CN109009098A - A kind of EEG signals characteristic recognition method under Mental imagery state - Google Patents
A kind of EEG signals characteristic recognition method under Mental imagery state Download PDFInfo
- Publication number
- CN109009098A CN109009098A CN201810792231.9A CN201810792231A CN109009098A CN 109009098 A CN109009098 A CN 109009098A CN 201810792231 A CN201810792231 A CN 201810792231A CN 109009098 A CN109009098 A CN 109009098A
- Authority
- CN
- China
- Prior art keywords
- eeg signals
- electrode
- energy
- mental imagery
- classification
- 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 56
- 230000003340 mental effect Effects 0.000 title claims abstract description 25
- 238000001228 spectrum Methods 0.000 claims abstract description 15
- 238000000605 extraction Methods 0.000 claims abstract description 8
- 238000012706 support-vector machine Methods 0.000 claims abstract description 7
- 230000008569 process Effects 0.000 claims abstract description 5
- 238000004458 analytical method Methods 0.000 claims description 16
- 239000000284 extract Substances 0.000 claims description 8
- 238000004070 electrodeposition Methods 0.000 claims description 7
- 238000004364 calculation method Methods 0.000 claims description 6
- 230000008859 change Effects 0.000 claims description 6
- 230000005611 electricity Effects 0.000 claims description 6
- 238000013461 design Methods 0.000 claims description 4
- 238000010606 normalization Methods 0.000 claims description 3
- 230000001537 neural effect Effects 0.000 abstract description 3
- 230000006870 function Effects 0.000 description 19
- 238000012360 testing method Methods 0.000 description 7
- 210000004556 brain Anatomy 0.000 description 6
- 238000005516 engineering process Methods 0.000 description 6
- 230000008093 supporting effect Effects 0.000 description 4
- 238000012545 processing Methods 0.000 description 3
- 238000012549 training Methods 0.000 description 3
- 206010033799 Paralysis Diseases 0.000 description 2
- 230000008901 benefit Effects 0.000 description 2
- 230000003925 brain function Effects 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000002474 experimental method Methods 0.000 description 2
- 230000001771 impaired effect Effects 0.000 description 2
- 230000001976 improved effect Effects 0.000 description 2
- 230000036651 mood Effects 0.000 description 2
- 210000002569 neuron Anatomy 0.000 description 2
- 230000002269 spontaneous effect Effects 0.000 description 2
- 230000009466 transformation Effects 0.000 description 2
- 208000012661 Dyskinesia Diseases 0.000 description 1
- 230000002159 abnormal effect Effects 0.000 description 1
- 206010002026 amyotrophic lateral sclerosis Diseases 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000010835 comparative analysis Methods 0.000 description 1
- 238000002790 cross-validation Methods 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 201000010099 disease Diseases 0.000 description 1
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 1
- 230000009977 dual effect Effects 0.000 description 1
- 235000013399 edible fruits Nutrition 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 230000000763 evoking effect Effects 0.000 description 1
- 230000001939 inductive effect Effects 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 201000006417 multiple sclerosis Diseases 0.000 description 1
- 210000005036 nerve Anatomy 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 230000002093 peripheral effect Effects 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000001953 sensory effect Effects 0.000 description 1
- 230000004936 stimulating effect Effects 0.000 description 1
- 230000000638 stimulation Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000002123 temporal effect Effects 0.000 description 1
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
- A61B5/72—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
- A61B5/7235—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
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- General Health & Medical Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Signal Processing (AREA)
- Veterinary Medicine (AREA)
- Public Health (AREA)
- Animal Behavior & Ethology (AREA)
- Surgery (AREA)
- Heart & Thoracic Surgery (AREA)
- Molecular Biology (AREA)
- Psychiatry (AREA)
- Medical Informatics (AREA)
- Biophysics (AREA)
- Pathology (AREA)
- Biomedical Technology (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Physiology (AREA)
- Psychology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Fuzzy Systems (AREA)
- Mathematical Physics (AREA)
- Measurement And Recording Of Electrical Phenomena And Electrical Characteristics Of The Living Body (AREA)
Abstract
The invention discloses the EEG signals characteristic recognition methods under a kind of Mental imagery state, comprising the following steps: S1: EEG signals data information of the acquisition subject in the case where imagining motion conditions；S2: the energy spectrum of EEG signals electrode is calculated using Welch method；S3: being arranged the customized information of the optimal electrode of EEG signals, chooses the highest electrode of classification accuracy rate for different subjects；S4: the characteristic value of EEG signals or so imagination is extracted using the synchronization/method that desynchronizes；S5: tagsort is carried out using EEG signals of the optimal classification function to extraction, assorting process is optimized using the three-stage classification method based on support vector machines.This method is by calculating energy spectrum and the characteristic value information of the electrode of EEG signals to which to EEG signals progress tagsort, wherein Mental imagery EEG feature extraction and sorting technique can be used for neural rehabilitation field.
Description
Technical field
The present invention relates to the EEG signals feature knowledges under signal processing technology field more particularly to a kind of Mental imagery state
Other method.
Background technique
Mental imagery EEG feature extraction and sorting technique can be used for neural rehabilitation field, for those paralysis or sternly
The people of dyskinesia again, especially brain function be complete but people (such as multiple sclerosis, amyotrophic lateral sclerosis funiculus lateralis medullae spinalis that can not move
The patient of the diseases such as hardening), it provides and a kind of is regained around impaired neuron to limbs with the extraneous new way exchanged
Or the control of artificial limb.The technology can be used for the fields such as traffic, military affairs, amusement and recreation, and application prospect is extensive.Currently, brain is electric
The main feature extracting method of signal has a time domain approach, multidimensional statistics analysis method, frequency-domain analysis method, and Time Domain Analysis is
Early stage carries out the main means of eeg analysis, directly extracts wave character from time domain, passes through the geometry character to brain wave waveform
The analysis of matter obtains some important brain electricity temporal signatures, such as wave amplitude, mean value, variance, the degree of bias；Multidimensional statistics analysis method
It is the independent element for extracting source signal from the mixed signal observed, isolates electrocardio, eye electricity and power frequency in EEG signals
The noise signals such as interference, the topographic map and power spectrum for extracting independent element are as feature.But utilize statistical characteristic analysis non-flat
Steady signal, it tends to be difficult to obtain the information of most worthy；The EEG signals that amplitude is changed over time be transformed to electroencephalogram power with
The spectrogram of frequency variation, discloses the rule of signal in terms of frequency domain.It can reflect the relatively strong and weak of frequency content, show Energy distribution
Feature.The good resolution of such method frequency, but the resolution ratio in time domain is bad, the analysis suitable for stationary signal.It handles non-
When stationary signal, reach preferable classifying quality if necessary, need for EEG signals to be divided into data segment short enough, to meet
The requirement of stationarity.
In addition the research of Mental imagery EEG signals is mostly in laboratory stage, classification accuracy rate at present in the prior art
It is one of important evaluation index.It is primarily present following problems at present:
1. the mode for inducing signal needs to improve
Evoked brain potential signal needs additional stimulating apparatus, and to rely on some feeling access of user, therefore be applicable in
Range is subject to certain restrictions.In addition, being not suitable for long-term continuous use, subject can otherwise generated dependent event and adapts to energy
Power or physical fatigue are to occur the change of related potential, there are also to be solved for this contradiction.The EEG signals of self start type system are complete
Subject's spontaneous brain electricity is come from entirely, external offer stimulation is not provided, therefore does not need the sensory nerve effect of subject, but due to
Its is non-stationary, vulnerable to environment and mood etc. influence, characteristic present is not significant, thus this system to signal processing method require compared with
Height, current recognition correct rate are lower.
2. the acquisition mode of signal needs to improve
EEG signals belong to complicated non-stationary signal, and very faint, and are highly prone to extraneous interference, therefore such as
The problems such as what more reasonable experimental program of design, raising signal-to-noise ratio, has to be solved.The acquisition of EEG signals mainly passes through plant at present
Enter formula electrode and two kinds of external electrode, external electrode can be used pastes the forms such as electrode slice or wearing electrode cap, the electricity used one by one
Number of poles is usually very much, and the acquisition of eeg data cannot be completed alone by subject.In addition, the optimum electrode position of every subject
Difference is set, personalization and the precision of electrode position is cannot achieve at present, influences signal testing precision.
3. EEG signals mode identification method needs to improve
Non-stationary signal generally comprises low frequency and radio-frequency component signal, and the feature for classification is usually contained in part
When-frequency information in, different moments include different frequency informations, wherein implicit signal component usually influences each other.Current
Method is primarily present the information for being difficult to analyze non-stationary signal, information is easily lost, cannot be considered in terms of the resolutions of the time domain and the frequency domain, Wu Fajian
The problems such as caring for analysis speed and precision.
The speed and precision of 4 analyses needs to improve
It is unable to Accurate Analysis human thinking activity, and task type is more, analysis accuracy rate is lower.Such as two class thinking
Task, accuracy rate is up to 80% or more, and three classes thinking task, only can reach 70% or so.This with subject, signal acquisition,
Each link that signal analysis and processing process is related to has relationship, for example whether subject is familiar with data acquisition request and acquisition stream
Journey, body and mood etc. are with the presence or absence of abnormal, electrode position if appropriate for the subject, Signal Pretreatment, feature extraction
And whether tagsort algorithm is appropriately effective etc..
5 system rejection to disturbance abilities need to improve
When leaving laboratory environment, by the interference of external environment and self-condition, the spontaneous brain electricity of subject's generation
Signal also can constantly change, and be easy to be influenced by artefact, and system not only needs accurately to judge task execution and free time
State, and signal characteristic is rapidly and accurately extracted, current brain-computer interface technology cannot be met the requirements, and subject is also not
Peripheral equipment can flexibly and comfortably be controlled by brain-computer interface under true environment.
Problem above causes BCI technology to be still within laboratory stage at present, needs to signal processing algorithm, signal acquisition
The multinomial technology of the horizontal, selection of subject and training etc. across multiple fields is furtherd investigate.
Summary of the invention
According to problem of the existing technology, the invention discloses the EEG signals feature knowledges under a kind of Mental imagery state
Other method, comprising the following steps:
S1: EEG signals data information of the acquisition subject in the case where imagining motion conditions；
S2: the energy spectrum of EEG signals electrode is calculated using Welch method；
S3: being arranged the customized information of the optimal electrode of EEG signals, chooses classification accuracy rate highest for different subjects
Electrode；
S4: the characteristic value of EEG signals or so imagination is extracted using the synchronization/method that desynchronizes；
S5: tagsort is carried out using EEG signals of the optimal classification function to extraction, using based on support vector machines
Three-stage classification method optimizes assorting process.
Further, the energy spectrum specific algorithm for calculating EEG signals electrode using Welch method in S2 is as follows:
If clock signal F (n) overall length of EEG signals is N, clock signal is divided into L section, every segment length is M, every section it
Between some data be overlapped, equipped with 1/3 data be overlapped, then:
Window function ω (n) is added on each data segment, the average power spectra of every segment data
Wherein: U is normalization factor, calculation formula are as follows:
Further, the customized information of optimal electrode is set in the following way:
S31: using the energy of calculated EEG signals electrode as feature vector；
S32: the Euclidean distance between each feature vector is calculated；
S33: analysis Euclidean distance numerical value selects the corresponding electrode energy of minimum Eustachian distance as characteristic of division vector.
Further, S4 is specifically used such as using the characteristic value that the synchronization/method that desynchronizes extracts EEG signals or so imagination
Under type:
According to the average power spectra of EEG signalsSignal extracts left and right by the algorithm ERD/ERS that synchronize/desynchronizes
Imagine the energy difference of EEG signals, wherein the energy difference of different periods constitutes Mental imagery feature vector；
Algorithm ERD/ERS wherein synchronize/desynchronize in the following way:
Certain length period before event is occurred as reference time section, with the corresponding band in reference time section
Energy is reference, calculates band energy percentage change caused by Mental imagery, calculation formula are as follows:
Wherein A indicates the energy of specific band when imagination left hand and right hand movement, the energy of wave band corresponding with A during R is reference
Amount indicates that the energy of respective specific wave band in time period increases, i.e. event-related design when ERD is positive number；If be negative
Indicate that the energy of respective specific wave band in time period is reduced when number, i.e. Event-related desynchronization.
Further, tagsort is carried out to EEG signals based on optimal classification function in S5, is chosen most according in S3
Excellent electrode determines each highest electrode position of EEG signals sample classification accuracy, extracts EEG signals based on the position, will
Feature vector obtained in S2 and S4 is classified as the input vector of classification function.
By adopting the above-described technical solution, the EEG signals feature under a kind of Mental imagery state provided by the invention is known
Other method, this method is by calculating energy spectrum and the characteristic value information of the electrode of EEG signals to carry out feature to EEG signals
Classification, wherein Mental imagery EEG feature extraction and sorting technique can be used for neural rehabilitation field, for those paralysis or
The people of severe motor disabilities, especially brain function be complete but the people that can not move, therefore this method is provided and a kind of exchanged with extraneous
New way regain the control to limbs or artificial limb around impaired neuron.
Detailed description of the invention
In order to illustrate the technical solutions in the embodiments of the present application or in the prior art more clearly, to embodiment or will show below
There is attached drawing needed in technical description to be briefly described, it should be apparent that, the accompanying drawings in the following description is only this
The some embodiments recorded in application, for those of ordinary skill in the art, without creative efforts,
It is also possible to obtain other drawings based on these drawings.
Fig. 1 is the flow chart of the method for the present invention.
Specific embodiment
To keep technical solution of the present invention and advantage clearer, with reference to the attached drawing in the embodiment of the present invention, to this
Technical solution in inventive embodiments carries out clear and complete description:
EEG signals characteristic recognition method under a kind of Mental imagery state as shown in Figure 1, specifically includes the following steps:
S1: EEG signals data information of the acquisition subject in the case where imagining motion conditions.
S2: the energy spectrum of EEG signals electrode is calculated using Welch method.
If clock signal F (n) overall length of EEG signals is N, clock signal is divided into L section, every segment length is M, every section it
Between some data be overlapped, equipped with 1/3 data be overlapped, then:
Window function ω (n) is added on each data segment, the average power spectra of every segment data
Wherein: U is normalization factor, calculation formula are as follows:
S3: being arranged the customized information of the optimal electrode of EEG signals, chooses classification accuracy rate highest for different subjects
Electrode；
S31: using the energy of calculated EEG signals electrode as feature vector；
S32: the Euclidean distance between each feature vector is calculated；
Euclidean distance is defined as follows:
If two n-dimensional vector xi=(xi1,xi2,,xin)TAnd xj=(xj1,xj2,,xjn)TTwo objects are respectively indicated, they
Euclidean distance are as follows:
S33: analysis Euclidean distance numerical value selects the corresponding electrode energy of minimum Eustachian distance as characteristic of division vector.
S4: the characteristic value of EEG signals or so imagination is extracted using the synchronization/method that desynchronizes.
According to the average power spectra of EEG signalsSignal extracts left and right by the algorithm ERD/ERS that synchronize/desynchronizes
Imagine the energy difference of EEG signals, wherein the energy difference of different periods constitutes Mental imagery feature vector；
Algorithm ERD/ERS wherein synchronize/desynchronize in the following way:
Certain length period before event is occurred as reference time section, with the corresponding band in reference time section
Energy is reference, calculates band energy percentage change caused by Mental imagery, calculation formula are as follows:
Wherein A indicates the energy of specific band when imagination left hand and right hand movement, the energy of wave band corresponding with A during R is reference
Amount indicates that the energy of respective specific wave band in time period increases, i.e. event-related design when ERD is positive number；If be negative
Indicate that the energy of respective specific wave band in time period is reduced when number, i.e. Event-related desynchronization.
S5: tagsort is carried out using EEG signals of the optimal classification function to extraction, using based on support vector machines
Three-stage classification method optimizes assorting process.Tagsort is carried out to EEG signals based on optimal classification function, according to
The optimal electrode chosen in S3 is determined each highest electrode position of EEG signals sample classification accuracy, is mentioned based on the position
EEG signals are taken, the feature vector that obtains in S2 and S4 is classified as the input vector of classification function.
Mental imagery eeg signal classification belongs to small sample, Nonlinear Classification, separates left and right two classes imagination knot by operation
Fruit.Based on Mental imagery classification characteristics, this method selects support vector machine method to carry out tagsort.Support vector machine method is
A kind of typical two classes classification method, is established on the basis of Statistical Learning Theory and structural risk minimization, small in solution
Many distinctive advantages are shown in sample, non-linear and high dimensional pattern identification problem.
Further, the optimal function for calculating EEG signals solves its optimal classification surface in the following way:
Given training sample set { (xi,yi) | i=1,2 ... n }, wherein xiFor observation, yiFor corresponding classification number, take
Value is 0 or 1.The expression formula of linear discriminant function is g (x)=ω x+b, and wherein x is sample, and ω and b are calculating parameters.In d
In dimension space, sample x=(x1,x2,...xd) the distance between plane expression formula is d=with classifying | ωTX+b |/| | ω | |,
WhereinOptimal classification surface equation is ω x+b=0, is normalized, i.e. g (x)=1, can be by institute
The classifying face for having sample correctly to classify must satisfy:
yi[(ω·xi)+b] -1 >=0, i=1,2 ... n (3.8)
At this point, class interval is 2/ | | ω | |, the sample for setting up the formula is supporting vector.
