Individual difference-based sMRI image analysis method, system and terminal

Abstract

The invention discloses an sMRI image analysis method, system and terminal based on individual differences, and the method comprises the steps: obtaining individual difference sample data, and building a brain structure individual difference model according to the individual difference sample data; constructing an sMRI image analysis model, and training and testing the sMRI image analysis model to obtain a target model; obtaining a target sMRI image, target individual data and AD related SNPs of a user to be analyzed, obtaining a target ROI feature according to the target sMRI image, inputting the target individual data into the brain structure individual difference model to obtain a normal ROI feature, and calculating an ROI feature atrophy rate according to the target ROI feature and the normal ROI feature; and inputting the target ROI feature, the ROI feature atrophy rate and the AD-related SNPs into a target model for analysis to obtain a classification result of the target sMRI image. According to the method, the individual difference of the brain structure is evaluated, sMRI image analysis is carried out based on the individual difference of the brain structure, and the sMRI image classification accuracy is improved.

Description

Technical Field

[0001] The present invention relates to the field of image processing technology, and in particular to an sMRI image analysis method, system, terminal and computer-readable storage medium based on individual differences. Background Art

[0002] Individual differences in brain structure are significant. This difference is not only between men and women, but also exists between individuals. There are associations between brain structure and genes, aging, and modifiable factors (such as lifestyle, medical history, local environment, physical measures, and psychosocial factors). In fact, aging is also a significant risk factor for various brain structural abnormalities (such as hippocampal atrophy).

[0003] Structural magnetic resonance imaging (sMRI) is a method that uses magnetic resonance imaging technology to perform non-invasive three-dimensional imaging of the human or animal brain. It can provide high-precision anatomical information, such as the location, size and shape of structures such as gray matter, white matter and ventricles in the brain, and reveal information about the structural composition and development process of the brain.

[0004] In the existing technology, usually only structural magnetic resonance imaging technology is relied upon to obtain the brain structure and thus discover abnormal brain structure, but the significant individual differences in brain structure are not taken into account. When analyzing the brain structure, it is necessary to consider the significant individual differences in brain structure. Currently, there is no reference standard for quantifying individual differences. It is difficult to obtain accurate classification results of brain structure changes in different individuals based only on sMRI images.

[0005] Therefore, the prior art still needs to be improved and developed. Summary of the invention

[0006] The main purpose of the present invention is to provide an sMRI image analysis method, system, terminal and computer-readable storage medium based on individual differences, aiming to solve the problem that in the prior art, when analyzing sMRI images, the differences between individuals are not taken into account, resulting in inaccurate sMRI image classification results.

[0007] To achieve the above object, the present invention provides an sMRI image analysis method based on individual differences, the sMRI image analysis method based on individual differences comprising the following steps:

[0008] Acquiring individual difference sample data, developing a nomogram based on the individual difference sample data, and establishing a brain structure individual difference model based on the nomogram;

[0009] Constructing an sMRI image analysis model, training and testing the sMRI image analysis model to obtain a target model;

[0010] Obtaining a target sMRI image, target individual data, and AD-related SNPs of a user to be analyzed, obtaining a target ROI feature according to the target sMRI image, inputting the target individual data into the brain structure individual difference model to obtain a normal ROI feature, and calculating a ROI feature atrophy rate according to the target ROI feature and the normal ROI feature;

[0011] The target ROI features, the ROI feature atrophy rate and the AD-related SNPs are input into the target model for analysis to obtain a classification result of the target sMRI image, and an sMRI image analysis report based on individual differences is generated according to the classification result.

[0012] Optionally, in the sMRI image analysis method based on individual differences, the individual difference sample data includes gender, age, genetic SNPs and modifiable factors.

[0013] Optionally, the sMRI image analysis method based on individual differences, wherein the developing a nomogram based on the individual difference sample data and establishing a brain structure individual difference model based on the nomogram specifically includes:

[0014] Based on the genetic SNPs, a polygenic risk score is obtained by an allele count weighting method, and GWAS results related to AD are obtained in the whole human genome;

[0015] Based on the t-test method, multiple AD-related ROI features are identified on the ADNI data set, and a first nomogram of the multiple AD-related ROI features and the GWAS results is established according to the polygenic risk score;

[0016] The association between environmental factors and epigenetic modifications was identified by detecting the correlation between methylation and histone modification sites and phenotypic traits, and EWAS results related to AD were obtained;

[0017] The modifiable factors were subjected to t-test and Bonferroni correction according to the age and gender to obtain significant factors;

[0018] According to the EWAS results, a second nomogram of the plurality of AD-related ROI features and the significant factors is established by a multivariate linear regression method, and a brain structure individual difference model is established based on the first nomogram and the second nomogram.

[0019] Optionally, in the sMRI image analysis method based on individual differences, the target model includes: a graph convolutional neural network, a Transformer neural network and a multi-layer perceptron.

[0020] Optionally, the sMRI image analysis method based on individual differences, wherein the training and testing of the sMRI image analysis model specifically includes:

[0021] Collecting sMRI images and individual data of multiple users from a database, and obtaining AD-related SNPs, and obtaining ROI features and ROI feature atrophy rates of the multiple users according to the sMRI images and the individual data;

[0022] Taking the ROI features, the ROI feature atrophy rate and the AD-related SNPs as sample data, taking the sMRI image classification results corresponding to the plurality of users as labels, and establishing a data set according to the sample data and the labels;

[0023] The data set is divided into a training set and a test set according to a preset ratio, the training set is used to train the sMRI image analysis model to obtain a trained model, and the test set is used to perform performance evaluation on the trained model to obtain the target model that meets the requirements.

[0024] Optionally, the sMRI image analysis method based on individual differences, wherein the target ROI feature is obtained according to the target sMRI image, the target individual data is input into the brain structure individual difference model to obtain the normal ROI feature, and the ROI feature atrophy rate is calculated according to the target ROI feature and the normal ROI feature, specifically includes:

[0025] The target sMRI image is processed using the FreeSurfer software package to obtain gray matter volume, white matter volume, cerebral cortex area and cerebral cortex thickness, and multiple target ROI features are obtained according to the gray matter volume, the white matter volume, the cerebral cortex area and the cerebral cortex thickness;

[0026] Inputting the target individual data into the plurality of brain structure individual difference models for calculation, and outputting a plurality of normal ROI features, wherein the target individual data includes gender, age, genetic SNPs and modifiable factors;

[0027] The ROI feature shrinkage rate of each target ROI feature is calculated according to the multiple target ROI features and the multiple normal ROI features.

[0028] Optionally, the sMRI image analysis method based on individual differences, wherein the target ROI features, the ROI feature atrophy rate and the AD-related SNPs are input into the target model for analysis to obtain the classification result of the target sMRI image, specifically includes:

[0029] Inputting the target ROI feature, the ROI feature atrophy rate, and the AD-related SNPs into the graph convolutional neural network for feature extraction to obtain three intermediate features, and connecting the three intermediate features in series to obtain multi-omics features;

[0030] The multi-omics features are input into the Transformer neural network for correlation analysis to obtain target features, the target features are weighted summed through the multi-layer perceptron, and the result of the weighted summation is output using an activation function to obtain the classification result of the target sMRI image.

[0031] In addition, to achieve the above-mentioned purpose, the present invention further provides an sMRI image analysis system based on individual differences, wherein the sMRI image analysis system based on individual differences comprises:

[0032] An individual difference model building module is used to obtain individual difference sample data, develop a nomogram based on the individual difference sample data, and establish an individual difference model of brain structure based on the nomogram;

[0033] An image analysis model building module is used to build an sMRI image analysis model, train and test the sMRI image analysis model, and obtain a target model;

[0034] A data processing module, used to obtain a target sMRI image, target individual data and AD-related SNPs of a user to be analyzed, obtain a target ROI feature according to the target sMRI image, input the target individual data into the brain structure individual difference model to obtain a normal ROI feature, and calculate a ROI feature atrophy rate according to the target ROI feature and the normal ROI feature;

[0035] The model application module is used to input the target ROI features, the ROI feature atrophy rate and the AD-related SNPs into the target model for analysis, obtain the classification result of the target sMRI image, and generate an sMRI image analysis report based on individual differences according to the classification result.

[0036] In addition, to achieve the above-mentioned purpose, the present invention also provides a terminal, wherein the terminal includes: a memory, a processor, and an sMRI image analysis program based on individual differences stored in the memory and executable on the processor, wherein the sMRI image analysis program based on individual differences, when executed by the processor, implements the steps of the sMRI image analysis method based on individual differences as described above.

[0037] In addition, to achieve the above-mentioned purpose, the present invention also provides a computer-readable storage medium, wherein the computer-readable storage medium stores an sMRI image analysis program based on individual differences, and when the sMRI image analysis program based on individual differences is executed by a processor, the steps of the sMRI image analysis method based on individual differences as described above are implemented.

