CN113907768A - Electroencephalogram signal processing device based on matlab - Google Patents

Electroencephalogram signal processing device based on matlab Download PDF

Info

Publication number
CN113907768A
CN113907768A CN202111185158.7A CN202111185158A CN113907768A CN 113907768 A CN113907768 A CN 113907768A CN 202111185158 A CN202111185158 A CN 202111185158A CN 113907768 A CN113907768 A CN 113907768A
Authority
CN
China
Prior art keywords
data
time
matlab
unit
electroencephalogram
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202111185158.7A
Other languages
Chinese (zh)
Inventor
张继勇
舒洪睿
李小东
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Zhejiang Handrui Intelligent Technology Co Ltd
Original Assignee
Zhejiang Handrui Intelligent Technology Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Zhejiang Handrui Intelligent Technology Co Ltd filed Critical Zhejiang Handrui Intelligent Technology Co Ltd
Priority to CN202111185158.7A priority Critical patent/CN113907768A/en
Publication of CN113907768A publication Critical patent/CN113907768A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/24—Detecting, measuring or recording bioelectric or biomagnetic signals of the body or parts thereof
    • A61B5/316—Modalities, i.e. specific diagnostic methods
    • A61B5/369—Electroencephalography [EEG]
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/16—Devices for psychotechnics; Testing reaction times ; Devices for evaluating the psychological state
    • A61B5/165—Evaluating the state of mind, e.g. depression, anxiety
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/24—Detecting, measuring or recording bioelectric or biomagnetic signals of the body or parts thereof
    • A61B5/25—Bioelectric electrodes therefor
    • A61B5/279—Bioelectric electrodes therefor specially adapted for particular uses
    • A61B5/291—Bioelectric electrodes therefor specially adapted for particular uses for electroencephalography [EEG]
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
    • A61B5/7203—Signal processing specially adapted for physiological signals or for diagnostic purposes for noise prevention, reduction or removal
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
    • A61B5/7235—Details of waveform analysis
    • A61B5/725—Details of waveform analysis using specific filters therefor, e.g. Kalman or adaptive filters
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
    • A61B5/7235—Details of waveform analysis
    • A61B5/7253—Details of waveform analysis characterised by using transforms
    • A61B5/7257—Details of waveform analysis characterised by using transforms using Fourier transforms
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
    • G06F17/10—Complex mathematical operations
    • G06F17/14—Fourier, Walsh or analogous domain transformations, e.g. Laplace, Hilbert, Karhunen-Loeve, transforms
    • G06F17/141—Discrete Fourier transforms
    • G06F17/142—Fast Fourier transforms, e.g. using a Cooley-Tukey type algorithm
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
    • G06F17/10—Complex mathematical operations
    • G06F17/14—Fourier, Walsh or analogous domain transformations, e.g. Laplace, Hilbert, Karhunen-Loeve, transforms
    • G06F17/148—Wavelet transforms
    • H—ELECTRICITY
    • H03—ELECTRONIC CIRCUITRY
    • H03H—IMPEDANCE NETWORKS, e.g. RESONANT CIRCUITS; RESONATORS
    • H03H17/00—Networks using digital techniques
    • H03H17/02—Frequency selective networks
    • H03H17/04—Recursive filters
    • H—ELECTRICITY
    • H03—ELECTRONIC CIRCUITRY
    • H03H—IMPEDANCE NETWORKS, e.g. RESONANT CIRCUITS; RESONATORS
    • H03H17/00—Networks using digital techniques
    • H03H2017/0072—Theoretical filter design
    • H03H2017/009—Theoretical filter design of IIR filters

Abstract

The invention discloses a matlab-based electroencephalogram signal processing device, which comprises a data acquisition and calculation module and a data display module, wherein the design method of the data acquisition and calculation module comprises the following steps: s11, setting a server based on Java Socket of TCP, the server obtaining data from the local test file; s12, the TCP Socket server side sends the data to a Spark Streaming client side for real-time data calculation processing by using thread control according to the random speed of 50 to 100 transaction records per second; s13, the client receives the data, carries out data grouping by using a MapToPair operator, and then carries out data stateful calculation by using an updateStateByKey operator; after the next batch of data is calculated, updating the value of the state in the cache region, which is the same as the key of the cache region; s14, the client writes the calculated data into the database in batches, and the database continuously updates the state values according to keys.

