CN101833669A - Method for extracting characteristics of event related potential generated by using audio-visual combined stimulation - Google Patents
Method for extracting characteristics of event related potential generated by using audio-visual combined stimulation Download PDFInfo
- Publication number
- CN101833669A CN101833669A CN201010170300A CN201010170300A CN101833669A CN 101833669 A CN101833669 A CN 101833669A CN 201010170300 A CN201010170300 A CN 201010170300A CN 201010170300 A CN201010170300 A CN 201010170300A CN 101833669 A CN101833669 A CN 101833669A
- Authority
- CN
- China
- Prior art keywords
- time
- signal
- audio
- event related
- related potential
- 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
Images
Landscapes
- Measurement And Recording Of Electrical Phenomena And Electrical Characteristics Of The Living Body (AREA)
Abstract
The invention relates to the field of brain-computer interfaces. For providing a method for extracting characteristics of an event related potential generated by using audio-visual combined stimulation, which aims to achieve super-sensitivity and high space-time resolution measurement performance to further realize a high-quality multi-component multi-parameter imaging function, the invention adopts a technical scheme comprising the following steps after experimental data is acquired: taking time-domain information about a P300 signal at a CZ lead position as known priori knowledge to determine time when the P300 signal appears, and taking the time as time-domain constriction to construct a reference signal; and performing time-domain information-constrained independent component analysis on acquired brain wave data and extracting an information source most capable of representing signal characteristics. The method is mainly applied to the characteristic extraction of the event related potential generated by using the audio-visual combined stimulation.
Description
Technical field
The present invention relates to field of brain-computer interfaces, especially audio-visual combined stimulation produces the feature extracting method of event related potential.
Background technology
Brain-computer interface (Brain-Computer Interface, BCI) technology is by computer monitoring, identification human thinking idea signal mode, produce the instruction of may command and peripheral communication of manipulation or working equipment, to reach anticipation operation purpose or realization and external information communication function.BCI can help physical disabilities but the normal patient of mental awareness, they are repaired and extraneous information interchange ability to a certain extent, thereby improve their quality of life, this is one of initial purpose of BCI research, also is a direct application approach of being approved most at present.
With sensation, cognitive closely-related event related potential (ERP) be brain-computer interface (BCI) system sense of independence information transform through one of canonical form.Classical ERP principal ingredient comprises P1, N1, P2, N2, P3, and wherein first three kind is called exogenous components, and then two kinds are called endogenous component (as Fig. 1).The principal feature of these several compositions is: at first be not only the embodiment of the simple physiological activity of brain, and reflected some aspect of psychological activity; Secondly, drawing of they must have special stimulation arrangement, and is the variation of plural stimulation or stimulation.Thereby in fact for the research of ERP is exactly analysis to these principal ingredients.
Research for event related potential mainly is single pass stimulation at first, as visual stimulus, and acoustic stimuli, body sense stimulation etc.Along with the development of signal processing method and hardware technology, begin to stimulate conversion for the research of ERP to binary channels, as vision-acoustic stimuli, the sense of hearing-body sense stimulation etc.Audio-visual combined stimulation is one of focus of neuroelectricity physiological field in recent years, it is short and contain the characteristics of high-dimensional space distribution information to studies show that it has a wave amplitude height, latent period, can remedy and singlely feel to bring out that ERP information is very few, the defective of unfavorable identification, thereby improve information conversion rate and classification accuracy.
In theory, be not isolated fully the separation between the different sensation of the nervous system path, but have the phenomenon of integrating mutually.Many sensations are integrated has vital role for human perception, memory etc.Yu Liping has proposed the behavior and the psychologic effect of many sensations in its document, and has introduced the cross-module formula plasticity of sensory cortex.
Recently each sensory channel of having thought before having denied of external result of study is obtained the hypothesis of relatively independent information for outside things, and propositions such as Driver: the information of obtaining from a sensory channel can influence brain to the former judgement that belongs to another sensory channel information.Such as input visual information correspondingly in acoustic stimuli, will strengthen understanding to auditory information; Although body sense stimulation itself can not transmitted colouring information, applying the body sense at given position stimulates and also can improve the people for judgement of visual color or the like.These examples show that all the information content of a passage might influence the process of information processing of another sensory channel.
