CN117251807A - Motor imagery electroencephalogram signal classification method of neural network - Google Patents
Motor imagery electroencephalogram signal classification method of neural network Download PDFInfo
- Publication number
- CN117251807A CN117251807A CN202311533871.5A CN202311533871A CN117251807A CN 117251807 A CN117251807 A CN 117251807A CN 202311533871 A CN202311533871 A CN 202311533871A CN 117251807 A CN117251807 A CN 117251807A
- Authority
- CN
- China
- Prior art keywords
- motor imagery
- channels
- electroencephalogram signal
- neural network
- electroencephalogram
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 54
- 238000013528 artificial neural network Methods 0.000 title claims abstract description 29
- 238000012549 training Methods 0.000 claims abstract description 43
- 239000011159 matrix material Substances 0.000 claims abstract description 37
- 230000008569 process Effects 0.000 claims abstract description 12
- 238000003062 neural network model Methods 0.000 claims abstract description 10
- 238000012360 testing method Methods 0.000 claims description 30
- 210000004556 brain Anatomy 0.000 claims description 18
- 238000004364 calculation method Methods 0.000 claims description 17
- 238000012545 processing Methods 0.000 claims description 13
- 230000033001 locomotion Effects 0.000 claims description 10
- 230000004913 activation Effects 0.000 claims description 9
- 230000006870 function Effects 0.000 claims description 8
- 238000004590 computer program Methods 0.000 claims description 4
- 230000005540 biological transmission Effects 0.000 claims description 3
- 238000012937 correction Methods 0.000 claims description 3
- 238000001914 filtration Methods 0.000 claims description 3
- 238000005070 sampling Methods 0.000 claims description 3
- 230000011218 segmentation Effects 0.000 claims description 3
- 230000009286 beneficial effect Effects 0.000 abstract description 3
- 230000003925 brain function Effects 0.000 abstract description 3
- 230000001537 neural effect Effects 0.000 abstract description 3
- 238000004458 analytical method Methods 0.000 description 2
- 238000013461 design Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000002599 functional magnetic resonance imaging Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 240000000972 Agathis dammara Species 0.000 description 1
- 229920002871 Dammar gum Polymers 0.000 description 1
- 230000003044 adaptive effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 230000006998 cognitive state Effects 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 201000010099 disease Diseases 0.000 description 1
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 238000010030 laminating Methods 0.000 description 1
- 238000002582 magnetoencephalography Methods 0.000 description 1
- 238000002610 neuroimaging Methods 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 230000008058 pain sensation Effects 0.000 description 1
- 230000035479 physiological effects, processes and functions Effects 0.000 description 1
- 230000003334 potential effect Effects 0.000 description 1
- 238000011084 recovery Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/243—Classification techniques relating to the number of classes
- G06F18/2431—Multiple classes
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/24—Detecting, measuring or recording bioelectric or biomagnetic signals of the body or parts thereof
- A61B5/316—Modalities, i.e. specific diagnostic methods
- A61B5/369—Electroencephalography [EEG]
- A61B5/372—Analysis of electroencephalograms
- A61B5/374—Detecting the frequency distribution of signals, e.g. detecting delta, theta, alpha, beta or gamma waves
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
- A61B5/7235—Details of waveform analysis
- A61B5/7264—Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems
- A61B5/7267—Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems involving training the classification device
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2415—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on parametric or probabilistic models, e.g. based on likelihood ratio or false acceptance rate versus a false rejection rate
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
- G06F3/01—Input arrangements or combined input and output arrangements for interaction between user and computer
- G06F3/011—Arrangements for interaction with the human body, e.g. for user immersion in virtual reality
- G06F3/015—Input arrangements based on nervous system activity detection, e.g. brain waves [EEG] detection, electromyograms [EMG] detection, electrodermal response detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/082—Learning methods modifying the architecture, e.g. adding, deleting or silencing nodes or connections
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2203/00—Indexing scheme relating to G06F3/00 - G06F3/048
- G06F2203/01—Indexing scheme relating to G06F3/01
- G06F2203/011—Emotion or mood input determined on the basis of sensed human body parameters such as pulse, heart rate or beat, temperature of skin, facial expressions, iris, voice pitch, brain activity patterns
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/12—Classification; Matching
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Theoretical Computer Science (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- General Health & Medical Sciences (AREA)
- Evolutionary Computation (AREA)
- General Physics & Mathematics (AREA)
- Biomedical Technology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Molecular Biology (AREA)
- Biophysics (AREA)
- Heart & Thoracic Surgery (AREA)
- Psychiatry (AREA)
- Veterinary Medicine (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Public Health (AREA)
- Mathematical Physics (AREA)
- Animal Behavior & Ethology (AREA)
