A method for constructing default mode network based on magnetic nanoparticle imaging system
Through magnetic nanoparticle imaging technology and adversarial network reconstruction method, the problem of detection of local microvascular blood flow changes in the brain is solved, the detection accuracy and safety of the default mode network are improved, and the problems of low contrast and radioactive use of traditional methods are overcome.
Patent Information
- Application Number
- CN202210657567.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-06-10
- Publication Date
- 2024-11-22
- Estimated Expiration
- 2042-06-10
AI Technical Summary
The existing technology is unable to effectively detect changes in local microvascular blood flow in the brain, resulting in low detection rates when studying default mode networks and weak or lost network connections in functional connectivity studies, limiting the utility of individualized medicine.
Using magnetic nanoparticle imaging (MPI)-based method, image reconstruction is carried out by generating an adversarial network, three-dimensional reconstruction images are obtained, rigid registration and brain division are performed, correlation coefficients between signals are calculated, default mode network is drawn, and single-sample t is used. Test statistical analysis determines the significance of functional connections.
Improve contrast, avoid attenuation caused by BOLD signal, enhance the accuracy of the analysis, and achieve a safer and more reliable detection process due to the absence of radioactive drugs.
Smart Images

Figure CN115049044B_ABST
Abstract
Description
Background Art
[0002] When the brain is at rest and not engaged in specific behavioral tasks, the brain area corresponding to the default mode network (DMN) is highly active and the network is activated. At present, research on the brain in a quiet rest state, that is, in a resting state, has been shown to be related to a variety of neurological diseases, such as Alzheimer's disease, epilepsy, depression, etc., and DMN plays a core role in the study of the brain's resting state. At the same time, DMN directly affects people's psychological activities. It is mainly responsible for the brain's monitoring of the internal and external environment, emotional processing, self-introspection, and maintaining consciousness. Therefore, constructing the DMN network plays an important role in studying human health and disease.
[0003] When the brain is stimulated, the activity of the cerebral cortex in the corresponding area increases, resulting in an increase in the blood flow, flow velocity and cerebral vascular capacity (CBV) of the cortical microvessels in the brain functional area, while the increase in cellular oxygen consumption is relatively insignificant. The disproportionate increase of the two leads to an increase in oxygenated hemoglobin and a decrease in deoxygenated hemoglobin in the functional activity area. Traditional imaging methods for obtaining DMN include positron emission tomography (PET) technology and functional magnetic resonance imaging (fMRI). PET technology is based on the changes in blood flow in local areas of the brain, and reflects the neural activity in the area through the collected signals. Although PET technology has high contrast, the injected tracer is radioactive and has low spatial resolution. Functional magnetic resonance imaging (fMRI) based on blood oxygen level dependence (BOLD) has been widely used in the study of spontaneous activity of the resting human brain due to its non-destructiveness, high sensitivity and easy operation. It is also the most commonly used technology for constructing DMN networks. However, BOLD fMRI technology can only indirectly reflect neuronal activity, which is a comprehensive and complex effect of cerebral blood flow (CBF), blood volume and blood oxygen level caused by neuronal activity. Since the actual amount of oxygen remaining in the blood vessels at the brain activation site changes after stimulation, and the BOLD signal is sensitive to the amount of oxygen in hemoglobin, changes in oxygenated and deoxygenated hemoglobin at the brain activation site can be detected by BOLD. However, since changes in cellular oxygen consumption are relatively insignificant, and changes in the oxygen state of hemoglobin in the same activation area are affected not only by the microvessels surrounding the activated neural activity area, but also by the signals of the surrounding large blood vessels, it is impossible to distinguish the blood signals in the dilated microvessels and large blood vessels in the true neural activity area. Although fMRI is a versatile imaging method that can measure in vivo hemodynamic changes associated with brain activation, the CNR of fMRI remains problematic. Low CNR can lead to low detection rates of subtle activations in task-based studies and weak or lost network connections in functional connectivity studies. In response, group averaging is often used to discover group effects, which limits the utility of fMRI in personalized medicine to only the strongest activation paradigms and the strongest resting-state networks.
[0004] In summary, a non-radioactive imaging technique is needed to detect changes in blood flow in local brain microvessels to reflect brain activity.
[0005] Magnetic nanoparticle imaging (MPI) is a new imaging technique. The principle of this technique is to locate and quantify the tracer in the area of interest by directly scanning superparamagnetic oxide nanoparticles (SPION) as imaging tracers. It has the characteristics of linear quantification, high sensitivity, and high temporal resolution. In the experiment, SPION was injected into the blood as a tracer. MPI can directly monitor the relationship between functional CBV changes and time by detecting and tracking local SPION concentrations, thereby reflecting the activation of functional brain areas. Summary of the invention
[0006] In order to solve the above-mentioned problem in the prior art, that is, the prior art does not have a non-radioactive imaging technology to detect the change state of local microvascular blood flow in the brain, the present invention provides a default mode network construction method based on magnetic nanoparticle imaging, comprising:
[0007] Apply tracer to the target to be modeled and set scanning parameters;
[0008] Acquire the MPI one-dimensional voltage signal of the target to be modeled in a resting state based on the scanning parameters;
[0009] Based on the MPI one-dimensional voltage signal, image reconstruction is performed by generating an adversarial network to obtain a three-dimensional reconstructed image;
[0010] Based on the three-dimensional reconstructed image, rigid registration is performed and brain regions are divided, thereby obtaining a time series of partitioned signals;
[0011] Calculating the correlation coefficient between signals based on the partition signal time series to obtain a functional connection matrix, and drawing a default mode network on the three-dimensional reconstructed image according to the functional connection matrix;
[0012] The correlation coefficient of the functional connection matrix is converted into a z value, and a single-sample t-test statistical analysis is performed on the z value to obtain the significance of the target functional connection to be modeled.
[0013] In some preferred embodiments, the three-dimensional reconstructed image is obtained by:
[0014] The MPI one-dimensional voltage signal is a one-dimensional voltage signal vector obtained by scanning frame by frame. The MPI one-dimensional voltage signal is reconstructed through a generative adversarial network at a speed of 10 frames per second, and a three-dimensional reconstructed image is obtained every 4 seconds.
[0015] In some preferred embodiments, the partition signal time series is obtained by:
[0016] Based on the three-dimensional reconstructed image, rigid registration is performed with an existing standard anatomical brain region template image to obtain a registered three-dimensional reconstructed image;
[0017] Selecting a seed region of a preset size from the registered three-dimensional reconstructed image according to an existing standard anatomical brain region template image, and then obtaining a partitioned reconstructed image;
[0018] The partition reconstructed image is composed of a plurality of voxel signals, and each voxel signal is arranged according to time to obtain a partition signal time series.
[0019] In some preferred embodiments, the functional connectivity matrix is calculated as follows:
[0020] Based on the partition signal time series, the Pearson correlation coefficient of each voxel signal and the other voxel signals of the same partition is calculated, where the i-th voxel ROI i With the j-th voxel ROI j The correlation coefficient is:
[0021]
[0022] Among them, ROI i (t) represents the value of the tth time point in the partition signal time series of voxel i, ROI j (t) represents the value of the tth time point in the partition signal time series of voxel j, Represents voxel ROI i The average voxel signal at the last T time points, Represents voxel ROI j The average voxel signal of the last T time points, N represents the number of voxel values in the region;
[0023] The functional connectivity matrix was obtained by combining all voxel signals with the Pearson correlation coefficients of the remaining voxel signals.
[0024] In some preferred embodiments, the functional connectivity matrix is converted into a z value, and a single sample t-test statistical analysis is performed on the z value, specifically:
[0025] Performing Fisher z-value transformation on the functional connectivity matrix to obtain a z-value;
[0026] The z value was subjected to a one-sample t-test statistical analysis with a P value of 0.05 to obtain the significance of functional connectivity.
[0027] In some preferred embodiments, the setting of scanning parameters is specifically:
[0028] The excitation magnetic field was set to 5 mT, the excitation frequency to 25 Hz, and the magnetic field gradients to: Z direction: 0.2 T / m, X direction: 0.1 T / m, Y direction: 0.1 T / m.
[0029] In some preferred embodiments, the tracer is administered to the target to be modeled, specifically, the tracer is injected based on the body weight of the target to be modeled, with a dose of 8-12 mg / kg, and the target to be modeled is placed in a resting state.
[0030] Another aspect of the present invention provides a default mode network construction system based on magnetic nanoparticle imaging, comprising:
[0031] a scan preparation module configured to apply a tracer to the target to be modeled and set scan parameters;
[0032] An MPI scanning module, configured to obtain an MPI one-dimensional voltage signal of a resting state of a target to be modeled based on the scanning parameters;
[0033] A three-dimensional reconstruction module, configured to perform image reconstruction based on the MPI one-dimensional voltage signal through a generative adversarial network to obtain a three-dimensional reconstructed image;
[0034] A brain region division module, configured to perform rigid registration and divide brain regions based on the three-dimensional reconstructed image, thereby obtaining a time series of partition signals;
[0035] a functional connection matrix acquisition module, configured to calculate the correlation coefficient between signals based on the partition signal time series, obtain the functional connection matrix, and draw the default mode network on the three-dimensional reconstructed image according to the functional connection matrix;
[0036] The significance authentication module is configured to convert the correlation coefficient of the functional connection matrix into a z value, and perform a single sample t-test statistical analysis on the z value to obtain the significance of the target functional connection to be modeled.
[0037] The third aspect of the present invention proposes an electronic device, comprising: at least one processor; and a memory communicatively connected to at least one of the processors; wherein the memory stores instructions executable by the processor, and the instructions are used to be executed by the processor to implement the above-mentioned default mode network construction method based on magnetic nanoparticle imaging.
[0038] In a fourth aspect of the present invention, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores computer instructions, and the computer instructions are used to be executed by the computer to implement the above-mentioned default mode network construction method based on magnetic nanoparticle imaging.
[0039] Beneficial effects of the present invention:
[0040] (1) The present invention uses MPI, a new imaging technology that can directly reflect changes in blood volume, to obtain brain signals, and uses a generative adversarial network to reconstruct images, thereby avoiding the problem of attenuation of fMRI signals caused by BOLD signals in traditional technical means;
[0041] (2) The present invention uses MPI to obtain brain signals, and there is no background signal of the brain in the image, which improves the contrast of the acquired image and thus improves the accuracy of the analysis;
[0042] (3) The present invention avoids the use of radioactive drugs, and the implementation process is safer and more reliable. BRIEF DESCRIPTION OF THE DRAWINGS
[0043] Other features, objects and advantages of the present application will become more apparent by reading the detailed description of non-limiting embodiments made with reference to the following drawings:
[0044] Figure 1 is a schematic flow chart of a method for constructing a default mode network based on magnetic nanoparticle imaging in an embodiment of the present invention;
[0045] Figure 2 Schematic diagram of the training process of the generative adversarial model in an embodiment of the present invention;
[0046] Figure 3 is a schematic diagram of the model structure of the generator in an embodiment of the present invention;
[0047] Figure 4 It is a schematic diagram of the structure of the VGG16 model of the generative adversarial model of the present invention;
[0048] Figure 5 is a schematic diagram of the principle of rigid registration in an embodiment of the present invention;
[0049] Figure 6 is a schematic diagram of the principle of obtaining a time series of partition signals according to a three-dimensional reconstructed image in an embodiment of the present invention;
[0050] Figure 7 Schematic diagram of the principle of calculating the correlation coefficient between signals by the default mode network construction method based on magnetic nanoparticle imaging in an embodiment of the present invention;
[0051] Figure 8 is a schematic diagram of the effect of the default mode network generated in an embodiment of the present invention;
[0052] Fig. 9 It is a structural diagram of a computer system of a server for implementing the method, system, and device embodiments of the present application. DETAILED DESCRIPTION
[0053] The present application will be further described in detail below in conjunction with the accompanying drawings and embodiments. It is to be understood that the specific embodiments described herein are only used to explain the relevant invention, rather than to limit the invention. It is also necessary to explain that, for ease of description, only the parts related to the relevant invention are shown in the accompanying drawings.
[0054] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.
[0055] The present invention provides a default mode network construction method based on magnetic nanoparticle imaging. The method uses MPI, a new imaging technology that can directly reflect changes in blood volume, to obtain brain signals, and reconstructs images through a generative adversarial network, thereby avoiding the attenuation problem of fMRI signals caused by BOLD signals used in traditional technical means, and improving the accuracy of the default mode network.
[0056] A method for constructing a default mode network based on magnetic nanoparticle imaging of the present invention comprises:
[0057] Apply tracer to the target to be modeled and set scanning parameters;
[0058] Acquire the MPI one-dimensional voltage signal of the target to be modeled in a resting state based on the scanning parameters;
[0059] Based on the MPI one-dimensional voltage signal, image reconstruction is performed by generating an adversarial network to obtain a three-dimensional reconstructed image;
[0060] Based on the three-dimensional reconstructed image, rigid registration is performed and brain regions are divided, thereby obtaining a time series of partitioned signals;
[0061] Calculating the correlation coefficient between signals based on the partition signal time series to obtain a functional connection matrix, and drawing a default mode network on the three-dimensional reconstructed image according to the functional connection matrix;
[0062] The correlation coefficient of the functional connection matrix is converted into a z value, and a single-sample t-test statistical analysis is performed on the z value to obtain the significance of the target functional connection to be modeled.
[0063] In order to more clearly illustrate the default mode network construction method based on magnetic nanoparticle imaging of the present invention, the following is combined with Figure 1 And each step in the embodiment of the present invention is described in detail.
[0064] The default mode network construction method based on magnetic nanoparticle imaging according to the first embodiment of the present invention includes steps S100 to S600, each of which is described in detail as follows:
[0065] Step S100, applying a tracer to the target to be modeled and setting scanning parameters;
[0066] In this embodiment, the tracer is applied to the target to be modeled, specifically, the tracer is injected based on the weight of the target to be modeled, the dosage is 8-12 mg / kg, and the target to be modeled is in a resting state. The specific operation steps are: select macaques as experimental subjects, recruit 20 macaques in total, including 5 female monkeys and 15 male monkeys, and select macaques without serious physical diseases, brain diseases and drug and alcohol dependence history. According to the weight of the macaque, a corresponding dose of SPION is injected into the calf vein. After the injection, wait for about 15 minutes to allow SPION to circulate to the brain with the vein. In order to collect data in the resting state, the macaque is injected with anesthetics and the head of the macaque is placed in the MPI imaging cavity with an aperture of 19 cm and a length and width of 15*15 cm. The head of the macaque is fixed by using equipment to ensure that the head does not move during data acquisition. Set the acquisition parameters of the MPI device, such as gradient magnetic field, excitation frequency, etc. After the setting is completed, turn on the device and start signal acquisition.
[0067] In this embodiment, the scanning parameters are set as follows:
[0068] The excitation magnetic field was set to 5 mT, the excitation frequency to 25 Hz, and the magnetic field gradients to: Z direction: 0.2 T / m, X direction: 0.1 T / m, Y direction: 0.1 T / m.
[0069] Step S200, acquiring an MPI one-dimensional voltage signal of a resting state of a target to be modeled based on the scanning parameters;
[0070] Step S300, based on the MPI one-dimensional voltage signal, image reconstruction is performed by generating an adversarial network to obtain a three-dimensional reconstructed image;
[0071] In this embodiment, the three-dimensional reconstructed image is obtained by:
[0072] The MPI one-dimensional voltage signal is a one-dimensional voltage signal vector obtained by scanning frame by frame. The MPI one-dimensional voltage signal is reconstructed through a generative adversarial network at a speed of 10 frames per second, and a three-dimensional reconstructed image is obtained every 4 seconds.
[0073] In this embodiment, the image reconstruction of the MPI one-dimensional voltage signal is performed by generating an adversarial network, and the specific steps include:
[0074] The MPI one-dimensional voltage signal is used as the input signal and input into the generative adversarial network to obtain a three-dimensional reconstructed image;
[0075] The generative adversarial network includes a generator and a discriminator;
[0076] The generator includes a fully connected layer, three encoding modules, and a linear layer connected in sequence; the encoding module includes a Transformer encoder and an upsampling unit;
[0077] The discriminator includes N convolutional layers and fully connected layers connected in sequence, where N represents a positive integer.
[0078] The training method of the generated adversarial network is as follows: Figure 2 As shown, specifically:
[0079] A100 generates different types of binary images through a pre-selected model in MATLAB; simulates the binary images to generate corresponding one-dimensional MPI signals; the binary images include geometric images, letter images, and resolution images;
[0080] In this embodiment, MATLAB is first used to generate 10,000 binary images of three different models, specifically: binary images (i.e., binary images include geometric images, letter images, and resolution images, such as Figure 2 As shown), use simulation software to generate a corresponding one-dimensional signal (i.e., an MPI one-dimensional voltage signal) from the binary image; wherein, the simulation software in the present invention preferably uses MPIRF simulation software, and in other embodiments, it can be selected according to actual needs.
[0081] A200, taking the binary image as the true value label and combining it with the corresponding MPI one-dimensional voltage signal, constructs the training sample and obtains the training data set;
[0082] In this embodiment, 8000 pairs of 10000 paired data (i.e., MPI one-dimensional voltage signals and their corresponding binary image true value labels) are used as training data, and the remaining 2000 pairs are used as test data sets. At the same time, 10 groups of data of different models are actually collected using the MPI device as test data sets.
[0083] A300, input the one-dimensional MPI signal into a pre-constructed generator to generate a two-dimensional MPI image as a predicted image; at the same time, input the binary image corresponding to the one-dimensional MPI signal input into the pre-constructed VGG16 model to obtain the category of the binary image, and the category of the binary image is used as a conditional constraint of the generator to enable image generation to proceed in a specified direction; in this embodiment, the generative adversarial network model includes a generator and a discriminator; the generator is constructed based on a fully connected layer, three encoding modules, and a linear inverse flattening layer connected in sequence; the encoding module includes a Transformer encoder; three upsampling units, the upsampling units are located between two Transformer encoders, and an upsampling unit (i.e. Figure 3 The 2× upsampling in , 2× means increasing the dimension of the image, for example, the matrix dimension before input is 8×8, and after upsampling it becomes 16×16), such as Figure 3 As shown, the linear inverse flattening layer is used to change the dimension of the feature matrix encoded by the Transformer encoder to convert it into a two-dimensional image;
[0084] The advantages of the generator constructed above are: 1. It can reduce the amount of calculation and memory; it can gradually increase the resolution of features but reduce the dimension of the embedding layer (better feature extraction and improved image resolution). The generator is mainly used to reconstruct the input one-dimensional MPI signal to obtain the corresponding two-dimensional MPI image (i.e. Figure 2 ).
[0085] While generating the two-dimensional MPI image, the binary image corresponding to the one-dimensional MPI signal of the input generator is input into the pre-built VGG16 model to obtain the category of the binary image, such as Figure 4 As shown. Since there are three types of images, we need to automatically classify the images first. For example, the classification result of letter images is 00, the classification result of resolution images is 01, and the classification result of geometric images is 10. If there is no way to distinguish the category, it is 11.
[0086] In addition, the overall structure of the VGG16 model is as follows Figure 4 As shown ( Figure 4 The size in the above example indicates the size, conv indicates the convolutional layer, pool indicates the pooling layer, and fc indicates the fully connected layer. Figure 4 3×3conv in the model has 64 channels in total, indicating a convolutional layer with a convolution kernel size of 3×3). The convolution kernel size (kernels) of the convolutional layer in the VGG16 model is 3×3, and the stride is 1, so there is no information loss. During training, the multi-classification cross loss function is used for optimization. The loss function of VGG16 is shown in formula (1):
[0087] Loss=α1*loss1+α2*loss2+α3*loss3 (1)
[0088] Among them, loss i (i=1, 2, 3) represent the cross entropy loss functions of the letter branch, resolution branch and geometry branch respectively, and α1, α2, α3 represent the preset weights corresponding to different branches. Existing research shows that setting appropriate weights for different branches in multi-tasks can improve the recognition accuracy of the entire or a single task.
[0089] A400, based on the category of the binary image and the predicted image, a loss value is obtained through a pre-constructed generator loss function, and the network parameters of the generator are updated; in this embodiment, the generator loss function is shown in formula (2):
[0090]
[0091] in, represents the loss function corresponding to the generator, G represents the generator, D represents the discriminator, y represents the category of the binary image, x represents the real image, G(z|y) represents the image generated by the generator, z represents the one-dimensional MPI signal of the input, represents the probability distribution function of x, λ G represents the constant coefficient, L MAE represents the mean absolute error function.
[0092] A500, inputting the predicted image and the true value label of the corresponding binary image into a pre-built discriminator to discriminate whether the predicted image is true or false;
[0093] In this embodiment, the discriminator is constructed based on N convolutional layers and fully connected layers connected in sequence; N is a positive integer. In the present invention, N is preferably set to 4.
[0094] The predicted image and the true value label of the corresponding binary image are input into the pre-built discriminator to determine whether the predicted image is true or false.
[0095] A600, combining the true and false discrimination result corresponding to the predicted image, obtaining a loss value through a pre-constructed discriminator loss function, and updating the network parameters of the discriminator;
[0096] In this embodiment, different discriminator loss functions are constructed for different inputs of the generator.
[0097] If the input of the generator is a one-dimensional MPI signal of a letter image, the discriminator loss function is as shown in formula (3):
[0098]
[0099] in, represents the loss function corresponding to the discriminator, z represents the input MPI one-dimensional voltage signal, represents the probability distribution function of z, L BCE represents the binary cross entropy loss function, t letter represents the features of the extracted letter image, D represents the discriminator, and λ D represents a preset second constant coefficient;
[0100] If the input of the generator is a one-dimensional MPI signal of a geometric image, the discriminator loss function is as shown in formula (4):
[0101]
[0102] Among them, t geometry Indicates the features of extracted geometric images;
[0103] If the input of the generator is a one-dimensional MPI signal of a resolution image, the loss function of the discriminator is as shown in formula (5):
[0104]
[0105] Among them, t resolution Represents the features of the extracted resolution image.
[0106] A700, loop A300-A600 until a trained generative adversarial network model is obtained.
[0107] Step S400, based on the three-dimensional reconstructed image, rigid registration is performed and brain regions are divided, thereby obtaining a time series of partition signals; this embodiment also includes: before performing rigid registration, a step of preprocessing the three-dimensional reconstructed image is also included, and the preprocessing includes time correction, head motion correction, registration, spatial normalization, smoothing and filtering;
[0108] In this embodiment, the partition signal time series is specifically obtained by:
[0109] Based on the three-dimensional reconstructed image, rigid registration is performed with an existing standard anatomical brain region template image to obtain a registered three-dimensional reconstructed image; the standard anatomical brain region template used in this embodiment is monkey F99;
[0110] In this embodiment, the rigid registration is as follows Figure 5 The specific methods shown are:
[0111] Assuming that there is only a translation change between the 3D reconstructed image and the standard anatomical brain region template, the translation change is selected as the registration method, and the registration formula is shown in formula (6):
[0112]
[0113] Among them, x′, y′ and z′ represent the coordinates of the pixels on the standard anatomical brain region template, x, y and z represent the coordinate values of the pixels on the 3D reconstructed image, Δ x , Δ y and Δ z is the translational variation coefficient to be solved;
[0114] The registration formula is calculated by comparing the three-dimensional reconstructed image with any three coordinate values of the standard anatomical brain region template.
[0115] The registered 3D reconstructed image is used to obtain a seed region of a preset size based on the existing standard anatomical brain region template image, and then a partitioned reconstructed image is obtained; the DMN network mainly includes the posterior cingulate cortex, precuneus, medial prefrontal cortex, inferior parietal lobule and bilateral temporal cortex, and these regions are selected as the regions of interest for analysis; taking the posterior cingulate cortex as an example, the posterior cingulate cortex is selected as the center point, and a brain region with a preset size of 10 mm is selected as the seed region based on the center point;
[0116] The partition reconstruction image is composed of multiple voxel signals, and each voxel signal is arranged according to time to obtain a partition signal time series. In this embodiment, the size of each voxel is 1mm*1mm*1mm. The process of obtaining the partition signal time series based on the partition reconstruction image is as follows: Figure 6 As shown;
[0117] Step S500, based on the partition signal time series, calculate the signal correlation coefficient to obtain the functional connection matrix, such as Figure 7 As shown, the default mode network is drawn on the three-dimensional reconstructed image according to the functional connection matrix; in this embodiment, the functional connection between voxels is drawn on the MPI image, and the default mode network finally obtained is as follows Figure 8 As shown, Figure 8 L represents the left hemisphere of the brain, and R represents the right hemisphere of the brain; the final default mode network reflects the activation of functional connection areas in the resting state;
[0118] In this embodiment, the functional connection matrix is as follows: Figure 7 As shown, the calculation method is:
[0119] Based on the partition signal time series, the Pearson correlation coefficient of each voxel signal and the other voxel signals of the same partition is calculated, where the i-th voxel ROI i With the j-th voxel ROI j The correlation coefficient is shown in formula (7):
[0120]
[0121] Among them, ROI i (t) represents the value of the tth time point in the partition signal time series of voxel i, ROI j (t) represents the value of the tth time point in the partition signal time series of voxel j, Represents voxel ROI i The average voxel signal at the last T time points, Represents voxel ROI jThe average voxel signal of the last T time points, N represents the number of voxel values in the region;
[0122] The functional connectivity matrix is obtained by combining the Pearson correlation coefficients of all voxel signals with the remaining voxel signals. The Pearson correlation coefficient reflects the synchronization between the two signals. The functional connectivity matrix finally obtained is a symmetrical correlation coefficient matrix, in which the larger the value, the stronger the functional connection; the more non-zero values, the more activation of the corresponding area.
[0123] The functional connection matrix of each brain region can be calculated by the calculation method of the functional connection matrix, and then the default mode network of all brain regions can be drawn;
[0124] Step S600, converting the correlation coefficient of the functional connection matrix into a z value, and performing a single-sample t-test statistical analysis on the z value to obtain the significance of the target functional connection to be modeled.
[0125] In this embodiment, the functional connectivity matrix is converted into a z value, and a single sample t test statistical analysis is performed on the z value, specifically:
[0126] Performing Fisher z-value transformation on the functional connectivity matrix to obtain a z-value;
[0127] The z value was subjected to a single sample t-test statistical analysis with a P value of 0.05 to obtain the significance of the target functional connection to be modeled. The significance of the target functional connection to be modeled was used to verify the reliability of the default mode network constructed by the present invention.
[0128] As an imaging technology for constructing the default mode network, MPI has the following advantages: 1. MPI imaging equipment has high sensitivity and can directly reflect changes in blood volume, while avoiding the attenuation of fMRI signals caused by BOLD signals. 2. MPI only images SPIONs, and there is no background signal from the brain in the image, so it has a high contrast. 3. SPIONs are a safe tracer that does not contain radioactive substances and can be metabolized in the body. At the same time, SPIONs have been used in clinical practice. The comparison of the effects of various methods is shown in Table 1.
[0129] Table 1 Comparison of the effects of MPI method, FMRI method and PET method
[0130] MPI fMRI PET Spatial resolution 10mm 2mm 20mm Temporal resolution >10 frames / s 25 frames / s A few minutes Sensitivity high Higher Low radioactivity none none have
[0131] Although PET technology can reflect changes in blood flow in the brain through glucose metabolism, the injected tracer is radioactive and the spatial resolution of PET is low. Although fMRI can reflect changes in oxygenated hemoglobin and deoxygenated hemoglobin in the blood flow, it cannot distinguish between signals from microvessels and large blood vessels in active brain areas, resulting in positioning errors.
[0132] Although the various steps in the above embodiment are described in the above-mentioned order, those skilled in the art can understand that in order to achieve the effect of this embodiment, different steps do not have to be executed in such an order. They can be executed simultaneously (in parallel) or in a reverse order. These simple changes are within the scope of protection of the present invention.
[0133] The default mode network construction system based on magnetic nanoparticle imaging according to the second embodiment of the present invention comprises:
[0134] a scan preparation module configured to apply a tracer to the target to be modeled and set scan parameters;
[0135] An MPI scanning module, configured to obtain an MPI one-dimensional voltage signal of a resting state of a target to be modeled based on the scanning parameters;
[0136] A three-dimensional reconstruction module, configured to perform image reconstruction based on the MPI one-dimensional voltage signal through a generative adversarial network to obtain a three-dimensional reconstructed image;
[0137] A brain region division module, configured to perform rigid registration and divide brain regions based on the three-dimensional reconstructed image, thereby obtaining a time series of partition signals;
[0138] A functional connection matrix acquisition module is configured to calculate the correlation coefficient between signals based on the partition signal time series, obtain the functional connection matrix, and draw the default mode network on the MPI image according to the functional connection matrix;
[0139] The significance verification module converts the correlation coefficient of the functional connection matrix into a z value, and performs a single-sample t-test statistical analysis on the z value to obtain the significance of the target functional connection to be modeled.
[0140] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process and related instructions of the system described above can refer to the corresponding process in the aforementioned method embodiment, and will not be repeated here.
[0141] It should be noted that the default mode network construction system based on magnetic nanoparticle imaging provided in the above embodiment is only illustrated by the division of the above functional modules. In practical applications, the above functions can be assigned to different functional modules as needed, that is, the modules or steps in the embodiments of the present invention can be decomposed or combined. For example, the modules in the above embodiment can be combined into one module, or further divided into multiple sub-modules to complete all or part of the functions described above. The names of the modules and steps involved in the embodiments of the present invention are only for distinguishing the modules or steps, and are not regarded as improper limitations of the present invention.
[0142] An electronic device according to the third embodiment of the present invention comprises: at least one processor; and a memory communicatively connected to at least one of the processors; wherein the memory stores instructions executable by the processor, and the instructions are used to be executed by the processor to implement the above-mentioned default mode network construction method based on magnetic nanoparticle imaging.
[0143] A computer-readable storage medium according to a fourth embodiment of the present invention stores computer instructions, and the computer instructions are used to be executed by the computer to implement the above-mentioned default mode network construction method based on magnetic nanoparticle imaging.
[0144] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process and related instructions of the storage device and processing device described above can refer to the corresponding process in the aforementioned method embodiment and will not be repeated here.
[0145] Those skilled in the art should be able to appreciate that the modules and method steps of each example described in conjunction with the embodiments disclosed herein can be implemented with electronic hardware, computer software or a combination of the two, and the programs corresponding to the software modules and method steps can be placed in random access memory (RAM), internal memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disks, removable disks, CD-ROMs, or any other form of storage medium known in the technical field. In order to clearly illustrate the interchangeability of electronic hardware and software, the composition and steps of each example have been generally described in the above description according to the function. Whether these functions are performed in electronic hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of the present invention.
[0146] Reference below Fig. 9 , which shows a schematic diagram of the structure of a computer system of a server for implementing the method, system, and device embodiments of the present application. Fig. 9 The server shown is merely an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.
[0147] like Fig. 9 As shown, the computer system includes a central processing unit (CPU) 901, which can perform various appropriate actions and processes according to the program stored in the read-only memory (ROM) 902 or the program loaded from the storage part 908 to the random access memory (RAM) 903. Various programs and data required for system operation are also stored in the RAM 903. The CPU 901, ROM 902 and RAM 903 are connected to each other through a bus 904. An input / output (I / O) interface 905 is also connected to the bus 904.
[0148] The following components are connected to the I / O interface 905: an input section 906 including a keyboard, a mouse, etc.; an output section 907 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker; a storage section 908 including a hard disk, etc.; and a communication section 909 including a network interface card such as a LAN (Local Area Network) card, a modem, etc. The communication section 909 performs communication processing via a network such as the Internet. A drive 910 is also connected to the I / O interface 905 as needed. A removable medium 911, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 910 as needed so that a computer program read therefrom is installed into the storage section 908 as needed.
[0149] In particular, according to an embodiment of the present disclosure, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present disclosure includes a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network through the communication part 909, and / or installed from the removable medium 911. When the computer program is executed by the central processing unit (CPU) 901, the above-mentioned functions defined in the method of the present application are executed. It should be noted that the above-mentioned computer-readable medium of the present application can be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium can be, for example, - but not limited to - a system, device or device of electricity, magnetism, light, electromagnetic, infrared, or semiconductor, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to, an electrical connection with one or more conductors, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present application, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, an apparatus or a device. In the present application, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, which carries a computer-readable program code. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. A computer-readable signal medium may also be any computer-readable medium other than a computer-readable storage medium, which may send, propagate, or transmit a program for use by or in combination with an instruction execution system, an apparatus or a device. The program code contained on the computer-readable medium may be transmitted using any appropriate medium, including but not limited to: wireless, wire, optical cable, RF, etc., or any suitable combination of the above.
[0150] Computer program code for performing the operations of the present application may be written in one or more programming languages or a combination thereof, including object-oriented programming languages, such as Java, Smalltalk, C++, and conventional procedural programming languages, such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider).
[0151] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present application. In this regard, each square box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the module, the program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the square box can also occur in a sequence different from that marked in the accompanying drawings. For example, two square boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each square box in the block diagram and / or flow chart, and the combination of the square boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.
[0152] The terms "first", "second", etc. are used to distinguish similar objects rather than to describe or indicate a particular order or sequence.
[0153] The term "comprise" or any other similar term is intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus / device that includes a list of elements includes not only those elements but also other elements not expressly listed, or also includes elements inherent to such process, method, article, or apparatus / device.
[0154] So far, the technical solutions of the present invention have been described in conjunction with the preferred embodiments shown in the accompanying drawings. However, it is easy for those skilled in the art to understand that the protection scope of the present invention is obviously not limited to these specific embodiments. Without departing from the principle of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will fall within the protection scope of the present invention.
Claims
1. A method for constructing a default mode network based on magnetic nanoparticle imaging, characterized in that: The method comprises: Apply tracer to the target to be modeled and set scanning parameters; Acquire the MPI one-dimensional voltage signal of the target to be modeled in a resting state based on the scanning parameters; Based on the MPI one-dimensional voltage signal, image reconstruction is performed by generating an adversarial network to obtain a three-dimensional reconstructed image; Based on the three-dimensional reconstructed image, rigid registration is performed and brain regions are divided, thereby obtaining a time series of partitioned signals; Calculating the correlation coefficient between signals based on the partition signal time series to obtain a functional connection matrix, and drawing a default mode network on the three-dimensional reconstructed image according to the functional connection matrix; The functional connection matrix is calculated as follows: Based on the partition signal time series, the Pearson correlation coefficient of each voxel signal and the other voxel signals of the same partition is calculated, where the i-th voxel ROI i With the j-th voxel ROI j The correlation coefficient is: Among them, ROI i (t) represents the value of the tth time point in the partition signal time series of voxel i, ROI j (t) represents the value of the tth time point in the partition signal time series of voxel j, Represents voxel ROI i The average voxel signal at the last T time points, Represents voxel ROI j The average value of the voxel signal at the last T time points, N represents the number of voxel values in the seed area of the preset size; The functional connectivity matrix was obtained by combining all voxel signals with the Pearson correlation coefficients of the remaining voxel signals; The correlation coefficient of the functional connection matrix is converted into a z value, and a single-sample t-test statistical analysis is performed on the z value to obtain the significance of the target functional connection to be modeled.
2. The method for constructing a default mode network based on magnetic nanoparticle imaging according to claim 1, characterized in that: The three-dimensional reconstructed image is obtained by: The MPI one-dimensional voltage signal is a one-dimensional voltage signal vector obtained by scanning frame by frame. The MPI one-dimensional voltage signal is reconstructed through a generative adversarial network at a speed of 10 frames per second, and a three-dimensional reconstructed image is obtained every 4 seconds.
3. The method for constructing a default mode network based on magnetic nanoparticle imaging according to claim 1, characterized in that: The specific method for obtaining the partition signal time series is as follows: Based on the three-dimensional reconstructed image, rigid registration is performed with an existing standard anatomical brain region template image to obtain a registered three-dimensional reconstructed image; The registered three-dimensional reconstructed image is used to obtain a seed region of a preset size according to an existing standard anatomical brain region template image, thereby obtaining a partitioned reconstructed image; The partition reconstructed image is composed of a plurality of voxel signals, and each voxel signal is arranged according to time to obtain a partition signal time series.
4. The method for constructing a default mode network based on magnetic nanoparticle imaging according to claim 1, characterized in that: The functional connectivity matrix is converted into a z value, and a single sample t test statistical analysis is performed on the z value, specifically: Performing Fisher z-value transformation on the functional connectivity matrix to obtain a z-value; The z values were subjected to one-sample t-test statistical analysis with a P value of 0.05 to obtain the default mode network.
5. The method for constructing a default mode network based on magnetic nanoparticle imaging according to claim 1, characterized in that: The scanning parameters are specifically set as follows: The excitation magnetic field was set to 5 mT, the excitation frequency to 25 Hz, and the magnetic field gradients to: Z direction: 0.2 T / m, X direction: 0.1 T / m, Y direction: 0.1 T / m.
6. The method for constructing a default mode network based on magnetic nanoparticle imaging according to claim 1, characterized in that: The tracer is administered to the target to be modeled, specifically, the tracer is injected based on the body weight of the target to be modeled, with a dosage of 8-12 mg / kg, and the target to be modeled is placed in a resting state.
7. A default mode network construction system based on magnetic nanoparticle imaging, characterized in that: The system comprises: a scan preparation module configured to apply a tracer to the target to be modeled and set scan parameters; An MPI scanning module, configured to obtain an MPI one-dimensional voltage signal of a resting state of a target to be modeled based on the scanning parameters; A three-dimensional reconstruction module, configured to perform image reconstruction based on the MPI one-dimensional voltage signal through a generative adversarial network to obtain a three-dimensional reconstructed image; A brain region division module, configured to perform rigid registration and divide brain regions based on the three-dimensional reconstructed image, thereby obtaining a time series of partition signals; a functional connection matrix acquisition module, configured to calculate the correlation coefficient between signals based on the partition signal time series, obtain the functional connection matrix, and draw the default mode network on the three-dimensional reconstructed image according to the functional connection matrix; The functional connection matrix is calculated as follows: Based on the partition signal time series, the Pearson correlation coefficient of each voxel signal and the other voxel signals of the same partition is calculated, where the i-th voxel ROI i With the j-th voxel ROI j The correlation coefficient is: Among them, ROI i (t) represents the value of the tth time point in the partition signal time series of voxel i, ROI j (t) represents the value of the tth time point in the partition signal time series of voxel j, Represents voxel ROI i The average voxel signal at the last T time points, Represents voxel ROI j The average value of the voxel signal at the last T time points, N represents the number of voxel values in the seed area of the preset size; The functional connectivity matrix was obtained by combining all voxel signals with the Pearson correlation coefficients of the remaining voxel signals; The significance authentication module converts the correlation coefficient of the functional connection matrix into a z value, and performs a single-sample t-test statistical analysis on the z value to obtain the significance of the target functional connection to be modeled.
8. An electronic device, characterized in that: include: at least one processor; And a memory communicatively connected to at least one of the processors; wherein the memory stores instructions executable by the processor, and the instructions are used to be executed by the processor to implement the default mode network construction method based on magnetic nanoparticle imaging as described in any one of claims 1-6.
9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions are used to be executed by the computer to implement the default mode network construction method based on magnetic nanoparticle imaging according to any one of claims 1 to 6.
