Electroencephalogram signal processing method and device based on brain function complex network and electronic equipment

Abstract

The invention provides an electroencephalogram signal processing method and device based on a brain function complex network and electronic equipment, and relates to the technical field of electroencephalogram signal processing.The method comprises the steps that electroencephalogram data of a sub-brain concussion group and a healthy control group are obtained, and the electroencephalogram data are preprocessed; separating the preprocessed electroencephalogram data according to an attention sub-network, respectively calculating phase lock values of the electroencephalogram data under different frequency bands, and generating a sub-network function connection matrix; based on a head model and a brain atlas, mapping the electroencephalogram data to a source space through a beam forming algorithm, and generating a whole brain function connection matrix; extracting a global index through the whole brain function connection matrix, and extracting a node index through the sub-network function connection matrix; and establishing a classification network model, fusing the global indexes and the node indexes, inputting the fused global indexes and node indexes into the trained classification network model, and outputting classification results of the sub-brain concussion group and the health control group. Compared with a traditional mode, the accuracy and generalization ability of the classification network model are greatly improved.

Description

Technical Field

[0001] The present invention relates to the technical field of electroencephalogram (EEG) signal processing, and in particular, to an EEG signal processing method, device and electronic device based on a brain functional complex network. Background Art

[0002] Subconcussion refers to a head impact that does not cause a known or clinically diagnosed concussion. This phenomenon can also occur during rapid acceleration and deceleration of the body or trunk, especially when the brain moves freely within the cranial cavity, resulting in a "drift" phenomenon. The main impact of subconcussion stems from its repeated occurrence, and cumulative exposure may lead to harmful consequences. Increasing evidence suggests that repetitive subconcussion is prevalent in contact sports and explosive impacts and may trigger a series of subacute and chronic complications.

[0003] However, in the prior art, the research on the neural mechanism of subconcussion mainly relies on clinical symptom observation, neuropsychological scale assessment and conventional electroencephalogram analysis, generally using single-modal features, which are difficult to effectively integrate global topological attributes and node-level functional characteristics, resulting in insufficient accuracy and generalization ability for the identification of the neural mechanism of subconcussion. Summary of the Invention

[0004] The purpose of the present invention is to provide an EEG signal processing method, device and electronic device based on a brain functional complex network to solve the problem of insufficient accuracy and generalization ability for the identification of the neural mechanism of subconcussion in the prior art mentioned in the above background art.

[0005] To achieve the above purpose, the present invention provides the following technical solution: An EEG signal processing method based on a brain functional complex network, the steps include: acquiring EEG data of a subconcussion group and a healthy control group, and preprocessing the EEG data; separating the preprocessed EEG data according to attention subnets, and respectively calculating the phase locking value of the EEG data in different frequency bands to generate a subnet functional connectivity matrix; based on a head model and a brain atlas, mapping the EEG data to the source space through a beamforming algorithm to generate a whole-brain functional connectivity matrix; extracting global indexes through the whole-brain functional connectivity matrix, and extracting node indexes through the subnet functional connectivity matrix; establishing a classification network model, fusing the global indexes and the node indexes through graph theory and inputting them into the trained classification network model, and outputting the classification results of the subconcussion group and the healthy control group.

[0006] Optionally, the steps of obtaining the electroencephalogram (EEG) data of the sub-concussion group and the healthy control group and preprocessing the EEG data specifically include: aligning the physical positions of the EEG electrodes with the standard brain region template for channel localization, and obtaining the EEG data of the sub-concussion group and the healthy control group; performing 0.5 Hz - 30 Hz band-pass filtering and 50 Hz notch filtering, downsampling, whole-brain average reference conversion, segmentation and baseline correction, bad segment deletion, removal of electrooculogram and electromyogram components, and removal of overly large amplitude values on the collected EEG data.

[0007] Optionally, the segmentation step specifically includes: under different stimulus cue conditions, taking the stimulus cue as the origin, intercepting the time period from 500 ms before the stimulus to 500 ms after the stimulus; under different stimulus target conditions, taking the target stimulus as the origin, intercepting the time period from 1000 ms before the stimulus to 1000 ms after the stimulus; the baseline correction step specifically includes: under different stimulus cue conditions, performing baseline correction on the 500 ms before the stimulus; under different stimulus target conditions, performing baseline correction on the time period from 1000 ms before the stimulus to 500 ms before the stimulus.

[0008] Optionally, the step of separating the preprocessed EEG data according to the attention sub-network specifically includes: separating the preprocessed EEG data according to the alert network, the orienting network, and the executive control network.

[0009] Optionally, the step of separately calculating the phase locking value of the EEG data in different frequency bands specifically includes: separately calculating the phase locking value of the EEG data in the δ (1 Hz - 4 Hz), θ (4 Hz - 8 Hz), α (8 Hz - 13 Hz), and β (13 Hz - 30 Hz) frequency bands, and its calculation formula is: where PLV is the phase locking value, N is the number of time points, e is the base of the natural logarithm, i is the imaginary unit, and ΔΦ t is the phase difference at each time point.

[0010] Optionally, the step of mapping the EEG data to the source space through the beamforming algorithm based on the head model and the brain atlas specifically includes: constructing a BEM head model and an AAL-116 brain atlas based on the MNI standard space, combining the brain electrical signal position information of the EEG electrodes to construct a forward model, using the beamforming algorithm to perform preliminary localization on the brain electrical signals, quantifying the functional connection strength between different brain regions according to the localization results using the phase locking value, and aggregating the brain electrical signals to the 116 standard brain regions defined by the AAL-116 atlas through the nearest neighbor interpolation method.

[0011] Optionally, the global indicators include clustering coefficient, network average shortest path length, global efficiency, and local efficiency; the node indicators include node degree, betweenness centrality, node efficiency, node clustering coefficient, and node local efficiency.

[0012] Optionally, the step of fusing the global index and the node index through graph theory and inputting them into the trained classification network model specifically includes: combining the global index and the node index of all bands to obtain a target feature vector, and inputting the target feature vector into the classification network model.

[0013] On the other hand, the present invention also provides an electroencephalogram (EEG) signal processing device based on a brain functional complex network, including: an acquisition module, which acquires EEG data of a sub-concussion group and a healthy control group and preprocesses the EEG data; a small matrix generation module, which is used to separate the preprocessed EEG data according to an attention sub-network, calculate the phase locking value of the EEG data in different frequency bands respectively, and generate a sub-network functional connection matrix; a large matrix generation module, which is used to map the EEG data to the source space through a beamforming algorithm based on a head model and a brain atlas to generate a whole-brain functional connection matrix; a feature extraction module, which is used to extract a global index through the whole-brain functional connection matrix and extract a node index through the sub-network functional connection matrix; a classification module, which is used to establish a classification network model, fuse the global index and the node index through graph theory and input them into the trained classification network model, and output the classification results of the sub-concussion group and the healthy control group.

[0014] On the other hand, the present invention also provides an electronic device, including a memory and a processor, where the memory stores a computer program, and when the processor executes the computer program, the steps of the above-mentioned EEG signal processing method based on a brain functional complex network are implemented.

[0015] Compared with the prior art, the beneficial effects of the present invention are:

[0016] By separating the EEG data of the three major attention sub-networks of the alert network, the orienting network, and the executive control network, and combining the analysis of multiple frequency bands of δ / θ / α / β, a multi-dimensional analysis of sub-concussion attention network damage is realized; compared with the traditional single-frequency band or single-network analysis method, it can more comprehensively capture the abnormal neural oscillation patterns caused by repetitive head impacts, providing a new perspective for revealing the neurophysiological mechanism of sub-concussion.

[0017] Adopting the beamforming algorithm in combination with the BEM head model and the AAL-116 brain atlas, the scalp electrophysiological signals are accurately mapped to the source space, and the spatial resolution reaches 116 brain regions, which can accurately locate the brain regions and effectively overcome the interference of the volume conduction effect.

[0018] By integrating global indicators and node indicators through graph theory, a multi-scale network feature set is constructed. Compared with single-indicator analysis, it can not only characterize the changes in the small-world properties of the overall network but also locate key brain regions, comprehensively analyze the neural network features of sub-concussion, and improve the robustness of model classification.

[0019] By integrating multi-dimensional graph theory features through a classification network model, a "electrophysiological signal - graph theory network - machine learning" trinity analysis framework is established, providing a new paradigm for the research of neural markers of mild brain injury, promoting the leap of concussion research from the behavioral level to the neural network mechanism level, realizing automatic classification of sub-concussion, and greatly improving the accuracy and generalization ability of the classification network model compared with traditional methods. Brief Description of the Drawings

[0020] Figure 1 It is a schematic flowchart of the method steps of the present invention.

[0021] Figure 2 It is a schematic structural diagram of the device of the present invention.

[0022] In the figure: 10 - acquisition module, 20 - small matrix generation module, 30 - large matrix generation module, 40 - feature extraction module, 50 - classification module. Detailed Embodiments

[0023] Next, the solutions of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all embodiments.

[0024] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and do not necessarily need to describe a specific order or sequence. It should be understood that such data can be interchanged under appropriate circumstances for the embodiments of the present application described here. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or units does not necessarily have to be limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products, or devices.

[0025] Those skilled in the art of the present technology can understand that, unless specifically stated, the singular forms "a", "an", "the" and "said" used herein may also include the plural forms. It should be further understood that the term "comprising" used in the specification of the present application means the presence of features, integers, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or their groups. It should be understood that when we say that an element is "connected" or "coupled" to another element, it can be directly connected or coupled to other elements, or there may also be intermediate elements. In addition, the "connection" or "coupling" used herein may include wireless connection or wireless coupling. The phrase "and / or" used herein includes all or any unit and all combinations of one or more associated listed items.

[0026] Those skilled in the art of the present technology can understand that, unless otherwise defined, all terms used herein (including technical terms and scientific terms) have the same meaning as the general understanding of those of ordinary skill in the art to which the present application belongs. It should also be understood that terms such as those defined in a general dictionary should be understood to have a meaning consistent with the meaning in the context of the prior art, and will not be interpreted with an idealized or overly formal meaning unless specifically defined as here.

[0027] It should be understood that the sequence numbers and magnitudes of the steps in this embodiment do not imply the sequence of execution. The execution sequence of each process is determined by its function and internal logic, and should not constitute any limitation to the implementation process of the embodiments of the present application.

[0028] It should be noted that, without conflict, the embodiments in the present application and the features in the embodiments can be combined with each other. The present application will be described in detail below with reference to the drawings and in combination with the embodiments.

[0029] Please refer to Figure 1 , a method for processing electroencephalogram signals based on a complex network of brain functions according to the present invention, the steps include: acquiring electroencephalogram data of a sub-concussion group and a healthy control group, and preprocessing the electroencephalogram data; separating the preprocessed electroencephalogram data according to an attention sub-network, and respectively calculating the phase locking values of the electroencephalogram data in different frequency bands to generate a sub-network functional connectivity matrix; based on a head model and a brain atlas, mapping the electroencephalogram data to the source space through a beamforming algorithm to generate a whole-brain functional connectivity matrix; extracting global indicators through the whole-brain functional connectivity matrix, and extracting node indicators through the sub-network functional connectivity matrix; establishing a classification network model, fusing the global indicators and the node indicators through graph theory and inputting them into the trained classification network model, and outputting the classification results of the sub-concussion group and the healthy control group.

[0030] Specifically, the present application analyzes the electroencephalogram (EEG) data of the three major attention sub-networks, namely the alerting network, the orienting network, and the executive control network, and combines the multi-bands of δ / θ / α / β for analysis, achieving a multi-dimensional analysis of sub-concussive attention network damage. Compared with the traditional single-band or single-network analysis methods, it can capture more comprehensively the abnormal neural oscillation patterns caused by repetitive head impacts, providing a new perspective for revealing the neurophysiological mechanisms of sub-concussions. By using the beamforming algorithm in combination with the boundary element method (BEM) head model and the automated anatomical labeling (AAL)-116 brain atlas, the scalp electrophysiological signals are accurately mapped to the source space with a spatial resolution of 116 brain regions, enabling accurate localization of brain regions and effectively overcoming the interference of volume conduction effects. By integrating global metrics and node metrics through graph theory, a multi-scale network feature set is constructed. Compared with single-metric analysis, it can not only characterize the changes in the small-world properties of the overall network but also locate key brain regions, comprehensively analyzing the neural network characteristics of sub-concussions and enhancing the robustness of model classification. By fusing multi-dimensional graph theory features through a classification network model, a "trinity" analysis framework of "electrophysiological signal - graph theory network - machine learning" is established, providing a new paradigm for the research on neural markers of mild traumatic brain injury, promoting the leap of concussion research from the behavioral level to the neural network mechanism level, realizing automatic classification of sub-concussions, and greatly improving the accuracy and generalization ability of the classification network model compared with traditional methods.

[0031] In some embodiments, the steps of obtaining the EEG data of the sub-concussion group and the healthy control group and preprocessing the EEG data specifically include: aligning the physical positions of the EEG electrodes with the standard brain region template for channel localization, and obtaining the EEG data of the sub-concussion group and the healthy control group; performing 0.5 Hz - 30 Hz band-pass filtering and 50 Hz notch filtering, downsampling, whole-brain average reference conversion, segmentation and baseline correction, bad segment deletion, removal of electrooculogram and electromyogram components, and removal of overly large amplitude values on the collected EEG data.

[0032] Specifically, this application uses 32 electrodes to cover the main regions of the cerebral cortex, including the frontal lobe, parietal lobe, temporal lobe, and occipital lobe. The physical positions of the EEG electrodes are aligned with the standard brain region template for channel localization, ensuring that the physical position of each electrode precisely matches the standard brain region template, providing an accurate data basis for subsequent source localization and functional connectivity calculation. Band-pass filtering from 0.5 Hz to 30 Hz can retain the brain electrical rhythms related to cognitive activities, filter out low-frequency drift and high-frequency noise, and notch filtering at 50 Hz can eliminate mains interference and improve signal purity. By performing 0.5 Hz - 30 Hz band-pass filtering and 50 Hz notch filtering on the collected EEG data, the signal-to-noise ratio can be improved, and the contamination of functional connectivity metrics by artifacts can be avoided. Downsampling reduces the sampling rate from the original value to 250 Hz, reducing the data volume and computational complexity while retaining the key frequency band information. The re-reference step changes the reference electrode from a single point to a whole-brain average reference, eliminating the influence of the reference electrode position on signal distribution, making the potential differences between brain regions more truly reflect neural activities, and avoiding false functional connectivity or network topology deviations caused by improper selection of the reference electrode. The segmentation step divides the EEG data into independent time periods under different stimulus cue conditions or different stimulus target conditions. For example, under different stimulus cue conditions, taking the appearance of the stimulus cue (no cue, double cue, central cue, and spatial cue) as the origin, the analysis time period is from 500 ms before the stimulus appears to 500 ms after the stimulus appears; under different stimulus target conditions, taking the appearance of the stimulus target (congruent / incongruent) as the origin, the analysis time period is from 1000 ms before the stimulus appears to 1000 ms after the stimulus appears, facilitating the direct comparison of network differences under different conditions. The baseline correction step eliminates the interference of the resting state potential on the task-related signals by subtracting the average potential during the pre-stimulus baseline period, ensuring that the functional connectivity metrics reflect the task-related dynamic changes and avoiding false increases or decreases in metrics such as phase-locked values caused by baseline drift. The bad segment deletion step deletes the trials containing artifacts (such as saccades, electromyograms, and electrode detachment) to ensure data quality. The step of removing electrooculogram and electromyogram components uses independent component analysis to separate and eliminate non-brain-derived components such as electrooculogram and electromyogram, which can greatly improve signal purity. The over-amplitude removal step deletes the trials with abnormal amplitudes with ±70 μV as the standard to exclude extreme noise. The preprocessing of EEG data can provide high-quality data input for subsequent tasks.

[0033] In some embodiments, the segmentation step specifically includes: under different stimulus cue conditions, taking the stimulus cue as the origin, intercepting the time period from 500 ms before the stimulus to 500 ms after the stimulus; under different stimulus target conditions, taking the target stimulus as the origin, intercepting the time period from 1000 ms before the stimulus to 1000 ms after the stimulus; the baseline correction step specifically includes: under different stimulus cue conditions, performing baseline correction on the 500 ms before the stimulus; under different stimulus target conditions, performing baseline correction on the time period from 1000 ms before the stimulus to 500 ms before the stimulus.

[0034] Specifically, under different stimulus cue conditions or under different stimulus target conditions, the EEG data is segmented into independent time periods. For example, under different stimulus cue conditions, taking the appearance of the stimulus cue (no cue, double cue, central cue, and spatial cue) as the origin, the time period from 500 ms before the stimulus appearance to 500 ms after the stimulus appearance is the analysis time period; under different stimulus target conditions, taking the appearance of the stimulus target (congruent / incongruent) as the origin, the time period from 1000 ms before the stimulus appearance and 1000 ms after the stimulus appearance is the analysis time period, which is convenient for directly comparing the network differences under different conditions. The baseline correction step eliminates the interference of the resting state potential on the task-related signals by subtracting the average potential of the baseline period before the stimulus, ensuring that the functional connectivity index reflects the task-related dynamic changes and avoiding the false increase or decrease of indicators such as phase locking value caused by baseline drift.

[0035] In some embodiments, the step of separating the preprocessed EEG data according to the attention subnetwork specifically includes: separating the preprocessed EEG data according to the vigilance network, the orienting network, and the executive control network.

[0036] Specifically, separating the preprocessed EEG data according to the vigilance network, the orienting network, and the executive control network deconstructs complex cognitive functions into quantifiable and analyzable neural network units, which not only improves the scientific rigor of sub-concussion research but also provides clear neurobiological targets for subsequent research, skillfully linking electrophysiological signals, cognitive functions, and nerve injuries together, enabling the precise matching of EEG data to cognitive function dimensions and enhancing sensitivity.

[0037] In some embodiments, the step of separately calculating the phase locking value of the EEG data in different frequency bands specifically includes: separately calculating the phase locking value of the EEG data in the δ (1 Hz - 4 Hz), θ (4 Hz - 8 Hz), α (8 Hz - 13 Hz), and β (13 Hz - 30 Hz) frequency bands, and its calculation formula is: In the formula, PLV is the phase locking value, N is the number of time points, e is the base of the natural logarithm, i is the imaginary unit, and ΔΦ t is the phase difference at each time point.

[0038] It should be understood that neural oscillations in different frequency bands are closely related to specific cognitive functions and pathological states. By calculating the PLV in different frequency bands, the damage patterns of sub-concussion to different neural networks can be analyzed specifically. Among them, δ (1Hz - 4Hz) is related to slow-wave activity, sleep regulation, and compensatory mechanisms after brain injury. Abnormal elevation may reflect chronic fatigue or white matter injury; θ (4Hz - 8Hz) is related to attention, working memory, and the activity of the hippocampal-frontal pathway. Reduced θ synchrony in the orienting network may indicate a deficit in spatial attention; α (8Hz - 13Hz) is related to cortical inhibition, sensory gating, and the resting-state activity of the default mode network. Insufficient α inhibition in the alert network may reflect a decrease in arousal; β (13Hz - 30Hz) is related to motor control, higher cognitive control, and the activity of the basal ganglia-cortical loop. Abnormal β synchrony in the executive control network may indicate impaired conflict monitoring function.

[0039] Specifically, by calculating the phase-locking value in different frequency bands, more refined feature inputs can be provided for subsequent graph theory feature fusion and classification network models, enhancing the robustness of the classification network model and providing multi-dimensional and interpretable features for the model.

[0040] In some embodiments, the step of mapping the EEG data to the source space by the beamforming algorithm based on the head model and brain atlas specifically includes: constructing a forward model by combining the BEM head model constructed based on the MNI standard space and the AAL-116 brain atlas with the position information of the EEG signals of the EEG electrodes, using the beamforming algorithm to perform preliminary localization on the EEG signals, quantifying the functional connection strength between different brain regions according to the localization results, and aggregating the EEG signals to the 116 standard brain regions defined by the AAL-116 atlas through the nearest neighbor interpolation method.