- Surgery (AREA)
- Medical Informatics (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Pathology (AREA)
- Software Systems (AREA)
- Psychology (AREA)
- Human Computer Interaction (AREA)
- Neurosurgery (AREA)
- Neurology (AREA)
- Dermatology (AREA)
- Probability & Statistics with Applications (AREA)
- Computing Systems (AREA)
- Computational Linguistics (AREA)
- Fuzzy Systems (AREA)
- Physiology (AREA)
- Signal Processing (AREA)
- Measurement And Recording Of Electrical Phenomena And Electrical Characteristics Of The Living Body (AREA)
Abstract
The invention provides a motor imagery electroencephalogram signal classification method of a neural network, which comprises the following steps of: s1, acquiring electroencephalogram signals of a plurality of channels; s2, selecting a reference channel; s3, constructing an adjacent matrix A containing N-order neighbor connection relations by using standardized mutual information; s4, training a neural network model by using the adjacency matrix A obtained in the S3 to generate a motor imagery electroencephalogram signal recognition model; s5, acquiring real-time motor imagery electroencephalogram signals of a plurality of channels of a patient; s6, identifying the motor imagery electroencephalogram signals. The invention has the beneficial effects that: the connecting channel selection is performed by using the characteristics of the brain function connecting network in the motor imagery process, so that the method has neurophysiologic significance, meanwhile, the number of used channel connections is small, the calculated amount in the network model training can be effectively reduced, the self-adaptive matrix is introduced in the GCN network training process, the adjacency matrix can be adjusted for training, the undirected graph is converted into a weighted directed graph, and the influence relation among the channels can be effectively adjusted.
Description
Technical Field
The invention belongs to the technical field of brain electrical signals, and particularly relates to a motor imagery brain electrical signal classification method of a neural network.
Background
The motor imagery technique is developed by researching event related desynchronization and event related synchronization phenomena, and the core of the technique is to research potential activities of the brain in imagery rather than actual movements, motor imagery brain electrical signals can reflect the activities of the brain in imagery, different motor imagery tasks can trigger responses of different brain cortex areas, and after the motor imagery brain electrical signals are classified, a computer can generate control signals for a brain-computer interface system.
The motor imagery technique has been widely used in medical fields such as rehabilitation, and can convert motor imagery signals into control signals through a motor imagery brain-computer interface system so as to realize control of external devices, for example, for limb motor recovery in rehabilitation, generate motor imagery signals by training a patient when imagining movements, and then convert the signals into limb motor control signals so as to help the patient recover limb movement ability, and can also be used in fields such as pain management, and the like, and relieve pain sensation through imagination.
The brain connection network refers to the connection relation between different areas in the brain, and can be analyzed and researched by using graph theory and network science methods, and the construction and analysis of the brain connection network are generally based on brain imaging technologies, such as functional magnetic resonance imaging (fMRI), electroencephalogram (EEG), and Magnetoencephalography (MEG), etc., and through the analysis of the brain connection network, the structure and function of the brain can be deeply known, and the change and influence of the brain connection network under different diseases and cognitive states can be explored.
With the continuous development and perfection of MI technology, the application prospect in the medical field is wider. However, the current task of online identifying motor imagery of patients and the identification and classification of multiple motor imagery brain electrical signals are the problems to be solved in rehabilitation robot design, and further research and expansion of methods are needed.
Disclosure of Invention
In view of the above, the present invention aims to provide a motor imagery electroencephalogram signal classification method of a neural network, so as to solve at least one of the above technical problems.
In order to achieve the above purpose, the technical scheme of the invention is realized as follows:
the first aspect of the invention provides a motor imagery electroencephalogram signal classification method of a neural network, which comprises the following steps of:
s1, acquiring electroencephalogram signals of a plurality of channels;
s2, selecting an important electroencephalogram channel as a reference channel in the motor imagery process through mutual information characteristics and network centrality of a brain connection network;
s3, constructing an adjacent matrix A containing N-order neighbor connection relations by using standardized mutual information;
s4, training a neural network model by using the adjacency matrix A obtained in the S3 to generate a motor imagery electroencephalogram signal recognition model;
s5, acquiring real-time motor imagery electroencephalogram signals of a plurality of channels of a patient;
s6, inputting the acquired real-time motor imagery electroencephalogram signals into a motor imagery electroencephalogram signal identification model, and identifying the motor imagery electroencephalogram signals.
