EEG signal emotion recognition method and system based on double-layer brain network

Abstract

The present invention provides an EEG signal emotion recognition method and system based on a double-layer brain network, including removing electrooculogram, downsampling and baseline correction of the acquired EEG signal; using wavelet packet transform to decompose the preprocessed EEG signal into frequency bands and using mutual information to construct a connectivity matrix, and using Prim algorithm to construct a brain network based on a minimum spanning tree and extract features based on the matrix; using threshold selection to construct a brain network based on threshold selection and extract features for the connectivity matrix. The features extracted from the double-layer brain network are subjected to feature fusion based on the Bayesian weighted average method to obtain a fused feature vector; the fused feature vector is input into a random forest classifier model to obtain an emotion recognition result. The present invention effectively improves the accuracy of EEG signal emotion recognition by fusing the double-layer brain network features.

Description

Technical Field

[0001] The present invention relates to the technical field of emotion recognition, and in particular to an electroencephalogram (EEG) signal emotion recognition method and system based on a double-layer brain network. Background Art

[0002] The statements in this section merely provide background art related to the present invention and do not necessarily constitute prior art.

[0003] In recent years, the rapid development of artificial intelligence (AI) technology has made it possible for computers to understand human emotions. In fields such as computer science and cognitive science, emotion recognition has become a hot topic of research. Human emotions can be identified through external facial expressions, behavioral postures, or internal physiological signals. However, the external characteristics are easily disguised and hidden, and cannot accurately reflect the most real internal emotional state. Recognizing emotions through various physiological signals is more objective and reliable. Among them, electroencephalogram (EEG) signals have the characteristics of real-time and accuracy, and are particularly concerned and used by researchers.

[0004] Emotion recognition based on EEG signals is a complex task, and it is difficult to achieve a high accuracy rate of emotion recognition through features of a single dimension. Therefore, how to extract the features that best represent the emotional state from EEG signals and how to integrate features of different dimensions become key challenges.

[0005] The inventors have found that the existing emotion recognition methods have the following technical problems:

[0006] (1) In the field of EEG emotion recognition, commonly used emotion analysis methods such as time domain, frequency domain and nonlinear are all based on extracting features on a single channel. However, the process of emotion generation is accompanied by the transmission of information between different brain regions in the brain. The single-channel-based method ignores the information interaction between brain regions. Relevant research in cognitive neuroscience has found that the expression of emotions has structural and functional connections between multiple brain regions. By transmitting and integrating information from different brain regions, the current emotional state is reflected in facial expressions and behaviors. How to efficiently and accurately measure the degree of correlation between EEG signals between different brain regions is a major technical difficulty.

[0007] (2) The single-dimensional features extracted from EEG signals contain less emotion-related information. By fusing features of different dimensions, EEG signals can be represented more comprehensively. However, how to adopt a feature fusion strategy to maximize the accuracy of EEG signal emotion recognition is a technical difficulty that needs to be solved. Summary of the invention

[0008] In order to address the deficiencies in the prior art, the present invention provides an EEG signal emotion recognition method and system based on a double-layer brain network. The scheme constructs a double-layer brain network by adopting a minimum spanning tree and threshold selection, and fuses the features of the double-layer brain network based on the Bayesian weighted averaging method, and performs emotion recognition based on the fused features, thereby effectively improving the accuracy of emotion recognition.

[0009] In order to achieve the above object, the present invention adopts the following technical solution:

[0010] According to a first aspect of an embodiment of the present invention, there is provided a method for recognizing emotions from electroencephalogram signals based on a double-layer brain network, comprising:

[0011] Acquire EEG signals and perform corresponding preprocessing;

[0012] The preprocessed EEG signals were decomposed by wavelet packet transform and the connectivity matrix was constructed by mutual information.

[0013] Based on the connectivity matrix, respectively constructing a brain network based on a minimum spanning tree and a brain network based on threshold selection, and performing feature extraction respectively;

[0014] Perform feature fusion on the features extracted from the double-layer brain network to obtain a fused feature vector;

[0015] The fused feature vector is input into the pre-trained emotion recognition model to obtain the emotion recognition result.