Optimal classification surface needs set up (3.8), while are minimized (3.9).
Therefore, a Lagrange function is defined, converts problem to the minimum for seeking w and b Lagrange function:
In formula, αiIt is Lagrange coefficient corresponding with each sample.
One dual problem is converted to using Lagrange multiplier, establishes objective function (3.11):
Assuming that α * is optimal solution, then:
In addition, can be acquired by following formula by the thresholding b*, b* that any one supporting vector acquires classification:
In formula: (xi,xj) it is any one supporting vector.
By calculating above, the expression formula of optimal classification function is obtained are as follows:
Non-linear behavior is presented in EEG signals, therefore is another by non-linear conversion by non-linear to linear transformation
Linear problem in a higher dimensional space, seeks optimal classification surface in transformation space.
At this point, the expression formula of optimal classification function are as follows:
Wherein, n is the number of supporting vector, K (xi, x) and it is Radial basis kernel function, expression formula is K (xiX)=exp (- γ
||xi-x||2)。
Further, seek the optimized parameter of optimal classification surface
The major parameter of SVM model is penalty factor and kernel functional parameter γ, next, being led to using cross validation method
It crosses trellis search method and parameter optimization is carried out to test data, search range is 2-5-25, step-length 0.5.By 270 groups of experiment numbers
According to two parts are divided into, a part is used as training set, and another part is as test set.Wherein 180 groups of data are used as instruction for random selection
Practice collection, 90 groups of data of another part are as test set.It after being completed, then repartitions, when all samples are all completed
Afterwards, it selects so that optimized parameter of highest that group of parameter of test group classification accuracy rate as SVM model, if obtaining highest just
Really the parameter of rate has multiple, then selecting that the smallest group of C is optimized parameter.If γ corresponding with C have it is multiple, selection search
Rope to first group of C and γ combination parameter as optimal parameter.For the parameter that this method uses for C=8, these are joined in γ=0.5
Number is applied in optimal classification function.
Classified based on optimal classification function according to the optimal electrode chosen in S3, determines each sample classification accuracy
Highest electrode position extracts EEG signals based on the position, obtains feature vector as classification function in S2 and S4
Input vector is classified.
Two signal testing moment of front are chosen as reference time point.In an experiment, it is believed that only held in certain imagination
When continuous a period of time (at least 1 second or more), the imagination content is effective.In order to avoid being interfered caused by user's mistake imagination, every time
The characteristic vector for participating in operation includes three segment signal information.
The Mental imagery signal at three time points is subjected to operation using the method in S5, comparative analysis classification results, with
The most result of number of repetition is exported as final classification result.
Classify by optimizing electrode signal, and based on improved support vector machine method, accuracy rate output more direct than tradition
As a result method is obviously improved, and in 270 test samples, average correct classification rate highest improves 24%.
The foregoing is only a preferred embodiment of the present invention, but scope of protection of the present invention is not limited thereto,
Anyone skilled in the art in the technical scope disclosed by the present invention, according to the technique and scheme of the present invention and its
Inventive concept is subject to equivalent substitution or change, should be covered by the protection scope of the present invention.
Claims (5)
1. the EEG signals characteristic recognition method under a kind of Mental imagery state, it is characterised in that the following steps are included:
S1: EEG signals data information of the acquisition subject in the case where imagining motion conditions；
S2: the energy spectrum of EEG signals electrode is calculated using Welch method；
S3: being arranged the customized information of the optimal electrode of EEG signals, chooses the highest electricity of classification accuracy rate for different subjects
Pole；
S4: the characteristic value of EEG signals or so imagination is extracted using the synchronization/method that desynchronizes；
S5: tagsort is carried out using EEG signals of the optimal classification function to extraction, using three sections based on support vector machines
Formula classification method optimizes assorting process.
2. the EEG signals characteristic recognition method under a kind of Mental imagery state according to claim 1, feature also exist
In: the energy spectrum specific algorithm for calculating EEG signals electrode using Welch method in S2 is as follows:
If clock signal F (n) overall length of EEG signals is N, clock signal is divided into L sections, every segment length is M, is had between every section
A part of data are to be overlapped, and the data equipped with 1/3 are overlapped, then:
Window function ω (n) is added on each data segment, the average power spectra of every segment data
Wherein: U is normalization factor, calculation formula are as follows:
3. the EEG signals characteristic recognition method under a kind of Mental imagery state according to claim 1, feature also exist
In；The customized information of optimal electrode is set in the following way:
S31: using the energy of calculated EEG signals electrode as feature vector；
S32: the Euclidean distance between each feature vector is calculated；
S33: analysis Euclidean distance numerical value selects the corresponding electrode energy of minimum Eustachian distance as characteristic of division vector.
4. the EEG signals characteristic recognition method under a kind of Mental imagery state according to claim 1, feature also exist
In: the characteristic value of EEG signals or so imagination is extracted specifically in the following way using the synchronization/method that desynchronizes in S4:
According to the average power spectra of EEG signalsSignal extracts the left and right imagination by the algorithm ERD/ERS that synchronize/desynchronizes
The energy difference of EEG signals, wherein the energy difference of different periods constitutes Mental imagery feature vector；
Algorithm ERD/ERS wherein synchronize/desynchronize in the following way:
Certain length period before event is occurred as reference time section, with the energy of the corresponding band in reference time section
For reference, band energy percentage change caused by Mental imagery, calculation formula are calculated are as follows:
Wherein A indicates that the energy of specific band when imagination left hand and right hand movement, R are the energy of wave band corresponding with A during referring to, when
When ERD is positive number, indicate that the energy of respective specific wave band in time period increases, i.e. event-related design；When if it is negative
Indicate that the energy of respective specific wave band in time period is reduced, i.e. Event-related desynchronization.
5. the EEG signals characteristic recognition method under a kind of Mental imagery state according to claim 1, feature also exist
In: tagsort is carried out to EEG signals based on optimal classification function in S5, according to the optimal electrode chosen in S3, is determined each
The highest electrode position of EEG signals sample classification accuracy extracts EEG signals based on the position, will be obtained in S2 and S4
Feature vector is classified as the input vector of classification function.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810792231.9A CN109009098B (en) | 2018-07-18 | 2018-07-18 | Electroencephalogram signal feature identification method under motor imagery state |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810792231.9A CN109009098B (en) | 2018-07-18 | 2018-07-18 | Electroencephalogram signal feature identification method under motor imagery state |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109009098A true CN109009098A (en) | 2018-12-18 |
| CN109009098B CN109009098B (en) | 2021-05-28 |
Family
ID=64642269
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810792231.9A Expired - Fee Related CN109009098B (en) | 2018-07-18 | 2018-07-18 | Electroencephalogram signal feature identification method under motor imagery state |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109009098B (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110674725A (en) * | 2019-09-20 | 2020-01-10 | 电子科技大学 | Equipment signal type identification method based on multi-dimensional feature vector combination of detection signals |
| CN111543988A (en) * | 2020-05-25 | 2020-08-18 | 五邑大学 | Adaptive cognitive activity recognition method and device and storage medium |
| CN111950455A (en) * | 2020-08-12 | 2020-11-17 | 重庆邮电大学 | Motion imagery electroencephalogram characteristic identification method based on LFFCNN-GRU algorithm model |
| CN113855023A (en) * | 2021-10-26 | 2021-12-31 | 深圳大学 | Lower limb movement BCI electrode selection method and system based on iteration tracing |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2011123059A1 (en) * | 2010-03-31 | 2011-10-06 | Agency For Science, Technology And Research | Brain- computer interface system and method |
| CN102722728A (en) * | 2012-06-11 | 2012-10-10 | 杭州电子科技大学 | Motion image electroencephalogram classification method based on channel weighting supporting vector |
-
2018
- 2018-07-18 CN CN201810792231.9A patent/CN109009098B/en not_active Expired - Fee Related
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2011123059A1 (en) * | 2010-03-31 | 2011-10-06 | Agency For Science, Technology And Research | Brain- computer interface system and method |
| CN102722728A (en) * | 2012-06-11 | 2012-10-10 | 杭州电子科技大学 | Motion image electroencephalogram classification method based on channel weighting supporting vector |
Non-Patent Citations (5)
| Title |
|---|
| 冯焕清等: "用于BCI设计的运动想象脑电信号分析方法", 《2007中国生物医学工程联合学术年会》 * |
| 夏阳等: "《神经信息学基础》", 30 September 2015, 电子科技大学出版社 * |
| 宋承天等: "《随机信号分析与估计》", 30 April 2018, 北京理工大学出版社 * |
| 康志亮等: "融合小波和神经网络的发动机转子系统故障诊断", 《农机化研究》 * |
| 郭硕达: "基于BCI技术的手部运动功能康复研究", 《中国优秀硕士学位论文全文数据 医药卫生科技辑》 * |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110674725A (en) * | 2019-09-20 | 2020-01-10 | 电子科技大学 | Equipment signal type identification method based on multi-dimensional feature vector combination of detection signals |
| CN110674725B (en) * | 2019-09-20 | 2022-06-03 | 电子科技大学 | Equipment signal type identification method based on multi-dimensional feature vector combination of detection signals |
| CN111543988A (en) * | 2020-05-25 | 2020-08-18 | 五邑大学 | Adaptive cognitive activity recognition method and device and storage medium |
| CN111950455A (en) * | 2020-08-12 | 2020-11-17 | 重庆邮电大学 | Motion imagery electroencephalogram characteristic identification method based on LFFCNN-GRU algorithm model |
| CN111950455B (en) * | 2020-08-12 | 2022-03-22 | 重庆邮电大学 | Motion imagery electroencephalogram characteristic identification method based on LFFCNN-GRU algorithm model |
| CN113855023A (en) * | 2021-10-26 | 2021-12-31 | 深圳大学 | Lower limb movement BCI electrode selection method and system based on iteration tracing |
| CN113855023B (en) * | 2021-10-26 | 2023-07-04 | 深圳大学 | Iterative tracing-based lower limb movement BCI electrode selection method and system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109009098B (en) | 2021-05-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106886792B (en) | Electroencephalogram emotion recognition method for constructing multi-classifier fusion model based on layering mechanism | |
| CN110070105B (en) | Electroencephalogram emotion recognition method and system based on meta-learning example rapid screening | |
| Coyle et al. | A time-series prediction approach for feature extraction in a brain-computer interface | |
| Cai et al. | Study on feature selection methods for depression detection using three-electrode EEG data | |
| Ma et al. | Deep channel-correlation network for motor imagery decoding from the same limb | |
| Hussein et al. | Robust detection of epileptic seizures based on L1-penalized robust regression of EEG signals | |
| CN106108894A (en) | A kind of emotion electroencephalogramrecognition recognition method improving Emotion identification model time robustness | |
| Miao et al. | A spatial-frequency-temporal optimized feature sparse representation-based classification method for motor imagery EEG pattern recognition | |
| CN114533086B (en) | Motor imagery brain electrolysis code method based on airspace characteristic time-frequency transformation | |
| CN109009098A (en) | A kind of EEG signals characteristic recognition method under Mental imagery state | |
| CN108577865A (en) | A kind of psychological condition determines method and device | |
| Du et al. | A method for detecting high-frequency oscillations using semi-supervised k-means and mean shift clustering | |
| CN111956219A (en) | Electroencephalogram signal-based emotion feature identification method and system and emotion feature identification and adjustment system | |
| CN108280414A (en) | A kind of recognition methods of the Mental imagery EEG signals based on energy feature | |
| Geng et al. | Epileptic seizure detection based on improved wavelet neural networks in long-term intracranial EEG | |
| Haider et al. | Performance enhancement in P300 ERP single trial by machine learning adaptive denoising mechanism | |
| Ramos-Aguilar et al. | Analysis of EEG signal processing techniques based on spectrograms | |
| CN108875799A (en) | A kind of Mental imagery classifying identification method based on improvement S-transformation | |
| Ma et al. | A classification algorithm of an SSVEP brain-Computer interface based on CCA fusion wavelet coefficients | |
| Mahmoudi et al. | Multi-class EEG classification of motor imagery signal by finding optimal time segments and features using SNR-based mutual information | |
| Tian et al. | Study on classification of left-right hands motor imagery EEG signals based on CNN | |
| Kim et al. | eRAD-Fe: Emotion recognition-assisted deep learning framework | |
| CN110338760A (en) | A kind of three classification method of schizophrenia based on electroencephalogram frequency domain data | |
| Placidi et al. | Classification strategies for a single-trial binary Brain Computer Interface based on remembering unpleasant odors | |
| Wang et al. | A multiple autocorrelation analysis method for motor imagery EEG feature extraction |
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 | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20210528 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |