High signal-to-noise ratio thin slice t1 parameter quantitative imaging method based on radio frequency encoding pulse

Abstract

The application provides a high signal-to-noise ratio thin layer T1 parameter quantitative imaging method based on radio frequency coding pulses, a new two-dimensional radio frequency coding pulse based on amplitude and phase modulation is designed through required scanning parameters, and a high signal-to-noise ratio high layer direction resolution image can be obtained through Tikhonov reconstruction, and meanwhile, the radio frequency excitation field B1 + The uneven, layer excitation flip angle profile uneven and crosstalk problem between adjacent layers are corrected by using a dictionary matching method to obtain a T1 quantitative image. The application simultaneously considers the signal-to-noise ratio and layer direction resolution of the reconstructed image by designing a new radio frequency pulse, and corrects multiple factors influencing the T1 parameter quantitative accuracy. Compared with other traditional methods, the application can provide an accurate high signal-to-noise ratio thin layer T1 quantitative image within a clinically acceptable time.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of MRI imaging, and particularly relates to a high signal-to-noise ratio thin layer T1 parameter quantitative imaging method based on a radio frequency coding pulse. BACKGROUND

[0002] Quantitative magnetic resonance imaging is a non-invasive imaging method that can provide quantitative parameter images reflecting physiological characteristics of brain tissue. Traditional magnetic resonance images are qualitative weighted images, which are easily affected by sequence parameter settings, hardware configurations and other factors, and cannot accurately measure tissue parameters. Quantitative images can better monitor changes in pathological tissues, and the results have higher repeatability and comparability, and have important application value in clinical diagnosis.

[0003] Longitudinal relaxation time T1 is an important parameter in quantitative magnetic resonance imaging research, directly reflecting the inherent physiological and physical characteristics of the tissue and being sensitive to pathological tissue changes. In the brain, the size of the T1 parameter is closely related to the myelin sheath, and can be used for the diagnosis of diseases such as multiple sclerosis, Parkinson's disease, epilepsy, etc. The traditional gold standard for T1 parameter quantification is the inversion recovery sequence, but this sequence has the problems of long scanning time and being easily affected by motion and other factors, and is not suitable for clinical application. Variable flip angle method (VFA) [Cheng, H.-L.M. and Wright, G.A. (2006), Rapid high-resolution T1 mapping by variable flip angles: Accurate and precise measurements in the presence of radio frequency field inhomogeneity. Magn. Reson. Med., 55: 566-574] calculates the T1 quantitative image by acquiring images of gradient echo sequences under different flip angle settings. Since the gradient echo sequence has relatively high temporal resolution and short acquisition time, it can acquire images within a reasonable time in clinical practice, and is therefore commonly used in basic and clinical research. In order to improve the temporal resolution and reduce the error caused by incomplete destruction of transverse residual magnetization on T1 parameter quantification, a two-dimensional layer interleaved acquisition method can be used. At the same time, the crosstalk between adjacent layers, B1 inhomogeneity and other factors can affect the accuracy of the T1 parameter quantification. In order to solve this problem, the application adopts a method of using a radio frequency coding pulse to encode the signal of each layer, and then reconstructs the T1 quantitative image of each layer by using the radio frequency coding pulse. +The field inhomogeneity and the inhomogeneity of the layer excitation flip angle profile will make the collected signal deviate from the setting, resulting in a large error in T1 estimation. In order to improve the accuracy of T1 parameter estimation, the Bloch equation can be used to simulate the actual layer excitation flip angle profile [Svedin, B. T., Parker, D. L.: Technical Note: The effect of 2D excitation profile on T1 measurement accuracy using the variable flip angle method with an average flip angle assumption. Med. Phys. 44, 5930-5937 (2017)], and then combined with the collected B1 + The quantitative map is obtained by the method of dictionary matching.

[0004] The resolution of the layer direction of the two-dimensional acquisition mode cannot be set too high, but for some clinical imaging applications, higher layer direction resolution is needed to assist diagnosis. In order to balance the layer direction resolution and the signal-to-noise ratio, a new type of radio frequency pulse coding technology based on amplitude and phase coding is introduced, which simultaneously acquires a mixed signal of a thick layer by radio frequency coding of the layer signal, and then reconstructs a high signal-to-noise ratio thin layer signal. By combining this technology with the two-dimensional gradient echo sequence, not only the signal-to-noise ratio of the thin layer signal can be improved, but also the power between different excitation pulses can be balanced by optimizing the flip angle combination of the excitation pulse, thereby reducing the influence of magnetization transfer on T1 parameter estimation. SUMMARY

[0005] In view of the above, the present application provides a high signal-to-noise ratio thin layer T1 parameter quantitative imaging method based on radio frequency coding pulse, which can simultaneously correct the deviations caused by B1 + field inhomogeneity and the inhomogeneity of the layer excitation flip angle profile, reduce the influence of interlayer crosstalk and magnetization transfer on T1 parameter estimation, and also obtain a high signal-to-noise ratio thin layer T1 quantitative map.

[0006] A high signal-to-noise ratio thin layer T1 parameter quantitative imaging method based on radio frequency coding pulse, comprising the following steps:

[0007] (1) Design and generate a two-dimensional radio frequency coding pulse based on M thin layers after amplitude modulation and phase modulation, which is superimposed by a radio frequency pulse, for the required tissue parameters and scanning parameters. The tissue parameters are longitudinal relaxation time T1, and M is the number of thin layers contained in each thick layer;

[0008] (2) combine the two-dimensional radio frequency encoding pulse with a two-dimensional gradient echo sequence to generate a magnetic resonance data acquisition sequence and import into a magnetic resonance scanner, use the magnetic resonance scanner to divide the brain part of the subject into a plurality of thick layer regions combined by thin layers located at different layer positions, then simultaneously scan and reconstruct to obtain L brain thick layer images excited by different radio frequency encoding pulses, L=M*N, N is the number of flip angles;

[0009] (3) separate the L brain thick layer images using matrix inversion reconstruction to obtain M thin layer images under N different flip angle excitations;

[0010] (4) consider the radio frequency excitation field B1 + non-uniform, layer excitation flip angle profile non-uniform, and signal crosstalk between adjacent thick layers, design a corrected steady-state signal equation, and give the dynamic variation range and discretization step length of the required quantitative tissue parameter T1 based on the equation, establish a thick layer dictionary reflecting the signal intensity change with respect to different B1 + size; +

[0011] (5) based on the matrix inversion reconstruction, separate the thick layer dictionary to obtain a thin layer dictionary reflecting the signal intensity change with respect to different B1 + size, then acquire a brain B1 + quantitative map, and select the B1 + value of each pixel point in the B1 + quantitative map to select the corresponding thin layer dictionary with the B1 + size;

[0012] (6) match the signal size of each pixel point in the thin layer image with respect to the flip angle change with the flip angle signal in the thin layer dictionary one by one, thereby indexing the specific tissue parameter T1 for each pixel point, and further obtaining a quantitative image of the tissue parameter T1.

[0013] Further, the two-dimensional radio frequency encoding pulse designed in step (1) is superimposed by M two-dimensional thin layer pulses distributed along the layer direction, which can simultaneously excite all tissues in the thick layer region composed of the positions of these thin layers at a time. The position distribution of these thin layers is continuous or discontinuous; the amplitude and phase of each two-dimensional thin layer pulse are designed according to the required parameters, including the number of flip angles N, the flip angle size α n (n=1, 2, …, N), and the number of thin layers contained in each thick layer M.

[0014] Further, the amplitude modulation has (N!) M combination ways, based on the fact that each two-dimensional thin layer needs to be flipped N times with flip angle sizes α n ​The principle of amplitude modulation of (n = 1, 2, …, N), each combination contains N amplitude modulation mode, in each amplitude modulation mode, the need to set the amplitude of the two-dimensional thin layer pulse at M different positions in a thick layer region, corresponding to a combination of flip angle size.

[0015] Further, the phase modulation, that is, the phase modulation of each two-dimensional thin layer pulse in the thick layer region is π relative to other thin layer size, and each two-dimensional thin layer pulse in the thick layer region is phase modulated once, a total of M phase modulations.

[0016] Further, the amplitude and phase modulation process in step (1) is performed simultaneously for all M two-dimensional thin layer pulses in a thick layer region, that is, M phase modulations are required for each of the N amplitude modulation modes contained in each amplitude combination, resulting in L = M * N radio frequency encoding pulses; then using these radio frequency encoding pulses to excite the tissue to be scanned, the repetition interval between adjacent two-dimensional radio frequency encoding pulses in the same layer is TR, and different thick layer interleaved excitation is used in the same TR to realize whole brain coverage scanning.

[0017] Further, the matrix inversion reconstruction method used in step (3) is based on Tikhonov regularization reconstruction, and the specific expression is as follows:

[0018] X = (A T A+λI) -1 A T Y

[0019] Where: Y is the thick layer image of the brain, X is the thin layer image, A is the radio frequency encoding matrix (obtained according to the phase of the radio frequency encoding pulse), λ is the Tikhonov regularization parameter, and I is the unit matrix.

[0020] Further, the radio frequency excitation field B1 + The correction of non-uniform, layer excitation flip angle profile non-uniform and signal crosstalk between adjacent thick layers is obtained by Bloch equation simulation to get the distribution of flip angle along the layer direction under the current scanning parameter setting, and then a dictionary is established based on the steady-state signal equation. The expression of the corrected steady-state signal equation is as follows:

[0021]

[0022] Where: E 1 / 2=exp(-TR / 2T1), E1=exp(-TR / T1), S is the intensity of the acquired brain thick layer image signal, M0 is the initial net magnetization vector, K is the number of discrete sampling points, Δz is the interval between adjacent sampling points in the slice direction, TR is the repetition interval between adjacent two-dimensional radio frequency coded pulses, i is the imaginary unit, k is the sequence number of the discrete sampling point in the slice direction, α1(k) is the flip angle of the slice corresponding to the kth discrete sampling point, α2(k) represents the flip angle of the pulse at the position of the two adjacent slices above and below the slice corresponding to the kth discrete sampling point when it was last excited, and Φ(k) is the phase of the transverse magnetization vector at the position of the slice corresponding to the kth discrete sampling point.

[0023] Furthermore, in step (4), information about different B1 is established. + The size of the thick dictionary is obtained through B1 + This is achieved by correcting the flip angle in the steady-state signal equation. The correction expression is as follows:

[0024]

[0025] Where: α norm Set the flip angle value, α corr The corrected flip angle size, r represents the position of any voxel in space in a thick brain image, B1 + (r) represents the radio frequency excitation field intensity at position r.

[0026] Furthermore, in step (5), brain B1 + Quantitative graphs are obtained by using B1 + Quantitative magnetic resonance data acquisition sequences are imported into an magnetic resonance scanner, which is then used to scan and reconstruct multi-section brain radiofrequency excitation fields (B1) of the subject's brain. + Quantitative images.

[0027] This invention designs a novel two-dimensional radio frequency coding sequence based on amplitude and phase coding. After the acquired image is reconstructed using Tikhonov regularization, it is matched with a thin-layer dictionary that has also undergone Tikhonov regularization to obtain an accurate high signal-to-noise ratio thin-layer T1 quantitative map. This process requires first acquiring B1... + Quantitative imaging is performed, and the flip angle profile of the excitation layer is simulated based on the Bloch equation. Then, a dictionary is established based on the steady-state signal equation corrected for interlayer crosstalk. Combining the performance of this invention in water film and human data experiments, and comparing it with the traditional T1 quantitative gold standard method, it is demonstrated that this invention can achieve quantitative imaging results consistent with the traditional gold standard method in a shorter time and with higher layer-direction resolution. This has significant practical application value for the diagnosis of diseases such as multiple sclerosis and Parkinson's disease. BRIEF DESCRIPTION OF DRAWINGS

[0028] Figure 1 The excitation profile of eight new 2D RF encoding pulses designed by combining amplitude modulation with phase modulation using the flip angle alternation in the example one of Table 1, wherein the excited thick layer is composed of thin layers continuously distributed in four layer positions.

[0029] Figure 2 The flowchart of the high SNR thin layer T1 quantitative algorithm based on dictionary matching of the present application.

[0030] Figure 3 (a) is the T1 quantitative map result obtained by using the conventional uncorrected variable flip angle method (2D VFA) on the water film experiment.

[0031] Figure 3 (b) is the T1 quantitative map result obtained by using the dictionary matching based correction variable flip angle method (DM based VFA) on the water film experiment.

[0032] Figure 3 (c) is the T1 quantitative map result obtained by using the method of the present application on the water film experiment.

[0033] Figure 3 (d) is the T1 quantitative map result obtained by using the gold standard method (IR) on the water film experiment.

[0034] Figure 4 (a) is the T1 quantitative map result obtained by using the DM based VFA method on the 19th layer (ascending) and the 23rd layer (descending) selected from the experiment on the human brain.

[0035] Figure 4 (b) is the T1 quantitative map result obtained by using the method of the present application on the 19th layer (ascending) and the 23rd layer (descending) selected from the experiment on the human brain. DETAILED DESCRIPTION

[0036] In order to more specifically describe the present application, the technical solutions of the present application will be described in detail below in combination with the drawings and specific embodiments.

[0037] The high resolution thin layer T1 parameter quantitative imaging method based on new RF encoding pulses of the present application comprises the following steps:

[0038] (1) Design of new 2D RF encoding pulse sequence based on amplitude and phase modulation.

[0039] 1.1 A radio frequency coded pulse consists of M two-dimensional thin-layer pulses distributed along the planar direction. It can simultaneously excite all tissue within a thick-layer region comprised of the locations of these thin layers. The distribution of these thin layers can be continuous or discontinuous. The amplitude and phase of each two-dimensional thin-layer pulse can be designed according to required parameters, including the number of flip angles N (N>=2) and the flip angle magnitude α. n (n=1,2,…,N), the number of thin layers M contained in each thick layer (M>=2).

[0040] There are N! total amplitude modulation methods. M This combination method is based on the fact that each two-dimensional thin layer needs to be flipped N times with an angle of α. n The principle of amplitude modulation is that each combination includes N amplitude tuning modes. Specific amplitude modulation combinations can be selected as needed. For example, to reduce the influence of magnetization transfer on T1 quantization, the amplitude modulation mode can be set to alternately change the amplitude of two-dimensional thin-layer pulses at M different locations within the same thick-layer region. That is, different flip angles are used to alternately excite the thin-layer at different locations, making the power of the excitation pulses between different excitation modes more balanced, thereby reducing the magnetization transfer effect caused by differences in excitation power.

[0041] 1.3 Phase modulation is performed by setting one of the two-dimensional thin-layer pulses in each thick layer region to a phase of π relative to the other thin layers; in this way, phase modulation is performed once for each thin layer in the thick layer, and a total of M phase modulations are required.

[0042] 1.4 The above amplitude and phase modulation process requires all M two-dimensional thin-layer pulses within a thick layer to be performed simultaneously. That is, for each amplitude combination method, the N amplitude modulation modes need to be phase modulated M times, resulting in a total of L = M * N RF coded pulses.

[0043] Table 1 shows specific implementation examples of two amplitude modulation methods using N=2 flip angles (flip angles of α1 and α2 respectively). Each example includes two amplitude modulation modes. Example 1 uses an amplitude modulation method with alternating α1 and α2, and Example 2 uses an amplitude modulation method with only α1 and α2. The RF coded pulses designed using the alternating α1 and α2 amplitude modulation method in Example 1 of Table 1 sequentially excite a thick layer composed of two-dimensional thin layers continuously distributed at M=4 layer positions. Each amplitude modulation mode performs M phase modulations, meaning each thin layer undergoes one phase modulation with a phase of π relative to other thin layers. The resulting 8 RF pulse excitation profiles are shown below. Figure 1 As shown.

[0044] Table 1

[0045]

[0046]

[0047] (2) The designed new two-dimensional radio frequency coding pulse based on amplitude and phase modulation is combined with a two-dimensional gradient echo sequence to generate a magnetic resonance data acquisition sequence, which is imported into a magnetic resonance scanner to scan a subject, and a total of L thick layer images can be acquired.

[0048] We respectively carried out experiments on water film and human brain to verify the effectiveness of the application, and compared the T1 quantitative results of the traditional uncorrected variable flip angle method (2D VFA), the corrected variable flip angle method based on the dictionary matching method of the application (DM based VFA) and the gold standard method (IR) to verify the accuracy of the application.

[0049] 2.1 Water film experiment verification.

[0050] The scanning experiment parameters of the sequence of the application are: TR=0.194s, in-plane resolution=1x1mm 2 , reconstruction matrix=174x192x114mm 3 , thick layer thickness=6mm, each thick layer is composed of 4 thin layers, thin layer thickness=1.5mm, radio frequency pulse duration=4ms, radio frequency pulse time bandwidth product=4, flip angle=[5°, 10°, 30°, 34°, 40°, 50°], and the average acquisition time of each thin layer is about 15s.

[0051] The scanning experiment parameters of the traditional uncorrected variable flip angle method (2D VFA) are the same as those of the corrected method based on the dictionary matching method of the application except that the radio frequency pulse duration=2ms and the radio frequency pulse time bandwidth product=2.7, and the difference between the two lies in the T1 parameter quantitative reconstruction method, the former adopts the traditional nonlinear least squares method, and the latter adopts the corrected method based on the dictionary matching method of the application.

[0052] The scanning experiment parameters of the gold standard method (IR) are: TR=5000ms, in-plane resolution 1x1mm 2 , reconstruction matrix=150x192x5mm 3 , layer thickness=5mm, TI=[200, 500, 800, 1200, 1700]ms, and the acquisition time of one layer is 507s.

[0053] 2.2 Human brain experiment verification.

[0054] Under the approval of the ethics committee, we carried out magnetic resonance data scanning on a healthy volunteer.

[0055] The scan experimental parameters of the correction variable flip angle method (DM based VFA) based on the dictionary matching method of the present application are the same as those used in the water film experiment, except that the flip angle = [5°, 20°, 34°, 50°].

[0056] In order to ensure the signal-to-noise ratio of the traditional uncorrected variable flip angle method (2D VEA) in the human brain experiment, the layer thickness is set to 3 mm, and other parameters are the same as those used in the water film experiment.

[0057] The gold standard method (IR) directly uses the results in the literature (Dieringer MA, Deimling M, Santoro D, Wuerfel J, Madai VI, et al.: Rapid Parametric Mapping of the Longitudinal Relaxation Time T1 Using Two-Dimensional Variable Flip Angle Magnetic Resonance Imaging at 1.5 Tesla, 3 Tesla, and 7 Tesla. PLoS ONE. 9, e91318 (2014)), and the layer thickness is also 5 mm.

[0058] (3) The B1 + The quantitative magnetic resonance data acquisition sequence Turbo-FLASH is imported into the magnetic resonance scanner to scan the subject, and the radio frequency excitation field B1 + The quantitative image is obtained.

[0059] (4) The distribution of the flip angle along the layer direction under the current scan parameters (including the flip angle size, the thin layer thickness, the number of thin layers in a thick layer, the radio frequency pulse duration, and the radio frequency pulse bandwidth) is simulated by the Bloch equation simulation, and the obtained layer excitation flip angle profile is as shown in Figure 1 .

[0060] (5) The high signal-to-noise ratio thin layer T1 quantitative image is obtained based on the dictionary matching method, and the process is as shown in Figure 2 .

[0061] 5.1 First, the L thick layer images collected in step (2) are separated and reconstructed using Tikhonov regularization to obtain the separated high signal-to-noise ratio thin layer images, and the separation and reconstruction model is as follows:

[0062] X = (A T A+λI) -1 A T Y

[0063] where Y is the acquired L RF-encoding thick-slab images, X is the separated reconstructed thin-slab images for N different flip angles, A is the RF-encoding matrix, λ is the Tikhonov regularization parameter, and I is the identity matrix.

[0064] 5.2 Considering the problem of inter-slice crosstalk existing in 2D acquisition mode, a new type of 2D RF-encoding pulse inter-slice crosstalk correction steady-state signal equation is designed, and its expression is as follows:

[0065]

[0066] where E 1 / 2 = exp(-TR / 2T1), E1= exp(-TR / T1), S is the acquisition signal intensity, M0 is the initial net magnetization vector, K is the number of discrete sampling points, Δz is the interval of adjacent sampling points in the slice direction, k is the serial number of the discrete sampling point in the slice direction, TR is the repetition time, α1 is the flip angle size of the current excitation slice, α2 is the flip angle size of the pulse at the position of the current slice on the upper and lower two adjacent slices when the pulse was excited last time, and Φ is the phase of the transverse magnetization vector at the position of the current slice.

[0067] 5.3 Based on the above inter-slice crosstalk correction steady-state signal equation, the simulated slice excitation flip angle profile obtained in step (4) and the parameters T1 and B1 + , a thick-slab dictionary reflecting the signal intensity change with the flip angle for different B1 + sizes is established. The specific establishment method is as follows: first, set the dynamic change range of the parameter T1 to be 0.1-5 s, wherein the discretization step length in the interval of 0.1-2 s is 0.02, the discretization step length in the interval of 2-3 s is 0.1, and the discretization step length in the interval of 3-5 s is 0.2; set the dynamic range and the discretization step length of B1 + to be 0.1-2 and 0.01, respectively; and further correct the simulated slice excitation flip angle profile obtained in step (4) by using the B1 + inhomogeneity correction formula to obtain the flip angle expression under different B1 + sizes, as follows:

[0068]

[0069] where α norm is the set value of the flip angle, α corr is the corrected flip angle size, and r is the position of the voxel.

[0070] Then, based on the above T1 discrete values, the flip angle values under different B1 + sizes, and the inter-slice crosstalk correction steady-state signal equation, a thick-slab dictionary reflecting the signal intensity change with the flip angle for different B1 +A thick dictionary of varying sizes is used, and this thick dictionary is then separated and reconstructed using Tikhonov regularization to establish a thin dictionary suitable for the method of this invention. Based on B1 obtained in step (3)... + B1 of each pixel in the quantitative image + Value, select the corresponding B1 + A thin-layer dictionary of varying sizes is established. Finally, the thin-layer image after reconstruction and separation of the thick-layer image is matched with the established thin-layer dictionary to obtain the T1 quantitative map.

[0071] The following experiments were conducted on a water film and the human brain to verify the accuracy and signal-to-noise ratio improvement of the method of this invention. Figure 3 (a)~ Figure 3 (d) These are comparisons of T1 quantification charts obtained from experiments on water films using different methods. Figure 3 (a) The traditional, uncorrected variable flip angle method (2D VFA) is used. Figure 3 (b) The method used is the DM-based VFA, which is based on the dictionary matching method of this invention. Figure 3 (c) The method of the present invention is used. Figure 3 (d) The gold standard method (IR) was used. We selected twelve regions of interest (ROIs) for statistical analysis, as shown in Table 2. The experimental statistical results show that the method of the present invention is closer to the results of the gold standard method (IR), and the signal-to-noise ratio is also higher than that of the traditional variable flip angle method, which verifies that the method of the present invention has good accuracy and high signal-to-noise ratio for water film experiments.

[0072] Table 2

[0073]

[0074]

[0075] like Figure 4 (a) and Figure 4 (b) shows the quantitative results of T1 parameters of layers 19 and 23 obtained by the corrected variable flip angle method (DMbased VFA) based on the dictionary matching method of the present invention and the method of the present invention in human brain experiments. We selected four regions of interest (ROIs) in the figure for statistical analysis. The results are shown in Table 3. This proves that the method framework of the present invention is also applicable to the traditional variable flip angle method. Under the same method framework, the sequence of the present invention can achieve a signal-to-noise ratio comparable to that of the traditional method at a layer thickness of 3 mm with a layer thickness of 1.5 mm. This also further proves that the results of the method of the present invention are in good agreement with those of the cited literature.

[0076] The above experiments were all completed on a Siemens 3T Prisma scanner with the following parameter definitions in the examples:

[0077] T1: longitudinal relaxation time, the time required for the longitudinal magnetization vector to recover to 63% of the original from 0%.

[0078] TR: repetition time, the time interval between two adjacent excitations of the sequence.

[0079] TI: the time interval between the 180° inversion pulse and the 90° excitation pulse.

[0080] B1 + : the radio frequency excitation field strength, B1 + Quantitative maps usually only show the case of radio frequency excitation field distribution.

[0081] Table 3

[0082]

[0083] The above description of the embodiments is for the purpose of enabling a person of ordinary skill in the art to understand and apply the present application. Those skilled in the art can obviously make various modifications to the above embodiments, and apply the general principles described herein to other embodiments without having to go through creative labor. Therefore, the present application is not limited to the above embodiments, and any improvements and modifications made to the present application by those skilled in the art based on the disclosure of the present application should be within the scope of protection of the present application.

Claims

1. A high signal-to-noise ratio thin layer T1 parameter quantitative imaging method based on radio frequency encoding pulse, comprising the following steps: (1) Designing a two-dimensional radio frequency encoding pulse based on M thin layers after amplitude modulation and phase modulation of the radio frequency pulse generated based on the tissue parameters and scanning parameters required for quantification, wherein the tissue parameter is longitudinal relaxation time T1, and M is the number of thin layers contained in each thick layer; The amplitude modulation has a total of (N!) M The combination mode is based on the principle that each two-dimensional thin layer needs to be subjected to N times of flip angle modulation with an angle of α n Each combination mode contains N amplitude modulation modes, and in each amplitude modulation mode, the amplitude of the two-dimensional thin layer pulse at M different positions in a thick layer region is set, corresponding to a combination of a flip angle. The phase modulation is to perform phase modulation of π size on any two-dimensional thin layer pulse in each thick layer region relative to other thin layers, and each two-dimensional thin layer pulse in the thick layer region is phase modulated once, and a total of M phase modulations are required; (2) Combining the two-dimensional radio frequency encoding pulse with a two-dimensional gradient echo sequence to generate a magnetic resonance data acquisition sequence and import it into a magnetic resonance scanner, using the magnetic resonance scanner to divide the brain part of the subject into a plurality of thick layer regions combined by thin layers located at different layer positions, and then simultaneously scanning and reconstructing to obtain L brain thick layer images excited by different radio frequency encoding pulses, L=M*N, and N is the number of flip angles; (3) Separating the L brain thick layer images using matrix inversion reconstruction to obtain M thin layer images under N different flip angles; (4) Consideration of radio frequency excitation field B1 + Non-uniform, slice-wise excitation flip angle profile, and signal crosstalk between adjacent thick slices. Design corrected steady-state signal equation, and give B1 + and the dynamic range and discretization step size of the desired quantitative tissue parameter T1, build a dictionary of thick slices reflecting signal intensity as a function of flip angle and with respect to different B1 + sizes; (5) Separate the thick layer dictionary based on matrix inversion reconstruction method to obtain thin layer dictionary reflecting signal intensity changes with flip angle and about different B1 + size, then collect the brain B1 + quantitative map, and select the corresponding B1 + size thin layer dictionary according to the B1 + value of each pixel point in the B1 + size thin layer dictionary; (6) Matching the change relationship between the signal size of each pixel point in the thin layer image and the flip angle signal in the thin layer dictionary one by one, thereby indexing the specific tissue parameter T1 for each pixel point, and further obtaining the quantitative image of the tissue parameter T1.

2. The high SNR thin-slice Tl parameter quantitative imaging method of claim 1, characterized in that: The two-dimensional radio frequency encoding pulse designed in the step (1) is superimposed by M two-dimensional thin layer pulses distributed along the slice direction, all tissues in the thick layer region composed of the positions of the thin layers can be excited simultaneously at one time, the position distribution of the thin layers is continuous or discontinuous; the amplitude and phase of each two-dimensional thin layer pulse are designed according to the required parameters, including the number N of flip angles, the size α of flip angles n , and the number M of thin layers contained in each thick layer.

3. The high SNR thin-slice Tl parameter quantitative imaging method of claim 1, characterized in that: The amplitude and phase modulation process in step (1) is performed simultaneously on all M two-dimensional thin layer pulses in a thick layer region, i.e. for each amplitude combination containing N amplitude modulation modes, M phase modulations are required, and a total of L radio frequency encoding pulses are obtained; then using these radio frequency encoding pulses to excite the tissue to be scanned, the repetition interval between adjacent two-dimensional radio frequency encoding pulses in the same layer during each excitation is a TR, and different thick layer interleaved excitation is used in the same TR to achieve whole brain coverage scanning.

4. The high SNR thin-slice Tl parameter quantitative imaging method of claim 1, characterized in that: The matrix inversion reconstruction method used in step (3) is based on Tikhonov regularization reconstruction, and the specific expression is as follows: X = (A T A + λI) -1 A T Y Wherein: Y is the brain thick layer image, X is the thin layer image, A is the radio frequency encoding matrix, λ is the Tikhonov regularization parameter, and I is the unit matrix.

5. The high SNR thin-slice Tl parameter quantitative imaging method of claim 1, wherein: The radio frequency excitation field B1 in the step (4) + The correction of the non-uniform, slice excitation flip angle profile and signal crosstalk between adjacent thick slices is achieved by simulating the flip angle profile along the slice direction under the current scan parameter settings using Bloch equation simulation, and then a dictionary is established based on the steady-state signal equation. The expression of the corrected steady-state signal equation is as follows: wherein: E 1 / 2 = exp(-TR / 2T1), E1= exp(-TR / T1), S is the intensity of the acquired thick slice image signal, M0is the initial net magnetization vector, K is the number of discrete sampling points, Δz is the interval of adjacent sampling points in the slice direction, TR is the repetition interval between adjacent two-dimensional radio frequency encoding pulses, i is the imaginary unit, k is the serial number of the slice direction discrete sampling point, α1(k) is the flip angle size of the kth discrete sampling point corresponding slice, α2(k) represents the flip angle size of the pulse at the slice position of the upper and lower two adjacent slices of the kth discrete sampling point corresponding slice when the pulse was excited last time, and Φ(k) is the phase of the transverse magnetization vector at the slice position of the kth discrete sampling point.

6. The high SNR thin-slice Tl parameter quantitative imaging method of claim 1, wherein: The step (4) establishes about different B1 + The thick layer dictionary of different sizes is established by B1 + The correction of the flip angle in the steady-state signal equation is realized, and the correction expression is as follows: where: a norm is the flip angle setting, a corr is the corrected flip angle size, r denotes the position of any voxel in the brain slab image in space, B1 + (r) is the radio frequency excitation field strength at position r.

7. The high SNR thin-slice Tl parameter quantitative imaging method of claim 1, wherein: the brain B1 in step (5) + The quantitative maps were obtained by importing the magnetic resonance data acquisition sequence for B1 + quantification into a magnetic resonance scanner, scanning the subject's brain with the scanner and reconstructing the obtained multi-slice brain radio frequency excitation field B1 + quantification maps.

Images