CN114757334A - Model construction method and device, storage medium and electronic equipment - Google Patents
Model construction method and device, storage medium and electronic equipment Download PDFInfo
- Publication number
- CN114757334A CN114757334A CN202210289158.XA CN202210289158A CN114757334A CN 114757334 A CN114757334 A CN 114757334A CN 202210289158 A CN202210289158 A CN 202210289158A CN 114757334 A CN114757334 A CN 114757334A
- Authority
- CN
- China
- Prior art keywords
- model
- brain
- neural network
- image data
- network
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000010276 construction Methods 0.000 title claims abstract description 22
- 238000013528 artificial neural network Methods 0.000 claims abstract description 115
- 210000004556 brain Anatomy 0.000 claims abstract description 105
- 238000012421 spiking Methods 0.000 claims abstract description 60
- 238000005481 NMR spectroscopy Methods 0.000 claims abstract description 36
- 230000003925 brain function Effects 0.000 claims abstract description 19
- 239000011159 matrix material Substances 0.000 claims abstract description 18
- 238000012216 screening Methods 0.000 claims abstract description 13
- 210000002569 neuron Anatomy 0.000 claims description 53
- 238000000034 method Methods 0.000 claims description 40
- 238000012545 processing Methods 0.000 claims description 32
- 230000003956 synaptic plasticity Effects 0.000 claims description 32
- 230000002401 inhibitory effect Effects 0.000 claims description 29
- 230000002964 excitative effect Effects 0.000 claims description 28
- 238000012937 correction Methods 0.000 claims description 15
- 238000007781 pre-processing Methods 0.000 claims description 10
- 206010001497 Agitation Diseases 0.000 claims description 5
- 238000009499 grossing Methods 0.000 claims description 5
- 230000002123 temporal effect Effects 0.000 claims description 5
- 238000001914 filtration Methods 0.000 claims description 4
- 230000011218 segmentation Effects 0.000 claims description 3
- 238000004590 computer program Methods 0.000 claims description 2
- 230000033001 locomotion Effects 0.000 claims description 2
- 230000001105 regulatory effect Effects 0.000 claims description 2
- 230000000946 synaptic effect Effects 0.000 description 21
- 238000004891 communication Methods 0.000 description 12
- 210000000225 synapse Anatomy 0.000 description 12
- 230000008569 process Effects 0.000 description 11
- 238000002599 functional magnetic resonance imaging Methods 0.000 description 10
- 230000006870 function Effects 0.000 description 8
- 238000011161 development Methods 0.000 description 6
- 230000018109 developmental process Effects 0.000 description 6
- 238000012986 modification Methods 0.000 description 6
- 230000004048 modification Effects 0.000 description 6
- 238000010586 diagram Methods 0.000 description 5
- 239000012528 membrane Substances 0.000 description 5
- 230000001242 postsynaptic effect Effects 0.000 description 4
- 238000011084 recovery Methods 0.000 description 4
- 230000036982 action potential Effects 0.000 description 3
- 230000008878 coupling Effects 0.000 description 3
- 238000010168 coupling process Methods 0.000 description 3
- 238000005859 coupling reaction Methods 0.000 description 3
- 230000004886 head movement Effects 0.000 description 3
- 230000001537 neural effect Effects 0.000 description 3
- 238000002610 neuroimaging Methods 0.000 description 3
- 238000010606 normalization Methods 0.000 description 3
- ZLMJMSJWJFRBEC-UHFFFAOYSA-N Potassium Chemical compound [K] ZLMJMSJWJFRBEC-UHFFFAOYSA-N 0.000 description 2
- 238000013178 mathematical model Methods 0.000 description 2
- 229910052700 potassium Inorganic materials 0.000 description 2
- 239000011591 potassium Substances 0.000 description 2
- 210000005215 presynaptic neuron Anatomy 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 238000006467 substitution reaction Methods 0.000 description 2
- DGAQECJNVWCQMB-PUAWFVPOSA-M Ilexoside XXIX Chemical compound C[C@@H]1CC[C@@]2(CC[C@@]3(C(=CC[C@H]4[C@]3(CC[C@@H]5[C@@]4(CC[C@@H](C5(C)C)OS(=O)(=O)[O-])C)C)[C@@H]2[C@]1(C)O)C)C(=O)O[C@H]6[C@@H]([C@H]([C@@H]([C@H](O6)CO)O)O)O.[Na+] DGAQECJNVWCQMB-PUAWFVPOSA-M 0.000 description 1
- 206010063385 Intellectualisation Diseases 0.000 description 1
- 208000012902 Nervous system disease Diseases 0.000 description 1
- 238000004422 calculation algorithm Methods 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000001149 cognitive effect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000010304 firing Methods 0.000 description 1
- 230000000670 limiting effect Effects 0.000 description 1
- 210000005036 nerve Anatomy 0.000 description 1
- 230000008062 neuronal firing Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000004091 panning Methods 0.000 description 1
- 230000037361 pathway Effects 0.000 description 1
- 230000003518 presynaptic effect Effects 0.000 description 1
- 230000002787 reinforcement Effects 0.000 description 1
- 230000035945 sensitivity Effects 0.000 description 1
- 229910052708 sodium Inorganic materials 0.000 description 1
- 239000011734 sodium Substances 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
- 230000003313 weakening effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/049—Temporal neural networks, e.g. delay elements, oscillating neurons or pulsed inputs
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Magnetic Resonance Imaging Apparatus (AREA)
Abstract
The disclosure discloses a model construction method and device, a storage medium and electronic equipment, which are used for constructing a spiking neural network brain model based on biological brain topological constraint, relate to the technical field of computational neurology and solve the problem that the spiking neural network brain model lacks biological rationality. The model construction method comprises the following steps: performing brain region division on functional nuclear magnetic resonance image data to be processed to obtain M brain region image data; generating M model nodes based on the M brain area image data; generating N model edges based on a correlation coefficient matrix among the M model nodes; based on a preset network topology threshold value, screening the N model edges to obtain S model edges meeting preset conditions; generating topological constraint of a brain-like model based on the biological brain function network based on the M model nodes and the S model edges; and constructing a brain-like model based on topological constraint. The present disclosure can improve the biological rationality of a brain-like model constructed based on a spiking neural network.
Description
Technical Field
The disclosure belongs to the technical field of computational neurology, and particularly relates to a model construction method and device, a storage medium and electronic equipment.
Background
In recent years, the computational neurology is rapidly developed, and the application of the artificial neural network model is more and more extensive. The impulse neural network as a new generation artificial network model has strong processing capability on complex nonlinear space-time information, plays an important role in the field of computational neurology, and is a necessary theoretical and model basis of the computational neurology. However, with the development of the computational neurology department towards the intelligent direction, the brain-like model is raised, and the brain-like model constructed based on the impulse neural network lacks of biological brain structure constraints, so that the problem of insufficient biological rationality is increasingly highlighted, thereby limiting the development of the impulse neural network in the computational neurology department.
Disclosure of Invention
In view of this, the present disclosure provides a model construction method and apparatus, a storage medium, and an electronic device, which are used to construct a spiking neural network brain-like model based on a biological brain topological constraint, so as to solve the problem that the existing spiking neural network brain-like model lacks biological rationality.
In a first aspect, an embodiment of the present disclosure provides a model construction method, configured to construct a spiking neural network brain-like model based on a biological brain topological constraint. The model construction method comprises the following steps: performing brain region division on functional nuclear magnetic resonance image data to be processed to obtain M brain region image data; generating M model nodes based on the M brain area image data, wherein the model nodes represent brain areas containing the brain area image data corresponding to the model nodes; generating N model edges based on a correlation coefficient matrix among the M model nodes, wherein the correlation coefficient matrix is used for representing the brain function network connection strength among the M model nodes; based on a preset network topology threshold, screening N model edges to obtain S model edges meeting preset conditions, wherein S is a positive integer less than or equal to N; generating topological constraint based on a biological brain function network of the impulse neural network brain-like model based on the M model nodes and the S model edges; and constructing a pulse neural network brain-like model based on topological constraint.
With reference to the first aspect, in certain implementations of the first aspect, constructing a spiking neural network brain-like model based on topological constraints includes: generating network nodes of a pulse neural network brain-like model based on a preset second-order neuron model and M model nodes, wherein the preset second-order neuron model comprises an Izhikevich neuron model; and constructing the impulse neural network brain-like model based on the network nodes and the topological constraints of the impulse neural network brain-like model.
With reference to the first aspect, in certain implementations of the first aspect, constructing the spiking neural network brain-like model based on network nodes and topological constraints of the spiking neural network brain-like model includes: generating a network edge of a pulse neural network brain-like model based on a preset synaptic plasticity model and the S model edges; and constructing the impulse neural network brain-like model based on the network edges of the impulse neural network brain-like model, the network nodes of the impulse neural network brain-like model and topological constraints.
With reference to the first aspect, in certain implementations of the first aspect, the preset synaptic plasticity model includes a synaptic plasticity model that is jointly modulated by excitability and inhibitivity, and before generating the network edges of the brain-like model based on the preset synaptic plasticity model and the S model edges, the method further includes: determining the number ratio of excitatory neurons to inhibitory neurons contained in the synaptic plasticity model based on neuroanatomy experimental data; and generating a preset synaptic plasticity model based on the quantity proportion.
With reference to the first aspect, in certain implementations of the first aspect, the model construction method further includes: the preset network topology threshold is determined based on parameters that characterize the topology of the network. Wherein the parameters characterizing the network topology include at least one of network density, average nodularity, small-world attributes, and non-scale attributes.
With reference to the first aspect, in certain implementations of the first aspect, M is 980, and brain region division is performed on the functional nuclear magnetic resonance image data to be processed to obtain M brain region image data, including: and (3) adopting a Zalesky _980 template to divide the brain area of the functional nuclear magnetic resonance influence data to be processed to obtain 980 brain area image data.
With reference to the first aspect, in certain implementations of the first aspect, before performing brain region segmentation on the functional mri image data to be processed to obtain M brain region image data, the method further includes: acquiring initial functional nuclear magnetic resonance image data of a subject; and preprocessing the initial functional nuclear magnetic resonance image data to obtain functional nuclear magnetic resonance image data to be processed. The preprocessing comprises temporal layer correction processing and spatial standardization processing, and also comprises head motion correction processing, smoothing processing and filtering processing.
In a second aspect, an embodiment of the present disclosure provides a model building apparatus for building a spiking neural network brain-like model based on biological brain topological constraint, where the model building apparatus includes: the brain area dividing module is used for performing brain area division on the functional nuclear magnetic resonance image data to be processed to obtain M brain area image data; the first generation module is used for generating M model nodes based on the M brain area image data, wherein the model nodes represent brain areas containing the brain area image data corresponding to the model nodes; the second generation module is used for generating N model edges based on a correlation coefficient matrix among the M model nodes, wherein the correlation coefficient matrix is used for representing the brain function network connection strength among the M model nodes; the screening module is used for screening the N model edges based on a preset network topology threshold value to obtain S model edges meeting preset conditions, wherein S is a positive integer less than or equal to N; the third generation module is used for generating topological constraint based on the biological brain function network of the impulse neural network brain-like model based on the M model nodes and the S model edges; and the construction module is used for constructing a pulse neural network brain-like model based on topological constraint.
In a third aspect, an embodiment of the present disclosure provides an electronic device, which includes a processor; a memory for storing processor executable instructions, wherein the processor is adapted to perform the method of the first aspect.
In a fourth aspect, an embodiment of the present disclosure provides a computer-readable storage medium, which stores a computer program for executing the method mentioned in the first aspect.
The model construction method provided by the embodiment of the disclosure is used for constructing the pulse neural network brain-like model based on the biological brain topological constraint, and the topological constraint of the model is determined based on the functional nuclear magnetic resonance image data, so that the embodiment of the disclosure can make full use of the topological characteristic of the biological brain functional network, and further construct the pulse neural network by using the biological brain functional network as the topological constraint, thereby achieving the purpose of improving the biological rationality of the brain-like model based on the pulse neural network. By combining the embodiment of the disclosure, the biological rationality of the brain-like model based on the impulse neural network can be effectively improved, so that the impulse neural network can be applied more widely in the computational neurology department.
Drawings
The above and other objects, features and advantages of the present disclosure will become more apparent by describing in more detail embodiments of the present disclosure with reference to the attached drawings. The accompanying drawings are included to provide a further understanding of the embodiments of the disclosure and are incorporated in and constitute a part of this specification, illustrate embodiments of the disclosure and together with the description serve to explain the principles of the disclosure and not to limit the disclosure.
Fig. 1 is a schematic view of an application scenario provided by an embodiment of the present disclosure.
Fig. 2 is a schematic flow chart of a model building method according to an embodiment of the present disclosure.
Fig. 3 is a schematic flowchart of a process for constructing a spiking neural network brain-like model based on topological constraints according to an embodiment of the present disclosure.
Fig. 4 is a schematic flow chart illustrating a process of constructing a spiking neural network brain model based on network nodes and topology constraints of the spiking neural network brain model according to an embodiment of the present disclosure.
Fig. 5 is a schematic flow chart illustrating a process of constructing a spiking neural network brain-like model based on network nodes and topology constraints of the spiking neural network brain-like model according to another embodiment of the present disclosure.
Fig. 6 is a schematic flowchart illustrating a process of determining a preset network topology threshold according to an embodiment of the present disclosure.
Fig. 7 is a schematic flow chart of a model building method according to another embodiment of the present disclosure.
Fig. 8 is a schematic structural diagram of a model building apparatus according to an embodiment of the present disclosure.
Fig. 9 is a schematic structural diagram of an electronic device according to an embodiment of the disclosure.
Detailed Description
The technical solutions in the embodiments of the present disclosure will be described clearly and completely with reference to the drawings in the embodiments of the present disclosure, and it is obvious that the described embodiments are only some embodiments of the present disclosure, rather than all embodiments.
The computational neuroscience integrates multiple fields of cognitive neuroscience, electronic engineering, information science, computer science and the like in a cross way, aims to explain a series of phenomena related to a biological brain nervous system from the perspective of multidisciplinary, and plays an important role in the development of future brain-like intelligence and nervous system disease research. With the development of the information industry, the neuroscience of computing is more and more emphasized, and the development of the information science and the brain science is further promoted. Therefore, the application of the artificial neural network model is more and more extensive, and the impulse neural network is used as a new generation artificial network model, different from the traditional artificial neural network model, has strong processing capability on complex nonlinear space-time information, plays an important role in the field of the computational neurology, and is a necessary theory and model basis of the computational neurology. However, as the computational neurology department develops towards the direction of intellectualization, the brain-like model rises, and the brain-like model constructed based on the impulse neural network lacks biological brain structure constraint and the problem of insufficient biological rationality is increasingly highlighted, so that the development of the impulse neural network in the computational neurology department is limited.
Therefore, the problem to be solved urgently is how to improve the biological rationality of the brain-like model constructed based on the impulse neural network. In order to solve the above problem, an embodiment of the present disclosure provides a model construction method, which is used for constructing a spiking neural network brain-like model, so as to solve the problem that the existing spiking neural network brain-like model lacks biological rationality.
An application scenario of the embodiment of the present disclosure is briefly described below with reference to fig. 1.
Fig. 1 is a schematic diagram of an application scenario provided by an embodiment of the present disclosure. As shown in fig. 1, the scene is a scene for constructing a spiking neural network brain-like model. Specifically, the scene for constructing the impulse neural network brain-like model includes a server 110, a user terminal 120 and a storage device 130 for functional magnetic resonance image data, wherein the user terminal 120 is respectively in communication connection with the server 110, and the server 110 is configured to execute the model construction method according to the embodiment of the present disclosure. Illustratively, the server 110 is configured to perform: performing brain region division on functional nuclear magnetic resonance image data to be processed to obtain M brain region image data; generating M model nodes based on the M brain area image data, wherein the model nodes represent brain areas containing the brain area image data corresponding to the model nodes; generating N model edges based on a correlation coefficient matrix among the M model nodes, wherein the correlation coefficient matrix is used for representing the brain function network connection strength among the M model nodes; based on a preset network topology threshold, screening N model edges to obtain S model edges meeting preset conditions, wherein S is a positive integer less than or equal to N; generating topological constraint based on a biological brain function network of the impulse neural network brain-like model based on the M model nodes and the S model edges; and constructing a pulse neural network brain-like model based on the topological constraint of the brain-like model.
Illustratively, in the practical application process, the user uses the user terminal 120 to issue an instruction for constructing the impulse neural network brain-like model to the server 110. After the server 110 receives the instruction, it calls out some functional nuclear magnetic resonance image data of a in the storage device 130, generates a spiking neural network brain-like model constructed for some a based on the functional nuclear magnetic resonance image data, and then outputs the model to the user terminal 120, so that the user terminal 120 applies the spiking neural network brain-like model.
Illustratively, the user terminal 120 mentioned above includes, but is not limited to, a computer terminal such as a desktop computer, a notebook computer, etc. The data stored in the storage device 130 includes, but is not limited to, neuroimaging Tools & Resources laboratories (NITRC), public neuroimaging data in a neuroimaging database, functional mri data input by a user, and the like.
The model construction method of the present disclosure is briefly described below with reference to fig. 2 to 7.
Fig. 2 is a schematic flow chart of a model building method according to an embodiment of the present disclosure. As shown in fig. 2, the model construction method provided by the embodiment of the present disclosure includes the following steps.
Step S210, performing brain region division on the functional nuclear magnetic resonance image data to be processed to obtain M brain region image data. Wherein M is a positive integer.
Illustratively, the selected Functional Magnetic Resonance Imaging (FMRI) data to be processed is image data of a healthy adult male selected from a NITRC public nerve image database, and then the acquired image data is divided into brain areas to obtain M brain area image data.
Step S220, based on the M brain region image data, M model nodes are generated.
Illustratively, the model nodes include corresponding brain region image data, that is, each brain region is taken as a node of a brain function network, and each node represents the corresponding brain region after the Functional Magnetic Resonance Image (FMRI) data is divided. For example, M brain regions are used as nodes of a brain function network according to M brain region image data to generate M network nodes.
Step S230, based on the correlation coefficient matrix between the M model nodes, N model edges are generated.
Illustratively, the correlation coefficient matrix between the model nodes is used for representing the brain function network connection strength between the M model nodes, i.e. the functional connection strength between the nodes is determined by Pearson correlation coefficients between different node average time sequences.
Illustratively, by Pearson correlation coefficient calculation, a symmetric correlation coefficient matrix of M × M is obtained, where the correlation coefficient other than 0 is 2 × N, i.e., the edges of N models are obtained. The mathematical expression of Pearson correlation coefficient is specifically as follows.
Wherein x isi(t) and xj(t) are average time sequences of the node i and the node j at the moment t respectively;
and
average time sequences of the node i and the node j are respectively; r isijIs the correlation coefficient between the node i and the node j; t is the number of time points.
And S240, screening the N model edges based on a preset network topology threshold value to obtain S model edges meeting preset conditions, wherein S is a positive integer less than or equal to N.
Illustratively, the preset network topology threshold is selected from thresholds meeting the biological brain network, and the connection condition of the network nodes is obtained according to the preset network topology threshold. And screening the N model edges according to the connection condition of the network nodes to obtain S model edges meeting a preset network topology threshold.
And S250, generating topological constraint based on the biological brain function network of the impulse neural network brain model based on the M model nodes and the S model edges.
Illustratively, according to M model nodes and S model edges, a binary matrix, which is a topological constraint based on a biological brain function network of the spiking neural network brain-like model, can be obtained.
And step S260, constructing a pulse neural network brain-like model based on topological constraint.
Illustratively, based on the above topological constraints, a brain-like model based on a spiking neural network is constructed.
In the practical application process, firstly, brain area division is carried out on functional nuclear magnetic resonance image data to be processed to obtain M brain area image data, M model nodes are generated based on the M brain area image data, N model edges are generated based on a correlation coefficient matrix between the M model nodes, then the N model edges are screened based on a preset network topology threshold value to obtain S model edges meeting preset conditions, then topological constraints based on a biological brain function network of the impulse neural network brain model are generated based on the M model nodes and the S model edges, and finally, the impulse neural network brain model is constructed based on the topological constraints.
Because the network topology adopted for constructing the brain-like model is obtained according to the functional nuclear magnetic resonance image data, namely the brain-like model based on the biological brain structure is constrained by the network topology, the embodiment of the disclosure can generate the brain-like model based on the impulse neural network for the network topology constraint according to the biological brain structure, so that the impulse neural network can reflect the real brain network connection, the impulse neural network has biological rationality, and the problem that the current brain-like model based on the impulse neural network lacks biological rationality is solved.
Fig. 3 is a schematic flowchart illustrating a process of constructing a neural network-like brain model based on topological constraints according to an embodiment of the present disclosure. The embodiment shown in fig. 3 is extended based on the embodiment shown in fig. 2, and the differences between the embodiment shown in fig. 3 and the embodiment shown in fig. 2 will be emphasized below, and the descriptions of the same parts will not be repeated.
As shown in fig. 3, in the embodiment of the present disclosure, the step of constructing the spiking neural network brain-like model based on the topological constraint includes the following steps.
And S310, generating network nodes of the impulse neural network brain-like model based on a preset second-order neuron model and the M model nodes.
Illustratively, the Izhikevich neuron model is selected from the preset second-order neuron model, the Izhikevich neuron model can well reflect the discharge characteristic of biological neurons, the time complexity is low, and the method is suitable for large-scale network construction. In this embodiment, the Izhikevich neuron model is selected to include two discharge modes, including a regular discharge mode and a low-threshold discharge mode, where the regular discharge mode simulates the discharge of excitatory neurons and the low-threshold discharge mode simulates the discharge of inhibitory neurons. Dimensionless parameters in the mathematical model of the Izhikevich neuron model are respectively selected to have different values so as to represent two discharge modes. It is understood that the predetermined second-order neuron model includes, but is not limited to, an Izhikevich neuron model.
Illustratively, the mathematical model of the Izhikevich neuron model is shown below.
if
Wherein, VIIs the neuronal membrane voltage; u is the membrane voltage recovery variable; i is the sum of the external input current and the current conducted through the plurality of synapses. a is the time scale of the recovery variable u; b is the sensitivity of the recovery variable u to fluctuations in the domain of the membrane voltage; c is the reset value of the membrane voltage caused by rapid high-threshold potassium conductance; d is the reset value of the recovery variable caused by slow high threshold potassium and sodium conductance. a, b, c and d are dimensionless parameters, and various discharge modes of neurons can be simulated by adjusting the values of the dimensionless parameters. In summary, in the embodiment, the excitatory neuron discharge is simulated in the regular discharge mode, and the parameters are selected as follows: a is 0.02, b is 0.2, c is-65, d is 8; simulating inhibitory neuron discharge in a low-threshold discharge mode, and selecting the following parameters: a is 0.02, b is 0.25, c is-65, and d is 2.
And step S320, constructing the impulse neural network brain-like model based on the network nodes and the topological constraints of the impulse neural network brain-like model.
Illustratively, in the brain-like model constructed under the topological constraint, the Izhikevich neuron model is set as a brain-like model network node to form the impulse neural network brain-like model.
Because the Izhikevich neuron model can well reflect the discharge characteristics of the biological neurons, the selected Izhikevich neuron model can express two discharge modes of the biological neurons, and can well express excitatory neuron discharge and inhibitory neuron discharge, and the biological rationality of the constructed brain-like model is further increased from the perspective of network nodes. It can be seen that the embodiments of the present disclosure can further increase the biological reasonableness of the brain-like model based on the spiking neural network in the direction of the nodes of the brain-like model based on the topological constraint.
Fig. 4 is a schematic flow chart of constructing a brain-like model based on network nodes and topological constraints of a spiking neural network brain-like model according to an embodiment of the present disclosure. The embodiment shown in fig. 4 is extended from the embodiment shown in fig. 3, and the differences between the embodiment shown in fig. 4 and the embodiment shown in fig. 3 will be mainly described below, and the description of the same parts will not be repeated.
As shown in fig. 4, in the embodiment of the present disclosure, the step of constructing the spiking neural network brain-like model based on the network nodes and the topological constraints of the spiking neural network brain-like model includes the following steps.
And step S410, generating a network edge of the spiking neural network brain-like model based on the preset synaptic plasticity model and the S model edges.
Illustratively, the preset synaptic plasticity model is a prominent plasticity model selected for co-regulation of excitability and inhibitivity, and both excitatory synapses and inhibitive synapses regulate the spiking neural network through changes in synaptic conductance. According to the regulation rule, influence parameters of excitatory synapses and inhibitory synapses in both excitatory and inhibitory states are preset, including inversion potential, excitatory synaptic weight, inhibitory synaptic weight, decay constant of excitatory synaptic conductance, decay constant of inhibitory synaptic conductance, maximum modification value of excitatory synaptic conductance, minimum modification value of excitatory synaptic conductance, and maximum modification value of inhibitory synaptic conductance and minimum modification value of inhibitory synaptic conductance. And obtaining a synaptic plasticity model according to the parameter setting, and combining the obtained synaptic plasticity model with the S model edges to generate an edge of the brain-like model.
Illustratively, the synaptic output current is approximately linear with the input voltage, which is mathematically described as follows.
Isyn＝gsyn(t)(E-Vj(t))
Wherein, IsynIs the synaptic current; gsynIs synaptic conductance; vj(t) is the membrane potential of the postsynaptic neuron; e is the inversion potential. In this example, the inverse potential E of the excitatory synapse is selectedex0mV, reversal potential E of inhibitory synapsesinIs-70 mV. Both excitatory and inhibitory synapses effect modulation of the spiking neural network by changes in synaptic conductance, which decay exponentially when the postsynaptic neuron j does not receive an action potential of the presynaptic neuron i, as shown below.
Wherein, gexDenotes excitatory synaptic weight, ginRepresenting inhibitory synaptic weights; tau isexAnd τinThe decay constants of excitatory synaptic conductance and inhibitory synaptic conductance are indicated, respectively.
When the post-synaptic neuron j receives the action potential of the pre-synaptic neuron i, excitatory synaptic conductance and inhibitory synaptic conductance changes are as follows.
Wherein,
and
for excitatory and inhibitory conductance increments caused by action potentials, respectively, by an excitatory correction function wijAnd an inhibitory correction function mijAnd (6) carrying out adjustment. Excitability correction function wijAnd an inhibitory correction function mijThe mathematical description of (a) is as follows.
Wherein A is+And A-Maximum and minimum modification values for excitatory synaptic conductance, respectively; b is+And B-Maximum and minimum modification of inhibitory synaptic conductance, respectively. Δ t is the firing interval between pre-synaptic and post-synaptic neurons. Tau is+And τ-The time interval ranges of neuronal firing at synapse reinforcement and synapse weakening, respectively. In summary, in this embodiment, the parameters are selected as follows:
and
0.015 and 0, respectively; tau is+＝τ-＝20ms，A+＝0.1，A-＝0.105，B+＝0.02，B-＝0.003
And step S420, constructing the impulse neural network brain model based on the network edges of the impulse neural network brain model, the network nodes of the impulse neural network brain model and the topological constraint.
Exemplarily, in the constructing of the brain-like model based on the topological constraint and the network node, the synaptic plasticity model is set as an edge of the brain-like model network to form the spiking neural network brain-like model.
According to the biological synapse research result, the brain-like model regulated by the excitatory synapses and the inhibitory synapses together has biological completeness, the biological completeness of the pulse neural network brain-like model is improved from the perspective of network edges, and the biological rationality of the pulse neural network brain-like model is improved. Therefore, the biological reasonableness of the brain-like model based on the spiking neural network can be further increased in the direction of the edge of the brain-like model based on the network nodes and the topological constraint of the brain-like model based on the spiking neural network.
The specific execution flow of the spiking neural network brain model based on the network nodes and the topological constraints of the spiking neural network brain model is further described with reference to fig. 5.
Fig. 5 is a schematic flow chart illustrating a process of constructing a spiking neural network brain-like model based on nodes and topological constraints of a spiking neural network brain-like model network according to another embodiment of the present disclosure. The embodiment shown in fig. 5 is extended based on the embodiment shown in fig. 4, and the differences between the embodiment shown in fig. 5 and the embodiment shown in fig. 4 will be emphasized below, and the descriptions of the same parts will not be repeated.
As shown in fig. 5, in another embodiment of the present disclosure, the preset synaptic plasticity model includes a synaptic plasticity model adjusted by excitability and inhibitivity, and the following steps are further included before the step of constructing the spiking neural network brain-like model based on network nodes and topological constraints of the spiking neural network brain-like model.
Step S510, determining a number ratio of excitatory neurons to inhibitory neurons included in the synaptic plasticity model based on the neuro-anatomical experimental data.
Illustratively, based on the experimental data of the neurolysis planning, the ratio of excitatory neurons to inhibitory neurons can be chosen to be 4: 1, determining a model ratio of excitatory neurons and inhibitory neurons contained in the synaptic plasticity model.
Step S520, generating a preset synaptic plasticity model based on the quantity ratio.
Illustratively, in the above ratio of excitatory neurons to inhibitory neurons, the model of excitatory neurons and inhibitory neurons in the synaptic plasticity model is as follows 4: 1, and generating a preset synaptic plasticity model.
The quantity ratio of excitatory neurons and inhibitory neurons of the synaptic plasticity model is based on the experimental data of the neural gouging, so that the synaptic plasticity model has more biological rationality. Therefore, the embodiment of the disclosure can construct the spiking neural network brain-like model based on the network nodes and network topology constraints of the spiking neural network brain-like model, and provides a premise for further increasing the biological rationality of the spiking neural network brain-like model by adopting the synaptic plasticity model with more biological rationality.
Fig. 6 is a schematic flowchart illustrating a process of determining a preset network topology threshold according to an embodiment of the present disclosure. The embodiment shown in fig. 6 is extended based on the embodiment shown in fig. 2, and the differences between the embodiment shown in fig. 6 and the embodiment shown in fig. 2 will be emphasized below, and the descriptions of the same parts will not be repeated.
As shown in fig. 6, in the embodiment of the present disclosure, the step of determining the preset network topology threshold includes the following steps.
Step S610, adjusting the threshold value within a certain range, and acquiring parameters expressing network topology characteristics generated under different threshold values.
Illustratively, the range of the adjustment threshold is selected to be 0.1-0.6, the step length is set to be 0.1, and different parameters for expressing the network topology characteristics can be generated according to different thresholds. In this embodiment, parameters expressing network topology characteristics are selected as network density, average node degree, small world attribute and power law index, and different network density, average node degree, small world attribute and power law index are obtained according to different thresholds.
Illustratively, the network density is derived from a definition formula, which is shown below.
Where N represents the number of nodes of the network and L represents the number of contiguous edges actually present in the network.
Illustratively, the average node degree is obtained according to a definition formula, which is shown below.
Wherein D isiIs the degree of node i. DiThe mathematical expression of (a) is as follows.
Wherein h isijThis indicates whether or not a connection exists between the node i and the node j, and is 1 if a connection exists, or 0 if no connection exists.
Illustratively, the small world attribute is determined by σ, and when σ >1, the network is represented to have the small world attribute, whose mathematical expression is as follows.
Wherein, CrealAnd CrandomRespectively representing the clustering coefficients of the constructed network and the random network; l isrealAnd LrandomRespectively representing the shortest path lengths of the constructed network and the random network.
Step S620, determining a preset network topology threshold according to the obtained different network topology characteristic parameters.
Illustratively, according to the characteristics of the biological brain topology, the network density generally ranges from 5% to 40%, the network has a small world attribute and a scale-free attribute, the power law index is about 2, the clustering coefficient is high, and according to the characteristics of the biological brain topology, the threshold selected in this embodiment is comprehensively considered to be 0.2.
The selection of the network topology threshold value is based on the characteristics of the biological brain network topology, so the biological rationality of the impulse neural network brain model is increased.
In the embodiment of the present disclosure, the step of performing brain region division on the functional nuclear magnetic resonance image data to be processed to obtain M brain region image data includes performing brain region division on the functional nuclear magnetic resonance image data to be processed by using a Zalesky — 980 template to obtain 980 brain region image data.
The specific execution flow of the model building is further explained in conjunction with fig. 7.
Fig. 7 is a schematic flow chart of a model building method according to another embodiment of the present disclosure. The embodiment shown in fig. 7 is extended from the embodiment shown in fig. 2, and the differences between the embodiment shown in fig. 7 and the embodiment shown in fig. 2 will be mainly described below, and the description of the same parts will not be repeated.
As shown in fig. 7, in another embodiment of the present disclosure, before performing brain region segmentation on functional mri data to be processed to obtain M brain region image data, the following steps are further included.
In step S710, initial functional mri data of the subject is obtained.
Step S720, preprocessing the initial functional nuclear magnetic resonance image data to obtain to-be-processed functional nuclear magnetic resonance image data.
Illustratively, the initial functional magnetic resonance image data is preprocessed, including temporal layer correction processing and spatial normalization processing. According to practical application, the initial functional nuclear magnetic resonance image data has time shift between layers, and time layer correction processing is required. Spatial normalization is to transform the original functional mri data into a standard MNI space in order to eliminate differences in the shape and size of the brain. In this embodiment, the preprocessing of the initial functional mri image data includes not only temporal layer correction processing and spatial normalization processing, but also panning correction processing, smoothing processing, and filtering processing. During the period of acquiring the functional nuclear magnetic resonance image data, the subject inevitably has a head movement phenomenon, so the head movement correction processing is carried out to eliminate the influence of the head movement phenomenon on the image positioning; in order to reduce the influence of random noise and improve the signal-to-noise ratio of data, smoothing processing needs to be performed on functional nuclear magnetic resonance image data; in order to reduce low-frequency drift and high-frequency physiological noise, a band-pass filter is selected to filter the functional nuclear magnetic resonance image data.
Method embodiments of the present disclosure are described in detail above in conjunction with fig. 2-7, and apparatus embodiments of the present disclosure are described in detail below in conjunction with fig. 8 and 9. Furthermore, it is to be understood that the description of the method embodiments corresponds to the description of the apparatus embodiments, and therefore reference may be made to the method embodiments in the foregoing for parts that are not described in detail.
Fig. 8 is a schematic structural diagram of a model building apparatus according to an embodiment of the present disclosure. As shown in fig. 8, the model building apparatus provided by the embodiment of the present disclosure includes a brain region dividing module 810, a first generating module 820, a second generating module 830, a screening module 840, a third generating module 850, and a building module 860.
Specifically, the brain region dividing module 810 is configured to perform brain region division on the functional nuclear magnetic resonance image data to be processed, so as to obtain M brain region image data. The first generating module 820 is configured to generate M model nodes based on the M brain region image data, where the model nodes represent brain regions containing the brain region image data corresponding to the model nodes. The second generating module 830 is configured to generate N model edges based on a correlation coefficient matrix between the M model nodes, where the correlation coefficient matrix is used to represent brain function network connection strength between the M model nodes. The screening module 840 is configured to screen the N model edges based on a preset network topology threshold to obtain S model edges meeting a preset condition, where S is a positive integer less than or equal to N. The third generation module 850 is configured to generate a biological brain function network-based topological constraint of the spiking neural network brain-like model based on the M model nodes and the S model edges. The building module 860 is configured to build a spiking neural network brain-like model based on the topological constraint.
In some embodiments, the building module 860 is further configured to generate network nodes of the spiking neural network brain-like model based on a preset second-order neuron model and the M number of model nodes, where the preset second-order neuron model includes an Izhikevich neuron model; and constructing the impulse neural network brain-like model based on the network nodes and the topological constraints of the impulse neural network brain-like model.
In some embodiments, the building module 860 is further configured to generate a network edge of the spiking neural network brain-like model based on the synaptic plasticity model and the S model edges; and constructing the impulse neural network brain-like model based on the network edges of the impulse neural network brain-like model, the network nodes of the impulse neural network brain-like model and topological constraints.
In some embodiments, the building module 860 is further configured to determine a preset synaptic plasticity model. Wherein determining a preset synaptic plasticity model comprises: determining the number ratio of excitatory neurons to inhibitory neurons contained in the synaptic plasticity model based on neuroanatomy experimental data; and generating a preset synaptic plasticity model based on the quantity proportion.
In some embodiments, the screening module 840 is further configured to determine a preset network topology threshold. Specifically, the network threshold is determined according to parameters capable of characterizing network topology characteristics, wherein the parameters characterizing the network topology characteristics comprise at least one of network density, average node degree, small world attributes and non-scale attributes.
In some embodiments, the first generating module 820 is further configured to preset M to 980, and generate M image data. Wherein predetermine M and be 980, generate M image data, include: and dividing the brain region of the functional nuclear magnetic resonance image data to be processed to obtain 980 brain region image data.
In some embodiments, the first generating module 820 is further configured to generate functional mri data to be processed. Specifically, acquiring initial functional nuclear magnetic resonance image data of a subject; and preprocessing the initial functional nuclear magnetic resonance image data to obtain functional nuclear magnetic resonance image data to be processed, wherein the preprocessing comprises temporal layer correction processing and spatial standardization processing, and the preprocessing further comprises dynamic header correction processing, smoothing processing and filtering processing.
Fig. 9 is a schematic structural diagram of an electronic device according to an embodiment of the disclosure. The electronic device 900 shown in fig. 9 (the electronic device 900 may be specifically a computer device) includes a memory 901, a processor 902, a communication interface 903, and a bus 904. The memory 901, the processor 902 and the communication interface 903 are connected to each other by a bus 904.
The Memory 901 may be a Read Only Memory (ROM), a static Memory device, a dynamic Memory device, or a Random Access Memory (RAM). The memory 901 may store a program, and the processor 902 and the communication interface 903 are used to perform the steps of the model construction method of the embodiments of the present disclosure when the program stored in the memory 901 is executed by the processor 902.
The processor 902 may be a general-purpose Central Processing Unit (CPU), a microprocessor, an Application Specific Integrated Circuit (ASIC), a Graphics Processing Unit (GPU), or one or more Integrated circuits, and is configured to execute related programs to implement the functions that are required to be executed by each Unit in the model building apparatus according to the embodiment of the present disclosure.
The processor 902 may also be an integrated circuit chip having signal processing capabilities. In implementation, the steps of the model building method of the present disclosure may be performed by instructions in the form of hardware integrated logic circuits or software in the processor 902. The processor 902 may 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 device, discrete Gate or transistor logic, or discrete hardware components. The various methods, steps, and logic blocks disclosed in the embodiments of the present disclosure may be implemented or performed. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like. The steps of the method disclosed in connection with the embodiments of the present disclosure may be directly implemented by a hardware decoding processor, or implemented by a combination of hardware and software modules in the decoding processor. The software module may be located in ram, flash memory, rom, prom, or eprom, registers, etc. storage media as is well known in the art. The storage medium is located in the memory 901, and the processor 902 reads information in the memory 901, and completes, in combination with hardware thereof, functions required to be executed by units included in the model building apparatus according to the embodiment of the present disclosure, or executes a model building method according to the embodiment of the method of the present disclosure.
The communication interface 903 enables communication between the electronic device 900 and other devices or communication networks using transceiver means, such as, but not limited to, transceivers. For example, the processing functional magnetic resonance image data signal may be acquired through the communication interface 903.
Bus 904 may include a pathway to transfer information between various components of electronic device 900 (e.g., memory 901, processor 902, communication interface 903).
It should be noted that although the electronic device 900 shown in fig. 9 only shows a memory, a processor and a communication interface, in a specific implementation procedure, a person skilled in the art will understand that the electronic device 900 also comprises other components necessary for realizing normal operation. Also, those skilled in the art will appreciate that the electronic device 900 may also include hardware components that implement other additional functions, according to particular needs. Furthermore, those skilled in the art will appreciate that electronic device 900 may also include only those components necessary to implement the embodiments of the present disclosure, and need not include all of the components shown in FIG. 9.
Those of ordinary skill in the art will appreciate that the various illustrative elements and algorithm steps described in connection with the embodiments disclosed herein may be implemented as electronic hardware, or combinations of computer software and electronic hardware. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the implementation. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present disclosure.
It is clear to those skilled in the art that, for convenience and brevity of description, the specific working processes of the above-described systems, apparatuses and units may refer to the corresponding processes in the foregoing method embodiments, and are not described herein again.
In the several embodiments provided in the present disclosure, it should be understood that the disclosed system, apparatus, and method may be implemented in other ways. For example, the above-described apparatus embodiments are merely illustrative, and for example, a division of a unit is merely a logical division, and an actual implementation may have another division, for example, a plurality of units or components may be combined or integrated into another system, or some features may be omitted, or not executed. In addition, the shown or discussed mutual coupling or direct coupling or communication connection may be an indirect coupling or communication connection through some interfaces, devices or units, and may be in an electrical, mechanical or other form.
The units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of the embodiment.
In addition, functional units in the embodiments of the present disclosure may be integrated into one processing unit, or each unit may exist alone physically, or two or more units are integrated into one unit.
The functions, if implemented in the form of software functional units and sold or used as a stand-alone product, may be stored in a computer readable storage medium. Based on such understanding, the technical solution of the present disclosure may be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the method according to the embodiments of the present disclosure. And the aforementioned storage medium includes: u disk, removable hard disk, read only memory, random access memory, magnetic or optical disk, etc. for storing program codes.
The above description is only for the specific embodiments of the present disclosure, but the scope of the present disclosure is not limited thereto, and any person skilled in the art can easily conceive of the changes or substitutions within the technical scope of the present disclosure, and all the changes or substitutions should be covered within the scope of the present disclosure. Therefore, the protection scope of the present disclosure shall be subject to the protection scope of the claims.
Claims (10)
1. A model construction method is used for constructing a spiking neural network brain-like model based on biological brain topological constraint, and the method comprises the following steps:
performing brain region division on functional nuclear magnetic resonance image data to be processed to obtain M brain region image data;
generating M model nodes based on the M brain area image data, wherein the model nodes represent brain areas containing the brain area image data corresponding to the model nodes;
generating N model edges based on a correlation coefficient matrix among the M model nodes, wherein the correlation coefficient matrix is used for representing the brain function network connection strength among the M model nodes;
based on a preset network topology threshold value, screening the N model edges to obtain S model edges meeting preset conditions, wherein S is a positive integer less than or equal to N;
generating a biological brain function network-based topological constraint of the spiking neural network brain-like model based on the M model nodes and the S model edges;
and constructing the impulse neural network brain-like model based on the topological constraint.
2. The model building method according to claim 1, wherein building the spiking neural network brain-like model based on the topological constraint comprises:
generating network nodes of the spiking neural network brain-like model based on a preset second-order neuron model and the M model nodes, wherein the preset second-order neuron model comprises an Izhikevich neuron model;
and constructing the spiking neural network brain-like model based on the network nodes of the spiking neural network brain-like model and the topological constraint.
3. The model building method according to claim 2, wherein the building the spiking neural network brain-like model based on the network nodes and the topological constraints of the spiking neural network brain-like model comprises:
generating a network edge of the spiking neural network brain-like model based on a preset synaptic plasticity model and the S model edges;
and constructing the 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 the topological constraint.
4. The model building method according to claim 3, wherein the preset synaptic plasticity model comprises a synaptic plasticity model regulated jointly by excitability and inhibitivity, and further comprises, before the generating the network edge of the brain-like model based on the preset synaptic plasticity model and the S model edges:
determining the number ratio of excitatory neurons to inhibitory neurons contained in the synaptic plasticity model based on neuro-anatomical experimental data;
and generating the preset synaptic plasticity model based on the quantity proportion.
5. The model building method according to any one of claims 1 to 4, wherein the preset network topology threshold value is determined based on a parameter capable of characterizing network topology, wherein the parameter characterizing network topology includes at least one of network density, average node degree, small world attribute and non-scale attribute.
6. The model building method according to any one of claims 1 to 4, wherein M is 980, and the obtaining M brain region image data by performing brain region division on the functional nuclear magnetic resonance image data to be processed comprises:
and carrying out brain area division on the functional nuclear magnetic resonance image data to be processed by adopting a Zalesky _980 template to obtain 980 brain area image data.
7. The model building method according to any one of claims 1 to 4, wherein before the performing brain region segmentation on the functional nuclear magnetic resonance image data to be processed to obtain M brain region image data, the method further comprises:
acquiring initial functional nuclear magnetic resonance image data of a subject;
and preprocessing the initial functional nuclear magnetic resonance image data to obtain the functional nuclear magnetic resonance image data to be processed, wherein the preprocessing comprises temporal layer correction processing and spatial standardization processing, and the preprocessing further comprises head motion correction processing, smoothing processing and filtering processing.
8. A model construction device for constructing a spiking neural network brain-like model based on biological brain topological constraint, the device comprising:
the brain area dividing module is used for performing brain area division on the functional nuclear magnetic resonance image data to be processed to obtain M brain area image data;
a first generation module, configured to generate M model nodes based on the M brain region image data, where the model nodes represent brain regions including brain region image data corresponding to the model nodes;
a second generation module, configured to generate N model edges based on a correlation coefficient matrix between the M model nodes, where the correlation coefficient matrix is used to represent brain function network connection strengths between the M model nodes;
the screening module is used for screening the N model edges based on a preset network topology threshold value to obtain S model edges meeting preset conditions, wherein S is a positive integer smaller than or equal to N;
a third generation module, configured to generate a biological brain function network-based topological constraint of the spiking neural network brain-like model based on the M model nodes and the S model edges;
and the construction module is used for constructing the impulse neural network brain-like model based on the topological constraint.
9. An electronic device, comprising:
a processor;
a memory for storing the processor-executable instructions,
wherein the processor is configured to perform the model construction method of any of the preceding claims 1 to 7.
10. A computer-readable storage medium, characterized in that the storage medium stores a computer program for executing the model construction method according to any one of the preceding claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210289158.XA CN114757334A (en) | 2022-03-23 | 2022-03-23 | Model construction method and device, storage medium and electronic equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210289158.XA CN114757334A (en) | 2022-03-23 | 2022-03-23 | Model construction method and device, storage medium and electronic equipment |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114757334A true CN114757334A (en) | 2022-07-15 |
Family
ID=82327993
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210289158.XA Pending CN114757334A (en) | 2022-03-23 | 2022-03-23 | Model construction method and device, storage medium and electronic equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114757334A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2024046463A1 (en) * | 2022-09-02 | 2024-03-07 | 深圳忆海原识科技有限公司 | Model construction method, apparatus and platform, electronic device and storage medium |
| WO2024119338A1 (en) * | 2022-12-05 | 2024-06-13 | 中国科学院深圳先进技术研究院 | Knowledge- and data-driven brain network computational method and apparatus, electronic device, and storage medium |
-
2022
- 2022-03-23 CN CN202210289158.XA patent/CN114757334A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2024046463A1 (en) * | 2022-09-02 | 2024-03-07 | 深圳忆海原识科技有限公司 | Model construction method, apparatus and platform, electronic device and storage medium |
| WO2024119338A1 (en) * | 2022-12-05 | 2024-06-13 | 中国科学院深圳先进技术研究院 | Knowledge- and data-driven brain network computational method and apparatus, electronic device, and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Fu et al. | Lightweight pyramid networks for image deraining | |
| Ioannidis et al. | Semi-blind inference of topologies and dynamical processes over dynamic graphs | |
| Rozell et al. | Sparse coding via thresholding and local competition in neural circuits | |
| CN114757334A (en) | Model construction method and device, storage medium and electronic equipment | |
| Paiva et al. | A comparison of binless spike train measures | |
| Shen et al. | Nonlinear structural vector autoregressive models with application to directed brain networks | |
| Kong et al. | Image fusion technique based on non-subsampled contourlet transform and adaptive unit-fast-linking pulse-coupled neural network | |
| CN110138595A (en) | Time link prediction technique, device, equipment and the medium of dynamic weighting network | |
| CN108022171B (en) | Data processing method and equipment | |
| CN112468326A (en) | Access flow prediction method based on time convolution neural network | |
| CN114861838B (en) | Intelligent classification method for pulsatile neural brains based on neuron complex dynamics | |
| Leung et al. | On the selection of weight decay parameter for faulty networks | |
| Tang et al. | Spatio-temporal latent graph structure learning for traffic forecasting | |
| CN114998659A (en) | Image data classification method for training impulse neural network model on line along with time | |
| JP2022522807A (en) | Legendre memory unit for recurrent neural networks | |
| CN117574059A (en) | High-resolution brain-electrical-signal deep neural network compression method and brain-computer interface system | |
| CN111714124A (en) | Magnetic resonance film imaging method, device, imaging equipment and storage medium | |
| CN115358485A (en) | Traffic flow prediction method based on graph self-attention mechanism and Hox process | |
| US10643092B2 (en) | Segmenting irregular shapes in images using deep region growing with an image pyramid | |
| Henderson et al. | Spike event based learning in neural networks | |
| Mishra et al. | Chebyshev functional link artificial neural networks for denoising of image corrupted by salt and pepper noise | |
| Cipriani et al. | From NeurODEs to AutoencODEs: a mean-field control framework for width-varying neural networks | |
| CN117636626A (en) | Heterogeneous map traffic prediction method and system for strengthening road peripheral space characteristics | |
| Kong et al. | A novel ConvLSTM with multifeature fusion for financial intelligent trading | |
| Pachitariu et al. | Recurrent linear models of simultaneously-recorded neural populations |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |