CN110251124B - Method and system for determining effective brain network - Google Patents

Method and system for determining effective brain network Download PDF

Info

Publication number
CN110251124B
CN110251124B CN201910653898.5A CN201910653898A CN110251124B CN 110251124 B CN110251124 B CN 110251124B CN 201910653898 A CN201910653898 A CN 201910653898A CN 110251124 B CN110251124 B CN 110251124B
Authority
CN
China
Prior art keywords
brain network
eeg
eeg data
determining
useful
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN201910653898.5A
Other languages
Chinese (zh)
Other versions
CN110251124A (en
Inventor
张雪英
张静
张卫
回海生
黄丽霞
李凤莲
陈桂军
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Taiyuan University of Technology
Original Assignee
Taiyuan University of Technology
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Taiyuan University of Technology filed Critical Taiyuan University of Technology
Priority to CN201910653898.5A priority Critical patent/CN110251124B/en
Publication of CN110251124A publication Critical patent/CN110251124A/en
Application granted granted Critical
Publication of CN110251124B publication Critical patent/CN110251124B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • AHUMAN NECESSITIES
    • A61MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61BDIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00Measuring for diagnostic purposes; Identification of persons
    • A61B5/24Detecting, measuring or recording bioelectric or biomagnetic signals of the body or parts thereof
    • A61B5/316Modalities, i.e. specific diagnostic methods
    • AHUMAN NECESSITIES
    • A61MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61BDIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00Measuring for diagnostic purposes; Identification of persons
    • A61B5/24Detecting, measuring or recording bioelectric or biomagnetic signals of the body or parts thereof
    • A61B5/316Modalities, i.e. specific diagnostic methods
    • A61B5/369Electroencephalography [EEG]
    • AHUMAN NECESSITIES
    • A61MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61BDIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00Measuring for diagnostic purposes; Identification of persons
    • A61B5/24Detecting, measuring or recording bioelectric or biomagnetic signals of the body or parts thereof
    • A61B5/316Modalities, i.e. specific diagnostic methods
    • A61B5/369Electroencephalography [EEG]
    • A61B5/377Electroencephalography [EEG] using evoked responses
    • A61B5/38Acoustic or auditory stimuli
    • AHUMAN NECESSITIES
    • A61MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61BDIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00Measuring for diagnostic purposes; Identification of persons
    • A61B5/40Detecting, measuring or recording for evaluating the nervous system
    • A61B5/4058Detecting, measuring or recording for evaluating the nervous system for evaluating the central nervous system
    • A61B5/4064Evaluating the brain
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00Pattern recognition
    • G06F18/20Analysing
    • G06F18/21Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/214Generating training patterns; Bootstrap methods, e.g. bagging or boosting

Landscapes

  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Public Health (AREA)
  • Animal Behavior & Ethology (AREA)
  • Surgery (AREA)
  • Molecular Biology (AREA)
  • Medical Informatics (AREA)
  • General Health & Medical Sciences (AREA)
  • Heart & Thoracic Surgery (AREA)
  • Biomedical Technology (AREA)
  • Pathology (AREA)
  • Biophysics (AREA)
  • Veterinary Medicine (AREA)
  • Psychology (AREA)
  • Neurology (AREA)
  • Psychiatry (AREA)
  • Theoretical Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • General Engineering & Computer Science (AREA)
  • Evolutionary Computation (AREA)
  • Evolutionary Biology (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Artificial Intelligence (AREA)
  • General Physics & Mathematics (AREA)
  • Acoustics & Sound (AREA)
  • Neurosurgery (AREA)
  • Physiology (AREA)
  • Measurement And Recording Of Electrical Phenomena And Electrical Characteristics Of The Living Body (AREA)

Abstract

The invention discloses a method and a system for determining an effective brain network. The determination method comprises the following steps: acquiring EEG data acquired by EEG acquisition equipment; screening useful electrode channels; preprocessing EEG data; carrying out frequency decomposition on the preprocessed EEG data by adopting a wavelet transform method; constructing an initial effective brain network; mapping the initial effective brain network to an FCM model to obtain a brain network FCM; training the weight and the direction of the edge of the brain network FCM to obtain a stable brain network FCM; mapping the stable brain network FCM to an effective brain network model to obtain an updated effective brain network; updating the updated validity network according to a set weight threshold value to obtain a steady-state validity brain network; and determining the steady-state effectiveness brain network as a final effectiveness brain network. The invention can improve the accuracy of the effective brain network, so that the effective brain network can more accurately reflect the activity of the cerebral cortex.

