CN115105093A - EEG signal classification and identification method based on power spectral density predetermined frequency band - Google Patents
EEG signal classification and identification method based on power spectral density predetermined frequency band Download PDFInfo
- Publication number
- CN115105093A CN115105093A CN202210594845.2A CN202210594845A CN115105093A CN 115105093 A CN115105093 A CN 115105093A CN 202210594845 A CN202210594845 A CN 202210594845A CN 115105093 A CN115105093 A CN 115105093A
- Authority
- CN
- China
- Prior art keywords
- signal
- frequency
- spectral density
- band
- power spectral
- 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.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 31
- 230000003595 spectral effect Effects 0.000 title claims abstract description 27
- 238000001914 filtration Methods 0.000 claims abstract description 19
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 18
- 238000000605 extraction Methods 0.000 claims abstract description 18
- 238000003062 neural network model Methods 0.000 claims abstract description 7
- 238000001228 spectrum Methods 0.000 claims abstract description 4
- 239000011159 matrix material Substances 0.000 claims description 33
- 238000013528 artificial neural network Methods 0.000 claims description 11
- 238000004364 calculation method Methods 0.000 claims description 5
- 230000008901 benefit Effects 0.000 abstract description 2
- 238000012545 processing Methods 0.000 description 10
- 210000004556 brain Anatomy 0.000 description 8
- 230000002087 whitening effect Effects 0.000 description 8
- 230000033001 locomotion Effects 0.000 description 7
- 230000008569 process Effects 0.000 description 7
- 230000033764 rhythmic process Effects 0.000 description 6
- 238000005516 engineering process Methods 0.000 description 4
- 238000007781 pre-processing Methods 0.000 description 4
- 238000004458 analytical method Methods 0.000 description 3
- 239000000284 extract Substances 0.000 description 3
- 230000006870 function Effects 0.000 description 3
- 210000002569 neuron Anatomy 0.000 description 3
- 238000007635 classification algorithm Methods 0.000 description 2
- 238000013145 classification model Methods 0.000 description 2
- 238000002474 experimental method Methods 0.000 description 2
- 230000003993 interaction Effects 0.000 description 2
- 238000010606 normalization Methods 0.000 description 2
- 230000002093 peripheral effect Effects 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 238000012706 support-vector machine Methods 0.000 description 2
- 230000009466 transformation Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000007177 brain activity Effects 0.000 description 1
- 210000003710 cerebral cortex Anatomy 0.000 description 1
- 238000000354 decomposition reaction Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000005611 electricity Effects 0.000 description 1
- 230000005281 excited state Effects 0.000 description 1
- 230000003155 kinesthetic effect Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003387 muscular Effects 0.000 description 1
- 238000003909 pattern recognition Methods 0.000 description 1
- 210000001428 peripheral nervous system Anatomy 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 238000000611 regression analysis Methods 0.000 description 1
- 230000000284 resting effect Effects 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 230000035807 sensation Effects 0.000 description 1
- 230000002269 spontaneous effect Effects 0.000 description 1
- 238000012549 training Methods 0.000 description 1
- 230000017105 transposition Effects 0.000 description 1
Images
Classifications
-
- 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]
-
- 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
-
- 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
- A61B5/7203—Signal processing specially adapted for physiological signals or for diagnostic purposes for noise prevention, reduction or removal
-
- 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
- A61B5/7225—Details of analog processing, e.g. isolation amplifier, gain or sensitivity adjustment, filtering, baseline or drift compensation
-
- 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
-
- 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
- G06N3/084—Backpropagation, e.g. using gradient descent
Landscapes
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Biomedical Technology (AREA)
- General Health & Medical Sciences (AREA)
- Biophysics (AREA)
- Molecular Biology (AREA)
- Veterinary Medicine (AREA)
- Medical Informatics (AREA)
- Pathology (AREA)
- Surgery (AREA)
- Animal Behavior & Ethology (AREA)
- Heart & Thoracic Surgery (AREA)
- Public Health (AREA)
- Psychiatry (AREA)
- Signal Processing (AREA)
- Artificial Intelligence (AREA)
- Physiology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Psychology (AREA)
- Evolutionary Computation (AREA)
- Theoretical Computer Science (AREA)
- Mathematical Physics (AREA)
- Computational Linguistics (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Software Systems (AREA)
- Power Engineering (AREA)
- Fuzzy Systems (AREA)
- Measurement And Recording Of Electrical Phenomena And Electrical Characteristics Of The Living Body (AREA)
Abstract
The invention relates to an EEG signal classification and identification method based on a power spectral density predetermined frequency band, which comprises the following steps: acquiring an original signal, and performing filtering pretreatment on the original signal by using a band-pass filter; carrying out power spectral density estimation on the preprocessed signal to obtain a frequency corresponding to the maximum power value on a frequency spectrum; determining a frequency band according to a set length by taking the frequency corresponding to the maximum value as a center; performing secondary band-pass filtering on the preprocessed signal under the frequency band to obtain a target frequency component; performing feature extraction operation on the target frequency components on a space domain by using a common space mode algorithm to obtain target features; and classifying the target characteristics by using a neural network model, and outputting to obtain an EEG signal identification result. Compared with the prior art, the method has the advantages of improving the decoding accuracy of the motor imagery EEG signal and the like.
Description
Technical Field
The invention relates to the field of electroencephalogram signal processing, in particular to an EEG signal classification and identification method based on a power spectral density predetermined frequency band.
Background
In recent years, Brain Computer Interface (BCI) technology has been widely studied, and has been actively promoted in the rehabilitation process of patients after stroke. The brain-computer interface technology is a novel human-computer interaction technology which does not depend on a conventional brain information output channel (including a peripheral nervous system and muscular tissues), and can enable a user to directly interact with an external environment through the brain or control various types of external control equipment, such as a wheelchair, a robot and the like. Motor imagery EEG (brain waves) is an important paradigm in brain-computer interface technology, which has the advantage of not requiring additional external stimuli, but requires a lengthy training process on the subject. The motor imagery EEG signal is an unstable signal, and there is a large difference between different subjects, and the performance of the same subject at different times will also have a certain difference, so that the EEG signal is also easily interfered by extraneous signals, such as power frequency interference of external environment, electromyographic signals generated by body movement, and the like. Therefore, improving the decoding accuracy of the motor imagery EEG signal is still a research difficulty in the field of brain-computer interface.
Patent CN202010022435.1 discloses an online processing method for a motor imagery EEG signal, which uses a band-pass filter, a co-space mode and a support vector machine to realize classification and identification of the motor imagery EEG signal. However, the method is insufficient for the extraction process of the components related to motor imagery in the EEG signal, and the decoding accuracy of the final EEG signal is affected.
Patent CN201810044806.9 discloses an energy feature-based identification method for motor imagery electroencephalogram signals, which extracts energy features in EEG signals, and classifies the EEG signals by using a support vector machine based on a radial basis kernel function, in the classification and identification process, only the energy features of the signals are used, and spatial features among channels are not considered, so that the decoding accuracy of the final EEG signals is affected.
Disclosure of Invention
The invention aims to overcome the defects in the prior art and provide an EEG signal classification and identification method based on a power spectral density predetermined frequency band, which improves the traditional identification and classification algorithm combining a co-space mode and a neural network classifier and finally improves the decoding accuracy of the motor imagery EEG signal.
The purpose of the invention can be realized by the following technical scheme:
an EEG signal classification identification method based on a power spectral density predetermined frequency band, comprising:
s1, acquiring an original signal, and performing filtering pretreatment on the original signal by using a band-pass filter;
s2, carrying out power spectral density estimation on the preprocessed signal to obtain a frequency corresponding to the maximum power value on the frequency spectrum;
s3, determining a frequency band according to a set length by taking the frequency corresponding to the maximum value as a center;
s4, carrying out secondary band-pass filtering on the preprocessed signals under the frequency band to obtain target frequency components;
s5, performing feature extraction operation on the target frequency components on the airspace by using a common space mode algorithm to obtain target features;
and S6, classifying the target features by using the neural network model, and outputting to obtain an EEG signal recognition result.
Further, in step S1, a butterworth bandpass digital filter of 1 to 30Hz is used to perform filtering operation on the original signal, so as to remove baseline drift of the original signal and interference components of the power frequency signal.
Further, in step S2, a power spectral density calculation is performed on each channel of the preprocessed signal, so as to obtain a frequency corresponding to a maximum power value therein.
Further, in step S3, the set length of the frequency band is determined to be 3-6 Hz.
Further, in step S4, the secondary bandpass filtering is to perform a filtering operation on the preprocessed signal using a butterworth bandpass digital filter in the range of the frequency band.
Further, in step S5, the co-spatial mode algorithm includes: and diagonalizing the covariance matrix of the target frequency components to generate a group of optimal spatial filters, then finding a projection matrix, and projecting each target frequency component to the same public space, wherein the variance value of each target frequency component can be maximally distinguished in the public space, so that the target characteristics are obtained.
Further, in step S6, the classifier in the neural network model employs an error back propagation neural network.
Compared with the prior art, the invention has the following beneficial effects:
1. the invention provides a strategy for predetermining a frequency band based on power spectral density, and extracts components most relevant to motor imagery in an EEG signal to perform feature extraction and feature classification operation, so that the classification performance of the traditional motor imagery EEG signal processing algorithm using a common space mode and a neural network is enhanced, and the decoding accuracy of the motor imagery EEG signal is finally improved.
2. The invention can effectively classify and recognize the motor imagery EEG signal, judge the tested motor intention, convert the intention into a corresponding control instruction and transmit the control instruction to the relevant peripheral equipment, and the peripheral equipment can execute the corresponding operation after receiving the instruction.
Drawings
Fig. 1 is a schematic structural diagram of a brain-computer interface system according to the present invention.
FIG. 2 is a flowchart illustrating an EEG signal classification and identification method according to the present invention.
Detailed Description
The invention is described in detail below with reference to the figures and specific embodiments. The present embodiment is implemented on the premise of the technical solution of the present invention, and a detailed implementation manner and a specific operation process are given, but the scope of the present invention is not limited to the following embodiments.
The embodiment provides an EEG signal classification and identification method based on a power spectral density predetermined frequency band, which is applied to a brain-computer interface system of a certain model. The classification identification of the EEG signal is a core part in the whole brain-computer interface system, and the identification accuracy determines the capability of a tested object to directly interact with the outside through the brain electricity, which affects the performance of the whole system. However, because the EEG signal has the characteristics of instability, low signal-to-noise ratio and high complexity, the accurate analysis and processing of the EEG signal often suffer from a lot of difficulties, and the recognition accuracy is difficult to maintain a high value, the EEG signal decoding process is still a challenging research work. For the problem, the embodiment provides an EEG signal classification and identification method based on a power spectral density predetermined frequency band, which improves the traditional recognition and classification method combining a common spatial mode and a neural network classifier, and improves the classification accuracy of EEG signals.
As shown in fig. 1, a brain-computer interface system can be generally divided into three parts: the device comprises an electroencephalogram signal acquisition module, a signal processing and identifying module and a control signal output/external equipment execution module.
The signal processing and identifying module is the most core module in the brain-computer interface system and comprises a series of processing on the original brain electrical signals obtained in the previous step. The stage comprises three steps of preprocessing, feature extraction and feature classification, wherein noise contained in an original signal is filtered as much as possible, key components reflecting brain activities are left, and feature components related to the attempted movement intention, such as energy information of the spontaneous electroencephalogram/rhythm of Motor Imagery (MI) and feature information of amplitude, phase and the like in induced electroencephalogram, are obtained through feature extraction operation. And finally, classifying the characteristics through a classification identification algorithm so as to identify the tested intention, and further converting the tested intention into a corresponding control signal to be transmitted to an application and outside interaction module. In this embodiment, the signal processing module is focused on, and power spectral density is used to improve the recognition and classification algorithm combining the conventional common spatial mode and the neural network classifier, thereby improving the recognition capability of the system on the EEG signal.
Specifically, the signal processing and identifying module executes an EEG signal classification and identification method based on a predetermined frequency band of power spectral density through software, as shown in fig. 2, and includes the following steps:
firstly, signal preprocessing:
and acquiring an original signal, and performing filtering pretreatment on the original signal by using a band-pass filter.
Secondly, feature extraction:
extracting frequency components of a small section of interval near the average value of the frequency with the maximum energy displayed in the respective power spectral densities of all channels of the preprocessed signals by using a method for pre-determining a frequency band by using the power spectral densities, wherein the extracted frequency components are target frequency components and serve as input of next feature extraction;
and performing feature extraction operation on the target frequency components on a space domain by using a common space mode algorithm to obtain target features.
Thirdly, feature classification:
and classifying the target characteristics by using a neural network model, and outputting to obtain an EEG signal identification result.
In the signal preprocessing process, a 1-30 Hz Butterworth band-pass digital filter is used for filtering an original signal, and the main function is to remove interference components such as baseline drift of the signal and a 50Hz power frequency signal. The butterworth bandpass digital filter may be represented as a function of the square of the amplitude with respect to frequency, as follows:
where N denotes the order of the digital filter, ω denotes the signal frequency, ω c The cut-off frequency is indicated and H (×) the amplitude.
The principle of the classification and identification method of the embodiment is as follows: the motor imagery EEG signal is closely Related to the motor-Related mu rhythm (8-13 Hz) and beta rhythm (13-30 Hz) in the alpha rhythm, and the specific expression is that when the tested person executes or imagines the specific limb movement, the motor sensation area Related to the movement in the cerebral cortex is in an excited state, the energy of the mu/beta rhythm is reduced, and the phenomenon is called Event-Related Desynchronization (ERD); after a few seconds the region resumes a resting state again, the μ/β rhythm energy rises, a phenomenon known as Event-Related Synchronization (ERS). Studies have shown that ERD and ERS phenomena do not occur independently, and tend to occur concomitantly when performing or imagining a particular limb movement. Taking left/right hand motor imagery as an example, when the subject is subjected to a left hand motor imagery task, the motion sensing area of the right half brain, namely the area near the C4 electrode, is subjected to an obvious ERD phenomenon; at the same time, the kinesthetic region of the left brain, i.e., the region near the C3 electrode, is subject to significant ERS. When the subject is trying to perform the right hand motor imagery task, the region where the ERD/ERS phenomenon occurs is just the opposite. Thus, the frequency value with the largest energy in the channel signals of C3, C4 and the like which are highly relevant to the motor imagery task can be determined by using the power spectral density. Then, a frequency band is taken near the value, the preprocessed signal is subjected to secondary band-pass filtering, signal components which are most relevant to motor imagery in the EEG signal can be extracted, and then feature extraction and feature classification operations can be carried out on the signal components, so that the decoding of the attempted movement intention is realized. In the feature extraction process, a common space mode algorithm is used, and the key idea of the algorithm is to analyze signals in a common space and fully consider the spatial features among all channels. In summary, the algorithm for analyzing and processing the EEG signals according to the motor imagery provided by the embodiment can significantly improve the accuracy of signal decoding.
In the EEG signal classification recognition method:
the feature extraction process specifically comprises the following steps:
step 1, power spectral density estimation is carried out on the preprocessed signals, and a spectrogram is drawn.
And 2, determining a frequency band according to the frequency corresponding to the maximum power in the spectrogram, namely determining the frequency band according to a set length by taking the frequency corresponding to the maximum value as a center. The set length of the frequency band is determined to be 3-6 Hz, and 4Hz is preferably adopted in the embodiment.
And 3, carrying out secondary band-pass filtering on the preprocessed signals under the frequency band to obtain target frequency components. The second-pass band filtering is to perform a filtering operation on the preprocessed signal within the frequency band using a butterworth band-pass digital filter.
And 4, performing feature extraction operation on the target frequency components on the airspace by using a common space mode algorithm to obtain target features.
In the execution process of the feature extraction algorithm, the frequency corresponding to the maximum power value in the power spectral density analysis of each channel needs to be determined, and the corresponding formula is as follows:
f max ＝argmaxPSD(f)
in determining the frequency interval of the second band-pass filtering, f is required max For the central extension interval, the corresponding formula is:
f band ＝[f max -2,f max +2]
the digital filter used in the second bandpass filtering operation is the same as that used in the preprocessing step.
The signal components after the second band-pass filtering have higher correlation with the motor imagery, and the signal components are input to the subsequent feature extraction step, so that more effective features can be extracted. In the feature extraction module, a common spatial mode algorithm is used, the algorithm can effectively extract spatial features in signals and is widely used for identification and classification of EEG signals, and the detailed principle and the specific steps of the algorithm are as follows:
the common spatial mode (CSP) is a spatial filter and is widely used in feature extraction of a motor imagery EEG signal. The key idea of the algorithm is to analyze the signal in co-space. In a classification scene of left/right hand motor imagery EEG signals, two types of signals are used, two covariance matrixes are diagonalized, a group of optimal spatial filters are generated, a projection matrix is found, the two types of signals are projected to the same public space, and in the space, the variance values of the two types of signals can be maximally distinguished. If the signal is in more than two motor imagery task scenes, a cascade mode can be used, and a plurality of two classification modules are adopted to realize multi-classification processing of the signal. The detailed steps of the algorithm are as follows:
for example, for a classification scenario of left/right hand motor imagery EEG signals, assume that the target frequency component derived from a single EEG experiment data is E ∈ R N*M Wherein N is the number of electrode channels, M is the number of single data sampling points, and the covariance matrix normalization operation is firstly carried out on each experimental data:
wherein, C represents a result matrix obtained after the covariance matrix normalization operation is carried out on each experimental data, and T represents the transposition operation is carried out on the matrix.
Then, within the class, the left-hand and right-hand cases are respectively averaged,
which represents the average of the left hand,
representing the right-hand average, and then adding the two to obtain a synthesized spatial covariance matrix:
wherein,
representing the left-hand average normalized spatial covariance matrix,
representing the right-hand mean normalized spatial covariance matrix, C C Is represented by
And
a synthesized spatial covariance matrix.
Then to C C Decomposing the eigenvalues, and arranging the eigenvectors according to the descending order of the eigenvalues:
wherein Uc representsEigenvector matrix, λ C Representing a matrix of eigenvalues
According to the whitening matrix
To pair
And
and carrying out whitening transformation to obtain a whitened left-hand/right-hand average covariance matrix as follows:
wherein,
representing the left-hand average normalized spatial covariance matrix,
representing the right-hand average normalized spatial covariance matrix, P being the whitening matrix, S L And S R The whitened left-hand and right-hand mean covariance matrices, respectively.
Based on the characteristics of the whitening transformation, the whitened matrix S L And S R Can be represented by a common feature vector a:
S L ＝Aλ L A T
S R ＝Aλ R A T
λ L +λ R ＝I
wherein A is a common feature vector, λ L Is the whitened matrix S L Eigenvalue matrix in eigenvalue decomposition, λ R Is the whitened matrix S R And I is an identity matrix.
From the above equation, the matrix S is obtained after left-hand whitening L Taking the maximum eigenvalue, the matrix S after right-hand whitening R Minimum eigenvalues must be achieved and therefore two types of signals can be distinguished. And multiplying the transpose of the common feature vector A by the whitening matrix P to obtain a projection matrix W, then projecting the single EEG experimental data E by using the projection matrix, and finally calculating a variance value to obtain the feature processed by the CSP. The characteristic value obtained by CSP calculation can be used as the input of the feedforward neural network classification model. The detailed calculation steps are as follows:
W＝A T P
Z p ＝WE,p＝1,2,3,...,2m
wherein, A is a common feature vector, P is a whitening matrix, W is a projection matrix, E is a target frequency component obtained by single EEG experimental data, P is a feature sequence number, Z is a result matrix of E after projection, var (x) is a variance value of a calculation vector, f p The characteristic value is finally calculated.
In this embodiment, an error back propagation neural network, referred to as a bp (back propagation) neural network, is selected as the classifier in the neural network model. The BP neural network has the characteristics of simple structure, nonlinearity, strong adaptability and the like, and is widely applied to the fields of pattern recognition, regression analysis and the like.
In the model structure of the BP neural network, there are three basic components, namely, an input layer, a hidden layer, and an output layer. In the classification model of this embodiment, the input layer includes 3 neurons, the middle layer is a single hidden layer, which includes 10 neurons, and the output layer includes 2 neurons.
The foregoing detailed description of the preferred embodiments of the invention has been presented. It should be understood that numerous modifications and variations could be devised by those skilled in the art in light of the present teachings without departing from the inventive concepts. Therefore, the technical solutions available to those skilled in the art through logic analysis, reasoning and limited experiments based on the prior art according to the concept of the present invention should be within the scope of protection defined by the claims.
Claims (7)
1. A power spectral density predetermined frequency band-based EEG signal classification and identification method is characterized by comprising the following steps:
s1, acquiring an original signal, and performing filtering pretreatment on the original signal by using a band-pass filter;
s2, carrying out power spectrum density estimation on the preprocessed signal to obtain the frequency corresponding to the maximum power value on the frequency spectrum;
s3, determining a frequency band according to a set length by taking the frequency corresponding to the maximum value as a center;
s4, carrying out secondary band-pass filtering on the preprocessed signals under the frequency band to obtain target frequency components;
s5, performing feature extraction operation on the target frequency components on the airspace by using a common space mode algorithm to obtain target features;
and S6, classifying the target features by using the neural network model, and outputting to obtain an EEG signal recognition result.
2. The EEG signal classification and identification method based on the power spectral density predetermined frequency band as claimed in claim 1, wherein in step S1, a Butterworth band-pass digital filter of 1-30 Hz is used to filter the original signal to remove the baseline drift of the original signal and the interference component of the power frequency signal.
3. The EEG signal classification and identification method based on power spectral density predetermined frequency bands according to claim 1, wherein in step S2, power spectral density calculation is performed on each channel of the preprocessed signal to obtain the frequency corresponding to the maximum power value.
4. The EEG signal classification and identification method based on the power spectral density predetermined frequency band according to claim 1, wherein in step S3, the set length of the frequency band is determined to be 3-6 Hz.
5. The EEG signal classification and identification method based on power spectral density predetermined frequency bands according to claim 1, characterized in that in step S4, the secondary band-pass filtering is to filter the preprocessed signal by using a Butterworth band-pass digital filter in the frequency band.
6. The EEG signal classification and identification method based on power spectral density predetermined frequency bands according to claim 1, wherein in step S5, the co-spatial mode algorithm comprises: and diagonalizing the covariance matrix of the target frequency components to generate a group of optimal spatial filters, then finding a projection matrix, and projecting the target frequency components to the same public space, wherein the variance values of the target frequency components can be maximally distinguished in the public space, so that the target characteristics are obtained.
7. The EEG signal classification and identification method based on power spectral density predetermined frequency bands according to claim 1, wherein in step S6, the classifier in the neural network model employs an error back propagation neural network.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210594845.2A CN115105093A (en) | 2022-05-27 | 2022-05-27 | EEG signal classification and identification method based on power spectral density predetermined frequency band |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210594845.2A CN115105093A (en) | 2022-05-27 | 2022-05-27 | EEG signal classification and identification method based on power spectral density predetermined frequency band |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115105093A true CN115105093A (en) | 2022-09-27 |
Family
ID=83327393
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210594845.2A Pending CN115105093A (en) | 2022-05-27 | 2022-05-27 | EEG signal classification and identification method based on power spectral density predetermined frequency band |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115105093A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115982558A (en) * | 2023-03-17 | 2023-04-18 | 同心智医科技(北京)有限公司 | Electroencephalogram movement intention classification model establishing method and application thereof |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105809124A (en) * | 2016-03-06 | 2016-07-27 | 北京工业大学 | DWT- and Parametric t-SNE-based characteristic extracting method of motor imagery EEG(Electroencephalogram) signals |
| CN108542383A (en) * | 2017-09-25 | 2018-09-18 | 同济大学 | EEG signal identification method, system, medium based on Mental imagery and equipment |
| CN108888264A (en) * | 2018-05-03 | 2018-11-27 | 南京邮电大学 | EMD and CSP merges power spectral density brain electrical feature extracting method |
| CN109480834A (en) * | 2018-12-20 | 2019-03-19 | 浙江大学 | A kind of Method of EEG signals classification based on quick multiple dimension empirical mode decomposition |
| CN112949533A (en) * | 2021-03-15 | 2021-06-11 | 成都信息工程大学 | Motor imagery electroencephalogram identification method based on relative wavelet packet entropy brain network and improved version lasso |
-
2022
- 2022-05-27 CN CN202210594845.2A patent/CN115105093A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105809124A (en) * | 2016-03-06 | 2016-07-27 | 北京工业大学 | DWT- and Parametric t-SNE-based characteristic extracting method of motor imagery EEG(Electroencephalogram) signals |
| CN108542383A (en) * | 2017-09-25 | 2018-09-18 | 同济大学 | EEG signal identification method, system, medium based on Mental imagery and equipment |
| CN108888264A (en) * | 2018-05-03 | 2018-11-27 | 南京邮电大学 | EMD and CSP merges power spectral density brain electrical feature extracting method |
| CN109480834A (en) * | 2018-12-20 | 2019-03-19 | 浙江大学 | A kind of Method of EEG signals classification based on quick multiple dimension empirical mode decomposition |
| CN112949533A (en) * | 2021-03-15 | 2021-06-11 | 成都信息工程大学 | Motor imagery electroencephalogram identification method based on relative wavelet packet entropy brain network and improved version lasso |
Non-Patent Citations (2)
| Title |
|---|
| 齐鹏, 茹画, 高凌云, 等.: "心理疲劳的神经机制——脑连接组的新见解", 《ENGINEERING》, vol. 5, no. 2, 31 December 2019 (2019-12-31), pages 276 - 286 * |
| 齐鹏: "三冗余系统在硝酸四合一机组中的应用研究", 《万方》, 15 June 2015 (2015-06-15) * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115982558A (en) * | 2023-03-17 | 2023-04-18 | 同心智医科技(北京)有限公司 | Electroencephalogram movement intention classification model establishing method and application thereof |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Vigário et al. | Independent component approach to the analysis of EEG and MEG recordings | |
| Xu et al. | High accuracy classification of EEG signal | |
| CN114533086B (en) | Motor imagery brain electrolysis code method based on airspace characteristic time-frequency transformation | |
| CN112515685B (en) | Multi-channel electroencephalogram signal channel selection method based on time-frequency co-fusion | |
| CN103425249A (en) | Electroencephalogram signal classifying and recognizing method based on regularized CSP and regularized SRC and electroencephalogram signal remote control system | |
| CN109558863B (en) | Electroencephalogram signal feature representation and extraction method and device and storage medium | |
| CN102306303B (en) | Electroencephalography signal characteristic extraction method based on small training samples | |
| CN111184509A (en) | Emotion-induced electroencephalogram signal classification method based on transfer entropy | |
| CN109657646B (en) | Method and device for representing and extracting features of physiological time series and storage medium | |
| Kang et al. | A method of denoising multi-channel EEG signals fast based on PCA and DEBSS algorithm | |
| CN112890827B (en) | Electroencephalogram identification method and system based on graph convolution and gate control circulation unit | |
| Shenoy et al. | An iterative optimization technique for robust channel selection in motor imagery based brain computer interface | |
| CN114218986A (en) | State classification method based on EEG electroencephalogram data | |
| CN115105093A (en) | EEG signal classification and identification method based on power spectral density predetermined frequency band | |
| Jiang et al. | Analytical comparison of two emotion classification models based on convolutional neural networks | |
| Radman et al. | Generalized sequential forward selection method for channel selection in EEG signals for classification of left or right hand movement in BCI | |
| Geng et al. | [Retracted] A Fusion Algorithm for EEG Signal Processing Based on Motor Imagery Brain‐Computer Interface | |
| Alizadeh et al. | Multi-task EEG signal classification using correlation-based IMF selection and multi-class CSP | |
| Tang et al. | Research on extraction and classification of EEG features for multi-class motor imagery | |
| CN116369950B (en) | Target detection method based on electroencephalogram tracing and multi-feature extraction | |
| Al Shiam et al. | Electroencephalography based motor imagery classification using unsupervised feature selection | |
| CN116662782A (en) | MSFF-SENET-based motor imagery electroencephalogram decoding method | |
| CN113662561B (en) | Electroencephalogram feature extraction method and device of subband cascade co-space mode | |
| Mu et al. | A channel selection method for motor imagery EEG based on Fisher score of OVR-CSP | |
| CN113780162A (en) | Motor imagery electroencephalogram decoding method combined with multi-period and multi-frequency band CSP algorithm |
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 |