Summary of the invention
For overcoming the deficiencies in the prior art, remedy the defective of prior art on spatial sampling density, make system with ultra-sensitivity and high time-space resolution measurement performance, realize high-quality polycomponent, multiparameter imaging function then.For achieving the above object, the technical solution used in the present invention is: audio-visual combined stimulation produces the feature extracting method of event related potential, after the collection experimental data, carries out the following step:
At first utilize the lead time-domain information of place's P300 signal of CZ to determine the time that the p300 signal occurs, and it is constructed reference signal as time-domain constraints as known priori; Then the eeg data that collects is subjected to the independent component analysis of time-domain information constraint, extracts an information source that best embodies signal characteristic.
Described time-domain constraints condition is expressed as:
D(ω
j)=ξ-ε(ω
j)≤0 (1)
ω j separates a certain row vector that mixes in the matrix W in the formula, and waits the amount of finding the solution y
j=ω
jX; ε (ω
j) characterize estimate and reference signal r between similarity degree; ξ is its threshold value.
It is described that to be subjected to the independent component analysis of time-domain information constraint be to adopt the FICA algorithm, with following formula approximate representation negentropy:
J(y)≈ρ{E[G(y)-G(g)]}
2 (2)
ρ is positive constant in the formula, and g is the gaussian signal of zero-mean, unit variance, and y is an information source to be extracted, and it also is a unit variance, and G () can get any non-quadratic function;
Use y
jAnd square error between r or simple crosscorrelation are as the tolerance of similarity degree, and when adopting the latter to measure, simple crosscorrelation is more similar, has:
D(ω
j)=ξ-E[CCR(r,ω
jx)]≤0 (3)
In the formula: CCR represents simple crosscorrelation, and this moment, ξ was the lower limit of optimum simple crosscorrelation, so algorithm can be expressed as:
Select ω
jMake F (ω
j)=ρ { E[G (ω
jX)]-E[G (g)]
2Greatly (4-1)
Constraint condition
Latter two constraint condition is requirement y
jEqual 1 with the variance of r;
Formula (4-1), (4-1) formula belong to affined optimization problem, adopt the Lagrangian function of augmentation, find the solution in conjunction with learning algorithm adjusting w and Lagrangian parameter like newton.
System of the present invention adopts optical fiber coupling noncontact mode to realize the measurement of time domain DOT/FMT prototype, and its characteristics are:
Utilization is subjected to the independent component analysis of time-domain information constraint, has reduced the dimension of signal on the one hand; On the other hand, as the reference signal, can reduce the interference of other compositions and noise, help the pattern-recognition in later stage according to the time domain characteristics of institute's information extraction.
Description of drawings
Each principal ingredient synoptic diagram of Fig. 1 event related potential.First three composition P1 wherein, N1, P2 belongs to exogenous components, influenced by the environmental stimuli parameter.Three composition N2 then, P3 then belongs to endogenous component, and main sensation with the people is relevant with cognition.
Fig. 2 experiment flow synoptic diagram.Experiment comprises 5 session, and each session has different goal stimulus numerals.Each session comprises 15 block.Comprise 9 different trial among each block.Each trial is respectively the flicker of a numeral and corresponding Chinese speech pronunciation.As glimmered 2 o'clock, with the pronunciation of " er ".
This experiment of Fig. 3 shows the 3*3 matrix of from 1 to 9, nine numeral, each trial one of them numeral of glimmering at random on the screen, other numerals then are grey.The demonstration of a trial in the experiment.The numeral of glimmering among the figure is " 5 ", and the while is with the Chinese speech pronunciation of " wu ".
Fig. 4 electrode for encephalograms is settled scheme.
The simple block diagram explanation of Fig. 5 ICA
Employed reference signal during Fig. 6 signal Processing.Because the p300 signal produces between the 450ms at 250ms after the stimulation.So adopt in this experiment each trial begin back 250ms to the positive pulse of 450ms as the reference signal.
The lead signal of place goal stimulus and non-goal stimulus sequence of Fig. 7 Cz passes through result behind the coherence average respectively.
After Fig. 8 handled through CICA, goal stimulus and non-goal stimulus sequence signal be the resulting signal of coherence average respectively.
Embodiment
The present invention has designed based on the independent component analysis (constrained IndependentComponent Analysis, event related potential p300 new feature extraction method CICA) that are subjected to the time-domain information constraint.The time-domain information that this method is at first utilized CZ to lead and located the P300 signal is determined the time that the p300 signal occurs as known priori.And it is constructed reference signal as time-domain constraints, thus the eeg data that collects is subjected to the independent component analysis of time-domain information constraint, extract an information source that best embodies signal characteristic.
For overcoming under the single channel pattern, the relation of the mutual restriction that forms between these three factors of recognition accuracy of the proportioning number of target stimulation and task stimulation during the appearance speed of target stimulation, task stimulate, and the conversion rate of lifting effective information.The present invention proposes the new method of extracting event related potential P300 feature under the audio-visual combined stimulation based on the independent component analysis that is subjected to time restriction (CICA) method.
The present invention has designed based on the independent component analysis (constrained IndependentComponent Analysis, event related potential p300 new feature extraction method CICA) that are subjected to the time-domain information constraint.The time-domain information that this method is at first utilized CZ to lead and located the P300 signal is determined the time that the p300 signal occurs as known priori.And it is constructed reference signal as time-domain constraints, thus the eeg data that collects is subjected to the independent component analysis of time-domain information constraint, extract an information source that best embodies signal characteristic.
1 experimental program
1.1 experimental data collection
This paper experimental data is taken from 8 dextromanual adult healthy volunteers and (is this school university student, 6 male 4 woman, 24.3 ± 3.1 years old mean age), what the stimulation generation equipment of experiment used is STIM 2 systems, and acquisition system is then used the Scan4.3 digital collection system of Neuroscan.Breadboard electromagnetic screen and soundproof effect are all good, and (indoor ground unrest is about 31dB, and background illumination is 2cd/m
2).In the experimentation, the experimenter to feel comfortable but the posture that does not influence data acquisition sit in an armchair, in the face of the experimental duties prompt on 1 meter left and right sides, dead ahead distance display screen.
Study subject is being watched on the computer screen 3*3 matrix of 9 numerals of from 1 to 9 attentively in the audio-visual combined stimulation.Whole experiment is divided into 5 session.(as Fig. 2) each session is divided into 15 block, all can point out next numeral that group session should be noted that before each session.Each block is and the flicker at random and the sound of 9 numerals of from 1 to 9 constitute, each single flicker experiment, and 9 digital random in the matrix are glimmered one of them, and the experimenter also can hear corresponding sound simultaneously.Flicker or sounding duration are 300ms, at interval flicker next random digit (as Fig. 3) behind the 200ms.
1.2 data recording and pre-service
64 lead the brain wave acquisition electrode position as shown in Figure 4 in the experiment, and the experiment electrode used therein is Ag/AgCI electrode (impedance is less than 5000 ohm); As the reference level, brain electricity sample frequency is 1000Hz with the bilateral mastoid process, and filter pass band is 0.5~100Hz, and adopts the 50Hz trapper to remove power frequency and disturb.
2 are subjected to the independent component analysis of time-domain information constraint
2.1 independent component analysis and essence thereof
Independent component analysis (ICA) is the signal processing method that grows up in the later stage nineties 20th century of signal Processing field.The signal decomposition that it will originally be mixed by several independent signal sources becomes several compositions independently mutually, promptly these information sources is decomposed and comes.Fig. 5 can simple declaration ICA basic processing process.Lead observation X is that a plurality of information source S combine (X=AS) through hybrid matrix A more.And the main task of independent component analysis is: under the condition of S and A the unknown, ask for one and separate mixed matrix B, making X pass through its back gained output Y (Y=BX) is the best approximation of S.Require each component among the Y separate, but allow Y to follow each the component ordering among the S not necessarily identical, engineer's scale is also not necessarily identical.
The cardinal rule that ICA decomposes can be divided into two roughly.
1, non-linear decorrelation is found the solution a mixed battle array B and is made its any amount output y
i, y
j(not only itself is uncorrelated between the i ≠ j), and through the component g (y after the nonlinear transformation
i) and h (y
j) also uncorrelated.Function g, it is suitable that h will select.
2, make and export Gaussization as far as possible.Under the constant condition of variance of certain component y of output, will import each component of X and do linear combination
Each weight b of optimized choice
i, make y Gaussization as far as possible, then each local maximum of the non-Gauss of y provides an isolated component.
