CN110584600A - High-frequency oscillation rhythm detection method based on intelligent algorithm optimized fuzzy clustering - Google Patents

High-frequency oscillation rhythm detection method based on intelligent algorithm optimized fuzzy clustering Download PDF

Info

Publication number
CN110584600A
CN110584600A CN201910747243.4A CN201910747243A CN110584600A CN 110584600 A CN110584600 A CN 110584600A CN 201910747243 A CN201910747243 A CN 201910747243A CN 110584600 A CN110584600 A CN 110584600A
Authority
CN
China
Prior art keywords
algorithm
clustering
frequency oscillation
formula
cluster
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.)
Granted
Application number
CN201910747243.4A
Other languages
Chinese (zh)
Other versions
CN110584600B (en
Inventor
吴敏
万雄波
方泽林
杜玉晓
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.)
China University of Geosciences
Original Assignee
China University of Geosciences
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 China University of Geosciences filed Critical China University of Geosciences
Priority to CN201910747243.4A priority Critical patent/CN110584600B/en
Publication of CN110584600A publication Critical patent/CN110584600A/en
Application granted granted Critical
Publication of CN110584600B publication Critical patent/CN110584600B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

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
    • A61B5/40—Detecting, measuring or recording for evaluating the nervous system
    • A61B5/4076—Diagnosing or monitoring particular conditions of the nervous system
    • A61B5/4094—Diagnosing or monitoring seizure diseases, e.g. epilepsy
    • 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
    • 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
    • A61B5/7271—Specific aspects of physiological measurement analysis

Landscapes

  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Engineering & Computer Science (AREA)
  • Public Health (AREA)
  • Molecular Biology (AREA)
  • Veterinary Medicine (AREA)
  • General Health & Medical Sciences (AREA)
  • Physics & Mathematics (AREA)
  • Animal Behavior & Ethology (AREA)
  • Biophysics (AREA)
  • Pathology (AREA)
  • Biomedical Technology (AREA)
  • Heart & Thoracic Surgery (AREA)
  • Medical Informatics (AREA)
  • Surgery (AREA)
  • Physiology (AREA)
  • Psychiatry (AREA)
  • Neurology (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Signal Processing (AREA)
  • Artificial Intelligence (AREA)
  • Neurosurgery (AREA)
  • Psychology (AREA)
  • Measurement And Recording Of Electrical Phenomena And Electrical Characteristics Of The Living Body (AREA)

Abstract

The invention provides a high-frequency oscillation rhythm detection method based on intelligent algorithm optimization fuzzy clustering, which is used for detecting high-frequency oscillation rhythms based on a fuzzy clustering method. The method comprises the following basic steps: selecting average singular value MSV and line length flPower ratio R and spectral centroid fcForming a characteristic vector for the characteristics of the epileptic electroencephalogram signals as the input of a clustering algorithm; adopting a simulated annealing genetic algorithm in an intelligent algorithm to optimize a fuzzy clustering algorithm to obtain an optimized parameter vc(ii) a According to the optimization parameter vcObtaining an optimized result; and selecting a median and a quartile range to analyze the statistical characteristics of each class, and detecting the high-frequency oscillation rhythm. The invention has the beneficial effects that: the detection precision of the high-frequency oscillation rhythm of the epileptic brain electrical signals is improved, and doctors are helped to diagnose epilepsy and excise epileptogenic foci.

