Brain structure-based brain model construction method and device

Abstract

The present disclosure discloses a model construction method and device, a storage medium and an electronic device, which are used for constructing a brain-like model of a pulse neural network based on biological brain topology constraints, and relate to the technical field of computational neuroscience. The present disclosure solves the problem of lack of biological rationality of the brain-like model of the pulse neural network. The model construction method comprises: dividing brain regions of to-be-processed functional magnetic resonance imaging data to obtain M brain region image data; generating M model nodes based on the M brain region image data; generating N model edges based on a correlation coefficient matrix between the M model nodes; screening the N model edges based on a preset network topology threshold to obtain S model edges meeting a preset condition; generating a topology constraint of a brain-like model based on a biological brain function network based on the M model nodes and the S model edges; and constructing the brain-like model based on the topology constraint. The present disclosure can improve the biological rationality of the brain-like model constructed based on the pulse neural network.

Description

Technical Field

[0001] This disclosure belongs to the field of computational neuroscience technology, specifically relating to a method and apparatus for constructing a brain-like model based on fMRI brain structure. Background Technology

[0002] In recent years, computational neuroscience has developed rapidly, and the application of artificial neural network models has become increasingly widespread. Spiking neural networks, as a new generation of artificial network models, possess powerful processing capabilities for complex nonlinear spatiotemporal information and play a crucial role in the field of computational neuroscience, serving as a necessary theoretical and model foundation. However, with the development of computational neuroscience towards intelligentization and the rise of brain-like models, the lack of biological brain structural constraints and insufficient biological rationality in brain-like models built based on spiking neural networks has become increasingly prominent, thus limiting the development of spiking neural networks in computational neuroscience. Summary of the Invention

[0003] In view of this, this disclosure provides a method and apparatus for constructing a brain-like model based on fMRI brain structure, which is used to construct a spiking neural network brain-like model based on biological brain topological constraints, in order to solve the problem that existing spiking neural network brain-like models lack biological rationality.

[0004] In a first aspect, this disclosure provides a method for constructing a brain-like model based on fMRI brain structure, used to construct a spiking neural network brain-like model based on biological brain topological constraints. The method includes: dividing the fMRI data to be processed into brain regions to obtain M brain region image data; generating M model nodes based on the M brain region image data, wherein each model node represents a brain region containing the corresponding brain region image data; generating N model edges based on the correlation coefficient matrix between the M model nodes, wherein the correlation coefficient matrix represents the brain functional network connection strength between the M model nodes; filtering the N model edges based on a preset network topology threshold to obtain S model edges that meet preset conditions, wherein S is a positive integer less than or equal to N; generating topological constraints for the spiking neural network brain-like model based on biological brain functional networks based on the M model nodes and S model edges; and constructing the spiking neural network brain-like model based on the topological constraints.

[0005] In conjunction with the first aspect, in some implementations of the first aspect, a spiking neural network brain-like model is constructed based on topological constraints, including: generating network nodes of the spiking neural network brain-like model based on a preset second-order neuron model and M model nodes, wherein the preset second-order neuron model includes the Izhikevich neuron model; and constructing the spiking neural network brain-like model based on the network nodes and topological constraints of the spiking neural network brain-like model.

[0006] In conjunction with the first aspect, in some implementations of the first aspect, a spiking neural network brain-like model is constructed based on the network nodes and topological constraints of the spiking neural network brain-like model, including: generating network edges of the spiking neural network brain-like model based on a preset synaptic plasticity model and S model edges; and constructing the spiking neural network brain-like model based on the network edges, network nodes, and topological constraints of the spiking neural network brain-like model.

[0007] In conjunction with the first aspect, in some implementations of the first aspect, the preset synaptic plasticity model includes a synaptic plasticity model that is jointly regulated by excitability and inhibition. Before generating the network edges of the brain-like model based on the preset synaptic plasticity model and S model edges, the method further includes: determining the ratio of the number of excitatory neurons to inhibitory neurons included in the synaptic plasticity model based on neuroanatomical experimental data; and generating the preset synaptic plasticity model based on the ratio.

[0008] In conjunction with the first aspect, in some implementations of the first aspect, the model construction method further includes: a preset network topology threshold determined based on parameters characterizing network topology characteristics. These parameters characterizing network topology characteristics include at least one of network density, average node degree, small-world property, and scale-free property.

[0009] In conjunction with the first aspect, in some implementations of the first aspect, M is 980, and the fMRI data to be processed is divided into brain regions to obtain M brain region image data, including: using the Zalesky_980 template to divide the fMRI data to be processed into brain regions to obtain 980 brain region image data.

[0010] In conjunction with the first aspect, in some implementations of the first aspect, before dividing the fMRI data to be processed into brain regions to obtain M brain region image data, the method further includes: acquiring the initial fMRI data of the subject; and preprocessing the initial fMRI data to obtain the fMRI data to be processed. The preprocessing includes temporal slice correction and spatial normalization, and also includes head motion correction, smoothing, and filtering.