[0041] Specifically, the head model is a three-dimensional head structure model in neuroimaging research, which is a mathematical model used to describe the propagation path of electromagnetic signals in the head and connect the signals recorded on the scalp with the brain source activities. By combining the head anatomical structure information, such as the conductive properties of the scalp, skull, cerebrospinal fluid, and brain tissue, the head model can solve the forward problem, that is, predicting the scalp signal distribution from the electroencephalogram (EEG) signals, and solve the inverse problem, that is, inferring the EEG signals from the scalp signals. This application uses a Boundary Element Model (BEM) volume conductor model constructed based on the MNI standard space, which can accurately describe the conductive properties of different tissues in the head and provides a reliable basis for source analysis and functional connectivity construction. The brain atlas is used to identify and locate the spatial positions and their relationships of different brain regions, and it includes the anatomical structure, functional characteristics, and connection relationships between regions of the brain. The AAL atlas divides the brain into 116 regions, each region has its unique number. This application uses the AAL-116 brain atlas in the functional connectivity part, combines the EEG signal position information of EEG electrodes to construct a forward model, uses the beamforming algorithm to preliminarily locate the EEG signals, quantifies the functional connectivity strength between different brain regions according to the positioning results using the phase-locking value, aggregates the EEG signals to the 116 standard brain regions defined by the AAL-116 atlas through the nearest neighbor interpolation method, and converts the original EEG signals from the electrode space to the anatomical space, solving the problem of spatial ambiguity of the signal source and providing a key basis for quantifying the functional connectivity between brain regions and revealing the neural mechanism of sub-concussion.

[0042] In some embodiments, the global metrics include clustering coefficient, network average shortest path length, global efficiency, and local efficiency; the node metrics include node degree, betweenness centrality, node efficiency, node clustering coefficient, and node local efficiency.

[0043] Specifically, the clustering coefficient reflects the density of connections between adjacent regions of nodes and measures the ability of the entire network to transmit information over short distances. If the clustering coefficient of the sub-concussion group is significantly reduced, it may reflect impaired local information integration ability in the attention network, such as the alert network and the orientation network. The average shortest path length of the network describes the average of the shortest paths between any two nodes in the network and measures the global information transmission efficiency. If the average shortest path length of the sub-concussion group is significantly increased, it may indicate a decline in the global information integration ability of the brain, leading to impaired attention and executive function. The global efficiency describes the average reciprocal of the shortest path lengths between any two nodes in the network and measures the long-distance information transmission ability of the entire network. If the global efficiency of the sub-concussion group is significantly reduced, it may reveal impaired global information transmission ability in the brain network. The local efficiency (Local efficiency, Eloc) reflects the efficiency of the sub-network of a node's neighborhood after the node is removed and measures the local information integration ability between brain regions and the functional modularity characteristics between regions. If the local efficiency of the sub-concussion group is significantly reduced, it may indicate abnormal local functional modularity in the attention network, such as the orientation and executive control networks, leading to a decline in cognitive flexibility.

[0044] Specifically, the node degree represents the number of edges or neighboring nodes directly connected to a node and measures the direct connectivity of a node in the network. If the node degree of specific brain regions in the sub-concussion group, such as the frontal lobe or parietal lobe, is significantly reduced, it may indicate impaired direct information input / output ability in these regions, leading to abnormal attention or executive function. The betweenness centrality examines the contribution of each node to the shortest paths between all other pairs of points and represents the importance of a node as a hub for information flow in the network. If the betweenness centrality of the default network in the sub-concussion group, such as the precuneus, is significantly reduced, it may indicate impaired ability to coordinate the default mode and task-related networks, leading to a decline in cognitive flexibility. The node efficiency reflects the efficiency of information transmission between a node and other nodes and measures the information transmission ability of a single node in the network. It is the manifestation of global efficiency at the node level. If the node efficiency of the sensorimotor area in the sub-concussion group, such as the precentral gyrus, is significantly reduced, it may be related to the decline in its motor control ability. The node clustering coefficient reflects whether the direct neighbors of a node are connected to each other and measures the density of connections within the neighborhood of a single node. If the node clustering coefficient of the temporal lobe in the sub-concussion group is significantly reduced, it may indicate a decline in the local network efficiency related to speech or memory processing. The node local efficiency reflects the ability of other nodes within the neighborhood of a node to maintain information flow when the node is removed and measures the information processing efficiency of a single node in its neighborhood network. If the node local efficiency of the anterior cingulate gyrus in the sub-concussion group is significantly reduced, it may indicate impaired stability in its emotion regulation or conflict monitoring network.

[0045] In some embodiments, the step of fusing the global metrics and the node metrics through graph theory and inputting them into the trained classification network model specifically includes: combining the global metrics and the node metrics of all bands to obtain a target feature vector, and inputting the target feature vector into the classification network model.

[0046] Specifically, through graph theory, the global metrics and the node metrics are fused. The global metrics provide an overall perspective of the network, and the node metrics supplement local details. The combination of the two reveals the complexity of the network. In graph theory, a complex network can be represented as a graph. The brain network is abstracted as a graph, consisting of nodes and edges, that is, brain regions and functional connections. Through graph theory, the complex brain system can be abstracted into a simple geometric graphical representation, namely many nodes and the relationships between the nodes, cleverly fusing the global metrics and the node metrics together, providing accurate and multi-dimensional feature inputs for the subsequent model classification task.

[0047] On the other hand, please refer to Figure 2 , the present invention also provides an electroencephalogram signal processing device based on a brain functional complex network, including: an acquisition module 10, which acquires electroencephalogram data of a sub-concussion group and a healthy control group, and preprocesses the electroencephalogram data; a small matrix generation module 20, which is used to separate the preprocessed electroencephalogram data according to the attention sub-network, and respectively calculate the phase locking values of the electroencephalogram data in different frequency bands to generate a sub-network functional connection matrix; a large matrix generation module 30, which is used to map the electroencephalogram data to the source space through a beamforming algorithm based on a head model and a brain atlas to generate a whole-brain functional connection matrix; a feature extraction module 40, which is used to extract global metrics through the whole-brain functional connection matrix, and extract node metrics through the sub-network functional connection matrix; a classification module 50, which is used to establish a classification network model, fuse the global metrics and the node metrics and input them into the trained classification network model, and output the classification results of the sub-concussion group and the healthy control group.

[0048] On the other hand, the present invention also provides an electronic device, including a memory and a processor, where the memory stores a computer program, and when the processor executes the computer program, the steps of the above-mentioned electroencephalogram signal processing method based on a brain functional complex network are implemented.

[0049] When the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods of the various embodiments of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as USB flash drives, mobile hard disks, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical discs.

[0050] Those of ordinary skill in the art can understand that to implement all or part of the processes in the above-described embodiment methods, it can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-described various methods. Among them, any reference to a memory, storage, database, or other medium used in the various embodiments provided by the present invention can include non-volatile and / or volatile memories. Non-volatile memories can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memories can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and Rambus dynamic RAM (RDRAM), etc.

[0051] The above are only the embodiments of the present invention, and do not limit the patent scope of the present invention. Any equivalent transformation made by using the content of the specification and drawings of the present invention, directly or indirectly applied in the related technical fields, shall be equally included in the patent protection scope of the present invention.

Claims

1. A method for processing electroencephalogram signals based on a complex network of brain functions, characterized in that the steps Including: Obtaining electroencephalogram (EEG) data of the sub-concussion group and the healthy control group, and preprocessing the EEG data; Separating the preprocessed EEG data according to the attention sub-networks, and respectively calculating the phase-locking values of the EEG data in different frequency bands to generate a sub-network functional connectivity matrix; Based on a head model and a brain atlas, mapping the EEG data to the source space through a beamforming algorithm to generate a whole-brain functional connectivity matrix; Extracting global indices through the whole-brain functional connectivity matrix, and extracting node indices through the sub-network functional connectivity matrix; Establishing a classification network model, fusing the global indices and the node indices through graph theory and inputting them into the trained classification network model, and outputting the classification results of the sub-concussion group and the healthy control group.

2. The electroencephalogram signal processing method based on a brain function complex network according to claim 1, wherein The steps of obtaining the EEG data of the sub-concussion group and the healthy control group and preprocessing the EEG data specifically include: Aligning the physical positions of the EEG electrodes with a standard brain region template for channel localization, and obtaining the EEG data of the sub-concussion group and the healthy control group; Performing 0.5Hz - 30Hz band-pass filtering and 50Hz notch filtering, downsampling, whole-brain average reference conversion, segmentation and baseline correction, bad segment deletion, removal of electrooculogram and electromyogram components, and removal of overly large amplitude values on the collected EEG data.

3. The electroencephalogram signal processing method based on the brain function complex network according to claim 2, wherein The segmentation step specifically includes: Under different stimulus cue conditions, taking the stimulus cue as the origin, intercepting the time period from 500ms before the stimulus to 500ms after the stimulus; Under different stimulus target conditions, taking the target stimulus as the origin, intercepting the time period from 1000ms before the stimulus to 1000ms after the stimulus; The baseline correction step specifically includes: Under different stimulus cue conditions, performing baseline correction on the 500ms before the stimulus; Under different stimulus target conditions, performing baseline correction on the time period from 1000ms before the stimulus to 500ms before the stimulus.

4. The EEG signal processing method based on the brain function complex network according to claim 1, characterized in that The step of separating the preprocessed EEG data according to the attention sub-networks specifically includes: Separating the preprocessed EEG data according to the vigilance network, the orienting network, and the executive control network.

5. The electroencephalogram signal processing method based on a brain function complex network according to claim 1, characterized in that The steps of respectively calculating the phase-locking values of the EEG data in different frequency bands specifically include: Respectively calculating the phase-locking values of the EEG data in the δ (1Hz - 4Hz), θ (4Hz - 8Hz), α (8Hz - 13Hz), and β (13Hz - 30Hz) frequency bands, and its calculation formula is: In the formula, PLV is the phase lock value, N is the number of time points, e is the base of the natural logarithm, i is the imaginary unit, and ΔΦ t is the phase difference at each time point.

6. The electroencephalogram signal processing method based on a brain function complex network according to claim 5, characterized in that The step of mapping the EEG data to the source space through a beamforming algorithm based on a head model and a brain atlas specifically includes: Based on the BEM head model and the AAL-116 brain atlas constructed in the MNI standard space, combining the brain electrical signal position information of the EEG electrodes to construct a forward model, using the beamforming algorithm to perform preliminary localization on the brain electrical signals, quantifying the functional connection strength between different brain regions according to the localization results using the phase-locking value, and aggregating the brain electrical signals to the 116 standard brain regions defined by the AAL-116 atlas through the nearest neighbor interpolation method.

7. The electroencephalogram signal processing method based on a brain function complex network according to claim 1, wherein The global metrics include clustering coefficient, average shortest path length of the network, global efficiency, and local efficiency; the node metrics include node degree, betweenness centrality, node efficiency, node clustering coefficient, and node local efficiency.

8. The electroencephalogram signal processing method based on the brain function complex network according to claim 1, characterized in that, The step of fusing the global metrics and the node metrics through graph theory and inputting them into the trained classification network model specifically includes: Combining the global metrics and the node metrics of all bands to obtain a target feature vector, and inputting the target feature vector into the classification network model.

9. An electroencephalogram signal processing device based on a complex network of brain functions, characterized in that It includes: An acquisition module that acquires electroencephalogram (EEG) data of the sub-concussion group and the healthy control group, and preprocesses the EEG data; A small matrix generation module that separates the preprocessed EEG data according to the attention sub-network, calculates the phase-locking value of the EEG data in different frequency bands respectively, and generates a sub-network functional connectivity matrix; A large matrix generation module that maps the EEG data to the source space through a beamforming algorithm based on a head model and a brain atlas to generate a whole-brain functional connectivity matrix; A feature extraction module that extracts global metrics through the whole-brain functional connectivity matrix and extracts node metrics through the sub-network functional connectivity matrix; A classification module that establishes a classification network model, fuses the global metrics and the node metrics through graph theory and inputs them into the trained classification network model, and outputs the classification results of the sub-concussion group and the healthy control group.

10. An electronic device, comprising a memory and a processor, the memory storing a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the electroencephalogram signal processing method based on a brain functional complex network according to any one of claims 1 to 8.

Images