[0038] In the present invention, a target fundus image and target non-imaging data are obtained, the fundus image is input into an encoder to output visual features; the visual features are used as features of nodes of a feature hypergraph to construct a hyperedge group of a feature hypergraph, and a first association matrix of the feature hypergraph is obtained; non-imaging data are classified to obtain multiple groups to obtain hyperedges of an attribute hypergraph, and a hyperedge group of an attribute hypergraph is constructed according to multiple hyperedges to obtain a second association matrix of the attribute hypergraph; the first association matrix and the second association matrix are spliced ​​to obtain a target association matrix of the target hypergraph; the nodes of the feature hypergraph and the target association matrix are input into a weighted hypergraph convolution module to obtain target node features, the probability of each category is obtained according to the target node features, and an sMRI image analysis report based on individual differences is generated according to the probability. The present invention shows excellent performance when processing sMRI image analysis tasks based on individual differences, and provides a new perspective and a powerful tool for high-order relationship modeling. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 It is a flow chart of a preferred embodiment of the sMRI image analysis method based on individual differences of the present invention;

[0040] Figure 2 It is the overall framework diagram of the individual difference model of brain structure in the sMRI image analysis method based on individual differences of the present invention;

[0041] Figure 3 It is a schematic diagram of the realization principle of the target model in the sMRI image analysis method based on individual differences of the present invention;

[0042] Figure 4 It is a schematic diagram of the principle of a preferred embodiment of the sMRI image analysis system based on individual differences of the present invention;

[0043] Figure 5 Schematic diagram of the operating environment of a preferred embodiment of the terminal of the present invention. DETAILED DESCRIPTION

[0044] The present application provides an sMRI image analysis method based on individual differences and related equipment. In order to make the purpose, technical solution and effect of the present application clearer and more specific, the present application is further described in detail with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0045] It will be understood by those skilled in the art that, unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as those generally understood by those skilled in the art to which this application belongs. It should also be understood that terms such as those defined in general dictionaries should be understood to have meanings consistent with those in the context of the prior art, and will not be interpreted with idealized or overly formal meanings unless specifically defined as here.

[0046] In addition, if there are descriptions involving "first", "second", etc. in the embodiments of the present invention, the descriptions of "first", "second", etc. are only used for descriptive purposes and cannot be understood as indicating or suggesting their relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In addition, the technical solutions between the various embodiments can be combined with each other, but they must be based on the ability of ordinary technicians in the field to implement them. When the combination of technical solutions is contradictory or cannot be implemented, it should be deemed that such a combination of technical solutions does not exist and is not within the scope of protection required by the present invention.

[0047] The sMRI image analysis method based on individual differences described in the preferred embodiment of the present invention is as follows: Figure 1 As shown, the sMRI image analysis method based on individual differences comprises the following steps:

[0048] Step S10: acquiring individual difference sample data, developing a nomogram based on the individual difference sample data, and establishing a brain structure individual difference model based on the nomogram.

[0049] SNPs (single nucleotide polymorphism) mainly refers to DNA sequence polymorphisms caused by variations in single nucleotides at the genomic level. GWAS (Genome-Wide Association Studies) refers to finding existing sequence variations within the human genome, that is, screening out disease-related SNPs from single nucleotide polymorphisms (SNPs). It is understandable that due to the complexity of brain development and aging, the brain structure difference SNPs determined therefrom show significant variability. Therefore, the present application constructs a model of individual differences in brain structure, and by sequentially inputting the user's gender, age, polygenic risk score (PRS) and modifiable factors into the established difference model, the user's estimated ROI (region of interest) characteristic normal value can be obtained.

[0050] Specifically, Figure 2 As shown, based on the genetic SNPs, a polygenic risk score was obtained by an allele count weighting method, and GWAS results associated with AD were obtained across the entire human genome.

[0051] Based on the t-test method, multiple AD (Alzheimer's disease) related ROI features are identified on the ADNI data set, and a first nomogram of the multiple AD related ROI features and the AD sequence variation is established according to the polygenic risk score.

[0052] In this embodiment, based on the genetic SNPs, polygenic risk scores (PRS) are obtained by an allele count weighted method, and the polygenic risk scores are used to establish an association between multiple (24 in this embodiment) ROI features and genetic variations based on GWAS results (AD-related GWAS results).

[0053] The PRS is the sum of the effects of a set of genetic variants on a trait, weighted by the allele counts: Among them, ES i represents the effect size, i represents the ordinal number of the SNP, C i represents the number of alleles of the i-th SNP.

[0054] Further, in this embodiment, a biological sample library and ADNI data set of a certain country are obtained, and the minor allele frequency of 0.001, the missing genotype of 0.02 and the aggregation of 250kb are filtered. The potential SNPs of the sample library of a certain country after filtering are 11282349, and the ADNI data set after filtering is 10648207. Multiple (24) AD-related ROI features are identified from the biological sample library and ADNI data set of a certain country after filtering. A first nomogram of multiple AD-related ROI features and AD sequence variation is established according to the polygenic risk score PRS.

