Graph convolution method and device for enhancing rs-fMRI feature analysis specificity ROI

Abstract

The invention provides an image convolution method and device for enhancing rs-fMRI feature analysis specificity ROI, and relates to the technical field of artificial intelligence and medical image.The method comprises the steps that resting state magnetic resonance imaging data are acquired; extracting a blood oxygen level dependent signal of each region of interest in the resting state magnetic resonance imaging data; performing chaotic processing on the blood oxygen level dependent signal to obtain a blood oxygen level dependent chaotic signal; dynamically fusing the blood oxygen level dependent signal and the blood oxygen level dependent chaotic signal to obtain a fused signal matrix; obtaining a sample sequence of the fusion signal matrix based on a preset time window, and constructing an adjacent matrix based on the sample sequence; inputting data corresponding to the adjacent matrix into the neural network model, and outputting a calculation result; and obtaining a specific region-of-interest pair based on a calculation result. The defects of the existing method for extracting potential brain features are overcome, so that the classification accuracy is improved, and the specific ROI of the neurological system disease is positioned.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence and medical imaging technology, and in particular to a graph convolution method, device, electronic device, and computer-readable storage medium for enhancing specific ROIs for rs-fMRI feature analysis. Background Art

[0002] Neurological disorders encompass a wide range of conditions, such as autism spectrum disorder and cognitive impairment, and are caused by a combination of environmental, genetic, and psychological factors. Identifying specific biomarkers for these disorders can greatly aid in targeted research. The lack of accurate pathological identification as a diagnostic reference hinders targeted diagnosis.

[0003] Resting-state functional magnetic resonance imaging (rs-fMRI), a non-invasive technique that measures spontaneous brain activity via the blood-oxygen-level-dependent (BOLD) signal, is increasingly being used to study neurological disorders. However, current research remains limited in accurately identifying specific regions of interest (ROIs) for various psychiatric disorders.

[0004] The ability of graph neural networks to diagnose neurological diseases by constructing graph nodes for each ROI is increasing. However, existing studies have ignored the fact that the BOLD signals of each ROI may be interfered with by surrounding factors during the acquisition process, resulting in deviations in the collected BOLD signals and inaccurate subsequent ROI analysis.

[0005] The deep learning technology currently used for feature extraction is not sufficient. Although many existing methods have achieved a high diagnostic accuracy, the retention and enhancement of original signal features still needs to be improved. Summary of the Invention

[0006] In view of this, the present invention provides a graph convolution method, apparatus, device and computer-readable storage medium for enhancing rs-fMRI feature analysis-specific ROI, which can solve the problems that the current deep learning technology used is not sufficient for feature extraction and the preservation and enhancement of original signal features are not perfect.

[0007] Some embodiments of the present invention provide a graph convolution method for enhancing specific ROIs for rs-fMRI feature analysis. The present invention is described below from multiple aspects, and the embodiments and beneficial effects of the following aspects can be referenced to each other.

[0008] In a first aspect, the present invention provides a graph convolution method for enhancing rs-fMRI feature analysis-specific ROI, comprising:

[0009] Acquiring resting-state magnetic resonance imaging data;

[0010] Extracting blood oxygen level-dependent signals from various regions of interest in resting-state magnetic resonance imaging data;

[0011] Performing chaotic processing on the blood oxygen level dependent signal based on the chaotic model to obtain a blood oxygen level dependent chaotic signal;

[0012] Dynamically fuse the blood oxygen level dependent signal with the blood oxygen level dependent chaotic signal to obtain a fused signal matrix;

[0013] Acquiring a sample sequence of the fusion signal matrix based on a preset time window, and constructing an adjacency matrix based on the sample sequence, wherein the adjacency matrix is used to characterize the time transformation relationship of the blood oxygen level dependent signal in the region of interest;

[0014] Input the data corresponding to the adjacency matrix into the neural network model and output the calculation results;

[0015] Based on the calculation results, specific regions of interest are obtained.

[0016] In a possible implementation of the first aspect, extracting the blood oxygen level dependent signal of each region of interest in the resting-state magnetic resonance imaging data includes:

[0017] performing data correction based on resting-state magnetic resonance imaging data to obtain corrected data;

[0018] Filling the corrected data into the automatic anatomical labeling template to obtain filled data;

[0019] The filled data are extracted based on the automatic anatomical labeling template to obtain the blood oxygen level dependent signals of each region of interest.

[0020] In a possible implementation of the first aspect, performing chaotic processing on the blood oxygen level dependent signal based on a chaotic model to obtain the blood oxygen level dependent chaotic signal includes:

[0021] The blood oxygen level-dependent signal of the target area with the highest functional connectivity in the region of interest is selected as the data input of the initial chaotic model;

[0022] The chaotic model is obtained by dynamically modulating the periodic control parameters of the initial chaotic model based on the blood oxygen level dependency signal of the target area;

[0023] The differential equation of the chaotic model is solved to obtain the blood oxygen level dependent chaotic signal.

[0024] In a possible implementation of the first aspect, the blood oxygen level dependent signal and the blood oxygen level dependent chaotic signal are dynamically fused to obtain a fused signal matrix, including:

[0025] Normalize the blood oxygen level dependence chaotic signal;

[0026] Select a component of the blood oxygen level dependent chaotic signal;

[0027] The components are dynamically fused with the blood oxygen level dependent signal to obtain a fused signal matrix.

[0028] In a possible implementation of the first aspect, inputting data corresponding to the adjacency matrix into a neural network model and outputting a calculation result includes:

[0029] Perform local feature aggregation based on the data corresponding to the adjacency matrix to obtain local features;

[0030] Perform global feature enhancement based on local features to obtain enhanced features;

[0031] Feature extraction is performed based on enhanced features to obtain multi-scale features;

[0032] The calculation results are obtained based on multi-scale features.

[0033] In a possible implementation of the first aspect, obtaining a specific region of interest pair based on the calculation result includes:

[0034] Outputting an adjacency matrix of abnormal resting-state magnetic resonance imaging data based on the calculation results;

[0035] The adjacency matrix of abnormal resting-state MRI data was compared with that of the control group to obtain specific region-of-interest pairs.

[0036] In a possible implementation of the first aspect, after extracting the blood oxygen level dependent signal of each region of interest in the resting-state magnetic resonance imaging data, the method further includes:

[0037] The blood oxygen level-dependent signals of each region of interest were normalized.

