CN104622467A - Method for detecting electroencephalogram signal complexity abnormity of Alzheimer disease - Google Patents

Abstract

The invention discloses a method for detecting the electroencephalogram signal complexity abnormity of the Alzheimer disease. The method includes the following steps of collecting original electroencephalogram signals through a UEA-FZ electroencephalogram amplifier device, intercepting the stable electroencephalogram signals through an EEGLAB software packet on an MATLAB software platform, extracting sub-rhythms of the electroencephalogram signals through a signal processing toolbox of the MATLAB software platform, conducting coarse graining on the electroencephalogram signals of each sub-rhythm, and calculating the multi-variable and multi-dimension weighted value permutation entropy of each sub-rhythm of the electroencephalogram signals. The method has the advantages that the complexity characteristic of the electroencephalogram signals can be effectively extracted, and the higher anti-noise performance is achieved compared with the prior art; meanwhile, the complexity characteristic of the electroencephalogram signals is detected at different moments and in different spatial dimensions, reliable data support is provided for medical diagnosis, and clinical guidance is provided for the diagnosis of the Alzheimer disease.

Description

A kind of method detecting the EEG signals complexity exception of alzheimer's disease

Technical field

The present invention relates to a kind of method of the detection alzheimer's disease exception based on EEG signals, particularly by calculating the multivariate multi-gauge strip weights ordering entropy of EEG signals, be used for a kind of method detecting the EEG signals complexity exception of alzheimer's disease of carrying out.

Background technology

Alzheimer's disease, is commonly called as senile dementia, is a kind of nervous system degenerative disease, is also modal dementia.The old people of its major effect over-65s, estimates be multiplied in its prevalence of future 50 years ^[1,2].But present stage does not cure the means of alzheimer's disease, and a large amount of medicines is merely able to the deterioration delaying this disease.Alzheimer disease is divided into four different stages.First stage is the mild cognitive impairment stage, is usually expressed as the loss of memory, but can not affect daily life significantly.Then being slight and the moderate alzheimer's disease stage, is finally severe alzheimer's disease ^[3], the even basic physiological activity of the memory of patient, perception, thinking is all badly damaged.

The medical diagnosis of alzheimer's disease is very difficult, and the performance of alzheimer's disease is easily considered to caused by naturally-aged usually.The diagnosis of medical science normally adopts multiple method to combine, and as will through psychology test, blood testing, spinal fluid be tested, and neurological detects, and imaging technique etc. are diagnosed.In the last few years, EEG signals is utilized to pay close attention to widely to diagnose the research of alzheimer's disease to cause.Due to brain wave acquisition equipment economy, portable, and noinvasive safety, EEG signals temporal resolution is very high, and therefore brain electricity has become a kind of means of potential diagnosis alzheimer's disease.

At present, research shows that the EEG signals of alzheimer's disease mainly contains the exception of three aspects compared with normal person: the slowing down of EEG signals, the reduction of complexity and the reduction of EEG signals synchronicity ^[9].Wherein, the detection method of complexity has a lot, cuts both ways, but still a kind of algorithm of not generally acknowledging can as the best approach of diagnosis alzheimer's disease complexity exception.Entropy is the algorithm of a class based on chaology, and the method as reflection EEG signals kinetic complexity characteristic is widely studied.Ordering entropy algorithm is a kind of method carrying out signal calculated complexity based on arrangement ordinal number, and its computational speed is fast, anti-interference, without the need to model.Ordering entropy has been used in the detection of epilepsy, anesthesia, Different Cognitive state and alzheimer's disease, and achieves good Detection results.

Consult list of references and include [1] Mattson M.Pathways towards and away from Alzheimer ' s disease.Nature vol.430 (2004); [2] Meek PD, McKeithan K, Shumock GT.Economics considerations of Alzheimer ' s disease.Pharmacotherapy 18 (2Pt 2): 68-73 (1998); [3] Shimokawa A, Yatomib N, Anamizuc S, Toriid S, Isonod H, Sugaid Y, et al.Influence of deteriorating ability of emotional comprehension on interpersonal behavior in Alzheimer-type dementia.Brain Cogn 47 (3): 423-433 (2001); [4] Small BJ, Gagnon E, Robinson B.Early identification of cognitive deficits:preclinical Alzheimer ' s disease and mild cognitive impairment.Geriatrics 62 (4): 19-23 (2007); [5] Palmer K, Berger AK, Monastero R, Winblad B, l, Fratiglioni L.Predictors of progression from mild cognitive impairment to Alzheimer disease.Neurology 68 (19): 1596-1602 (2007); [6] Arnaiz E, Almkvist O.Neuropsychological features of mild cognitive impairment and preclinical Alzheimer ' s disease.Acta Neurol Scand Suppl 179:34-41 (2003); [7] Weiner WM.Editorial:Imaging and Biomarkers Will be Usedfor Detection and Monitoring Progression of Early Alzheimer ' s Disease.J Nutr Health Aging 4:332 (2009); [8] Sunderland T, Hampel H, Takeda M, Putnam KT, Cohen RM.Biomarkers in the diagnosis of Alzheimer ' s disease:are we ready? J Geriatr Psychiatry Neurol 19 (3): 172-9 (2006); [9] Dauwels J, Vialatte F, Cichocki A, Diagnosis of Alzheimer ' s Disease from EEG Signals:Where Are We Standing? Current Alzheimer Research, 7,487-505 (2010).

But there is following three point problem in the detection based on ordering entropy: one be ordering entropy algorithm calculate be the complexity singly leading brain electricity, calculate process in, do not consider the dependency of leading between eeg data more; Two are, the calculating of ordering entropy is only limitted to single time scale, and the dynamic characteristic of Nonlinear Time Series is reflected in multiple time scale; Three are, EEG signals has the interference such as noise, and ordering entropy algorithm have ignored the impact of noise for EEG signals, and inaccurate complexity therefore can be caused to estimate.Therefore, for above problem, propose a kind of algorithm of multivariate multi-gauge strip weights ordering entropy, for the detection of alzheimer's disease EEG signals complexity exception.A kind of method being mainly used in the complexity exception detecting alzheimer's disease EEG signals is provided, for the medical diagnosis of alzheimer's disease provides objective basis, simultaneously also can as the potential method of the electroencephalogramsignal signal analyzing of other mental sickness.

Summary of the invention

For existing methodical deficiency, the object of this invention is to provide a kind of method detecting the EEG signals complexity exception of alzheimer's disease, be beneficial to from multiple brain district, the complexity abnormal conditions of multiple time scale angle inspection Alzheimer's eeg data, the accuracy of further raising electroencephalogramsignal signal analyzing, thus carry out the Data support that clinical diagnosis provides more favourable for doctor.

For achieving the above object, the technical solution used in the present invention is to provide a kind of method detecting the EEG signals complexity exception of alzheimer's disease, the method carries out collection and the analysis of EEG signals, and the method comprises the following steps:

1. UEA-FZ eeg amplifier device is utilized to gather original EEG signals, first UEA-FZ eeg amplifier device is connected with the USB of computer, the lead end meeting the brain electrode cap of international standard is inserted the contact conductor interface of UEA-FZ eeg amplifier device, brain electrode cap has correctly been worn on the scalp surface of measured, then the signal acquiring system of the UEA-FZ eeg amplifier device that described computer has been installed is run, to realize the Real-time Collection of original EEG signals, record and storage;

2. utilize the EEGLAB software kit of MATLAB software platform to intercept steady EEG signals, EEGLAB software kit is the software kit of the EEG Processing be widely used, have signal visual, intercept, pretreated several functions;

3. the Matlab DSPToolBox of MATLAB software platform is utilized to extract the sub-rhythm and pace of moving things of EEG signals, frequency partition according to EEG signals is delta rhythm (0-4Hz), theta rhythm (4-8Hz), alpha rhythm (8-12Hz) and beta response (12-30Hz), utilize the finite impulse response in MATLAB software platform (Finite Impulse Response) band filter to extract the different rhythm and pace of moving things of EEG signals;

4. coarse process is carried out to the EEG signals of each rhythm and pace of moving things, the given one EEG signals { x led _i, i=1,2 ..., N}, by the method for averaging under different scale, primary signal step 1. gathered is converted into the signal of coarse, the time series of each coarse $y_r^{\left(\mathrm{\ϵ}\right)}=\frac{1}{\mathrm{\ϵ}}\underset{i=(r-1)\mathrm{\ϵ}+1}{\overset{\mathrm{r\ϵ}}{\mathrm{\Σ}}}{x}_i,1\≤r\≤\frac{N}{\mathrm{\ϵ}},$ Primary signal is when ε=1;

5. finally calculate the multivariate multi-gauge strip weights ordering entropy of the sub-rhythm and pace of moving things of each EEG signals: for the eeg data section V intercepted capable N row, V represents the number of leading of brain electricity, the sampling number of EEG signals that what N represented is, wherein the signal of every a line is expressed as { x _i, i=1,2 ..., N}: first, is embedded in a m-dimensional space: X by the signal of every a line _i=[x _i, x _{i+ τ}..., x _{i+ (m-1) τ}], i=1,2 ..., N-(m-1) τ; Then, for given Arbitrary Digit i, sequence X _i=[x _i, x _{i+ τ}..., x _{i+ (m-1) τ}] in element arrange according to ascending order: x just sorts according to the order of corresponding j; Then, this vectorial X is calculated _i=[x _i, x _{i+ τ}..., x _{i+ (m-1) τ}] variance; Then, with [j ₁, j ₂..., j _m] symbol sebolic addressing occur number of times with variance be multiplied as Weighted Coefficients number of times, ask the Weighted Coefficients distribution probability p of this symbol sebolic addressing _w1, p _w2... p , so for singly leading time series { x _i, i=1,2 ..., the Weighted Coefficients ordering entropy of N} is just defined as then, for leading EEG signals, when calculating the distribution probability of each symbol sebolic addressing of each eeg data led, it is considered that [j more ₁, j ₂..., j _m] sequence accounts for the ratio of all symbol sebolic addressing Weighted Coefficients summations of all leading, and calculates the multivariate Weighted Coefficients sequence entropy under single time scale thus; Then, for the EEG signals of each yardstick after coarse calculate its multivariate Weighted Coefficients ordering entropy, draw multivariate multi-gauge strip weights sequence entropy thus.

Patient is in the entropy of healthy normal person of the same age in contrast, judges the brain electrical anomaly degree of patient, and the final entropy chart exporting multivariate multi-gauge strip weights sequence entropy and drafting, compares the value of normal person and patient, detects the brain electricity complexity abnormal problem of patient.

Effect of the present invention is the complexity diversity of EEG signals under zones of different, Different time scales that the method may be used for the different rhythm and pace of moving things detecting alzheimer's disease, be specially adapted to the research across brain district, for the diagnosis of alzheimer's disease provides objective basis.The multivariate multi-gauge strip weights ordering entropy of EEG signals can also be used for the analysis of brain cognitive activities, different physiological status and other mental sickness.

Brain is a complicated nonlinear system, and EEG signals is the reflection of brain cortex neural discharge activities.The inner synapse cell structural damage of Alzheimer's brain, can cause the ANOMALOUS VARIATIONS of EEG signals complexity.The method is conducive to extracting in a noisy environment more accurately, has the complexity of the multi-lead EEG signals of dependency.Compared with prior art, innovative point comprises the method:

(1) consider lead more EEG signals respectively lead between dependency, from the multivariable sequence entropy of angle calculation of all leading, reflection be complexity change spatially, be conducive to the analysis of signal complexity across brain district.

(2) calculate entropy from multiple time scale, be conducive to fully reflecting the complexity change be implied in time series.

(3) weaken the interference of not removable noise for EEG signals, improve the noise immunity of existing ordering entropy algorithm.

Accompanying drawing explanation

Fig. 1 is workflow schematic diagram of the present invention;

Fig. 2 is the flow chart calculating multivariate multi-gauge strip weights ordering entropy;

Fig. 3 is average and the standard deviation figure of the electric multivariate multi-gauge strip weights ordering entropy in territory, occipital region, top of alpha rhythm brain of the present invention.

Detailed description of the invention

By reference to the accompanying drawings a kind of method detecting the EEG signals complexity exception of alzheimer's disease of the present invention is illustrated.

The present invention is a kind of method detecting the EEG signals complexity exception of alzheimer's disease, and the method carries out collection and the analysis of EEG signals, and the method comprises the following steps:

1. UEA-FZ eeg amplifier device is utilized to gather original EEG signals, first UEA-FZ eeg amplifier device is connected with the USB of computer, the lead end meeting the brain electrode cap of international standard is inserted the contact conductor interface of UEA-FZ eeg amplifier device, brain electrode cap has correctly been worn on the scalp surface of measured, then the signal acquiring system of the UEA-FZ eeg amplifier device that described computer has been installed is run, to realize the Real-time Collection of original EEG signals, record and storage.

2. utilize the EEGLAB software kit of MATLAB software platform to intercept steady EEG signals, EEGLAB software kit is the software kit of the EEG Processing be widely used, have signal visual, intercept, pretreated several functions.

3. the Matlab DSPToolBox of MATLAB software platform is utilized to extract the sub-rhythm and pace of moving things of EEG signals, frequency partition according to EEG signals is delta rhythm (0-4Hz), theta rhythm (4-8Hz), alpha rhythm (8-12Hz) and beta response (12-30Hz), utilize the finite impulse response in MATLAB software platform (Finite Impulse Response) band filter to extract the different rhythm and pace of moving things of EEG signals;

4. coarse process is carried out to the EEG signals of each rhythm and pace of moving things, the given one EEG signals { x led _i, i=1,2 ..., N}, by the method for averaging under different scale, primary signal step 1. gathered is converted into the signal of coarse, the time series of each coarse $y_r^{\left(\mathrm{\ϵ}\right)}=\frac{1}{\mathrm{\ϵ}}\underset{i=(r-1)\mathrm{\ϵ}+1}{\overset{\mathrm{r\ϵ}}{\mathrm{\Σ}}}{x}_i,1\≤r\≤\frac{N}{\mathrm{\ϵ}},$ Primary signal is when ε=1.

5. finally calculate the multivariate multi-gauge strip weights ordering entropy of the sub-rhythm and pace of moving things of each EEG signals: for the eeg data section V intercepted capable N row, V represents the number of leading of brain electricity, the sampling number of EEG signals that what N represented is, wherein the signal of every a line is expressed as { x _i, i=1,2 ..., N}: first, is embedded in a m-dimensional space: X by the signal of every a line _i=[x _i, x _{i+ τ}..., x _{i+ (m-1) τ}], i=1,2 ..., N-(m-1) τ; Then, for given Arbitrary Digit i, sequence X _i=[x _i, x _{i+ τ}..., x _{i+ (m-1) τ}] in element arrange according to ascending order: x just sorts according to the order of corresponding j; Then, this vectorial X is calculated _i=[x _i, x _{i+ τ}..., x _{i+ (m-1) τ}] variance; Then, with [j ₁, j ₂..., j _m] symbol sebolic addressing occur number of times with variance be multiplied as Weighted Coefficients number of times, ask the Weighted Coefficients distribution probability p of this symbol sebolic addressing _w1, p _w2... p , so for singly leading time series { x _i, i=1,2 ..., the Weighted Coefficients ordering entropy of N} is just defined as then, for leading EEG signals, when calculating the distribution probability of each symbol sebolic addressing of each eeg data led, it is considered that [j more ₁, j ₂..., j _m] sequence accounts for the ratio of all symbol sebolic addressing Weighted Coefficients summations of all leading, and calculates the multivariate Weighted Coefficients sequence entropy under single time scale thus; Then, for the EEG signals of each yardstick after coarse calculate its multivariate Weighted Coefficients ordering entropy, draw multivariate multi-gauge strip weights sequence entropy thus.

Patient is in the entropy of healthy normal person of the same age in contrast, judges the brain electrical anomaly degree of patient, and the final entropy chart exporting multivariate multi-gauge strip weights sequence entropy and drafting, compares the value of normal person and patient, detects the brain electricity complexity abnormal problem of patient.

A kind of methodological function detecting the EEG signals complexity exception of alzheimer's disease of the present invention is achieved in that

First be gather original eeg data.This example has carried out the collection of EEG signals to 14 alzheimer's disease patients and 14 normal healthy peoples of the same age.What adopt is the commercially available known products UEA-FZ eeg amplifier device of Beijing Xin Tuo instrument company, washes hair before measuring and dries up.UEA-FZ eeg amplifier device is utilized to gather original EEG signals, first device is connected with the USB of computer, the contact conductor interface of the lead end insertion apparatus of the brain electrode cap of international standard will be met, brain electrode cap has correctly been worn on the scalp surface of measured, then the signal acquiring system of the UEA-FZ eeg amplifier device that computer has been installed is run, to realize the Real-time Collection of original EEG signals, record and storage; A part of EEG signals for intercepting on the left of Fig. 2.The EEGLAB software kit of MATLAB software platform is utilized to intercept steady EEG signals, it is commercially available known products that MATLAB software platform will be pointed out, EEGLAB software kit is the software kit of the EEG Processing be widely used, and has the several functions such as signal is visual, intercepting, pretreatment.

Then, FIR filter is utilized to extract the sub-rhythm and pace of moving things to the EEG signals comparatively stably intercepted, realize for delta rhythm (0-4Hz), theta rhythm (4-8Hz), the extraction of alpha rhythm (8-12Hz) and beta response (12-30Hz).

Antithetical phrase circadian signal carries out coarse, and the eeg data of each being led utilizes the method for getting average at multiple yardstick, obtains the EEG signals under different scale.The given one EEG signals { x led _i, i=1,2 ..., N}, by the method for averaging under different scale, is converted into the signal of coarse by primary signal.The time series of each coarse primary signal is when ε=1.

For under different scale, the EEG signals after coarse calculates multivariate multi-gauge strip weights sequence entropy.For the eeg data section intercepted (be V capable N row, V represents the number of leading of brain electricity, the sampling number of EEG signals that what N represented is), wherein the signal of every a line is expressed as { x _i, i=1,2 ..., N}.First, the signal of every a line is embedded in a m-dimensional space: X _i=[x _i, x _{i+ τ}..., x _{i+ (m-1) τ}], i=1,2 ..., N-(m-1) τ.For given Arbitrary Digit i, sequence X _i=[x _i, x _{i+ τ}..., x _{i+ (m-1) τ}] in element arrange according to ascending order: x just sorts according to the order of corresponding j, an arbitrary vectorial X _i=[x _i, x _{i+ τ}..., x _{i+ (m-1) τ}] (j can be mapped as uniquely ₁, j ₂... j _m).The variance of compute vector.For the concrete symbol of m (1,2 ... m) be m! Plant the one in arrangement mode.

When calculating the distribution probability of often kind of arrangement mode, utilize (j ₁, j ₂... j _m) number of times that occurs is with corresponding variance be multiplied as Weighted Coefficients number of times, ask this symbol sebolic addressing (j ₁, j ₂... j _m) Weighted Coefficients distribution probability p _w1, p _w2... p , wherein U=N-(m-1) τ, $\mathrm{\Π}={\left\{{\mathrm{\π}}_j\right\}}_{j=1}^{m!},1_{A\left(v\right)}=\left\{\begin{array}{cc}1,& v\∈A\\ 0,& v\∉A\end{array}\right..$ Weights are calculated by the variance of vector so for singly leading time series { x _i, i=1,2 ..., the Weighted Coefficients ordering entropy of N} is just defined as for leading EEG signals more, when calculating the distribution probability of each symbol sebolic addressing of each eeg data led, it is considered that this sequence accounts for the ratio of all symbol sebolic addressing Weighted Coefficients summations of all leading, calculate the multivariate Weighted Coefficients sequence entropy under single time scale thus.What Fig. 2 represented is the flow process calculating multivariate multi-gauge strip weights ordering entropy, and the EEG signals for each yardstick of other after coarse calculates its multivariate Weighted Coefficients ordering entropy, draws multivariate multi-gauge strip weights sequence entropy thus.

Multivariate multi-gauge strip weights ordering entropy is subject to the impact of Embedded dimensions m, if m is too small, then can not reflecting time sequence well, in principle, data length be at least greater than m! , therefore m selects excessive, can exceed data length, strengthen computation time simultaneously.Suggestion m selects 3 ..., the integer between 7 analyzes eeg data.Delay time T is fixed as 1, and yardstick ε can get 1 ..., the integer between 10.

Fig. 3 is average and the standard deviation figure of the electric multivariate multi-gauge strip weights ordering entropy in territory, occipital region, top of alpha rhythm brain of the present invention.In figure, abscissa is time scale, and vertical coordinate is entropy, and red line represents alzheimer patient data, and blue line represents data of normal people.This figure is that the alpha rhythm brain electricity that alzheimer's disease group and healthy normal person of the same age organize is pushing up multivariate multi-gauge strip weights ordering entropy average and the analysis of variance diagram in territory, occipital region, and yardstick is 1-9.As seen from the figure, the complexity of normal brain electricity all higher than Alzheimer's, can reflect the exception of patient's brain electricity under each yardstick of territory, occipital region, top.

According to the research of existing brain neurophysiology generation mechanism, brain is a complicated nonlinear system, EEG signals can reflect the dynamics of this nonlinear system, the physiological brain structure generation pathological changes of Alzheimer's, cause its function impaired, be reflected in brain electricity complexity and reduce in this characteristic.Utilize multivariate multi-gauge strip weights ordering entropy can reflect the complexity abnormal conditions of alzheimer's disease brain electricity in Different brain region better, for the physiological diagnosis of alzheimer's disease provides foundation.

Be described in detail of the present invention by reference to the accompanying drawings above, obvious the present invention is not limited to this, and the various forms of changes carried out within the scope of the present invention all do not exceed protection scope of the present invention.

Claims (1)

1. detect a method for the EEG signals complexity exception of alzheimer's disease, the method carries out collection and the analysis of EEG signals, and the method comprises the following steps:

1. UEA-FZ eeg amplifier device is utilized to gather original EEG signals, first UEA-FZ eeg amplifier device is connected with the USB of computer, the lead end meeting the brain electrode cap of international standard is inserted the contact conductor interface of UEA-FZ eeg amplifier device, brain electrode cap has correctly been worn on the scalp surface of measured, then the signal acquiring system of the UEA-FZ eeg amplifier device that described computer has been installed is run, to realize the Real-time Collection of original EEG signals, record and storage;

2. utilize the EEGLAB software kit of MATLAB software platform to intercept steady EEG signals, EEGLAB software kit is the software kit of the EEG Processing be widely used, have signal visual, intercept, pretreatment several functions;

3. the Matlab DSPToolBox of MATLAB software platform is utilized to extract the sub-rhythm and pace of moving things of EEG signals, frequency partition according to EEG signals is delta rhythm (0-4Hz), theta rhythm (4-8Hz), alpha rhythm (8-12Hz) and beta response (12-30Hz), utilize the finite impulse response in MATLAB software platform (Finite Impulse Response) band filter to extract the different rhythm and pace of moving things of EEG signals;

4. coarse process is carried out to the EEG signals of each rhythm and pace of moving things, the given one EEG signals { x led _i, i=1,2 ..., N}, by the method for averaging under different scale, the EEG signals after step 2. being intercepted is converted into the signal of coarse, the time series of each coarse when ε=1, be the primary signal treating coarse;

5. the multivariate multi-gauge strip weights ordering entropy of the sub-rhythm and pace of moving things of each EEG signals is calculated: for the capable N row of the eeg data section V intercepted, V represents the number of leading of brain electricity, and N represents the sampling number of EEG signals, and wherein the signal of every a line is expressed as { x _i, i=1,2 ..., N}: first, is embedded in a m-dimensional space: X by the signal of every a line _i=[x _i, x _{i+ τ}..., x _{i+ (m-1) τ}], i=1,2 ..., N-(m-1) τ; Then, for given Arbitrary Digit i, sequence X _i=[x _i, x _{i+ τ}..., x _{i+ (m-1) τ}] in element arrange according to ascending order: x just sorts according to the order of corresponding j; Then, this vectorial X is calculated _i=[x _i, x _{i+ τ}..., x _{i+ (m-1) τ}] variance; Then, with [j ₁, j ₂..., j _m] symbol sebolic addressing occur number of times with variance be multiplied as Weighted Coefficients number of times, ask the Weighted Coefficients distribution probability p of this symbol sebolic addressing _w1, p _w2... p , so for singly leading time series { x _i, i=1,2 ..., the Weighted Coefficients ordering entropy of N} is just defined as then, for leading EEG signals, when calculating the distribution probability of each symbol sebolic addressing of each eeg data led, it is considered that [j more ₁, j ₂..., j _m] sequence accounts for the ratio of all symbol sebolic addressing Weighted Coefficients summations of all leading, and calculates the multivariate Weighted Coefficients sequence entropy under single time scale thus; Then, for the EEG signals of each yardstick after coarse calculate its multivariate Weighted Coefficients ordering entropy, draw multivariate multi-gauge strip weights sequence entropy thus;

The entropy of contrast patient and healthy normal person of the same age, judges the brain electrical anomaly degree of patient, and the final entropy chart exporting multivariate multi-gauge strip weights sequence entropy and drafting, compares the value of normal person and patient, detects the brain electricity complexity abnormal problem of patient.