[0055] Furthermore, the association between environmental factors and epigenetic modifications was identified by detecting the correlation between methylation and histone modification sites and phenotypic characteristics, and EWAS results related to AD were obtained.

[0056] It is understandable that GWAS searches for genes associated with human diseases by studying genetic variations at a large number of gene loci, and its basic idea is to analyze the correlation between genetic variations and phenotypic characteristics (such as disease status, height, etc.). GWAS usually uses gene chips or sequencing technology, which can cover the entire genome or most of the genome, so it can fully explore the relationship between genes and diseases, but the single gene loci discovered by GWAS usually have only a slight effect on the risk of disease.

[0057] In contrast, EWAS (Epigenome-Wide Association Study) focuses on epigenetic research, i.e., epigenetic modifications in gene expression regulation processes such as DNA methylation and histone modification. EWAS uses a method similar to GWAS to identify the association between environmental factors and epigenetic modifications by detecting the correlation between methylation and histone modification sites and phenotypic characteristics. Unlike GWAS, EWAS requires the use of higher-resolution technology to detect epigenetic modifications, so a larger sample size and more sophisticated data analysis methods are required. EWAS can identify key epigenetic markers that have a more direct impact on the occurrence and development of diseases. Therefore, in this embodiment, the association between environmental factors and epigenetic modifications is identified by detecting the correlation between methylation and histone modification sites and phenotypic characteristics, and EWAS results related to AD are obtained, thereby providing more accurate results for image classification.

[0058] Furthermore, the modifiable factors were subjected to t-test and Bonferroni correction according to the age and gender to obtain significant factors.

[0059] Specifically, a t-test was performed for each factor according to age and gender, and then the most significant factor after Bonferroni correction was selected and the EIS (Environmental Impact Score) was calculated, which was used to represent the sum of the effects of the most significant risk modifiable factors:

[0060] EIS=β1x1+β2x2+…+β p x p ;

[0061] Among them, β1,β2,…β p are coefficients, x1, x2, … x p were significant modifiable variables from the clinical variables. In addition, all significant factors were applied together to establish a multiple linear regression to compare the effect size of each factor.

[0062] Furthermore, based on the EWAS results, a second nomogram of the multiple AD-related ROI features and the significant factors is established by a multivariate linear regression method, and a brain structure individual difference model is established based on the first nomogram and the second nomogram.

[0063] Specifically, based on the EWAS results, multiple linear regression was used to establish the association between each ROI feature and the risk modifiable factors. The multiple linear regression formula used was as follows:

[0064] y=β0+β1x1+β2x2+…+β p x p +∈; where β0 is a constant and ε is the random error (residual) term of the model.

[0065] Finally, multiple (24) brain structure individual difference models can be established based on the first nomogram and the second nomogram. Each brain structure individual difference model represents an AD-related ROI feature and outputs an estimated (normal) ROI feature value after inputting individual data (gender, age, genetic SNPs and modifiable factors).

[0066] Step S20: construct an sMRI image analysis model, train and test the sMRI image analysis model to obtain a target model.

[0067] like Figure 3As shown, the target model includes: Graph Convolutional Network (GCN), Transformer neural network and Multilayer Perceptron (MLP). The input features of the target model include ROI features from sMRI (including ROI volume, cortical surface area and cortical thickness), atrophy rate of AD-related ROI features (obtained from the brain structure individual difference model) and well-known AD-related SNPs. Among them, GCN and Transformer have personalized loss functions and learn deep features.

[0068] The training and testing of the sMRI image analysis model specifically includes:

[0069] Collecting sMRI images and individual data of multiple users from a database, and obtaining AD-related SNPs, and obtaining ROI features and ROI feature atrophy rates of the multiple users according to the sMRI images and the individual data;

[0070] Taking the ROI features, the ROI feature atrophy rate and the AD-related SNPs as sample data, taking the sMRI image classification results corresponding to the plurality of users as labels, and establishing a data set according to the sample data and the labels;

[0071] The data set is divided into a training set and a test set according to a preset ratio, the training set is used to train the sMRI image analysis model to obtain a trained model, and the test set is used to perform performance evaluation on the trained model to obtain the target model that meets the requirements.

[0072] It is understood that the AD-related SNPs were obtained from previous studies, including 89 common risk SNPs. After comparison and screening with various databases, 82 common risk SNPs were retained, including well-known AD-related SNPs (such as ApoEε3, ApoEε4, APP, PSEN1, PSEN2, et.al). Plink software was used to encode the screened SNP data and uniformly encode the 012 genotype.

