CN102436302B - Method for brain-computer interface based on amplitude modulated visual evoked potential - Google Patents

Abstract

The invention discloses a method for a brain-computer interface based on amplitude modulated visual evoked potential. The method comprises the following steps of: firstly, connecting an electroencephalogram signal acquisition system; optimally selecting the most-sensitive frequency of amplitude frequency response of a tested person as carrier frequency; maximizing a signal-noise ratio of a response signal; secondly, providing a selection range of modulation wave frequency, designing an AMVEP (Amplitude Modulated Visual Evoked Potential) normal form and generating a stimulation sequence of an amplitude modulated visual evoked potential brain-computer interface normal form; and finally, realizing the identification of single signal or a small amount of signals of the normal form AMVEP. The invention provides the normal form based on the amplitude modulated visual evoked potential and a method for analysis and characteristic extraction of a modulation signal corresponding to the normal form AMVEP. The identification of single signal or a small amount of signals of the normal form AMVEP based on the amplitude modulated visual evoked potential is realized; and the method disclosed by the invention has the advantages of simpleness in operation, fewer electrodes and more target numbers.

Description

Based on which amplitude modulation VEP brain-computer interface method

Technical field

The present invention relates to the brain-computer interface technical field, be specifically related to based on which amplitude modulation VEP brain-computer interface method.

Background technology

The EEG signals that is based on brain-computer interface realizes that brain and computing machine or other electronic equipments directly exchange the system of communication and control.Brain-computer interface (Brain-Computer Interface, be called for short BCI) as man machine interface (Human-computer interface, be called for short HCI) in a kind of, the information that can send brain directly is converted to and can drives the external unit order, and the musculatures such as limbs that replace the people realize that the people exchanges and to the control of external environment condition with extraneous, owing to not relying on conventional brain output channel, the brain of behaving has been opened up the approach that brand-new and an external world carry out communication and control, make the idea of utilizing people's brain signal directly to control external unit become possibility.In recent years, brain-computer interface (BCI) technical development is very fast, in fields such as biomedicine, virtual reality, Entertainment, rehabilitation project and space flight, military affairs, embodies important value.

Stable state vision inducting electricity (Steady State Visually Evoked Potential, SSVEP) be that the brain vision system continues the periodically response of visual stimulus to outside, it is input signal commonly used in brain-computer interface BCI system, compare P300, the event related synchronization, the signals such as spontaneous brain electricity, Steady State Visual Evoked Potential (SSVEP) is many owing to having the number of targets of control, rate of information transmission is high and antijamming capability is strong, training time is short, recording electrode is few, the outstanding advantages that waits simple to operate, show following research potential and potential practical value, therefore become the important research normal form that there is wide application prospect and using value in practical BCI system.

In current SSVEP-BCI system applies, SSVEP mainly utilizes the following low frequency region of 30Hz, adopt the cursor flicker stimulation mode of a goal task of a frequency representation, restriction due to frequency field, frequency resolution and the response amplitude of target performance, in order to guarantee the recognition correct rate of goal task, make the task object number that can present limited.Simultaneously, traditional SSVEP identification efficiency is low, needs repeatedly stimulus signal cumulative, has had a strong impact on the real-time identification ability.Therefore, the target numbers that increase can present, the rate of information transmission that promotes the SSVEP technology becomes the key problem of current SSVEP-BCI system.

Summary of the invention

In order to overcome the shortcoming of above-mentioned prior art, the invention provides based on which amplitude modulation VEP brain-computer interface method, normal form (Amplitude Modulated Visual Evoked Potential based on the which amplitude modulation VEP has been proposed, be called for short AMVEP), and modulated signal analysis and the feature extracting method adapt with it proposed, the AMVEP single of the normal form of realization based on the which amplitude modulation VEP or few signal identification, have advantages of simple to operate, electrode number is few and number of targets is many.

In order to achieve the above object, the technical scheme that the present invention takes is:

Based on which amplitude modulation VEP brain-computer interface method, comprise the following steps:

Step 1, at the occipital region of subject's head D Oz position of sound production potential electrode A, one-sided ear-lobe position of sound production reference electrode B at subject's head D, at the frons Fpz of subject's head D position of sound production ground electrode C, the first input end E1 of the output terminal access eeg amplifier E of potential electrode A, the second input end E2 of the output terminal access eeg amplifier E of reference electrode B, the 3rd input end E3 of the output terminal access eeg amplifier E of ground electrode C, the output terminal of eeg amplifier E is connected with the input end of computing machine F, the output terminal of computing machine F and computer screen G are connected to form the eeg signal acquisition system,

Step 2, stimulate normal form to be presented in face of the experimenter by computer screen G the denotable unifrequency of computer screen G, subject's head D distance calculation screen G is 50～100 centimetres, bring out potential response amplitude susceptibility rule according to the experimenter under different unifrequencys stimulate, optimize and select responsive stimuli responsive frequency to stimulate the carrier frequency of normal form as the modulation VEP, optimizing selection principle is: according to experimenter's difference, make the amplitude-frequency response figure of the optional frequency of this experimenter, from amplitude-frequency response figure, select the corresponding frequency values of amplitude maximal value as carrier frequency, select the most responsive frequency of experimenter's amplitude-frequency response as carrier frequency, maximize the signal to noise ratio (S/N ratio) of response signal,

Step 3, according to the selected carrier frequency of step 2, and the factor that affects the selection of modulating wave modulating frequency, the range of choice of modulating wave modulating frequency is proposed, because the factor that affects modulating wave comprises screen refresh rate and spectral resolution, so carrier frequency must meet sampling thheorem to the modulating wave modulating frequency, the method that the modulating wave modulating frequency is selected specifically comprises following content:

Amplitude modulated signal: y (t)=C+C*sin (2* π * f _m* Δ t _m-pi/2) (1)

Sampling thheorem: $\frac{f_{\mathrm{ms}}}{f_m}\≥2.56---\left(2\right)$

In the amplitude modulated signal computing formula, the transformation relation of parameter is as follows:

F _Ms→ f ₀, that is: ${\mathrm{\&Delta;t}}_m\&RightArrow;{\mathrm{\&Delta;<figure-callout\; id="0"\; label="t"\; filenames="HDA0000088735540000012.png,HDA0000088735540000021.png"\; state="\{\{state\}\}">t</figure-callout>}}_0=\frac{1}{{\mathrm{<figure-callout\; id="0"\; label="f"\; filenames="HDA0000088735540000012.png,HDA0000088735540000021.png"\; state="\{\{state\}\}">f</figure-callout>}}_0}=\frac{n}{f_r}---\left(3\right)$

${\mathrm{\&Delta;<figure-callout\; id="0"\; label="t"\; filenames="HDA0000088735540000012.png,HDA0000088735540000021.png"\; state="\{\{state\}\}">t</figure-callout>}}_0={\mathrm{\&Delta;t}}_m*\mathrm{\&lambda;}\&DoubleRightArrow;\frac{1}{\mathrm{\&lambda;}}=\frac{{\mathrm{\&Delta;t}}_m}{{\mathrm{\&Delta;<figure-callout\; id="0"\; label="t"\; filenames="HDA0000088735540000012.png,HDA0000088735540000021.png"\; state="\{\{state\}\}">t</figure-callout>}}_0}=\frac{1}{f_{\mathrm{ms}}}/\frac{n}{f_r}=\frac{f_r}{f_{\mathrm{ms}}*n}---\left(4\right)$

Derive and learn that final frequency of modulated wave is by above formula (1)～(4):

$X=\frac{f_m}{\mathrm{\&lambda;}}=\frac{f_m*f_r}{f_{\mathrm{ms}}*n}\&DoubleRightArrow;\frac{f_{\mathrm{ms}}}{f_m}=\frac{f_r}{n*x}\&GreaterEqual;2.56\&DoubleRightArrow;x\&le;\frac{f_r}{n*2.56}$

Wherein:

C: constant

F _Ms: the sample frequency of amplitude modulated signal

F _m: the frequency of amplitude modulated signal

F ₀: carrier frequency

Δ t ₀: the sampling interval of carrier signal

: the sampling interval of amplitude modulated signal

λ: dimensionless factor

X: modulation wave frequency

N: the frame number that means the required screen-refresh of one-period

F _r: screen refresh rate

Learn that by the above process of shifting onto the scope of frequency of modulated wave should be:

$0\&le;x\&le;\frac{f_r}{n*2.56}=\frac{{\mathrm{<figure-callout\; id="0"\; label="f"\; filenames="HDA0000088735540000012.png,HDA0000088735540000021.png"\; state="\{\{state\}\}">f</figure-callout>}}_0}{2.56};$

Step 4, according to the selected carrier frequency of step 2 optimization principles, and the range of choice of frequency of modulated wave that step 3 proposes, design AMVEP normal form, produce the stimulus sequence of which amplitude modulation VEP brain-computer interface normal form, by computer screen G, be presented in face of the experimenter, the computing formula of stimulus sequence that produces the AMVEP normal form is as follows:

$y_{\mathrm{AMVEP}}\left(t\right)=C+C*\sin (2*\mathrm{\&pi;}*\frac{n*x}{f_r}-\mathrm{\&pi;}/2);$

Wherein, C is constant, and x is designed frequency of modulated wave, and knows according to step 3:

$x\&le;\frac{f_r}{n*2.56};$

Step 5, first by step 2～3, determine modulating-coding AMVEP normal form, by step 4, it is presented in face of the experimenter again, the final AMVEP response signal that obtains the experimenter by the described eeg signal acquisition system of step 1, the modulated signal analysis and the feature extracting method that adapt are with it proposed, realize single or few the signal identification of AMVEP, specifically comprise the following steps:

◆ the AMVEP response eeg data to the different stimulated sequence that collects carries out bandpass filtering;

◆ adopt Hilbert (Hilbert) demodulation method to carry out amplitude demodulation to filtered data;

◆ the data after adopting fast fourier transform (Fast Fourier transform, FFT) method to amplitude demodulation are asked its spectrum maximum;

◆ according to spectrum results, obtain frequency values corresponding to maximum value in frequency spectrum, be the frequency of modulated wave value.

There is the problem that system stability is not high, rate of information transmission is low that target numbers is few, stimulation time is long, experimental paradigm simply causes in the BCI system based on SSVEP at present that the present invention is directed to, proposed to bring out current potential brain-computer interface normal form and the feature extraction demodulation method that adapts with it based on modulation, realize single or few the signal identification of AMVEP, take efficient and can't harm as final goal for realizing building, there is brain-computer interface technology simple to operate, that electrode number is few, number of targets is many and opened up new thinking, shown following superiority:

1, modulation VEP brain-computer interface normal form

Propose modulation VEP brain-computer interface normal form, greatly increased the optional number of targets of AMVEP, and utilized the strong advantage of responsive carrier frequency antijamming capability, improved the signal to noise ratio (S/N ratio) of response signal.

2, which amplitude modulation is brought out the feature extracting method of potential response

Propose features of response based on AMVEP and propose individual features extraction signal processing scheme, realize single or few the signal identification of AMVEP, improve the identification efficiency of brain-computer interface, the high efficiency of transmission of guarantee brain-computer interface information.

The accompanying drawing explanation

Fig. 1 is hardware connection diagram of the present invention.

Fig. 2 is stimulus sequence schematic diagram of the present invention.

Fig. 3 is amplitude-frequency curve chart that the CRT unifrequency of 100Hz refresh rate stimulates lower AMVEP response.

Fig. 4 is based on the process flow diagram of which amplitude modulation VEP EEG feature extraction method.

Fig. 5-a is the time domain waveform figure after the data filtering of 1.25Hz modulation.

Fig. 5-b is the FFT spectrogram of data after filtering.

Fig. 5-c is the amplitude envelope FFT spectrogram of data after filtering.

Embodiment

Below in conjunction with accompanying drawing, the present invention is described in further detail.

Based on which amplitude modulation VEP brain-computer interface method, comprise the following steps:

Step 1, with reference to Fig. 1, at the occipital region of subject's head D Oz position of sound production potential electrode A, one-sided ear-lobe position of sound production reference electrode B at subject's head D, at the frons Fpz of subject's head D position of sound production ground electrode C, the first input end E1 of the output terminal access eeg amplifier E of potential electrode A, the second input end E2 of the output terminal access eeg amplifier E of reference electrode B, the 3rd input end E3 of the output terminal access eeg amplifier E of ground electrode C, the output terminal of eeg amplifier E is connected with the input end of computing machine F, the output terminal of computing machine F and computer screen G are connected to form the eeg signal acquisition system,

Step 2, see figures.1.and.2, stimulate normal form to be presented in face of the experimenter by computer screen G the denotable unifrequency of computer screen G, subject's head D distance calculation screen G is 50～100 centimetres, wherein, optional frequency is subject to the screen refresh rate restriction, reach as high as 100Hz and test at present CRT monitor screen refresh rate used setting, therefore as shown in table 1 in the selectable frequency of 6～50Hz, bring out potential response amplitude susceptibility rule according to the experimenter under different unifrequencys stimulate, optimize and select responsive stimuli responsive frequency to stimulate the carrier frequency of normal form as the modulation VEP, optimizing selection principle is: according to experimenter's difference, make the amplitude-frequency response figure of the optional frequency of this experimenter, from amplitude-frequency response figure, select the corresponding frequency values of amplitude maximal value as carrier frequency, select the most responsive frequency of experimenter's amplitude-frequency response as carrier frequency, maximize the signal to noise ratio (S/N ratio) of response signal, with reference to Fig. 3, the responsive stimuli responsive frequency of the experimenter of known participation experiment is 12.5Hz,

Table 1: test CRT monitor used and can mean frequency and refresh the frame number corresponding relation

Step 3, according to the selected carrier frequency of step 2, and the factor that affects the selection of modulating wave modulating frequency, the range of choice of modulating wave modulating frequency is proposed, because the factor that affects modulating wave comprises screen refresh rate and spectral resolution, so carrier frequency must meet sampling thheorem to the modulating wave modulating frequency, the method that the modulating wave modulating frequency is selected specifically comprises following content:

Amplitude modulated signal: y (t)=C+C*sin (2* π * f _m* Δ t _m-pi/2) (1)

Sampling thheorem: $\frac{f_{\mathrm{ms}}}{f_m}\&GreaterEqual;2.56---\left(2\right)$

In the amplitude modulated signal computing formula, the transformation relation of parameter is as follows:

F _Ms→ f ₀, that is: ${\mathrm{\&Delta;t}}_m\&RightArrow;{\mathrm{\&Delta;<figure-callout\; id="0"\; label="t"\; filenames="HDA0000088735540000012.png,HDA0000088735540000021.png"\; state="\{\{state\}\}">t</figure-callout>}}_0=\frac{1}{{\mathrm{<figure-callout\; id="0"\; label="f"\; filenames="HDA0000088735540000012.png,HDA0000088735540000021.png"\; state="\{\{state\}\}">f</figure-callout>}}_0}=\frac{n}{f_r}---\left(3\right)$

${\mathrm{\&Delta;<figure-callout\; id="0"\; label="t"\; filenames="HDA0000088735540000012.png,HDA0000088735540000021.png"\; state="\{\{state\}\}">t</figure-callout>}}_0={\mathrm{\&Delta;t}}_m*\mathrm{\&lambda;}\&DoubleRightArrow;\frac{1}{\mathrm{\&lambda;}}=\frac{{\mathrm{\&Delta;t}}_m}{{\mathrm{\&Delta;<figure-callout\; id="0"\; label="t"\; filenames="HDA0000088735540000012.png,HDA0000088735540000021.png"\; state="\{\{state\}\}">t</figure-callout>}}_0}=\frac{1}{f_{\mathrm{ms}}}/\frac{n}{f_r}=\frac{f_r}{f_{\mathrm{ms}}*n}---\left(4\right)$

Derive and learn that final frequency of modulated wave is by above formula (1)～(4):

$X=\frac{f_m}{\mathrm{\&lambda;}}=\frac{f_m*f_r}{f_{\mathrm{ms}}*n}\&DoubleRightArrow;\frac{f_{\mathrm{ms}}}{f_m}=\frac{f_r}{n*x}\&GreaterEqual;2.56\&DoubleRightArrow;x\&le;\frac{f_r}{n*2.56}$

Wherein:

C: constant

F _Ms: the sample frequency of amplitude modulated signal

F _m: the frequency of amplitude modulated signal

F ₀: carrier frequency

Δ t ₀: the sampling interval of carrier signal

: the sampling interval of amplitude modulated signal

λ: dimensionless factor

X: modulation wave frequency

N: the frame number that means the required screen-refresh of one-period

F _r: screen refresh rate

Learn that by the above process of shifting onto the scope of frequency of modulated wave should be:

$0\&le;x\&le;\frac{f_r}{n*2.56}=\frac{{\mathrm{<figure-callout\; id="0"\; label="f"\; filenames="HDA0000088735540000012.png,HDA0000088735540000021.png"\; state="\{\{state\}\}">f</figure-callout>}}_0}{2.56};$

As screen refresh rate f _r=100Hz;

Carrier frequency $f_{}$ ${}_0=\frac{f_r}{n}=\frac{100}{8}=12.5;$

The scope of frequency of modulated wave should be:

$0\&le;x\&le;\frac{12.5}{2.56}$

Step 4, with reference to Fig. 1, Fig. 2 and Fig. 3, according to the selected carrier frequency of step 2 optimization principles, and the range of choice of frequency of modulated wave that step 3 proposes, design AMVEP normal form, produce the stimulus sequence of which amplitude modulation VEP brain-computer interface normal form, be presented in face of the experimenter by computer screen G, the computing formula of stimulus sequence that produces the AMVEP normal form is as follows:

$y_{\mathrm{AMVEP}}\left(t\right)=C+C*\sin (2*\mathrm{\&pi;}*\frac{n*x}{f_r}-\mathrm{\&pi;}/2);$

Wherein, C is constant, and x is designed frequency of modulated wave, and knows according to step 3:

$x\&le;\frac{f_r}{n*2.56};$

Step 5, with reference to Fig. 4, first by step 2～3, determine modulating-coding AMVEP normal form, by step 4, it is presented in face of the experimenter again, the final AMVEP response signal that obtains the experimenter by the described eeg signal acquisition system of step 1, propose the modulated signal analysis and the feature extracting method that adapt with it, realize single or few the signal identification of AMVEP, specifically comprise the following steps:

◆ the AMVEP response eeg data to the different stimulated sequence that collects carries out bandpass filtering, with reference to Fig. 5-a;

◆ adopt the Hilbert demodulation method to carry out amplitude demodulation to filtered data;

◆ the data after adopting the FFT method to amplitude demodulation are asked its spectrum maximum, with reference to Fig. 5-b;

◆ according to spectrum results, obtain frequency values corresponding to maximum value in frequency spectrum, be the frequency of modulated wave value, with reference to Fig. 5-c.

Claims (1)

1. based on which amplitude modulation VEP brain-computer interface method, it is characterized in that, comprise the following steps:

Step 1, in the occipital region of subject's head (D) Oz position of sound production potential electrode (A), one-sided ear-lobe position of sound production reference electrode (B) in subject's head (D), at the frons Fpz of subject's head (D) position of sound production ground electrode (C), the first input end (E1) of the output terminal access eeg amplifier (E) of potential electrode (A), second input end (E2) of the output terminal access eeg amplifier (E) of reference electrode (B), the 3rd input end (E3) of the output terminal access eeg amplifier (E) of ground electrode (C), the output terminal of eeg amplifier (E) is connected with the input end of computing machine (F), the output terminal of computing machine (F) and computer screen (G) are connected to form the eeg signal acquisition system,

Step 2, stimulate normal form to be presented in face of the experimenter by computer screen (G) the denotable unifrequency of computer screen (G), subject's head (D) distance calculation screen (G) is 50～100 centimetres, bring out potential response amplitude susceptibility rule according to the experimenter under different unifrequencys stimulate, optimize and select responsive stimuli responsive frequency to stimulate the carrier frequency of normal form as the modulation VEP, optimizing selection principle is: according to experimenter's difference, make the amplitude-frequency response figure of the optional frequency of this experimenter, from amplitude-frequency response figure, select the corresponding frequency values of amplitude maximal value as carrier frequency, select the most responsive frequency of experimenter's amplitude-frequency response as carrier frequency, maximize the signal to noise ratio (S/N ratio) of response signal,

Step 3, according to the selected carrier frequency of step 2, and the factor that affects the selection of modulating wave modulating frequency, the range of choice of modulating wave modulating frequency is proposed, because the factor that affects modulating wave comprises screen refresh rate and spectral resolution, so carrier frequency must be closed sampling thheorem to the modulating wave modulating frequency, the method that the modulating wave modulating frequency is selected specifically comprises following content:

Amplitude modulated signal: y (t)=C+C*sin (2* π * f _m* Δ t _m-pi/2) (1)

Sampling thheorem: $\frac{f_{\mathrm{ms}}}{f_m}\&GreaterEqual;2.56---\left(2\right)$

In the amplitude modulated signal computing formula, the transformation relation of parameter is as follows:

F _Ms→ f ₀, that is: ${\mathrm{\&Delta;t}}_m\&RightArrow;{\mathrm{\&Delta;t}}_0=\frac{1}{f_0}=\frac{n}{f_r}---\left(3\right)$

${\mathrm{\&Delta;t}}_0={\mathrm{\&Delta;t}}_m*\mathrm{\&lambda;}\&DoubleRightArrow;\frac{1}{\mathrm{\&lambda;}}=\frac{{\mathrm{\&Delta;t}}_m}{{\mathrm{\&Delta;t}}_0}=\frac{1}{f_{\mathrm{ms}}}/\frac{n}{f_r}=\frac{f_r}{f_{\mathrm{ms}}*n}---\left(4\right)$

Derive and learn that final frequency of modulated wave is by above formula (1)～(4):

$X=\frac{f_m}{\mathrm{\&lambda;}}=\frac{f_m*f_r}{f_{\mathrm{ms}}*n}\&DoubleRightArrow;\frac{f_{\mathrm{ms}}}{f_m}=\frac{f_r}{n*x}\&GreaterEqual;2.56\&DoubleRightArrow;x\&le;\frac{f_r}{n*2.56}$

Wherein:

C: constant

F _Ms: the sample frequency of amplitude modulated signal

F _m: the frequency of amplitude modulated signal

F ₀: carrier frequency

Δ t ₀: the sampling interval of carrier signal

: the sampling interval of amplitude modulated signal

λ: dimensionless factor

X: frequency of modulated wave

N: the frame number that means the required screen-refresh of one-period

F _r: screen refresh rate

Learn that by the above process of shifting onto the scope of frequency of modulated wave should be:

$0\&le;x\&le;\frac{f_r}{n*2.56}=\frac{f_0}{2.56};$

Step 4, according to the selected carrier frequency of step 2 optimization principles, and the range of choice of frequency of modulated wave that step 3 proposes, design is based on which amplitude modulation VEP (Amplitude Modulated Visual Evoked Potential, abbreviation AMVEP) normal form, produce the stimulus sequence of which amplitude modulation VEP brain-computer interface normal form, be presented in face of the experimenter by computer screen (G), the computing formula of stimulus sequence that produces the AMVEP normal form is as follows:

$y_{\mathrm{AMVEP}}\left(t\right)=C+C*\sin (2*\mathrm{\&pi;}*\frac{n*x}{f_r}-\mathrm{\&pi;}/2);$

Wherein, C is constant, and x is designed frequency of modulated wave, and knows according to step 3:

$x\&le;\frac{f_r}{n*2.56};$

Step 5, first by step 2～3, determine modulating-coding AMVEP normal form, by step 4, it is presented in face of the experimenter again, the final AMVEP response signal that obtains the experimenter by the described eeg signal acquisition system of step 1, the modulated signal analysis and the feature extracting method that adapt are with it proposed, realize single or few the signal identification of AMVEP, specifically comprise the following steps:

◆ the AMVEP response eeg data to the different stimulated sequence that collects carries out bandpass filtering;

◆ adopt Hilbert transform (Hilbert) demodulation method to carry out amplitude demodulation to filtered data;

◆ the data after adopting fast fourier transform (Fast Fourier transform, FFT) method to amplitude demodulation are asked its spectrum maximum;

◆ according to spectrum results, obtain frequency values corresponding to maximum value in frequency spectrum, be the frequency of modulated wave value.

Images