parentheses are known measurement conditions and
Po(r) is precisely the set of values returned in the real file
I(r). Consequently, a velocity image V(r) can be produced on
the CRT 27 as follows:
V(r) - J(r)/[(2~yT FO)I(r)l (12)
Thus in a single scan of the subject, data files Sl(t) and
S2(t) can be created and stored in the memory 26. From the
files I(r) and iJ(r) which are produced by the Fourier inversion
of these data files, three separate images can be produced with
-17-
minimal computation. The first image Po(r) is the conventional
spin-density NMR image as modulated by phenomena such as Tl
and T2. The second image is the same spin-density image pO(r)
modulate~ by the magnitude of average spin velocity V(r) in
the direction selected by the flow sensitizing field gradient
F. The third image is the magnitude of spin velocity V(r~ in
the direction of F throughout the region of the NMR scan.
It should be apparent to those skilled in the art that
other images of medical value can be constructed from this measured
data with further computation. For example, the exchange flow
of molecules into or out of a specified volume may be calculated
by integrating the velocity modulated spin-density values ~(r,v)
over the surface area of the volume. The same flow rate through
a specified plane may also be calculated by integrating across
the surface of the plane. Such measurements may provide, for
example, the quantity of blood flowing through a specific vessel.
It is important to note that because an anatomical image may
be produced from the same data, the location (~) of the particular
volume or surface of interest in the patient can be precisely
located by the NMR scanner operator.
While the most important applica-tion of the invention is
presently believed to be the measurement of velocity, the inven-
tion may be extended to measure "higher order" motion such as
acceleration. For velocity sensitization, the flow sensitizing
field gradient ~ must not only be of alternating polarity, but
its wave form should by symmetrical. That is, the velocity
sensitizing field gradient F should be a mirror image about
horizontal and vertical axes of symmetry. Referring to Fig.
8A, for example, the field gradient wave form 50 alternates
in polarity and is anti-symmetrically mirrored about an axis
of symmetry 51. Needless to say, the integral of this wave
-18-
~2V3~
form 50 over the interval TA to TB is zero and it therefore
satisfies the baisc requirement for motion sensitization. In
contrast, the motion sensitization gradient field F' produced
by the wave form 52 in Fig. 8B is symmetrically mirrored about
the vertical axis 51. This wave form will sensitize the subse-
quent NMR signal to acceleration. Note that the integral of
the wave form 52 over the time period TA to TB is zero, thus
satisfying the basic motion sensitization requirement.
It should also be understood that the alternating polarity
requirement for the motion sensitizing field gradient F is refer-
enced to the gyromagnetic nuclei - not the table 6. Thus, if
a 180 echo pulse of excitation energy is applied to the gyro-
magnetic nuclei, their phase-po]arity is effectively reversed
and the second half-cycle of the motion sensitizing field gradient
~ need not be reversed in polarity. This is illustrated in
Fig. 8, where the wave form 53 of the field F is the equivalent
of the wave form 50 in Fig. 8A, when an echo pulse 54 is generated
at the axis of symmetry 55. This technique may be useful with
systems which do not enable the gradient fields to be reversed
in polarity.
-19-
