JPS61194339A - Nuclear magnetic resonance image pick-up apparatus - Google Patents

Abstract

PURPOSE:To obtain an image reduced in artifact by the re-constitution of a Fourier method using a FID signal even if there is spacial non-uniformity in a main magnetic field, by using two kinds of gradient magnetic fields equal in magnitude but different in an inclination direction. CONSTITUTION:Two gradient magnetic fields equal in magnitude but mutually different in the inclination direction of a magnetic field gradient to measure the FID signals of the same object (a). A zero value is embedded in the negative time region of the FID signals and, corresponding to the inclination direction of the magnetic field gradient at the measuring time of FID signals, one complex image issubjected to Fourier conversion as it is and the other is subjected to Fourier conversion so that a reading time axis receives positive and negative reversal. Further, the position and density strain correction in a reading direction is applied to the output from a Fourier converter means corresponding to preliminarily calculated main magnetic field non-uniform distribution and two complex images are added and the added complex image is subjected to image processing to obtain an actual number image.

Description

Detailed Description of the Invention (Industrial Application Field) The present invention relates to a nuclear magnetic resonance imaging apparatus (hereinafter referred to as nuclear magnetic resonance).
In particular, FID (Free Induct)
NMfl is an improved reconstruction method that reconstructs images using the Fourier method using ion [)edcay) signals.
ll! This invention relates to an i-image device.

(Prior Art) Conventionally, there is a Fourier imaging method using the Fourier method as a method for obtaining images in a nuclear magnetic resonance imaging apparatus. The NMR signal used in the Fourier method may be either an FrD signal or an echo signal, but since FID signals cannot collect data in the negative time domain, echo signals are generally used.

However, compared to echo signals, FID signals have less attenuation in the transverse relaxation time T2, do not require an R pulse signal for inversion, and have a high S/N ratio per time. It can be said that it is more desirable to use the FID signal.

(Problems to be Solved by the Invention) By the way, in the conventional method of using FID signals for Fourier reconstruction, the negative time domain is estimated and compensated for using complex conjugate data folded symmetrically about the origin. If there is spatial inhomogeneity in the main magnetic field, its effect appears as a discontinuity in the waveform at the time origin, so as described in more detail below,
Shading and differential artifacts occurred in the reconstructed image.

That is, as shown in FIG. 8, an object to be measured as shown in (a) is scanned and FID signals as shown in (b) are obtained according to each warb amount. In order to perform a two-dimensional inverse Fourier transform, the data in (b) is folded back symmetrically to the origin as shown in (c) to create new 1, r data. By subjecting the new data in (c) to two-dimensional inverse Fourier transformation and absolute value processing, a reconstructed image as in (d) is obtained.

Furthermore, if we look at this mathematically, it is as follows.

If the reconstructed image is g(x, y), then =ri(×.,'1)f(x・,y) m(7/f)(y
, y)T) Here, xO is the solution of It is a non-uniform component of the magnetic field, with the χ axis in the readout direction and the y axis in the warp direction.

From the above, when the positional and concentration distortions caused by the non-uniformity of the main magnetic field are corrected, Xo becomes X and α(Xo.

y) becomes 1, the corrected image is LCx, y) = fCx, y) ctn (tfD(x,
y > ding) However, 80 = D (χ, y) - D (x, y
). The cosine of the first term results in shading, but when D is small, it becomes 1/cos, <10
(D) It can be corrected by multiplying. However, there is no means to separate the second term artifact image from the first term image, and the artifact image cannot be removed.

In view of these points, an object of the present invention is to provide a nuclear magnetic resonance imaging apparatus that can obtain 1q image with few artifacts by Fourier method reconstruction using FrD signals even when there is spatial non-uniformity in the main magnetic field. Our goal is to provide the following.

In order to achieve such an object, the present invention generates two types of gradient magnetic fields that are equal in magnitude but have different inclination directions of the magnetic field gradients, and each part is configured to measure the FID signal of the same counter-stain. A sequence control means that embeds a zero value in the negative time region of the FID signal or assumes that there is a zero, and depending on the direction of the gradient of the magnetic field gradient at the time of FID signal measurement, one is left as it is and the other is read out. A Fourier transform means that performs Fourier transform on a complex image so that the time axis is reversed; and a position and density in a readout direction according to a predetermined main magnetic field non-uniform distribution with respect to the output of the Fourier transform means Rikihi et al. distortion correction means for performing distortion correction;
a series of FID signals measured with a positive readout gradient and a series of FID signals measured with a negative readout gradient. By using a series of two types of ID signals, the Fourier transform means and the distortion correction means separately reconstruct and distortion correct the FID signals of each C, and the addition means performs complex addition of the two images. The present invention is characterized in that the influence of the non-uniformity of the main magnetic field is reduced, and then the complex image is image-processed in the image processing means to obtain a real image.

(Example) The present invention will be described in detail below with reference to the drawings. FIG. 1 is a borrowed diagram of the main parts of an NMR tomography apparatus for carrying out the method of the present invention. In the figure, 1 is a magnet assembly, which has a space (hole) inside for inserting an object, and a main magnetic field coil that surrounds this space and applies a constant magnetic field to the object. , a gradient to generate a gradient magnetic field! J 3J2 coil (X configured to be able to generate gradient magnetic fields individually)
gradient magnetic field coil, y gradient magnetic field coil, Z gradient magnetic field coil), an RF transmitting coil that provides a high-frequency pulse to excite the spin of atomic nuclei in the object, and NM from the object.
A receiving coil and the like for detecting the R signal are arranged.

The main magnetic field coils, GX, GVr, GZ gradient magnetic field cofuls, RF transmitting coil J5, and NMR signal receiving coil are the main magnetic field power supply 2, Gx, Gy, respectively.

It is connected to a Gz gradient magnetic field driver 3, an RF power amplifier 4 and a preamplifier 5. Reference numeral 10 denotes a sequence storage circuit that controls the generation sequence of gradient magnetic fields and high-frequency magnetic fields, and also controls the timing when A/D converting the obtained NMR signal.

6 is a gate modulation circuit, 7 is an R F that generates a high frequency signal.
It is an oscillation circuit. The gate modulation circuit 6 modulates the high frequency signal output from the RF oscillation circuit 7 using the timing signal from the sequence storage circuit 10 to generate high frequency pulses. This high frequency pulse is applied to the RFffi force amplifier 4.

Reference numeral 8 denotes a phase detector, which refers to the output signal of the RF oscillation circuit 7 and detects the phase of the NMR signal detected by the receiving coil and sent via the 1y+ amplifier 5.

Reference numeral 11 denotes an A/D converter, which converts the NMR signal obtained through the phase detector 8 from analog to digital.

13 includes a total of n machines 1! l! In order to realize various scans in the device, scan conditions are stored in a sequence storage circuit 10.
It has a calculation processing function that reconstructs the distribution of information regarding resonance energy into an image from observation data input from an A/D converter, and a function that sends and receives information to and from the operation console 12.

The reconstructed image obtained by the processing device 13 is displayed on the display device 9.

FIG. 2 shows an embodiment of the configuration of an image processing function section that obtains a reconstructed image from a measured FID signal. In the same figure, 21 is an input means, and two measured pIo (
FID signal FI measured with gs or positive readout slope
Dp and receives an FID signal FIDn measured with a negative readout slope. 22 is a Fourier transform means which performs Fourier transform or inverse Fourier transform according to the data given from the input means 21. The result is a complex image. 23 is a distortion correction means that receives each complex pixel 1r+ (b), corrects the complex image by referring to the main magnetic field non-uniform distribution obtained in advance, and creates another complex image 1p, In according to the sign of the readout gradient. 24 is an addition means which adds these two complex images and increases one complex image I2 by 1η. 25 is an image processing means which processes the complex pixel (#I2) into a real number image I3. It is.

The operation in such a configuration will be explained next.

A 90° pulse as shown in FIG. 3(A) is generated through the gate modulation circuit 6 under the control of the sequence storage circuit 10, and is applied to the RF transmitting coil via the RF power amplifier 4 to excite the object.

At this time, a gradient magnetic field Gz is also applied (see figure (b)),
Selectively excite only spins within a specific slice plane.

Next, as shown in (c) and (d), the gradient magnetic field Gx.

Gy is added to generate an FrD signal as shown in (e). This FID signal is detected by a receiving coil, a preamplifier 52, a phase detector 8. The signal is sent to the processing device 13 via the A/D converter 11.

In this way, a series of data is collected while changing the size of Gy (warp mother) for each view, and a reconstructed image is obtained.The operation will be explained in more detail below with reference to the flowchart in Figure 4. do.

For the same object, FID signals are measured using two types of readout gradients that have the same magnitude but different signs.

Among these FrD signals, the one measured with a positive readout slope is first input to the input means 21.

Here, the FID signal is the warp amount -(N-1)ΔG
The figure shows the entire measurement that was repeated in 2N steps from y to NΔGy. This FID signal (t, h) (1 is time, k is warb step) is defined as follows: f (x, y) is the proton density distribution, D (x, y) is the independent component of the static magnetic field, and warp If the time used for
+ (+) (t≧Q, (N-1)≦broiled≦K). This is reconstructed by a reconstruction means. Reconstruction using the Fourier transform method is performed by two-dimensional inverse Fourier transform, and as shown in FIG. 5(b), the left half plane is filled with zero values for calculation. This process is based on the unit step function 1-1(t
) was introduced, the warb step was continuously observed, and fiGy
Approximating Gy to Gy, it is expressed by the following equation. Let the reconstructed image at this time be Q+ (X, V).

x e”””d CICn<i) cl CT6y
T) However, Xo is x□ −x+D (Xo, y)
/Gx−〇 solution α<xo, y).

The Q+ (X, y) image is subjected to position correction (Xo→X) and density distortion correction (multiplying by 1/α(x, y)) by the distortion correction means 23. The corrected image h+ (X, y) is expressed by the following equation.

However, △D=D (χ, y) τD (x, y) Next,
FID signal F measured with negative readout slopes of equal magnitude
2 (t, A) is input into the input means 21. The FrD signal F2 (t, ri is expressed as follows.

(t≧0.-(I-1)≦moxibustion≦N) This signal is reconstructed by the Fourier transform means 22.

In this case, a Fourier transform normalized by 1/2π is performed in the axial direction, and a conventional inverse Fourier transform is performed in the axial direction. This operation has the effect of reversing the X-axis direction of the reconstructed image. Here, too, zero values are filled in the half plane of F2 (t, k) where t≦O ((e) in FIG. 5). By performing the same approximation as in the previous case, the reconstructed image g2 expressed by the following equation
Obtain (X, V).

X e"""d(1'extM(7%T) However, X
l is the solution of ni m
Correct the degree. Post-correction image h2 (x.

y) is as follows.

However, △D=D (χ, y) − D (x, y) Then,
When hl (x, y) and h2 (x, y) are added as a multiple image ((h) in Figure 5), the image h (x, y
) ((su) in Figure 5) becomes as follows. - 4c (x, y) = ICXpl) 10 right LCX, l no α (
x, y) ~1. Approximating β(x, y) to 1 gives the following. The second term of equation (8) is the same as the first term, f(x,
y)e-7f(XX,"l)T It is obtained by multiplying d by an extremely small value ΔD and performing an approximate quadratic differentiation, and its absolute value is smaller than the first term. This value does not pose a problem in practice. This reduction has been confirmed in simulation experiments.

Finally, the image processing means 25 takes the absolute value of h (x, y) and provides 1 for the reconstructed image (() in FIG. 5).

Note that the present invention is not limited to the above embodiments, and may be implemented as follows.

(2) When reconstructing an FID signal measured with a negative readout gradient, instead of performing Fourier transform in the readout time direction, the data in the readout direction may be reversed and then inversely Fourier transformed. In this case, the inverse Fourier transform in the warp direction is combined with the normal two-dimensional inverse Fourier transform. Both are mathematically equivalent.

(2) The magnitude of the read gradient may be set to different values for positive and negative values.

In this case, when performing Fourier transformation so that the sizes of the reconstructed images are equal, it is necessary to perform correction such as padding with zero values to an appropriate length.

■Instead of converting the complex image created by adding two images into a real number by absolute value processing, as shown in Figure 6, phase rotation is performed to obtain the original image (first term component of equation 8). and an artifact image (the second term component of equation (8)) caused by the main magnetic field inhomogeneity and the discontinuity at the time origin of the FID signal may be separated.

This means that the phases of the first and second terms in equation (8) are almost orthogonal, that is, arg (second term/first term)7-
This takes advantage of the fact that π/2.

■As shown in Figure 7, half-plane data consisting of one type of FID signal may be subjected to two-dimensional inverse Fourier transform, the complex image may be phase rotated, and separation may be performed using the same orthogonality as in (■) above. . With this method, artifact images cannot be separated in low spatial frequency components, and the image quality is inferior to that using two types of FID signals, but for images with many high spatial frequency components, images that are suitable for practical use can be obtained and can be used. be.

(Effects of the Invention) As explained above, the present invention has the following effects.

■Since FID signals can be collected earlier after excitation than echo signals, in the present invention that uses FID signals, ■2 reconstructed images with less attenuation and a large S/N ratio can be obtained. It will be done.

■Since the influence of main magnetic field inhomogeneity is small, it is possible to weaken the gradient magnetic field strength.At this time, the frequency region occupied by the FID signal becomes narrower, so the bandwidth of the low-pass filter can be narrowed. , the received 1gfff noise is reduced, and the S/N ratio of the reconstructed image can be further improved.

[Brief explanation of drawings]

Figures 1 and 2 show nuclear magnetic resonance imaging II according to the present invention.
3 is a diagram showing a pulse sequence, FIG. 4 is a flowchart 1 for explaining the operation, and FIG. 5 is an explanatory diagram for explaining the flow of the operation. 6 and 7 are explanatory diagrams for explaining the flow of operations in other embodiments of the present invention, and FIG. 8 is a flowchart for explaining the flow of image reconstruction operations in a conventional apparatus. . 1... Magnet assembly, 2... Main magnetic field power supply,
3... Gradient magnetic field drive circuit, 4... RF power amplifier,
5... Preamplifier, 6... Gate modulation circuit, 7...
・RFF! Oscillator circuit, 8: Phase detector, 9: Display device, 10: Sequence memory circuit, 11: △/
D converter, 12... Operation console, 13... Processing device, 21... Input means, 22... Fourier transform means, 23... Distortion correction means, 24... Addition means, 25
...Image processing means. Figure 3 r; T,) FED i,, fs0! lN no λ
, r (-----
Figure 7 (han (d) kune affection [y + furin 4 to i 1>)
Ku1 [Installation (I family) No. 0F31:, If11'TI book (d) 6 figure 1 entry) (c)

Claims (1)

[Claims] FID by applying high frequency pulses and gradient magnetic fields to the target object.
A nuclear magnetic resonance imaging device that generates a signal, detects this FID signal, and uses the detected signal to obtain an image of the tissue of a target object, in which the sizes are equal but the directions of inclination of the magnetic field gradients are different from each other. Two different types of gradient magnetic fields are generated to obtain FID of the same object.
A sequence control means for controlling each part so as to measure the signal; , a Fourier transform means that Fourier transforms a complex image so that one remains unchanged and the other has a readout time axis inverted between positive and negative; A positive readout comprising: a distortion correction means for correcting position and density distortion in a readout direction; an addition means for adding two complex images; and an image processing means for processing the complex image to calculate a real image. Using two types of FID signals, a series of FID signals measured with a gradient and a series of FID signals measured with a negative readout gradient, each FID signal is reconstructed and distorted separately in the Fourier transform means and the distortion correction means, and the The adding means adds complex numbers to these two images to reduce the influence of the main magnetic field inhomogeneity, and then the image processing means processes the complex images to obtain a real number image. Nuclear magnetic resonance imaging device.