Description

Electroencephalogram signal processing device based on matlab
Technical Field
The invention belongs to the technical field of signal analysis, and relates to an electroencephalogram signal processing device based on matlab.
Background
The brain electrical signals are rhythmic bioelectrical activity on the surface of the human scalp that is spontaneously generated by neurons. The German neurologist Hans Berger firstly acquires an electric signal in 1926, and the EEG signal can reflect the cognitive and emotional conditions of human beings, so that the German neurologist Hans Berger is always concerned by the human beings. Has wide application in brain diseases such as epilepsy and mental diseases such as anxiety and depression.
With the rise of computers, the analysis of electroencephalogram signals by using computers has become a trend. The superiority of the computer capable of rapidly processing a large amount of data prompts a new algorithm for researching electroencephalogram signals.
(1) C4.5 decision tree algorithm
The decision tree is a graphical solution that can intuitively use an analysis probability by constructing a decision tree to obtain a probability that an expected value of a net present value is zero or more on the basis of the probabilities of occurrence of various cases. This decision peak is plotted graphically like a tree, hence the name decision tree, which is a predictive value in machine learning, representing an intermediate mapping between object attributes and object values [13 ]. Its advantages are easily understood classification rule and high accuracy. The disadvantage is that during the process of manufacturing the tree, multiple sequential scans and sequences of data are required, resulting in inefficient algorithms that cannot run if the training set is large enough to fit in memory.
(2) Naive Bayes algorithm
The core of the naive Bayes algorithm is a Bayes formula, and the naive Bayes algorithm is a classification model with wider application. Derived from classical mathematical theory, can generally bring stable classification efficiency on classification problems. Not much estimation parameters are needed and missing data can be processed. Compared with other classification theories, the naive Bayes algorithm has the minimum error rate, and the naive Bayes model has the most favorable performance under the condition of smaller correlation. However, due to the estimation parameters required by the classification, the data missing is not sensitive enough, the calculation method is relatively simple, and meanwhile, if the data correlation is high, the classification efficiency of the naive Bayes algorithm is not as good as that of the decision tree algorithm.
Disclosure of Invention
In view of the above circumstances and problems, the technical solution of the present invention includes a matlab-based electroencephalogram signal processing apparatus, which includes a data collector, a denoising unit, a nonlinear unit and a time-frequency domain unit, wherein an output of the data collector is connected to an input of the denoising unit, an output of the denoising unit is connected to inputs of the nonlinear unit and the time-frequency domain unit, respectively, wherein,
the data acquisition unit comprises three electrodes for acquiring electroencephalogram signals: fp1, Fpz, Fp2 and a pinna electrode as a circuit;
the denoising unit adopts wavelet three-layer decomposition, and comprises the steps of performing wavelet transformation on signals acquired by a data acquisition unit to obtain a wavelet coefficient; denoising the signal through different characteristics of the signal and noise in a wavelet transform domain; reconstructing the denoised signal;
the nonlinear unit comprises calculation of correlation dimension, Co complexity and Renyi entropy;
and the time-frequency domain unit is used for carrying out sectional processing on the obtained signals through an IIR fourth-order filter.
Preferably, the sampling frequency of the data collector is 250 Hz.
Preferably, the denoising unit automatically denoises the signal by using a wden function, performs threshold processing by using a thselect command, and performs threshold processing on the intensity of the signal noise by using a ddencomp command.
Preferably, the correlation dimension is a geometric measure of the dynamics complexity, which is used for the near-four digits of the phase space, and reflects the correlation degree between the dynamic features of the electroencephalogram signal and the electroencephalogram sequence, and is calculated by: the electroencephalogram data sequence obtains the element dimension by taking the equal spacing amount tau from the N data points { xi, i ═ 1, … … N }, and using an m-dimensional Euclidean space:
X(i)＝{x(i),x(i+τ),…,x[i+(m-1)τ]},i＝1,2,…N-(m-1)τ
integral function of correlation
Figure BDA0003298866620000031
r is the radial distance around each reference point x (i), M is the number of data points in phase space, x (i) -x (j) is the euclidean form, θ (x) represents a step function;
the calculation formula of the correlation dimension CD is:
Figure BDA0003298866620000032
preferably, the Co complexity reflects the degree of irregularity, and is a representation of the randomness of the sequence, and the complex sequence is decomposed into a regular activity and a random activity, and is equal to the area ratio between the random activity sequence and the time axis in value, and the area ratio between the whole complex activity sequence and the time axis is calculated by the following calculation:
let the time series of EEG signals be x (N) { x (0), x (1), … …, x (N-1) }, N ═ 0,1,2, … …, N-1, N sample points;
firstly, performing Fast Fourier Transform (FFT) on x (n):
Figure BDA0003298866620000033
calculating the mean square value G of X (k)N，
Figure BDA0003298866620000034
Replacing x (k) or less with 0 to obtain a new spectral order y (k):
Figure BDA0003298866620000035
then Y (n) is obtained through the inverse FFT of Y (k), and the complexity of Co is defined as:
Figure BDA0003298866620000036
wherein, y (n) is defined as the electroencephalogram rule active part, and x (n) -y (n) is defined as the electroencephalogram sequence random part.
Preferably, the Renyi entropy includes amplitude information and frequency information of the signal, and the Renyi entropy is used for analyzing a time series of a non-stationary process or a non-gaussian process, and is calculated by:
Figure BDA0003298866620000041
α denotes the generalized entropy of order α, and pi is the probability of each subinterval.
Preferably, the sampling frequency of the IIR fourth-order filter in the time-frequency domain unit is kept consistent with the frequency when data is collected, and the number of sampling points is 30000.
Preferably, the passband of the IIR fourth-order filter is set to be 0.5-4 Hz in the time-frequency domain unit, and a Delta waveband time domain map and a Delta waveband frequency domain map are obtained.
Preferably, the passband of the IIR fourth-order filter is set to be 4-8 Hz in the time-frequency domain unit, and a Theta-band time-domain map and a Theta-band frequency-domain map are obtained.
Preferably, the passband of the IIR fourth-order filter is set to be 8-14 Hz in the time-frequency domain unit, and an Alpha waveband time domain map and a frequency domain map are obtained.
The invention has the following specific beneficial effects:
1) the difference of the electroencephalogram signals can be displayed more intuitively through time-frequency transformation;
2) placing the electroencephalogram signals of the depression group and the contrast group in the same image for visual comparison;
3) the difference between the nonlinear index and the linear index is displayed in a more three-dimensional manner by combining the nonlinear index and the linear index;
4) the noise removal by utilizing matlab through wavelet transformation is more efficient;
5) and the different wave bands are divided for comparative study, so that the difference is more visually exposed.
Drawings
FIG. 1 is a structural block diagram of a matlab-based electroencephalogram signal processing device according to an embodiment of the present invention;
FIG. 2 is an exploded view of the electroencephalogram signal wavelet of the matlab-based electroencephalogram signal processing device according to the embodiment of the present invention;
FIG. 3 is a non-linear (CD, Co, Renyi entropy) difference diagram of three electroencephalogram channels of a depressed group and a control group of the matlab-based electroencephalogram signal processing device according to the embodiment of the invention;
FIG. 4 is a time domain and frequency domain diagram of Delta wave bands of a depression group (MDD) and a control group (HC) of the matlab-based electroencephalogram signal processing device according to the embodiment of the invention;
FIG. 5 is a time domain and frequency domain diagram of the depressive group (MDD) and the control group (HC) Theta band of the matlab-based electroencephalogram signal processing apparatus according to the embodiment of the present invention;
FIG. 6 is a time domain and frequency domain diagram of the depressed group (MDD) and the control group (HC) Alpha wave band of the matlab-based electroencephalogram signal processing device according to the embodiment of the invention;
FIG. 7 is a time domain and frequency domain diagram of Beta wave bands of a depression group (MDD) and a control group (HC) of the matlab-based electroencephalogram signal processing device according to the embodiment of the invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
On the contrary, the invention is intended to cover alternatives, modifications, equivalents and alternatives which may be included within the spirit and scope of the invention as defined by the appended claims. Furthermore, in the following detailed description of the present invention, certain specific details are set forth in order to provide a better understanding of the present invention. It will be apparent to one skilled in the art that the present invention may be practiced without these specific details.
The invention adopts the denoising of the maximum value of the multi-scale wavelet transform mode and has the characteristic of multi-resolution. The denoising unit has the capacity of representing the local characteristics of signals in two time-frequency domains, and has the characteristic that a time window can be changed in a stretching way along with a flat road in wavelet transformation relative to Fourier change. This is exactly the rule that the duration of the high frequency signal is short and the duration of the low frequency signal is long in practical problems.
The invention compares the electroencephalogram signals by using matlab, and places the depression group and the control group in the same picture for comparative study, thereby facilitating visual observation.
According to the method, the results are more three-dimensional and convincing through the Correlation Dimension (CD), the Co complexity and the Renyi entropy combined with the frequency domain image contrast research of the electroencephalogram signals.
Referring to fig. 1, a block diagram of a matlab-based electroencephalogram signal processing apparatus according to an embodiment of the present invention is shown, and a structural block diagram of the matlab-based electroencephalogram signal processing apparatus includes a data collector 10, a denoising unit 20, a nonlinear unit 31, and a time-frequency domain unit 32, an output of the data collector 10 is connected to an input of the denoising unit 20, an output of the denoising unit 20 is respectively connected to inputs of the nonlinear unit 31 and the time-frequency domain unit 32, wherein,
the data collector 10 includes three electrodes for collecting electroencephalogram signals: fp1, Fpz, Fp2 and a pinna electrode as a circuit;
the denoising unit 20 adopts wavelet three-layer decomposition, including performing wavelet transformation on the signals acquired by the data acquisition unit 10 to obtain wavelet coefficients; denoising the signal through different characteristics of the signal and noise in a wavelet transform domain; reconstructing the denoised signal; the electroencephalogram data contains various kinds of noise. The wearable device is placed at Fp1, Fpz and Fp2 of the forehead, is easily influenced by eye electrical or myoelectrical noise, and in addition, 50Hz power frequency interference exists around the acquisition environment. The invention adopts Mallat to put forward a de-noising method of multi-scale wavelet transform modulus maximum, it has characteristic of multiresolution, and have the ability to characterize the local characteristic of the signal in two domains of time and frequency, relative to Fourier's change, the wavelet transform has characteristic that the time window can change with the level road is flexible, when analyzing the low frequency (corresponding to the large scale) signal, its time window is very large, and when analyzing the high frequency (corresponding to the small scale) signal, its time window is reduced. This is exactly the rule that the duration of the high frequency signal is short and the duration of the low frequency signal is long in practical problems. Denoising based on wavelet threshold requires the following three steps: wavelet transforming the signal to obtain a wavelet coefficient; secondly, denoising the signal in a wavelet transform domain according to different characteristics of the signal and noise; finally, the de-noised signals need to be reconstructed.
In the denoising unit, the invention uses a wden function to automatically denoise signals, and in the threshold processing, the invention selects a thselect command and then calculates the threshold of the intensity of signal noise through a ddencomp command. The wavelet toolkit in Matlab provides the functions and commands we need. The invention selects wavelet three-layer decomposition to denoise the electroencephalogram signal, and randomly selects a group of obtained data to simulate in order to verify the effect of wavelet denoising. Referring to the three-layer decomposition result of the electroencephalogram signal in fig. 2, it can be seen from the figure that a good denoising effect can be achieved by selecting a proper wavelet basis and a proper threshold according to the characteristics of noise. The noise reduction of the non-stationary signal by the wavelet can achieve the incomparable advantages of the traditional filter, can reflect the transient abnormal phenomenon in the signal, can effectively separate the useful components and the noise in the signal, and is a superior method in the noise reduction process of the signal.
The non-linear unit 31 includes calculating the correlation dimension, Co complexity and Renyi entropy;
the time-frequency domain unit 32 performs a segmentation process on the obtained signal through an IIR fourth-order filter.
The sampling frequency of data collector 10 is 250 Hz.
The denoising unit 20 includes performing automatic denoising on the signal by using wden function, performing threshold processing by using thselect command, and thresholding the intensity of the signal noise by using ddencomp command.
The correlation dimension CD is a geometric measurement of dynamics complexity, is used for the approximate four digits of a phase space, reflects the correlation degree of the dynamic characteristics of the electroencephalogram signal and the electroencephalogram sequence, and is obtained by the following calculation: the electroencephalogram data sequence uses an m-dimensional Euclidean space to obtain the element dimension of N data points { xi, i ═ 1, … … N }:
X(i)＝{x(i),x(i+τ),…,x[i+(m-1)τ]},i＝1,2,…N-(m-1)τ
integral function of correlation
Figure BDA0003298866620000071
r is the radial distance around each reference point x (i), M is the number of data points in phase space, x (i) -x (j) is the euclidean form, θ (x) represents a step function;
the calculation formula of the correlation dimension CD is:
Figure BDA0003298866620000072
the Co complexity reflects the degree of irregularity, which is a representation of the randomness of the sequence, and the complex sequence is decomposed into regular activities and random activities, and is equal to the area ratio between the random activity sequence and the time axis in value and the area ratio between the whole complex activity sequence and the time axis, and is calculated by the following steps:
let a time series of EEG (Electroencephalogram) signals be x (N) { x (0), x (1), … …, x (N-1) }, N ═ 0,1,2, … …, N-1, N sample points;
firstly, performing Fast Fourier Transform (FFT) on x (n):
Figure BDA0003298866620000081
calculating the mean square value G of X (k)N，
Figure BDA0003298866620000082
Replacing x (k) or less with 0 to obtain a new spectral order y (k):
Figure BDA0003298866620000083
then Y (n) is obtained through the inverse FFT of Y (k), and the complexity of Co is defined as:
Figure BDA0003298866620000084
wherein, y (n) is defined as the electroencephalogram rule active part, and x (n) -y (n) is defined as the electroencephalogram sequence random part.
The Renyi entropy includes amplitude information and frequency information of a signal, and is used for analyzing a time sequence of a non-stationary process or a non-gaussian process and is calculated by the following steps:
Figure BDA0003298866620000085
α denotes the generalized entropy of order α, and pi is the probability of each subinterval.
As can be seen, CD is a geometric measure of kinetic complexity and is an indispensable quantitative index in nonlinear electroencephalogram analysis. CD is considered to be a reflection of the complexity of cortical dynamics under electroencephalographic recording. Thus, higher CD reflects increased brain neural activity. The Co complexity refers to the area ratio between a random activity sequence and a time axis, and is also a measure of the complexity of the brain electrical activity. The more complex and irregular the brain electrical activity, the higher the complexity of Co. The general form of the Renyi entropy Shannon entropy contains the amplitude information and the frequency information of the signal and reflects the time-frequency information of the electroencephalogram signal. The higher Renyi entropy value represents the synthesis complexity of the electroencephalogram signal. In summary, the selected non-linear characteristics reflect the complexity of brain activity. The greater the cognitive load of the brain, the more complex the mental activities and the greater the eigenvalues.
See fig. 3 for the difference in calculated non-linear characteristics in each brain electrical channel (Fp1, Fpz, Fp2) for the depressed group MDD versus the control group HC. From the same species, it can be seen that the differences of the values of the nonlinear features of the depressed groups relative to the control group are positive numbers, which indicates that the nonlinear characteristics of the depressed groups are higher than those of the control group, meaning that the depressed groups have greater brain load relative to the control group. In the invention, the recorded data is the electroencephalogram data of a study subject under a quiet eye-closing condition, and the above results show that the brain activity of a depression patient is in a quiet eye-closing and relaxing state but is not relaxed as normal people.
The sampling frequency of the IIR fourth-order filter in the time-frequency domain unit 32 is kept consistent with the frequency at the time of data acquisition, and the number of sampling points is 30000. The electroencephalogram signal is a very weak bioelectricity signal and is easily interfered by irrelevant noise. The human body itself also has other bioelectrical interferences, such as electrocardio, electrooculogram, and myoelectricity, due to external interferences such as various noises, electromagnetic radiation, etc. These are the interferences to be removed when studying electroencephalogram signals. The invention filters the noise of the EEG signal by using a wavelet transform method.
Compared with the common people, the electroencephalogram signals of the depressed people have different differences in different wave bands, and the electroencephalogram signals are divided into five wave bands for comparison research, wherein the five wave bands respectively comprise: delta wave (Delta) with a frequency in the range of 0.5 to 4 Hz; theta waves (θ) with frequencies in the range of 4 to 8 Hz; alpha waves (α) in the frequency range of 8 to 14 Hz; beta wave (β) with a frequency in the range of 14 to 30 Hz; gamma wave (Gamma), frequency range is 30 to 50 Hz. The results can be more comprehensive.
Referring to fig. 4, the passband of the IIR fourth-order filter in the time-frequency domain unit 32 is set to 0.5 to 4Hz, so as to obtain a Delta band time domain map and a Delta band frequency domain map. It is found from fig. 4 that the amplitude of the electroencephalogram of the MDD in the depression group is significantly stronger than that of the HC in the control group, which indicates that the electroencephalogram activity of the Delta band depression group is more severe. The enhancement of slow wave Delta activity indicates a brain injury, and the more intense the Delta wave activity the more severe the brain injury.
Referring to fig. 5, the passband of the IIR fourth-order filter is set to 4-8 Hz in the time-frequency domain unit 32, so as to obtain a Theta-band time-domain map and a Theta-band frequency-domain map. It can be seen from fig. 5 that the intensity of brain electrical activity of MDD Theta band in depressed group is significantly higher than that of control group HC.
Referring to fig. 6, the passband of the IIR fourth-order filter in the time-frequency domain unit 32 is set to 8 to 14Hz, and an Alpha band time domain map and a frequency domain map are obtained. As shown in FIG. 6, the activity intensity of the EEG signal in Alpha band of the control group HC was higher than that of MDD in the depressed group. When the Alpha level is higher, the cerebral cortex is closer to the idle state, namely, the power spectrum of Alpha waves is inversely proportional to the activity of the brain, the Alpha waves are main brain waveforms of adults in the resting state, and the Alpha activity of the depression group MDD in the resting state is more intense, so that the brain of a depression patient still keeps a more tense state in the resting state, and the study on the nonlinear characteristics of electroencephalogram signals is also proved.
Referring to fig. 7, the passband of the IIR fourth-order filter in the time-frequency domain unit 32 is set to 14-30 Hz, and a Beta-band time-domain map and a frequency-domain map are obtained. The electroencephalogram signals of the Beta waves are compared to the electroencephalogram signals of the depression patients, the electroencephalogram signals of the depression patients are more intense relative to the activities of normal people, the Beta waves are main electroencephalograms of the people under the condition of nervous excitement, the stronger the activity of the Beta waves is, the more the individual is in the nervous anxiety, and the electroencephalogram signals are consistent with the clinical manifestations of the depression patients.

Claims (10)

1. The matlab-based electroencephalogram signal processing device is characterized by comprising a data acquisition unit, a denoising unit, a nonlinear unit and a time-frequency domain unit, wherein the output of the data acquisition unit is connected with the input of the denoising unit, the output of the denoising unit is respectively connected with the input of the nonlinear unit and the input of the time-frequency domain unit, wherein,
the data acquisition unit comprises three electrodes for acquiring electroencephalogram signals: fp1, Fpz, Fp2 and a pinna electrode as a circuit;
the denoising unit adopts wavelet three-layer decomposition, and comprises the steps of performing wavelet transformation on signals acquired by a data acquisition unit to obtain a wavelet coefficient; denoising the signal through different characteristics of the signal and noise in a wavelet transform domain; reconstructing the denoised signal;
the nonlinear unit comprises calculation of correlation dimension, Co complexity and Renyi entropy;
and the time-frequency domain unit is used for carrying out sectional processing on the obtained signals through an IIR fourth-order filter.
2. The matlab-based electroencephalogram signal processing device according to claim 1, wherein the sampling frequency of the data collector is 250 Hz.
3. The matlab-based electroencephalogram signal processing device according to claim 1, wherein the denoising unit comprises a wden function for automatically denoising the signal, a thselect command for thresholding, and a ddencomp command for thresholding the intensity of the signal noise.
4. The matlab-based electroencephalogram signal processing device according to claim 1, wherein the correlation dimension is a geometrical measure of dynamics complexity, is used for a near four-digit number of a phase space, reflects the correlation degree of dynamic features of the electroencephalogram signal and the electroencephalogram sequence, and is obtained by the following calculation: the electroencephalogram data sequence obtains the element dimension by taking the equal spacing amount tau from the N data points { xi, i ═ 1, … … N }, and using an m-dimensional Euclidean space:
X(i)＝{x(i),x(i+τ),…,x[i+(m-1)τ]},i＝1,2,…N-(m-1)τ
integral function of correlation
Figure FDA0003298866610000011
r is the radial distance around each reference point x (i), M is the number of data points in phase space, x (i) -x (j) is the euclidean form, θ (x) represents a step function;
the calculation formula of the correlation dimension CD is:
Figure FDA0003298866610000021
5. the matlab-based electroencephalogram signal processing apparatus according to claim 1, wherein the Co complexity reflects degree of irregularity, and is a representation of sequence randomness, and the complex sequence is decomposed into regular activity and random activity, and is equal to the area ratio between the random activity sequence and the time axis in value, and the area ratio between the whole complex activity sequence and the time axis is calculated by:
let the time series of EEG signals be x (N) { x (0), x (1), … …, x (N-1) }, N ═ 0,1,2, … …, N-1, N sample points;
firstly, performing Fast Fourier Transform (FFT) on x (n):
Figure FDA0003298866610000022
calculating the mean square value G of X (k)N，
Figure FDA0003298866610000023
Replacing x (k) or less with 0 to obtain a new spectral order y (k):
Figure FDA0003298866610000024
then Y (n) is obtained through the inverse FFT of Y (k), and the complexity of Co is defined as:
Figure FDA0003298866610000025
wherein, y (n) is defined as the electroencephalogram rule active part, and x (n) -y (n) is defined as the electroencephalogram sequence random part.
6. The matlab-based electroencephalogram signal processing device according to claim 1, wherein the Renyi entropy includes amplitude information and frequency information of the signal, and is used for analyzing a time sequence of a non-stationary process or a non-Gaussian process, and is calculated by:
Figure FDA0003298866610000031
α is expressed as a generalized entropy of order α, and pi is the probability of each subinterval.
7. The matlab-based electroencephalogram signal processing device according to claim 1, wherein the sampling frequency of the IIR fourth-order filter in the time-frequency domain unit is consistent with the frequency when data is acquired, and the number of sampling points is set to 30000.
8. The matlab-based electroencephalogram signal processing device according to claim 7, wherein the passband of the IIR fourth-order filter in the time-frequency domain unit is set to be 0.5-4 Hz, and a Delta-band time domain map and a Delta-band frequency domain map are obtained.
9. The matlab-based electroencephalogram signal processing device according to claim 7, wherein the passband of the IIR fourth-order filter in the time-frequency domain unit is set to be 4-8 Hz, and a Theta-band time-domain map and a frequency-domain map are obtained.
10. The matlab-based electroencephalogram signal processing device according to claim 7, wherein the passband of the IIR fourth-order filter in the time-frequency domain unit is set to be 8-14 Hz, and an Alpha band time domain map and a frequency domain map are obtained.
CN202111185158.7A 2021-10-12 2021-10-12 Electroencephalogram signal processing device based on matlab Pending CN113907768A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202111185158.7A CN113907768A (en) 2021-10-12 2021-10-12 Electroencephalogram signal processing device based on matlab

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202111185158.7A CN113907768A (en) 2021-10-12 2021-10-12 Electroencephalogram signal processing device based on matlab

Publications (1)

Publication Number Publication Date
CN113907768A true CN113907768A (en) 2022-01-11

Family

ID=79239419

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202111185158.7A Pending CN113907768A (en) 2021-10-12 2021-10-12 Electroencephalogram signal processing device based on matlab

Country Status (1)

Country Link
CN (1) CN113907768A (en)

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106236027A (en) * 2016-08-23 2016-12-21 兰州大学 Depressed crowd's decision method that a kind of brain electricity combines with temperature
WO2017016086A1 (en) * 2015-07-30 2017-02-02 华南理工大学 Depression evaluating system and method based on physiological information
CN111956219A (en) * 2020-08-27 2020-11-20 济南大学 Electroencephalogram signal-based emotion feature identification method and system and emotion feature identification and adjustment system

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2017016086A1 (en) * 2015-07-30 2017-02-02 华南理工大学 Depression evaluating system and method based on physiological information
CN106236027A (en) * 2016-08-23 2016-12-21 兰州大学 Depressed crowd's decision method that a kind of brain electricity combines with temperature
CN111956219A (en) * 2020-08-27 2020-11-20 济南大学 Electroencephalogram signal-based emotion feature identification method and system and emotion feature identification and adjustment system

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
赵庆林: "可穿戴脑电传感器关键问题研究及应用验证", 《中国博士学位论文全文数据库 医药卫生科技辑》, no. 3, pages 17 - 18 *

Similar Documents

Publication Publication Date Title
Djemal et al. EEG-based computer aided diagnosis of autism spectrum disorder using wavelet, entropy, and ANN
Peng et al. Removal of ocular artifacts in EEG—An improved approach combining DWT and ANC for portable applications
Ahirwal et al. Power spectrum analysis of EEG signals for estimating visual attention
Khoshnevis et al. Applications of higher order statistics in electroencephalography signal processing: a comprehensive survey
CN111310570B (en) Electroencephalogram signal emotion recognition method and system based on VMD and WPD
EP1788937A4 (en) Method for adaptive complex wavelet based filtering of eeg signals
Patil et al. Quality advancement of EEG by wavelet denoising for biomedical analysis
CN112861625B (en) Determination method for stacked denoising self-encoder model
CN112426162A (en) Fatigue detection method based on electroencephalogram signal rhythm entropy
Jindal et al. Migraine disease diagnosis from EEG signals using Non-linear Feature Extraction Technique
Daud et al. Wavelet based filters for artifact elimination in electroencephalography signal: A review
Sreekrishna et al. Real time cascaded moving average filter for detrending of electroencephalogram signals
Saini et al. Discriminatory features based on wavelet energy for effective analysis of electroencephalogram during mental tasks
CN112244880B (en) Emotion-induced electroencephalogram signal analysis method based on variable-scale symbol compensation transfer entropy
CN113723557A (en) Depression electroencephalogram classification system based on multiband time-space convolution network
Upadhyay et al. Application of wavelet fractal features for the automated detection of epileptic seizure using electroencephalogram signals
Korde et al. Analysis of eeg signals and biomedical changes due to meditation on brain by using ica for feature extraction
CN113907768A (en) Electroencephalogram signal processing device based on matlab
CN111493864B (en) EEG signal mixed noise processing method, equipment and storage medium
Atmaji et al. Three-class classification of eeg signals using support vector machine methods
Longxin et al. Research on Removing Ocular Artifacts from Multi-Channel EEG signals
Nath et al. Wavelet Based De-noising of EEG Signal Acquired from Tele-serial Addicted Persons
Zhang et al. Method of removing noise from EEG signals based on HHT method
Ullah et al. Identification of Brain disorders by Sub-band Decomposition of EEG signals and Measurement of Signal to Noise Ratio
Li et al. Ocular Artifact Removal Algorithm of Single Channel EEG Based on CEEMDAN-ICA-WTD

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination