Method for characterizing the diffusion properties at the intravoxel micrometer scale by magnetic resonance imaging

Abstract

A method for characterizing the microscale diffusion characteristics of a single voxel in magnetic resonance imaging, comprising: S1, constructing a probability density function of a plurality of single-voxel diffusion characteristic distributions; S2, designing a special sequence to scan and image a target object to obtain scan data for constructing the probability density function of the plurality of single-voxel diffusion characteristic distributions; and S3, calculating the probability density function of the plurality of single-voxel diffusion characteristic distributions in step S1 according to the imaging sequence parameters and the scan data in step S2 to obtain the microscale diffusion characteristic results of the single voxel. The method for characterizing the microscale diffusion characteristics of a single voxel in magnetic resonance imaging solves the problem of how to use a single voxel in millimeter-scale magnetic resonance imaging to characterize the complex diffusion characteristics associated with the microscale tissue structure in the single voxel.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of magnetic resonance imaging technology, in particular to a method for characterizing the micron-scale diffusion properties within a single voxel in magnetic resonance imaging. BACKGROUND

[0002] Diffusion Magnetic Resonance Imaging (Diffusion MRI) can image the diffusion motion of water molecules in biological tissues, reflecting the functional information related to the microstructure properties of tissues. Since the diffusion motion of water molecules in biological tissues is closely related to various life activities, Diffusion MRI has extremely important basic research value and clinical application value. After more than thirty years of development, Diffusion MRI technology has made important breakthroughs in many fields such as neuroscience and oncology.

[0003] Since the 1980s of last century, scientists have continuously promoted the development of Diffusion MRI technology, and have successively proposed various Diffusion MRI techniques. Le Bihan et al. proposed diffusion weighted imaging technology in 1986, which uses the apparent diffusion coefficient (ADC) value of each voxel to reflect the average diffusion properties in the voxel. Considering that the diffusion motion of water molecules and blood microcirculation in a single voxel both contribute to the ADC value, in order to distinguish the diffusion motion of water molecules and blood microcirculation in a single voxel, scientists further proposed intravoxel incoherent motion magnetic resonance imaging technology, which reflects the diffusion motion of water molecules through the diffusion coefficient, and reflects the blood microcirculation through the pseudo-diffusion coefficient and perfusion fraction. Basser et al. proposed diffusion tensor imaging in 1994, which usually uses the partial anisotropy parameter to characterize the anisotropy degree of biological tissues, and uses the average diffusion coefficient parameter to represent the average water molecule diffusion coefficient measured in multiple directions. Jensen et al. proposed diffusion kurtosis imaging technology for quantifying non-Gaussian diffusion effects in 2005, which uses the average kurtosis to characterize the degree of water molecule diffusion displacement deviating from the Gaussian distribution. In 2011, Shao et al. proposed a filtering exchange imaging technique, which can obtain the diffusion characteristic information related to the transmembrane water exchange of cells in tissues, and then reflect the changes of the transmembrane water exchange function of cells. In 2019, Shao et al. combined Diffusion MRI technology with the filtering exchange imaging technique, and proposed a pseudo-continuous arterial spin labeling technique containing a diffusion preparation module, to obtain the diffusion characteristic information related to the water exchange across the blood-brain barrier. The above-mentioned representative Diffusion MRI techniques can reflect the functional information related to the microstructure characteristics of tissues in a certain extent from the perspective of the respective diffusion characteristic parameter values, and have been widely applied in medical basic and clinical research.

[0004] In the occurrence and development process of various diseases, the changes of tissue microstructure and its characteristics will lead to the changes of water molecule diffusion characteristics in tissues, including the diffusion characteristics related to various components in cells, the diffusion characteristics related to various components in extracellular matrix, the diffusion characteristics related to the water molecule permeability of cell membrane, the diffusion characteristics related to the volume fraction of cells in tissues, and various main diffusion characteristics related to the microstructure characteristics of tissues. When the various main diffusion characteristics related to the microstructure characteristics of tissues change, the diffusion motion state of water molecules in biological tissues will change, and it will be reflected in the Diffusion MRI image results. However, the above-mentioned Diffusion MRI techniques reflect the combined effect of various main diffusion characteristics related to the microstructure characteristics of tissues in a single voxel by assigning one or several diffusion characteristic parameter values to a single voxel, and cannot effectively distinguish the diffusion characteristics related to various microstructure characteristics of tissues in a single voxel.

[0005] If the information reflecting the respective changes of various main diffusion characteristics related to the microstructure characteristics of tissues in a single voxel can be distinguished, it is possible to distinguish the changes of cell membrane function characteristics, the intracellular microenvironment and the extracellular matrix microenvironment through Diffusion MRI images, which has great scientific value for exploring the dynamic change process of disease occurrence and development and dynamically revealing the mechanism of disease in vivo. However, so far, how to use millimeter-scale magnetic resonance imaging single voxel to represent the complex diffusion characteristics related to the microstructure of micron-scale tissues in a single voxel has been a scientific problem that has not been solved.

[0006] Therefore, in view of the deficiencies in the prior art, it is necessary to provide a method for representing the micron-scale diffusion characteristics in a magnetic resonance imaging single voxel to overcome the deficiencies in the prior art. SUMMARY

[0007] The present application aims at avoiding the shortcomings of the prior art and providing a method for characterizing the microscale diffusion characteristics in a single voxel in magnetic resonance imaging, which obtains the complex diffusion characteristic information associated with the microstructure of the tissue in the single voxel at the microscale by using a plurality of probability density functions for characterizing the diffusion characteristic distribution in the single voxel.

[0008] The present application provides a method for characterizing the microscale diffusion characteristics in a single voxel in magnetic resonance imaging, which comprises the following steps:

[0009] S1, constructing a plurality of probability density functions for the diffusion characteristic distribution in the single voxel;

[0010] S2, designing a special sequence to perform scanning imaging on a target object to obtain scanning data for constructing the plurality of probability density functions for the diffusion characteristic distribution in the single voxel;

[0011] S3, according to the imaging sequence parameters and the scanning data in step S2, calculating the plurality of probability density functions for the diffusion characteristic distribution in the single voxel in step S1 to obtain the diffusion characteristic results in the single voxel at the microscale.

[0012] Preferably, in the method for characterizing the microscale diffusion characteristics in a single voxel in magnetic resonance imaging, the probability density function for the diffusion displacement distribution of water molecules in the single voxel, the probability density function for the diffusion coefficient distribution of water molecules in the single voxel and the probability density function for the diffusion anisotropy distribution of water molecules in the single voxel are constructed in step S1.

[0013] The specific process for constructing the probability density function for the diffusion displacement distribution of water molecules in the voxel is as follows:

[0014] The diffusion process of water molecules is described by using the probability density function for the diffusion displacement of water molecules within a given diffusion time, and the probability density function for the diffusion displacement of water molecules from a starting point x0 to a terminal point x1 within a diffusion time t d is represented as P r (x1, x0, t d ).

[0015] In the case of a single voxel, the k-space signal S(k, q) and the probability density function P(x1, x0, t d ) for characterizing the diffusion displacement distribution of water molecules in the single voxel satisfy the relationship of formula (1):

[0016]

[0017] In formula (1), the vector k = γt s G / 2π, q = γδG d / 2π, γ represents the magnetic rotation ratio, G represents the spatial encoding gradient, G d represents the diffusion encoding gradient, and t swhere δ represents the duration of the diffusion encoding gradient, and p(x0) represents the nuclear spin density within a single voxel;

[0018] The inverse Fourier transform of S(q) directly gives P(x1, x0, t d ); however, P(x1, x0, t d ) represents the water molecule displacement which contains both the water molecule diffusion motion and the water molecule displacement caused by the blood microcirculation; in order to obtain the probability density function P r (x1, x0, t d ) of the water molecule diffusion displacement distribution within a voxel, S(q) needs to be divided into S d (q) and S m (q), S d (q) and S m (q) represent the water molecule diffusion motion related to the microstructure and the pseudo-diffusion motion magnetic resonance signal related to the blood microcirculation, respectively; then the inverse Fourier transform of S d (q) is performed to obtain P r (x1, x0, t d ).

[0019] The specific processing process is as follows:

[0020] The inverse Fourier transform of the k-space signal S(k, q) gives the image signal S(q), and then S(q) is divided into two parts by formula (2) to obtain S d (q), and the inverse Fourier transform of S d (q) gives P r (x1, x0, t d );

[0021] S(q) = S d (q) + S m (q) …… formula (2);

[0022] where S d (q) and S m (q) represent the water molecule diffusion motion related to the microstructure and the pseudo-diffusion motion magnetic resonance signal related to the blood microcirculation, respectively.

[0023] The specific process of constructing the probability density function of the water molecule diffusion coefficient distribution within a voxel is as follows:

[0024] The diffusion magnetic resonance signal within a single voxel is the sum of the diffusion magnetic resonance signals generated by all nuclear spins within the voxel;

[0025] To further analyze the diffuse magnetic resonance signal within a single voxel, we consider that the nuclear spins within the voxel are distributed in different spatial locations, and that each nuclear spin is constrained and hindered by the microstructure. To examine the distribution of the dispersion coefficient D of each nuclear spin, the probability density function of the water molecule dispersion coefficient distribution within a single voxel is expressed as P. d (D), P d In (D), the subscript d is the lowercase of the diffusion coefficient D, indicating that the probability is related to the diffusion coefficient;

[0026] To distinguish the contributions of intravoxel water molecule diffusion motion and blood microcirculation to the diffusion coefficient D, let D... * D represents the diffusion coefficient related to the diffusion motion of water molecules, and D′ represents the pseudo-diffusion coefficient related to blood microcirculation; for D * Discretization is performed in D * Divide the range into M discrete D * Values, respectively using Let m be a natural number from 1 to M; after discretization, the probability density function P is... d (D) is also discretized as

[0027] When the number of scans is N, the relationship in equation (3) exists:

[0028]

[0029] Satisfying Relationships S n Sn represents the diffuse magnetic resonance signal of a single voxel obtained in the nth scan, S0 represents the magnetic resonance signal of a single voxel obtained without applying a diffuse coding gradient, and b n It is the diffusion weighting factor used in the nth scan, and the coefficient f represents the proportion of magnetic resonance signal attenuation caused by blood microcirculation to the total magnetic resonance signal attenuation.

[0030] The acquired diffuse magnetic resonance signal is fitted to Equation (3) using the non-negative regularized least squares method. The expression for the non-negative regularized least squares fitting problem is shown in Equation (4):

[0031]

[0032] The first and second terms in equation (4) represent the fitting error and regularization constraint, respectively, and μ represents the regularization factor.

[0033] get Again By performing Nth-order polynomial fitting, the probability density function P of the dispersion coefficient distribution of water molecules within a continuous voxel is obtained. d (D), the fitting formula is as follows:

[0034]

[0035] N is the fitting degree, N is determined according to actual situation, ω j represents the coefficient of polynomial fitting;

[0036] The process of constructing the probability density function of the intra- voxel water molecule diffusion anisotropy distribution is as follows:

[0037] The water molecule diffusion anisotropy is measured by the relative anisotropy (RA), and the relative anisotropy RA is defined as:

[0038]

[0039] wherein λ is the eigenvalue of the diffusion tensor D, Var(λ) represents the variance of the eigenvalue λ of the diffusion tensor D, is the average diffusion coefficient;

[0040] The probability density function of the intra- voxel water molecule anisotropy distribution is represented by P a (RA), and the steps of calculating P a (RA) are as follows: firstly, the probability density function of the diffusion tensor distribution needs to be obtained, and then based on the probability density function of the diffusion tensor distribution, P a (RA) is obtained, and the specific calculation process is as follows:

[0041] Firstly, the method represented by formula (3) and formula (4) is used to obtain the discrete intra- voxel water molecule diffusion tensor distribution probability density function Then, the respective discrete diffusion tensor corresponding eigenvalue λ, variance of eigenvalue Var(λ) and average diffusion coefficient are obtained, and the respective correlation anisotropy RA of the discrete diffusion tensor is calculated according to formula (6), and based on the diffusion tensor and the correlation anisotropy RA one-to-one corresponding relationship, the discrete intra- voxel water molecule anisotropy distribution probability density function P a (RA * ) is obtained; and then P a (RA * ) is subjected to Y times polynomial fitting to obtain the continuous intra- voxel water molecule anisotropy distribution probability density function P a (RA), and the fitting formula is as follows:

[0042]

[0043] In equation (7), Y represents the number of fitting iterations. The value of Y is determined based on the actual situation, and λ j These represent the coefficients of the polynomial fit.

[0044] Preferably, the above-described method for characterizing the diffusion properties of intracellular elements at the micrometer scale in magnetic resonance imaging (MRI) uses P... r (x1,x0,t d The first quartile, median, third quartile, and the dispersion shift of water molecules with the highest probability were used as quantitative indicators.

[0045] P d (D) The first quartile, median, third quartile, and the water molecule dispersion coefficient with the highest probability are used as quantitative indicators;

[0046] P a The first quartile, median, third quartile, and the relative anisotropy RA with the highest probability (RA) are used as quantitative indicators.

[0047] Preferably, in the above-mentioned method for characterizing the diffusion properties of intracellular micrometer-scale magnetic resonance imaging, a dedicated sequence is designed in S2, specifically including:

[0048] Design a diffusion-weighted imaging stimulated echo acquisition sequence to provide scanning data for constructing the probability density function of the diffusion displacement distribution of water molecules within a voxel;

[0049] The design series includes diffusion-weighted imaging stimulated echo acquisition sequence, diffusion-weighted imaging oscillating gradient spin echo acquisition sequence, and diffusion-weighted imaging pulse gradient spin echo acquisition sequence, to provide scanning data for constructing the probability density function of the diffusion coefficient distribution of water molecules within voxels.

[0050] We designed a diffusion-weighted imaging dual-pulse gradient spin echo acquisition sequence to provide scanning data for constructing the probability density function of the diffusion anisotropic distribution of water molecules within a voxel.

[0051] Preferably, in the above-mentioned method for characterizing the diffusion characteristics of intracellular micron-scale magnetic resonance imaging, the diffusion-weighted imaging stimulated echo acquisition sequence can simultaneously achieve low b-value and high b-value sampling.

[0052] The diffusion-weighted imaging oscillating gradient spin echo acquisition sequence, diffusion-weighted imaging pulse gradient spin echo acquisition sequence, and diffusion-weighted imaging stimulated echo acquisition sequence respectively achieve diffusion measurement of short diffusion time (usually <10ms), medium diffusion time (usually between 10ms and 100ms), long diffusion time (usually >100ms) and multiple b values.

[0053] The diffusion weighted imaging double-pulse gradient spin echo acquisition sequence comprises two diffusion weighted imaging pulse gradient spin echo acquisition sequence modules, and each of the two diffusion weighted imaging pulse gradient spin echo acquisition sequence modules has a respective b value and diffusion gradient direction.

[0054] Preferably, in the actual scanning process, the above-mentioned magnetic resonance imaging intravoxel microscale diffusion property characterization method also optimizes the key parameters of the diffusion weighted imaging stimulated echo acquisition sequence, the diffusion weighted imaging oscillating gradient spin echo acquisition sequence, the diffusion weighted imaging pulse gradient spin echo acquisition sequence, and the diffusion weighted imaging double-pulse gradient spin echo acquisition sequence. The optimized sequence parameters specifically refer to the diffusion time t d , echo time, b value, diffusion gradient direction, and mixing time, and the specific parameters are as follows:

[0055] A. During scanning, the diffusion time t d is adjusted to correspond to the change of different microstructure properties. d is Δ-δ / 3, δ is the duration of a single diffusion gradient, and Δ is the time interval between two diffusion gradients.

[0056] B. The echo time is adjusted to change the sensitivity of the diffusion magnetic resonance signal to the diffusion properties associated with each microstructure property.

[0057] C. The size and number of the diffusion weighting factor b value are adjusted to change the acquired diffusion magnetic resonance signal.

[0058] D. The diffusion gradient direction is adjusted to change the sensitivity of the diffusion magnetic resonance signal to the diffusion properties associated with each microstructure property.

[0059] E. The mixing time of the diffusion weighted imaging stimulated echo acquisition sequence and the diffusion weighted imaging double-pulse gradient spin echo acquisition sequence is adjusted to change the sensitivity of the diffusion magnetic resonance signal to the diffusion properties associated with each microstructure property.

[0060] A method for characterizing intravoxel microscale diffusion properties in magnetic resonance imaging, characterized by comprising the following steps: S1, constructing a plurality of intravoxel diffusion property distribution probability density functions; S2, designing a special sequence to scan the target object to obtain scan data for constructing the plurality of intravoxel diffusion property distribution probability density functions; S3, according to the imaging sequence parameters and scan data of step S2, corresponding calculation of the plurality of intravoxel diffusion property distribution probability density functions in step S1, to obtain the intravoxel microscale diffusion property results. The application obtains scan data by designing a diffusion weighted imaging special sequence related to tissue microstructure characteristics, and constructs a plurality of intravoxel diffusion property distribution probability density functions related to various main tissue microstructure characteristics, such as "intravoxel water molecule diffusion displacement distribution probability density function, intravoxel water molecule diffusion coefficient distribution probability density function, and intravoxel water molecule diffusion anisotropy distribution probability density function". The application proposes a method for characterizing intravoxel microscale diffusion properties in magnetic resonance imaging, constructs a plurality of intravoxel diffusion property distribution probability density functions related to various main tissue microstructure characteristics, and designs a special sequence for constructing a plurality of intravoxel diffusion property distribution probability density functions. The method for characterizing intravoxel microscale diffusion properties in magnetic resonance imaging solves the problem of how to use millimeter-scale magnetic resonance imaging single voxel to characterize the complex diffusion properties related to microscale tissue structure in single voxel.

[0061] Drawings

[0062] The application is further illustrated by the drawings, but the contents of the drawings do not constitute any limitation on the application.

[0063] Figure 1 It is a flowchart of a method for characterizing intravoxel microscale diffusion properties in magnetic resonance imaging.

[0064] Figure 2 It is an example of a diffusion weighted imaging stimulated echo acquisition sequence.

[0065] Figure 3 It is an example of a diffusion weighted imaging pulsed gradient spin echo acquisition sequence.

[0066] Figure 4 It is an example of a diffusion weighted imaging oscillating gradient spin echo acquisition sequence.

[0067] Figure 5 It is an example of a diffusion weighted imaging double pulsed gradient spin echo acquisition sequence.

[0068] Figure 6 This is an application diagram of the resonance imaging method for characterizing the diffusion properties of intracellular micron-scale particles in embodiment 4 of the present invention. Wherein, Figure 6 A shows a representative MRI image. Figure 6 The boxes in A represent individual voxels; Figure 6 B and Figure 6 C represents what? Figure 6 The probability density function P of the diffuse displacement distribution of water molecules within voxels 1 and 2 in A. r (x1,x0,t d The probability density function P of the dispersion coefficient distribution of water molecules within a voxel d (D), and the probability density function P of the anisotropic distribution of water molecules within the voxel. a (RA). Figure 6 D shows the relationship with Figure 6 A staining image of a tissue section of the same size as voxel 3 in A. Figure 6 E and Figure 6 G represents what... Figure 6 A schematic diagram showing the possible arrangement of cells, extracellular matrix, etc. in regions 1 and 2 of D. Figure 6 F shows what Figure 6 The distribution of possible diffuse displacement, diffusion coefficient, and relative anisotropy values ​​within the spatial region from point a to point f in E. Figure 6 H shows what Figure 1 The distribution of possible diffuse displacement, diffusion coefficient, and relative anisotropy values ​​within the spatial region from point a to point g in G. Detailed Implementation

[0069] The present invention will be further described in conjunction with the following embodiments.

[0070] Example 1.

[0071] The objective of this invention is to characterize tissue microstructure using magnetic resonance imaging (MRI) and achieve MRI tomography of tissue microstructure. To achieve this objective, it is necessary to address the problem of how to use millimeter-scale MRI monomers to characterize the complex diffusion characteristics within monomers that correlate with micrometer-scale tissue structures.

[0072] The present application needs to design a plurality of probability density functions for representing the distribution of diffusion characteristics in a voxel, to obtain the complex diffusion characteristic information associated with the microstructure of the tissue in a single voxel. Specifically, how to construct an imaging method capable of representing the diffusion characteristics associated with the microstructure of the tissue by diffusion magnetic resonance imaging. Specifically, it includes but is not limited to representing the complex changes of diffusion characteristics associated with the microstructure of the tissue, such as "the diffusion characteristics associated with various components in cells, the diffusion characteristics associated with various components in extracellular matrix, the diffusion characteristics associated with the permeability of cell membrane water molecules, and the diffusion characteristics associated with the volume fraction of cells in tissue". The representation method includes using a plurality of probability density functions for representing the distribution of diffusion characteristics in a voxel, such as "the probability density function of water molecule diffusion displacement distribution in a single voxel, the probability density function of water molecule diffusion coefficient distribution in a single voxel, and the probability density function of water molecule diffusion anisotropy distribution in a single voxel", to obtain the complex diffusion characteristic information associated with the microstructure of the tissue in a single voxel.

[0073] The present application is a method for representing the diffusion characteristics in a single voxel in a magnetic resonance imaging microscale, which includes the following steps:

[0074] S1, construct a plurality of probability density functions for representing the distribution of diffusion characteristics in a single voxel. Specifically, a plurality of probability density functions for representing the distribution of diffusion characteristics in a single voxel can be constructed, such as "the probability density function of water molecule diffusion displacement distribution in a single voxel, the probability density function of water molecule diffusion coefficient distribution in a single voxel, and the probability density function of water molecule diffusion anisotropy distribution in a single voxel".

[0075] S2, design a special sequence to scan and image the target object, to obtain the scanning data for constructing a plurality of probability density functions for representing the distribution of diffusion characteristics in a single voxel.

[0076] Design a special sequence to provide scanning data for constructing a plurality of probability density functions for representing the distribution of diffusion characteristics in a single voxel. Specifically, it includes: designing a special diffusion weighted imaging stimulated echo acquisition sequence to provide scanning data for constructing the probability density function of water molecule diffusion displacement distribution in a voxel; designing a special series sequence, including the above-mentioned diffusion weighted imaging stimulated echo acquisition sequence, diffusion weighted imaging oscillating gradient spin echo acquisition sequence, and diffusion weighted imaging pulse gradient spin echo acquisition sequence, to provide scanning data for constructing the probability density function of water molecule diffusion coefficient distribution in a voxel; and designing a special diffusion weighted imaging double-pulse gradient spin echo acquisition sequence to provide scanning data for constructing the probability density function of water molecule diffusion anisotropy distribution in a voxel.

[0077] S3, according to the imaging sequence parameters and scanning data of step S2, calculate the plurality of probability density functions for representing the distribution of diffusion characteristics in a single voxel in step S1, to obtain the diffusion characteristic results in a single voxel in a microscale.

[0078] The application provides a method for characterizing the microscale diffusion characteristics in a single voxel in magnetic resonance imaging, constructs a plurality of probability density functions of the diffusion characteristics distribution in a single voxel which can reflect the diffusion characteristic changes associated with various main microstructure characteristics of tissues, and designs a special sequence for providing scanning data for constructing the probability density functions of the diffusion characteristics distribution in a plurality of single voxels.

[0079] Embodiment 2.

[0080] A method for characterizing the microscale diffusion characteristics in a single voxel in magnetic resonance imaging, comprising the following steps:

[0081] S1, constructing a plurality of probability density functions of the diffusion characteristics distribution in a single voxel. Figure 2 S1, constructing a plurality of probability density functions of the diffusion characteristics distribution in a single voxel.

[0082] The specific process of constructing the probability density function of the diffusion displacement distribution of water molecules in a voxel is as follows:

[0083] The diffusion process of water molecules is described by using the probability density function of the diffusion displacement of water molecules in a given diffusion time. In the diffusion time t d , the probability density function of water molecules moving from the starting point x0 to the ending point x1 is represented as P r (x1, x0, t d ).

[0084] In the case of a single voxel, the k-space signal S(k, q) and the probability density function P(x1, x0, t d ) representing the diffusion displacement distribution of water molecules in a single voxel satisfy the relationship of formula (1):

[0085] S(k, q) = ∫∫ρ(x0)P(x1, x0, t d )exp(i2π(k·x0+q·(x1-x0)))dx1dx0...... formula (1)

[0086] In formula (1), the vector k = γt s G / 2π, q = γδG d / 2π, γ represents the magnetic spin ratio, G represents the spatial encoding gradient, G d represents the diffusion encoding gradient, t sδ represents the duration of the spatial coding gradient, ρ(x0) represents the duration of the diffuse coding gradient, and ρ(x0) represents the nuclear spin density within a monomer.

[0087] The image signal S(q) is obtained by performing an inverse Fourier transform on the k-space signal S(k,q). Then, S(q) is divided into two parts using equation (2) to obtain S. d (q), then for S d (q) Perform an inverse Fourier transform to obtain P r (x1,x0,t d );

[0088] S(q)=S d (q)+S m (q)……Equation (2);

[0089] Among them, S d (q) and S m (q) represents the magnetic resonance signals of water molecule diffusion motion related to microstructure and pseudo-diffusion motion related to blood microcirculation, respectively.

[0090] To obtain the P values ​​corresponding to the changes in diffusion properties associated with the four main tissue microstructure properties, r (x1,x0,t d The variation pattern of P is considered while analyzing the entire probability density function distribution curve. r (x1,x0,t d The first quartile, median, third quartile, and the dispersion shift of water molecules with the highest probability were used as quantitative indicators.

[0091] The specific process for constructing the probability density function of the dispersion coefficient distribution of water molecules within a voxel is as follows:

[0092] To further analyze the diffuse magnetic resonance signal within a single voxel, we consider that the nuclear spins within the voxel are distributed in different spatial locations, and that each nuclear spin is constrained and hindered by the microstructure. To examine the distribution of the dispersion coefficient D of each nuclear spin, the probability density function of the water molecule dispersion coefficient distribution within a single voxel is expressed as P. d (D), P d In (D), the subscript d is the lowercase of the diffusion coefficient D, indicating that the probability is related to the diffusion coefficient;

[0093] To distinguish the contributions of intravoxel water molecule diffusion motion and blood microcirculation to the diffusion coefficient D, let D... * D represents the diffusion coefficient related to the diffusion motion of water molecules, and D′ represents the pseudo-diffusion coefficient related to blood microcirculation; for D * Discretization is performed in D * Divide the range into M discrete D * Values, respectively using Let m be a natural number from 1 to M; after discretization, the probability density function P is... d (D) is also discretized as

[0094] When the number of scans is N, the relationship in equation (3) exists:

[0095]

[0096] Satisfying Relationships S n Sn represents the diffuse magnetic resonance signal of a single voxel obtained in the nth scan, S0 represents the magnetic resonance signal of a single voxel obtained without applying a diffuse coding gradient, and b n It is the diffusion weighting factor used in the nth scan, and the coefficient f represents the proportion of magnetic resonance signal attenuation caused by blood microcirculation to the total magnetic resonance signal attenuation.

[0097] The acquired diffuse magnetic resonance signal is fitted to Equation (3) using the non-negative regularized least squares method. The expression for the non-negative regularized least squares fitting problem is shown in Equation (4):

[0098]

[0099] The first and second terms in equation (4) represent the fitting error and regularization constraint, respectively, and μ represents the regularization factor.

[0100] get Again By performing Nth-order polynomial fitting, the probability density function P of the dispersion coefficient distribution of water molecules within a continuous voxel is obtained. d (D), the fitting formula is as follows:

[0101]

[0102] In equation (5), N represents the number of fitting iterations, which is determined based on the actual situation. j These represent the coefficients of the polynomial fit.

[0103] To further obtain the P values ​​corresponding to the changes in diffusion properties associated with the four main tissue microstructure properties, d The variation pattern of (D) should be considered while analyzing the entire probability density function distribution curve, taking into account P. d (D) uses the first quartile, median, third quartile, and the water molecule dispersion coefficient with the highest probability as quantitative indicators.

[0104] The process of constructing the probability density function of the diffuse anisotropic distribution of water molecules within a voxel is as follows:

[0105] The water molecule diffusion anisotropy is measured by the relative anisotropy (RA), which is defined as:

[0106]

[0107] where λ is the eigenvalue of the diffusion tensor, Var(λ) represents the variance of the eigenvalue λ of the diffusion tensor, is the average diffusion tensor;

[0108] The probability density function of the intravoxel water molecule anisotropy distribution is represented by P a (RA), and the steps for calculating P a (RA) are as follows: firstly, the probability density function of the diffusion tensor distribution is obtained, and then P a (RA) is obtained based on the probability density function of the diffusion tensor distribution. The specific calculation process is as follows:

[0109] Firstly, the method represented by formula (3) and formula (4) is used to obtain the probability density function of the discrete intravoxel water molecule diffusion tensor distribution Then, the corresponding eigenvalue λ, the variance Var(λ) of the eigenvalue λ, and the average diffusion coefficient of each discrete diffusion tensor are obtained, and the relative anisotropy RA corresponding to each discrete diffusion tensor is calculated according to formula (6), and the relationship between the diffusion tensor and the relative anisotropy RA is one-to-one, so that the probability density function P a (RA * ) of the discrete intravoxel water molecule anisotropy distribution is obtained; then P a (RA * ) is subjected to Y times polynomial fitting to obtain the continuous intravoxel water molecule anisotropy distribution probability density function P a (RA), and the fitting formula is as follows:

[0110]

[0111] Y in formula (7) is the number of fitting, and the value of Y is determined according to the actual situation, λ j represents the coefficient of polynomial fitting.

[0112] In order to further obtain the change rule of P a (RA) corresponding to the change of the diffusion characteristics related to the four main tissue microstructure characteristics, the first quartile, the median, the third quartile, and the relative anisotropy RA with the maximum probability of P a (RA) are considered as quantitative indexes when analyzing the whole probability density function distribution curve.

[0113] After the probability density function of the multiple single voxel diffusion characteristics distribution is constructed, the next step is to design a special sequence S2.

[0114] S2, design a special sequence, scan and image the target object to obtain the scanning data for constructing the probability density function of the multiple single voxel diffusion characteristics distribution.

[0115] Design a series of sequences, including a diffusion weighted imaging stimulated echo acquisition sequence, a diffusion weighted imaging oscillating gradient spin echo acquisition sequence, and a diffusion weighted imaging pulsed gradient spin echo acquisition sequence, to provide scanning data for constructing the probability density function of the voxel water molecule diffusion coefficient distribution.

[0116] Design a diffusion weighted imaging double pulsed gradient spin echo acquisition sequence to provide scanning data for constructing the probability density function of the voxel water molecule diffusion anisotropy distribution.

[0117] The diffusion weighted imaging stimulated echo acquisition sequence can simultaneously realize low b value and high b value sampling; the diffusion weighted imaging oscillating gradient spin echo acquisition sequence, the diffusion weighted imaging pulsed gradient spin echo acquisition sequence, and the diffusion weighted imaging stimulated echo acquisition sequence realize diffusion measurement of multiple b values with short diffusion time (usually <10 ms), medium diffusion time (usually between 10 ms and 100 ms), long diffusion time (usually >100 ms), respectively.

[0118] The diffusion weighted imaging double pulsed gradient spin echo acquisition sequence includes two diffusion weighted imaging pulsed gradient spin echo acquisition sequence modules, and the two diffusion weighted imaging pulsed gradient spin echo acquisition sequence modules have their own b values and diffusion gradient directions.

[0119] In the actual scanning process, the key parameters of the diffusion weighted imaging stimulated echo acquisition sequence, the diffusion weighted imaging oscillating gradient spin echo acquisition sequence, the diffusion weighted imaging pulsed gradient spin echo acquisition sequence, and the diffusion weighted imaging double pulsed gradient spin echo acquisition sequence are also optimized. The optimized sequence parameters specifically refer to the diffusion time t d , the echo time, the b value, the diffusion gradient direction, and the mixing time, which are as follows:

[0120] A. During scanning, corresponding to the changes of different microstructure characteristics, the diffusion time t d is adjusted, t d = Δ- δ / 3, δ is the duration of a single diffusion gradient, and Δ is the time interval between two diffusion gradients.

[0121] B. By adjusting the echo time, the sensitivity of the diffusion magnetic resonance signal to the diffusion characteristics associated with each microstructure characteristic is changed.

[0122] C. Adjusting the size and number of diffusion weighting factor b values, changing the acquired diffusion magnetic resonance signals;

[0123] D. Adjusting the diffusion gradient direction, changing the sensitivity of the diffusion magnetic resonance signals to the diffusion characteristics associated with each microstructure characteristic.

[0124] E. Adjusting the mixing time of the diffusion-weighted imaging stimulated echo acquisition sequence and the diffusion-weighted imaging double-pulse gradient spin echo acquisition sequence, changing the sensitivity of the diffusion magnetic resonance signals to the diffusion characteristics associated with each microstructure characteristic.

[0125] Through this step, the sequence design is completed, and the target object is scanned and imaged to obtain corresponding scan data.

[0126] S3, according to the imaging sequence parameters and scan data of step S2, the probability density function of each single voxel-in diffusion characteristic distribution designed in step S1 is brought in to obtain the probability density function of a plurality of voxel-in diffusion characteristic distributions. Thus, the micron-scale diffusion characteristic result in a single voxel is obtained.

[0127] The present application proposes a kind of magnetic resonance imaging single voxel-in micron-scale diffusion characteristic characterization method, constructs the probability density function of a plurality of voxel-in diffusion characteristic distributions that can reflect the diffusion characteristics change associated with various main tissue microstructure characteristics, designs the special sequence for providing scan data for constructing the probability density function of a plurality of voxel-in diffusion characteristic distributions.The magnetic resonance imaging single voxel-in micron-scale diffusion characteristic characterization method of the present application solves the problem of how to use millimeter-scale magnetic resonance imaging single voxel to characterize the complex diffusion characteristics associated with micron-scale tissue structure in single voxel.

[0128] Embodiment 3.

[0129] A kind of magnetic resonance imaging single voxel-in micron-scale diffusion characteristic characterization method, other features are same with embodiment 2, the difference is as follows: the specific way of designing special sequence is as follows.

[0130] In order to obtain P r (x1,x0,t d ), up to dozens of b values are required, and diffusion magnetic resonance signals are acquired along multiple directions, covering the range from low b value to high b value, and the sampling time is long, and the gradient amplitude is high. Therefore, the present embodiment designs a special diffusion-weighted imaging stimulated echo acquisition sequence capable of simultaneously sampling low b value and high b value. For example, Figure 3 is a kind of diffusion-weighted imaging stimulated echo acquisition sequence timing.

[0131] In order to obtain P d(D), different diffusion time scales and multiple b values are required for diffusion measurement. In this embodiment, special diffusion weighted imaging oscillating gradient spin echo acquisition sequence, diffusion weighted imaging pulsed gradient spin echo acquisition sequence, diffusion weighted imaging stimulated echo acquisition sequence are designed to realize short diffusion time (usually <10 ms), medium diffusion time (usually between 10 ms and 100 ms), long diffusion time (usually >100 ms) respectively. Take Figure 4 and Figure 5 as examples to show the timing of diffusion weighted imaging pulsed gradient spin echo acquisition sequence and the timing of diffusion weighted imaging oscillating gradient spin echo acquisition sequence respectively.

[0132] In order to obtain P a (RA), the diffusion magnetic resonance signal acquired needs to be sensitive to the difference of diffusion characteristics in different directions within the voxel. In this embodiment, special diffusion weighted imaging double pulsed gradient spin echo acquisition sequence is designed. Diffusion weighted imaging double pulsed gradient spin echo acquisition sequence contains two diffusion weighted imaging pulsed gradient spin echo acquisition sequence modules, both of which have their own b values and diffusion gradient directions. Take Figure 6 as an example to show the timing of diffusion weighted imaging double pulsed gradient spin echo acquisition sequence.

[0133] In this embodiment, the sequence parameters are also optimized. In the actual scanning process, the key parameters of diffusion weighted imaging stimulated echo acquisition sequence, diffusion weighted imaging oscillating gradient spin echo acquisition sequence, diffusion weighted imaging pulsed gradient spin echo acquisition sequence, diffusion weighted imaging double pulsed gradient spin echo acquisition sequence are further optimized. The optimized sequence parameters specifically refer to diffusion time t d , echo time, b value, diffusion gradient direction, mixing time. Specifically as follows:

[0134] A. In the scanning process, a shorter diffusion time t d (t d = Δ- δ / 3, δ is the duration of a single diffusion gradient, Δ is the time interval between two diffusion gradients) is set, then the diffusion displacement of water molecules generated within the diffusion time t d is shorter, and the obstacles encountered in the diffusion process are relatively less; the diffusion time t d is extended, then the obstacles encountered by water molecules in the diffusion process will increase. Corresponding to the changes of different microstructure characteristics, the selection of diffusion time t d will directly affect the detectable diffusion characteristics of water molecules.

[0135] B. When different echo times are set, the permeability of cell membrane water molecules and the transverse relaxation time of tissues will directly affect the amplitude of diffusion magnetic resonance signals. By adjusting the echo time, the sensitivity of diffusion magnetic resonance signals to the diffusion characteristics associated with each microstructure characteristic can be changed.

[0136] C. By adjusting the size and number of diffusion weighting factor b values, the acquired diffusion magnetic resonance signals can be directly changed.

[0137] D. By adjusting the diffusion gradient direction, the sensitivity of diffusion magnetic resonance signals to the diffusion characteristics associated with each microstructure characteristic can be changed.

[0138] E. By adjusting the mixing time of diffusion-weighted imaging stimulated echo acquisition sequences and diffusion-weighted imaging double-pulse gradient spin echo acquisition sequences, the sensitivity of diffusion magnetic resonance signals to the diffusion characteristics associated with each microstructure characteristic can be changed.

[0139] The present application proposes a method for characterizing the microscale diffusion characteristics within a single voxel of magnetic resonance imaging, constructs a plurality of intravoxel diffusion characteristic distribution probability density functions that can reflect the changes of diffusion characteristics associated with various main tissue microstructure characteristics, and designs a special sequence for providing scanning data for constructing the plurality of intravoxel diffusion characteristic distribution probability density functions. The method for characterizing the microscale diffusion characteristics within a single voxel of magnetic resonance imaging can use millimeter-scale magnetic resonance imaging single voxels to characterize the complex diffusion characteristics within a single voxel associated with microscale tissue structures.

[0140] Example 4.

[0141] The method for characterizing the microscale diffusion characteristics within a single voxel of magnetic resonance imaging is applied to actual scene applications.

[0142] Taking human abdominal Diffusion MRI as an example, a 3T magnetic resonance imaging instrument is used for scanning. The respiratory triggering technique is used to reduce image artifacts caused by abdominal movement. In order to obtain P r (x1,x0,t d ), a diffusion-weighted imaging stimulated echo acquisition sequence is used to acquire magnetic resonance signals, and the scanning parameters are as follows: repetition time = 7s, echo time = 112ms, mixing time = 41ms, δ = 16ms, Δ = 74.5ms, diffusion gradients are applied along three orthogonal directions, the maximum diffusion gradient strength is 60mT / m, the highest b value is 4600s / mm 2 , 31 q values linearly increase from 0 to 414cm -1 , the corresponding b value range is 0 to 4600s / mm 2 , the diffusion time is set to 80, 150, 260 and 400ms. In order to obtain P d(D) Simultaneously, diffusion-weighted imaging (DWI) oscillatory gradient spin echo acquisition sequences, DWI pulsed gradient spin echo acquisition sequences, and DWI stimulated echo acquisition sequences were used to acquire magnetic resonance signals, achieving short, medium, and long diffusion time measurements, respectively. The diffusion times for the DWI oscillatory gradient spin echo acquisition sequence were set to 4.3, 5.1, 6.5, and 9.3 ms, and the diffusion times for the DWI pulsed gradient spin echo acquisition sequence were set to 21, 35, 46, and 60 ms. Other parameters were the same as those for the DWI stimulated echo acquisition sequence. To obtain P... a (RA) was performed using diffusion-weighted imaging with a dual-pulse gradient spin echo acquisition sequence of magnetic resonance signals. The scanning parameters were as follows: repetition time = 3 s, echo time = 50 ms, blending time = 21 ms, δ = δ1 = 5 ms, Δ = Δ1 = 16 ms. Diffusion gradients were applied along 21 directions, with six b values ​​of 0, 500, 1000, 1500, 2000, and 2500 s / mm. 2 .

[0143] After the magnetic resonance scan is completed, the invented reconstruction algorithm is used to process the magnetic resonance signals acquired using different sequences, and P is obtained respectively. r (x1,x0,t d ), P d (D) and P a The three probability density functions (RA) not only reveal the variation law of the function curve, but also provide quantitative parameters of each function, such as the first quartile, median, third quartile, water molecule dispersion displacement with the highest probability, water molecule dispersion coefficient D, and relative anisotropy RA.

[0144] Figure 6 This is an application diagram of the resonance imaging method for characterizing the diffusion properties of intracellular micron-scale particles in embodiment 4 of the present invention. Wherein, Figure 6 A shows a representative MRI image. Figure 6 The boxes in A represent individual voxels. Figure 6 B and Figure 6 C represents what Figure 6 The probability density function P of the diffuse displacement distribution of water molecules within voxels 1 and 2 in A. r (x1,x0,t d The probability density function P of the dispersion coefficient distribution of water molecules within a voxel d (D), and the probability density function P of the anisotropic distribution of water molecules within the voxel. a (RA). Figure 6 D shows the connection with Figure 6 A staining image of a tissue section of the same size as voxel 3 in A. Figure 6 E andFigure 6 G respectively show that Figure 6 D shows the schematic diagram of the arrangement of cells, extracellular matrix and the like in region 1 and region 2, which indicates that there are differences in cell types, cell sizes, cell shapes, extracellular matrix proteins and the like at different spatial positions in a single voxel, finally resulting in the distribution of water molecule diffusion displacement, water molecule diffusion coefficient and water molecule diffusion anisotropy in a single voxel. Figure 6 F shows that Figure 6 E shows the distribution of diffusion displacement, diffusion coefficient and relative anisotropy value in the spatial region from point a to point f. Figure 6 H shows that ​ G shows the distribution of diffusion displacement, diffusion coefficient and relative anisotropy value in the spatial region from point a to point g.

[0145] It can be seen that the method for characterizing the micron-scale diffusion characteristics of a magnetic resonance imaging single voxel can use a millimeter-scale magnetic resonance imaging single voxel to characterize the complex diffusion characteristic problem associated with the micron-scale tissue structure in a single voxel.

[0146] It should be noted that the above examples are only used to illustrate the technical solutions of the present application, and are not intended to limit the protection scope of the present application. Although the present application has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present application can be modified or replaced by equivalents without departing from the essence and scope of the technical solutions of the present application.

Claims

1. A method for characterizing the diffusion properties of intracellular elements at the micrometer scale in magnetic resonance imaging, characterized in that, Includes the following steps: S1. Construct probability density functions for the diffusion characteristics distribution within multiple individual voxels; S2. Design a dedicated sequence to scan and image the target object, and obtain scan data that constructs the probability density function of the diffusion characteristics within a single voxel. S3. Based on the imaging sequence parameters and scanning data of step S2, calculate the probability density function of the diffusion characteristics distribution within various individual voxels in step S1 to obtain the diffusion characteristics results at the micrometer scale within individual voxels. S2 includes a dedicated sequence design, specifically: Design a diffusion-weighted imaging stimulated echo acquisition sequence to provide scanning data for constructing the probability density function of the diffusion displacement distribution of water molecules within a voxel; The design series includes diffusion-weighted imaging stimulated echo acquisition sequence, diffusion-weighted imaging oscillating gradient spin echo acquisition sequence, and diffusion-weighted imaging pulse gradient spin echo acquisition sequence, to provide scanning data for constructing the probability density function of the diffusion coefficient distribution of water molecules within voxels. We designed a diffusion-weighted imaging dual-pulse gradient spin echo acquisition sequence to provide scanning data for constructing the probability density function of the diffusion anisotropic distribution of water molecules within a voxel.

2. The method for characterizing the diffusion properties of intracellular micrometer-scale magnetic resonance imaging cells according to claim 1, characterized in that: In S1, we construct the probability density function of the dispersion displacement distribution of water molecules within a single voxel, the probability density function of the dispersion coefficient distribution of water molecules within a single voxel, and the probability density function of the dispersion anisotropy distribution of water molecules within a single voxel. The specific process for constructing the probability density function of the diffuse displacement distribution of water molecules within a voxel is as follows: The dispersion process of water molecules is described by the probability density function of the dispersion displacement of water molecules within a given dispersion time. Inside, water molecules start from the origin Move to the finish line The probability density function is expressed as ; In the case of a single voxel, the k-space signal The probability density function characterizing the displacement distribution of water molecules within a monomer. The relationship between them satisfies equation (1): …...Equation (1) In equation (1), the vector , , Represents the reciprocal of the diffusion displacement. G represents the gyromagnetic ratio, and G represents the spatial coding gradient. Represents the diffuse coding gradient. t s Represents the duration of the spatially encoded gradient. The duration of the diffuse coding gradient is represented. Represents the nuclear spin density within a monomer; For k-space signals The image signal is obtained by performing an inverse Fourier transform. Then, through equation (2) Divided into two parts And then Perform inverse Fourier transform to obtain ; ... Equation (2); in, and These represent magnetic resonance signals related to the diffusion motion of water molecules associated with microstructure and the pseudo-diffusion motion associated with blood microcirculation, respectively. The specific process for constructing the probability density function of the dispersion coefficient distribution of water molecules within a voxel is as follows: The diffuse magnetic resonance signal within a single voxel is the sum of the diffuse magnetic resonance signals generated by all nuclear spins within the voxel; To further analyze the diffuse magnetic resonance signal within a single voxel, we consider that the nuclear spins within the voxel are distributed in different spatial locations, and that each nuclear spin is constrained and hindered by the microstructure. To examine the distribution of the dispersion coefficient D of each nuclear spin, the probability density function of the water molecule dispersion coefficient distribution within a single voxel is expressed as follows: , subscript d It is the lowercase version of the diffusion coefficient D, indicating that the probability is related to the diffusion coefficient; To distinguish the contributions of intravoxel water molecule diffusion motion and blood microcirculation to the diffusion coefficient D, let This represents the dispersion coefficient related to the dispersion motion of water molecules. Indicates the pseudo-diffusion coefficient related to blood microcirculation; for Discretization is performed, in Divide the range into M discrete Values, respectively using It means that among them For each number, it is a natural number from 1 to M; after discretization, the probability density function is... It is also discretized into ; When the number of scans is N, the relationship in equation (3) exists: ... Equation (3); Satisfying Relationships ,Bundle Discretize it into M parts according to its maximum value, from the lowest value to the highest value, where M is the number of discretized values. This represents the diffuse magnetic resonance signal of a single voxel obtained in the nth scan. This represents the magnetic resonance signal of a single voxel obtained without applying a diffusion coding gradient. It is the diffusion weighting factor used in the nth scan, and the coefficient is... This represents the proportion of magnetic resonance signal attenuation caused by blood microcirculation to the total magnetic resonance signal attenuation; The acquired diffuse magnetic resonance signal was fitted to Equation (3) using the non-negative regularized least squares method. The expression for the non-negative regularized least squares fitting problem is shown in Equation (4): ... Equation (4); The first and second terms in equation (4) represent the fitting error and regularization constraint, respectively. Represents the regularization factor; get And then By performing Nth-order polynomial fitting, the probability density function of the dispersion coefficient distribution of water molecules within a continuous voxel is obtained. The fitting formula is as follows: ...Equation (5); In equation (5), N represents the number of fitting iterations, which is determined based on the actual situation. The coefficients represent the polynomial fit. m , j It is a quantity index in series calculation; The process of constructing the probability density function of the diffuse anisotropic distribution of water molecules within a voxel is as follows: The diffuse anisotropy of water molecules is measured by relative anisotropy (RA). RA Defined as: ... Equation (6); in, These are the eigenvalues ​​of the diffusion tensor D. Eigenvalues ​​representing the diffuse tensor D variance It is the average dispersion coefficient; The probability density function of the anisotropic distribution of water molecules within a voxel is used Indicates, calculation The steps are as follows: First, the probability density function of the diffuse tensor distribution needs to be obtained. Then, based on the probability density function of the diffuse tensor distribution, the following steps are taken: The specific calculation process is as follows: First, the probability density function of the discrete water molecule dispersion tensor distribution within the voxel is obtained using the methods represented by equations (3) and (4). Then, the discrete diffuse tensors are obtained respectively. Corresponding eigenvalues eigenvalues variance and average dispersion coefficient And calculate each discrete diffusion tensor according to equation (6). Corresponding related anisotropy RA Based on diffusion tensor With related anisotropy RA By establishing a one-to-one correspondence, the probability density function of the anisotropic distribution of water molecules within a discrete voxel can be obtained. ; then By performing a Y-order polynomial fitting, the probability density function of the anisotropic distribution of water molecules within a continuous voxel is obtained. The fitting formula is as follows: ... Equation (7); In equation (7), Y represents the number of fitting iterations. The value of Y is determined based on the actual situation. The coefficients represent the polynomial fit. m is the number used in polynomial fitting. m The value of the discretized RA.

3. The method for characterizing the diffusion properties of intracellular micrometer-scale magnetic resonance imaging (MRI) cells according to claim 2, characterized in that: Will The first quartile, median, third quartile, and the dispersion shift of water molecules with the highest probability were used as quantitative indicators. Will The first quartile, median, third quartile, and the water molecule dispersion coefficient with the highest probability were used as quantitative indicators. Will The first quartile, median, third quartile, and the relative anisotropy with the highest probability. RA As a quantitative indicator.

4. The method for characterizing the diffusion properties of intracellular micrometer-scale magnetic resonance imaging (MRI) cells according to claim 2 or 3, characterized in that: The diffusion-weighted imaging stimulated echo acquisition sequence can simultaneously achieve low b-value and high b-value sampling. The diffusion-weighted imaging oscillating gradient spin echo acquisition sequence, the diffusion-weighted imaging pulse gradient spin echo acquisition sequence, and the diffusion-weighted imaging stimulated echo acquisition sequence respectively realize diffusion measurement with short diffusion time, medium diffusion time, long diffusion time, and multiple b values. The diffusion-weighted imaging dual-pulse gradient spin echo acquisition sequence includes two diffusion-weighted imaging pulse gradient spin echo acquisition sequence modules, each with its own b-value and diffusion gradient direction.

5. The method for characterizing the diffusion properties of intracellular micrometer-scale magnetic resonance imaging (MRI) cells according to claim 4, characterized in that: During the actual scanning process, key parameters of the diffusion-weighted imaging stimulated echo acquisition sequence, diffusion-weighted imaging oscillating gradient spin echo acquisition sequence, diffusion-weighted imaging pulse gradient spin echo acquisition sequence, and diffusion-weighted imaging dual-pulse gradient spin echo acquisition sequence were optimized. Specifically, the optimized sequence parameters refer to the diffusion time. echo time b Value, diffusion gradient direction, and mixing time are as follows: A. During scanning, adjust the diffusion time to correspond to different changes in microstructural properties. , , It is the duration of a single diffuse gradient. The time interval between two diffusion gradients; B. By adjusting the echo time, the sensitivity of the diffuse magnetic resonance signal to the diffusion characteristics associated with various microstructure properties can be changed; C. Adjust the diffusion weighting factor b The magnitude and number of values ​​alter the acquired diffuse magnetic resonance signal; D. Adjusting the direction of the diffusion gradient changes the sensitivity of the diffusion magnetic resonance signal to the diffusion characteristics associated with various microstructure properties; E. Adjust the mixing time of the diffusion-weighted imaging stimulated echo acquisition sequence and the diffusion-weighted imaging dual-pulse gradient spin echo acquisition sequence to change the sensitivity of the diffusion magnetic resonance signal to the diffusion characteristics associated with various microstructure properties.

Images