[0016] As an optional implementation, the construction of a brain network based on a minimum spanning tree adopts Prim's algorithm.

[0017] As an optional implementation, the feature fusion method adopts a Bayesian weighted average-based method.

[0018] As an optional implementation, the emotion recognition model adopts a random forest classifier model.

[0019] As a further limitation, the preprocessing includes but is not limited to electrooculogram removal, downsampling and baseline correction.

[0020] According to a second aspect of an embodiment of the present invention, there is provided an EEG signal emotion recognition system based on a double-layer brain network, comprising:

[0021] The data acquisition module is configured to: acquire EEG signals and perform corresponding preprocessing;

[0022] The connectivity matrix construction module is configured to: perform frequency band decomposition on the preprocessed EEG signal using wavelet packet transform and construct a connectivity matrix using mutual information;

[0023] A feature extraction module is configured to: construct a minimum spanning tree-based brain network and a threshold-selected brain network based on the connectivity matrix, and perform feature extraction on each of them;

[0024] The feature fusion module is configured to: perform feature fusion on the features extracted from the double-layer brain network to obtain a fused feature vector;

[0025] The emotion recognition module is configured to: input the fused feature vector into a pre-trained emotion recognition model to obtain an emotion recognition result.

[0026] According to a third aspect of an embodiment of the present invention, a computer-readable storage medium is provided, on which a program is stored, and when the program is executed by a processor, the steps in the method for emotion recognition of electroencephalogram signals based on a double-layer brain network as described above are implemented.

[0027] According to a fourth aspect of an embodiment of the present invention, there is provided an electronic device, comprising a memory, a processor, and a program stored in the memory and executable on the processor, wherein when the processor executes the program, the steps in the method for emotion recognition of electroencephalogram signals based on a double-layer brain network as described above are implemented.

[0028] Compared with the prior art, the present invention has the following beneficial effects:

[0029] (1) The scheme described in the present invention respectively constructs a two-layer brain network using a minimum spanning tree and threshold selection, and fuses the features of the two-layer brain network based on the Bayesian weighted average method, and proposes an EEG signal emotion recognition method based on a two-layer brain network. Five features, including leaf score, intermediate coreness, tree hierarchy, feature path length and degree, are automatically extracted from the brain network based on the minimum spanning tree, and three features, including global clustering coefficient, local clustering coefficient and global efficiency, are automatically extracted from the brain network based on threshold selection. In view of the problem that direct feature fusion has little effect on the improvement of EEG signal emotion recognition, a fusion method based on Bayesian weighted average is adopted, which can balance the contribution of different features and improve the accuracy of emotion recognition after fusion.

[0030] (2) The scheme described in the present invention removes eye contact, downsamples and corrects the baseline of the obtained EEG signal, which can effectively remove the influence of noise in the EEG signal; the brain network feature extraction based on the minimum spanning tree is to construct a connectivity matrix that characterizes the relationship between channels of the preprocessed EEG signal using mutual information, and then use the Prim algorithm to obtain the brain network based on the minimum spanning tree and extract five features; the brain network feature extraction based on threshold selection is to construct a brain network based on threshold selection using threshold selection on the connectivity matrix and extract three features; the feature fusion part is to fuse the eight features extracted from the double-layer brain network based on the Bayesian weighted average method to obtain a fused feature vector; the emotion recognition part is to input the fused feature vector into the random forest classifier model to obtain the emotion recognition result, thereby improving the accuracy of emotion recognition. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] The accompanying drawings in the specification, which constitute a part of the present invention, are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute improper limitations on the present invention.

[0032] Figure 1 This is a flow chart of the emotion recognition method based on a double-layer brain network provided in Embodiment 1 of the present invention;

[0033] Figure 2 A schematic diagram of a six-layer wavelet packet transform provided in Embodiment 1 of the present invention;

[0034] Figure 3 A schematic diagram of a connectivity matrix provided in Embodiment 1 of the present invention;

[0035] Figure 4 This is a schematic diagram of a brain network based on a minimum spanning tree provided in Embodiment 1 of the present invention;

[0036] Figure 5 This is a schematic diagram of a brain network based on threshold selection provided in Example 1 of the present invention. DETAILED DESCRIPTION

[0037] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0038] It should be noted that the following detailed descriptions are all illustrative and intended to provide further explanation of the present invention. Unless otherwise specified, all technical and scientific terms used herein have the same meanings as those commonly understood by those skilled in the art to which the present invention belongs.

[0039] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit exemplary embodiments according to the present invention. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprising" and / or "including" are used in this specification, it indicates the presence of features, steps, operations, devices, components and / or combinations thereof.

[0040] In the absence of conflict, the embodiments of the present invention and the features of the embodiments may be combined with each other.

[0041] Embodiment 1:

[0042] The purpose of this embodiment is to provide an EEG signal emotion recognition method based on a double-layer brain network.

[0043] like Figure 1 As shown, a method for emotion recognition of EEG signals based on a double-layer brain network includes:

[0044] Acquire EEG signals and perform corresponding preprocessing;

[0045] The preprocessed EEG signals were decomposed by wavelet packet transform and the connectivity matrix was constructed by mutual information.

[0046] Based on the connectivity matrix, respectively constructing a brain network based on a minimum spanning tree and a brain network based on threshold selection, and performing feature extraction respectively;

[0047] Perform feature fusion on the features extracted from the double-layer brain network to obtain a fused feature vector;

[0048] The fused feature vector is input into the pre-trained emotion recognition model to obtain the emotion recognition result.

[0049] As an optional implementation, the construction of a brain network based on a minimum spanning tree adopts Prim's algorithm.

[0050] As an optional implementation, the feature fusion method adopts a Bayesian weighted average-based method.

[0051] As an optional implementation, the emotion recognition model adopts a random forest classifier model.

[0052] As a further limitation, the preprocessing includes but is not limited to electrooculogram removal, downsampling and baseline correction.

[0053] Specifically, the method includes:

[0054] S1: Remove the oculomotor activity from the original EEG signal, and perform downsampling and baseline correction;

[0055] S2: The preprocessed EEG signal is decomposed into four frequency bands: Theta, Alpha, Beta and Gamma by wavelet packet transform, and the connectivity matrix is ​​constructed by mutual information in each frequency band. In order to remove redundant information, the Prim algorithm is used to construct a brain network based on the minimum spanning tree on the basis of the connectivity matrix, and five features are extracted: leaf score, intermediate coreness, tree level, characteristic path length and degree, to form the feature vector x 1 ;

[0056] S3: The connectivity matrix is ​​processed by the threshold selection method to obtain the brain network based on threshold selection, and three features, namely the global clustering coefficient, the local clustering coefficient and the global efficiency, are extracted to form the feature vector x 2 ;

[0057] S4: The feature vector x is fused using the Bayesian weighted average method 1 With the eigenvector x 2 Perform feature fusion and obtain feature vector x after fusion;

[0058] S5: Input the feature vector x into the random forest classification model to obtain the emotion recognition result.

[0059] In S1, the public EEG dataset DEAP is used in this example. The DEAP dataset collects 32 channels of EEG signals and 8 channels of other physiological signals from 32 subjects. Each subject conducted 40 experiments, and the data recording time for each experiment was the same, all 63 seconds. It takes a certain amount of time to induce emotions under video stimulation, so intercepting the data segment in the middle of the EEG signal can contain more emotion-related information, thereby improving the effect of emotion recognition. The EEG signals from 24 to 43 seconds are intercepted, and the intercepted data is subtracted from the resting state data of the previous three seconds to achieve baseline correction.

[0060] In S2: The preprocessed EEG signal is transformed by wavelet packet transform and decomposed into four frequency bands: Theta, Alpha, Beta and Gamma. For the downsampled 128Hz EEG signal, the Nyquist sampling frequency is 64Hz according to the sampling theorem, as shown in Figure 2 As shown in , a 6-layer wavelet packet transform is used to decompose it. The relationship between EEG channels is calculated through mutual information to obtain a 32×32 connectivity matrix. Part of the information intercepted in the matrix is ​​shown in Figure 3 As shown. The mutual information is calculated by the following formula:

[0061]

[0062] Among them, MI XY represents the mutual information between EEG channel X and channel Y, Pij represents the joint probability density, P i represents the probability density.

[0063] In order to extract the key information in the connectivity matrix, the Prim algorithm was used to process the 32×32 connectivity matrix, retaining the 31 maximum weights while ensuring that all channel nodes have edges connected to them. The 32 channels were used as nodes of the brain network, and the 31 maximum weights were used as edges of the brain network to construct a brain network based on the minimum spanning tree. Figure 4 As shown in Figure 1, the brain network based on the minimum spanning tree is visualized. In order to measure the topological relationship of the brain network, five features, leaf score, intermediate coreness, tree level, characteristic path length and degree, are extracted based on the brain network based on the minimum spanning tree to obtain the feature vector x 1 The calculation formulas for the five features are as follows:

[0064] (1) Leaf Fraction (LF): The number of nodes with degree 1 divided by the total number of nodes.

[0065]

[0066] Among them, F represents the number of nodes with degree 1, and N represents the total number of nodes.

[0067] (2) Betweenness Centrality (BC): The number of paths passing through node i, excluding paths with node i as the start and end node.

[0068]

[0069] Among them, ρ ab (i) represents the number of shortest paths from node a to b that passes through node i, ρ ab Represents the number of shortest paths from node a to b.

[0070] (3) Tree Level (TL): measures the relationship between shortening the diameter and preventing central nodes from being overloaded.

[0071]

[0072] Among them, F represents the number of nodes with degree 1, M represents the total number of all connections, and BC max Indicates the maximum value of betweenness coreness.

[0073] (4) Characteristic Path Length (CPL): measures the information transmission capacity of the brain network.

[0074]

[0075] Among them, n represents the number of nodes, i and j represent different nodes, and d ij Represents the distance between nodes i and j in the network.

[0076] (5) Degree (DEG): the number of edges connected to a node i.

[0077] DEG=∑ j∈N w ij (6)

[0078] In S3: For the 32×32 connectivity matrix, the threshold selection method is used to process it. If the weight in the connectivity matrix is ​​greater than the threshold, the weight is set to 1, otherwise it is set to 0. A value of 1 represents that there is a connection between the corresponding two EEG channels, and 0 represents no connection. The 32 EEG channels are used as nodes of the brain network, and the weights are used as edges of the brain network to obtain a brain network based on threshold selection. Figure 5 As shown in Figure 1, the brain network based on threshold selection is visualized. In order to characterize the topological structure of the brain network, three features, namely global clustering coefficient, local clustering coefficient and global efficiency, are extracted based on the threshold-based selection of the brain network, and the feature vector x is obtained. 2 The calculation formulas for the three features are as follows:

[0079] (1) Global Clustering Coefficient (GCC): measures the degree of clustering of nodes in the brain network.

[0080]

[0081] Where n represents the number of nodes, t i Indicates the number of triangle structures near the i node, DEG o Indicates the degree of the i-node.

[0082] (2) Local Clustering Coefficient (LCC): measures the degree of clustering between a node and its neighboring nodes.

[0083]

[0084] (3) Global Efficiency (GE): measures the ability of brain networks to transmit and process information.

[0085]

[0086] Among them, d ijRepresents the distance between nodes i and j in the network.

[0087] In S4: the feature vector x is fused using the Bayesian weighted average method 1 With the eigenvector x 2 Perform feature fusion and obtain the feature vector x. The weight calculation formula is as follows:

[0088]

[0089] Among them, λ i represents the weight of feature i, p i represents the posterior probability of feature i, p MST represents the posterior probability of the feature based on the minimum spanning tree brain network, p TS represents the posterior probability of the feature for selecting brain networks based on the threshold.

[0090] In S5: the feature vector x after Bayesian fusion is input into the random forest classification model to obtain the result of emotion recognition.

[0091] Embodiment 2:

[0092] The purpose of this embodiment is to provide an EEG signal emotion recognition system based on a double-layer brain network.

[0093] An EEG signal emotion recognition system based on a double-layer brain network, comprising:

[0094] The data acquisition module is configured to: acquire EEG signals and perform corresponding preprocessing;

[0095] The connectivity matrix construction module is configured to: perform frequency band decomposition on the preprocessed EEG signal using wavelet packet transform and construct a connectivity matrix using mutual information;

[0096] A feature extraction module is configured to: construct a minimum spanning tree-based brain network and a threshold-selected brain network based on the connectivity matrix, and perform feature extraction on each of them;

[0097] The feature fusion module is configured to: perform feature fusion on the features extracted from the double-layer brain network to obtain a fused feature vector;

[0098] The emotion recognition module is configured to: input the fused feature vector into a pre-trained emotion recognition model to obtain an emotion recognition result.

[0099] Furthermore, the functional implementation of each module in the system is the same as the EEG signal emotion recognition method based on a double-layer brain network provided in Example 1, so it will not be repeated here.

[0100] Embodiment three:

[0101] A computer-readable storage medium is provided in the third embodiment of the present invention, on which a program is stored. When the program is executed by a processor, the steps in the method for emotion recognition of electroencephalogram signals based on a double-layer brain network as described in the first embodiment of the present invention are implemented.

[0102] Embodiment 4:

[0103] In a fourth embodiment of the present invention, an electronic device is provided, comprising a memory, a processor, and a program stored in the memory and executable on the processor. When the processor executes the program, the steps in the method for emotion recognition of electroencephalogram signals based on a double-layer brain network as described in the first embodiment of the present invention are implemented.

[0104] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of hardware embodiments, software embodiments, or embodiments combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage and optical storage, etc.) containing computer-usable program code.

[0105] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0106] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0107] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0108] A person skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program, and the program can be stored in a computer-readable storage medium, and when the program is executed, it can include the processes of the embodiments of the above-mentioned methods. The storage medium can be a disk, an optical disk, a read-only memory (ROM) or a random access memory (RAM), etc.

[0109] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for emotion recognition of EEG signals based on a double-layer brain network, characterized in that: include: Acquire EEG signals and perform corresponding preprocessing; The preprocessed EEG signals were decomposed by wavelet packet transform and the connectivity matrix was constructed by mutual information. The method of using wavelet packet transform to perform frequency band decomposition and using mutual information to construct a connectivity matrix is ​​as follows: using wavelet packet transform to decompose the preprocessed EEG signal into four frequency bands: Theta, Alpha, Beta and Gamma, and using mutual information to construct a connectivity matrix in each frequency band; Based on the connectivity matrix, respectively constructing a brain network based on a minimum spanning tree and a brain network based on threshold selection, and performing feature extraction respectively; Brain network extraction features based on minimum spanning tree, including but not limited to leaf score, betweenness coreness, tree hierarchy, characteristic path length and degree; Brain network extraction features based on threshold selection, including but not limited to global clustering coefficient, local clustering coefficient and global efficiency; The leaf fraction refers to the number of nodes with degree 1 divided by the total number of nodes. Where F represents the number of nodes with degree 1, N represents the total number of nodes, and M represents the total number of all connections; the intermediate coreness refers to the number of paths passing through node i, excluding the paths with node i as the starting and ending nodes. Among them, ρ ab (i) represents the number of shortest paths from node a to b that passes through node i, ρ ab represents the number of shortest paths from node a to node b; the level of the tree refers to the relationship between shortening the diameter and preventing the central node from being overloaded, Among them, F represents the number of nodes with degree 1, M represents the total number of all connections, and BC max represents the maximum value of the mediating coreness; the characteristic path length is used to measure the information transmission ability of the brain network. Among them, n represents the number of nodes, i and j represent different nodes, and d ij represents the distance between nodes i and j in the network; the degree refers to the number of edges connected to a node i, DEG = ∑ j∈N w ij ; Perform feature fusion on the features extracted from the double-layer brain network to obtain a fused feature vector; The fused feature vector is input into the pre-trained emotion recognition model to obtain the emotion recognition result.

2. The method for emotion recognition based on EEG signals of a double-layer brain network as claimed in claim 1, characterized in that: The brain network based on the minimum spanning tree is constructed using the Prim algorithm.

3. The method for emotion recognition based on EEG signals of a double-layer brain network as claimed in claim 1, characterized in that: The feature fusion method adopts a Bayesian weighted average method.

4. The method for emotion recognition based on EEG signals of a double-layer brain network as claimed in claim 1, characterized in that: The emotion recognition model adopts a random forest classifier model.

5. The method for emotion recognition based on EEG signals of a double-layer brain network as claimed in claim 1, characterized in that: The preprocessing includes but is not limited to electrooculogram removal, downsampling and baseline correction.

6. An EEG signal emotion recognition system based on a double-layer brain network, characterized in that: include: The data acquisition module is configured to: acquire EEG signals and perform corresponding preprocessing; The connectivity matrix construction module is configured to: perform frequency band decomposition on the preprocessed EEG signal using wavelet packet transform, and construct a connectivity matrix using mutual information; the frequency band decomposition using wavelet packet transform, and constructing a connectivity matrix using mutual information, specifically: perform frequency band decomposition on the preprocessed EEG signal using wavelet packet transform, decompose it into four frequency bands of Theta, Alpha, Beta and Gamma, and construct a connectivity matrix using mutual information on each frequency band; A feature extraction module is configured to: construct a minimum spanning tree-based brain network and a threshold-selected brain network based on the connectivity matrix, and perform feature extraction on each of them; Brain network extraction features based on minimum spanning tree, including but not limited to leaf score, betweenness coreness, tree hierarchy, characteristic path length and degree; Brain network extraction features based on threshold selection, including but not limited to global clustering coefficient, local clustering coefficient and global efficiency; The leaf fraction refers to the number of nodes with degree 1 divided by the total number of nodes. Where F represents the number of nodes with degree 1, N represents the total number of nodes, and M represents the total number of all connections; the intermediate coreness refers to the number of paths passing through node i, excluding the paths with node i as the starting and ending nodes. Among them, ρ ab (i) represents the number of shortest paths from node a to b that passes through node i, ρ ab represents the number of shortest paths from node a to node b; the level of the tree refers to the relationship between shortening the diameter and preventing the central node from being overloaded, Among them, F represents the number of nodes with degree 1, M represents the total number of all connections, and BC mac represents the maximum value of the mediating coreness; the characteristic path length is used to measure the information transmission ability of the brain network. Among them, n represents the number of nodes, i and j represent different nodes, and d ij represents the distance between nodes i and j in the network; the degree refers to the number of edges connected to a node i, DEG = ∑ j∈N w ij ; The feature fusion module is configured to: perform feature fusion on the features extracted from the double-layer brain network to obtain a fused feature vector; The emotion recognition module is configured to: input the fused feature vector into a pre-trained emotion recognition model to obtain an emotion recognition result.

7. A computer-readable storage medium having a program stored thereon, characterized in that: When the program is executed by a processor, the steps in the method for recognizing emotions of EEG signals based on a double-layer brain network as described in any one of claims 1 to 5 are implemented.

8. An electronic device comprising a memory, a processor, and a program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the steps in the EEG signal emotion recognition method based on a double-layer brain network as described in any one of claims 1 to 5 are implemented.

Images