Description

Method and system for determining effective brain network
Technical Field
The invention relates to the field of electroencephalogram signal processing, in particular to a method and a system for determining an effective brain network.
Background
An Electroencephalogram (EEG) signal can well represent an emotional brain cognitive process due to high time resolution, and the combination of the EEG signal and the construction of a brain network is an important means for revealing potential neural connection when the brain processes emotion. At present, a topological structure of a brain network constructed based on an EEG is analyzed by using a graph theory method, different emotional states are distinguished by taking network attributes as brain cognitive features, and the method is widely applied to the field of emotion recognition.
Typically, when a brain network is constructed, EEG electrode channels are selected as nodes of the brain network, and correlations between the channels are taken as edges of the brain network. The Brain Network is divided into a Functional Brain Network (FBN) and an Effective Brain Network (EBN) according to the connection mode between nodes in the Brain Network. The FBN is a undirected network and reflects the statistical connection relationship among different brain area nodes. The EBN is a directed network, and can reflect the mutual influence among nodes and represent the flow direction of information.
At present, a functional brain network is constructed based on EEG, the topological structure and network attributes of the brain network are analyzed from the perspective of a complex network, and the network attributes are used for identifying characteristics of different emotions. In some technologies, a resting-state functional brain network is constructed for a preprocessed BOLD (blood oxygen saturation) signal by sequentially adopting methods such as a sliding window sampling technology, a pearson correlation detection technology, a genetic algorithm and the like, and is used for researching a working mechanism in a brain and analyzing brain diseases. However, the functional brain network ignores important biological interpretation in the nervous system, i.e. the trend of nerve fibers or the information conduction of neuron activity, and does not reflect the interaction between nodes in the brain region well.
Compared with a functional brain network, the effective brain network can better depict the interaction of nodes in the brain network and more deeply reflect the information transmission and functional activity rule of a brain system due to the fact that the direction information of edges is added. For the effective brain network, researchers use the relation between the brain structure and the function and the characteristic that the ant colony algorithm is easy to perform information fusion, and the effective brain network which accords with the physiological result of the brain is obtained by fusing the brain structure information and the function information into the ant colony algorithm to search the effective connection of the brain network. However, the effectiveness brain network constructed above assumes that the nodes are independent from each other, and then establishes effectiveness connection for every two nodes, and does not consider the fact that all nodes have synergy and interaction. Therefore, the cortical activity represented by the EEG signal is not accurately reflected.
Therefore, the brain networks constructed based on EEG are mostly functional brain networks, and important biological interpretation in the nervous system, namely the trend of nerve fibers or information conduction of neuron activity, is ignored. The existing effective brain network assumes that all nodes are mutually independent, then establishes effective connection for every two nodes, and does not consider the fact that all nodes have synergy and interaction, so that the brain cortex activity induced by emotional voice cannot be accurately reflected.
Disclosure of Invention
The invention aims to provide a method and a system for determining an effective brain network, which are used for improving the accuracy of the effective brain network and enabling the effective brain network to reflect the activity of a cerebral cortex more accurately.
In order to achieve the purpose, the invention provides the following scheme:
a method of determining an effective brain network, comprising:
acquiring EEG data acquired by EEG acquisition equipment; the EEG data comprises EEG signals acquired by a plurality of electrode channels;
screening useful electrode channels of the EEG acquisition equipment according to the EEG data; the useful electrode channels are all electrode channels of the electroencephalogram acquisition equipment left by removing the useless electrode channels and damaging the electrode channels;
preprocessing the EEG data to obtain preprocessed EEG data; the preprocessed EEG data comprises preprocessed EEG data corresponding to each useful electrode channel;
carrying out frequency decomposition on the preprocessed EEG data by adopting a wavelet transform method to obtain decomposed EEG data; the decomposed EEG data comprises decomposed EEG signals corresponding to each useful electrode channel;
constructing an initial effectiveness brain network according to the decomposed EEG data;
mapping the initial validity brain network to an FCM model to obtain a brain network FCM;
training the weight and the direction of the edge of the brain network FCM until reaching a steady state to obtain a steady state brain network FCM;
mapping the steady state brain network FCM to an effective brain network model to obtain an updated effective brain network;
updating the updated validity network according to a set weight threshold value to obtain a steady-state validity brain network;
and determining the steady-state validity brain network as a final validity brain network corresponding to the electroencephalogram acquisition equipment.
Optionally, the preprocessing the EEG data to obtain preprocessed EEG data specifically includes:
extracting useful EEG data; the useful EEG data is EEG data corresponding to all useful electrode channels;
averaging the EEG data corresponding to all the useful electrode channels at the same moment, and subtracting the EEG data corresponding to each useful electrode channel from the average value to obtain an EEG reference value corresponding to each useful electrode channel; further obtaining EEG reference data corresponding to the useful EEG data, wherein the EEG reference data comprise EEG reference values corresponding to all useful electrode channels at all moments;
intercepting part of data of the EEG reference data according to the research requirement on the EEG signal to obtain preliminary preprocessing EEG data;
and filtering the preliminary pretreatment EEG data by adopting a band-pass filter to obtain the pretreated EEG data.
Optionally, the decomposed EEG data specifically includes: EEG data of five frequency bands of 0.5Hz-4Hz, 4Hz-8Hz, 8Hz-13Hz, 13Hz-22Hz and 22Hz-30 Hz.
Optionally, the constructing an initial validity brain network according to the decomposed EEG data specifically includes:
determining the useful electrode channels as nodes of a brain network, and determining decomposed EEG data corresponding to each useful electrode channel as input signals of the brain network;
determining the correlation degree between different nodes by adopting an effective connection method according to the nodes of the brain network and the input signals to obtain correlation matrixes among all useful electrode channels;
determining the weight and the direction of the edge of the brain network according to the incidence matrix;
and constructing the initial effective brain network according to the weight and the direction of the nodes and the edges of the brain network.
Optionally, the updating the updated validity network according to the set weight threshold to obtain a steady-state validity brain network specifically includes:
determining the weight of the edge of which the weight is smaller than the set weight threshold value in the updated validity brain network as 0, and determining the weight of the edge of which the weight is larger than the set weight threshold value in the updated validity brain network as 1 to obtain a new incidence matrix; the incidence matrix is a matrix formed by the incidence degrees among different nodes;