Further, the step S3 includes the following steps:
s31, initializing an adjacent matrix A and recording the connection relation between the channels;
s32, counting 1-order neighbor channels of a reference channel in each test standardized mutual information network;
s33, setting the mutual information value of the selected 1-order neighbor channel in each test network to 0;
s34, selecting a 1-order neighbor channel of the 1-order neighbor channel as a 2-order neighbor channel of a reference channel by using the method from S32 to S33, and adding corresponding connection on an adjacent matrix A;
s35, repeating the step S34 until an N-order neighbor relation is constructed, and obtaining an adjacent matrix A recording the N-order neighbor connection relation among the channels.
Further, the step S32 includes the following steps:
s321, selecting k channels with the maximum mutual information value with the reference channel as candidate 1-order neighbor channels;
s322, counting the occurrence times of each channel in candidate 1-order neighbor channels of all test times;
s323, selecting k channels with the largest occurrence number of each reference channel from all candidate 1-order neighbor channels according to the statistical result of S322 as real 1-order neighbor channels of the reference channel;
s324, setting a corresponding connection value in the adjacent matrix A to 1, namely connecting the reference channel with the 1-order neighbor channel.
Further, the step S1 includes the following steps:
s11, acquiring a plurality of types of motor imagery electroencephalogram signals of 64 leads by using an electroencephalogram acquisition device;
s12, performing 8-30Hz filtering processing on the acquired motor imagery electroencephalogram signals;
s13, performing 250Hz downsampling on the data obtained in the step S12;
s14, taking an electroencephalogram signal mean value of 200ms before the test as a baseline, and subtracting the baseline of each electrode from the electroencephalogram signal of each electrode obtained in the S13 to obtain an electroencephalogram signal after baseline correction;
s15, carrying out data segmentation processing according to time on the data obtained in the step S14;
s16, the data obtained in the S14 are processed according to the following steps: the scale of 1 is divided into training set data and test set data.
Further, the step S4 includes the following steps:
s41, inputting training set data obtained in the S1 and an adjacency matrix A obtained in the S35 into a neural network model;
s42, converting the adjacent matrix A obtained in the S35 into a Laplace standardized matrix through calculation;
s43, carrying out 2d convolution, standardization, activation and dropout processing on the training set data obtained in the step S1;
s44, performing graph convolution calculation on the training set data processed in the S43;
s45, carrying out 2d convolution, standardization, activation and dropout processing on the training set data subjected to the graph convolution calculation in S44;
s46, scaling the data obtained in the S45 by using a full connection layer, and giving probabilities of which category the different test motor imagery motion data belong to by using a softmax activation function;
s47, superposing the network modules of S43, S44, S45 and S46 to form a GCN neural network;
s48, training a neural network model, namely using cross Entropy as a loss function, using a self-adaptive time estimation method in a training process to perform gradient descent, and obtaining a motor imagery electroencephalogram signal recognition model through multiple iterations in a data transmission process which is one training iteration;
s49, comparing the motor imagery motion prediction result obtained in the S46 with a training set data real label, calculating the accuracy of a motor imagery electroencephalogram recognition model, and selecting the motor imagery electroencephalogram recognition model with the highest accuracy as a final motor imagery electroencephalogram recognition model;
s410, inputting the test set data obtained in the S1 into a final motor imagery electroencephalogram signal recognition model obtained in the S49, comparing the output result with a test set data label, and judging the accuracy of the final motor imagery electroencephalogram signal recognition model.
Furthermore, the sampling rate of the electroencephalogram signal acquisition system in the step S1 is 1000Hz.
Further, the step S2 includes the following steps:
s21, defining an electrode coverage area corresponding to each EEG lead as a node;
s22, calculating an entropy value of an electroencephalogram signal during motor imagery;
the mutual information between the two lead signals can be calculated by utilizing the entropy values of the two signals;
normalizing mutual information between two lead signals;
s23, calculating standardized mutual information between two channels in each test time to obtain a motor imagery electroencephalogram signal standardized mutual information network;
s24, calculating the intermediacy centrality of all nodes of the electroencephalogram signal standardized mutual information network;
s25, counting channels with medium centrality values larger than 0 in all test orders, and performing descending order arrangement on the channels to generate channel importance ordering;
s26, sorting channel importance middle and frontThe channels are used as important electroencephalogram channels, and the important electroencephalogram channels are used as reference channels.
A second aspect of the present invention provides an electronic device, including a processor and a memory communicatively connected to the processor and configured to store instructions executable by the processor, where the processor is configured to execute the motor imagery electroencephalogram classification method of the neural network described in the first aspect.