[0073] Furthermore, the sMRI images are processed using the FreeSurfer software package (freesurfer is a toolkit for analyzing and visualizing brain structural imaging and functional imaging, which can process MRI, fMRI data, and measure brain anatomical data, etc.) to obtain gray matter volume, white matter volume, cerebral cortical area, and cerebral cortical thickness. ROI features of multiple users are obtained based on the gray matter volume, white matter volume, cerebral cortical area, and cerebral cortical thickness, and individual data of the multiple users are respectively input into multiple brain structure individual difference models for calculation to obtain multiple estimated ROI features; finally, the ROI feature atrophy rate is obtained based on the ROI features of the multiple users and the multiple estimated ROI features.

[0074] Furthermore, the ROI features, the ROI feature atrophy rate and the AD-related SNPs are used as sample data, and the sMRI image classification results corresponding to the multiple users are used as labels corresponding to the sample data, and a data set is established according to the sample data and the labels. The data set is divided into a training set and a test set according to a preset ratio (for example, a ratio of 7:3), the sMRI image analysis model is trained using the training set to obtain a trained model, and the performance of the trained model is evaluated using the test set to detect the correctness of the prediction results of the trained model, so as to obtain the target model that meets the requirements.

[0075] Step 30: obtain the target sMRI image, target individual data and AD-related SNPs of the user to be analyzed, obtain the target ROI feature according to the target sMRI image, input the target individual data into the brain structure individual difference model to obtain the normal ROI feature, and calculate the ROI feature atrophy rate according to the target ROI feature and the normal ROI feature.

[0076] Specifically, a target sMRI image of the user to be analyzed is obtained by nuclear magnetic resonance technology, AD-related SNPs are obtained from previous studies, and the target individual data of the user to be analyzed is collected. The target sMRI image is processed using the FreeSurfer software package to obtain gray matter volume, white matter volume, cerebral cortex area and cerebral cortex thickness, and multiple target ROI features are obtained according to the gray matter volume, the white matter volume, the cerebral cortex area and the cerebral cortex thickness.

[0077] Furthermore, the target individual data are respectively input into a plurality of the brain structure individual difference models for calculation, and a plurality of normal ROI features are output, wherein the target individual data include gender, age, genetic SNPs and modifiable factors; as an example, the target individual data are respectively input into 24 brain structure individual difference models for calculation, and 24 normal ROI features can be output.

[0078] Further, the ROI feature shrinkage rate of each target ROI feature is calculated based on the plurality of target ROI features and the plurality of normal ROI features. According to the individual difference model, the estimated normal values ​​of the 24 ROI features of each user can be obtained.

[0079] The shrinkage rate of feature j of the target user is: Among them, j is the ordinal number of 24 ROI features, Atr j represents the shrinkage rate of the jth ROI feature of the target user, F j represents the value of the jth target ROI feature of the target user, Represents the normal value (estimated value) of the j-th ROI feature of the target user.

[0080] Step 40: input the target ROI features, the ROI feature atrophy rate and the AD-related SNPs into the target model for analysis to obtain a classification result of the target sMRI image, and generate an sMRI image analysis report based on individual differences according to the classification result.

[0081] like Figure 3 As shown, specifically, the target ROI feature, the ROI feature atrophy rate and the AD-related SNPs are input into the graph convolutional neural network for feature extraction to obtain three intermediate features, and the three intermediate features are connected in series to obtain multi-omics features.

[0082] Furthermore, the multi-omics features are input into the Transformer neural network for correlation analysis to obtain target features, the target features are weighted summed through the multi-layer perceptron, and the result of the weighted summation is output using an activation function to obtain a classification result of the target sMRI image.

[0083] It can be understood that in this embodiment, Transformer is used to analyze the correlation of multiple omics features. Its core component (self-attention mechanism) is expressed as:

[0084]

[0085] where Q∈R M×N , K∈R M×N and V∈RM×N are the outputs of the three feed-forward networks. is a scaling factor. The self-attention mechanism has a similar architecture to GCN, where the learnable and trainable weight term (i.e. It can also be viewed as an adjacency matrix. The two matrices (i.e., Q and K T ) captures the intrinsic relationship between features.

[0086] It can be seen that the present invention evaluates the individual differences in brain structure and obtains a difference model. According to the normal brain structure characteristics estimated by the difference model, the atrophy rate of these key brain structure characteristics is calculated, and the graph convolutional network (GCN) and Transformer model are used as prediction networks. The calculated atrophy rate characteristics and the obtained AD-related gene characteristics are used as network inputs to learn deep features and further achieve accurate classification of sMRI images.

[0087] Furthermore, if Figure 4 As shown, based on the above-mentioned sMRI image analysis method based on individual differences, the present invention also provides a sMRI image analysis system based on individual differences, wherein the sMRI image analysis system based on individual differences includes:

[0088] An individual difference model building module 51 is used to obtain individual difference sample data, develop a nomogram based on the individual difference sample data, and establish an individual difference model of brain structure based on the nomogram;

[0089] An image analysis model building module 52 is used to build an sMRI image analysis model, train and test the sMRI image analysis model, and obtain a target model;

[0090] The data processing module 53 is used to obtain the target sMRI image, target individual data and AD-related SNPs of the user to be analyzed, obtain the target ROI feature according to the target sMRI image, input the target individual data into the brain structure individual difference model to obtain the normal ROI feature, and calculate the ROI feature atrophy rate according to the target ROI feature and the normal ROI feature;

[0091] The model application module 54 is used to input the target ROI features, the ROI feature atrophy rate and the AD-related SNPs into the target model for analysis, obtain the classification result of the target sMRI image, and generate an sMRI image analysis report based on individual differences according to the classification result.

[0092] Furthermore, if Figure 5As shown, based on the above-mentioned sMRI image analysis method and system based on individual differences, the present invention also provides a terminal accordingly, and the terminal includes a processor 10, a memory 20 and a display 30. Figure 5 Only some components of the terminal are shown, but it should be understood that it is not required to implement all of the components shown, and more or fewer components may be implemented instead.

[0093] In some embodiments, the memory 20 may be an internal storage unit of the terminal, such as a hard disk or memory of the terminal. In other embodiments, the memory 20 may also be an external storage device of the terminal, such as a plug-in hard disk, a smart memory card (SmartMedia Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), etc. equipped on the terminal. Further, the memory 20 may also include both an internal storage unit of the terminal and an external storage device. The memory 20 is used to store application software and various types of data installed in the terminal, such as the program code of the installation terminal. The memory 20 may also be used to temporarily store data that has been output or is to be output. In one embodiment, an sMRI image analysis program 40 based on individual differences is stored on the memory 20, and the sMRI image analysis program 40 based on individual differences can be executed by the processor 10, thereby realizing the sMRI image analysis method based on individual differences in the present application.

[0094] In some embodiments, the processor 10 may be a central processing unit (CPU), a microprocessor or other data processing chip, used to run the program code or process data stored in the memory 20, such as executing the sMRI image analysis method based on individual differences.

[0095] In some embodiments, the display 30 may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, an OLED (Organic Light-Emitting Diode) touch device, etc. The display 30 is used to display information on the terminal and to display a visual user interface. The components 10-30 of the terminal communicate with each other via a system bus.

[0096] In one embodiment, when the processor 10 executes the sMRI image analysis program 40 based on individual differences in the memory 20, the following steps are implemented:

[0097] Acquiring individual difference sample data, developing a nomogram based on the individual difference sample data, and establishing a brain structure individual difference model based on the nomogram;

[0098] Constructing an sMRI image analysis model, training and testing the sMRI image analysis model to obtain a target model;

[0099] Obtaining a target sMRI image, target individual data, and AD-related SNPs of a user to be analyzed, obtaining a target ROI feature according to the target sMRI image, inputting the target individual data into the brain structure individual difference model to obtain a normal ROI feature, and calculating a ROI feature atrophy rate according to the target ROI feature and the normal ROI feature;

[0100] The target ROI features, the ROI feature atrophy rate and the AD-related SNPs are input into the target model for analysis to obtain a classification result of the target sMRI image, and an sMRI image analysis report based on individual differences is generated according to the classification result.

[0101] The individual difference sample data include gender, age, genetic SNPs and modifiable factors.

[0102] The developing of a nomogram based on the individual difference sample data and establishing a brain structure individual difference model based on the nomogram specifically include:

[0103] Based on the genetic SNPs, a polygenic risk score is obtained by an allele count weighting method, and GWAS results related to AD are obtained in the whole human genome;

[0104] Based on the t-test method, multiple AD-related ROI features are identified on the ADNI data set, and a first nomogram of the multiple AD-related ROI features and the GWAS results is established according to the polygenic risk score;

[0105] The association between environmental factors and epigenetic modifications was identified by detecting the correlation between methylation and histone modification sites and phenotypic traits, and EWAS results related to AD were obtained;

[0106] The modifiable factors were subjected to t-test and Bonferroni correction according to the age and gender to obtain significant factors;

[0107] According to the EWAS results, a second nomogram of the plurality of AD-related ROI features and the significant factors is established by a multivariate linear regression method, and a brain structure individual difference model is established based on the first nomogram and the second nomogram.

[0108] Among them, the target model includes: graph convolutional neural network, Transformer neural network and multi-layer perceptron.

[0109] The training and testing of the sMRI image analysis model specifically includes:

[0110] Collecting sMRI images and individual data of multiple users from a database, and obtaining AD-related SNPs, and obtaining ROI features and ROI feature atrophy rates of the multiple users according to the sMRI images and the individual data;

[0111] Taking the ROI features, the ROI feature atrophy rate and the AD-related SNPs as sample data, taking the sMRI image classification results corresponding to the plurality of users as labels, and establishing a data set according to the sample data and the labels;

[0112] The data set is divided into a training set and a test set according to a preset ratio, the training set is used to train the sMRI image analysis model to obtain a trained model, and the test set is used to perform performance evaluation on the trained model to obtain the target model that meets the requirements.

[0113] The step of obtaining a target ROI feature according to the target sMRI image, inputting the target individual data into the brain structure individual difference model to obtain a normal ROI feature, and calculating the ROI feature atrophy rate according to the target ROI feature and the normal ROI feature specifically includes:

[0114] The target sMRI image is processed using the FreeSurfer software package to obtain gray matter volume, white matter volume, cerebral cortex area and cerebral cortex thickness, and multiple target ROI features are obtained according to the gray matter volume, the white matter volume, the cerebral cortex area and the cerebral cortex thickness;

[0115] Inputting the target individual data into the plurality of brain structure individual difference models for calculation, and outputting a plurality of normal ROI features, wherein the target individual data includes gender, age, genetic SNPs and modifiable factors;

[0116] The ROI feature shrinkage rate of each target ROI feature is calculated according to the multiple target ROI features and the multiple normal ROI features.

[0117] The step of inputting the target ROI features, the ROI feature atrophy rate and the AD-related SNPs into the target model for analysis to obtain the classification result of the target sMRI image specifically includes:

[0118] Inputting the target ROI feature, the ROI feature atrophy rate, and the AD-related SNPs into the graph convolutional neural network for feature extraction to obtain three intermediate features, and connecting the three intermediate features in series to obtain multi-omics features;

[0119] The multi-omics features are input into the Transformer neural network for correlation analysis to obtain target features, the target features are weighted summed through the multi-layer perceptron, and the result of the weighted summation is output using an activation function to obtain the classification result of the target sMRI image.

[0120] In summary, the present invention provides an sMRI image analysis method based on individual differences and related equipment, the method comprising: obtaining individual difference sample data, establishing a brain structure individual difference model according to the individual difference sample data; constructing an sMRI image analysis model, training and testing the sMRI image analysis model, and obtaining a target model; obtaining a target sMRI image, target individual data, and AD-related SNPs of a user to be analyzed, obtaining a target ROI feature according to the target sMRI image, inputting the target individual data into the brain structure individual difference model to obtain a normal ROI feature, and calculating the ROI feature atrophy rate according to the target ROI feature and the normal ROI feature; inputting the target ROI feature, the ROI feature atrophy rate, and the AD-related SNPs into the target model for analysis, and obtaining a classification result of the target sMRI image. The present invention evaluates individual differences in brain structure, and performs sMRI image analysis based on individual differences in brain structure, thereby improving the accuracy of sMRI image classification.

[0121] It should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or terminal including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or terminal. In the absence of further restrictions, an element defined by the sentence "comprises a ..." does not exclude the presence of other identical elements in the process, method, article or terminal including the element.

[0122] Of course, those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided by the present invention can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0123] It should be understood that the application of the present invention is not limited to the above examples. For ordinary technicians in this field, improvements or changes can be made based on the above description. All these improvements and changes should fall within the scope of protection of the claims attached to the present invention.

Claims

1. A sMRI image analysis method based on individual differences, characterized in that: The sMRI image analysis method based on individual differences comprises: Acquiring individual difference sample data, developing a nomogram based on the individual difference sample data, and establishing a brain structure individual difference model based on the nomogram; Constructing an sMRI image analysis model, training and testing the sMRI image analysis model to obtain a target model; Obtaining a target sMRI image, target individual data, and AD-related SNPs of a user to be analyzed, obtaining a target ROI feature according to the target sMRI image, inputting the target individual data into the brain structure individual difference model to obtain a normal ROI feature, and calculating a ROI feature atrophy rate according to the target ROI feature and the normal ROI feature; The target ROI features, the ROI feature atrophy rate and the AD-related SNPs are input into the target model for analysis to obtain a classification result of the target sMRI image, and an sMRI image analysis report based on individual differences is generated according to the classification result.

2. The sMRI image analysis method based on individual differences according to claim 1, characterized in that: The individual difference sample data include gender, age, genetic SNPs and modifiable factors.

3. The sMRI image analysis method based on individual differences according to claim 2, characterized in that: The developing a nomogram according to the individual difference sample data and establishing a brain structure individual difference model according to the nomogram specifically includes: Based on the genetic SNPs, a polygenic risk score is obtained by an allele count weighting method, and GWAS results related to AD are obtained in the whole human genome; Based on the t-test method, multiple AD-related ROI features are identified on the ADNI data set, and a first nomogram of the multiple AD-related ROI features and the GWAS results is established according to the polygenic risk score; The association between environmental factors and epigenetic modifications was identified by detecting the correlation between methylation and histone modification sites and phenotypic traits, and EWAS results related to AD were obtained; The modifiable factors were subjected to t-test and Bonferroni correction according to the age and gender to obtain significant factors; According to the EWAS results, a second nomogram of the plurality of AD-related ROI features and the significant factors is established by a multivariate linear regression method, and a brain structure individual difference model is established based on the first nomogram and the second nomogram.

4. The sMRI image analysis method based on individual differences according to claim 3, characterized in that: The target model includes: graph convolutional neural network, Transformer neural network and multi-layer perceptron.

5. The sMRI image analysis method based on individual differences according to claim 1, characterized in that: The training and testing of the sMRI image analysis model specifically includes: Collecting sMRI images and individual data of multiple users from a database, and obtaining AD-related SNPs, and obtaining ROI features and ROI feature atrophy rates of the multiple users according to the sMRI images and the individual data; Taking the ROI features, the ROI feature atrophy rate and the AD-related SNPs as sample data, taking the sMRI image classification results corresponding to the plurality of users as labels, and establishing a data set according to the sample data and the labels; The data set is divided into a training set and a test set according to a preset ratio, the training set is used to train the sMRI image analysis model to obtain a trained model, and the test set is used to perform performance evaluation on the trained model to obtain the target model that meets the requirements.

6. The sMRI image analysis method based on individual differences according to claim 4, characterized in that: The step of obtaining a target ROI feature according to the target sMRI image, inputting the target individual data into the brain structure individual difference model to obtain a normal ROI feature, and calculating the ROI feature atrophy rate according to the target ROI feature and the normal ROI feature specifically includes: The target sMRI image is processed using the FreeSurfer software package to obtain gray matter volume, white matter volume, cerebral cortex area and cerebral cortex thickness, and multiple target ROI features are obtained according to the gray matter volume, the white matter volume, the cerebral cortex area and the cerebral cortex thickness; Inputting the target individual data into the plurality of brain structure individual difference models for calculation, and outputting a plurality of normal ROI features, wherein the target individual data includes gender, age, genetic SNPs and modifiable factors; The ROI feature shrinkage rate of each target ROI feature is calculated according to the multiple target ROI features and the multiple normal ROI features.

7. The sMRI image analysis method based on individual differences according to claim 6, characterized in that: The step of inputting the target ROI features, the ROI feature atrophy rate and the AD-related SNPs into the target model for analysis to obtain the classification result of the target sMRI image specifically includes: Inputting the target ROI feature, the ROI feature atrophy rate, and the AD-related SNPs into the graph convolutional neural network for feature extraction to obtain three intermediate features, and connecting the three intermediate features in series to obtain multi-omics features; The multi-omics features are input into the Transformer neural network for correlation analysis to obtain target features, the target features are weighted summed through the multi-layer perceptron, and the result of the weighted summation is output using an activation function to obtain the classification result of the target sMRI image.

8. A sMRI image analysis system based on individual differences, characterized in that: The sMRI image analysis system based on individual differences comprises: An individual difference model building module is used to obtain individual difference sample data, develop a nomogram based on the individual difference sample data, and establish an individual difference model of brain structure based on the nomogram; An image analysis model building module is used to build an sMRI image analysis model, train and test the sMRI image analysis model, and obtain a target model; A data processing module, used to obtain a target sMRI image, target individual data and AD-related SNPs of a user to be analyzed, obtain a target ROI feature according to the target sMRI image, input the target individual data into the brain structure individual difference model to obtain a normal ROI feature, and calculate a ROI feature atrophy rate according to the target ROI feature and the normal ROI feature; The model application module is used to input the target ROI features, the ROI feature atrophy rate and the AD-related SNPs into the target model for analysis, obtain the classification result of the target sMRI image, and generate an sMRI image analysis report based on individual differences according to the classification result.

9. A terminal, characterized in that: The terminal includes: a memory, a processor, and an sMRI image analysis program based on individual differences stored in the memory and executable on the processor. When the sMRI image analysis program based on individual differences is executed by the processor, the steps of the sMRI image analysis method based on individual differences as described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores an sMRI image analysis program based on individual differences, and when the sMRI image analysis program based on individual differences is executed by a processor, the steps of the sMRI image analysis method based on individual differences as described in any one of claims 1 to 7 are implemented.

Images