and obtaining the steady-state effectiveness brain network according to the new incidence matrix.
The present invention also provides a system for determining an effective brain network, comprising:
the EEG data acquisition module is used for acquiring EEG data acquired by the EEG acquisition equipment; the EEG data comprises EEG signals acquired by a plurality of electrode channels;
the screening module is used for screening the useful electrode channels of the EEG acquisition equipment according to the EEG data; the useful electrode channels are all electrode channels of the electroencephalogram acquisition equipment left by removing the useless electrode channels and damaging the electrode channels;
the preprocessing module is used for preprocessing the EEG data to obtain preprocessed EEG data; the preprocessed EEG data comprises preprocessed EEG data corresponding to each useful electrode channel;
the frequency decomposition module is used for carrying out frequency decomposition on the preprocessed EEG data by adopting a wavelet transform method to obtain decomposed EEG data; the decomposed EEG data comprises decomposed EEG signals corresponding to each useful electrode channel;
the initial significance brain network construction module is used for constructing an initial significance brain network according to the decomposed EEG data;
the brain network FCM building module is used for mapping the initial effective brain network to an FCM model to obtain the brain network FCM;
the training module is used for training the weight and the direction of the edge of the brain network FCM until reaching a steady state to obtain the steady state brain network FCM;
the validity brain network updating module is used for mapping the steady state brain network FCM to a validity brain network model to obtain an updated validity brain network;
the steady-state validity brain network construction module is used for updating the updated validity network according to a set weight threshold value to obtain a steady-state validity brain network;
and the final effectiveness brain network determining module is used for determining the steady-state effectiveness brain network as a final effectiveness brain network corresponding to the electroencephalogram acquisition equipment.
Optionally, the preprocessing module specifically includes:
a useful EEG data extraction unit for extracting useful EEG data; the useful EEG data is EEG data corresponding to all useful electrode channels;
the average reference unit is used for averaging the EEG data corresponding to all the useful electrode channels at the same moment, and subtracting the EEG data corresponding to each useful electrode channel from the average value to obtain an EEG reference value corresponding to each useful electrode channel; further obtaining EEG reference data corresponding to the useful EEG data, wherein the EEG reference data comprise EEG reference values corresponding to all useful electrode channels at all moments;
the data interception unit is used for intercepting partial data of the EEG reference data according to the research requirement on the EEG signal to obtain preliminary preprocessing EEG data;
and the filtering unit is used for filtering the preliminary pretreatment EEG data by adopting a band-pass filter to obtain the pretreated EEG data.
Optionally, the EEG data decomposed by the frequency decomposition module specifically includes: EEG data of five frequency bands of 0.5Hz-4Hz, 4Hz-8Hz, 8Hz-13Hz, 13Hz-22Hz and 22Hz-30 Hz.
Optionally, the initial validity brain network construction module specifically includes:
the brain network node and input signal determining unit is used for determining the useful electrode channels as nodes of the brain network and determining decomposed EEG data corresponding to each useful electrode channel as input signals of the brain network;
the incidence matrix determining unit is used for determining the incidence between different nodes by adopting an effective connection method according to the nodes of the brain network and the input signals to obtain the incidence matrix between all useful electrode channels;
the brain network side determining unit is used for determining the weight and the direction of the side of the brain network according to the incidence matrix;
and the initial effectiveness brain network construction unit is used for constructing the initial effectiveness brain network according to the weight and the direction of the nodes and the edges of the brain network.
Optionally, the steady-state validity brain network constructing module specifically includes:
a weight updating unit, configured to determine a weight of a side of the updated validity brain network whose weight is smaller than the set weight threshold as 0, and determine a weight of a side of the updated validity brain network whose weight is larger than the set weight threshold as 1, so as to obtain a new incidence matrix; the incidence matrix is a matrix formed by the incidence degrees among different nodes;
and the steady-state validity brain network construction unit is used for obtaining the steady-state validity brain network according to the new incidence matrix.
According to the specific embodiment provided by the invention, the invention discloses the following technical effects:
compared with the existing brain network, the steady-state effectiveness brain network constructed by the invention takes the synergy and interaction among all nodes into consideration, introduces the FCM model with reasoning function, trains and determines the structure and weight of the brain network, so that the constructed steady-state effectiveness brain network can better reflect the cognitive process and expression habit of human beings and can more accurately reflect the cortical activity induced by emotional voice. Further, when the method is applied, the network characteristics of the steady-state effective brain network are extracted to serve as the brain cognitive characteristics and are used in the emotion voice recognition field, and the voice emotion recognition performance is further improved.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings needed to be used in the embodiments will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings without inventive exercise.
FIG. 1 is a schematic flow chart of a method for determining an effective brain network according to the present invention;
FIG. 2 is an example of EEG data acquired by an EEG acquisition device in accordance with the present invention;
FIG. 3 is a schematic diagram of an initial validity brain network of the present invention;
FIG. 4 is an example of a final validity brain network in the present invention;
fig. 5 is a schematic structural diagram of the system for determining an effective brain network according to the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
In order to make the aforementioned objects, features and advantages of the present invention comprehensible, embodiments accompanied with figures are described in further detail below.
Fig. 1 is a schematic flow chart of the method for determining an effective brain network according to the present invention. As shown in fig. 1, the method for determining the validity brain network includes the following steps:
step 100: and acquiring EEG data acquired by the EEG acquisition equipment. The EEG data comprises EEG signals acquired by a plurality of electrode channels. Generally, emotion voice signals are used as stimulation materials, an electroencephalograph with 64 electrode channels is adopted to record emotion EEG signals corresponding to different emotions induced by a human brain in the process of emotion cognitive activities, and the EEG signals mainly record the change of potential differences of electrode detection points on the surface of a scalp along with time. Fig. 2 is an example of EEG data acquired by an electroencephalogram acquisition device in the present invention, and in this example, EEG data recorded by an electroencephalograph recorder with 64 electrode channels is used.
Step 200: and screening useful electrode channels of the EEG acquisition equipment according to the EEG data. The useful electrode channels are all electrode channels of the electroencephalogram acquisition equipment left by removing the useless electrode channels and damaging the electrode channels. The useless electrode channel refers to an electrode channel which does not acquire an EEG signal, data acquired by a damaged electrode channel is obviously changed compared with data of other channels, and in order to improve accuracy, the electrode channel corresponding to signals such as ocular artifacts in the EEG signal is generally classified as the damaged electrode channel. Before signal acquisition, the electrode channel of the electroencephalogram acquisition equipment can be detected in advance, and a damaged electrode channel and a useless electrode channel can be determined.
Step 300: and preprocessing the EEG data to obtain preprocessed EEG data. The pre-processed EEG data comprises pre-processed EEG data corresponding to each useful electrode channel. Specifically, the process of preprocessing EEG data is as follows:
(a) and positioning an electrode channel corresponding to the EEG data, and extracting useful EEG data. The useful EEG data here is EEG data corresponding to all useful electrode channels.
(b) Useful EEG data are averaged at the same time. And averaging the EEG data corresponding to all the useful electrode channels at the same moment, and subtracting the EEG data corresponding to each useful electrode channel from the average value to obtain an EEG reference value corresponding to each useful electrode channel. And obtaining EEG reference values corresponding to all useful electrode channels at all moments in sequence, and further obtaining EEG reference data corresponding to the useful EEG data.
(c) According to the research requirement on the EEG signal, partial data of the EEG reference data are intercepted, and preliminary preprocessing EEG data are obtained. This step is intended to segment the EEG data and based on the change in EEG average data, remove data longer before the start of stimulation and longer after the end of stimulation, typically intercepting the EGG signal during the 200ms period before the start of stimulation to 800ms period after stimulation as preliminary pre-processing EEG data.
(d) And filtering the preliminary pretreatment EEG data by adopting a band-pass filter to obtain the pretreated EEG data. For example, a 0.5Hz-30Hz bandpass filter may be used.
Step 400: and (3) carrying out frequency decomposition on the preprocessed EEG data by adopting a wavelet transform method to obtain decomposed EEG data. The decomposed EEG data comprises decomposed EEG signals for each useful electrode channel.
For example, when the wavelet transform method is used, the EEG signal of each electrode channel can be decomposed into the EEG signals of the five different frequency bands of δ (0.5-4Hz), θ (4-8Hz), α (8-13Hz), β (13-22Hz) and γ (22-30Hz) according to the magnitude of the frequency, which are respectively expressed as: x is the number ofδ(t),xθ(t),xα(t),xβ(t) and xγ(t) of (d). Continuous wavelet transform W of signal x (t)ψIs defined as:
Figure BDA0002136215020000081
where a is a scale factor, b is a time shift factor,
Figure BDA0002136215020000082
is the window function introduced.
Step 500: an initial significance brain network is constructed from the decomposed EEG data. The method specifically comprises the following steps:
(a) determining the useful electrode channels as nodes of the brain network, and determining decomposed EEG data corresponding to each useful electrode channel as input signals of the brain network;
(b) and determining the correlation degree between different nodes by adopting an effective connection method according to the nodes and the input signals of the brain network to obtain a correlation matrix among all useful electrode channels. The effectiveness connection method adopts a Glangel causal model to analyze and measure the relevance (causal relationship) between different nodes, and determines the relevance as the weight and the direction of an edge between two nodes. And calculating GC values between any two electrodes to obtain a correlation matrix of N by N, wherein N is the number of the electrodes. The GC values between the electrodes represent the causal relationship of the interaction between the two electrode channels, and are calculated as follows:
suppose X1,tAnd X2,tIs the EEG signal on two electrodes, then the two electrodes give a time invariant model expression in the univariate case:
Figure BDA0002136215020000091
Figure BDA0002136215020000092
considering that the sequences may be influenced by each other in the past, the time series X is combined1,tAnd X2,tThe time-invariant model can be built as follows:
Figure BDA0002136215020000093
Figure BDA0002136215020000094
where t is 1,2 … m denotes the regression order, ε1,t2,tAnd epsilon'1,t,ε′2,tIs white noise.
At this time, X is defined2To X1The cause and effect relationship of (A) is as follows:
Figure BDA0002136215020000095
(c) and determining the weight and the direction of the edge of the brain network according to the incidence matrix.
(d) And constructing an initial effective brain network according to the weight and the direction of the nodes and the edges of the brain network. FIG. 3 is a schematic diagram of an initial effective brain network of the present invention, as shown in FIG. 3, considering that the EEG electrode channel number is usually a multiple of 8, and the node number is 8, that is, the node of the brain network is C1~C8Node CiAnd CjCause and effect relationship between the two is represented by Wi,jThe values are represented (i.e., weights). There are three types of causal relationships: when W isi,j> 0, represents CiIs caused by a change injDegree of equidirectional variation; when W isi,j< 0, represents CiIs caused by a change injThe degree of change in the opposite direction; when W isi,j0 represents CiAnd CjNo causal relationship exists.
Step 600: and mapping the initial effective brain network to the FCM to obtain the brain network FCM. And mapping the weights and the directions of the nodes and the edges of the initial Effective Brain Network (EBN) into the FCM one by one, namely, taking the nodes of the initial EBN as the nodes of the FCM, taking the edges of the brain network as the edges of the FCM, and constructing the FCM to obtain the brain network FCM.
Step 700: and training the weight and the direction of the edge of the brain network FCM until reaching a steady state to obtain the steady-state brain network FCM. Suppose a node of the brain network FCM is C1~CnNode CiAnd CjCause and effect relationship between the two is represented by Wi,jRepresentation (i.e., weight). When W isi,j> 0, represents CiIs caused by a change injDegree of equidirectional variation; when W isi,j< 0, represents CiIs caused by a change injThe degree of change in the opposite direction; when W isi,j0 represents CiAnd CjNo causal relationship exists. The training process is as follows:
(a) the initialized network state value when T is 0 is shown as follows:
C(0)=(C1(0),C2(0)...,Cn(0))
(b) the state values of n nodes at time T ═ T are shown in the following equation:
C(t)=(C1(t),C2(t)...,Cn(t))
(c) the transfer relationship between the electrode channels at time t is:
Ct=WtCt-1
Ctrepresenting the input of the electrode channel at time t, WtAnd representing a directed weighted connection matrix between channels at the time t, and weighting the connection coefficients of adjacent points by using a Gaussian radial basis function for weight estimation, so that the smoothness of time-varying connection can be ensured. Thus, WtCan be estimated as
Figure BDA0002136215020000101
Wherein, N represents the channel number of the electroencephalogram signal in the model, lambda represents a regression parameter, and the Gaussian radial basis kernel function is as follows:
Figure BDA0002136215020000102
(d) through a limited number of iterations, the brain network FCM can reach two states: (i) the state value of the node reaches a fixed value, namely a hidden mode or a fixed-point attractor; (ii) the node state values remain cycled between a number of fixed state values, referred to as a finite cycle. When the brain network FCM reaches the (i) or (ii) state, a steady state/equilibrium state is reached, at which time a steady state brain network FCM is obtained.
Step 800: and mapping the stable brain network FCM to the effective brain network model to obtain the updated effective brain network. And mapping by taking the weight and the direction of the side of the stable brain network FCM as the new weight and the new direction of the side of the effective brain network to obtain the updated effective brain network.