2.2 be subjected to the ICA analytical procedure of time-domain information constraint
Generally each component decomposes out simultaneously behind ICA.Will must priori from picking out that the isolated component that will extract still needs.When number of active lanes is many, be difficult to from numerous isolated components, find needed component.The ICA that is subjected to the time-domain information constraint then is that the time domain priori of utilizing generation to extract information source reaches the purpose of extracting required component as constraint condition.The method belongs to extracts isolated component one by one, therefore is adapted to adopt quick ICA (FICA) algorithm.
The FICA algorithm is from " at first extract the most non-Gauss component " (component of negentropy maximum just) this thinking.But the most non-Gauss's component may not be exactly to want the signal that extracts just, therefore needs to introduce constraint condition in searching process, makes and wants that the signal that extracts is at first extracted.FICA is with following formula approximate representation negentropy:
J(y)≈ρ{E[G(y)-G(g)]}
2 (1)
ρ is positive constant in the formula, and g is the gaussian signal of zero-mean, unit variance.Y is an information source to be extracted, and it also is a unit variance.G () can get any non-quadratic function.
In order to make optimizing process converge on the isolated component of desired extraction, need a reference signal r very approaching (t) with this isolated component.R (t) might not equal information source to be extracted, also not necessarily will mate finely with this information source, but it will have the ability to make the direction convergence of optimizing process towards hope.
Constraint condition can usually be expressed as:
D(ω
j)=ξ-ε(ω
j)≤0 (2)
ω j separates a certain row vector that mixes in the matrix W in the formula, and waits the amount of finding the solution y
j=ω
jX; ε (ω
j) characterize estimate and reference signal r between similarity degree; ξ is its threshold value.
Common available y
jAnd square error between r or simple crosscorrelation are as the tolerance of similarity degree.When adopting the latter to measure, simple crosscorrelation is more similar.Therefore have:
D(ω
j)=ξ-E[CCR(r,ω
jx)]≤0 (3)
In the formula: CCR represents simple crosscorrelation, and this moment, ξ was the lower limit of optimum simple crosscorrelation.So algorithm can be expressed as: select ω
jMake F (ω
j)=ρ { E[G (ω
jX)]-E[G (g)]
2Greatly (4-1)
Constraint condition
Latter two constraint condition is requirement y
jEqual 1. with the variance of r
Formula (4-1), (4-2) formula belong to affined optimization problem, adopt the Lagrangian function of augmentation, find the solution in conjunction with learning algorithm adjusting w and Lagrangian parameter like newton.
The present invention adopts a pulse that stimulates the beginning back 250ms-450ms time period as reference signal r (t).As shown below.
Beneficial effect
Because event related potential p300 characteristic mainly concentrates on below the 10HZ, so, carry out the filtering of 1-10Hz earlier for the data that collect, remove other unwanted information and interference of noise.Afterwards the data that should comprise p300 and should not comprise p300 are carried out coherence average respectively.For each test, 5 goal stimulus numerals are arranged.Each goal stimulus numeral has 15 block respectively, has only one to be goal stimulus among 9 trial of each block.So total total 15*5 trial comprises p300 in the whole experiment, 600 trial do not comprise the p300 signal.
Fig. 7 be with Cz lead the resulting data of place's goal stimulus and non-goal stimulus respectively the coherence average gained (get each stimulation begin to after the data of 700ms).Can see that thus obtained giving prominence to for p300 information in the event related potential of goal stimulus generation, still, the difference of goal stimulus and non-goal stimulus is not clearly.In addition, also have other compositions and interference of noise.
Fig. 8 utilizes the independent component analysis that is subjected to the time-domain information constraint to handle the result of back coherence average.Can significantly find out the p300 feature of goal stimulus sequence among the figure.And the difference of goal stimulus and non-goal stimulus is apparent in view, and has removed the influence of other noises to signal.
Utilization is subjected to the independent component analysis of time-domain information constraint, has reduced the dimension of signal on the one hand.Be that data with 64 dimensions have dropped to 1 dimension in this patent, only get in the isolated component and reference signal relation the most non-the closest gaussian component.On the other hand, as the reference signal, can reduce the interference of other compositions and noise according to the time domain characteristics of institute's information extraction.Be beneficial to the pattern-recognition in later stage.
The present invention has designed based on the independent component analysis (constrained IndependentComponent Analysis, event related potential p300 new feature extraction method CICA) that are subjected to the time-domain information constraint.For feature extraction, this method has reduced the dimension of signal on the one hand, has strengthened the p300 feature on the other hand, helps later pattern-recognition.
Claims (3)
1. an audio-visual combined stimulation produces the feature extracting method of event related potential, it is characterized in that, after gathering experimental data, carry out the following step: at first utilize the lead time-domain information of place's P300 signal of CZ to determine the time that the p300 signal occurs, and it is constructed reference signal as time-domain constraints as known priori; Then the eeg data that collects is subjected to the independent component analysis of time-domain information constraint, extracts an information source that best embodies signal characteristic.
2. a kind of audio-visual combined stimulation according to claim 1 produces the feature extracting method of event related potential, it is characterized in that the time-domain constraints condition is expressed as:
D(ω
j?)=ξ-ε(ω
j)≤0 (1)
ω j separates a certain row vector that mixes in the matrix W in the formula, and waits the amount of finding the solution y
j=ω
jX; ε (ω
i) characterize estimate and reference signal r between similarity degree; ξ is its threshold value.
3. a kind of audio-visual combined stimulation according to claim 1 produces the feature extracting method of event related potential, it is characterized in that, the independent component analysis that is subjected to the time-domain information constraint is to adopt the FICA algorithm, with following formula approximate representation negentropy:
J(y)≈ρ{E[G(y)-G(g)]}
2 (2)
ρ is positive constant in the formula, and g is the gaussian signal of zero-mean, unit variance, and y is an information source to be extracted, and it also is a unit variance, and G () can get any non-quadratic function;
Use y
jAnd square error between r or simple crosscorrelation are as the tolerance of similarity degree, and when adopting the latter to measure, simple crosscorrelation is more similar, has:
In the formula: CCR represents simple crosscorrelation, and this moment, ξ was the lower limit of optimum simple crosscorrelation, so algorithm can be expressed as:
Select ω
jMake F (ω
j)=ρ { E[G (ω
jX)]-E[G (g)]
2Greatly (4-1)
Latter two constraint condition is requirement y
jEqual 1 with the variance of r;
Formula (4-1), (4-2) belong to affined optimization problem, adopt the Lagrangian function of augmentation, find the solution in conjunction with learning algorithm adjusting w and Lagrangian parameter like newton.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201010170300A CN101833669A (en) | 2010-05-13 | 2010-05-13 | Method for extracting characteristics of event related potential generated by using audio-visual combined stimulation |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201010170300A CN101833669A (en) | 2010-05-13 | 2010-05-13 | Method for extracting characteristics of event related potential generated by using audio-visual combined stimulation |
Publications (1)
Publication Number | Publication Date |
---|---|
CN101833669A true CN101833669A (en) | 2010-09-15 |
Family
ID=42717732
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201010170300A Pending CN101833669A (en) | 2010-05-13 | 2010-05-13 | Method for extracting characteristics of event related potential generated by using audio-visual combined stimulation |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN101833669A (en) |
Cited By (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102184017A (en) * | 2011-05-13 | 2011-09-14 | 天津大学 | Lead optimizing method for P300 brain-computer interface |
CN102184019A (en) * | 2011-05-16 | 2011-09-14 | 天津大学 | Method for audio-visual combined stimulation of brain-computer interface based on covert attention |
CN102793540A (en) * | 2012-06-14 | 2012-11-28 | 天津大学 | Method for optimizing audio-visual cognitive event-related potential experimental paradigm |
CN104615887A (en) * | 2015-02-05 | 2015-05-13 | 广州市润杰医疗器械有限公司 | Voice-image matching method for detecting cognitive function of child |
CN104899573A (en) * | 2015-06-18 | 2015-09-09 | 福州大学 | P300 feature extraction method based on wavelet transformation and Fisher criterion |
CN111012342A (en) * | 2019-11-01 | 2020-04-17 | 天津大学 | Audio-visual dual-channel competition mechanism brain-computer interface method based on P300 |
