CN102488518B

CN102488518B - Electroencephalogram detection method and device by utilizing fluctuation index and training for promotion

Info

Publication number: CN102488518B
Application number: CN 201110416444
Authority: CN
Inventors: 周卫东; 陈爽爽
Original assignee: Shandong University
Current assignee: Shandong University
Priority date: 2011-12-14
Filing date: 2011-12-14
Publication date: 2013-09-04
Anticipated expiration: 2031-12-14
Also published as: CN102488518A

Abstract

The invention relates to an electroencephalogram detection method and an electroencephalogram detection device by utilizing a fluctuation index and training for promotion. Electroencephalogram data which is acquired and processed is subjected to feature extraction by using the fluctuation index with a good feature effect, and the extracted feature vector is input into a classifier obtained through a method of training for promotion, so that marks of abnormal electroencephalogram signals are obtained. Therefore, the workload of clinicians for judging large-scale electroencephalogram data is relieved, and the timeliness of abnormal electroencephalogram detection is improved.

Description

A method and device for EEG detection using volatility index and lifting training

技术领域 technical field

本发明公开了一种利用波动指数和提升训练的脑电检测方法和装置，属于脑电检测技术领域。The invention discloses an EEG detection method and device using fluctuation index and lifting training, and belongs to the technical field of EEG detection.

技术背景 technical background

癫痫是一种以脑部神经元反复突然过度放电所致的间歇性中枢神经系统功能失调为特征的脑部疾患。目前为止，癫痫检测主要是医务工作者依靠经验对脑电图(EEG)进行目测来完成，查看EEG中是否含有癫痫样放电等特征波，其工作量大，容易造成医务工作者疲倦而产生误判。因此，在癫痫检测中，自动检测系统检测脑电的准确性有着越重要的地位，它可极大提高对EEG的检测效率。Epilepsy is a brain disorder characterized by intermittent central nervous system dysfunction caused by repeated and sudden excessive discharge of brain neurons. So far, epilepsy detection is mainly done by medical workers relying on experience to visually inspect the electroencephalogram (EEG) to check whether there are characteristic waves such as epileptiform discharges in the EEG. sentenced. Therefore, in the detection of epilepsy, the accuracy of the automatic detection system to detect EEG has a more important position, which can greatly improve the detection efficiency of EEG.

自上世纪六十年代起，自动癫痫检测技术就受到了广泛的关注，这一领域的众多学者提出了多种自动检测脑电的方法。其主流算法有支持向量机(SVM)和神经网络等。而支持向量机是借助二次规划来求解支持向量，求解二次规划将涉及高阶矩阵的计算，矩阵的存储和计算将耗费大量的机器内存和运算时间。CN1253762A(99124032.4)所公开的一种全自动定量检测脑电图中癫痫样放电的装置采用了神经网络及前馈逆传播(BP)学习算法。神经网络必须进行多次重复学习，训练速度缓慢，计算效率低。同时由于BP算法是一种局部搜索的优化算法，用它来解决复杂非线性函数的全局极值，很有可能陷入局部极值从而导致训练失败。Since the 1960s, automatic epilepsy detection technology has received widespread attention, and many scholars in this field have proposed a variety of methods for automatic detection of EEG. Its mainstream algorithms include support vector machine (SVM) and neural network. The support vector machine uses quadratic programming to solve the support vector, and solving the quadratic programming will involve the calculation of high-order matrices, and the storage and calculation of the matrix will consume a lot of machine memory and computing time. CN1253762A (99124032.4) discloses a device for fully automatic quantitative detection of epileptiform discharges in electroencephalograms, which uses a neural network and a feedforward backpropagation (BP) learning algorithm. The neural network must learn many times, the training speed is slow, and the calculation efficiency is low. At the same time, since the BP algorithm is an optimization algorithm for local search, if it is used to solve the global extremum of complex nonlinear functions, it is very likely to fall into the local extremum and lead to training failure.

发明内容 Contents of the invention

针对现有技术的不足，本发明提出了一种利用波动指数和提升训练的脑电检测方法，该方法将提取到的脑电信号波动指数作为输入参数，送入由提升训练的分类器进行计算，获得输出概率值，将输出概率值与预设阈值进行比较，从而获得脑电检测结果。Aiming at the deficiencies in the prior art, the present invention proposes an EEG detection method using fluctuation index and lifting training, which uses the extracted EEG signal fluctuation index as an input parameter, and sends it to a classifier trained by lifting for calculation , to obtain an output probability value, and compare the output probability value with a preset threshold to obtain an EEG detection result.

本发明还提供一种利用上述方法进行检测脑电的装置。The present invention also provides a device for detecting brain electricity by using the above method.

本发明的技术方案如下：Technical scheme of the present invention is as follows:

一种利用波动指数和提升训练的脑电检测方法，步骤如下：An EEG detection method using volatility index and lifting training, the steps are as follows:

1)利用脑电放大器和数据采集卡采集脑电信号，将采集到的脑电信号通过A/D转换，存储到计算机中；1) Utilize the EEG amplifier and the data acquisition card to collect EEG signals, and store the EEG signals collected in the computer through A/D conversion;

2)计算机对脑电信号进行滤波和去噪；2) The computer filters and denoises the EEG signal;

3)计算机提取脑电信号各通道各小波层的波动指数；3) The computer extracts the fluctuation index of each wavelet layer of each channel of the EEG signal;

4)将步骤3)提取到的波动指数输入分类器进行计算，得到输出概率值；4) input the volatility index extracted in step 3) into the classifier for calculation, and obtain the output probability value;

5)将输出概率值与预设阈值进行比较，获得脑电检测结果并标记：5) Compare the output probability value with the preset threshold, obtain the EEG detection result and mark it:

输出概率值大于预设阈值，则判断检测脑电为异常，标记为1；If the output probability value is greater than the preset threshold, it is judged that the detected EEG is abnormal and marked as 1;

输出概率值小于或等于预设阈值，则判断检测脑电为正常，标记为-1。If the output probability value is less than or equal to the preset threshold, it is judged that the detected EEG is normal, and it is marked as -1.

优选的，所述的预设阈值为0.5。Preferably, the preset threshold is 0.5.

步骤1)中所述的脑电放大器为Neurofile NT脑电放大器，所述的数据采集卡为16位A/D转换数据采集卡，采样频率为256Hz。The EEG amplifier described in step 1) is a Neurofile NT EEG amplifier, and the described data acquisition card is a 16-bit A/D conversion data acquisition card, and the sampling frequency is 256Hz.

步骤2)中所述的计算机对脑电信号进行滤波和去噪的方法步骤如下：The computer described in step 2) filters and denoises the method steps of EEG signal as follows:

采集一段长度为LEN的脑电信号，利用Daubechies-4小波进行S层小波分解，优选S＝5；随后对分解后的脑电信号进行信号重构，提取重构信号的3-30Hz频段，即第3、4、5层重构信号a_j，n，a_j，n代表长度为LEN的脑电信号第j通道信号x_j的第n层小波重构信号，其中j＝1，2，...，C，n＝3，4，5；C是通道数，优选C＝6。Collect an EEG signal with a length of LEN, use Daubechies-4 wavelet to decompose the S-level wavelet, preferably S=5; then perform signal reconstruction on the decomposed EEG signal, and extract the 3-30Hz frequency band of the reconstructed signal, namely The 3rd, 4th, 5th layer reconstruction signal a _{j, n} , a _{j, n} represents the nth layer wavelet reconstruction signal of the jth channel signal x _j of the EEG signal whose length is LEN, where j=1, 2,. . . . , C, n=3, 4, 5; C is the number of channels, preferably C=6.

优选LEN＝1024。Preferably LEN=1024.

步骤3)中所述的提取脑电信号各通道各小波层的波动指数的方法为：Step 3) described in extracting the method for the fluctuation index of each wavelet layer of each channel of EEG signal is:

利用公式(1)计算步骤2)脑电信号中第j通道信号x_j的第n层小波重构信号的波动指数wav_j，n为Use formula (1) to calculate step 2) the wavelet index wav _j of the nth layer wavelet reconstruction signal of the jth channel signal x _j in the EEG signal, n is

${wav wav}_{j j,, n no} = = {Σ Σ}_{l l = = 11}^{LEN LEN - - 11} | | {a a}_{j j,, n no} ((l l + + 11)) - - {a a}_{j j,, n no} ((l l)) | | - - - - - - ((11)) . .$

步骤4)中所述的通过分类器计算输出概率值的方法为：The method for calculating the output probability value by the classifier described in step 4) is:

将步骤3)中的波动指数wav_j，n作为特征向量w送入分类器F，利用公式(2)Send the volatility index wav _j,n in step 3) as a feature vector w into the classifier F, using the formula (2)

$p p ((y the y = = 11 | | w w)) = = \frac{{e e}^{F f ((w w))}}{{e e}^{F f ((w w))} + + {e e}^{- - F f ((w w))}} - - - - - - ((22))$

得到长度LEN的脑电信号为异常脑电的概率P。The probability P that the EEG signal of length LEN is abnormal is obtained.

步骤4)中所述的分类器是按以下提升训练方法获得，具体实现步骤为：The classifier described in step 4) is obtained by the following promotion training method, and the specific implementation steps are:

a)W为分类器训练所用分段的脑电数据，W＝{w_i∈R^k，i＝1，2，L，N}，其中K＝C×S，C是EEG的通道数，S是小波层数，N为数据段数，每段长度为LEN；Y为对应标记量，Y＝{y_i∈{-1，1}，i＝1，2，L，N}，标记量为-1表示正常脑电，标记量为1表示异常脑电；w_i为第i段脑电信号各通道第3、4、5层小波重构信号的波动指数wav_j，n组成的特征向量；F_m表示m步后建立的分类器；设定迭代次数为M；设定第i段脑电信号特征向量w_i属于异常脑电的初始概率为p₀(y_i＝1|w_i)＝0.5，i＝1，2，L，N；设定第i段脑电信号特征向量w_i的初始分类器为F₀(w_i)＝0，i＝1，2，L，N；N＝900，M＝180。a) W is the segmented EEG data used for classifier training, W={w _i ∈ R ^k , i=1, 2, L, N}, where K=C×S, C is the number of EEG channels, S is the number of wavelet layers, N is the number of data segments, and the length of each segment is LEN; Y is the corresponding label quantity, Y={y _i ∈{-1,1}, i=1,2,L,N}, the label quantity is - 1 means normal EEG, and 1 means abnormal EEG; w _i is the wavelet index wav _{j of the 3rd, 4th, and 5th layer wavelet reconstruction signals of each channel of the i-th segment of the EEG signal, and the eigenvector composed of n} ; F _m represents the classifier established after m steps; set the number of iterations as M; set the initial probability that the i-th EEG signal feature vector w _i belongs to abnormal EEG is p ₀ (y _i =1|w _i )=0.5 , i=1, 2, L, N; set the initial classifier of the i-th EEG signal feature vector w _i as F ₀ (w _i )=0, i=1, 2, L, N; N=900 , M=180.

b)m表示迭代步数，从m＝1开始进行以下循环迭代：b) m represents the number of iteration steps, and the following loop iterations are performed from m=1:

i.计算分类器F_m似然函数的一阶导数

i. Calculate the first derivative of the likelihood function of the classifier F _m

其中p_m-1(y_i＝1|w_i)表示第m-1步迭代后特征向量w_i属于异常脑电的概率值；Where p _m-1 (y _i =1|w _i ) represents the probability value that the feature vector w _i belongs to abnormal EEG after the m-1th iteration;

ii.通过最小二乘法由w_i对

拟合得到回归系数r，用f(w_i)表示第i段脑电信号特征向量w_i的弱分类器：ii. By the method of least squares from w _i to

The regression coefficient r is obtained by fitting, and f(w _i ) is used to represent the weak classifier of the i-th EEG signal feature vector w _i :

f(w_i)＝r^Tw_i，i＝1，2，L，N；f(w _i )=r ^T w _i , i=1, 2, L, N;

iii.得到m次迭代后选用的弱分类器f_m iii. Get the weak classifier f _m selected after m iterations

iv.通过训练数据推出的伯努利回归函数L(F_m；W，Y)可以表示为：iv. The Bernoulli regression function L(F _m ; W, Y) derived from the training data can be expressed as:

$L L (({F f}_{m m};; W W,, Y Y)) = = log log (({Π Π}_{i i = = 11}^{N N} {p p}_{m m} {(({y the y}_{i i} = = 11 | | {w w}_{i i}))}^{{y the y}_{i i}} {p p}_{m m} {(({y the y}_{i i} = = 00 | | {w w}_{i i}))}^{11 - - {y the y}_{i i}})),, i i = = 1,2 1,2,, L L,, N N;;$

v.计算第m步后弱分类器加权系数γ_m为v. Calculate the weighting coefficient γ _m of the weak classifier after the mth step as

${γ γ}_{m m} = = arg arg \underset{γ γ}{max max} L L (({F f}_{m m - - 11} + + γ γ {f f}_{m m};; W W,, Y Y));;$

vi.更新分类器vi. Update the classifier

F_m＝F_m-1+εγ_mf_m F _m ＝F _m-1 +εγ _m f _m

其中ε为一个极小的值，ε＝0.05；Where ε is a very small value, ε=0.05;

vii.由分类器F_m计算特征向量w_i属于异常脑电的概率值：vii. Calculate the probability value that the feature vector w _i belongs to abnormal EEG by the classifier F _m :

${p p}_{m m} (({y the y}_{i i} = = 11 | | {w w}_{i i})) = = \frac{{e e}^{{F f}_{m m} (({w w}_{i i}))}}{{e e}^{{F f}_{m m} (({w w}_{i i}))} + + {e e}^{- - {F f}_{m m} (({w w}_{i i}))}},, i i = = 1,2 1,2,, L L,, N N$

其中，F_m(w_i)表示m步后对应训练数据w_i的分类器；Among them, F _m (w _i ) represents the classifier corresponding to the training data w _i after m steps;

viii.令m＝m+1，重复进行上述循环，如果m＝M则循环迭代结束，得到分类器F＝F_M。viii. Let m=m+1, repeat the above cycle, if m=M, the cycle iteration ends, and the classifier F=F _M is obtained.

一种利用上述方法进行脑电检测的装置，包括以电路连接的脑电放大器、数据采集卡和计算机，所述计算机中内置有利用波动指数和提升训练方法检测脑电的脑电检测模块，利用脑电放大器和数据采集卡对脑电信号进行采集后传输到计算机中，利用波动指数和提升训练方法检测脑电的脑电检测模块对脑电信号进行滤波和去噪处理；提取每段脑电信号的波动指数作为特征向量；将特征向量送入以提升训练方法所获得的分类器中，获输出概率值；将输出概率值与预设阈值比较，得脑电检测结果并加以标记。A kind of device that utilizes above-mentioned method to carry out EEG detection, comprises the EEG amplifier connected with the circuit, data acquisition card and computer, and the EEG detection module that utilizes fluctuation index and lifting training method to detect EEG is built-in in the described computer, utilizes The EEG amplifier and the data acquisition card collect the EEG signal and transmit it to the computer, and use the fluctuation index and the improvement training method to detect the EEG EEG detection module to filter and denoise the EEG signal; extract each EEG signal The fluctuation index of the signal is used as the feature vector; the feature vector is sent to the classifier obtained by the improved training method to obtain the output probability value; the output probability value is compared with the preset threshold value to obtain the EEG detection result and mark it.

本发明的有益的效果是：The beneficial effects of the present invention are:

利用特征效果较好的波动指数对采集并经预处理后的脑电数据进行特征提取，将提取的特征向量送入由提升训练方法获得的分类器中，从而得到对异常脑电信号的标记，不但减轻了临床医生对大规模脑电数据进行判别的工作量，而且提高了对异常脑电检测的时效性。Use the fluctuation index with better characteristic effects to extract the features of the collected and preprocessed EEG data, and send the extracted feature vectors to the classifier obtained by the promotion training method, so as to obtain the marking of abnormal EEG signals, It not only reduces the workload of clinicians to discriminate large-scale EEG data, but also improves the timeliness of abnormal EEG detection.

附图说明 Description of drawings

图1为本发明的脑电检测方法的流程图；Fig. 1 is the flowchart of EEG detection method of the present invention;

图2为本发明的脑电检测装置的硬件连接图；Fig. 2 is a hardware connection diagram of the EEG detection device of the present invention;

图3为实施例1中所述脑电信号的波动指数，其中两竖线间为异常脑电持续时间；Fig. 3 is the fluctuation index of EEG signal described in embodiment 1, wherein between two vertical lines is abnormal EEG duration;

图4为图3所述脑电信号的分类结果：其中，1表示异常脑电，即癫痫样放电；-1表示正常脑电；两竖线间为异常脑电持续时间。FIG. 4 is the classification result of the EEG signal in FIG. 3 : 1 indicates abnormal EEG, that is, epileptiform discharge; -1 indicates normal EEG; and the duration of abnormal EEG is between two vertical lines.

具体实施方式 Detailed ways

下面结合附图与实例对本发明做进一步说明，显然本发明并不限于此。The present invention will be further described below in conjunction with the accompanying drawings and examples, obviously the present invention is not limited thereto.

实施例1、Embodiment 1,

如图1所示，一种利用波动指数和提升训练的脑电检测方法，步骤如下：As shown in Figure 1, an EEG detection method using volatility index and boost training, the steps are as follows:

1)利用Neurofile NT脑电放大器和16位A/D转换数据采集卡采集脑电信号，采样频率为256Hz，将采集到的脑电信号通过A/D转换，存储到计算机中。1) Use Neurofile NT EEG amplifier and 16-bit A/D conversion data acquisition card to collect EEG signals, the sampling frequency is 256Hz, and store the collected EEG signals in the computer through A/D conversion.

2)计算机对脑电信号进行滤波和去噪，其方法步骤如下：2) The computer filters and denoises the EEG signal, and the method steps are as follows:

采集一段长度为LEN＝1024的脑电信号，利用Daubechies-4小波进行S层小波分解，优选S＝5；随后对分解后的脑电信号进行信号重构，提取重构信号的3-30Hz频段，即第3、4、5层重构信号a_j，n，a_j，n代表长度为LEN的脑电信号第j通道信号x_j的第n层小波重构信号，其中j＝1，2，...，C，n＝3，4，5；C是通道数，C＝6。Collect an EEG signal with a length of LEN=1024, use Daubechies-4 wavelet to decompose S-level wavelet, preferably S=5; then reconstruct the decomposed EEG signal, and extract the 3-30Hz frequency band of the reconstructed signal , that is, the 3rd, 4th, and 5th layer reconstruction signals a _{j, n} , a _{j, n} represent the nth layer wavelet reconstruction signal of the jth channel signal x _j of the EEG signal whose length is LEN, where j=1, 2 ,..., C, n=3, 4, 5; C is the number of channels, C=6.

3)计算机提取脑电信号各通道各小波层的波动指数的方法为：3) The method for the computer to extract the fluctuation index of each channel and each wavelet layer of the EEG signal is:

${wav wav}_{j j,, n no} = = {Σ Σ}_{l l = = 11}^{LEN LEN - - 11} | | {a a}_{j j,, n no} ((l l + + 11)) - - {a a}_{j j,, n no} ((l l)) | | - - - - - - ((11));;$

图3中所示为脑电信号第1通道第3层小波重构信号的波动指数。Fig. 3 shows the fluctuation index of the wavelet reconstructed signal of the first channel and the third layer of the EEG signal.

所述的通过分类器计算输出概率值的方法为：The method for calculating the output probability value by the classifier is:

得到长度LEN的脑电信号为异常脑电的概率P；The probability P that the EEG signal of length LEN is abnormal is obtained;

所述的分类器是按以下提升训练方法获得，具体实现步骤为：The classifier is obtained by the following promotion training method, and the specific implementation steps are:

a)W为分类器训练所用分段的脑电数据，W＝{w_i∈R^k，i＝1，2，L，N}，其中K＝C×S，C是EEG的通道数，S是小波层数，N为数据段数，每段长度为LEN；Y为对应标记量，Y＝{y_i∈{-1，1}，i＝1，2，L，N}，标记量为-1表示正常脑电，标记量为1表示异常脑电；w_i为第i段脑电信号各通道第3、4、5层小波重构信号的波动指数wav_j，n组成的特征向量；F_m表示m步后建立的分类器；设定迭代次数为M；设定第i段脑电信号特征向量w_i属于异常脑电的初始概率为p₀(y_i＝1|w_i)＝0.5，i＝1，2，L，N；设定第i段脑电信号特征向量w_i的初始分类器为F₀(w_i)＝0，i＝1，2，L，N；N＝900，M＝180；a) W is the segmented EEG data used for classifier training, W={w _i ∈ R ^k , i=1, 2, L, N}, where K=C×S, C is the number of EEG channels, S is the number of wavelet layers, N is the number of data segments, and the length of each segment is LEN; Y is the corresponding label quantity, Y={y _i ∈{-1,1}, i=1,2,L,N}, the label quantity is - 1 means normal EEG, and 1 means abnormal EEG; w _i is the wavelet index wav _{j of the 3rd, 4th, and 5th layer wavelet reconstruction signals of each channel of the i-th segment of the EEG signal, and the eigenvector composed of n} ; F _m represents the classifier established after m steps; set the number of iterations as M; set the initial probability that the i-th EEG signal feature vector w _i belongs to abnormal EEG is p ₀ (y _i =1|w _i )=0.5 , i=1, 2, L, N; set the initial classifier of the i-th EEG signal feature vector w _i as F ₀ (w _i )=0, i=1, 2, L, N; N=900 , M=180;

i.计算分类器F_m似然函数的一阶导数

ii.通过最小二乘法由w_i对

f(w_i)＝r^Tw_i，i＝1，2，L，N；f(w _i )=r ^T w _i , i=1, 2, L, N;

vi.更新分类器vi. Update the classifier

F_m＝F_m-1+εγ_mf_m F _m ＝F _m-1 +εγ _m f _m

输出概率值小于或等于预设阈值，则判断检测脑电为正常，标记为-1；所述的预设阈值为0.5。If the output probability value is less than or equal to the preset threshold, it is judged that the detected EEG is normal and marked as -1; the preset threshold is 0.5.

图4为图3所述脑电信号的标记结果。FIG. 4 is the labeling result of the EEG signal described in FIG. 3 .

实施例2、Embodiment 2,

一种利用实施例1所述方法进行脑电检测的装置，包括以电路连接的脑电放大器、数据采集卡和计算机，所述计算机中内置有利用波动指数和提升训练方法检测脑电的脑电检测模块，利用脑电放大器和数据采集卡对脑电信号进行采集后传输到计算机中，利用波动指数和提升训练方法检测脑电的脑电检测模块对脑电信号进行滤波和去噪处理；提取每段脑电信号的波动指数作为特征向量；将特征向量送入以提升训练方法所获得的分类器中，获输出概率值；将输出概率值与预设阈值比较，得脑电检测结果并加以标记。A device utilizing the method described in Embodiment 1 for EEG detection, comprising an EEG amplifier connected with a circuit, a data acquisition card and a computer, the computer is built-in an EEG that utilizes fluctuation index and lifting training method to detect EEG The detection module uses the EEG amplifier and data acquisition card to collect the EEG signal and transmits it to the computer, and uses the fluctuation index and the improvement training method to detect the EEG. The EEG detection module performs filtering and denoising processing on the EEG signal; The fluctuation index of each EEG signal is used as the feature vector; the feature vector is sent to the classifier obtained by the improved training method, and the output probability value is obtained; the output probability value is compared with the preset threshold value, and the EEG detection result is obtained. mark.

利用本发明对21例癫痫患者的脑电进行检测，对癫痫样异常放电检测的准确率达94％，每小时错误检测次数为0.2次。Using the invention to detect the EEG of 21 cases of epilepsy patients, the accuracy rate of detecting epileptiform abnormal discharge reaches 94%, and the number of wrong detections per hour is 0.2 times.

Claims

1. a brain electro-detection method of utilizing the index of oscillation and training for promotion is carried out the device of brain electro-detection, it is characterized in that, described device comprises eeg amplifier, data collecting card and the computer that connects with circuit, be built-in with in the described computer and utilize the index of oscillation and training for promotion method to detect the brain electro-detection module of brain electricity, be transferred in the computer after utilizing eeg amplifier and data collecting card that EEG signals is gathered, the brain electro-detection module of utilizing the index of oscillation and training for promotion method to detect the brain electricity is carried out filtering and denoising to EEG signals; Extract the index of oscillation of every section EEG signals as characteristic vector; Characteristic vector is sent in the grader that is obtained with the training for promotion method, obtained the output probability value; With output probability value and predetermined threshold value relatively, get brain electro-detection result and labelling in addition;

Wherein said eeg amplifier and the data collecting card of utilizing gathered EEG signals, and the EEG signals that collects is changed by A/D, stores in the computer; Described eeg amplifier is Neurofile NT eeg amplifier, and described data collecting card is 16 A/D translation data capture cards, and sample frequency is 256Hz;

The brain electro-detection module of the wherein said index of oscillation and training for promotion method detection brain electricity is as follows to the method step that EEG signals is carried out filtering and denoising:

Gather the EEG signals that a segment length is LEN, utilize the Daubechies-4 small echo to carry out S layer wavelet decomposition; Subsequently the EEG signals after decomposing is carried out signal reconstruction, extract the 3-30Hz frequency range of reconstruction signal, i.e. the 3rd, 4,5 layer of reconstruction signal a _{J, n}, a _{J, n}Representing length is the EEG signals j channel signal x of LEN _jN layer wavelet reconstruction signal, j=1 wherein, 2 ..., C, n=3,4,5; C is port number; Described port number C=6; Described LEN=1024;

The method that the brain electro-detection module of the described index of oscillation and training for promotion method detection brain electricity is extracted the index of oscillation of each each wavelet layer of passage of EEG signals is:

Utilize j channel signal x in formula (1) the calculation procedure EEG signals _jThe index of oscillation wav of n layer wavelet reconstruction signal _{J, n}For

{wav}_{j, n} = Σ_{l = 1}^{LEN - 1} | a_{j, n} (l + 1) - a_{j, n} (l) | - - - (1);

Described method by classifier calculated output probability value is:

With index of oscillation wav _{J, n}Send into grader F as characteristic vector w, utilize formula (2)

p (y = 1 | w) = \frac{e^{F (w)}}{e^{F (w)} + e^{- F (w)}} - - - (2)

The EEG signals that obtains length LEN is the probability P of unusual brain electricity;

Described grader is to obtain by following training for promotion method, and the specific implementation step is:

A) W is the eeg data that grader is trained used segmentation, W={w _i∈ R ^k, i=1,2 ..., N}, K=C * S wherein, C is the port number of EEG, and S is the small echo number of plies, and N is the data hop count, and every segment length is LEN; Y is the correspondence markings amount, Y={y _i∈ 1,1}, and i=1,2 ..., N}, labelled amount is-1 expression normal brain activity electricity, labelled amount is the unusual brain of 1 expression; w _iBe the index of oscillation wav of each passage the 3rd of i section EEG signals, 4,5 layers of wavelet reconstruction signal _{J, n}The characteristic vector of forming; F _mThe grader that expression m sets up the step back; The setting iterations is M; Set i section EEG signals characteristic vector w _iThe initial probability that belongs to unusual brain electricity is p ₀(y _i=1|w _i)=0.5, i=1,2 ..., N; Set i section EEG signals characteristic vector w _iThe preliminary classification device be F ₀(w _i)=0, i=1,2 ..., N; N=900, M=180;

B) m represents the iteration step number, begins to carry out following loop iteration from m=1:

I. calculate grader F _mThe first derivative of likelihood function

{\tilde{y}}_{i} = 2 (y_{i} - p_{m - 1} (y_{i} = 1 | w_{i})), i = 1,2, \cdot \cdot \cdot, N

P wherein _M-1(y _i=1|w _i) characteristic vector w after expression m-1 step iteration _iThe probit that belongs to unusual brain electricity;

Ii. by method of least square by w _iRight

Match obtains regression coefficient r, with f (w _i) expression i section EEG signals characteristic vector w _iWeak Classifier:

f(w _i)＝r ^Tw _i,i＝1,2,…,N；

Iii. obtain the Weak Classifier f that selects for use after m iteration _m

f_{m} = \arg \min_{f} Σ_{i = 1}^{N} {({\tilde{y}}_{i} - f (w_{i}))}^{2};

Iv. the Bernoulli Jacob's regression function L (F that releases by training data _m; W Y) can be expressed as:

L (F_{m}; W, Y) = \log (Π_{i = 1}^{N} p_{m} {(y_{i} = 1 | w_{i})}^{y_{i}} p_{m} {(y_{i} = 0 | w_{i})}^{1 - y_{i}}), i = 1,2, \cdot \cdot \cdot, N;

V. calculate m step back Weak Classifier weight coefficient γ _mFor

γ_{m} = \arg \max_{γ} L (F_{m - 1} + γ f_{m}; W, Y);

Vi. upgrade grader

F _m＝F _m-1+εγ _mf _m

Wherein ε is a minimum value, ε=0.05;

Vii. by grader F _mCalculated characteristics vector w _iThe probit that belongs to unusual brain electricity:

p_{m} (y_{i} = 1 | w_{i}) = \frac{e^{F_{m} (w_{i})}}{e^{F_{m} (w_{i})} + e^{- F_{m} (w_{i})}}, i = 1,2, \cdot \cdot \cdot, N

Wherein, F _m(w _i) the corresponding training data w in expression m step back _iGrader;

Viii. make m=m+1, repeat above-mentioned circulation, if m=M then loop iteration finish, obtain grader F=F _MDescribed predetermined threshold value is 0.5.

2. a kind of brain electro-detection method of utilizing the index of oscillation and training for promotion according to claim 1 is carried out the device of brain electro-detection, it is characterized in that, utilizes the Daubechies-4 small echo to carry out S layer wavelet decomposition, described S=5.