Step 900: and updating the updated validity network according to the set weight threshold value to obtain the steady-state validity brain network. The set weight threshold value can be ln (n), or other values can be set according to actual requirements, the weight of the side of the updated validity brain network where the weight is smaller than the set weight threshold value is determined as 0, the weight of the side of the updated validity brain network where the weight is larger than the set weight threshold value is determined as 1, a new incidence matrix is obtained, and then the steady-state validity brain network is obtained. The incidence matrix is a matrix formed by the incidence degrees among different nodes. And under the condition of ensuring the steady-state validity brain network connectivity of each frequency band, selecting the threshold with the maximum value as the set weight threshold, wherein the accuracy is highest at the moment.
Step 1000: and determining the steady-state effective brain network as a final effective brain network corresponding to the electroencephalogram acquisition equipment. Fig. 4 is an example of the final validity brain network in the present invention, and as shown in fig. 4, the present example is the final validity brain network in the case where the delta band has a threshold value of 0.5.
In the present invention, after the step 400 performs frequency decomposition on the preprocessed EEG data, the subsequent steps 500-1000 are all operations performed for each frequency band, for example, the step 400 is decomposed into 5 frequency bands, and then the EEG data of each frequency band passes through the steps 500-1000 to obtain a final validity brain network, and finally, 5 final validity brain networks corresponding to the 5 frequency bands are obtained.
Corresponding to the method for determining the validity brain network shown in fig. 1, the present invention further provides a system for determining the validity brain network, and fig. 5 is a schematic structural diagram of the system for determining the validity brain network according to the present invention. As shown in fig. 5, the system for determining an effective brain network includes the following structures:
an EEG data acquisition module 501, configured to acquire EEG data acquired by an electroencephalogram acquisition device; the EEG data comprises EEG signals acquired by a plurality of electrode channels;
a screening module 502 for screening a useful electrode channel of the electroencephalogram acquisition device according to the EEG data; the useful electrode channels are all electrode channels of the electroencephalogram acquisition equipment left by removing the useless electrode channels and damaging the electrode channels;
a preprocessing module 503, configured to preprocess the EEG data to obtain preprocessed EEG data; the preprocessed EEG data comprises preprocessed EEG data corresponding to each useful electrode channel;
a frequency decomposition module 504, configured to perform frequency decomposition on the preprocessed EEG data by using a wavelet transform method to obtain decomposed EEG data; the decomposed EEG data comprises decomposed EEG signals corresponding to each useful electrode channel;
an initial significance brain network construction module 505, configured to construct an initial significance brain network according to the decomposed EEG data;
a brain network FCM construction module 506, configured to map the initial validity brain network to an FCM model to obtain a brain network FCM;
a training module 507, configured to train a weight and a direction of an edge of the brain network FCM until a steady state is reached, so as to obtain a steady-state brain network FCM;
an effectiveness brain network updating module 508, configured to map the steady-state brain network FCM to an effectiveness brain network model, so as to obtain an updated effectiveness brain network;
a steady-state validity brain network constructing module 509, configured to update the updated validity network according to a set weight threshold, so as to obtain a steady-state validity brain network;
and a final validity brain network determination module 5010, configured to determine the steady-state validity brain network as a final validity brain network corresponding to the electroencephalogram acquisition device.
As another embodiment, the preprocessing module 503 specifically includes:
a useful EEG data extraction unit for extracting useful EEG data; the useful EEG data is EEG data corresponding to all useful electrode channels;
the average reference unit is used for averaging the EEG data corresponding to all the useful electrode channels at the same moment, and subtracting the EEG data corresponding to each useful electrode channel from the average value to obtain an EEG reference value corresponding to each useful electrode channel; further obtaining EEG reference data corresponding to the useful EEG data, wherein the EEG reference data comprise EEG reference values corresponding to all useful electrode channels at all moments;
the data interception unit is used for intercepting partial data of the EEG reference data according to the research requirement on the EEG signal to obtain preliminary preprocessing EEG data;
and the filtering unit is used for filtering the preliminary pretreatment EEG data by adopting a band-pass filter to obtain the pretreated EEG data.
As another embodiment, the EEG data decomposed by the frequency decomposition module 504 specifically includes: EEG data of five frequency bands of 0.5Hz-4Hz, 4Hz-8Hz, 8Hz-13Hz, 13Hz-22Hz and 22Hz-30 Hz.
As another embodiment, the initial validity brain network constructing module 505 specifically includes:
the brain network node and input signal determining unit is used for determining the useful electrode channels as nodes of the brain network and determining decomposed EEG data corresponding to each useful electrode channel as input signals of the brain network;
the incidence matrix determining unit is used for determining the incidence between different nodes by adopting an effective connection method according to the nodes of the brain network and the input signals to obtain the incidence matrix between all useful electrode channels;
the brain network side determining unit is used for determining the weight and the direction of the side of the brain network according to the incidence matrix;
and the initial effectiveness brain network construction unit is used for constructing the initial effectiveness brain network according to the weight and the direction of the nodes and the edges of the brain network.
As another embodiment, the steady-state validity brain network constructing module 509 specifically includes:
a weight updating unit, configured to determine a weight of a side of the updated validity brain network whose weight is smaller than the set weight threshold as 0, and determine a weight of a side of the updated validity brain network whose weight is larger than the set weight threshold as 1, so as to obtain a new incidence matrix; the incidence matrix is a matrix formed by the incidence degrees among different nodes;
a steady state validity brain network construction unit for obtaining the steady state validity brain network according to the new incidence matrix
The embodiments in the present description are described in a progressive manner, each embodiment focuses on differences from other embodiments, and the same and similar parts among the embodiments are referred to each other. For the system disclosed by the embodiment, the description is relatively simple because the system corresponds to the method disclosed by the embodiment, and the relevant points can be referred to the method part for description.
The principles and embodiments of the present invention have been described herein using specific examples, which are provided only to help understand the method and the core concept of the present invention; meanwhile, for a person skilled in the art, according to the idea of the present invention, the specific embodiments and the application range may be changed. In view of the above, the present disclosure should not be construed as limiting the invention.

Claims (6)

1. A method for determining an effective brain network, comprising:
acquiring EEG data acquired by EEG acquisition equipment; the EEG data comprises EEG signals acquired by a plurality of electrode channels;
screening useful electrode channels of the EEG acquisition equipment according to the EEG data; the useful electrode channels are all electrode channels of the electroencephalogram acquisition equipment left by removing the useless electrode channels and damaging the electrode channels;
preprocessing the EEG data to obtain preprocessed EEG data; the preprocessed EEG data comprises preprocessed EEG data corresponding to each useful electrode channel;
carrying out frequency decomposition on the preprocessed EEG data by adopting a wavelet transform method to obtain decomposed EEG data; the decomposed EEG data specifically comprises EEG data of five frequency bands;
the decomposed EEG data comprises decomposed EEG signals corresponding to each useful electrode channel;
constructing an initial effectiveness brain network according to the decomposed EEG data;
the constructing of the initial significance brain network according to the decomposed EEG data specifically comprises:
determining the useful electrode channels as nodes of the brain network, and determining decomposed EEG data corresponding to each useful electrode channel as input signals of the brain network;
determining the correlation degree between different nodes by adopting an effective connection method according to the nodes of the brain network and the input signals to obtain correlation matrixes among all useful electrode channels; the effectiveness connection method adopts a Glange causal model to analyze and measure the association degree between different nodes, and determines the association degree as the weight and the direction of an edge between two nodes; calculating GC values between any two electrodes to obtain a correlation matrix of N x N, wherein N is the number of the electrodes; the GC values between the electrodes represent the causal relationship of the interaction between the two electrode channels, and are calculated as follows:
suppose X1,tAnd X2,tIs the EEG signal on two electrodes, then the two electrodes give a time invariant model expression in the univariate case:
Figure FDA0003423471150000021
Figure FDA0003423471150000022
considering that the sequences may be influenced by each other in the past, the time series X is combined1,tAnd X2,tThe time-invariant model is established as follows:
Figure FDA0003423471150000023
Figure FDA0003423471150000024
where t is 1,2 … m denotes the regression order, ε1,t2,tAnd epsilon'1,t,ε′2,tIs white noise;
at this time, X is defined2To X1The cause and effect relationship of (A) is as follows:
Figure FDA0003423471150000025
determining the weight and the direction of the edge of the brain network according to the incidence matrix;
constructing an initial effective brain network according to the weight and the direction of the nodes and the edges of the brain network; node CiAnd CjCause and effect relationship between the two is represented by Wi,jValue representation, CiIs the ith node, CjIs the jth node, Wi,jIs the weight; there are three types of causal relationships: when W isi,j> 0, represents CiIs caused by a change injDegree of equidirectional variation; when W isi,j< 0, represents CiIs caused by a change injThe degree of change in the opposite direction; when W isi,j0 represents CiAnd CjNo causal relationship exists;
mapping the initial validity brain network to an FCM model to obtain a brain network FCM;
training the weight and the direction of the edge of the brain network FCM until reaching a steady state to obtain a steady state brain network FCM;
mapping the steady state brain network FCM to an effective brain network model to obtain an updated effective brain network;
updating the updated validity network according to a set weight threshold value to obtain a steady-state validity brain network;
and determining the steady-state validity brain network as a final validity brain network corresponding to the electroencephalogram acquisition equipment.
2. The method for determining an effective brain network according to claim 1, wherein the preprocessing the EEG data to obtain preprocessed EEG data specifically comprises:
extracting useful EEG data; the useful EEG data is EEG data corresponding to all useful electrode channels;
averaging the EEG data corresponding to all the useful electrode channels at the same moment, and subtracting the EEG data corresponding to each useful electrode channel from the average value to obtain an EEG reference value corresponding to each useful electrode channel; further obtaining EEG reference data corresponding to the useful EEG data, wherein the EEG reference data comprise EEG reference values corresponding to all useful electrode channels at all moments;
intercepting part of data of the EEG reference data according to the research requirement on the EEG signal to obtain preliminary preprocessing EEG data;
and filtering the preliminary pretreatment EEG data by adopting a band-pass filter to obtain the pretreated EEG data.
3. The method for determining an effectiveness brain network according to claim 1, wherein the updating the updated effectiveness brain network according to the set weight threshold to obtain a steady-state effectiveness brain network specifically comprises:
determining the weight of the edge of which the weight is smaller than the set weight threshold value in the updated validity brain network as 0, and determining the weight of the edge of which the weight is larger than the set weight threshold value in the updated validity brain network as 1 to obtain a new incidence matrix; the incidence matrix is a matrix formed by the incidence degrees among different nodes;
and obtaining the steady-state effectiveness brain network according to the new incidence matrix.
4. A system for determining an effective brain network, comprising:
the EEG data acquisition module is used for acquiring EEG data acquired by the EEG acquisition equipment; the EEG data comprises EEG signals acquired by a plurality of electrode channels;
the screening module is used for screening the useful electrode channels of the EEG acquisition equipment according to the EEG data; the useful electrode channels are all electrode channels of the electroencephalogram acquisition equipment left by removing the useless electrode channels and damaging the electrode channels;
the preprocessing module is used for preprocessing the EEG data to obtain preprocessed EEG data; the preprocessed EEG data comprises preprocessed EEG data corresponding to each useful electrode channel;
the frequency decomposition module is used for carrying out frequency decomposition on the preprocessed EEG data by adopting a wavelet transform method to obtain decomposed EEG data; the decomposed EEG data comprises decomposed EEG signals corresponding to each useful electrode channel; the EEG data decomposed by the frequency decomposition module specifically comprises EEG data of five frequency bands;
the initial significance brain network construction module is used for constructing an initial significance brain network according to the decomposed EEG data;
the initial effectiveness brain network construction module specifically comprises:
the brain network node and input signal determining unit is used for determining the useful electrode channels as nodes of the brain network and determining decomposed EEG data corresponding to each useful electrode channel as input signals of the brain network;
the incidence matrix determining unit is used for determining the incidence between different nodes by adopting an effective connection method according to the nodes of the brain network and the input signals to obtain the incidence matrix between all useful electrode channels; the effectiveness connection method adopts a Glange causal model to analyze and measure the association degree between different nodes, and determines the association degree as the weight and the direction of an edge between two nodes; calculating GC values between any two electrodes to obtain a correlation matrix of N x N, wherein N is the number of the electrodes; the GC values between the electrodes represent the causal relationship of the interaction between the two electrode channels, and are calculated as follows:
suppose X1,tAnd X2,tIs the EEG signal on two electrodes, then the two electrodes give a time invariant model expression in the univariate case:
Figure FDA0003423471150000041
Figure FDA0003423471150000042
considering that the sequences may be influenced by each other in the past, the time series X is combined1,tAnd X2,tThe time-invariant model is established as follows:
Figure FDA0003423471150000043
Figure FDA0003423471150000044
wherein t 1, 2.. m denotes the regression order, epsilon1,t,ε2,tAnd epsilon'1,t,ε′2,tIs white noise;
at this time, X is defined2To X1The cause and effect relationship of (A) is as follows:
Figure FDA0003423471150000051
the brain network side determining unit is used for determining the weight and the direction of the side of the brain network according to the incidence matrix;
the initial effectiveness brain network construction unit is used for constructing the initial effectiveness brain network according to the weight and the direction of the nodes and the edges of the brain network; node CiAnd CjCause and effect relationship between the two is represented by Wi,jValue representation, CiIs the ith node, CjIs the jth node, Wi,jIs the weight; there are three types of causal relationships: when W isi,j> 0, represents CiIs caused by a change injDegree of equidirectional variation; when W isi,j< 0, represents CiIs caused by a change injThe degree of change in the opposite direction; when W isi,j0 represents CiAnd CjNo causal relationship exists;
the brain network FCM building module is used for mapping the initial effective brain network to an FCM model to obtain the brain network FCM;
the training module is used for training the weight and the direction of the edge of the brain network FCM until reaching a steady state to obtain the steady state brain network FCM;
the validity brain network updating module is used for mapping the steady state brain network FCM to a validity brain network model to obtain an updated validity brain network;
the steady-state validity brain network construction module is used for updating the updated validity network according to a set weight threshold value to obtain a steady-state validity brain network;
and the final effectiveness brain network determining module is used for determining the steady-state effectiveness brain network as a final effectiveness brain network corresponding to the electroencephalogram acquisition equipment.
5. The system for determining an effectiveness brain network according to claim 4, wherein the preprocessing module specifically comprises:
a useful EEG data extraction unit for extracting useful EEG data; the useful EEG data is EEG data corresponding to all useful electrode channels;
the average reference unit is used for averaging the EEG data corresponding to all the useful electrode channels at the same moment, and subtracting the EEG data corresponding to each useful electrode channel from the average value to obtain an EEG reference value corresponding to each useful electrode channel;
further obtaining EEG reference data corresponding to the useful EEG data, wherein the EEG reference data comprise EEG reference values corresponding to all useful electrode channels at all moments;
the data interception unit is used for intercepting partial data of the EEG reference data according to the research requirement on the EEG signal to obtain preliminary preprocessing EEG data;
and the filtering unit is used for filtering the preliminary pretreatment EEG data by adopting a band-pass filter to obtain the pretreated EEG data.
6. The system for determining an effectiveness brain network according to claim 4, wherein the steady-state effectiveness brain network constructing module specifically comprises:
a weight updating unit, configured to determine a weight of a side of the updated validity brain network whose weight is smaller than the set weight threshold as 0, and determine a weight of a side of the updated validity brain network whose weight is larger than the set weight threshold as 1, so as to obtain a new incidence matrix; the incidence matrix is a matrix formed by the incidence degrees among different nodes;
and the steady-state validity brain network construction unit is used for obtaining the steady-state validity brain network according to the new incidence matrix.
CN201910653898.5A 2019-07-19 2019-07-19 Method and system for determining effective brain network Active CN110251124B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201910653898.5A CN110251124B (en) 2019-07-19 2019-07-19 Method and system for determining effective brain network

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201910653898.5A CN110251124B (en) 2019-07-19 2019-07-19 Method and system for determining effective brain network

Publications (2)

Publication Number Publication Date
CN110251124A CN110251124A (en) 2019-09-20
CN110251124B true CN110251124B (en) 2022-02-18

Family

ID=67927134

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201910653898.5A Active CN110251124B (en) 2019-07-19 2019-07-19 Method and system for determining effective brain network

Country Status (1)

Country Link
CN (1) CN110251124B (en)

Families Citing this family (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111543988B (en) * 2020-05-25 2021-06-08 五邑大学 Adaptive cognitive activity recognition method and device and storage medium
CN111616704A (en) * 2020-06-24 2020-09-04 天津大学 Research method for brain dynamic function network interaction mode of sentence listening task
CN112950908B (en) * 2021-02-03 2022-12-30 重庆川仪自动化股份有限公司 Data monitoring and early warning method, system, medium and electronic terminal
CN113143293B (en) * 2021-04-12 2023-04-07 天津大学 Continuous speech envelope nerve entrainment extraction method based on electroencephalogram source imaging
CN115778323B (en) * 2022-11-21 2024-04-26 山东大学 User rehabilitation level assessment system with multi-source data fusion function

Family Cites Families (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102722727B (en) * 2012-06-11 2014-03-05 杭州电子科技大学 Electroencephalogram feature extracting method based on brain function network adjacent matrix decomposition
CN103325119A (en) * 2013-06-27 2013-09-25 中国科学院自动化研究所 Default state brain network center node detecting method based on modality fusion
KR101653910B1 (en) * 2015-02-27 2016-09-06 연세대학교 원주산학협력단 Apparatus and method for neurofeedback using brain network characteristic analysis based on electroencephalogram
CN107242873B (en) * 2017-07-05 2020-06-09 成都信息工程大学 Brain network construction method based on functional magnetic resonance psychophysiological interaction
EP3706617A4 (en) * 2017-11-10 2021-08-18 LVIS Corporation Efficacy and/or therapeutic parameter recommendation using individual patient data and therapeutic brain network maps
CN108577835B (en) * 2018-05-17 2019-07-19 太原理工大学 A kind of brain function network establishing method based on micro- state

Also Published As

Publication number Publication date
CN110251124A (en) 2019-09-20

Similar Documents

Publication Publication Date Title
CN110251124B (en) Method and system for determining effective brain network
Sangaiah et al. An intelligent learning approach for improving ECG signal classification and arrhythmia analysis
CN110969108B (en) Limb action recognition method based on autonomic motor imagery electroencephalogram
US20200367800A1 (en) Method for identifying driving fatigue based on cnn-lstm deep learning model
CN114533086B (en) Motor imagery brain electrolysis code method based on airspace characteristic time-frequency transformation
CN110013250B (en) Multi-mode characteristic information fusion prediction method for suicidal behavior of depression
CN109871831B (en) Emotion recognition method and system
CN112450947B (en) Dynamic brain network analysis method for emotional arousal degree
CN111839506B (en) Mental load detection method and device
CN110543831A (en) brain print identification method based on convolutional neural network
CN115363531A (en) Epilepsy detection system based on bimodal electroencephalogram signal information bottleneck
Cordoș et al. ECG signal classification using convolutional neural networks for biometric identification
Sun et al. A novel complex network-based graph convolutional network in major depressive disorder detection
AU2020103949A4 (en) EEG Signal Mixed Noise Processing Method, Equipment and Storage Medium
Dzitac et al. Identification of ERD using fuzzy inference systems for brain-computer interface
CN111931578B (en) Electroencephalogram identification method based on minimum spanning tree and regional double-layer network
CN111887811B (en) Brain abnormal discharge detection method and system based on electroencephalogram signal characteristics
CN117883082A (en) Abnormal emotion recognition method, system, equipment and medium
CN111671421B (en) Electroencephalogram-based children demand sensing method
CN111736690A (en) Motor imagery brain-computer interface based on Bayesian network structure identification
Priyadharsini et al. An Efficient method for the removal of ECG artifact from measured EEG Signal using PSO algorithm
Sutharsan et al. Electroencephalogram signal processing with independent component analysis and cognitive stress classification using convolutional neural networks
CN115392287A (en) Electroencephalogram signal online self-adaptive classification method based on self-supervision learning
CN111493864B (en) EEG signal mixed noise processing method, equipment and storage medium
Wang et al. Functional brain network based multi-domain feature fusion of hearing-Impaired EEG emotion identification

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant