CN102609618A - Method for calculating brain asymmetric index based on information flow gain - Google Patents

Abstract

The invention discloses a method for calculating a brain asymmetric index based on information flow gain, which relates to the field of neural engineering. The method comprises the following steps: setting corresponding electrodes on a head of a testee, sending a parallel port synchronous signal to the electrodes through a computer and collecting brain wave of the testee through the electrodes; pre-processing the brain wave; calculating a brain function network connection matrix through a directed transfer function method according to the pre-processed brain wave; and calculating to obtain the brain asymmetric index based on the information flow gain according to the brain function network connection matrix. In the method, from the point of view of brain information exchange and integration, deep mechanisms of a brain are researched, and the brain asymmetric state is observed. The method is a supplement of brain function mechanism research methods and has important enlightenment for deeply searching the deep mechanisms of the brain.

Description

The asymmetric index calculation method of brain based on the information flow gain

Technical field

The present invention relates to neural field of engineering technology, particularly a kind of asymmetric index calculation method of brain based on the information flow gain.

Background technology

The main task of Modern Cognitive Neuroscience Research is the neuromechanism that cognitive function is understood in research.Present most brain function research all is to adopt the thinking of " cognitive subtraction " experimentize design and data analysis, and this thinking just is based on the research thinking of functional localization in brief.Whether the research method of " cognitive subtraction " is set up will be based on a hypothesis, and this hypothesis thinks that each functional areas of brain are modular, and promptly relevant with each functional areas cognitive function composition is to separate, independently with activity.

Part result of study confirms that above-mentioned hypothesis sets up, yet increasing research shows that even cognitive task is very simple, the integration and cooperation that its realization also depends on the brain domain of a plurality of separation could be accomplished.With respect to the research thinking of aforementioned " cognitive subtraction ", the concurrency of information flow between the different cerebral district and brain function network has become the emphasis that researchers pay close attention to.With respect to the research thinking of functional localization, this method thinking can be described as the thinking that function is integrated.In order to explore the brain function deeper mechanisms, a rational method is that comprehensive functions of use location and two kinds of thinkings of function integration are studied, and replenishes each other.

The above-mentioned functions integrated mechanism can be described with the brain network, its research to as if the different cerebral functional areas between reciprocation, and different cognitive tasks is to this interactive modulation and influence.Yet, also do not have a kind of angle research brain deeper mechanisms from brain information interaction, integration at present, and then the method for strong help is provided for brain science research.

Summary of the invention

The technical matters that (one) will solve

The technical matters that the present invention will solve is: how a kind of asymmetric index calculation method of brain based on information flow gain is provided, so that from the angle research brain deeper mechanisms of brain information interaction, integration.

(2) technical scheme

For solving the problems of the technologies described above, the present invention provides a kind of asymmetric index calculation method of brain based on the information flow gain, and it comprises step:

A: the head person to be measured is provided with corresponding electrode, sends the parallel port synchronizing signal through computing machine to said electrode, gathers person's to be measured brain wave through said electrode;

B: said brain wave is carried out pre-service;

C:, calculate brain function network connection matrix through oriented transfer function method according to pretreated brain wave;

D:, calculate brain asymmetric index based on the information flow gain according to said brain function network connection matrix.

Preferably, said steps A specifically comprises step:

A1: the full brain position person to be measured is provided with test electrode, in person's to be measured mastoid process position reference electrode is set, ground-electrode ground connection;

A2: to said test electrode, reference electrode and ground-electrode, send the parallel port synchronizing signal through computing machine;

A3: the brain wave that detects person to be measured through said test electrode, reference electrode and ground-electrode;

A4: with sending to said computing machine after amplification of said brain wave process and the analog to digital conversion.

Preferably, said step B specifically comprises step:

B1: said brain wave is carried out Filtering Processing;

B2: go baseline to handle to the brain wave after the Filtering Processing;

B3: go the electric artefact of eye to handle to removing the brain wave after baseline is handled;

B4: the moment alignment of the corresponding said parallel port synchronizing signal of the each brain electric potential in the brain wave that goes after the electric artefact of eye is handled.

Preferably, said step C specifically comprises step:

C1: from pretreated brain wave, choose frequency band to be calculated;

C2: use red pond quantity of information criterion to calculate the order that said pretreated brain wave is carried out match, use the brain wave data in the said frequency band of multivariate regression model match according to said order;

C3:,, calculate the corresponding brain function network connection matrix of each discrete frequency in the said frequency band respectively according to the brain wave data in the said frequency band after the match through oriented transfer function method;

C4: use the alternate data method successively the corresponding brain function network connection matrix of each discrete frequency in the said frequency band to be connected the conspicuousness checking, remove because the meaningless connection that randomness causes.

Preferably, said step D specifically comprises step:

D1: flow out the direction summation with the stream of the behavioural information in the said brain function network connection matrix and obtain flowing out the value of information, obtain flowing into the value of information with the information flow inflow direction summation of classifying as in the said brain function network connection matrix;

D2: with the middle line of test electrode longitudinally of person's to be measured brain is axle, and half brain about being divided into calculates the information flow gain of left half each test electrode of brain and the information flow gain of right half each test electrode of brain respectively; Said information flow gain is the outflow value of information on the corresponding test electrode and the ratio that flows into the value of information;

D3: according to the left-right symmetric principle, according to about the corresponding right information flow gain of test electrode, utilize the asymmetric index computing formula, calculate the brain asymmetric index of each test electrode successively to correspondence.

(3) beneficial effect

The asymmetric index calculation method of brain based on the information flow gain of the present invention utilizes oriented transport function to calculate brain function network connection matrix, through obtaining flowing out the value of information, flow into the value of information by row, column direction summation; Obtain the information flow gain through flowing out the value of information with the ratio that flows into the value of information, and then obtain brain asymmetric index based on the information flow gain according to asymmetric formula of index.Said method is observed the brain asymmetrical state from the deeper mechanisms of the angle research brain of brain information interaction, integration, is replenishing of brain function Mechanism Study method, for further investigation brain deep layer mechanism important enlightenment is arranged.

Description of drawings

Fig. 1 is the described asymmetric index calculation method process flow diagram of brain based on the information flow gain of the embodiment of the invention;

Fig. 2 be about the half brain division synoptic diagram that leads;

Fig. 3 a～b is based on the asymmetric index brain mapping of brain that the information flow gain calculating obtains.

Embodiment

Below in conjunction with accompanying drawing and embodiment, specific embodiments of the invention describes in further detail.Following examples are used to explain the present invention, but are not used for limiting scope of the present invention.

Fig. 1 is the described asymmetric index calculation method process flow diagram of brain based on the information flow gain of the embodiment of the invention.As shown in Figure 1, said method comprises:

Steps A: the head person to be measured is provided with corresponding electrode, sends the parallel port synchronizing signal through computing machine to said electrode, gathers person's to be measured brain wave through said electrode.

Said steps A specifically comprises:

Steps A 1: the full brain position person to be measured is provided with test electrode, in person's to be measured mastoid process position reference electrode is set, ground-electrode ground connection.Fig. 2 be about the half brain division synoptic diagram that leads, as shown in Figure 2, present embodiment adopts general eeg recording equipment, employings frequency is 1000Hz, 64 lead (test electrode) of employing during record, the position of leading 10-20 system setting in accordance with international practices.Wherein, it is not shown in Fig. 2 to be arranged on the reference electrode of mastoid process position.

Steps A 2: to said test electrode, reference electrode and ground-electrode, send the parallel port synchronizing signal through computing machine.

Steps A 3: the brain wave that detects person to be measured through said test electrode, reference electrode and ground-electrode.

Steps A 4: with sending to said computing machine after amplification of said brain wave process and the analog to digital conversion.

Step B: said brain wave is carried out pre-service.

Said step B specifically comprises:

Step B1: said brain wave is carried out Filtering Processing, disturb to get rid of the 50Hz power frequency.

Step B2: go baseline to handle to the brain wave after the Filtering Processing, remove baseline wander, so that brain wave becomes stably, average is close to 0 signal.

Step B3: go the electric artefact of eye to handle to removing the brain wave after baseline is handled, remove the corresponding data division of eye movement.

Step B4: the moment alignment of the corresponding said parallel port synchronizing signal of the each brain electric potential in the brain wave that goes after the electric artefact of eye is handled.

Step C:, calculate brain function network connection matrix through oriented transport function (DTF, directed transfer function) method according to pretreated brain wave.Said step C specifically comprises:

Step C1: from pretreated brain wave, choose frequency band to be calculated.Choosing of said frequency band looked different test assignment and difference, and such as the brain wave of test motion process, then corresponding frequency band is generally 8～13Hz; If the brain wave of test thinking processes, then corresponding frequency band is generally 0.4～5Hz.

Step C2: use red pond quantity of information criterion (AIC, Akaike information criterion) to calculate the order that said pretreated brain wave is carried out match, use the brain wave data in the said frequency band of multivariate regression model match according to said order.

Known time is t, and the number that leads is N (N is 64 in the present embodiment), and the brain wave data X in the said frequency band can be described as:

X＝[x ₁(t)，x ₂(t)，...，x _N(t)] ^T； (1)

Obtain after using the multivariate regression model match:

$\underset{k=0}{\overset{p}{\mathrm{\Σ}}}\mathrm{\Λ}\left(k\right)X(t-k)=E\left(t\right);---\left(2\right)$

Wherein, Λ (0)=I, the element among N * N matrix Λ (k) is the multivariate regression model parameter; E (t) is a many reference amounts zero-mean white noise vector; The said order that p is to use red pond quantity of information criterion to calculate, the size of order can influence the effect of match; K is a natural number, and 0≤k≤p.

Step C3:,, calculate the corresponding brain function network connection matrix of each discrete frequency in the said frequency band respectively according to the brain wave data in the said frequency band after the match through oriented transfer function method.Suppose that test assignment is the brain wave of test motion process, selected frequency band is 8～13Hz, and the frequency values (like 8Hz, 9Hz, 10Hz, 11Hz, 12Hz and 13Hz) that generally can choose round values point calculates corresponding brain function network connection matrix respectively.

At first above-mentioned (2) formula is transformed into frequency domain, obtains:

X(f)＝Λ ^-1(f)E(f)＝H(f)E(f)； (3)

Wherein, $\mathrm{\Λ}\left(f\right)=\underset{k=0}{\overset{p}{\mathrm{\Σ}}}\mathrm{\Λ}\left(k\right)e^{-j2\mathrm{\π\; F\Δ\; Tk}};$

And H (f) is the systems communicate matrix, element H among the H (f) _IjValue to have described with j be input, i is two of the output strength of joint between leading, i and j represent corresponding numbering of leading.Can find out that by above-mentioned (3) formula multivariate regression model is equivalent to a black box, noise is input, through over-fitting output signal.All spectrum signature information and lead between link information be included in the systems communicate matrix H (f).Thus, through oriented transfer function method, the brain function network connection matrix that calculates each discrete frequency correspondence in the said frequency band respectively is following:

${\mathrm{\θ}}_{\mathrm{ij}}^2\left(f\right)={\left|H_{\mathrm{ij}}\left(f\right)\right|}^2;---\left(4\right)$

Above-mentioned (4) formula is carried out obtaining after the normalization:

${\mathrm{\γ}}_{\mathrm{ij}}^2\left(f\right)={\left|H_{\mathrm{ij}}\left(f\right)\right|}^2/\underset{m=1}{\overset{N}{\mathrm{\Σ}}}{\left|H_{\mathrm{ij}}\left(f\right)\right|}^2;---\left(5\right)$

Step C4: use the alternate data method successively the corresponding brain function network connection matrix of each discrete frequency in the said frequency band to be connected the conspicuousness checking, remove because the meaningless connection that randomness causes.

Step D:, calculate brain asymmetric index based on the information flow gain according to said brain function network connection matrix.Said step D specifically comprises:

Step D1: flow out the direction summation with the stream of the behavioural information in the said brain function network connection matrix and obtain flowing out the value of information, obtain flowing into the value of information with the information flow inflow direction summation of classifying as in the said brain function network connection matrix.According to above-mentioned (5) formula, the computing formula of the outflow value of information of leading that can obtain being numbered m is following:

${\mathrm{flow}}_{\mathrm{OUT}}=\underset{m=1}{\overset{N}{\mathrm{\Σ}}}{\mathrm{\γ}}_{\mathrm{im}}^2;---\left(6\right)$

The computing formula of the inflow value of information of leading that obtains being numbered m is following:

${\mathrm{flow}}_{\mathrm{IN}}=\underset{j=1}{\overset{N}{\mathrm{\Σ}}}{\mathrm{\γ}}_{\mathrm{mj}}^2;---\left(7\right)$

Wherein, flow into the value of information and reflected to be that the target that function connects is led with the special leads, as the recipient's of other information of leading situation.On the contrary, the outflow value of information as the source, has reflected special leads from this and has led to other information transmitted comprehensive condition that leads.

Step D2: with the middle line of test electrode longitudinally of person's to be measured brain is axle, and half brain about being divided into calculates the information flow gain of left half each test electrode of brain and the information flow gain of right half each test electrode of brain respectively; Said information flow gain is the outflow value of information on the corresponding test electrode and the ratio that flows into the value of information.

Said information flow Calculation of Gain formula is following:

${\mathrm{\ρ}}_m=\frac{{\mathrm{flow}}_{\mathrm{OUT}}}{{\mathrm{flow}}_{\mathrm{IN}}}=\frac{\underset{i=1}{\overset{N}{\mathrm{\Σ}}}{\mathrm{\γ}}_{\mathrm{im}}^2}{\underset{j=1}{\overset{N}{\mathrm{\Σ}}}{\mathrm{\γ}}_{\mathrm{mj}}^2};---\left(8\right)$

The present embodiment method adopts 64 to lead altogether, and referring to Fig. 2, said 64 test electrode lines (being dotted line among Fig. 2) that lead with brain central authorities are axis of symmetry, by about be divided into two groups.According to symmetry principle, it is right that the corresponding test electrode on test electrode on left half brain and right half brain constitutes test electrode, like Fp1 among Fig. 2 and Fp2; It is right that test electrode on the axis of symmetry and himself constitute test electrode, like the Fpz among Fig. 2.

Step D3: according to the left-right symmetric principle, according to about the corresponding right information flow gain of test electrode, utilize the asymmetric index computing formula, calculate the brain asymmetric index of each test electrode successively to correspondence.

Said asymmetric index computing formula is following:

$\mathrm{AI}=\frac{1}{2}\×\frac{({\mathrm{\ρ}}_L-{\mathrm{\ρ}}_R)}{({\mathrm{\ρ}}_L+{\mathrm{\ρ}}_R)};---\left(9\right)$

Wherein, AI representes the brain asymmetric index of test electrode to correspondence, ρ _LThe information flow gain of expression test electrode centering left side test electrode, ρ _RThe information flow gain of expression test electrode centering right side test electrode.According to (9) formula, can know that the test electrode on the said axis of symmetry is 0 to the brain asymmetric index of correspondence.

Fig. 3 a～b is based on the asymmetric index brain mapping of brain that the information flow gain calculating obtains.Wherein, Fig. 3 a representes the asymmetric index brain mapping of brain that individual subjects obtains under two different frequency flicker pieces stimulate; Fig. 3 b representes that whole 11 experimenters stimulate the asymmetric index brain mapping of average brain that obtains down at two different frequencies flicker pieces, among Fig. 3 a and Fig. 3 b about two row results said two different frequencies of the correspondence piece that glimmers respectively.From Fig. 3 a and Fig. 3 b, can see three apparent in view difference sections: comprise the language district near forehead, the temporal lobe, and near the occipital region on the upper side.Have research to point out right front frontal region often greater than left front volume, occipital region, the left side is often greater than the right.In the occipital region, half brain asymmetry about forehead and temporal lobe can be found significantly, this kind asymmetry shows in the distribution of grey matter and white matter.In the brain mapping result of the inventive method, can see the result who conforms to basically with report.What is more important, the inventive method have confirmed the result that conforms to the brain structure from the information interaction angle.In Fig. 3 a and Fig. 3 b, can't see tangible asymmetric difference in the occipital region, this is because we have adopted the experimental paradigm of steady-state induced current potential; Though on structure; Occipital region, the brain left side is often greater than the right, and when carrying out stable state vision inducting current potential task, on the information interaction; The occipital region demonstrates the asymmetry characteristic that is different from the brain structure, and this is owing to due to the warbled reason.

The said asymmetric index calculation method of brain based on the information flow gain of the embodiment of the invention utilizes oriented transport function to calculate brain function network connection matrix, through obtaining flowing out the value of information, flow into the value of information by row, column direction summation; Obtain the information flow gain through flowing out the value of information with the ratio that flows into the value of information, and then obtain brain asymmetric index based on the information flow gain according to asymmetric formula of index.Said method is observed the brain asymmetrical state from the deeper mechanisms of the angle research brain of brain information interaction, integration, is replenishing of brain function Mechanism Study method, for further investigation brain deep layer mechanism important enlightenment is arranged.

Above embodiment only is used to explain the present invention; And be not limitation of the present invention; The those of ordinary skill in relevant technologies field under the situation that does not break away from the spirit and scope of the present invention, can also be made various variations and modification; Therefore all technical schemes that are equal to also belong to category of the present invention, and scope of patent protection of the present invention should be defined by the claims.

Claims (5)

1. the asymmetric index calculation method of brain based on the information flow gain is characterized in that, comprises step:

A: the head person to be measured is provided with corresponding electrode, sends the parallel port synchronizing signal through computing machine to said electrode, gathers person's to be measured brain wave through said electrode;

B: said brain wave is carried out pre-service;

C:, calculate brain function network connection matrix through oriented transfer function method according to pretreated brain wave;

D:, calculate brain asymmetric index based on the information flow gain according to said brain function network connection matrix.

2. the method for claim 1 is characterized in that, said steps A specifically comprises step:

A1: the full brain position person to be measured is provided with test electrode, in person's to be measured mastoid process position reference electrode is set, ground-electrode ground connection;

A2: to said test electrode, reference electrode and ground-electrode, send the parallel port synchronizing signal through computing machine;

A3: the brain wave that detects person to be measured through said test electrode, reference electrode and ground-electrode;

A4: with sending to said computing machine after amplification of said brain wave process and the analog to digital conversion.

3. method as claimed in claim 2 is characterized in that, said step B specifically comprises step:

B1: said brain wave is carried out Filtering Processing;

B2: go baseline to handle to the brain wave after the Filtering Processing;

B3: go the electric artefact of eye to handle to removing the brain wave after baseline is handled;

B4: the moment alignment of the corresponding said parallel port synchronizing signal of the each brain electric potential in the brain wave that goes after the electric artefact of eye is handled.

4. method as claimed in claim 3 is characterized in that, said step C specifically comprises step:

C1: from pretreated brain wave, choose frequency band to be calculated;

C2: use red pond quantity of information criterion to calculate the order that said pretreated brain wave is carried out match, use the brain wave data in the said frequency band of multivariate regression model match according to said order;

C3:,, calculate the corresponding brain function network connection matrix of each discrete frequency in the said frequency band respectively according to the brain wave data in the said frequency band after the match through oriented transfer function method;

C4: use the alternate data method successively the corresponding brain function network connection matrix of each discrete frequency in the said frequency band to be connected the conspicuousness checking, remove because the meaningless connection that randomness causes.

5. method as claimed in claim 4 is characterized in that, said step D specifically comprises step:

D1: flow out the direction summation with the stream of the behavioural information in the said brain function network connection matrix and obtain flowing out the value of information, obtain flowing into the value of information with the information flow inflow direction summation of classifying as in the said brain function network connection matrix;

D2: with the middle line of test electrode longitudinally of person's to be measured brain is axle, and half brain about being divided into calculates the information flow gain of left half each test electrode of brain and the information flow gain of right half each test electrode of brain respectively; Said information flow gain is the outflow value of information on the corresponding test electrode and the ratio that flows into the value of information;

D3: according to the left-right symmetric principle, according to about the corresponding right information flow gain of test electrode, utilize the asymmetric index computing formula, calculate the brain asymmetric index of each test electrode successively to correspondence.

Images