[0038] In a second aspect, the present invention provides a graph convolution device for enhancing rs-fMRI feature analysis-specific ROI, comprising:

[0039] an acquisition module, for acquiring resting-state magnetic resonance imaging data;

[0040] an extraction module for extracting blood oxygen level dependent signals of various regions of interest in resting-state magnetic resonance imaging data;

[0041] a chaos processing module, configured to perform chaos processing on the blood oxygen level dependent signal based on a chaos model to obtain a blood oxygen level dependent chaotic signal;

[0042] A fusion module is used to dynamically fuse the blood oxygen level dependent signal with the blood oxygen level dependent chaotic signal to obtain a fused signal matrix;

[0043] A construction module is used to obtain a sample sequence of the fusion signal matrix based on a preset time window, and to construct an adjacency matrix based on the sample sequence, wherein the adjacency matrix is used to characterize the time transformation relationship of the blood oxygen level dependent signal in the region of interest;

[0044] The classification module is used to input the data corresponding to the adjacency matrix into the neural network model and output the calculation results;

[0045] The analysis module is used to obtain specific regions of interest based on the calculation results.

[0046] In a third aspect, the present invention provides an electronic device, comprising: a processor; and a memory, wherein computer program instructions are stored in the memory.

[0047] When the computer program instructions are executed by the processor, the processor executes a graph convolution method based on chaotic system enhancement of rs-fMRI feature analysis specific ROI.

[0048] In a fourth aspect, the present invention provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, the processor executes a graph convolution method based on chaotic system enhanced rs-fMRI feature analysis specific ROI.

[0049] In a fifth aspect, the present invention discloses a device comprising:

[0050] The memory is used to store instructions executed by one or more processors of the device, and the processor is one of the processors of the device, which is used to execute the method disclosed in any one of the first to fourth aspects above.

[0051] The present invention can effectively fuse the blood oxygen level dependent chaotic signal with the blood oxygen level dependent signal of each region of interest, thereby reducing the influence of external factors on the blood oxygen level dependent signal, improving the accuracy of neural network calculation, and making the found difference-specific region of interest pairs more accurate. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 Flowchart of a graph convolution method for enhancing rs-fMRI feature analysis specific ROI according to an embodiment of the present invention;

[0053] Figure 2This is a model diagram of a graph convolution method for enhancing rs-fMRI feature analysis specific ROI according to an embodiment of the present invention;

[0054] Figure 3 This is a flowchart of step S120 according to an embodiment of the present invention;

[0055] Figure 4 This is a flowchart of step S130 according to an embodiment of the present invention;

[0056] Figure 5 This is a flowchart of step S140 according to an embodiment of the present invention;

[0057] Figure 6 This is a visualization experiment result diagram of an embodiment of the present invention;

[0058] Figure 7 This is a flowchart of step S160 according to an embodiment of the present invention;

[0059] Figure 8 This is a flowchart of step S170 according to an embodiment of the present invention;

[0060] Figure 9 is a block diagram of a device according to an embodiment of the present invention;

[0061] Figure 10 This is a block diagram of a SoC (System on Chip) according to an embodiment of the present invention. DETAILED DESCRIPTION

[0062] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0063] In order to facilitate the understanding of the technical solution of the present invention, the technical problem to be solved by the present invention is first described.

[0064] Resting-state magnetic resonance imaging (RSMRI), a non-invasive technique that measures spontaneous brain activity through blood oxygenation-dependent signals, is increasingly being used to study neurological disorders. However, current research remains limited in accurately identifying specific regions of interest (ROIs) for various psychiatric disorders.

[0065] As graph neural networks' ability to construct graph nodes for each ROI in the diagnosis of neurological diseases continues to improve, it has brought convenience to identifying specific ROIs for various mental illnesses. However, there are also some technical problems, such as:

[0066] (1) When BOLD signals are collected, they are affected by environmental noise (such as scanner artifacts, physiological noise, and head movement) and neurovascular coupling effects, which can lead to signal distortion and specific ROI identification deviation.

[0067] (2) Traditional graph neural networks rely on static adjacency matrices and do not fully consider the temporal transformation characteristics of the BOLD signal, resulting in insufficient spatiotemporal dynamic modeling of the functional connectivity network between ROIs.

[0068] (3) The original features are not sufficiently preserved, and a single modality (such as only BOLD signal) is difficult to fully characterize the functional characteristics of the ROI.

[0069] In order to solve the above technical problems, the present invention proposes a graph convolution method for enhancing the specific ROI of rs-fMRI feature analysis. This method performs nonlinear dynamic modulation of the BOLD signal through a chaotic model to enhance the signal's noise resistance and feature robustness; dynamically fuses the BOLD signal with the chaotic signal, and uses standardized components to achieve signal complementarity; extracts the sample sequence of the fused signal through a preset time window, constructs an adjacency matrix reflecting the dynamic evolution of the signal, captures the temporal dependence between ROIs, and improves the adaptability of the graph structure to the representation of dynamic brain networks; by comparing the differences in the adjacency matrices between the abnormal group and the control group, combined with the multi-scale feature classification of the neural network, screens out abnormal specific ROI pairs, thereby improving the reliability of biomarker identification.

[0070] The graph convolution method for enhancing rs-fMRI feature analysis specific ROI according to an embodiment of the present invention is described in detail below with reference to the accompanying drawings.

[0071] refer to Figure 1 and Figure 2 , Figure 1 A flowchart of a graph convolution method for enhancing rs-fMRI feature analysis specific ROI according to an embodiment of the present invention is shown. Figure 2 A model diagram of a graph convolution method for enhancing rs-fMRI feature analysis-specific ROI according to an embodiment of the present invention is shown.

[0072] The method includes steps S110 - S170 .

[0073] S110, acquiring resting-state magnetic resonance imaging data.

[0074] In an embodiment of the present invention, resting-state magnetic resonance imaging data of a subject who is awake but avoiding active thinking activities with eyes closed or open, without external tasks or stimulation, can be collected by an MRI scanner with an acquisition time of 5 to 15 minutes.

[0075] like Figure 2As shown, in the implementation process of the present invention, step S110 is equivalent to step A of inputting the rs-fMRI file to obtain resting-state magnetic resonance imaging data.

[0076] S120 , extracting blood oxygen level dependent signals of each region of interest in the resting-state magnetic resonance imaging data.

[0077] like Figure 2 As shown, in the implementation process of the present invention, step S120 is equivalent to step B of aligning the template to divide the ROI and step C of extracting the time series, which is used to align the corrected resting-state MRI data to the automatic anatomical labeling template and extract the time series of each ROI.

[0078] It should be noted that by performing data correction based on resting-state magnetic resonance imaging data, interference factors such as noise and artifacts in the original data can be effectively removed, thereby obtaining more accurate and reliable corrected data.

[0079] The corrected data is filled into the automatic anatomical labeling template, and with the help of the template's standardized structure and positioning information, accurate identification and positioning of various brain regions of interest can be achieved.

[0080] By extracting the filled data based on the automatic anatomical labeling template, the blood oxygen level-dependent signals of each region of interest can be obtained in a targeted manner, improving the efficiency and accuracy of signal extraction, avoiding the errors and subjectivity that may be caused by manual extraction, and providing key data support for in-depth research on brain functional activities.

[0081] S130 , performing chaotic processing on the blood oxygen level dependent signal based on a chaotic model to obtain a blood oxygen level dependent chaotic signal.

[0082] like Figure 2 As shown, in the implementation process of the present invention, step S130 is equivalent to step D in generating a chaotic signal for obtaining a blood oxygen level dependent chaotic signal.

[0083] It should be noted that the blood oxygen level dependent chaotic signal is a high-dimensional chaotic signal reconstructed through a chaotic model based on the nonlinear dynamic characteristics of the blood oxygen level dependent signal in resting-state magnetic resonance imaging data.

[0084] By processing blood oxygen level-dependent signals based on a chaotic model, it is possible to more sensitively capture subtle dynamic changes and complex features in the signal, helping to discover potential information that is difficult to detect using traditional linear analysis methods, thereby gaining a deeper understanding of the complex mechanisms of brain activity.

[0085] By dynamically modulating the periodic control parameters of the initial chaotic model based on the blood oxygen level-dependent signal in the target area, the chaotic model can better adapt to the characteristics and changing laws of the actual signal, and improve the model's fitting accuracy and prediction ability for the blood oxygen level-dependent signal.

[0086] S140 , dynamically fuse the blood oxygen level dependent signal with the blood oxygen level dependent chaotic signal to obtain a fused signal matrix.

[0087] like Figure 2 As shown, in the implementation process of the present invention, step S140 is equivalent to step D of generating a chaotic signal and diffusing it to the whole brain, and step E of dynamically fusing it with the standardized chaotic signal, which is used to perform whole-brain mapping on the nonlinear threshold component in the blood oxygen level-dependent chaotic signal to obtain a mapping signal, and dynamically fuse the mapping signal with the standardized chaotic signal.

[0088] It should be noted that the dynamic fusion of the blood oxygen level-dependent signal and the blood oxygen level-dependent chaotic signal allows for more precise integration of signal features across different dimensions, enabling the fused signal matrix to retain both the physiological information of the original signal and the dynamic characteristics of the chaotic signal, providing a more comprehensive and richer information foundation for subsequent analysis. This dynamic fusion approach can reflect the interactions and dynamic changes between the blood oxygen level-dependent and chaotic signals in real time, capturing the dynamic characteristics of brain activity under different physiological states.

[0089] S150 , acquiring a sample sequence of the fusion signal matrix based on a preset time window, and constructing an adjacency matrix based on the sample sequence.

[0090] Specifically, the size of the preset time window is equal to the size of the blood oxygen level dependent signal, and the adjacency matrix expression is:

[0091] A=corrcoef(Q)

[0092] Among them, A is the adjacency matrix and Q is the sample sequence.

[0093] like Figure 2 As shown, in the implementation process of the present invention, step S150 is equivalent to step F. The time window processes each ROI in turn to obtain a sample sequence of the fusion signal matrix and construct an adjacency matrix based on the sample sequence.

[0094] It should be noted that the adjacency matrix is a square matrix, that is, the number of rows is equal to the number of columns, and the element A(i, j) in the matrix represents the correlation between the i-th sample and the j-th sample in the sample sequence.

[0095] By obtaining the sample sequence of the fusion signal matrix based on a preset time window, the dynamic changes of the fusion signal within a specific time range can be captured, which helps to analyze the time-varying characteristics and short-term behavior of the signal.

[0096] By constructing an adjacency matrix based on the sample sequence, we can quantify the similarity or correlation between signals in different time windows, thereby constructing the adjacency matrix of the signal, providing a basis for studying the complex interactions of signals.

[0097] Since the size of the preset time window is equal to the size of the blood oxygen level dependent signal, it can ensure that the key information of the original signal is not lost during the analysis process, while avoiding signal distortion or information omission caused by a time window that is too large or too small.

[0098] S160, inputting the data corresponding to the adjacency matrix into the neural network model and outputting the calculation result.

[0099] like Figure 2 As shown, in the implementation process of the present invention, step S160 is equivalent to step G to generate feature F, which is used to determine whether the input rs-fMRI image is an abnormal image through a neural network based on the adjacency matrix.

[0100] It should be noted that the neural network models in the above steps include but are not limited to graph convolutional neural network models, graph attention network models, and graph sampling and aggregation models.

[0101] Preferably, the data corresponding to the adjacency matrix is input into the graph convolutional neural network model, and the calculation results are output.

[0102] It should be noted that the graph convolutional neural network model includes local feature aggregation, global feature enhancement and multi-scale feature extraction steps. Through the gradual processing of local feature aggregation, global feature enhancement and multi-scale feature extraction, it can deeply refine and enhance the blood oxygen level dependent signal features layer by layer, making the final multi-scale features richer and more representative.

[0103] Global feature enhancement is performed based on local feature aggregation, which not only retains the local detail information of the blood oxygen level-dependent signal, but also incorporates global context information, making the features more complete and accurate, and helping to comprehensively characterize the complex characteristics of the signal.

[0104] Through multi-scale feature extraction, the characteristics of the signal can be fully captured from different scales, avoiding the information loss caused by single-scale analysis, improving the ability to recognize complex patterns of blood oxygen level-dependent signals, and enhancing the reliability of calculation results.

[0105] Through deep processing of adjacency matrix data using neural networks, it is possible to explore the complex relationships and deep features hidden in the blood oxygen level dependency signal, improve the depth and accuracy of blood oxygen level dependency signal analysis, and provide more valuable information for subsequent research and applications.