CN113595569A (en) * | 2021-07-29 | 2021-11-02 | 中国人民解放军国防科技大学 | ICA-R algorithm-based 2FSK signal anti-interference method and device |
-
2010
- 2010-05-13 CN CN201010170300A patent/CN101833669A/en active Pending
Cited By (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102184017A (en) * | 2011-05-13 | 2011-09-14 | 天津大学 | Lead optimizing method for P300 brain-computer interface |
CN102184017B (en) * | 2011-05-13 | 2012-11-28 | 天津大学 | Lead optimizing method for P300 brain-computer interface |
CN102184019A (en) * | 2011-05-16 | 2011-09-14 | 天津大学 | Method for audio-visual combined stimulation of brain-computer interface based on covert attention |
CN102793540A (en) * | 2012-06-14 | 2012-11-28 | 天津大学 | Method for optimizing audio-visual cognitive event-related potential experimental paradigm |
CN104615887A (en) * | 2015-02-05 | 2015-05-13 | 广州市润杰医疗器械有限公司 | Voice-image matching method for detecting cognitive function of child |
CN104899573A (en) * | 2015-06-18 | 2015-09-09 | 福州大学 | P300 feature extraction method based on wavelet transformation and Fisher criterion |
CN104899573B (en) * | 2015-06-18 | 2018-06-15 | 福州大学 | P300 feature extracting methods based on wavelet transformation Yu Fisher criterion |
CN111012342A (en) * | 2019-11-01 | 2020-04-17 | 天津大学 | Audio-visual dual-channel competition mechanism brain-computer interface method based on P300 |
CN111012342B (en) * | 2019-11-01 | 2022-08-02 | 天津大学 | Audio-visual dual-channel competition mechanism brain-computer interface method based on P300 |
CN113595569A (en) * | 2021-07-29 | 2021-11-02 | 中国人民解放军国防科技大学 | ICA-R algorithm-based 2FSK signal anti-interference method and device |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110765920B (en) | Motor imagery classification method based on convolutional neural network | |
CN106569604B (en) | Audiovisual bimodal semantic matches and semantic mismatch collaboration stimulation brain-machine interface method | |
CN107024987B (en) | Real-time human brain attention testing and training system based on EEG | |
Kachenoura et al. | ICA: a potential tool for BCI systems | |
CN101833669A (en) | Method for extracting characteristics of event related potential generated by using audio-visual combined stimulation | |
CN106407733A (en) | Depression risk screening system and method based on virtual reality scene electroencephalogram signal | |
Pinheiro et al. | Wheelchair simulator game for training people with severe disabilities | |
CN111616681B (en) | Anesthesia state monitoring system based on portable electroencephalogram acquisition equipment and deep learning | |
CN110969108A (en) | Limb action recognition method based on autonomic motor imagery electroencephalogram | |
CN108520239B (en) | Electroencephalogram signal classification method and system | |
CN113031766B (en) | Method for decoding Chinese pronunciation through electroencephalogram | |
CN114533086A (en) | Motor imagery electroencephalogram decoding method based on spatial domain characteristic time-frequency transformation | |
CN106510702B (en) | The extraction of sense of hearing attention characteristics, identifying system and method based on Middle latency auditory evoked potential | |
Mustafa et al. | The analysis of eeg spectrogram image for brainwave balancing application using ann | |
CN101433460B (en) | Spatial filtering method of lower limb imaginary action potential | |
CN113576498B (en) | Visual and auditory aesthetic evaluation method and system based on electroencephalogram signals | |
Nguyen et al. | LIBS: a lightweight and inexpensive in-ear sensing system for automatic whole-night sleep stage monitoring | |
Paulraj et al. | A machine learning approach for distinguishing hearing perception level using auditory evoked potentials | |
CN113116306A (en) | Consciousness disturbance auxiliary diagnosis system based on auditory evoked electroencephalogram signal analysis | |
Fickling et al. | Distant sensor prediction of event-related potentials | |
Bai et al. | Domain-adaptive emotion recognition based on horizontal vertical flow representation of EEG signals | |
Kanaga et al. | A Pilot Investigation on the Performance of Auditory Stimuli based on EEG Signals Classification for BCI Applications | |
TWI620173B (en) | Method of mixing frequency following response evokded sounds with music to modulate neuronal networks associated with multiple intelligences in the brain | |
Kshirsagar et al. | Mental task classification using wavelet transform and support vector machine | |
Hoang | Multivariate features for multi-class brain computer interface systems |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
C12 | Rejection of a patent application after its publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20100915 |