阅读说明：本技术 一种增强磁共振谱仪动态范围的方法 (Method for enhancing dynamic range of magnetic resonance spectrometer ) 是由 吴林 张涛 尧德中 余洁 刘杭 于 2020-12-28 设计创作，主要内容包括：本发明公开一种增强磁共振谱仪动态范围的方法,应用于电子信息技术领域,针对现有技术获得高动态范围的核磁共振谱仪采集信号存在的问题；本发明通过预扫描得到不同增益设置下K空间的回波信号的幅值分布情况,正式扫描启动后,将K空间相位编码行分成N部分,每部分设置一个接收增益,使得相应部分的相位编码行对应的回波信号到达ADC进行模数转换时能够充分利用ADC的动态范围；采用本发明的方法降低了对磁共振谱仪接收机ADC的性能要求,磁共振接收机在固定增益控制下ADC的最佳采样位至少需要19位才能覆盖MR信号的全动态范围,接收机在可变增益控制下ADC的最佳采样位宽为15位以上即可覆盖MR信号的全动态范围。(The invention discloses a method for enhancing the dynamic range of a magnetic resonance spectrometer, which is applied to the technical field of electronic information and aims at solving the problem of acquiring signals of a nuclear magnetic resonance spectrometer with a high dynamic range in the prior art; the amplitude distribution condition of the echo signals of K space under different gain settings is obtained through pre-scanning, after formal scanning is started, the phase coding lines of the K space are divided into N parts, and each part is provided with a receiving gain, so that the dynamic range of an ADC (analog to digital converter) can be fully utilized when the echo signals corresponding to the phase coding lines of the corresponding part reach the ADC for analog to digital conversion; the method reduces the performance requirement on the ADC of the magnetic resonance spectrometer receiver, the optimal sampling bit of the ADC under the control of the fixed gain of the magnetic resonance receiver at least needs 19 bits to cover the full dynamic range of the MR signal, and the optimal sampling bit width of the ADC under the control of the variable gain of the receiver is more than 15 bits to cover the full dynamic range of the MR signal.)

1. A method of enhancing the dynamic range of a magnetic resonance spectrometer, comprising:

s1, obtaining the amplitude distribution condition of the echo signals of the K space under different gain settings through pre-scanning;

s2, iteratively correcting the gain difference and the phase corresponding to different gains;

and S3, after the formal scanning is started, dividing the K space phase coding line into N parts, and setting a receiving gain for each part to obtain complete K space data.

2. The method of claim 1, wherein the frequency encoding step number and the phase encoding step number associated with the scanning of step S1 are set to 1/4 of the formal scanning.

3. The method according to claim 2, wherein the step S1 is specifically performed by: n gain control parameters G1 and G2 … Gn are respectively set from large to small according to 6dB gain step; the phase coding comprises M steps, and a group of K space data is obtained through scanning under the control of each gain parameter; from the maximum gain G in turn₁Firstly, analyzing the K space data scanned and collected under the control of the corresponding gain parameters, and finding out all phase encoding lines without overflow of echo amplitudes; sequentially finding out all phase encoding lines with the echo amplitudes not overflowing from the remaining phase encoding steps by the subsequent gain control parameters;

note G₁To G_nThe remaining number of non-overflowed phase encoding steps under control of the gain parameter is M₁、M₂…M_nWherein M is₁+M₂…+M_n＝M。

4. The method according to claim 3, wherein the step S2 is specifically performed by: from maximum gain G₁And starting to compare the gain and the phase of the corresponding line of the echo acquired by the phase coding step number corresponding to each of the two adjacent gain control parameters to obtain a statistical gain difference factor and an echo phase difference under the control of the two adjacent gain control parameters.

5. The method according to claim 4, wherein step S3 comprises: the frequency encoding step number and the total phase encoding step number of each phase encoding line in the K space are expanded to be 4 times of those in the prescan.

6. The method of claim 5, wherein the step S3 further comprises normalizing the phase encoded line data under different gain step controls, specifically: gain G₁Phase encoding under control to 4M₁The step data remains unchanged; and normalizing the data corresponding to the phase coding steps under the control of the subsequent gain control parameters according to all the statistical gain difference factors and the echo phase difference before the gain control parameters.

7. The method of claim 6, wherein the normalization calculation process comprises: g_iControlled phase encoding to 4M_iMultiplication of complex signals at various points of the step data

Technical Field

The invention belongs to the technical field of electronic information, and particularly relates to a technology for enhancing the dynamic range of a magnetic resonance spectrometer.

Background

Obtaining signals acquired by a nuclear Magnetic Resonance (MRI) spectrometer with a high dynamic range is always a pursued target in the field of magnetic resonance. The Dynamic Range (DR) of the acquired signal is the power ratio of the maximum effective value to the minimum effective value. DR is typically expressed in decibels (db),smax is the maximum signal amplitude to be measured and epsilon is the minimum signal amplitude that can be identified from the noise floor. This ratio determines the number of bits required by an analog-to-digital converter (ADC) to convert an analog signal to a digital quantity, and it also dictates the channel DR requirements for transmitting this signal. The factor 2 in parentheses expresses the fact that the signal range can be positive or negative. The minimum number of bits J of the ADC required to meet the DR requirement of the input signal is given by: j is DR/6.02. Any global SNR-affecting parameter will similarly affect DR, so extending DR may improve the signal-to-noise ratio. In order to fully exploit the potential of an MRI spectrometer receiver, the ADC must capture the full signal in the analog-to-digital conversion stage, ranging from the central peak in k-space to the signal amplitude of the thermal noise floor of the system.

The following 3 methods are commonly used to improve the dynamic range of signals collected by the magnetic resonance spectrometer:

1) the effective DR of the ADC is enhanced at a sampling rate higher than the Nyquist rate, thereby reducing ADC quantization noise. Although the performance of the ADC is rapidly developed in recent years, it is still difficult to consider the two indexes of high bit width and high conversion rate at the same time, and when the bit width of the ADC is increased, the conversion rate is decreased, whereas when the conversion rate is increased, the bit width is decreased. Simulation results in papers C.H.OH, Y.C.RYU, J.H.HYUN, S.H.BAE, S.T.CHUNG, H.W.PARK, Y.G.KIM.dynamic Range extension of Receiver by Using Optimized Gain Adjustment for High-Field MRI, J.hubs in Magnetic Resonance resource Part A, Vol.36A (4) 243-254 (2010) show that the Receiver needs at least 19 bits of optimal sampling bits of ADC under the control of fixed Gain to meet the requirement of dynamic Range of Magnetic Resonance image. Direct radio frequency sampling technology is very commonly applied in modern magnetic resonance equipment, and the conversion rate of an ADC is usually more than 80 MHz. Therefore, it is difficult to satisfy the ADC specification at the same time, if the ADC specification is more than 19 bits wide, the conversion rate is more than 80MHz, and the cost is within the acceptable range of the commercial spectrometer.

2) The dynamic range is increased by nonlinear gradient pulses. The paper v.j.wedeen, y.s.chao, j.l.ackerman, Dynamic range compression in MRI by means of a nonlinear gradient pulse, j.magn.reson.med.6(1988) 287-295 mentions that the nonlinear gradient pulse converts the linear phase distribution of the scanned object into a nonlinear phase pulse, thereby compressing the maximum peak of the k-space data and increasing the Dynamic range of the signal. By changing the design of the gradient coils, the gradient in space is not linearly changing, resulting in a non-linear gradient field. However, it is difficult to design a nonlinear gradient coil capable of meeting the target requirement in engineering practice, and there is insensitivity to system error in linear gradient encoding, and at this time, the nonlinear compressed amplitude signal of the echo central signal may not be accurately recovered by using the nonlinear gradient pulse technique.

3) In view of some technical bottlenecks existing in the methods for acquiring DR signals by using the enhanced magnetic resonance spectrometer in the above 1) and 2), people transfer a research field of view to the scaling of signal amplitude values corresponding to the related local positions of magnetic resonance k space, wherein the signal rescaling is realized by different gain settings, so that an ADC can cover DR which is larger than the dynamic range of the ADC itself. The paper Katsumi Kose, K.Endoh, T.Inouye.Nonliner Amplitude Compression In Magnetic Resonance Imaging, Quantization Noise Reduction and Data Memory Saving, J.IEEE AES Magazine, June 1990 compresses the low frequency components of the K-space phase encoded lines and expands the high frequency components. Simulation shows that the non-linear amplification reduces the quantization noise by 10 to 20 times, and the reduced dynamic range is increased by about 4 bits. According to the papers j.bollenbeck, m.vehicle, r.oppelt, h.kroeckel, w.schnell.a high performance multi-channel RF Receiver for magnetic response Imaging Systems, j.proc.int.soc.mag.reson.med.med.13 (2005), a concrete implementation scheme of nonlinear compression is proposed by siemens 2005, a signal amplitude compressor/expander (compander) is designed before the MRI Receiver ADC, which obviates the need for internal automatic gain control of the Receiver, and when the ADC conversion rate is 10MS/s, the MRI DR can achieve 164dB/Hz, and the corresponding effective bit number (ENOB) is 16 bits. If there is no compander before the ADC, 18 bits of resolution are required to achieve a DR of 164 dB/Hz. Since the scheme provided by siemens corporation must perform dynamic compression of signal amplitude in analog circuits, such design is a very challenging task in engineering implementation, and besides the design index satisfying basic functions (DR and compression characteristics), the design index must also comprehensively consider the influences of factors such as non-linear compression signal recoverability, voltage offset, temperature stability and the like.

Disclosure of Invention

In order to solve the technical problem, the invention provides a method for enhancing the dynamic range of a magnetic resonance spectrometer, linear gradient pulses are used for encoding in the phase encoding direction and the frequency encoding direction, and a nonlinear compression analog circuit is not used for carrying out nonlinear amplification on signals acquired by different frequency encoding of one phase encoding line.

The technical scheme adopted by the invention is as follows: a method of enhancing the dynamic range of a magnetic resonance spectrometer, comprising:

s1, obtaining the amplitude distribution condition of the echo signals of the K space under different gain settings through pre-scanning;

s2, iteratively correcting the gain difference and the phase corresponding to different gains;

and S3, after the formal scanning is started, dividing the K space phase coding line into N parts, and setting a receiving gain for each part to obtain complete K space data.

The frequency coding step number and the phase coding step number of the scan are set to 1/4 of the main scan in step S1.

Step S1 specifically includes: n gain control parameters G1 and G2 … Gn are respectively set from large to small according to 6dB gain step; the phase coding comprises M steps, and a group of K space data is obtained through scanning under the control of each gain parameter; from the maximum gain G in turn₁Starting to analyze the K space data scanned and collected under the control of the corresponding gain parameters,finding out all phase encoding lines with the echo amplitudes not overflowing; sequentially finding out all phase encoding lines with the echo amplitudes not overflowing from the remaining phase encoding steps by the subsequent gain control parameters;

note G₁To G_nThe remaining number of non-overflowed phase encoding steps under control of the gain parameter is M₁、M₂…M_nWherein M is₁+M₂…+M_n＝M。

Step S2 specifically includes: from maximum gain G₁And starting to compare the gain and the phase of the corresponding line of the echo acquired by the phase coding step number corresponding to each of the two adjacent gain control parameters to obtain a statistical gain difference factor and an echo phase difference under the control of the two adjacent gain control parameters.

Step S3 includes: the frequency encoding step number and the total phase encoding step number of each phase encoding line in the K space are expanded to be 4 times of those in the prescan.

Step S3 further includes normalizing the phase encoded line data under different gain step controls, specifically: gain G₁Phase encoding under control to 4M₁The step data remains unchanged; and normalizing the data corresponding to the phase coding steps under the control of the subsequent gain control parameters according to all the statistical gain difference factors and the echo phase difference before the gain control parameters.

The normalized calculation process is as follows: g_iControlled phase encoding to 4M_iMultiplication of complex signals at various points of the step data

The invention has the beneficial effects that: the invention collects the magnetic resonance echo signals based on the conventional ADC (16bit wide &80MHz sampling rate), and uses linear gradient pulse to code in the phase coding and frequency coding directions, and the invention does not use a nonlinear compression analog circuit to carry out nonlinear amplification on the signals collected by different frequency codes of one phase coding line. The amplitude distribution condition of the echo signals of K space under different gain settings is obtained through pre-scanning, after formal scanning is started, the phase coding lines of the K space are divided into N parts, and each part is provided with a receiving gain, so that the dynamic range of an ADC (analog to digital converter) can be fully utilized when the echo signals corresponding to the phase coding lines of the corresponding part reach the ADC for analog to digital conversion; the invention has the following advantages:

1. the performance requirement of the magnetic resonance spectrometer receiver ADC is reduced, the optimal sampling bit of the magnetic resonance spectrometer receiver ADC under the fixed gain control at least needs 19 bits to cover the full dynamic range of the MR signal, and the optimal sampling bit width of the magnetic resonance spectrometer ADC under the variable gain control of the receiver ADC is more than 15 bits to cover the full dynamic range of the MR signal;

2. the invention improves the dynamic range of the image by optimizing the scanning parameters of the sequence under the condition of not modifying the circuit of the magnetic resonance receiver;

3. the scanning time of the whole sequence is only increased by about 5%, the signal-to-noise ratio of a reconstructed image is improved by 10-20 dB (the bit width of an ADC is equivalent to the improvement of 4 bits);

4. the iterative signal gain and phase correction method avoids the influence of a noise substrate on echoes with smaller amplitude and reduces correction errors.

Drawings

FIG. 1 is a flow chart of a protocol of the present invention;

FIG. 2 is a schematic K-space diagram of a pre-scan (prescan) provided by an embodiment of the present invention;

fig. 3 is a schematic diagram of iterative correction of gain difference and phase under different gain steps according to an embodiment of the present invention;

fig. 3(a) is a schematic diagram of obtaining a gain difference under different gain steps through a histogram, and fig. 3(b) is a schematic diagram of obtaining a phase under different gain steps through another histogram;

FIG. 4 is a K-space schematic of a corrected pre-scan (prescan);

FIG. 5 is a schematic diagram of a K space for formal scanning according to an embodiment of the present invention;

FIG. 6 is a simulation result of the method of the present invention;

fig. 6(a) is a relationship between different ADC bit widths and an acquisition signal-to-noise ratio under a fixed gain, and fig. 6(b) is a relationship between different ADC bit widths and an acquisition signal-to-noise ratio under a dynamic adjustment gain.

Detailed Description

In order to facilitate the understanding of the technical contents of the present invention by those skilled in the art, the present invention will be further explained with reference to the accompanying drawings.

The invention collects the magnetic resonance echo signals based on the conventional ADC (16bit wide &80MHz sampling rate), and uses linear gradient pulse to code in the phase coding and frequency coding directions, and the invention does not use a nonlinear compression analog circuit to carry out nonlinear amplification on the signals collected by different frequency codes of one phase coding line.

The amplitude distribution condition of the echo signals of the K space under different gain settings is obtained through pre-scanning, after formal scanning is started, the phase coding line of the K space is divided into N parts, and each part is provided with one receiving gain, so that the dynamic range of an ADC (analog-to-digital converter) can be fully utilized when the echo signals corresponding to the phase coding line of the corresponding part reach the ADC for analog-to-digital conversion.

The implementation process of the invention is shown in fig. 1, and comprises the following three steps:

1. in the prescan, as shown in fig. 2, the abscissa represents the number of frequency steps, the ordinate represents the number of phase encoding steps, and the scan matrix size is 1/4 of the main scan, that is, the number of frequency encoding steps and the number of phase encoding steps are set to 1/4 of the main scan. N gain control parameters G are respectively set from large to small according to 6dB gain step₁、G₂…G_nI.e. one scan each under the n receive gain control parameters; the advantage of this design can greatly reduce the time of pre-scanning, wherein the phase encoding has M steps, and a group of K space data is obtained by scanning under the control of each gain parameter. From the maximum gain G in turn₁Starting, analyzing the K space data scanned and collected under the control of the gain parameter, finding out all the phase encoding lines without the overflow of the echo amplitude, and summing up M₁Step phase coding; then analyze G₂Controlled down scanningCollected K-space data from the remaining M-M₁Finding out all the phase encode lines without overflow of echo amplitude from step phase encode, summing up M₂Step phase coding; respectively obtaining G according to the above rule₃To G_nThe remaining number of non-overflowed phase encoding steps under control of the gain parameter is M₃、M₄…M_nWherein M is₁+M₂…+M_n＝M。

2. And performing iterative correction of gain difference and phase under different gain steps. The data format is set to 32-bit floating point. G₁Under the control of M₁Step phase encoding of acquired echo and G₂Under control, these M are the same₁And gain and phase comparison are carried out on the corresponding line of the echo acquired by the step phase coding. As shown in FIG. 3, G is obtained from the histogram₁And G₂The statistical gain difference factor under control is Δ G₁(see FIG. 3 (a)), and G is also obtained from another histogram₂Relative to G₁The phase difference of the echo under control is phi₁(as shown in FIG. 3 (b)), and sequentially obtaining G₂And G₃Is Δ G₂，G₃Relative to G₂The phase difference of the echo under control is phi₂… …, and G_n-1And G_nIs Δ G_n-1，G_nRelative to G_n-1The phase difference of the echo under control is phi_n-1。

FIG. 4 is a diagram illustrating a gain difference and a phase difference between adjacent receiving gains obtained according to FIG. 3, and a method for implementing gain and phase compensation of each portion of K space under different receiving gain controls for each portion of K space, that is, data obtained under different receiving gain controls for each portion of K space is normalized to the same receiving gain; on the right side of fig. 4 are gain difference and phase difference compensation values for compensating the data corresponding to the parenthesized area in fig. 4.

3. Starting formal scanning, as shown in FIG. 5, expanding the frequency encoding step number and the total phase encoding step number of each phase encoding line in K space from M of pre-scanning to 4M, and obtaining gain G₁Phase encoding under control to 4M₁Step (G)₂Phase encoding under controlThe code is 4M₂Step by analogy, G_nThe controlled phase encoding step number is 4M_nAnd (5) carrying out the steps. Scanning and collecting to obtain complete K space data, and respectively normalizing the phase coding data under different gain step controls. Gain G₁Phase encoding under control to 4M₁The step data remains unchanged; g₂Controlled phase encoding to 4M₂Multiplication of complex signals at various points of the step dataPhase encoding to 4M under control of G3₃Multiplication of complex signals at various points of the step dataBy analogy, G_nControlled phase encoding to 4M_nMultiplication of complex signals at various points of the step dataG in FIG. 5₁,…,G_nIs the receive gain parameter for the corresponding K-space region.

In this embodiment, the abscissa direction of the K space is frequency encoding, the ordinate direction is phase encoding, and each row is one step of phase encoding.

As shown in fig. 6, the simulation result expresses the comparison relationship between the ADC bit width and the signal-to-noise ratio of the acquired signal under the fixed gain and the dynamic adjustment gain, fig. 6(a) is the relationship between the ADC bit width and the signal-to-noise ratio under the fixed gain, and fig. 6(b) is the relationship between the ADC bit width and the signal-to-noise ratio under the dynamic adjustment gain, where GC is the abbreviation of dynamic gain (gain control). It can be known from the simulation of fig. 6 that the dynamic range of the acquired signal, which is equivalent to the dynamic range of the acquired signal in which the gain is fixed and the bit width of the ADC is 19 bits, can be realized by the ADC with the bit width of 15 bits through the dynamic gain adjustment of the spectrometer during the scanning process, and the method does not need to modify any hardware of the spectrometer.

It will be appreciated by those of ordinary skill in the art that the embodiments described herein are intended to assist the reader in understanding the principles of the invention and are to be construed as being without limitation to such specifically recited embodiments and examples. Various modifications and alterations to this invention will become apparent to those skilled in the art. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the scope of the claims of the present invention.