[0011] Secondly, one embodiment of this disclosure provides a brain-like model construction device based on fMRI brain structure, used to construct a spiking neural network brain-like model based on biological brain topological constraints. The device includes: a brain region segmentation module for segmenting the fMRI data to be processed into brain regions, obtaining M brain region image data; a first generation module for generating M model nodes based on the M brain region image data, wherein each model node represents a brain region containing the corresponding brain region image data; a second generation module for generating N model edges based on the correlation coefficient matrix between the M model nodes, wherein the correlation coefficient matrix represents the brain functional network connection strength between the M model nodes; a filtering module for filtering the N model edges based on a preset network topology threshold, obtaining S model edges that meet preset conditions, wherein S is a positive integer less than or equal to N; a third generation module for generating topological constraints of the biological brain functional network for the spiking neural network brain-like model based on the M model nodes and S model edges; and a construction module for constructing the spiking neural network brain-like model based on the topological constraints. Thirdly, one embodiment of this disclosure provides an electronic device including a processor and a memory for storing processor-executable instructions, wherein the processor is configured to perform the methods mentioned in the first aspect. Fourthly, one embodiment of this disclosure provides a computer-readable storage medium storing a computer program configured to perform the methods mentioned in the first aspect.

[0012] The model construction method provided in this disclosure is used to construct a brain-like model of a spiking neural network based on biological brain topological constraints. Since the topological constraints of the model are determined based on functional magnetic resonance imaging (fMRI) data, this disclosure can fully utilize the topological characteristics of biological brain functional networks, and thus use these networks as topological constraints to construct a spiking neural network, thereby improving the biological rationality of the brain-like model based on the spiking neural network. Combined with this disclosure, the biological rationality of the brain-like model based on the spiking neural network can be effectively improved, thereby enabling the wider application of spiking neural networks in computational neuroscience. Attached Figure Description

[0013] The above and other objects, features, and advantages of this disclosure will become more apparent from the more detailed description of the embodiments thereof in conjunction with the accompanying drawings. The drawings are provided to further illustrate the embodiments of this disclosure and form part of the specification. They are used together with the embodiments of this disclosure to explain the disclosure and do not constitute a limitation thereof.

[0014] Figure 1 The diagram shown is an application scenario illustration provided by an embodiment of this disclosure.

[0015] Figure 2The diagram shown is a flowchart of a model building method provided in an embodiment of this disclosure.

[0016] Figure 3 The diagram shown is a flowchart illustrating the process of constructing a brain-like spiking neural network model based on topological constraints according to an embodiment of this disclosure.

[0017] Figure 4 The diagram shows a flowchart of constructing a spiking neural network brain-like model based on network nodes and topological constraints according to an embodiment of this disclosure.

[0018] Figure 5 The diagram shown is a flowchart illustrating the process of constructing a spiking neural network brain-like model based on network nodes and topological constraints according to another embodiment of this disclosure.

[0019] Figure 6 The diagram shown is a flowchart illustrating the process of determining a preset network topology threshold according to an embodiment of this disclosure.

[0020] Figure 7 The diagram shown is a flowchart of a model building method provided in another embodiment of this disclosure.

[0021] Figure 8 The diagram shown is a schematic diagram of the structure of a model building apparatus provided in an embodiment of this disclosure.

[0022] Figure 9 The diagram shown is a structural schematic of an electronic device provided in an embodiment of this disclosure. Detailed Implementation

[0023] The technical solutions of the present disclosure will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present disclosure, and not all embodiments.

[0024] Computational neuroscience integrates cognitive neuroscience, electrical engineering, information science, and computer science, aiming to explain a range of phenomena related to the biological brain's nervous system from a multidisciplinary perspective. It plays a crucial role in the future development of brain-like intelligence and research on nervous system diseases. With the development of the information industry, computational neuroscience has gained increasing attention, further promoting the development of information science and brain science. Therefore, the application of artificial neural network models is becoming increasingly widespread. Spiking neural networks, as a new generation of artificial network models, differ from traditional artificial neural network models. Spiking neural networks possess powerful processing capabilities for complex nonlinear spatiotemporal information and play a vital role in the field of computational neuroscience, serving as a necessary theoretical and model foundation. However, as computational neuroscience develops towards intelligence and brain-like models emerge, the lack of biological brain structural constraints and insufficient biological rationality in brain-like models built based on spiking neural networks is becoming increasingly prominent, thus limiting the development of spiking neural networks in computational neuroscience.

[0025] Therefore, improving the biological plausibility of brain-like models built on spiking neural networks is an urgent problem to be solved. To address this issue, this disclosure provides a model construction method for building a spiking neural network-based brain-like model, thereby resolving the lack of biological plausibility in existing spiking neural network-based brain-like models.

[0026] The following is combined with Figure 1 A brief introduction to the application scenarios of the embodiments of this disclosure is provided.

[0027] Figure 1 The illustration shows an application scenario provided by an embodiment of this disclosure. Figure 1As shown, this scenario is for constructing a spiking neural network brain-like model. Specifically, the scenario for constructing a spiking neural network brain-like model includes a server 110, a user terminal 120 that is communicatively connected to the server 110, and a storage device 130 for functional magnetic resonance imaging (fMRI) data. The server 110 is used to execute the model construction method mentioned in the embodiments of this disclosure. For example, server 110 is configured to perform the following: divide the functional magnetic resonance imaging data to be processed into brain regions to obtain M brain region image data; generate M model nodes based on the M brain region image data, wherein each model node represents a brain region containing the brain region image data corresponding to the model node; generate N model edges based on the correlation coefficient matrix between the M model nodes, wherein the correlation coefficient matrix is ​​used to represent the brain functional network connection strength between the M model nodes; filter the N model edges based on a preset network topology threshold to obtain S model edges that meet preset conditions, wherein S is a positive integer less than or equal to N; generate topological constraints of a spiking neural network brain-like model based on biological brain functional networks based on the M model nodes and S model edges; and construct a spiking neural network brain-like model based on the topological constraints of the brain-like model.

[0028] For example, in practical applications, a user sends an instruction from user terminal 120 to server 110 to construct a spiking neural network brain-like model for user A. After receiving the instruction, server 110 retrieves A's functional magnetic resonance imaging (fMRI) data from storage device 130, generates a spiking neural network brain-like model for A based on the fMRI data, and then outputs the model to user terminal 120 so that user terminal 120 can apply the spiking neural network brain-like model.

[0029] For example, the user terminal 120 mentioned above includes, but is not limited to, computer terminals such as desktop computers and laptops. The data stored in the storage device 130 mentioned above includes, but is not limited to, data from the publicly available neuroimaging database of Neuroimaging Tools & Resources Collaboratory (NITRC), user-input functional magnetic resonance imaging data, and other image data.

[0030] The following is combined with Figures 2 to 7 A brief introduction to the model construction method disclosed herein.

[0031] Figure 2 The diagram shown is a flowchart illustrating a model construction method provided in one embodiment of this disclosure. Figure 2 As shown in the embodiments of this disclosure, the model construction method includes the following steps.

[0032] Step S210: Divide the functional magnetic resonance imaging data to be processed into brain regions to obtain M brain region imaging data. Where M is a positive integer.

[0033] For example, the selected functional magnetic resonance imaging data to be processed is the image data of a healthy adult male selected from the NITRC public neuroimaging database. Then, the acquired image data is divided into brain regions to obtain M brain region image data.

[0034] Step S220: Based on the imaging data of M brain regions, generate M model nodes.

[0035] For example, model nodes contain corresponding brain region imaging data; that is, each brain region is used as a node in the brain functional network, and each node represents the brain region after being divided by functional magnetic resonance imaging data. For instance, based on M brain region imaging data, M brain regions are used as nodes in the brain functional network to generate M network nodes.

[0036] Step S230: Generate N model edges based on the correlation coefficient matrix between M model nodes.

[0037] For example, the correlation coefficient matrix between model nodes is used to represent the brain functional network connectivity strength between M model nodes, that is, the functional connectivity strength between nodes is determined by the Pearson correlation coefficient between the average time series of different nodes.

[0038] For example, by calculating the Pearson correlation coefficient using the formula, an M*M symmetric correlation coefficient matrix is ​​obtained, where the non-zero correlation coefficients are 2*N, representing the N edges of the model. The mathematical expression for the Pearson correlation coefficient is shown below.

[0039]

[0040] in, x i (t) and x j (t) They are nodes i and nodes j exist t Average time series at any given moment; and They are nodes respectively i and nodes j The average time series; r ij For nodes i and nodes j The correlation coefficient between them; T For time points.

[0041] Step S240: Based on a preset network topology threshold, filter the N model edges to obtain S model edges that meet the preset conditions, where S is a positive integer less than or equal to N.

[0042] For example, a preset network topology threshold is selected that conforms to the threshold of a biological brain network. Based on the preset network topology threshold, the connection status of network nodes is obtained. Based on the network node connection status, N model edges are filtered to obtain S model edges that conform to the preset network topology threshold.

[0043] Step S250: Based on M model nodes and S model edges, generate topological constraints for the spiking neural network brain-like model based on biological brain functional networks.

[0044] For example, based on M model nodes and S model edges, a binary matrix can be obtained, which is the topological constraint of the spiking neural network brain-like model based on the biological brain functional network.

[0045] Step S260: Construct a brain-like model of a spiking neural network based on topological constraints.

[0046] For example, based on the above topological constraints, a brain-like model based on a spiking neural network is constructed.

[0047] In practical applications, the functional magnetic resonance imaging (fMRI) data to be processed is first divided into brain regions to obtain M brain region image data. Based on the M brain region image data, M model nodes are generated. Then, based on the correlation coefficient matrix between the M model nodes, N model edges are generated. Then, based on a preset network topology threshold, the N model edges are filtered to obtain S model edges that meet the preset conditions. Subsequently, based on the M model nodes and S model edges, the topological constraints of the spiking neural network brain-like model based on the biological brain functional network are generated. Finally, based on the topological constraints, the spiking neural network brain-like model is constructed.

[0048] Since the network topology used to construct the brain-like model is obtained from functional magnetic resonance imaging data, that is, based on the biological brain structure as the network topology constraint, the embodiments of this disclosure can generate a brain-like model based on spiking neural networks based on the biological brain structure as the network topology constraint. This enables the spiking neural network to reflect the real brain network connections, making the spiking neural network more biologically plausible and solving the problem that current brain-like models based on spiking neural networks lack biological plausibility.

[0049] Figure 3 The diagram shown is a flowchart illustrating the construction of a brain-like spiking neural network model based on topological constraints, according to an embodiment of this disclosure. Figure 2 Extending from the illustrated embodiment Figure 3 The illustrated embodiment will be described in detail below. Figure 3 The illustrated embodiments and Figure 2 The differences between the embodiments shown are not repeated here, and the similarities are not repeated here.

[0050] like Figure 3 As shown in this embodiment, the steps for constructing a spiking neural network brain-like model based on topological constraints include the following steps.

[0051] Step S310: Based on the preset second-order neuron model and M model nodes, generate network nodes for the spiking neural network brain-like model.

[0052] For example, the preset second-order neuron model is the Izhikevich neuron model. The Izhikevich neuron model effectively reflects the firing characteristics of biological neurons, has low time complexity, and is suitable for constructing large-scale networks. Therefore, when constructing a spiking neural network brain-like model, the Izhikevich neuron model is used as the node of the spiking neural network brain-like model. In this embodiment, the selected Izhikevich neuron model includes two firing modes: a regular firing mode and a low-threshold firing mode. The regular firing mode simulates the firing of excitatory neurons, and the low-threshold firing mode simulates the firing of inhibitory neurons. The dimensionless parameters in the mathematical model of the Izhikevich neuron model are selected with different values ​​to represent the two firing modes. It can be understood that the preset second-order neuron model includes, but is not limited to, the Izhikevich neuron model.

[0053] For example, the mathematical model of the Izhikevich neuron model is shown below.

[0054]

[0055]

[0056]

[0057] in, V I It is the neuronal membrane voltage; u It is a membrane voltage recovery variable; I It is the sum of the external input current and the current conducted through multiple synapses. a It is a recovery variable u Time scale; b It is a recovery variable u Sensitivity to fluctuations in membrane voltage within the domain; c It is the reset value of the membrane voltage caused by rapid high threshold potassium conductivity; d It is the reset value of the recovery variable caused by slow, high-threshold potassium and sodium conductance. a , b, c and d All parameters are dimensionless, and their values ​​can be adjusted to simulate various firing patterns of neurons. In summary, this embodiment uses a regular firing pattern to simulate the firing of excitatory neurons, and the selected parameters are as follows: a =0.02, b =0.2, c =-65, d =8; To simulate inhibitory neuron firing in a low-threshold firing mode, the following parameters were selected: a =0.02, b =0.25, c =-65, d =2.

[0058] Step S320: Construct a spiking neural network brain-like model based on the network nodes and topological constraints of the spiking neural network brain-like model.

[0059] For example, in the brain-like model constructed under the above topological constraints, the Izhikevich neuron model is set as a brain-like model network node to form a spiking neural network brain-like model.

[0060] Because the Izhikevich neuron model can well reflect the firing characteristics of biological neurons, and the selected Izhikevich neuron model can express both firing modes of biological neurons, it can well express the firing of excitatory neurons and inhibitory neurons. From the perspective of network nodes, this further increases the biological rationality of the constructed brain-like model. Therefore, the embodiments of this disclosure can, based on topological constraints, further increase the biological rationality of the brain-like model based on spiking neural networks in the direction of the nodes of the brain-like model.

[0061] Figure 4 The diagram illustrates a flowchart of constructing a brain-like model based on network nodes and topological constraints of a spiking neural network, according to an embodiment of this disclosure. Figure 3 Extending from the illustrated embodiment Figure 4 The illustrated embodiment will be described in detail below. Figure 4 The illustrated embodiments and Figure 3 The differences between the embodiments shown are not repeated here, and the similarities are not repeated here.

[0062] like Figure 4 As shown in this embodiment, the steps for constructing a spiking neural network brain-like model based on network nodes and topological constraints of the spiking neural network brain-like model include the following steps.

[0063] Step S410: Based on the preset synaptic plasticity model and S model edges, generate the network edges of the spiking neural network brain-like model.

[0064] For example, the preset synaptic plasticity model is a prominent plasticity model that uses both excitatory and inhibitory regulation. Both excitatory and inhibitory synapses regulate the spiking neural network through changes in synaptic conductance. Based on the regulation rules, preset parameters are defined for the influence of excitatory and inhibitory synapses in both excited and inhibited states, including reversal potential, excitatory synaptic weight, inhibitory synaptic weight, attenuation constant of excitatory synaptic conductance, attenuation constant of inhibitory synaptic conductance, maximum correction value of excitatory synaptic conductance, minimum correction value of excitatory synaptic conductance, and maximum and minimum correction values ​​of inhibitory synaptic conductance. Based on these parameter settings, a synaptic plasticity model is obtained. This model is then combined with S model edges to generate the edges of a brain-like model.

[0065] For example, the synaptic output current is approximately linearly related to the input voltage, and its mathematical description is as follows.

[0066]

[0067] in, I syn It is synaptic current; g syn It is synaptic conductance; V j(t) This represents the membrane potential of the postsynaptic neuron. E This is the inversion potential. In this embodiment, the inversion potential of an excitatory synapse is selected. E ex The inversion potential of an inhibitory synapse is 0 mV. E in The voltage is -70mV. Both excitatory and inhibitory synapses regulate spiking neural networks by altering synaptic conductance. When the postsynaptic neuron... j No presynaptic neuron received i During the action potential, the synaptic conductance of excitatory and inhibitory synapses decays exponentially, as shown below.

[0068]

[0069]

[0070] in, g ex G represents the excitatory synaptic weight. in Indicates the inhibitory synaptic weight; τ ex and τ in These represent the attenuation constants of excitatory synaptic conductance and inhibitory synaptic conductance, respectively.

[0071] When postsynaptic neurons j Received from presynaptic neuron i During the action potential, the changes in excitatory and inhibitory synaptic conductance are shown below.

[0072]

[0073]

[0074] in, and The excitatory and inhibitory conductance increments caused by the action potential are respectively derived from the excitatory correction function. w ij and inhibition correction function m ij Adjustments are made. Excitability correction function. w i and inhibition correction function m ij The mathematical description of it is as follows.

[0075]

[0076]

[0077] in, A + and A - These are the maximum and minimum correction values ​​for excitatory synaptic conductance, respectively. B + and B - These are the maximum and minimum correction values ​​for inhibitory synaptic conductance, respectively. △t This refers to the firing interval between presynaptic and postsynaptic neurons. τ + and τ - These represent the time intervals between neuronal firing during synaptic enhancement and synaptic demagnetization, respectively. In summary, in this embodiment, the parameters are selected as follows: g ex and `g in The maximum and minimum values ​​are 0.015 and 0, respectively; τ + =τ - =20ms , A + =0.1, A - =0.105, B + =0.02, B- =0.003

[0078] Step S420: Construct a spiking neural network brain-like model based on the network edges, network nodes, and topological constraints of the spiking neural network brain-like model.

[0079] For example, in the above brain-like model constructed based on topological constraints and network nodes, the above synaptic plasticity model is set as the edge of the brain-like model network to form a spiking neural network brain-like model.

[0080] Research on biological synapses indicates that brain-like models regulated by both excitatory and inhibitory synapses possess greater biological completeness. From the perspective of network edges, this enhances the biological completeness and increases the biological plausibility of the spiking neural network-based brain-like model. Therefore, the embodiments disclosed in this paper, based on the network nodes and topological constraints of the spiking neural network-based brain-like model, further increase the biological plausibility of the spiking neural network-based brain-like model in the direction of its edges.

[0081] Combination Figure 5 This paper further explains the specific execution process of constructing a brain-like model based on the network nodes and topological constraints of a spiking neural network.

[0082] Figure 5 The diagram shown illustrates the process of constructing a spiking neural network-based brain-like model, including node and topological constraints, according to another embodiment of this disclosure. Figure 4 Extending from the illustrated embodiment Figure 5 The illustrated embodiment will be described in detail below. Figure 5 and Figure 4 The differences between the embodiments shown are not repeated here, and the similarities are not repeated here.

[0083] like Figure 5 As shown, in another embodiment of this disclosure, the preset synaptic plasticity model includes a synaptic plasticity model that is jointly regulated by excitability and inhibition. Before the step of constructing the spiking neural network brain-like model based on the network nodes and topological constraints of the spiking neural network brain-like model, the following steps are also included.

[0084] Step S510: Based on neuroanatomical experimental data, determine the ratio of excitatory neurons to inhibitory neurons included in the synaptic plasticity model.

[0085] For example, based on neuroanatomical experimental data, the ratio of excitatory neurons to inhibitory neurons can be selected as 4:1 to determine the model ratio of excitatory neurons and inhibitory neurons included in the synaptic plasticity model.

[0086] Step S520: Generate a preset synaptic plasticity model based on the quantity ratio.

[0087] For example, according to the ratio of excitatory neurons to inhibitory neurons mentioned above, the excitatory neurons and inhibitory neurons in the synaptic plasticity model are randomly distributed in a 4:1 ratio to generate a preset synaptic plasticity model.

[0088] Because the ratio of excitatory to inhibitory neurons in the synaptic plasticity model is based on neuroanatomical experimental data, the synaptic plasticity model is more biologically plausible. Therefore, the embodiments of this disclosure can construct a spiking neural network brain-like model based on the network nodes and network topology constraints of the spiking neural network brain-like model, employing a more biologically plausible synaptic plasticity model, thus providing a prerequisite for further increasing the biological plausibility of brain-like models constructed based on spiking neural networks.

[0089] Figure 6 The diagram shown is a flowchart illustrating the process of determining a preset network topology threshold according to an embodiment of this disclosure. Figure 2 Extending from the illustrated embodiment Figure 6 The illustrated embodiment will be described in detail below. Figure 6 The illustrated embodiments and Figure 2 The differences between the embodiments shown are not repeated here, and the similarities are not repeated here.

[0090] like Figure 6 As shown in this embodiment, the step of determining the preset network topology threshold includes the following steps.

[0091] Step S610: Adjust the threshold within a certain range to obtain parameters representing the network topology characteristics generated under different thresholds.

[0092] For example, the adjustment threshold is selected to be between 0.1 and 0.6, and the step size is set to 0.1. Different parameters expressing network topology characteristics can be generated according to different thresholds. In this embodiment, the parameters expressing network topology characteristics are selected as network density, average node degree, small-world property, and power-law exponent. Different network densities, average node degree, small-world property, and power-law exponents are obtained according to different thresholds.

[0093] For example, network density is derived from a defined formula, which is shown below.

[0094]

[0095] in, N Indicates the number of nodes in the network. L This represents the actual number of edges in the network.

[0096] For example, the average node degree is derived from the defined formula, which is shown below.

[0097]

[0098] in, D i For nodes i The degree. D i The mathematical expression for is shown below.

[0099]

[0100] in, h ij Represents a node i and nodes j The value is 1 if a connection exists between them, and 0 if no connection exists.

[0101] For example, the small-world property is used б To determine, when б> When 1, it indicates that the network has the small-world property, and its mathematical expression is as follows.

[0102]

[0103] in, C real and C random These represent the clustering coefficients of the constructed network and the random network, respectively; L real and L random These represent the shortest path lengths of the constructed network and the random network, respectively.

[0104] Step S620: Determine the preset network topology threshold based on the obtained different network topology characteristic parameters.

[0105] For example, based on the characteristics of biological brain topology, the network density is generally in the range of 5%-40%, the network has small-world properties and scale-free properties, the power-law exponent is around 2 and the clustering coefficient is relatively high. Considering the above characteristics of biological brain topology, this embodiment selects a threshold of 0.2.

[0106] Since the selection of the network topology threshold is based on the characteristics of biological brain network topology, the biological rationality of the spiking neural network brain-like model is increased.

[0107] In this embodiment of the disclosure, the step of dividing the functional magnetic resonance imaging data to be processed into brain regions to obtain M brain region imaging data includes using the Zalesky_980 template to divide the functional magnetic resonance imaging data to be processed into brain regions to obtain 980 brain region imaging data.

[0108] Combination Figure 7 This section further explains the specific execution process of building the model.

[0109] Figure 7 The diagram shown is a flowchart illustrating a model construction method provided in another embodiment of this disclosure. Figure 2 Extending from the illustrated embodiment Figure 7 The illustrated embodiment will be described in detail below. Figure 7 The illustrated embodiments and Figure 2 The differences between the embodiments shown are not repeated here, and the similarities are not repeated here.

[0110] like Figure 7 As shown, in another embodiment of this disclosure, before dividing the functional magnetic resonance imaging data to be processed into brain regions to obtain M brain region imaging data, the following steps are also included.

[0111] Step S710: Acquire the subject's initial functional magnetic resonance imaging data.

[0112] Step S720: Preprocess the initial functional magnetic resonance imaging data to obtain the functional magnetic resonance imaging data to be processed.

[0113] For example, the initial functional magnetic resonance imaging (fMRI) data is preprocessed, including temporal correction and spatial normalization. Depending on the application, the initial fMRI data exhibits temporal shifts between layers, necessitating temporal correction. Spatial normalization aims to eliminate differences in brain shape and size, converting the initial fMRI data to the standard MNI space. Considering all factors, this embodiment's preprocessing of the initial fMRI data includes not only temporal correction and spatial normalization but also head movement correction, smoothing, and filtering. During fMRI data acquisition, subjects inevitably experience head movements; therefore, head movement correction is performed to eliminate the impact of head movements on image localization. To reduce the influence of random noise and improve the signal-to-noise ratio, smoothing is applied to the fMRI data. To reduce low-frequency drift and high-frequency physiological noise, a bandpass filter is used to filter the fMRI data.

[0114] The above text combined Figures 2 to 7 The present disclosure describes in detail the method embodiments, which are then combined with the following. Figure 8 and Figure 9 The present disclosure provides a detailed description of the apparatus embodiments. Furthermore, it should be understood that the descriptions of the method embodiments correspond to the descriptions of the apparatus embodiments; therefore, any parts not described in detail can be found in the foregoing method embodiments.

[0115] Figure 8 The diagram shown is a structural schematic of a model building apparatus provided in an embodiment of this disclosure. Figure 8 As shown, the model building apparatus provided in this embodiment includes a brain region segmentation module 810, a first generation module 820, a second generation module 830, a screening module 840, a third generation module 850, and a building module 860.

[0116] Specifically, the brain region segmentation module 810 is used to segment the functional magnetic resonance imaging (fMRI) data to be processed into brain regions, obtaining M brain region image data. The first generation module 820 is used to generate M model nodes based on the M brain region image data, where each model node represents a brain region containing the corresponding brain region image data. The second generation module 830 is used to generate N model edges based on the correlation coefficient matrix between the M model nodes, where the correlation coefficient matrix represents the brain functional network connectivity strength between the M model nodes. The filtering module 840 is used to filter the N model edges based on a preset network topology threshold, obtaining S model edges that meet preset conditions, where S is a positive integer less than or equal to N. The third generation module 850 is used to generate topological constraints for the spiking neural network brain-like model based on the biological brain functional network, based on the M model nodes and S model edges. The construction module 860 is used to construct the spiking neural network brain-like model based on the topological constraints.

[0117] In some embodiments, the construction module 860 is further configured to generate network nodes of a spiking neural network brain-like model based on a preset second-order neuron model and M model nodes, wherein the preset second-order neuron model includes the Izhikevich neuron model; and construct a spiking neural network brain-like model based on the network nodes and topological constraints of the spiking neural network brain-like model.

[0118] In some embodiments, the construction module 860 is further configured to generate network edges of a spiking neural network brain-like model based on the synaptic plasticity model and S model edges; and construct a spiking neural network brain-like model based on the network edges of the spiking neural network brain-like model, the network nodes of the spiking neural network brain-like model, and topological constraints.

[0119] In some embodiments, the construction module 860 is further configured to determine a preset synaptic plasticity model. Determining the preset synaptic plasticity model includes: determining the ratio of excitatory neurons to inhibitory neurons included in the synaptic plasticity model based on neuroanatomical experimental data; and generating the preset synaptic plasticity model based on the ratio.

[0120] In some embodiments, the filtering module 840 is further configured to determine a preset network topology threshold. Specifically, the network threshold is determined based on parameters that characterize network topology characteristics, wherein the parameters characterizing network topology characteristics include at least one of network density, average node degree, small-world property, and scale-free property.

[0121] In some embodiments, the first generation module 820 is further configured to generate M image data sets, with a preset M of 980. Generating M image data sets with a preset M of 980 includes: dividing the functional magnetic resonance imaging data to be processed into brain regions to obtain 980 brain region image data sets.

[0122] In some embodiments, the first generation module 820 is further configured to generate functional magnetic resonance imaging (fMRI) data to be processed. Specifically, it acquires the initial fMRI data of the subject; preprocesses the initial fMRI data to obtain the fMRI data to be processed, wherein the preprocessing includes temporal correction and spatial normalization, and further includes head motion correction, smoothing, and filtering.

[0123] Figure 9 The diagram shown is a structural schematic of an electronic device provided in an embodiment of this disclosure. Figure 9 The electronic device 900 shown (which may specifically be a computer device) includes a memory 901, a processor 902, a communication interface 903, and a bus 904. The memory 901, processor 902, and communication interface 903 are interconnected via the bus 904.

[0124] The memory 901 may be a read-only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM). The memory 901 may store a program, and when the program stored in the memory 901 is executed by the processor 902, the processor 902 and the communication interface 903 are used to execute the various steps of the model building method of the embodiments of this disclosure.

[0125] The processor 902 may be a general-purpose central processing unit (CPU), microprocessor, application specific integrated circuit (ASIC), graphics processing unit (GPU), or one or more integrated circuits, used to execute relevant programs to achieve the functions required by each unit in the model building apparatus of this disclosure embodiment.

[0126] The processor 902 can also be an integrated circuit chip with signal processing capabilities. In implementation, each step of the model construction method of this disclosure can be completed by the integrated logic circuits in the hardware of the processor 902 or by instructions in software form. The processor 902 can also be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. It can implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of this disclosure. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the method disclosed in the embodiments of this disclosure can be directly embodied in the execution of a hardware decoding processor, or executed by a combination of hardware and software modules in the decoding processor. The software modules can be located in random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, or other mature storage media in the art. The storage medium is located in the memory 901. The processor 902 reads the information in the memory 901 and, in conjunction with its hardware, performs the functions required by the units included in the model building apparatus of this disclosure embodiment, or executes the model building method of this disclosure method embodiment.

[0127] The communication interface 903 uses transceiver devices, such as, but not limited to, transceivers, to enable communication between the electronic device 900 and other devices or communication networks. For example, functional magnetic resonance imaging (fMRI) data signals can be acquired and processed through the communication interface 903.

[0128] Bus 904 may include a pathway for transmitting information between various components of electronic device 900 (e.g., memory 901, processor 902, communication interface 903).

[0129] It should be noted that, although Figure 9The illustrated electronic device 900 only shows the memory, processor, and communication interface. However, those skilled in the art should understand that in specific implementations, the electronic device 900 may also include other devices necessary for normal operation. Furthermore, depending on specific needs, those skilled in the art should understand that the electronic device 900 may also include hardware devices for implementing other additional functions. Moreover, those skilled in the art should understand that the electronic device 900 may only include the devices necessary for implementing the embodiments of this disclosure, and may not necessarily include... Figure 9 All the devices shown.

[0130] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this disclosure.

[0131] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0132] In the embodiments provided in this disclosure, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0133] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0134] In addition, the functional units in the various embodiments of this disclosure can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0135] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this disclosure, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this disclosure. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory, random access memory, magnetic disks, or optical disks.

[0136] The above description is merely a specific embodiment of this disclosure, but the scope of protection of this disclosure is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this disclosure should be included within the scope of protection of this disclosure. Therefore, the scope of protection of this disclosure should be determined by the scope of the claims.

Claims

1. A model construction method, characterized in that, The method for constructing a brain-like spiking neural network model based on biological brain topological constraints includes: Initial functional magnetic resonance imaging (fMRI) data of the subjects were acquired; the initial fMRI data were preprocessed to obtain fMRI data to be processed, including temporal correction, spatial normalization, head motion correction, smoothing and filtering. Using the Zalesky_980 template, the functional magnetic resonance imaging data to be processed is divided into brain regions to obtain M brain region imaging data, where M is 980; Based on M brain region image data, M model nodes are generated, where each model node represents a brain region containing the brain region image data corresponding to the model node. Based on the correlation coefficient matrix between M model nodes, N model edges are generated. The correlation coefficient matrix is ​​used to represent the brain functional network connectivity strength between the M model nodes. The functional connectivity strength between nodes is determined by the Pearson correlation coefficient between the average time series of different nodes. Based on a preset network topology threshold, N model edges are filtered to obtain S model edges that meet preset conditions, where S is a positive integer less than or equal to N; the preset network topology threshold is determined based on parameters that can characterize network topology characteristics. Based on M model nodes and S model edges, a topological constraint for a spiking neural network brain-like model based on biological brain functional networks is generated. Constructing the aforementioned spiking neural network brain-like model specifically includes: Based on a pre-defined second-order neuron model and M model nodes, network nodes of a brain-like spiking neural network model are generated. Based on a preset synaptic plasticity model and S model edges, network edges of a spiking neural network brain-like model are generated; the preset synaptic plasticity model includes a synaptic plasticity model that is jointly regulated by excitability and inhibition. The spiking neural network brain-like model is constructed based on the network edges, network nodes, and topological constraints of the spiking neural network brain-like model.

2. The model construction method according to claim 1, characterized in that, The preset second-order neuron model includes the Izhikevich neuron model.

3. A model building apparatus for executing the model building method according to any one of claims 1 to 2, characterized in that, The device for constructing a brain-like spiking neural network model based on biological brain topological constraints includes: The brain region segmentation module is used to segment the functional magnetic resonance imaging data to be processed into brain regions, resulting in M ​​brain region imaging data. The first generation module is used to generate M model nodes based on the M brain region image data, wherein the model node represents the brain region containing the brain region image data corresponding to the model node. The second generation module is used to generate N model edges based on the correlation coefficient matrix between the M model nodes, wherein the correlation coefficient matrix is ​​used to represent the brain functional network connection strength between the M model nodes. The filtering module is used to filter the N model edges based on a preset network topology threshold to obtain S model edges that meet preset conditions, where S is a positive integer less than or equal to N. The third generation module is used to generate the topological constraints of the biological brain functional network for the spiking neural network brain-like model based on the M model nodes and the S model edges. The construction module constructs the spiking neural network brain-like model based on the topological constraints.

4. An electronic device, comprising: include: processor; Memory used to store the processor's executable instructions. The processor is used to execute the model building method according to any one of claims 1 to 2.

5. A computer readable storage medium, characterized in that, The storage medium stores a computer program for executing the model building method according to any one of claims 1 to 2.

Images