A third aspect of the present invention provides a server comprising at least one processor, and a memory communicatively coupled to the processor, the memory storing instructions executable by the at least one processor to cause the at least one processor to perform the motor imagery electroencephalogram classification method of the neural network of the first aspect.
A fourth aspect of the present invention provides a computer-readable storage medium storing a computer program which, when executed by a processor, implements the motor imagery electroencephalogram classification method of the neural network described in the first aspect.
Compared with the prior art, the motor imagery electroencephalogram signal classification method of the neural network has the following beneficial effects:
the motor imagery electroencephalogram signal classification method of the neural network provided by the invention has the advantages that the characteristics of the brain function connection network in the motor imagery process are used for connection channel selection, the neural physiology significance is realized, meanwhile, the number of used channel connections is small, the calculated amount in the network model training can be effectively reduced, the self-adaptive matrix is introduced in the GCN network training process, the adjacency matrix can be trained, the undirected graph is converted into the weighted graph, and the influence relation among the channels can be effectively adjusted.
Drawings
The accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this specification, illustrate embodiments of the invention and together with the description serve to explain the invention. In the drawings:
fig. 1 is a schematic flow chart of a classification method according to an embodiment of the invention.
Detailed Description
It should be noted that, without conflict, the embodiments of the present invention and features of the embodiments may be combined with each other.
The invention will be described in detail below with reference to the drawings in connection with embodiments.
Embodiment one:
as shown in fig. 1, a motor imagery electroencephalogram signal classification method of a neural network includes the following steps:
s1, acquiring electroencephalogram signals of a plurality of channels;
s2, selecting an important electroencephalogram channel as a reference channel in the motor imagery process through mutual information characteristics and network centrality of a brain connection network;
s3, constructing an adjacent matrix A containing N-order neighbor connection relations by using standardized mutual information;
s4, training a neural network model by using the adjacency matrix A obtained in the S3 to generate a motor imagery electroencephalogram signal recognition model;
s5, acquiring real-time motor imagery electroencephalogram signals of a plurality of channels of a patient;
s6, inputting the acquired real-time motor imagery electroencephalogram signals into a motor imagery electroencephalogram signal identification model, and identifying the motor imagery electroencephalogram signals.
S3 comprises the following steps:
s31, initializing an adjacent matrix A and recording the connection relation between the channels;
s32, counting 1-order neighbor channels of a reference channel in each test standardized mutual information network;
s33, setting the mutual information value of the selected 1-order neighbor channel in each test network to 0;
s34, selecting a 1-order neighbor channel of the 1-order neighbor channel as a 2-order neighbor channel of a reference channel by using the method from S32 to S33, and adding corresponding connection on an adjacent matrix A;
s35, repeating the step S34 until an N-order neighbor relation is constructed, and obtaining an adjacent matrix A recording the N-order neighbor connection relation among the channels.
Further, S32 includes the following steps:
s321, selecting k channels with the maximum mutual information value with the reference channel as candidate 1-order neighbor channels;
s322, counting the occurrence times of each channel in candidate 1-order neighbor channels of all test times;
s323, selecting k channels with the largest occurrence number of each reference channel from all candidate 1-order neighbor channels according to the statistical result of S322 as real 1-order neighbor channels of the reference channel;
s324, setting a corresponding connection value in the adjacent matrix A to 1, namely connecting the reference channel with the 1-order neighbor channel.
S1 comprises the following steps:
s11, acquiring a plurality of types of motor imagery electroencephalogram signals of 64 leads by using an electroencephalogram acquisition device;
s12, performing 8-30Hz filtering processing on the acquired motor imagery electroencephalogram signals;
s13, performing 250Hz downsampling on the data obtained in the step S12;
s14, taking an electroencephalogram signal mean value of 200ms before the test as a baseline, and subtracting the baseline of each electrode from the electroencephalogram signal of each electrode obtained in the S13 to obtain an electroencephalogram signal after baseline correction;
s15, carrying out data segmentation processing according to time on the data obtained in the step S14;
s16, the data obtained in the S14 are processed according to the following steps: the scale of 1 is divided into training set data and test set data.
Through extracting and processing the effective characteristics of the training set data, a foundation is laid for the subsequent neural network training.
S4 comprises the following steps:
s41, inputting training set data obtained in the S1 and an adjacency matrix A obtained in the S35 into a neural network model;
s42, converting the adjacent matrix A obtained in the step S35 into a Laplace standardized matrix through calculation, wherein the calculation formula is as follows:
；
wherein,is the sum of the adjacent matrix A and the identity matrix I;
s43, carrying out 2d convolution, standardization, activation and dropout processing on the training set data obtained in the step S1, wherein the formula is as follows:
；
wherein the method comprises the steps ofFor training set dataThe output at convolution layer 1;
s44, carrying out graph convolution calculation on the training set data processed in S43, wherein the calculation formula is as follows:
；
wherein,is data ofAt the output of the layer stack of the drawing,for the product of the dammar,are training parameters;
s45, carrying out 2d convolution, standardization, activation and dropout processing on the training set data subjected to the graph convolution calculation in S44, wherein the calculation formula is as follows:
；
is data ofLaminating the output of the layer on the graph;
s46, scaling the data obtained in the S45 by using a full connection layer, and giving the probability of which category the different test motor imagery motion data belong to by using a softmax activation function, wherein the calculation formula is as follows:
；
wherein,is a prediction result;
s47, superposing the network modules of S43, S44, S45 and S46 to form a GCN neural network;
s48, training a neural network model, namely using cross Entropy as a loss function, using a self-adaptive time estimation method in a training process to perform gradient descent, and obtaining a motor imagery electroencephalogram signal recognition model through multiple iterations in a data transmission process which is one training iteration;
s49, comparing the motor imagery motion prediction result obtained in the S46 with a training set data real label, and calculating the accuracy of a motor imagery electroencephalogram signal recognition model, wherein the calculation formula is as follows:
；
wherein acc is the accuracy of the model, TP is the positive sample predicted by the model as the positive class, TN is the negative sample predicted by the model as the negative class, FP is the negative sample predicted by the model as the positive class, FN is the positive sample predicted by the model as the negative class;
selecting a motor imagery electroencephalogram signal recognition model with highest accuracy as a final motor imagery electroencephalogram signal recognition model;
s410, inputting the test set data obtained in the S1 into a final motor imagery electroencephalogram signal recognition model obtained in the S49, comparing the output result with a test set data label, and judging the accuracy of the final motor imagery electroencephalogram signal recognition model.
And S1, the sampling rate of the electroencephalogram signal acquisition system is 1000Hz.
S2 comprises the following steps:
s21, defining an electrode coverage area corresponding to each EEG lead as a node;
s22, calculating an entropy value of an electroencephalogram signal during motor imagery, wherein a calculation formula is as follows:
；
x and Y are two different lead brain electrical signals in a motor imagery period, p (X) is the probability of the signal value X, namely the edge distribution probability of the signal X, and H (X) is the entropy of the signal;
the mutual information between the two lead signals can be calculated by using the entropy values of the two signals, and the calculation formula is as follows:
；
wherein I (X, Y) is mutual information between two lead signals, and H (X, Y) is joint entropy of the signals X and Y;
the mutual information between the two lead signals is standardized, and the calculation formula is as follows:
；
s23, calculating standardized mutual information between two channels in each test time to obtain a motor imagery electroencephalogram signal standardized mutual information network;
s24, calculating the intermediacy centrality of all nodes of the electroencephalogram signal standardized mutual information network;
the formula for calculating the intermediacy in S24 is as follows:
；
s, t, i represent nodes in the network,representing the number of all shortest paths from node s to node t,BC representing the number of paths through node i in all shortest paths from node s to node t i BC representing the extent to which node i acts as a bridge in the network i The larger the value is, the higher the importance of node i in the network is;
s25, counting channels with medium centrality values larger than 0 in all test orders, and performing descending order arrangement on the channels to generate channel importance ordering;
s26, sorting channel importance middle and frontThe channels are used as important electroencephalogram channels, and the important electroencephalogram channels are used as reference channels.
The invention has the beneficial effects that:
firstly, the characteristics of the brain function connection network in the motor imagery process are used for connection channel selection, so that the method has neurophysiologic significance, and meanwhile, the number of used channel connections is small, so that the calculated amount in the network model training can be effectively reduced;
secondly, an adaptive matrix is introduced in the GCN network training process, so that the adjacency matrix can be adjusted to train, an undirected graph is converted into a weighted directed graph, and the influence relation among channels can be effectively adjusted.
Thirdly, a trainable parameter theta is introduced, so that the data can be subjected to space-time optimization in a self-adaptive manner, and better classification performance is obtained.
The model provided by the invention has better performance in the aspect of motor imagery motion classification, and has better effects in different data sets and different EEG classification tasks in order to have wide application prospects in actual brain-computer interaction in consideration of the fact that channels and frequency bands play a great role in different EEG recognition tasks.
Embodiment two:
an electronic device includes a processor and a memory communicatively connected to the processor and configured to store instructions executable by the processor, where the processor is configured to execute the motor imagery electroencephalogram classification method of the neural network according to the first embodiment.
Embodiment III:
a server comprising at least one processor and a memory communicatively coupled to the processor, the memory storing instructions executable by the at least one processor to cause the at least one processor to perform a motor imagery electroencephalogram classification method of a neural network as in embodiment one.
Embodiment four:
a computer-readable storage medium storing a computer program which, when executed by a processor, implements the motor imagery electroencephalogram classification method of the neural network of the first embodiment.
Those of ordinary skill in the art will appreciate that the elements and method steps of each example described in connection with the embodiments disclosed herein may be implemented as electronic hardware, computer software, or combinations of both, and that the elements and steps of each example have been described generally in terms of functionality in the foregoing description to clearly illustrate this interchangeability of hardware and software. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the solution. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present invention.
In the several embodiments provided in this application, it should be understood that the disclosed methods and systems may be implemented in other ways. For example, the above-described division of units is merely a logical function division, and there may be another division manner when actually implemented, for example, a plurality of units or components may be combined or may be integrated into another system, or some features may be omitted or not performed. The units may or may not be physically separate, and components shown as units may or may not be physical units, may be located in one place, or may be distributed over a plurality of network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the embodiment of the present invention.
Finally, it should be noted that: the above embodiments are only for illustrating the technical solution of the present invention, and not for limiting the same; although the invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some or all of the technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit of the invention, and are intended to be included within the scope of the appended claims and description.
The foregoing description of the preferred embodiments of the invention is not intended to be limiting, but rather is intended to cover all modifications, equivalents, alternatives, and improvements that fall within the spirit and scope of the invention.
Claims (10)
1. The motor imagery electroencephalogram signal classification method of the neural network is characterized by comprising the following steps of:
s1, acquiring electroencephalogram signals of a plurality of channels;
s2, selecting an important electroencephalogram channel as a reference channel in the motor imagery process through mutual information characteristics and network centrality of a brain connection network;
s3, constructing an adjacent matrix A containing N-order neighbor connection relations by using standardized mutual information;
s4, training a neural network model by using the adjacency matrix A obtained in the S3 to generate a motor imagery electroencephalogram signal recognition model;
s5, acquiring real-time motor imagery electroencephalogram signals of a plurality of channels of a patient;
s6, inputting the acquired real-time motor imagery electroencephalogram signals into a motor imagery electroencephalogram signal identification model, and identifying the motor imagery electroencephalogram signals.
2. The motor imagery electroencephalogram signal classification method of a neural network according to claim 1, wherein the S3 comprises the steps of:
s31, initializing an adjacent matrix A and recording the connection relation between the channels;
s32, counting 1-order neighbor channels of a reference channel in each test standardized mutual information network;
s33, setting the mutual information value of the selected 1-order neighbor channel in each test network to 0;
s34, selecting a 1-order neighbor channel of the 1-order neighbor channel as a 2-order neighbor channel of a reference channel by using the method from S32 to S33, and adding corresponding connection on an adjacent matrix A;
s35, repeating the step S34 until an N-order neighbor relation is constructed, and obtaining an adjacent matrix A recording the N-order neighbor connection relation among the channels.
3. The motor imagery electroencephalogram signal classification method of a neural network according to claim 2, wherein: the step S32 includes the steps of:
s321, selecting k channels with the maximum mutual information value with the reference channel as candidate 1-order neighbor channels;
s322, counting the occurrence times of each channel in candidate 1-order neighbor channels of all test times;
s323, selecting k channels with the largest occurrence number of each reference channel from all candidate 1-order neighbor channels according to the statistical result of S322 as real 1-order neighbor channels of the reference channel;
s324, setting a corresponding connection value in the adjacent matrix A to 1, namely connecting the reference channel with the 1-order neighbor channel.
4. The motor imagery electroencephalogram signal classification method of a neural network according to claim 1, wherein the S1 comprises the steps of:
s11, acquiring a plurality of types of motor imagery electroencephalogram signals of 64 leads by using an electroencephalogram acquisition device;
s12, performing 8-30Hz filtering processing on the acquired motor imagery electroencephalogram signals;
s13, performing 250Hz downsampling on the data obtained in the step S12;
s14, taking an electroencephalogram signal mean value of 200ms before the test as a baseline, and subtracting the baseline of each electrode from the electroencephalogram signal of each electrode obtained in the S13 to obtain an electroencephalogram signal after baseline correction;
s15, carrying out data segmentation processing according to time on the data obtained in the step S14;
s16, the data obtained in the S14 are processed according to the following steps: the scale of 1 is divided into training set data and test set data.
5. The motor imagery electroencephalogram signal classification method of a neural network according to claim 2, wherein: the step S4 comprises the following steps:
s41, inputting training set data obtained in the S1 and an adjacency matrix A obtained in the S35 into a neural network model;
s42, converting the adjacent matrix A obtained in the S35 into a Laplace standardized matrix through calculation;
s43, carrying out 2d convolution, standardization, activation and dropout processing on the training set data obtained in the step S1;
s44, performing graph convolution calculation on the training set data processed in the S43;
s45, carrying out 2d convolution, standardization, activation and dropout processing on the training set data subjected to the graph convolution calculation in S44;
s46, scaling the data obtained in the S45 by using a full connection layer, and giving probabilities of which category the different test motor imagery motion data belong to by using a softmax activation function;
s47, superposing the network modules of S43, S44, S45 and S46 to form a GCN neural network;
s48, training a neural network model, namely using cross Entropy as a loss function, using a self-adaptive time estimation method in a training process to perform gradient descent, and obtaining a motor imagery electroencephalogram signal recognition model through multiple iterations in a data transmission process which is one training iteration;
s49, comparing the motor imagery motion prediction result obtained in the S46 with a training set data real label, calculating the accuracy of a motor imagery electroencephalogram recognition model, and selecting the motor imagery electroencephalogram recognition model with the highest accuracy as a final motor imagery electroencephalogram recognition model;
s410, inputting the test set data obtained in the S1 into a final motor imagery electroencephalogram signal recognition model obtained in the S49, comparing the output result with a test set data label, and judging the accuracy of the final motor imagery electroencephalogram signal recognition model.
6. The motor imagery electroencephalogram signal classification method of a neural network according to claim 1, wherein: and the sampling rate of the electroencephalogram signal acquisition system in the step S1 is 1000Hz.
7. The motor imagery electroencephalogram signal classification method of a neural network according to claim 1, wherein: the step S2 comprises the following steps:
s21, defining an electrode coverage area corresponding to each EEG lead as a node;
s22, calculating an entropy value of an electroencephalogram signal during motor imagery;
the mutual information between the two lead signals can be calculated by utilizing the entropy values of the two signals;
normalizing mutual information between two lead signals;
s23, calculating standardized mutual information between two channels in each test time to obtain a motor imagery electroencephalogram signal standardized mutual information network;
s24, calculating the intermediacy centrality of all nodes of the electroencephalogram signal standardized mutual information network;
s25, counting channels with medium centrality values larger than 0 in all test orders, and performing descending order arrangement on the channels to generate channel importance ordering;
s26, sorting channel importance middle and frontThe channels are used as important electroencephalogram channels, and the important electroencephalogram channels are used as reference channels.
8. An electronic device comprising a processor and a memory communicatively coupled to the processor for storing processor-executable instructions, characterized in that: the processor is configured to execute the motor imagery electroencephalogram signal classification method of the neural network according to any one of claims 1 to 7.
9. A server, characterized by: comprising at least one processor and a memory communicatively connected to the processor, the memory storing instructions executable by the at least one processor to cause the at least one processor to perform the motor imagery electroencephalogram classification method of the neural network of any one of claims 1-7.
10. A computer-readable storage medium storing a computer program, characterized in that: the computer program, when executed by a processor, implements the motor imagery electroencephalogram classification method of the neural network of any one of claims 1-7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311533871.5A CN117251807B (en) | 2023-11-17 | 2023-11-17 | Motor imagery electroencephalogram signal classification method of neural network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311533871.5A CN117251807B (en) | 2023-11-17 | 2023-11-17 | Motor imagery electroencephalogram signal classification method of neural network |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117251807A true CN117251807A (en) | 2023-12-19 |
| CN117251807B CN117251807B (en) | 2024-02-13 |
Family
ID=89135406
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311533871.5A Active CN117251807B (en) | 2023-11-17 | 2023-11-17 | Motor imagery electroencephalogram signal classification method of neural network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117251807B (en) |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112698720A (en) * | 2020-12-23 | 2021-04-23 | 杭州电子科技大学 | Movement imagery identification method based on mutual information feature extraction and multi-person fusion |
| CN114209320A (en) * | 2021-11-17 | 2022-03-22 | 山东师范大学 | Depression patient electroencephalogram recognition system based on mutual graph information maximization |
| CN114330338A (en) * | 2022-01-13 | 2022-04-12 | 东北电力大学 | Program language identification system and method fusing associated information |
| CN114360254A (en) * | 2022-03-17 | 2022-04-15 | 湖南工商大学 | Space-time-based traffic flow prediction method and related equipment |
| CN114387668A (en) * | 2021-12-28 | 2022-04-22 | 杭州电子科技大学 | Classification method based on multi-level neuromuscular coupling characteristic information fusion |
| WO2023098303A1 (en) * | 2021-12-01 | 2023-06-08 | 浙江大学 | Real-time epileptic seizure detecting and monitoring system for video electroencephalogram examination of epilepsy |
| CN116541751A (en) * | 2023-07-03 | 2023-08-04 | 中国医学科学院生物医学工程研究所 | Electroencephalogram signal classification method based on brain function connection network characteristics |
-
2023
- 2023-11-17 CN CN202311533871.5A patent/CN117251807B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112698720A (en) * | 2020-12-23 | 2021-04-23 | 杭州电子科技大学 | Movement imagery identification method based on mutual information feature extraction and multi-person fusion |
| CN114209320A (en) * | 2021-11-17 | 2022-03-22 | 山东师范大学 | Depression patient electroencephalogram recognition system based on mutual graph information maximization |
| WO2023098303A1 (en) * | 2021-12-01 | 2023-06-08 | 浙江大学 | Real-time epileptic seizure detecting and monitoring system for video electroencephalogram examination of epilepsy |
| CN114387668A (en) * | 2021-12-28 | 2022-04-22 | 杭州电子科技大学 | Classification method based on multi-level neuromuscular coupling characteristic information fusion |
| CN114330338A (en) * | 2022-01-13 | 2022-04-12 | 东北电力大学 | Program language identification system and method fusing associated information |
| CN114360254A (en) * | 2022-03-17 | 2022-04-15 | 湖南工商大学 | Space-time-based traffic flow prediction method and related equipment |
| CN116541751A (en) * | 2023-07-03 | 2023-08-04 | 中国医学科学院生物医学工程研究所 | Electroencephalogram signal classification method based on brain function connection network characteristics |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117251807B (en) | 2024-02-13 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113693613B (en) | Electroencephalogram signal classification method, electroencephalogram signal classification device, computer equipment and storage medium | |
| CN107126193B (en) | Multivariate causal relationship analysis method based on hysteresis order self-adaptive selection | |
| CN108960299B (en) | Method for identifying multi-class motor imagery electroencephalogram signals | |
| CN104771163A (en) | Electroencephalogram feature extraction method based on CSP and R-CSP algorithms | |
| CN110333783B (en) | Irrelevant gesture processing method and system for robust electromyography control | |
| CN112990008B (en) | Emotion recognition method and system based on three-dimensional characteristic diagram and convolutional neural network | |
| CN109063639A (en) | A kind of method of real-time prediction Brain behavior | |
| KR20190033972A (en) | Method and apparatus for an automatic artifact removal of EEG based on a deep leaning algorithm | |
| CN116861967A (en) | Data enhancement method and device based on impulse neural network | |
| CN113761777B (en) | HP-OVMD-based ultra-short-term photovoltaic power prediction method | |
| CN114027786B (en) | Sleep breathing disorder detection method and system based on self-supervision type memory network | |
| CN117574059A (en) | High-resolution brain-electrical-signal deep neural network compression method and brain-computer interface system | |
| KR20190073330A (en) | Method and apparatus for an automatic artifact removal of EEG based on a deep leaning algorithm | |
| CN116244633A (en) | Motor imagery electroencephalogram signal classification method and system | |
| CN116541751B (en) | Electroencephalogram signal classification method based on brain function connection network characteristics | |
| CN115859185A (en) | Electroencephalogram emotion recognition method based on pulse convolution neural network | |
| CN117251807B (en) | Motor imagery electroencephalogram signal classification method of neural network | |
| CN110432899B (en) | Electroencephalogram signal identification method based on depth stacking support matrix machine | |
| CN117224082A (en) | Dynamic brain function network learning method based on graph learning optimization algorithm | |
| CN115721323A (en) | Brain-computer interface signal identification method and system and electronic equipment | |
| Lu et al. | On model-guided neural networks for system identification | |
| CN116115240A (en) | Electroencephalogram emotion recognition method based on multi-branch chart convolution network | |
| CN114816076A (en) | Brain-computer interface computing processing and feedback system and method | |
| Luo et al. | Classification algorithm for motor imagery EEG signals based on parallel DAMSCN-LSTM | |
| Silaen et al. | EEG Signal Processing For Motor Imagery Direction of Hand Movement Using the Brain Computer Interface |
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 |