JPS6351850A - Nuclear magnetic resonance imaging method - Google Patents

Abstract

(57) [Abstract] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

DETAILED DESCRIPTION OF THE INVENTION [Industrial Application Field] The present invention relates to nuclear magnetic resonance imaging for obtaining chemical shift images of two components, and particularly to nuclear magnetic resonance imaging that can remove chemical shift artifacts.

[Prior Art] In general, nuclear magnetic resonance (NMR) is a phenomenon in which, when a certain atomic nucleus is placed in a static magnetic field, the nuclei move around the axis at a frequency proportional to the strength of the magnetic field. This is due to the fact that they precess. This frequency is known as the Larmor frequency and ωG = γH.

However, γ: the gyromagnetic ratio of the atomic nucleus; Ho: given by the strength of the magnetic field. In other words, if we apply a static magnetic field whose strength varies along a particular direction, the atomic nuclei at each position in that direction will precess at different frequencies. Therefore, the object has a gradient magnetic field (Badient
When a sufficiently strong high-frequency pulsed magnetic field is applied at the same time as a magnetic field), only the atomic nuclei whose spins precess at a frequency equal to that of the high-frequency pulse are rotated through 90° or 180°, and other can be isolated from the nucleus of the atom.

Next, a method for obtaining a tomographic image of a human body using this phenomenon will be explained.

Figure 3 shows, for example, A.
es) etc. in the American Journal of Radiology.
138, page 206, published in 1982, Vol. 138, p. 206 of ``Iology'', 1982, is a simplified block diagram showing a general nuclear magnetic resonance imaging apparatus, with a partial side view.

In the figure, (1> is a magnet, (2) is a human body as an object lying in the static magnetic field of magnet (1), and (3) is a human body.
2) is a high frequency coil wound around, (4) is a transceiver for transmitting electromagnetic waves to the high frequency coil (3) and receiving electromagnetic waves from the human body (2), and (5) is a magnet (1). and the high-frequency coil (3), the gradient magnetic field coil (6) is a gradient magnetic field coil power source for the gradient magnetic field coil (5), and (7) is the gradient magnetic field coil power source (6).
) and a control circuit that controls the transceiver (4), (8) is a computer connected to the control circuit (7), and (9) is a computer (8).
an image display connected to the

Next, the operation of the nuclear magnetic resonance imaging apparatus shown in FIG. 3 will be explained.

First, a uniform static magnetic field is applied to the human body (2) by a magnet (1), and an electromagnetic wave corresponding to the Zeeman energy of a specific atomic nucleus in the human body (2) is transmitted from the transmitting section of the transceiver (4) to a high-frequency coil (3). ).

This electromagnetic wave causes a specific atomic nucleus within the human body (2) to undergo a resonant transition from a ground state to an excited state.

Then, the irradiation of electromagnetic waves is stopped, and the electromagnetic waves emitted from the atomic nuclei in the human body (2) are detected by the receiving section of the transceiver (4) through the high frequency coil (3). The transceiver (4) has a built-in ^D converter for reception, and receives the magnetic resonance signal from the high frequency coil (3) according to a predetermined sampling frequency. At this time, by applying a gradient to the static magnetic field using a gradient magnetic field coil (5), it is determined from which position of the human body (2) the signal is coming.

On the other hand, the computer (8) controls, via the control circuit (7), the gradient magnetic field coil power source (6) and the transceiver (4) for supplying current to the gradient magnetic field coil (5), and performs fast Fourier transformation. The resulting image is displayed on the image display (9). The details of nuclear magnetic resonance imaging using Fourier transform are described in, for example, British Patent No. 2,079,946, so they will not be described here.

Next, a conventional nuclear magnetic resonance imaging method for obtaining a chemical shift image using the nuclear magnetic resonance imaging apparatus shown in FIG. 3 will be described. Note that chemical shift is a slight splitting of a nuclear spin resonance spectrum line, and for example, in the case of protons, the chemical shift difference between water and fat is 3.5 pp-.

The fourth country is, for example, the 1st edition of the 7th Nuclear Magnetic Resonance Medicine Research Conference Conference Abstracts (N14R Medicine, Vol. 6, Supplement-1, 1986) published by the Nuclear Magnetic Resonance Medicine Research Society.
34 is a pulse sequence diagram showing a general nuclear magnetic resonance imaging method for obtaining a chemical shift image, described on page 34. FIG. The pulse sequence diagram in Figure 4 is based on the first part of the above document.
Although not shown on page 34, this is a summary of its contents.

In the figure, G x , G y , and G z are gradient magnetic fields in three axes X, Y, and Z directions that are orthogonal to each other, and the Z axis is the thickness direction of the object. RF is high frequency pulse, S is NMR
The received signal is a spin echo signal. Also, for example, G
z L' represents "gradient magnetic field Gz in the a-th section".

Below, while referring to the flowchart in FIG.
A specific sequence operation of the conventional magnetic resonance imaging method using the pulse sequence shown in the figure will be explained.

First, gradient magnetic field QX==QX for frequency encoding.

(positive polarity), the first image m1 (x, y) is acquired in the first to fourth sections (step (11)).

J-Ward W Apply a 90° high-frequency pulse RF'' along with a gradient magnetic field Gz'' to the object. At this time, another gradient magnetic field Gx”
' and Gy"' are zero. Therefore, nuclear spins with a predetermined thickness in two directions are excited depending on the frequency of the 90° high-frequency pulse RFL11. This thickness is Frequency or gradient magnetic field G z ''
This can be changed by changing the amplitude of.

4th Section - Inquiry - In order to obtain a spin echo signal S (4゜) in the fourth section, which will be described later, by setting the gradient magnetic field Q zI 2 ] to zero, a positive gradient magnetic field Gx"' (-G xo) for phase correction is applied. As a result, the spins are dephased (the phase is disturbed) along the X direction.At the same time, in order to provide position information in the Y direction as phase information of the received signal S, a gradient magnetic field Gy'' is applied. This is called a phase encode, and this phase encode itv is equal to the time integral value of the gradient magnetic field G, + 2l.

-yth (interval) While maintaining the gradient magnetic fields Gx" and GET" at zero,
After a time τa has elapsed since the application of the 0° high-frequency pulse RFLl', a pulse 1 with an amplitude twice that of the 90° high-frequency pulse RF'' is applied.
An 80° high frequency pulse Rp11 is applied. At this time, the area of the shaded part of ZL11 is added to the gradient magnetic field, and the area of z11 is added to the gradient magnetic field.
The sum of the area up to the left hatched part, that is, the timing at which the 180° high-frequency pulse RF'3'' reaches its peak, is the gradient magnetic field G.
It is equal to the area of the hatched part on the right side of "z".

14 - While maintaining the gradient magnetic field %4] and y14] at zero, the positive gradient magnetic field Gx '4' (
-Gx. ) is applied while receiving the spin echo signal S34'.

The area of the shaded part of the gradient magnetic field Gx” applied at this time is:
It is equal to the area of the shaded part of the gradient magnetic field Gy, +2l.

Therefore, the spins dephased by the gradient magnetic field Gx'' are corrected to zero and imaged (rephased) at the timing when the spin echo signal S34 reaches its peak.
The timing at which the spin echo signal S il+ reaches its peak is 18 hours after the peak of the 180° high frequency pulse RF"'.

In this way, the received signal S consisting of a nuclear magnetic resonance signal is received as a spin echo signal S″′, but at this time, since the gradient magnetic field G has as frequency information.

Hereinafter, the first section to the fourth section are sequentially repeated to obtain a plurality of spin echo signals 344+, but only the second section is different in each sequence.

That is, for each sequence, the time integral value V of the gradient magnetic field G for phase encoding (2') is changed each time.The magnitude of the gradient magnetic field ay is shown by a solid line for the first time, and For example, by performing the following two-dimensional Fourier transform on the 128 spin echo signals S'41 obtained by repeating this 128 times with respect to the time U and the phase encode amount ν, First image m1
(x,y) can be formed.

S”(u,v)−iS+n1(x,y)exp[j7
(GxXu+yv)]dxdy For details on the two-dimensional Fourier transform, please refer to "Bipedal Literature" (British Patent No. 2079)
No. 946 specification), so it will be omitted here.

Subsequently, the same pulse sequence operation as in the first section to the fourth section described in F is repeated using a gradient magnetic field Q x - G X o (negative polarity) to obtain a second image m2 (x, y) ( Step (12)).

The difference in the pulse sequence in step (12) is as follows:
The value of the gradient magnetic field G in the second and fourth sections is -
It's just that it has Gxo.

Next, a chemical shift image L(x, y) of water and a chemical shift image l~(image f1(x, y) of fat are obtained (step (13))
.

First, the density distribution image of water is expressed as ``(x, yL).If the density distribution image of fat is f1(x, y), and the angular frequency corresponding to the chemical shift frequency Δr of water and fat is c(=2πΔ''), , first and second images m1 (x, y), m2 (x, y
) is II (x, y) = f0(x, y) ten f
+ (x+c,y)m2(x,y)-f0(x,y)
It is expressed as +f+(x-c, y).

By performing Fourier transform on the above equation to solve the simultaneous equations, and further performing inverse Fourier transform, the chemical shift image of water and fat ". (x, y), f1 (x, y) is f0 (x, y) =
g1(x,y)? -'[(cosuc/j-sinuc)G2(u,v
) co=lit(x,y)-7'-1[G2(u,v)/
j4anucl-=■f1(x,y)=? -'r(G
2(u,v)/j-sinuc)] −■. However, x and y are spatial coordinates, U is a time coordinate, and V is a coordinate of the gradient magnetic field Gy. Also, g1 (x, y) is the first,
The second sum image, g2(x,y), is the difference image, and g1(
x, y) - [m1 (x, y) + m2 (x, y)]/2
−・・0g2(x,y)−4m1(x,y)−m
It is expressed as 2 (x, y) 1/2 -= (2).

Further, A1 is an operator of inverse Fourier transform, and the time series G2(u,v) for determining the sampling point n is obtained by Fourier transform of the difference image g2(X,)l). That is, the difference image g2(x,y) is expressed as g2(x,y)-A1[G2(u,v)] by inverse Fourier transform of the time series G2(u,v).

According to formula 0~0, each image m1(x,y) and m2(x
, y), and after a predetermined calculation, U
By performing inverse Fourier transformation on and ν, chemical shift images of water and fat f. and f are obtained.

Figure 6 Ca> to (c) are explanatory diagrams showing an example in which cross-sectional images of protons were obtained by arranging two test tubes containing water and one test tube containing fat in a row. a) is the first image m
1 (x, y), (b) is the image ii when the gradient magnetic field Gx is set to negative polarity; (x, y), (c) is the second image m2
(x, y), since the gradient magnetic field Gx in the X direction is used for frequency encoding, the chemical shift images f0 and f1 of water and fat are That is, it deviates by about 3.5 ppm.

For example, when the static magnetic field strength is 1.5 T (Tesler 10,000 Gauss), the resonance frequency of protons is about 64 N II z, so 64 MtlzX 3.5 ppm = 224 Hz, and the fat image f is approximately 22
It will be shifted by 4Hz.

At this time, if 10011z is assigned to one pixel, this means that there is a shift of two pixels in the X direction on the image.

In FIG. 6(a) showing the first image m1 (x, y) containing chemical shift images f, , f of water and fat, the axis of the frequency f is the X axis, and the gradient magnetic field G -G X6
Therefore, f−γG×. The relationship X/2π holds true.

In addition, in Fig. 6(b) showing the chemical shift image ii'z (x, y) including 0 and f1 when the gradient magnetic field Gx = Gxo, it becomes r-r (G xo)X 72r. Since the direction of the X-axis is reversed, the image i(x,y) moves to the right side of the screen.However, since the chemical shift frequency f remains unchanged, the chemical shift image f1 of fat is about 22
4 Hz resonance frequency is low. Therefore, the chemical shift image f of fat is the chemical shift image f of water. Shift 1- to the lower frequency side (to the left).

The true second image m2 (x, y) is obtained by horizontally inverting the image mi; (X, y) in FIG. 6(b), as shown in FIG. 6(c).

Incidentally, the inverse Fourier transform of equation (2) and 0 in step (13) is performed at an arbitrary sampling frequency fs ('150 kl!
t, ), the processing is performed discretely according to the timing of the sampling point n determined by the time series a2(u,v).

However, for the time U that gives 5inuc=O, since it is a division by zero in the equations (■ and 0), 5inuc=0.
Approximately calculate as 01.

[Problems to be solved by the invention] As described above, in the conventional nuclear magnetic resonance imaging method, in order to prevent division by zero, each chemical is Since the shift image r0, f was obtained, the chemical shift image f. There is a problem in that band-like false images called artifacts occur in the image and fl.

This invention was made to solve the above-mentioned problems, and by appropriately selecting sampling points (calculation points), it is possible to perform nuclear magnetic resonance imaging without causing artifacts in chemical shift images. The purpose is to obtain.

Means for Solving Problem E] The nuclear magnetic resonance imaging method according to the present invention includes a first step of setting the sampling frequency "S" to fs=2lΔf/I11, and a step of setting the sampling frequency "S" to fs=2lΔf/I11, and After performing a predetermined operation, Fourier transform is performed to obtain a time series G having a time shift view δ.
2(u-δ, v), and an image obtained by performing a predetermined operation on the sum image g1 (x, y) and the time series G2 (u+δ, v) and then performing inverse Fourier transform. from,
and a sixth step of calculating chemical shift images for two formulas.

[Operation] In this invention, the sampling frequency is set to a value according to a predetermined rule, and the time series for determining sampling points has a time shift amount, thereby eliminating the need for approximate calculations, Obtain a chemical shift image of the formula.

[Example] Hereinafter, an example of the present invention will be described with reference to the drawings. 1st
The figure is a flowchart 1 showing the signal processing operation of an embodiment of the present invention, and steps (S2) and (S3) are as follows:
This step is similar to (11) and (12) in FIG. Further, FIG. 2(a) is an explanatory diagram showing sampling points when the time shift amount δ is zero, and FIG. 2(b) is an explanatory diagram showing sampling points in an embodiment of the present invention.

Note that the general nuclear magnetic resonance apparatus shown in FIG. 3 may be used as an apparatus for carrying out the present invention, and the apparatus is capable of controlling the transducer in the transmitter/receiver (4) and calculating the received signal S. It is only necessary that the program in the device (8) is changed. Therefore, for example, a magnet (1) applies a static magnetic field Ho along a direction perpendicular to an arbitrary tomographic plane of the object, that is, two directions (thickness direction), and also applies a gradient magnetic field Gx, as appropriate, in the x, y, and z directions.
Gy, G7. is applied, and the X axis (or Y axis) perpendicular to the Z axis
It transmits and receives high-frequency pulses along the Further, the pulse sequence used in one embodiment of the present invention is similar to that shown in FIG.

Further, as a received signal, that is, a spin echo signal S, both the e08 component and the sine component having a phase difference of 90" are received, and QD (QuadraLion Detection) is performed.
On) method is used to perform signal processing.

Next, the operation of one embodiment of the present invention will be described with reference to FIGS. 1 to 4 and FIGS. 6(a) to (c).

First, a transceiver (
4) 〇 converter sampling frequency 1゜, chemical shift frequency Δf, sampling point n and arbitrary integer m
, 1, rs=2lΔf/m where Δf: chemical shift frequency mn/I: set to be an integer (step (St)).

Furthermore, based on the pulse sequence shown in FIG. 4, using the Fourier transform method, the first image m1 (x, y
) is obtained (step (SZ)), and a second image m2 (x, y) when the gradient magnetic field Gx=Gxo is obtained as shown in FIG. 6(c) (step (S3)).

The first and second images TRI(X,y) thus obtained
, m2(x,y) chemical shift image f of water and fat
0, f1 is the chemical shift frequency Δf (=
3.5 ppm>, resulting in a deviation of 224 If z with respect to the proton resonance frequency.

Note that the first and second images m1 (x, y) and m2 (x, y
) is displayed as an absolute value to remove phase differences due to sequence timing errors.

Next, the sum image g1 (x, y) and the difference image g2 (x, y)
, g1 (x, y) - [m1 (x, y) + n + 2 (x,
y)]/2g2(x,y)=[m1(x,y)−+
n2(x,y)]/2 to calculate the step (S4)
), further, the time series G2 (u + δ, v) with the time shift amount δ (= 1/2 fs) is expressed as G2 ('10 δ, V) 1 [g2 (x, y) exp (-jx δ)] = 55g
2(x,y)exp←jxδ)exp[-j(ux+v
y)]dxdyHowever,? : Obtained from a two-dimensional Fourier transform operator (step (S5)).

Next, the chemical dift image f of water. (x, y) and the chemical shift image L(X, y) of fat, f0(x, y)' = g1(x, y)-l'-1[G2(11+
δ, v)/j・tan(u+δ)c month f1(x, y)
−|■″−' [G2(u+δ,v)/j-sin(u
+δ) cll (step (S8)). In this case, since the time coordinate (rooster δ) is shifted by δ, the phase rotation of the image due to inverse Fourier transform is corrected by absolute value processing.

By the way, the inverse Fourier transform in step (S6) is performed discretely at the timing of a plurality of sampling points n, that is, u=n/fs (where n=o, ±1, ±2, . . . ). ing.

Also, in step (Sl), the sampling frequency f
s is set in advance as fs=2l8f/m (where 1 is an integer) for the chemical shift frequency Δf (=c/2π), and for the sampling point n consisting of mn/l is set to be an integer. Therefore, uc= (n/ fs)42rΔf = [n/(2lΔf/m)]4rΔf −”mnπ/1, and for n where mn/l is an integer, 5inuc=
It can be seen that o.

Figure 2 (a) shows 1-6, m=1, time shift amount δ-0
It is an explanatory diagram showing sampling point n when n=
It is shown that 5inuc=o periodically at each point of o, soil 6, ±12, ±18, . . . .

In fact, when the static magnetic field strength is 1.5T, the sampling frequency is fs-50kHz, and the chemical shift frequency is f
Since both are 224Hz, the values of 1 and m are, for example, l1
1-1.1 = 116, sampling frequency f
It is set to s=51.97kllz.

FIG. 2(b) is an explanatory diagram showing sampling points n when the time shift amount δ of the time series G2 (u+δ, v) is set to 1/2 fs according to an embodiment of the present invention. In this case, the time shift amount δ is set to half of the sampling (calculation) point n, 1/'s, that is, |
+δ)c≠0 and jan(u+δ)c≠O. Therefore, the calculation of each chemical shift image r0, f1 in step (S6) does not include division by zero.

Note that in the above embodiment, the first and second images, (to, y),
Although l112(x, y) is expressed as an absolute value, it may be displayed as a real number by correcting the phase distortion.

Further, although absolute value calculation was used to calculate water and fat chemical shift 1 to image 0 and f1, the phase may be corrected by multiplying exp (j to δ).

In addition, we have explained the case where the two-dimensional Fourier transform method is used for a thin tomographic plane of the subject, but the three-dimensional Fourier transform method is used for a thick volume.
It goes without saying that the same effect can be obtained even if the dimensional Fourier transform method is applied to perform volumetric imaging.

Furthermore, chemical shift image f of water. and chemical shift image of fat f
By adding 1, a composite image of water and fat can be constructed, so a proton image without chemical shift artifacts can be obtained.

[Effect of the invention] As described above, according to the invention, the sampling frequency f
After setting s to fs=2lΔf/m and performing a predetermined operation on the difference image Fi2(x, y), the time series G2(+uδ, v
), and a step of calculating a two-component chemical shift image from an image obtained by performing predetermined operations on the sum image g1 (x, y) and the time series G2 and then performing inverse Fourier transformation. , since approximate calculations are not required, nuclear magnetic resonance imaging can be obtained without artifacts in chemical shift images.

[Brief explanation of drawings]

FIG. 1 is a flowchart showing the operation of an embodiment of the present invention, FIG. 2(a) is an explanatory diagram showing sampling points when the time shift amount is zero, and FIG. An explanatory diagram showing sampling points in one embodiment, FIG. 3 is a block diagram showing a partial side view of a general nuclear magnetic resonance imaging device, and FIG. 4 is a diagram for explaining a general nuclear magnetic resonance imaging method. A pulse sequence diagram, FIG. 5 is a flowchart diagram showing the conventional chemical shift image acquisition operation, and FIGS. 6(a) to (
c) is an explanatory diagram showing a general chemical shift image, Fig. 6(,) is the first image when the gradient magnetic field Cx is positive polarity, and Fig. 6(b) is the first image when the gradient magnetic field Gx is negative polarity. 6(c) shows a second image obtained by inverting the image in FIG. 6(b). 8... Received signal Is... Sampling frequency n... Sampling point (calculation point) δ... Time shift amount Δf... Chemical shift frequency C... Chemical shift angular frequency m1 (x, y) - First image m2 (x, y)... Second image g1 (x, y)... Sum image g2 (x,
y)...Difference image f,...Chemical shift image of water f,...Chemical shift image of fat Gx...Gradient magnetic field for frequency encoding Gx0...Magnitude of gradient magnetic field for frequency encoding GV・... Gradient magnetic field V for phase encoding ... Gradient magnetic field coordinates G2 (u + δ, v) for phase encoding ... Operator of time series A-Fourier transform ... Operator of inverse Fourier transform (Sl) ...First step (Sl)...Second step (S3)...Third step (S4)...
・Fourth step (S5)...Fifth step (S6
)...Sixth step In the figures, the same reference numerals indicate the same or corresponding parts.躬2 s (b) 范5 s 6 s (a) (b) (C) しf

Claims (11)

[Claims] (1) In nuclear magnetic resonance imaging to obtain two-component chemical shift images f_0 and f_1 having a chemical shift frequency Δf,
For arbitrary integers m and l and a predetermined sampling point n, the sampling frequency fs of the received signal is expressed as fs=2lΔ
f/m and mn/l: a first step of setting an integer, a second step of obtaining a first image m_1 by Fourier transformation by setting the gradient magnetic field for frequency encoding to positive polarity, and setting the gradient magnetic field to an integer. As a negative polarity, the second
a third step of obtaining an image m_2, and a sum image g_1 and a difference image g_ of the first and second images m_1 and m_2;
2, and after performing a predetermined operation on the difference image g_2, Fourier transform is performed to calculate the time shift amount δ.
a fifth step of calculating a time series G_2 having a time series G_2; and a fifth step of calculating a time series G_2 having the above 2
A nuclear magnetic resonance imaging method comprising: a sixth step of calculating chemical shift images f_0 and f_1 of components. (2) In the second and third steps, the first and second
2. The nuclear magnetic resonance imaging method according to claim 1, wherein the images m_1 and m_2 are obtained by absolute value display. (3) In the second and third steps, the first and second
2. The nuclear magnetic resonance imaging method according to claim 1, wherein the images m_1 and m_2 are obtained by displaying real numbers after phase correction. (4) In the fifth step, the time shift amount δ is 1
The nuclear magnetic resonance imaging method according to any one of claims 1 to 3, wherein the nuclear magnetic resonance imaging method is set to /2 fs. (5) Claims 1 to 4 characterized in that a two-dimensional Fourier transform method is used as the Fourier transform method.
Nuclear magnetic resonance imaging according to any of paragraphs. (6) In the fifth step, the frequency encoding coordinate is X, the phase encoding coordinate is Y, and the first image is m_1(
x, y), and the second image m_2(x, y), the time series G obtained by Fourier transformation of the difference image g_2(x, y)
Time series G with time shift amount δ for _2(u,v)
_2(u+δ,v) is G_2(u+δ,v)=■[g_2(x,y)exp(
−jxδ)] However, g_2(x, y)[m_1(x, y
)−m — 2 (x, y)]/2. The nuclear magnetic resonance imaging method according to claim 5, wherein the nuclear magnetic resonance imaging method is calculated from . (7) In the sixth step, the two-component chemical shift images f_0 and f_1 are obtained by setting the chemical shift angular frequency to c, and f
_0 (x, y) = g_1 (x, y) −｜■＾−＾1[G_2(u+δ, v)/j・tan(
u+δ)c] |f_1(x,y)=|■＾−＾1[G_
2(u+δ,v)/j·sin(u+δ)c]| The nuclear magnetic resonance imaging method according to claim 5 or 6, characterized in that it is calculated from |. (8) In the sixth step, the chemical shift images f_0 and f_1 of the two components are expressed as f_0(x,y)=g_1(x,y)−exp(jxδ
)■＾−＾1[G_2(u+δ,v)/j・tan(u
+δ)c] f_1(x,y)=exp(jxδ)■＾−＾1[G_
2(u+δ,v)/j−sin(u+δ)c] The nuclear magnetic resonance imaging method according to any one of claims 5 to 7, characterized in that the calculation is performed from the following. (9) In the sixth step, the inverse Fourier transform is performed discretely at calculation points u expressed by u=n/fs for any integer n. The nuclear magnetic resonance imaging method according to any one of items 1 to 8. (10) The nuclear magnetic resonance imaging method according to any one of claims 1 to 4, characterized in that a three-dimensional Fourier transform method is used as the Fourier transform method. (11) The nuclear magnetic resonance imaging method according to any one of claims 1 to 10, wherein in the sixth step, a proton image is obtained by combining two component chemical shift images.