A Classification Method for EEG Brain Networks Based on K-order Propagation Number and Dominance Entropy

Abstract

This invention discloses an EEG brain network classification method based on K-order propagation number and dominance entropy, comprising: collecting EEG data; constructing a brain network based on PLV phase-locked values ​​based on the EEG data; extracting important nodes of the brain network using the K-order propagation number algorithm to construct a weighted brain network; binarizing the weighted brain network and calculating the dominance entropy; constructing classification features based on the dominance entropy and performing classification using support vector machines. This invention simplifies the network structure, reduces computational costs, and introduces a novel dominance entropy attribute to brain network classification for the first time. Dominance entropy is sensitive to changes in network topology and can measure the degree of network connectivity to a certain extent. Furthermore, the classification features constructed based on dominance entropy in different frequency bands are simple and effective, improving classification speed and saving computational costs.

Description

Technical Field

[0001] This invention relates to the technical field of brain network classification, and more particularly to an EEG brain network classification method based on K-order propagation number and dominance entropy. Background Technology

[0002] Currently, the diagnosis of neurological disorders such as depression and anxiety, as well as the assessment of a person's emotional state, largely relies on methods such as scale measurements, which are easily influenced by the patient's subjectivity. EEG (Electroencephalogram), as a technique that objectively reflects brain activity, is being researched for distinguishing different brain states. Brain networks, as an application of complex network technology in neuroscience, play a crucial role in EEG classification research. By constructing brain networks and combining them with graph theory techniques to calculate network properties, specific classification features are constructed. These features are then fed into a classifier to classify the brain network. The classification results are objective and accurate, and hold promise as an auxiliary tool for doctors in diagnosing neurological diseases.

[0003] Current EEG classification techniques utilize network topology information, i.e., network attributes, for classification. This requires calculating all nodes in the network, making the method quite complex. Furthermore, the network attributes calculated are relatively traditional. The constructed classification features are also complex, increasing the computational cost of classification and reducing its speed. Summary of the Invention

[0004] The purpose of this section is to outline some aspects of embodiments of the present invention and to briefly describe some preferred embodiments. Simplifications or omissions may be made in this section, as well as in the abstract and title of this application, to avoid obscuring the purpose of these documents; however, such simplifications or omissions should not be construed as limiting the scope of the invention.

[0005] In view of the aforementioned existing problems, the present invention is proposed.

[0006] To address the aforementioned technical problems, this invention provides the following technical solution: It includes acquiring EEG brainwave data; constructing a brain network based on PLV phase-locked values ​​using the EEG brainwave data; extracting important nodes of the brain network using the K-order propagation number algorithm to form a weighted brain network; binarizing the weighted brain network and calculating the dominance entropy; constructing classification features based on the dominance entropy, and performing classification using a support vector machine.

[0007] As a preferred embodiment of the EEG brain network classification method based on K-order propagation number and dominance entropy described in this invention, the acquisition of EEG data includes: designing a working memory experimental paradigm for memorizing letter sequences, i.e., sequentially displaying three sets of randomly selected English letter sequences and requiring subjects to recall a set of letters in the corresponding order of appearance according to prompts; recruiting subjects to conduct experiments based on the working memory experimental paradigm, acquiring EEG data, i.e., EEG signals at electrode points, and defining the stage of the subject's letter memorization as the memory update state and the stage of the subject's recall of the corresponding letters as the memory retrieval state, with the time when the prompt appears as the boundary.

[0008] As a preferred embodiment of the EEG brain network classification method based on K-order propagation number and dominance entropy described in this invention, the method includes: selecting the θ and γ frequency bands to construct a brain network based on PLV phase-locked values; assuming that the EEG signals at two electrode points are x(t) and y(t), respectively, obtaining the corresponding analytic signal z through Hilbert transform. x (t), z y (t):

[0009] z x (t)=x(t)+iHT(x(t))

[0010] z y (t)=y(t)+iHT(y(t))

[0011] Calculate the analytic signal z x (t), z y The phase difference Δφ(t) is calculated based on this, and the phase lock value PLV is calculated accordingly. xy :

[0012]

[0013] Using electrode points as network nodes and the PLV value between electrode points as network edge connections, a brain network based on PLV phase-locked values ​​is constructed; where n is the total number of network nodes, t k For data points.

[0014] As a preferred embodiment of the EEG brain network classification method based on K-order propagation number and dominance entropy described in this invention, it further includes: all electrode points in the EEG data are placed according to the functional regions of the human brain.

[0015] As a preferred embodiment of the EEG brain network classification method based on K-order propagation number and dominance entropy described in this invention, the construction of a weighted brain network includes: calculating the shortest path between all pairs of network nodes based on the adjacency matrix A of the brain network, where the propagation time K takes the value of all shortest paths, i.e., K∈[0,d], and d is the network diameter; taking a network node v iThe number of network nodes that can be reached within the propagation time K is defined as the K-order propagation number. By combining the K-order propagation number with information entropy, we define the K-order structural entropy H. K :

[0016]

[0017] A comprehensive evaluation of K from 0 to d when v i The K-order structure entropy is used to calculate the network node v. i importance All network nodes are sorted in descending order of importance, and the top 8 network nodes and their connections are selected to construct the weighted brain network; where i and j represent the i-th and j-th network nodes, respectively.

[0018] As a preferred embodiment of the EEG brain network classification method based on K-order propagation number and dominance entropy described in this invention, wherein: network node v i importance include:

[0019]

[0020] Where, H = {H 0 H 1 ,...,H d} is the set of structural entropy. This represents the set of propagation counts for all network nodes.

[0021] As a preferred embodiment of the EEG brain network classification method based on K-order propagation number and dominance entropy described in this invention, the method includes: transforming a weighted brain network with 8 network nodes into an unweighted network G by binarization and removing isolated nodes; determining the minimum dominance set of the unweighted network G, where the number of network nodes is defined as γ(G); sequentially increasing the number of network nodes in the dominance set up to i, and determining whether it is a dominance set of G, to obtain the number d of dominance sets containing i network nodes. i (G); Define the total number of dominating sets γ of G. s (G); Merge the dominance set with the information entropy to calculate the dominance entropy I. dom (G):

[0022]

[0023] As a preferred embodiment of the EEG brain network classification method based on K-order propagation number and dominance entropy described in this invention, wherein: the total number of dominance sets γ of G is... s (G) includes,

[0024]

[0025] As a preferred embodiment of the EEG brain network classification method based on K-order propagation number and dominance entropy described in this invention, the construction of classification features includes concatenating the dominance entropy of the θ and γ frequency bands to form a feature vector set of memory update state and memory retrieval state.

[0026] As a preferred embodiment of the EEG brain network classification method based on K-order propagation number and dominance entropy described in this invention, the classification includes dividing the feature vector sets of memory update state and memory retrieval state into training set and validation set by ten-fold cross-validation, and then feeding them into support vector machine for classification.

[0027] The beneficial effects of this invention are as follows: 1. This invention fully evaluates the local and global topological characteristics of the network, selects important nodes of the network to construct a new brain network, and simplifies the network structure and reduces computational costs compared to the traditional method of calculating network attributes for all nodes. 2. This invention introduces a novel dominance entropy attribute to brain network classification for the first time. Dominance entropy is sensitive to changes in network topology and can measure the degree of network connectivity to a certain extent, making it a novel network topology attribute. 3. The classification features constructed based on dominance entropy in different frequency bands are simple and effective, improving classification speed and saving computational costs. 4. This invention was applied to the classification of EEG brain networks in memory update and retrieval states, achieving an accuracy of 78.27%. Attached Figure Description

[0028] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. Wherein:

[0029] Figure 1 This is a schematic diagram of the EEG brain network classification process described in the first embodiment of the present invention;

[0030] Figure 2 This refers to the classification feature set based on dominance entropy as described in the second embodiment of the present invention. Detailed Implementation

[0031] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the protection scope of the present invention.

[0032] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0033] Secondly, the term "an embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0034] This invention is described in detail with reference to the schematic diagrams. When detailing the embodiments of this invention, for ease of explanation, the cross-sectional views illustrating the device structure may be partially enlarged, not adhering to the usual scale. Furthermore, the schematic diagrams are merely examples and should not be construed as limiting the scope of protection of this invention. In actual fabrication, the three-dimensional spatial dimensions of length, width, and depth should be included.

[0035] Furthermore, in the description of this invention, it should be noted that the terms "upper," "lower," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. These terms are used solely for the convenience of describing the invention and for simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the invention. In addition, the terms "first," "second," or "third" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0036] Unless otherwise explicitly specified and limited, the terms "installation," "connection," and "joining" in this invention should be interpreted broadly. For example, they can refer to fixed connections, detachable connections, or integral connections; similarly, they can refer to mechanical connections, electrical connections, or direct connections, or indirect connections through an intermediate medium, or internal connections between two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0037] Example 1

[0038] Reference Figure 1 This is the first embodiment of the present invention, which provides an EEG brain network classification method based on K-order propagation number and dominance entropy, including:

[0039] S1: Collect EEG brainwave data.

[0040] It should be noted that EEG is a method of recording brain activity using electrophysiological indicators. During brain activity, the postsynaptic potentials of a large number of neurons occur synchronously and are summed to form the EEG. It records the changes in electrical waves during brain activity and is a comprehensive reflection of the electrophysiological activity of brain nerve cells on the cerebral cortex or scalp surface. It can be acquired using a specialized acquisition device.

[0041] Brain networks are network patterns formed by the integration of different cortical regions in the brain through structural or functional connections; they can also be formed by different electrode points establishing connections based on some linear or nonlinear statistical dependence between the signals on them.

[0042] To verify the effectiveness of this method for EEG brain network classification, this embodiment designed a working memory experiment paradigm for memorizing letter sequences. This involves sequentially displaying three sets of randomly selected English letter sequences and requiring participants to recall a set of letters in the corresponding order of appearance according to prompts. For example, in the working memory experiment paradigm, the letters az, cd, and rf appear in sequence. After the participants have finished memorizing, when prompt 2 appears, the participants need to record the letter sequence that appears the second time, i.e., cd.

[0043] Based on the working memory experimental paradigm, participants were recruited for the experiment, and EEG data, i.e., the brain signals of the electrode points, were collected. The stage when the participant memorized the letter was defined as the memory update state, and the stage when the participant recalled the corresponding letter was defined as the memory retrieval state. Based on the EEG data of these two states, the brain network classification of working memory states was completed.

[0044] S2: Construct a brain network based on PLV phase-locked values ​​using EEG data.

[0045] It should be noted that PLV (Phase Locking Value) is a functional connection index based on signal phase, and it actually measures the phase difference between two channel signals.

[0046] This embodiment obtained EEG signals from 62 electrode points. All electrode points were placed according to the functional regions of the human brain. Among the five classic EEG frequency bands δ (1-3Hz), θ (4-8Hz), α (9-13Hz), β (14-30Hz), and γ (30-60Hz), the θ and γ bands have been shown to be closely related to working memory. Therefore, this embodiment selected the θ and γ bands to construct a brain network based on PLV phase-locked values. Assuming that the EEG signals of two electrode points are x(t) and y(t), the corresponding analytic signal z is obtained through Hilbert transform. x (t), z y (t):

[0047] z x(t)=x(t)+iHT(x(t))

[0048] z y (t)=y(t)+iHT(y(t))

[0049] Calculate the analytic signal z x (t), z y The phase difference Δφ(t) is calculated based on this, and the phase lock value PLV is calculated accordingly. xy :

[0050]

[0051] In the formula, n is the total number of network nodes, n = 62, t k For data points.

[0052] A brain network based on PLV phase-locked values ​​was constructed by using 62 electrode points as network nodes and connecting the electrode points with PLV values ​​as network edges.

[0053] Among them, PLV ranges from 0 to 1. The larger the value, the stronger the phase synchronization of the two signals, that is, the tighter the connection between the two electrode points.

[0054] S3: Use the K-order propagation number algorithm to extract important nodes of the brain network and construct a weighted brain network.

[0055] It should be noted that the disease transmission process is abstracted based on the network topology, and each network node is set as an infection source. After a certain transmission time K, the number of infected nodes in the network is the Kth order transmission number.

[0056] The K-order propagation number algorithm abstracts the process of infectious disease transmission based on network topology. By designating each network node as a source of infection, and after a certain propagation time, the importance of a network node is determined based on the number of infected nodes in the network. The specific steps are as follows:

[0057] (1) Calculate the shortest path between all network node pairs based on the adjacency matrix A of the brain network. The propagation time K takes the value of all shortest paths, i.e. K∈[0,d], where d is the network diameter.

[0058] Preferably, the present invention introduces a propagation time K, the value of which increases from 0 to the network diameter, reflecting the local characteristics to the global characteristics of the network, and fully taking into account the network topology.

[0059] (2) Select a network node v i The number of network nodes that can be reached within the propagation time K is defined as the K-order propagation number.

[0060] (3) Combining the K-order propagation number with information entropy, we define the K-order structural entropy H. K :

[0061]

[0062] In the formula, i and j represent the i-th and j-th network nodes, respectively.

[0063] (4) Comprehensive evaluation of v from K to d i The K-order structure entropy is used to calculate the network node v. i importance

[0064]

[0065] Where, H = {H 0 H 1 ,...,H d} is the set of structural entropy. This represents the set of propagation counts for all network nodes.

[0066] Other network node importance algorithms can also be used to select important nodes here.

[0067] (5) Sort all network nodes in descending order of importance, and select the top 8 network nodes and their connections to construct a power brain network.

[0068] S4: Binarize the power brain network and calculate the dominance entropy.

[0069] Since dominance entropy is applicable to unweighted connected networks without isolated nodes, the weighted brain network with 8 network nodes is transformed into an unweighted network G by binarization and removal of isolated nodes. In the binarization operation, the threshold is defined as half the difference between the maximum and minimum values ​​in the network matrix plus the minimum value. Elements greater than the threshold are set to 1, and elements less than the threshold are set to 0.

[0070] The dominance entropy of a binary network G is obtained by combining the number of its dominance sets with different numbers of nodes with the information entropy. The specific steps are as follows:

[0071] (1) Determine the minimum dominance set of an unweighted network G, where the number of network nodes contained therein is defined as γ(G);

[0072] It should be noted that if all nodes V of the network are divided into two parts V1 and V2, and all network nodes in V2 are connected to at least one node in V1, then V1 is called a dominating set of the network; the minimum dominating set is the dominating set that contains the fewest network nodes.

[0073] (2) Increment the number of network nodes in the dominating set to i, and determine whether it is a dominating set of G. The number of dominating sets containing i nodes is d. i (G);

[0074] (3) Define the total number of dominating sets γ of G. s (G):

[0075]

[0076] (4) Integrate the dominance set and information entropy to calculate the dominance entropy I. dom (G):

[0077]

[0078] Information entropy is a relatively abstract concept in mathematics. It can be understood as the probability of the occurrence of a specific piece of information, and its formula is:

[0079]

[0080] In the formula, p(x) i ) is x i The probability of occurrence.

[0081] Better yet, the calculation of dominance entropy is based on the concept of dominance sets in graph theory, which fully examines the topological properties of the network and can be mapped to the interconnection, synchronization and control of various network nodes in the brain network, forming a new classification feature.

[0082] S5: Based on dominance entropy, construct classification features and use support vector machines for classification.

[0083] It should be noted that 10-fold cross-validation is used to test the accuracy of the algorithm. The dataset is divided into ten parts, and nine parts are used as training data and one part is used as test data in turn. Each test will yield the corresponding accuracy.

[0084] (1) The dominance entropy of the θ and γ bands is concatenated to form the feature vector set of memory update state and memory read state.

[0085] (2) The feature vector sets of memory update state and memory read state are divided into training set and validation set by ten-fold cross-validation and then fed into support vector machine for classification. Support vector machine is a generalized linear classifier that performs binary classification of data in a supervised learning manner.

[0086] (3) The final average classification accuracy is the classification accuracy of working memory states.

[0087] Preferably, this invention introduces a novel dominance entropy attribute into brain network classification for the first time. Dominance entropy is sensitive to changes in network topology and can measure the degree of network connectivity to a certain extent, making it a novel network topology attribute. The classification features constructed based on dominance entropy in different frequency bands are simple and effective, improving classification speed and saving computational costs.

[0088] Example 2

[0089] To verify and illustrate the technical effects of this method, this embodiment designs corresponding experiments to verify the actual effects of this method.

[0090] Experimental Design: Working memory experiment paradigm for memorizing letter sequences: Three sets of randomly selected English letter sequences are presented sequentially, and participants are asked to recall a certain set of letter sequences in the order they appear, according to prompts. For example, the experimental paradigm presents az, cd, and rf in sequence. After the participants have finished memorizing, when prompt 2 appears, the participants need to record the letter sequence that appears the second time, i.e., cd.

[0091] Based on this experimental paradigm, participants were recruited to conduct the experiment, and their EEG signals were recorded using an EEG acquisition device with 62 electrode points placed according to international standards. According to the time when the cues appeared in the experiment, the EEG signals before the cues were classified as working memory update state, and the EEG signals after the cues were classified as working memory retrieval state (only the EEG signals of correctly recalled results were analyzed).

[0092] Using the MATLAB software toolbox, the acquired EEG signals were preprocessed to remove noise. Two-state PLV brain networks were constructed using this method, with a total of 124 brain networks updated and 107 brain networks read. Using this method, a dominance entropy classification feature set was constructed using MATLAB software programming, such as... Figure 2 As shown ( Figure 2 In the diagram, the x-axis and y-axis represent two dimensions of the feature, with one point representing one classification feature; black represents features of the memory update state, and gray represents features of the read state, including 124 2D update state features and 107 2D read state features; using this method for classification, an average classification accuracy of 78.27% was obtained.

[0093] It should be recognized that embodiments of the present invention can be implemented or carried out by computer hardware, a combination of hardware and software, or by computer instructions stored in a non-transitory computer-readable storage medium. The method can be implemented using standard programming techniques—including a non-transitory computer-readable storage medium configured with a computer program, wherein such a storage medium causes the computer to operate in a specific and predefined manner—according to the methods and drawings described in the specific embodiments. Each program can be implemented in a high-level procedural or object-oriented programming language to communicate with the computer system. However, if desired, the program can be implemented in assembly or machine language. In any case, the language can be a compiled or interpreted language. Furthermore, for this purpose, the program can run on a programmed application-specific integrated circuit (ASIC).

[0094] Furthermore, the procedures described herein may be performed in any suitable order unless otherwise indicated herein or otherwise clearly contradicted by the context. The procedures described herein (or variations and / or combinations thereof) may be executed under the control of one or more computer systems configured with executable instructions, and may be implemented by hardware or a combination thereof as code (e.g., executable instructions, one or more computer programs, or one or more applications) that commonly executes on one or more processors. The computer program comprises a plurality of instructions executable by one or more processors.

[0095] Furthermore, the method can be implemented in any suitable type of computing platform, including but not limited to personal computers, minicomputers, mainframes, workstations, networked or distributed computing environments, standalone or integrated computer platforms, or in communication with charged particle tools or other imaging devices, etc. Aspects of the invention can be implemented as machine-readable code stored on a non-transitory storage medium or device, whether removable or integrated into a computing platform, such as a hard disk, optical read and / or write storage medium, RAM, ROM, etc., such that it is readable by a programmable computer, and when the storage medium or device is read by the computer, it can be used to configure and operate the computer to perform the processes described herein. Furthermore, the machine-readable code, or portions thereof, can be transmitted via wired or wireless networks. The invention described herein includes these and other different types of non-transitory computer-readable storage media when such media comprises instructions or programs that implement the steps described above in conjunction with a microprocessor or other data processor. When programmed according to the methods and techniques described herein, the invention also includes the computer itself. A computer program can be applied to input data to perform the functions described herein, thereby transforming the input data to generate output data stored in non-volatile memory. The output information can also be applied to one or more output devices such as a display. In a preferred embodiment of the invention, the converted data represents physical and tangible objects, including specific visual depictions of physical and tangible objects generated on a display.

[0096] As used herein, the terms “component,” “module,” “system,” etc., are intended to refer to a computer-related entity, which may be hardware, firmware, a combination of hardware and software, software, or running software. For example, a component may be, but is not limited to, a process running on a processor, a processor, an object, an executable file, a running thread, a program, and / or a computer. As an example, an application running on a computing device and the computing device itself can both be components. One or more components may reside in a running process and / or thread, and components may be located in a single computer and / or distributed among two or more computers. Furthermore, these components are capable of execution from various computer-readable media having various data structures thereon. These components may communicate locally and / or remotely via signals, such as based on one or more data packets (e.g., data from a component that interacts with a local system, another component in a distributed system, and / or signals that interact with other systems via a network such as the Internet).

[0097] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A classification method for EEG brain networks based on K-order propagation number and dominance entropy, characterized in that, include: Collect EEG brainwave data; A brain network based on PLV phase-locked values ​​was constructed using the EEG data. The important nodes of the brain network are extracted using the K-order propagation number algorithm, and a weighted brain network is constructed. The construction of the authoritative brain network includes: The adjacency matrix A of the brain network is used to calculate the shortest path between all pairs of network nodes. The propagation time K takes the value of all shortest paths, i.e., K∈[0,d], where d is the network diameter. Take a network node v i The number of network nodes that can be reached within the propagation time K is defined as the K-order propagation number. By combining the K-order propagation number with information entropy, we define the K-order structural entropy H. K : A comprehensive evaluation of K from 0 to d when v i The K-order structure entropy is used to calculate the network node v. i importance All network nodes are sorted in descending order of importance, and the top 8 network nodes and their connections are selected to construct the power brain network. Where i and j represent the i-th and j-th network nodes, respectively; Binarize the empowered brain network and calculate the dominance entropy; By binarizing and removing isolated nodes, the weighted brain network with 8 network nodes is transformed into an unweighted network G. The steps for calculating the dominance entropy include: Determine the minimum dominance set of an unweighted network G, where the number of network nodes contained therein is defined as γ(G); Divide all nodes V of the network into two parts V1 and V2. If all network nodes in V2 are connected to at least one node in V1, then V1 is called a dominating set of the network. The minimum dominating set is the dominating set that contains the fewest network nodes. By sequentially increasing the number of network nodes in the dominating set up to i, and determining whether each set is a dominating set of G, the number of dominating sets containing i nodes is d. i (G); Define the total number of dominating sets γ of G. s (G): The dominance set and information entropy are fused to calculate the dominance entropy I. dom (G): The formula for information entropy is: In the formula, p(x) i ) is x i The probability of occurrence; Based on the dominance entropy, classification features are constructed, and a support vector machine is used for classification.

2. The EEG brain network classification method based on K-order propagation number and dominance entropy as described in claim 1, characterized in that, Collecting EEG brainwave data includes, Design a working memory experiment paradigm for memorizing alphabetic sequences, which involves presenting three randomly selected sequences of English letters in sequence and asking participants to recall a set of letters in the corresponding order of appearance according to prompts; Based on the working memory experimental paradigm, participants were recruited for the experiment. EEG data, i.e., EEG signals at electrode points, were collected. The stage when the cue appeared was defined as the memory update state, and the stage when the participant recalled the corresponding letter was defined as the memory retrieval state.

3. The EEG brain network classification method based on K-order propagation number and dominance entropy as described in claim 2, characterized in that, include: A brain network based on PLV phase-locked values ​​is constructed by selecting the θ and γ frequency bands. Assuming that the EEG signals at two electrode points are x(t) and y(t), the corresponding analytic signal z is obtained through Hilbert transform. x (t), z y (t): z x (t)=x(t)+iHT(x(t)) z y (t)=y(t)+iHT(y(t)) Calculate the analytic signal z x (t), z y The phase difference Δφ(t) is calculated based on this, and the phase lock value PLV is calculated accordingly. xy : Using electrode points as network nodes and PLV values ​​between electrode points as network edge connections, a brain network based on PLV phase-locked values ​​is constructed. Where n is the total number of network nodes, t k For data points.

4. The EEG brain network classification method based on K-order propagation number and dominance entropy as described in claim 3, characterized in that, include: All electrode points in EEG data are placed according to the functional areas of the human brain.

5. The EEG brain network classification method based on K-order propagation number and dominance entropy as described in claim 4, characterized in that, Network node v i importance include: Where, H = {H 0 H 1 ,...,H d } is the set of structural entropy. This represents the set of propagation counts for all network nodes.

6. The EEG brain network classification method based on K-order propagation number and dominance entropy as described in claim 5, characterized in that, Constructing classification features includes: The dominance entropy of the θ and γ frequency bands is concatenated to form a feature vector set for the memory update state and the memory retrieval state.

7. The EEG brain network classification method based on K-order propagation number and dominance entropy as described in claim 6, characterized in that the classification... include: The feature vector sets of memory update state and memory read state are divided into training set and validation set by using ten-fold cross-validation, and then fed into support vector machine for classification.

Images