A Diffusion Magnetic Resonance Method for Measuring Lymphatic-Like Circulation Dependent on Heart Rate Period

Abstract

The present invention discloses a diffusion magnetic resonance method for measuring lymphatic circulation-like dependent on the heart rate cycle: using DTI to acquire brain magnetic resonance images; synchronously acquiring the signal time series of heart rate fluctuations; determining the peak coordinates of the cardiac fluctuation signals according to the signal time series of heart rate fluctuations; rearranging the specific positions of the brain magnetic resonance images within the heart rate signal cycle; performing re-interpolation and uniform sampling in the time dimension on the rearranged brain magnetic resonance images to generate a number of equidistant diffusion magnetic resonance image data within the heart rate cycle; calculating AD, RD, and MD in the diffusion magnetic resonance images pixel by pixel; generating the change characteristic curves of AD, RD, MD, and S within the heart rate cycle based on the mask images of specific regions in the individual space. Through the present invention, the flow rate and direction characteristic changes of cerebrospinal fluid in the perivascular space around large blood vessels in the brain within the cardiac cycle can be obtained, and the flow characteristics of cerebrospinal fluid in the perivascular space around small blood vessels within the cardiac cycle can also be measured.

Description

Technical Field

[0001] The present invention relates to the technical field of magnetic resonance imaging, and particularly relates to a diffusion magnetic resonance method for measuring lymphatic-like circulation dependent on the cardiac cycle. Background Art

[0002] The glymphatic system in the brain is a new mechanism for the brain to clear waste proposed in recent years (Iliff, Jeffrey J et al. “A paravascular pathway facilitates CSF flow through the brain parenchyma and the clearance of interstitial solutes, including amyloid β.” Science translational medicine vol. 4, 147 (2012): 147ra111. Iliff, Jeffrey J et al. “A paravascular pathway facilitates CSF flow through the brain parenchyma and the clearance of interstitial solutes, including amyloid β.”

[0003] Science translational medicine vol.4,147(2012):147ra111. Iliff, J.J et al. Cerebral arterial pulsation drives paravascular CSF-Interstitial fluid exchange in the murine brain. Journal of Neuroscience, 33(46),(2013)18190-18199.). Among them, cerebrospinal fluid (CSF) is an important clearance medium, specifically including: CSF enters the brain along the perivascular space around the arterial tree, and with the assistance of aquaporin-4 (AQP4) expressed by the end-feet of perivascular astrocytes, it further enters the interstitial space and forms convection in the interstitial space (thereby accelerating the clearance of interstitial molecules), and finally flows out along the perivascular space around the venous vessels. The current measurement of the glymphatic system in the brain is mainly based on invasive methods and is an indirect measurement method of the glymphatic system, with risks of inaccurate measurement and interference with the state of the glymphatic system, and it cannot be translated into normal population measurements. Examples of existing methods are as follows: (1) The optical imaging method based on tracers requires a craniotomy on the animal's head, which is an invasive method, and due to the limited traditional depth of light, the imaging range is only limited to the cortical surface; (2) The dynamic contrast-enhanced magnetic resonance imaging method based on contrast agents relies on the injection of magnetic resonance contrast agents into the cisterna magna and can only observe the macroscopic inflow and outflow velocities of the glymphatic system, and cannot reflect microscopic level information such as interstitial volume and interstitial fluid flow velocity; (3) Injecting tracers (such as TMA) into the brain parenchyma and using means such as microelectrodes to monitor the concentration of tracers point by point and in real time, so as to calculate the interstitial size and tracer flow velocity, etc. However, this method is invasive, and the processes of tracer injection and electrode implantation can cause local inflammatory reactions, and the polarity of AQP4 is damaged, thereby interfering with the glymphatic system.

[0004] However, there are currently few tools for non-invasive measurement of the glymphatic system in the human brain, and the demand for using non-invasive magnetic resonance measurement to evaluate the glymphatic system and explore its underlying mechanisms is increasing. As a non-invasive method for measuring the diffusion of water molecules, diffusion magnetic resonance technology has been increasingly applied to the measurement of the glymphatic system in recent years. With the continuous in-depth research on the glymphatic system in recent years, great progress has also been made in the flow measurement technology for arterial and venous PVS. In 2017, Taoka et al. (Taoka, T et al. Evaluation of glymphatic system activity with the diffusion MR technique: diffusion tensor image analysis along the perivascular space (DTI-ALPS) in Alzheimer’s disease cases. (2017). 172–178. https: / / doi.org / 10.1007 / s11604-017-0617-z) proposed the ALPS index based on DTI technology to characterize the outflow characteristics of venous PVS, opening the way for using diffusion magnetic resonance to measure the glymphatic system. Subsequently, in 2018, Harrison et al. (Harrison, I.F et al. Non-invasive imaging of CSF-mediated brain clearance pathways via assessment of perivascular fluid movement with diffusion tensor MRI. ELife, (2018). 7, e34028.) proposed using DTI technology with long TE and low b-value to evaluate arterial PVS flow by measuring the apparent diffusion coefficient D*. Subsequently, in 2020, Bito et al. (Bito Y et al. Low b-value diffusion tensor imaging for measuring pseudorandom flow of cerebrospinal fluid. Magn.Reson.Med., 2021, 86(3):1369–1382.) further enriched the theoretical basis of DTI technology with low b-value and proposed that the observed diffusion coefficient can directly reflect the liquid flow rate.On this basis, Wen QT et al. (Wen Q et al. Assessing pulsatile waveforms of paravascular cerebrospinal fluid dynamics using dynamic diffusion-weighted imaging (dDWI). Neuroimage, Elsevier Inc., 2022, 260 (July): 119464.) proposed a diffusion magnetic resonance detection method for measuring the cerebrospinal fluid flow characteristics in the perivascular spaces of the whole brain in 2022 and extracted the cardiac-dependent features therein. However, this study failed to reflect the direction-dependent characteristics of the perivascular fluid, but instead averaged the diffusion-weighted signals in simply three directions. In addition, the cerebrospinal fluid flow observed in this study was all around the large arteries and superficial arteries, and the cerebrospinal fluid flow characteristics around the perforating arteries were not observed. Summary of the Invention

[0005] The purpose of the present invention is to provide a diffusion magnetic resonance method for measuring the glymphatic circulation dependent on the heart rate cycle. Through the present invention, the flow rate and direction characteristic changes of the cerebrospinal fluid in the perivascular spaces of the large blood vessels in the brain during the cardiac cycle can be obtained, and it can also be used to measure the cerebrospinal fluid flow characteristics in the perivascular spaces of the microvessels during the cardiac cycle.

[0006] The present invention provides the following technical solutions:

[0007] A diffusion magnetic resonance method for measuring the glymphatic circulation dependent on the heart rate cycle, the method comprising:

[0008] (1) Using dynamic diffusion tensor magnetic resonance technology (DTI, Diffusion Tensor Imaging) to acquire brain magnetic resonance images; simultaneously obtaining the signal time series of heart rate fluctuations through a finger pulse oximetry monitoring device;

[0009] (2) Determining the peak coordinates of the cardiac fluctuation signal according to the signal time series of heart rate fluctuations; and rearranging the time coordinates of the brain magnetic resonance images acquired in step (1) according to their specific positions in the heart rate signal cycle;

[0010] (3) Performing time-dimensional re-interpolation and uniform sampling on the brain magnetic resonance images in step (2) to generate a number of equidistant diffusion magnetic resonance image data during the heart rate cycle;

[0011] (4) Based on the equidistant diffusion magnetic resonance image data in step (3), calculating the axial diffusion coefficient AD, the radial diffusion coefficient RD, and the mean diffusion coefficient MD, which characterize the diffusion characteristics, for each pixel in the diffusion magnetic resonance image;

[0012] (5) Based on the mask image of the specific area in the individual space, the axial diffusion coefficient AD, the radial diffusion coefficient RD, the mean diffusion coefficient MD and the characteristic curve of the diffusion magnetic resonance signal during the heart rate cycle are generated.

[0013] Among them, the signal time series of heart rate fluctuation is also called finger pulse blood oxygen signal; the peak coordinates of the heart fluctuation signal are also called pulse oxygen signal peak time coordinates; the heart rate cycle is also called cardiac cycle.

[0014] In step (1), the acquisition of the MRI dynamic diffusion tensor MRI technique sequence will include two parts: first, the acquisition of the MRI image will be completed without applying a gradient, and the acquisition will be repeated multiple times; second, the image acquisition will be performed in different diffusion weighted directions, wherein multiple images are acquired in each diffusion weighted direction.

[0015] Preferably, in step (1), the image sequence parameters of dynamic diffusion tensor imaging (DTI) magnetic resonance imaging technology are optimized (echo time, diffusion gradient direction and applied intensity, number of repeated acquisitions are optimized) to obtain brain magnetic resonance images. Specifically:

[0016] In step (1), the brain magnetic resonance image includes: a multi-diffusion-weighted brain magnetic resonance image acquired by using a long echo time low diffusion weighted magnetic resonance sequence technology, wherein: the long echo time TE is 100ms-200ms, and the diffusion weighted b value is 50s / mm 2 -300s / mm 2 ; Based on the dynamic diffusion tensor MRI sequence, brain MRI images without diffusion weighting were scanned.

[0017] The echo time in the dynamic diffusion tensor MRI sequence should be long enough to produce transverse relaxation contrast between different tissue components and further highlight the signal proportion of fluid in the perivascular space. Therefore, it should be designed in the range of (100-200ms). For gradient diffusion MRI, the diffusion gradient intensity should not be too large, so the diffusion weighted b value should be designed in the range of (50-300s / mm 2 ) range.

[0018] The present invention provides contrast imaging of the transverse relaxation (T2) mode using a long echo time (TE), making the magnetic resonance image more sensitive to the components of cerebrospinal fluid. At a long echo time, it is sensitive enough to the volume changes in the perivascular space. Using a low diffusion-weighted b value can make parameters such as the axial diffusion coefficient and the radial diffusion coefficient of diffusion magnetic resonance more sensitive to the pseudo-diffusion effect caused by slow flow. And the involved diffusion coefficient increases with the increase of the liquid flow rate.

[0019] In step (1), in the diffusion-weighted magnetic resonance sequence technology, the number of gradient directions of multi-diffusion weighting is greater than or equal to 6. Applying at least 6 directions of diffusion gradient weighting directions aims to reconstruct the tensor matrix and calculate the axial diffusion coefficient and the radial diffusion coefficient.

[0020] In step (1), the brain magnetic resonance images are acquired the same number of times for each gradient direction of multi-diffusion weighting; the number of repetitions of scanning the brain magnetic resonance images without diffusion weighting is not less than the number of repetitions of a single gradient direction.

[0021] Through multiple acquisitions of magnetic resonance images, the acquisition time of the magnetic resonance images randomly falls within different stages of the continuously recorded heart rate signal. That is, high-time-resolution signal sampling is achieved through repeated sampling. The principles of repeated sampling include: repeating the same number of times for each gradient direction of repeated scanning of the low diffusion-weighted b value. In addition, it is necessary to repeat scanning the magnetic resonance images without applying diffusion weighting, and the number of repetitions is not less than the number of repetitions of a single direction.

[0022] Preferably, in step (2), after removing the high-frequency noise signals in the signal time series (one-dimensional time signal) of heart rate fluctuations by filtering, the peak coordinates of the heart fluctuation signals are determined. By performing Fourier transform on the one-dimensional time signal, the high-frequency signals significantly greater than the heart pulsation frequency are removed, ensuring no noise interference in the subsequent automatic peak-seeking process.

[0023] Preferably, the filtering frequency in step (3) can be selected as the threshold higher than the upper limit of the heart rate period of normal people to filter out the high-frequency noise signals.

[0024] In step (2), the real scanning time of each layer of the brain magnetic resonance images is extracted (for example, using matlab to obtain the layer scanning time information in the image header file, and rearranging each layer of the brain magnetic resonance images according to the real scanning time corresponding to the signal time series of heart rate fluctuations (corresponding to the one-dimensional time series with the heart rate signal peak determined), so that each layer of each frame of the brain magnetic resonance images has a heart rate period, and the one heart rate period includes two adjacent wave peaks.

[0025] In step (2), rearrangement means returning the magnetic resonance signal to the heart rate cycle, and clarifying that each pixel point in each layer of each frame of the magnetic resonance data corresponds to a specific coordinate within a specific pulse cycle.

[0026] Determine the specific position of the acquisition time of each layer of magnetic resonance images in the peak-to-peak coordinates of heart rate fluctuations. It can be understood that: the coordinate close to the previous peak is -50%, and the coordinate close to the next peak is 50%. At the same time, the coordinate interval from -50% to 0% can be equivalent to the heart rate cycle from 50% to 100%. In this way, each layer of each frame of the magnetic resonance image has a specific heart rate cycle coordinate from 0% to 100%.

[0027] Taking the application of six gradients as an example, steps (2) and (3) are further described as follows:

[0028] Group the magnetic resonance images scanned with different parameters for all times, and divide them into a non-gradient application group and a gradient application group respectively; the signals in the gradient application group are further divided into six subgroups in six directions; then there are a total of seven groups. Rearrange all the magnetic resonance sequences within each group, and rearrange the magnetic resonance signals of each layer according to the specific heart rate cycle coordinates. Then, through the interpolation function built into the matlab program, 100 interpolations within the heart rate cycle from 0% to 100% are completed. After completing this step, 7 groups of magnetic resonance data are obtained, and each group contains 100 frames of magnetic resonance data, and these 100 frames of magnetic resonance data are evenly distributed within one cardiac cycle.

[0029] Preferably, in step (4), the diffusion tensor model is fitted using the non-linear least squares method to obtain the axial diffusivity AD, the radial diffusivity RD, and the mean diffusivity MD of each pixel point in the diffusion-weighted magnetic resonance image.

[0030] In step (4), specifically, based on the magnetic resonance images with uniform interpolation within the cardiac cycle obtained in step (6), the TORTOISE software is used to complete the extraction of relevant parameters such as the diffusion characteristics of the diffusion-weighted magnetic resonance.

[0031] Preferably, in step (4), preprocessing (eddy current, image acquisition distortion correction, etc.) can be performed on the magnetic resonance images.

[0032] In step (5), the mask image based on a specific region in the individual space includes: an arterial mask image and a white matter mask image.

[0033] In step (5), the mask image can have physiological meanings: one is an arterial vessel probability map registered from the standard space to the individual space, which represents the position of the large arteries visible by magnetic resonance and the cerebrospinal fluid in the perivascular space around the superficial arteries; the other is a specific white matter brain region registered from the standard space to the individual space, such as the corona radiata region of the white matter. This mask image represents the signal of cerebrospinal fluid in the perivascular space of small arteries / veins that are not visible by magnetic resonance.

[0034] The specific steps of step (5) include:

[0035] (5-1) Based on the T1 magnetic resonance image, register the standard brain atlas into the T1 structural image and retain its affine transformation matrix;

[0036] (5-2) Based on the affine transformation matrix from the standard space to the individual space obtained in step (5-1), register the white matter atlas in the standard space into the individual space and set a threshold to obtain a white matter mask image; register the arterial atlas in the standard space into the individual space and set a threshold to obtain an arterial mask image;

[0037] (5-3) Calculate the change characteristic curves of the axial diffusivity AD, the radial diffusivity RD, and the diffusion magnetic resonance signal S of the arterial mask image within the cardiac cycle, and calculate the average diffusivity MD and the change characteristic curves of the diffusion magnetic resonance signal of the white matter mask image within the cardiac cycle.

[0038] Specifically, in step (5-1), use the ANTS software to register the standard brain atlas (MNI152) into the T1 structural image based on the T1 magnetic resonance image.

[0039] Preferably, in step (5-3), the diffusion magnetic resonance signal of the arterial mask image is the diffusion magnetic resonance signal S at b = 150 s / mm 2 under b = 150 s / mm 2 ), and the diffusion magnetic resonance signal of the white matter mask image is the original diffusion magnetic resonance signal S at b = 0 s / mm 2 ).

[0040] Compared with the prior art, the technical effect of the present invention lies in:

[0041] The method provided by the present invention can non-invasively measure the fluid flow in the glymphatic circulation system. The magnetic resonance sequence in the method provided by the present invention adopts the diffusion magnetic resonance technology. Through multi-diffusion weighted direction encoding and synchronous signal acquisition of the heart rate, it realizes the specific detection of the cerebrospinal fluid in the perivascular space around the artery changing with the arterial fluctuation cycle rate and direction change (for example, AD can characterize the PVS flow parallel to the blood vessel, and RD can characterize the PVS flow perpendicular to the blood vessel), and can be applied to measure the cerebrospinal fluid flow in the perivascular space of small blood vessels (for example, for a smaller PVS space, the mean diffusion coefficient MD is used to represent its diffusion rate). Description of the Drawings

[0042] Figure 1 The general flow chart of a diffusion magnetic resonance method for measuring the glymphatic circulation dependent on the heart rate cycle provided for the embodiment;

[0043] Figure 2 A method paradigm provided for the embodiment and a display diagram of six-direction gradient space encoding designed;

[0044] Figure 3 A display diagram of the method for mapping the magnetic resonance time to the pulse cycle;

[0045] Figure 4 A display diagram of the mask image in the standard space;

[0046] Figure 5 A diagram showing the change trend of the original diffusion magnetic resonance signal within the heart rate cycle;

[0047] Figure 6 A diagram showing the change trend of the relevant parameters of the diffusion magnetic resonance within the heart rate cycle. Detailed Embodiments

[0048] To make the objectives, technical solutions and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and do not limit the protection scope of the present invention.

[0049] The technical principle of the present invention is as follows: By using a long echo time to increase the contrast of the cerebrospinal fluid component in the signal, it enables the measurement of the cerebrospinal fluid signal with a small proportion in the perivascular space using magnetic resonance technology; by using a low diffusion intensity b value, this diffusion magnetic resonance technology is not only sensitive to the diffusion of water molecules, but also has a high sensitivity to the pseudo-diffusion effect caused by the slow flow of the liquid; at a low diffusion intensity b value, the flow rate of the liquid shows that the apparent diffusion coefficient increases with the increase of the liquid flow rate, presenting a monotonically increasing characteristic.

[0050] As a specific implementation example, the diffusion magnetic resonance method of lymphatic circulation dependent on heart rate cycle provided by the present invention is applied to the brains of adult healthy subjects. The operation flow chart is shown in Figure 1 , and specifically includes the following steps:

[0051] Step 1: Place adult healthy subjects in a 3T magnetic resonance imaging system, use the center of the head as the scanning center point, and perform image acquisition of the head. In this implementation example, magnetic resonance data of 8 adult healthy subjects were collected as brain magnetic resonance images. A finger pulse oximetry monitoring device was attached to the thumb of all subjects to monitor their heart rate pulse signals (or signal time series of heart rate fluctuations or one-dimensional finger pulse heart rate and blood oxygen data).

[0052] Step 2: Perform brain magnetic resonance image acquisition in different diffusion weighted directions, and specifically, acquire multiple brain magnetic resonance images in each diffusion weighted direction; specifically:

[0053] Set the dynamic diffusion tensor magnetic resonance technology sequence, with the resolution set to 1.5×1.5mm 2 , the slice thickness is 1.5mm, set the middle position of the corpus callosum of the brain in the head-to-toe direction as the middle position of the scanning FOV (Field of View), and a total of 49 slices are acquired. Among them, the diffusion weighted b value should be set to 150s / mm 2 . The diffusion weighted directions are set to 6 gradient directions, namely: [1,0,1], [-1,0,1], [0,1,1], [0,1,-1], [1,1,0], [-1,1,0]. The diffusion weighted direction is set to Figure 2 one of the 6 directions in

[0054] Step 3: Acquire brain magnetic resonance images without applying gradients, and repeat the acquisition multiple times; specifically:

[0055] Based on the dynamic diffusion tensor magnetic resonance technology sequence, scan non-diffusion weighted images, and the diffusion weighted value should be set to 0s / mm 2 (without applying gradients), and repeat 60 times. Keep other parameter values of the sequence unchanged.

[0056] Step 4: Remove noise signals by high-frequency filtering and extract signal peak coordinates, specifically:

[0057] Use matlab software to extract the time peak coordinates (peak coordinates of the cardiac fluctuation signal) of the heart rate pulse signal obtained in Step 1: adopt the method of high-frequency filtering to remove the spike-like noise in the signal, and then find the peak time coordinates within each cardiac cycle (or heart rate cycle).

[0058] Step Five: Analyze the actual acquisition time of each layer in the header file of the magnetic resonance data (brain magnetic resonance image data). Specifically:

[0059] Use Matlab software to extract the information in the header file of the brain magnetic resonance image data obtained in Step Two and Step Three, and the specific actual time obtained from each layer of scanning can be read out simultaneously.

[0060] Step Six: Integrate the information of the peak time of finger pulse blood oxygen and the time information of each layer of magnetic resonance and map them to a heart rate cycle. Specifically:

[0061] Use Matlab software to map the sampling time of the data layer by layer of the brain magnetic resonance image data obtained in Step Five to the one-dimensional finger pulse heart rate blood oxygen data in Step Four. Just as Figure 3 The method shown in determines the specific position of the acquisition time of each layer of magnetic resonance image in the peak-to-peak coordinates of heart rate fluctuations: the coordinate close to the previous peak is 50%, and the central coordinate of each peak-to-peak coordinate is 100% (which is also 0% at the same time). In this way, each layer of each frame of the magnetic resonance image has a specific heart rate cycle coordinate from 0% to 100%; after reordering, the ordered data of the magnetic resonance image in the peak-to-peak interval of the blood oxygen signal can be obtained, and thus the valley-peak-valley time series signal convenient for display can be obtained.

[0062] Step Seven: Perform uniform resampling within the heart rate cycle. Specifically:

[0063] Use Matlab software to perform uniform resampling within a heart rate cycle on the reordered brain magnetic resonance image data in Step Six: through the interpolation function built into the Matlab program, 100 interpolations from 0 to 100% within the heart rate cycle are completed; after this step, seven groups of brain magnetic resonance image data evenly distributed within the heart rate cycle will be obtained, and each group contains 100 frames of magnetic resonance data, and these 100 frames of magnetic resonance data are evenly distributed within a cardiac cycle.

[0064] Step Eight: Fit the diffusion magnetic resonance data and extract the target parameters for the data at each resampling time point. Specifically:

[0065] Use the TORTOISE software to complete the extraction of relevant parameters such as the diffusion characteristics of the diffusion magnetic resonance in Step Seven: use the nonlinear least squares method to fit the diffusion tensor model and obtain the axial diffusivity (AD), radial diffusivity (RD), and mean diffusivity (MD) of each pixel point in the image.

[0066] Step 9: Use the ANTS software to generate white matter mask images and vascular probability mask images, and use the FSL software to complete the registration from T1 to diffusion magnetic resonance space specifically as follows:

[0067] (1) Use the ANTS software to complete the registration of the T1 structural image from the standard space to the individual space, and save the affine transformation matrix obtained during the registration process.

[0068] (2) Use the ANTS software to register the white matter standard atlas in the standard space into the individual space, determine the threshold, and save it as the mask image of the white matter region.

[0069] (3) Use the ANTS software to register the arterial vascular standard atlas in the standard space into the individual space, determine the threshold, and save it as the mask image of the vascular region.

[0070] (4) Use the FSL:epi_reg command to complete the registration from the T1 image to the diffusion magnetic resonance image, and retain the affine transformation matrix.

[0071] (5) Use the FSL:flirt command to register from the T1 space to the diffusion magnetic resonance data space in combination with the obtained white matter mask image and arterial vascular mask image.

[0072] Step 10: Use the matlab software to analyze the change characteristic curves of the dynamic diffusion tensor magnetic resonance technology-related parameters within the cardiac cycle under different mask images: Use the principle of multiplying the target parameter matrix by the mask image to complete the information extraction of AD, RD, MD, S(b = 0 s / mm 2 ) and S(b = 150 s / mm 2 ).

[0073] The diffusion magnetic resonance method for measuring the lymphatic circulation dependent on the heart rate cycle provided in this embodiment can generate magnetic resonance images that change with the arterial pulse by synchronously collecting magnetic resonance image data and pulse heart rate data. By evaluating the diffusion tensor magnetic resonance parameters and analyzing the mask region images of specific regions ( Figure 4 , taking the corona radiata region of the white matter as an example), the cerebrospinal fluid flow in the perivascular space around arteries of different sizes and the cerebrospinal fluid flow around arterioles in the white matter can be analyzed ( Figure 5 and Figure 6 ). It can be seen from the results that the axial diffusion coefficient (AD), radial diffusion coefficient (RD), and mean diffusion coefficient (MD) under different mask images all show a trend of changing with the heart rate. At the same time, the signal S(b = 0 s / mm 2 ) without applying the gradient also shows a trend of changing with the heart rate. Here, it can be considered that the reduction of the perivascular space caused by blood vessel dilation can reflect the blood vessel dilation process. ByFigure 5 and Figure 6 The results in Figure 6 also show that AD and RD reach their maximum during blood vessel dilation. Based on this phenomenon, it can be explained that the increase in the axial and radial flow rates of the perivascular interstitial fluid is caused by blood vessel dilation. Through the diffusion magnetic resonance method proposed in the present invention, direct data on the perivascular interstitial fluid driven by arterial pulsation can be obtained.

[0074] The specific embodiments described above have elaborated on the technical solutions and beneficial effects of the present invention. It should be understood that the above is only the most preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, supplements, equivalent replacements, etc. made within the scope of the principles of the present invention shall be included within the protection scope of the present invention.

Claims

1. A diffusion magnetic resonance method for measuring heart rate cycle-dependent lymphatic circulation, characterized in that: The method comprises: (1) Use dynamic diffusion tensor magnetic resonance technology to collect brain magnetic resonance images; synchronously use finger pulse blood oxygen monitoring equipment to obtain the signal time series of heart rate fluctuations; (2) determining the peak coordinates of the heart rate fluctuation signal according to the signal time series of the heart rate fluctuation; and rearranging the time coordinates of the brain magnetic resonance image acquired in step (1) according to the specific position within the heart rate signal cycle; (3) re-interpolating and uniformly sampling the brain magnetic resonance image in the time dimension in step (2) to generate a number of equally spaced diffusion magnetic resonance image data within the heart rate cycle; (4) Based on the equally spaced diffusion magnetic resonance image data in step (3), the axial diffusion coefficient AD, the radial diffusion coefficient RD and the average diffusion coefficient MD representing the diffusion characteristics in the diffusion magnetic resonance image are calculated pixel by pixel; (5) based on the mask image of the specific area in the individual space, generating the axial diffusion coefficient AD, the radial diffusion coefficient RD, the mean diffusion coefficient MD and the characteristic curve of the diffusion magnetic resonance signal during the heart rate cycle; In step (5), the mask image based on the specific area in the individual space includes: an artery mask image and a white matter mask image.

2. The diffusion magnetic resonance method for measuring heart rate cycle-dependent lymphatic circulation according to claim 1, characterized in that: In step (1), the brain magnetic resonance image includes: a multi-diffusion-weighted brain magnetic resonance image acquired by using a long echo time low diffusion weighted magnetic resonance sequence technology, wherein: the long echo time TE is 100ms-200ms, and the diffusion weighted b value is 50s / mm 2 -300s / mm 2 ; Based on the dynamic diffusion tensor MRI sequence, brain MRI images without diffusion weighting were scanned.

3. The diffusion magnetic resonance method for measuring heart rate cycle-dependent lymphatic circulation according to claim 2, characterized in that: The number of multi-diffusion weighted gradient directions in the diffusion-weighted magnetic resonance sequence technology is greater than or equal to 6.

4. The diffusion magnetic resonance method for measuring heart rate cycle-dependent lymphatic circulation according to claim 3, characterized in that: The brain magnetic resonance images are acquired by repeating the same number of times in each gradient direction of multiple diffusion weighting; the number of repetitions of scanning brain magnetic resonance images without diffusion weighting is not less than the number of repetitions in a single gradient direction.

5. The diffusion magnetic resonance method for measuring heart rate cycle-dependent lymphatic circulation according to claim 1, characterized in that: In step (2), the high-frequency noise signal in the time series of the heart rate fluctuation signal is removed by filtering, and then the peak coordinates of the heart rate fluctuation signal are determined.

6. The diffusion magnetic resonance method for measuring heart rate cycle-dependent lymphatic circulation according to claim 1, characterized in that: In step (2), the real scanning time of the brain magnetic resonance image is extracted layer by layer, and each layer of the brain magnetic resonance image is rearranged corresponding to the signal time series of the heart rate fluctuation according to the real scanning time, so that each layer of each frame of the brain magnetic resonance image has a heart rate cycle, and the heart rate cycle includes two adjacent peaks.

7. The diffusion magnetic resonance method for measuring heart rate cycle-dependent lymphatic circulation according to claim 1, characterized in that: In step (4), the diffusion tensor model is fitted using the nonlinear least squares method to obtain the axial diffusion coefficient AD, radial diffusion coefficient RD and mean diffusion coefficient MD of each pixel in the diffusion magnetic resonance image.

8. The diffusion magnetic resonance method for measuring heart rate cycle-dependent lymphatic circulation according to claim 1, characterized in that: The step (5) specifically comprises: (5-1) Based on the T1 magnetic resonance image, the standard brain atlas is registered to the T1 structural image and its affine transformation matrix is ​​retained; (5-2) Based on the affine transformation matrix from the standard space to the individual space obtained in step (5-1), the white matter atlas in the standard space is registered to the individual space, and a threshold is set to obtain a white matter mask image; the artery atlas in the standard space is registered to the individual space, and a threshold is set to obtain an artery mask image; (5-3) Calculate the axial diffusion coefficient AD, radial diffusion coefficient RD and diffusion magnetic resonance signal S of the arterial mask image during the heart rate cycle, and calculate the average diffusion coefficient MD of the white matter mask image and the diffusion magnetic resonance signal during the heart rate cycle.

9. The diffusion magnetic resonance method for measuring heart rate cycle-dependent lymphatic circulation according to claim 1 or 8, characterized in that: The diffusion magnetic resonance signal S of the artery mask image is b=150s / mm 2 The diffusion magnetic resonance signal under the white matter mask image is the diffusion magnetic resonance original signal.

Images