[0106] Inputting adjacency matrix data into a neural network model for automated feature extraction and analysis can reduce manual intervention, improve processing efficiency, and quickly find abnormal rs-fMRI image data from a large amount of rs-fMRI image data.

[0107] S170: Obtain a specific region of interest pair based on the calculation result.

[0108] like Figure 2 As shown, in the implementation process of the present invention, step S170 is equivalent to step H. If the patient is ill, a specific ROI is analyzed to find a differential ROI pair between the abnormal rs-fMRI image data and the control group image.

[0109] It should be noted that by comparing the adjacency matrix of abnormal resting-state MRI data with the adjacency matrix of the control group, specific region-of-interest pairs in the abnormal data can be accurately identified.

[0110] Next, Figure 1 Steps S120-S140, S160-S170 are further described in detail.

[0111] refer to Figure 3 , Figure 3 The flowchart of step S120 in the embodiment of the present invention is shown. Step S120 includes:

[0112] S121 , performing data correction based on the resting-state magnetic resonance imaging data to obtain corrected data.

[0113] It should be noted that the data content of the calibration data includes image data and header information. The image data is a three-dimensional or four-dimensional array used to represent the signal intensity of each voxel. The header information includes the affine transformation matrix, voxel size, slice order, and spatial direction. The file format is:

[0114] subject_01_resting_state_corrected.nii.gz, commonly used data correction methods include rigid body transformation, similarity transformation, nonlinear deformation, and affine transformation.

[0115] Preferably, the resting-state magnetic resonance imaging data is corrected by affine transformation to obtain corrected data.

[0116] It should be noted that the advantages of choosing affine transformation for data correction are:

[0117] (1) Affine transformation (12 degrees of freedom) can simultaneously model translation, rotation, scaling, and shearing. Therefore, compared with 6-degree-of-freedom rigid transformation or 7-degree-of-freedom similarity transformation, it has greater flexibility and can more accurately align brain structures with proportional differences or slight deformations between different subjects.

[0118] (2) Compared with nonlinear transformations (such as the high-dimensional deformation model provided by FNIRT), affine transformations are more efficient, have a moderate number of parameters, and reduce the risk of overfitting.

[0119] (3) Affine transformation can ensure the coherence of local structure during spatial changes without introducing large-scale non-physiological distortions. Therefore, it is more robust and reliable for subsequent macro-region-based processing tasks such as ROI time series extraction and brain network analysis.

[0120] S122, filling the corrected data into the automatic anatomical labeling template to obtain filled data.

[0121] It should be noted that the Automated Anatomical Labeling (AAL) template is a segmentation tool based on anatomical criteria that is used to divide the brain into multiple predefined regions of interest (ROIs). Its core purpose is to standardize brain region parcellation, facilitating functional connectivity analysis and data comparability across studies. Subsequent versions include the Automated Anatomical Labeling 2.0 template (AAL2) and the Automated Anatomical Labeling 3.0 template (AAL3).

[0122] Preferably, the correction data is aligned with AAL3 by a FLIRT tool to obtain aligned data, and the aligned data is resampled to AAL3 by a linear interpolation method to obtain filled data.

[0123] It should be noted that the advantage of choosing AAL3 is that the high-precision ROI division of AAL3 reduces signal aliasing, ensures the BOLD signal purity of the target area with the highest functional connectivity, and provides reliable input for the dynamic modulation of the hybrid model.

[0124] S123 , extracting the filled data based on the automatic anatomical labeling template to obtain a blood oxygen level dependent signal of each region of interest.

[0125] Preferably, the filling data is extracted by using the NiftiLabelsMasker tool based on AAL3 to obtain the blood oxygen level dependent signal of each region of interest.

[0126] It should be noted that the NiftiLabelsMasker tool in the Nilearn library was used to extract signals from functional images based on AAL3 brain region zoning. Compared with other common signal extraction methods (such as NiftiMasker, NiftiMapsMasker, or direct use of ROI averaging), NiftiLabelsMasker is particularly suitable for regional division based on discrete label maps (such as AAL3). It can efficiently average or otherwise aggregate the voxel signals within each brain region to ensure that the extracted results strictly correspond to the anatomical structure. At the same time, the tool has built-in preprocessing functions such as standardization (such as z-score), detrending, and smoothing, which facilitates the simultaneous completion of data normalization during the extraction process, thereby improving the accuracy and consistency of downstream analysis.

[0127] refer to Figure 4 , Figure 4 The flowchart of step S130 of the embodiment of the present invention is shown. Step S130 includes:

[0128] S131, selecting the blood oxygen level dependent signal of the target area with the highest functional connectivity in the region of interest as the data input of the initial chaotic model.

[0129] It should be noted that the target areas with the highest functional connectivity among the regions of interest include: precentral gyrus, insular cortex, dorsolateral prefrontal cortex, etc. The chaos model includes: Chaos model, Lorenz model, Duffing model, etc.

[0130] Preferably, the blood oxygen level dependent signal of the precentral gyrus is selected as the initial The data input of the chaos model, where the initial The differential equation of the chaos model is expressed as follows:

[0131]

[0132] Where x represents the fast oscillation component of the blood oxygen level-dependent chaotic signal, y represents the slow phase component of the blood oxygen level-dependent chaotic signal, z represents the nonlinear threshold component of the blood oxygen level-dependent chaotic signal, a represents the nonlinear control parameter, b represents the dissipation parameter, c represents the periodic control parameter, and dt represents the time step.

[0133] It should be noted that the advantages of selecting the blood oxygen level-dependent signal of the precentral gyrus as input are:

[0134] (1) The node degree of the precentral gyrus in the resting-state functional network ranks in the top 5% of the whole brain, indicating that the precentral gyrus is widely connected with multiple functional blocks and is suitable as the global dynamic feature input of the chaos model.

[0135] (2) The betweenness centrality value of the precentral gyrus ranks first in the sensorimotor network, indicating that the precentral gyrus plays a key relay role in cross-network transmission and can reflect the dynamic integration characteristics of the whole brain.

[0136] (3) The sample entropy of the BOLD signal in the precentral gyrus ranges from 0.25 to 0.35, which is in the optimal input range of the chaos model (0.2-0.5), which can avoid both excessive randomness and excessive regularization.

[0137] (4) The maximum Lyapunov exponent of the BOLD signal in the precentral gyrus (λ_max=0.05±0.01) was significantly positive (P<0.01), indicating that it has deterministic chaotic characteristics, while the chaotic characteristics of the limbic system (such as the amygdala λ_max=0.02±0.01) are weaker.

[0138] choose The advantages of chaos models are:

[0139] (1) The parameters of the chaotic model are dynamically adjusted to adapt to the dynamic differences of BOLD signals of different individuals.

[0140] (2) The topological structure of the chaotic model is sensitive to initial conditions, but has structural stability and is suitable for long-term signal analysis.

[0141] (3) Embedding BOLD signal The phase space of the chaotic model can preserve the key dynamic characteristics of the BOLD signal and avoid information loss.

[0142] (4) Pass The chaotic model performs a linear transformation on the BOLD signal, which can amplify pathology-specific features.

[0143] S132 , dynamically modulating the periodic control parameters of the initial chaotic model based on the blood oxygen level dependency signal of the target area to obtain a chaotic model.

[0144] Specifically, the blood oxygen level-dependent signal based on the precentral gyrus is used to determine the initial The periodic control parameter of the chaotic model is dynamically modulated. At this time, the periodic control parameter expression is as follows:

[0145] c=C base +k*S BOLD

[0146] Among them, c base is the center value of the chaotic interval, k is the scaling factor, S BOLD It is the blood oxygen level dependent signal in the precentral gyrus area.

[0147] S133, solving the differential equation of the chaotic model to obtain a blood oxygen level dependent chaotic signal.

[0148] Specifically, the differential equation based on the chaotic model is discretized by Euler to obtain the blood oxygen level dependent chaotic signal. At this time, the expression of the blood oxygen level dependent chaotic signal is as follows:

[0149]

[0150] Among them, n is a variable parameter.

[0151] In the above steps, The nonlinear control parameter a of the chaotic model is set to 0.2, the dissipation parameter b is set to 0.4, the period control parameter c is set to 1.0, the time step dt is set to 0.07, the initial state x0 of the fast oscillation component of the blood oxygen level dependent chaotic signal is set to 0.1, the initial state y0 of the slow phase component of the blood oxygen level dependent chaotic signal is set to 0.1, and the initial state z0 of the nonlinear threshold component of the blood oxygen level dependent chaotic signal is set to 0.1.

[0152] It should be noted that experiments have shown that when the initial states of the fast oscillation component, the slow phase component, and the nonlinear threshold component are set to 0.1, the nonlinear threshold component can closely follow the changes in the blood oxygen level-dependent signal. Because the chaos model is sensitive to these initial states, if the initial states of the fast oscillation component, the slow phase component, and the nonlinear threshold component are set to 0.2, the signal will exhibit significant fluctuations. However, when the initial states of the fast oscillation component, the slow phase component, and the nonlinear threshold component are set to 0.01, the changes are less noticeable. After debugging, 0.1 was determined to be the most suitable setting.

[0153] Setting the nonlinear control parameter a to 0.2 can keep the system in a chaotic state while ensuring the numerical stability of the system. Setting the dissipation parameter b to 0.4 can ensure that the system has a certain amount of energy dissipation, which helps to form a stable chaotic attractor and avoid divergence. Setting the periodic control parameter c to 1.0 ensures that the system has complex dynamic behavior and can maintain long-term chaotic oscillations. When c is low, the system may tend to periodic oscillations, and when it is high, it will lead to excessive divergence. Setting the time step dt to 0.07 is a value that strikes a balance between ensuring numerical stability and computational efficiency. Too large a step may lead to numerical instability or loss of chaotic characteristics, while too small a step will increase the computational overhead. This value has been verified by experiments to effectively capture Complex dynamic behavior of the system.

[0154] refer to Figure 5 , Figure 5The flowchart of step S140 in an embodiment of the present invention is shown. Step S140 includes:

[0155] S141, normalizing the blood oxygen level dependence chaotic signal.

[0156] It should be noted that the blood oxygen level-dependent signal is normalized to eliminate dimensional differences, enhance data comparability, and optimize the model convergence speed. Normalization methods include but are not limited to Min-Max normalization, decimal scaling normalization, Robus normalization, vector normalization, and zero-mean unit variance normalization.

[0157] Preferably, the blood oxygen level-dependent chaotic signal is normalized by a zero-mean unit variance normalization algorithm.

[0158] It should be noted that the advantages of zero-mean unit variance standardization are:

[0159] (1) Zero-mean variance normalization only translates and scales the data, does not change the distribution form of the blood oxygen level-dependent chaotic signal, and can retain the nonlinear dynamic characteristics of the blood oxygen level-dependent chaotic signal.

[0160] (2) Zero-mean variance normalization is sensitive to extreme outliers and can help detect residual noise.

[0161] In the above steps, the normalized blood oxygen level dependent chaotic signal is expressed as:

[0162]

[0163] Among them, s ij is the normalized blood oxygen level dependent chaotic signal, s ij is the blood oxygen level dependent chaotic signal, μ j is the mean of the nonlinear threshold component, σ j is the standard deviation of the nonlinear threshold component, ∈=10-8, used to prevent the denominator from being zero.

[0164] S142, selecting a component of the blood oxygen level dependent chaotic signal.

[0165] It should be noted that the blood oxygen level-dependent chaotic signal includes a fast oscillation component, a slow phase vector, and a nonlinear threshold component.

[0166] Preferably, the blood oxygen level is selected to depend on the nonlinear threshold component of the chaotic signal.

[0167] It should be noted that the reference Figure 6, shown is the visualization experiment result diagram, where x is the fast oscillation component, y is the slow phase component, and z is the nonlinear threshold component. The nonlinear threshold component of the blood oxygen level dependent chaotic signal is selected to be mapped to all regions of interest because in the visualization experiment, the nonlinear threshold component changes relatively smoothly and contains slow drift and mutation jumps, which is closer to the blood oxygen level dependent signal.

[0168] The advantages of choosing a nonlinear threshold component are:

[0169] (1) The nonlinear threshold component can capture the nonlinear changes in the BOLD signal, which are closely related to the complex physiological mechanisms of brain activity.

[0170] (2) By fusing with BOLD signals, we can better understand the complex dynamic behavior of the brain in different physiological states.

[0171] (3) The nonlinear threshold component can help remove noise from the signal and enhance the stability of the BOLD signal.

[0172] (4) By fusing the nonlinear threshold component, the overall quality of the BOLD signal can be improved, making it more reliable during analysis.

[0173] S143: Map the nonlinear threshold component to all regions of interest to obtain mapping signals.

[0174] It should be noted that the mapping signal is a nonlinear threshold characteristic data containing each ROI, usually a three-dimensional or four-dimensional array, and each element in the mapping signal represents the nonlinear threshold characteristic value of the corresponding region of interest.

[0175] In the above steps, the mapping signal expression is:

[0176] N roi =tile(N z ,(1,M))

[0177] Among them, N roi is the mapping signal, N z is the nonlinear threshold component, and M is the number of regions of interest.

[0178] S144: Perform signal strength adjustment based on the mapped signal through the scaling model to obtain an adjusted signal.

[0179] It should be noted that the adjustment signal is data with the same structure as the mapping signal, but the adjustment signal strength has been scaled according to certain rules.

[0180] The expression of the regulation signal is:

[0181] N'=λ*N roi

[0182] Among them, N' is the adjustment signal and λ is the intensity control value.

[0183] S145 , obtaining an optimal weight, and dynamically fusing the regulation signal and the blood oxygen level dependent signal based on the optimal weight to obtain a fused signal matrix.

[0184] It should be noted that the fusion signal matrix expression is:

[0185] fusion=ω * ·N′+(1-ω * )·real_signal

[0186] Among them, fusion is the fusion signal matrix, ω* is the optimal weight, and real_signal is the blood oxygen level dependent signal.

[0187] It should be noted that the optimal weight needs to be calculated through the weight optimization objective function, and the weight optimization objective function expression is:

[0188]

[0189] Among them, ω is the test weight and temp_fusion is the temporary fusion signal.

[0190] refer to Figure 7 , Figure 7 The flowchart of step S160 of the embodiment of the present invention is shown. Step S160 includes:

[0191] S161, local feature aggregation is performed based on the data corresponding to the adjacency matrix to obtain local features. It should be noted that the local feature expression is:

[0192] Z (1) =ReLU(A·(Q·W1+b1))

[0193] Among them, Z (1) is a local feature, W1∈R T*H is the learning weight, b1 is the bias term, T is the blood oxygen level dependent signal length, and H is the hidden layer dimension.

[0194] S162: Perform global feature enhancement based on the local features to obtain enhanced features.

[0195] It should be noted that the enhanced feature expression is:

[0196] Z (2) =LayerNorm((ReLU(W2Z (1) +b2)))+Z (1)

[0197] Among them, Z (2) To enhance the features, W2∈R H*H is the global enhancement weight, and b2 is the bias term.

[0198] S163: Perform feature extraction based on the enhanced features to obtain multi-scale features.

[0199] It should be noted that the multi-scale feature expression is:

[0200] F=MaxPool(Z (2) )

[0201] Among them, F is a multi-scale feature.

[0202] S164: Flatten the multi-scale features to obtain flattened features.

[0203] It should be noted that the flattened feature expression is:

[0204] F flat =Flatten(F)

[0205] Among them, F flat Flatten feature.

[0206] S165: Calculate the result through the fully connected layer based on the flattened features.

[0207] It should be noted that the calculation result expression is:

[0208] Y=softmax(W f ·F flat +b f )

[0209] in, is the learnable weight, b f ∈R C is the bias term, C is the number of categories of the calculation results, and N is the number of samples in the sample sequence.

[0210] refer to Figure 8 , Figure 8 The flowchart of step S170 of the embodiment of the present invention is shown. Step S170 includes:

[0211] S171, outputting an adjacency matrix of abnormal resting-state magnetic resonance imaging data based on the calculation results;

[0212] Specifically, a comparison threshold is preset, and it is determined whether the calculation result is greater than the comparison threshold. If so, the resting-state magnetic resonance imaging data is determined to be abnormal resting-state magnetic resonance imaging data, and the adjacency matrix of the abnormal resting-state magnetic resonance imaging data is output; otherwise, no processing is performed.

[0213] S172, comparing the adjacency matrix of the abnormal resting-state magnetic resonance imaging data with the adjacency matrix of the control group to obtain a comparison result.

[0214] It should be noted that the comparison result expression is:

[0215] FC Difference =|FC condition1 -FC condition2 |

[0216] Among them, FC Difference is the comparison result, FC condition1 Represents the condition1 adjacency matrix of abnormal resting-state magnetic resonance imaging data, FC condition2 Represents the condition2-th adjacency matrix of the control group, condition1 is the variable parameter, and condition2 is the variable parameter.

[0217] S173, presetting a second comparison threshold, and determining whether the comparison result is greater than the second comparison threshold. If so, outputting the adjacency matrix of the abnormal resting-state MRI data and the adjacency matrix of the control group to obtain a specific region of interest pair; if not, no processing is performed.

[0218] The present invention provides a graph convolutional neural network device for analyzing specific ROIs based on chaotic system-enhanced rs-fMRI features, comprising:

[0219] an acquisition module, for acquiring resting-state magnetic resonance imaging data;

[0220] an extraction module for extracting blood oxygen level dependent signals of various regions of interest in resting-state magnetic resonance imaging data;

[0221] a chaos processing module, configured to perform chaos processing on the blood oxygen level dependent signal based on a chaos model to obtain a blood oxygen level dependent chaotic signal;

[0222] A fusion module is used to dynamically fuse the blood oxygen level dependent signal with the blood oxygen level dependent chaotic signal to obtain a fused signal matrix;

[0223] A construction module is used to obtain a sample sequence of the fusion signal matrix based on a preset time window, and to construct an adjacency matrix based on the sample sequence, wherein the adjacency matrix is used to characterize the time transformation relationship of the blood oxygen level dependent signal in the region of interest;

[0224] The classification module is used to input the data corresponding to the adjacency matrix into the neural network model and output the calculation results;

[0225] The analysis module is used to obtain specific regions of interest based on the calculation results.

[0226] Each module in the device of the present invention executes the method of the above embodiment. Its specific functions and effects can be referred to the description of the above embodiment and will not be repeated here.

[0227] The present invention provides an electronic device, comprising: a processor; and a memory, wherein computer program instructions are stored in the memory.

[0228] When the computer program instructions are executed by the processor, the processor executes a graph convolution method based on chaotic system enhancement of rs-fMRI feature analysis specific ROI.

[0229] The present invention provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, the processor executes a graph convolution method for analyzing specific ROIs based on chaotic system-enhanced rs-fMRI features.

[0230] The present invention discloses a device, comprising:

[0231] The memory is used to store instructions executed by one or more processors of the device, and the processor is one of the processors of the device, which is used to execute the method disclosed in any one of the first to fourth aspects above.

[0232] Now refer to Figure 9 , shown is a block diagram of a device 1200 according to one embodiment of the present invention. The device 1200 may include one or more processors 1201 coupled to a controller hub 1203. For at least one embodiment, the controller hub 1203 communicates with the processors 1201 via a multi-drop bus such as a Front Side Bus (FSB), a point-to-point interface such as a QuickPath Interconnect (QPI), or a similar connection 1206. The processors 1201 execute instructions that control general types of data processing operations. In one embodiment, the controller hub 1203 includes, but is not limited to, a Graphics Memory Controller Hub (GMCH) (not shown) and an Input Output Hub (IOH) (which may be on separate chips) (not shown), wherein the GMCH includes memory and graphics controllers and is coupled to the IOH.

[0233] The device 1200 may also include a coprocessor 1202 and a memory 1204 coupled to the controller hub 1203. Alternatively, one or both of the memory and the GMCH may be integrated within the processor (as described in the present invention), with the memory 1204 and the coprocessor 1202 directly coupled to the processor 1201 and the controller hub 1203, with the controller hub 1203 and the IOH being in a single chip. The memory 1204 may be, for example, a dynamic random access memory (DRAM), a phase change memory (PCM), or a combination of the two. In one embodiment, the coprocessor 1202 is a special-purpose processor, such as, for example, a high-throughput MIC processor (Many Integerated Core, MIC), a network or communication processor, a compression engine, a graphics processor, a general-purpose graphics processor (GPGPU), or an embedded processor, etc. The optional nature of the coprocessor 1202 is indicated by a dotted line in Figure 8 middle.

[0234] The memory 1204, as a computer-readable storage medium, may include one or more tangible, non-transitory computer-readable media for storing data and / or instructions. For example, the memory 1204 may include any suitable non-volatile memory such as flash memory and / or any suitable non-volatile storage device, such as one or more hard disk drives (HDD(s)), one or more compact disc (CD) drives, and / or one or more digital versatile disc (DVD) drives.

[0235] In one embodiment, the device 1200 may further include a network interface controller (NIC) 1206. The network interface 1206 may include a transceiver for providing a radio interface for the device 1200, thereby communicating with any other suitable device (such as a front-end module, an antenna, etc.). In various embodiments, the network interface 1206 may be integrated with other components of the device 1200. The network interface 1206 may implement the functions of the communication unit in the above-mentioned embodiments.

[0236] Device 1200 may further include input / output (I / O) devices 1205. I / O 1205 may include: a user interface designed to enable a user to interact with device 1200; a peripheral component interface designed to enable peripheral components to interact with device 1200; and / or sensors designed to determine environmental conditions and / or location information related to device 1200.

[0237] It is worth noting that Figure 9 This is for illustrative purposes only. Figure 9 The device 1200 is shown to include multiple components such as a processor 1201, a controller hub 1203, and a memory 1204. However, in actual applications, the device using the methods of the present invention may only include a part of the components of the device 1200, for example, it may only include the processor 1201 and the NIC 1206. Figure 9 The properties of the optional components are shown by dashed lines. According to some embodiments of the present invention, memory 1204, a computer-readable storage medium, stores instructions that, when executed on a computer, cause system 1200 to perform the graph convolution method for analyzing specific ROIs based on chaotic system-enhanced rs-fMRI features according to the above-described embodiments. For details, reference may be made to the methods of the above-described embodiments, and no further description is given here.

[0238] Now refer to Figure 10 , which is a block diagram of a SoC (System on Chip) 1300 according to an embodiment of the present invention. Figure 10 In FIG, similar components have the same reference numerals. In addition, the dashed boxes are optional features of more advanced SoCs. Figure 10 In the embodiment, SoC 1300 includes: an interconnect unit 1350 coupled to an application processor 1310; a system agent unit 1380; a bus controller unit 1390; an integrated memory controller unit 1340; a set of one or more coprocessors 1320, which may include integrated graphics logic, an image processor, an audio processor, and a video processor; a static random access memory (SRAM) unit 1330; and a direct memory access (DMA) unit 1360. In one embodiment, coprocessors 1320 include specialized processors, such as, for example, a network or communication processor, a compression engine, a GPGPU, a high-throughput MIC processor, or an embedded processor.

[0239] The static random access memory (SRAM) unit 1330 may include one or more computer-readable media for storing data and / or instructions. The computer-readable storage medium may store instructions, specifically, temporary and permanent copies of the instructions. The instructions may include: when executed by at least one unit in the processor, causing the Soc 1300 to perform the graph convolution method based on chaotic system-enhanced rs-fMRI feature analysis of specific ROIs according to the above-mentioned embodiment. For details, please refer to the method of the above-mentioned embodiment and will not be repeated here.

[0240] The various embodiments of the mechanisms disclosed in the present invention can be implemented in hardware, software, firmware, or a combination of these implementation methods. The embodiments of the present invention can be implemented as a computer program or program code executed on a programmable system, which includes at least one processor, a storage system (including volatile and non-volatile memory and / or storage elements), at least one input device, and at least one output device.

[0241] Program code can be applied to input instructions to perform the functions described herein and generate output information. The output information can be applied to one or more output devices in a known manner. For purposes of the present invention, a processing system includes any system having a processor such as, for example, a digital signal processor (DSP), a microcontroller, an application specific integrated circuit (ASIC), or a microprocessor.

[0242] Program code can be implemented using a high-level programming language or an object-oriented programming language to facilitate processing system communications. Where necessary, program code can also be implemented using assembly language or machine language. In fact, the mechanism described in the present invention is not limited to the scope of any particular programming language. In either case, the language can be a compiled language or an interpreted language.

[0243] In some cases, the disclosed embodiments can be implemented in hardware, firmware, software or any combination thereof. The disclosed embodiments can also be implemented as instructions carried or stored on one or more temporary or non-temporary machine-readable (e.g., computer-readable) storage media, which can be read and executed by one or more processors. For example, instructions can be distributed over a network or through other computer-readable media. Therefore, machine-readable media can include any mechanism for storing or transmitting information in a machine (e.g., computer) readable form, including but not limited to, floppy disks, optical disks, optical disks, compact disc read-only memories (Compact Disc Read Only Memory, CD-ROMs), magneto-optical disks, read-only memories (ROM), random access memories (RAM), erasable programmable read-only memories (EPROM), electrically erasable programmable read-only memories (EEPROM), magnetic cards or optical cards, flash memory or tangible machine-readable memories for transmitting information (e.g., carrier waves, infrared signals, digital signals, etc.) using the Internet in electrical, optical, acoustic or other forms of propagation signals. Accordingly, machine-readable media includes any type of machine-readable media suitable for storing or transmitting electronic instructions or information in a form readable by a machine (eg, a computer).

[0244] In the accompanying drawings, some structural or method features may be shown in a particular arrangement and / or order. However, it should be understood that such a particular arrangement and / or order may not be required. Rather, in some embodiments, these features may be arranged in a manner and / or order different from that shown in the accompanying drawings. In addition, the inclusion of a structural or method feature in a particular figure does not imply that such a feature is required in all embodiments, and in some embodiments, such features may not be included or may be combined with other features.

[0245] It should be noted that the various units / modules mentioned in the various device embodiments of the present invention are all logical units / modules. Physically, a logical unit / module can be a physical unit / module, or a part of a physical unit / module, or can be implemented as a combination of multiple physical units / modules. The physical implementation of these logical units / modules themselves is not the most important. The combination of functions implemented by these logical units / modules is the key to solving the technical problems raised by the present invention. In addition, in order to highlight the innovative part of the present invention, the above-mentioned device embodiments of the present invention do not introduce units / modules that are not closely related to solving the technical problems raised by the present invention. This does not mean that other units / modules do not exist in the above-mentioned device embodiments.

[0246] It should be noted that in the examples and description of this patent, relational terms such as first and second, etc. are merely used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device that includes 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 device. In the absence of further restrictions, an element defined by the sentence "including one" does not exclude the presence of other identical elements in the process, method, article or device that includes the element.

[0247] While the present invention has been shown and described with reference to certain preferred embodiments thereof, it will be understood by those skilled in the art that various changes in form and details may be made therein without departing from the spirit and scope of the invention.

Claims

1. A graph convolution method for enhancing ROI-specific rs-fMRI feature analysis, characterized in that: include: Acquiring resting-state magnetic resonance imaging data; extracting a blood oxygen level dependent signal from each region of interest in the resting-state magnetic resonance imaging data; performing chaotic processing on the blood oxygen level dependent signal based on a chaotic model to obtain a blood oxygen level dependent chaotic signal; Dynamically fusing the blood oxygen level dependent signal with the blood oxygen level dependent chaotic signal to obtain a fused signal matrix; Acquiring a sample sequence of the fusion signal matrix based on a preset time window, and constructing an adjacency matrix based on the sample sequence, wherein the adjacency matrix is used to characterize a time transformation relationship of the blood oxygen level dependent signal in the region of interest; Inputting the data corresponding to the adjacency matrix into the neural network model and outputting the calculation results; A specific region of interest pair is obtained based on the calculation results.

2. The method according to claim 1, characterized in that The extracting of the blood oxygen level dependent signal of each region of interest in the resting-state magnetic resonance imaging data includes: performing data correction based on the resting-state magnetic resonance imaging data to obtain corrected data; Filling the corrected data into an automatic anatomical labeling template to obtain filled data; The filled data is extracted based on the automatic anatomical labeling template to obtain blood oxygen level dependent signals of the respective regions of interest.

3. The method according to claim 1, characterized in that The performing chaotic processing on the blood oxygen level dependent signal based on the chaotic model to obtain the blood oxygen level dependent chaotic signal includes: Selecting a blood oxygen level-dependent signal of a target region with the highest functional connectivity in the region of interest as data input for an initial chaotic model; Dynamically modulating the periodic control parameters of the initial chaotic model based on the blood oxygen level dependent signal of the target area to obtain the chaotic model; The differential equation of the chaotic model is solved to obtain the blood oxygen level dependent chaotic signal.

4. The method according to claim 1, wherein The dynamically fusing the blood oxygen level dependent signal and the blood oxygen level dependent chaotic signal to obtain a fused signal matrix includes: Normalizing the blood oxygen level dependent chaotic signal; selecting a component of the blood oxygen level dependent chaotic signal; The components are dynamically fused with the blood oxygen level dependent signal to obtain the fused signal matrix.

5. The method according to claim 1, wherein The step of inputting the data corresponding to the adjacency matrix into the neural network model and outputting the calculation result comprises: Performing local feature aggregation based on the data corresponding to the adjacency matrix to obtain local features; Performing global feature enhancement based on the local features to obtain enhanced features; Perform feature extraction based on the enhanced features to obtain multi-scale features; The calculation result is obtained based on the multi-scale features.

6. The method according to claim 1, characterized in that The obtaining of a specific region of interest pair based on the calculation result comprises: outputting an adjacency matrix of abnormal resting-state magnetic resonance imaging data based on the calculation results; The adjacency matrix of the abnormal resting-state magnetic resonance imaging data is compared with the adjacency matrix of the control group to obtain the specific region of interest pair.

7. The method according to claim 1, characterized in that After extracting the blood oxygen level dependent signal of each region of interest in the resting-state magnetic resonance imaging data, the method further includes: The blood oxygen level dependent signals of the respective regions of interest are normalized.

8. A graph convolution device for enhancing ROI-specific rs-fMRI feature analysis, characterized in that: include: an acquisition module, for acquiring resting-state magnetic resonance imaging data; an extraction module, configured to extract a blood oxygen level-dependent signal of each region of interest in the resting-state magnetic resonance imaging data; a chaos processing module, configured to perform chaos processing on the blood oxygen level dependent signal based on a chaos model to obtain a blood oxygen level dependent chaotic signal; a fusion module, configured to dynamically fuse the blood oxygen level dependent signal with the blood oxygen level dependent chaotic signal to obtain a fused signal matrix; a construction module, configured to obtain a sample sequence of the fused signal matrix based on a preset time window, and construct an adjacency matrix based on the sample sequence, wherein the adjacency matrix is used to characterize a time transformation relationship of the blood oxygen level dependent signal within the region of interest; A classification module, used to input the data corresponding to the adjacency matrix into a neural network model and output a calculation result; An analysis module is used to obtain specific region of interest pairs based on the calculation results.

9. An electronic device, characterized in that: include: processor; and a memory having computer program instructions stored therein, Wherein, when the computer program instructions are executed by the processor, the processor is caused to perform the method according to claims 1-7.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the processor is caused to perform the method according to claims 1 to 7.

Images