Description

High-frequency oscillation rhythm detection method based on intelligent algorithm optimized fuzzy clustering
Technical Field
The invention relates to the field of epilepsia electroencephalogram signal processing, in particular to a high-frequency oscillation rhythm detection method based on intelligent algorithm optimization fuzzy clustering.
Background
Epilepsy is a common nervous system disease and has characteristics of spontaneity and unpredictability. At the time of seizures, patients often exhibit abnormalities such as movement and behavior. The prevalence rate of epilepsy is 0.5% -1% worldwide, wherein the prevalence rate of epilepsy is 0.7% in China, and the prevalence rate still increases year by year. Most epileptic patients can be treated with anti-epileptic drugs, but about 30% of patients are unable to control seizures with drugs, and are therefore diagnosed as refractory epileptic patients.
Refractory epilepsy refers to patients who are treated by conventional and systematic antiepileptic drugs, the concentration of the antiepileptic drugs in blood is kept in an effective range, but the epileptic seizure of the patients can not be controlled, and the work, study or normal life of the patients is seriously affected; the attack frequency is more than 2-4 times per month, and the disease course is more than 4 years. For intractable epilepsy, one of the effective measures for controlling and curing epilepsy is to excise the epileptic focus, and the key to successful operation is the accurate positioning of the focus. Currently, the main means for locating epileptic lesions are electroencephalography examination, neuroelectrophysiology examination, nuclear medicine examination, magnetoencephalography examination, and the like. The electroencephalogram examination is the most common epileptic focus positioning means, and usually, electroencephalogram signals are recorded for a long time by implanting intracranial or subdural large electrodes, electroencephalogram changes in the disease process of an epileptic are captured to form an electroencephalogram, and then, the electroencephalogram examination is carried out by an experienced medical expert. The medical expert realizes the diagnosis of epilepsy by checking the electroencephalogram signals of epilepsy.
The focus of epilepsy is the pathological cerebral cortex area of patients in epileptic seizure stage or inter-seizure period. Generally, the epileptogenic focus is divided into an epileptogenic focus, an irritative focus, an epileptogenic lesion and a functional deficiency area, wherein the epileptogenic focus is the most effective and remarkable mark and can be regarded as a substitute index of the epileptogenic focus.
The traditional electroencephalogram method focuses on electroencephalogram signals below 40Hz, and the epileptic seizure onset region is located by extracting electroencephalogram frequency components in the frequency range. However, the low-frequency components detected by the method are easily interfered by other signals, so that the positioning result is wrong, and the resection operation is failed. At the same time, the method is very time-consuming, the positioning time is about 24-72 hours, and the operation risk is increased. Therefore, it is very urgent to find a new marker or a fast positioning method.
In the last two decades, more and more researchers have been focusing on the high frequency oscillation rhythm of epileptic brain electrical signals above 80 Hz. During the latent phase of an epileptic seizure, the patient's brain begins to undergo pathological changes. There is a significant difference in the high frequency oscillation rhythm in the brain before and after the lesion. A great deal of research shows that the high-frequency oscillation rhythm can be used as a biomarker of an epileptic seizure onset region, the high-frequency oscillation rhythm has higher incidence rate in the epileptic seizure onset region and obvious specificity, so that the rapid and accurate positioning of the seizure onset region can be realized by detecting the high-frequency oscillation rhythm. By accurately positioning the seizure onset region, the aim of epileptic focus excision surgery can be determined, and the risk of unrecoverable cranial nerve injury caused by the injury of a functional region in the surgery is reduced.
Disclosure of Invention
In order to solve the problems, the invention provides a high-frequency oscillation rhythm detection method based on intelligent algorithm optimization fuzzy clustering. A high-frequency oscillation rhythm detection method based on intelligent algorithm optimization fuzzy clustering mainly comprises the following steps:
s101: acquiring an epilepsia electroencephalogram signal time sequence of an epileptic patient, and calculating four characteristics of the electroencephalogram signal time sequence at different moments: mean singular value MSV, line length flPower ratio R and spectral centroid fcFurther form a plurality of feature vectors Vi(ii) a Each feature vector consists of four features at a moment;
s102: optimizing a fuzzy c-means clustering algorithm by adopting a simulated annealing algorithm and combining a genetic algorithm, and according to the plurality of eigenvectors ViObtaining an optimized clustering center vc；
S103: according to the optimized clustering center vcObtaining a clustering result;
s104: according to the clustering result, a boxplot method is adopted to carry out statistical analysis on four characteristics in each clustered class, and the high-frequency oscillation rhythm of the epileptic brain electrical signal is detected; the four characteristics in each class refer to the average singular value MSV and the line length f of each class after the clustering is obtainedlPower ratio R and spectral centroid fc。
Further, in step S101, the specific calculation steps of the average singular value are as follows:
s201: obtaining a time-frequency information matrix S (tau, f) of the epileptic brain electrical signal time sequence x (t) by using S transformation, wherein the calculation formula is shown as formula (1):
in the above formula, f is frequency, j is an imaginary unit, and τ is a position parameter in time domain t;
s202: performing singular value decomposition on the obtained time-frequency information matrix S (tau, f) to obtain singular values, wherein a calculation formula is shown as a formula (2):
S(τ,f)＝UΓVT (2)
wherein S (tau, f) is M multiplied by N, and U and V are orthogonal matrixes of M multiplied by M, N multiplied by N respectively; Γ has a size of M × N and is in the form ofΛr×r＝diag(σ1,σ2,...σr)，σiThe singular values are called as time frequency information matrix S (tau, f), and r is the total number of the singular values;
s203: determining the effective values and the number of the obtained singular values by using a k-medoids algorithm;
s204: and calculating the average value of the effective singular values to obtain an average singular value MSV.
Further, in step S203, the effective number of the obtained singular values is determined by using a k-medoids algorithm, which specifically includes the following steps:
s301: singular value σ obtained from S202i(i ═ 1, 2.., r), o is selected1＝σ1And o2∈{σ2，σ3，...σrTaking the k-medoids as the center of the k-medoids algorithm;
s302: respectively calculating the remaining singular values to the center o according to the formula (3)1And o2Distance ofSeparation device
In the formula (3), sj∈{σ2，σ3，...，σrAnd sj≠o2(ii) a If d is1(sj，o1)≤d2(sj，o2) A 1 is tojIs assigned to1Centered cluster C1Performing the following steps; otherwise will sjIs assigned to2Centered cluster C2Performing the following steps;
s303: from cluster C2In (1), a value o is randomly selected2', as a clustering center, cluster C1Center in (A) is o1′＝o1(ii) a Calculating new o according to the method in S3021' and o2' Cluster C1' and C2'; and the following values were calculated:
wherein, if J is less than 0, use o2' as a New Cluster center instead of o2(ii) a Otherwise, the original center o is retained2；
S304: looping steps S302 and S303 until the newly generated cluster center does not change any more; containing sigma1The cluster is needed, and the obtained cluster number is q (q is more than or equal to 1 and less than or equal to r-1), namely q effective singular values sigmak(k＝1，2...，q)。
Further, in step S204, the calculation formula of the average value of the effective singular values is shown as formula (3):
in step S101, the line length flIs represented by equation (6):
in the formula (6), LlIs x (k) length, and x (k) is amplitude of epileptic brain electrical signal time sequence;
the calculation formula of the power ratio R is shown in formula (7):
in the formula (7), P[80-200]Is the power, P, of the suspected high frequency oscillation rhythm in ripple (ripple, 80-200 Hz) bandwidth[250-500]The power of the suspected high-frequency oscillation rhythm is within a fast ripple (250-500 Hz) bandwidth; the ripple bandwidth is 80-200 Hz; the fast ripple bandwidth is 250-500 Hz;
the frequency spectrum centroid fcIs represented by equation (8):
in equation (8), T is the sampling period, L is the window length, M (k) is the multi-window power spectral density estimate, andw (L) is a Hamming window of length L.
Further, in step S102, the specific steps of optimizing the fuzzy c-means clustering algorithm by using the simulated annealing algorithm in the intelligent algorithm in combination with the genetic algorithm are as follows:
s401: initializing control parameters: size of population individual NpMaximum number of evolutionary events NmaxCross probability PcProbability of variation PmAnnealing initiation temperature T0Coefficient of cooling tkTermination temperature Tend；
S402: randomly generating NpGroup initial clustering center vcC, C is the number of clusters, and an initial population Chrom is generated, and a membership value μ of the ith eigenvector belonging to the C-th class is calculated according to formula (9)ic：
In the formula (9), muicThe conditions are satisfied:c is the number of clusters and p is a weighted fuzzy parameter, usually taking the value 2, ViIs the extracted four-dimensional feature vector and is composed of the extracted average singular value, the line length, the power ratio, and the spectrum centroid, i ═ 1,2f，NfThe length of a preset four-dimensional feature vector; calculating to obtain a membership value muicThen, the j (j ═ 1, 2., N) is calculated according to the formula (8)p) Fitness f of group individualj：
S403: setting a cycle count variable gen as 0;
s404: performing selective crossover and mutation genetic operations on the population Chrom, and carrying out selection crossover and mutation genetic operations on new NpCalculating clustering center, membership degree and fitness value f 'of individual'j(ii) a If'j＞fjReplacing the new individual with the new individual and going to step S405; otherwise, with probabilityReceiving new individuals and discarding old individuals;
s405: if gen < NmaxIf so, go to S404; otherwise, go to S406;
s406: if Ti＜TendIf yes, the algorithm successfully returns the global optimal solution to the step S407; otherwise, executing a cooling operation Ti+1＝tkTiGo to S403;
s407: the resulting clustering center vcTo optimize the resulting parameters.
Further, in step S103, according to the optimized parameters, the specific steps of obtaining the clustering result are as follows:
s501: obtaining initialization parameters based on a fuzzy c-means clustering algorithm by adopting a simulated annealing algorithm and combining genetic algorithm optimization;
s502: calculating a four-dimensional feature vector V by equation (11) based on the initialization parametersiDegree of membership μ to class C (C ═ 1,2, …, C)iC：
In the above formula, vcRefers to the center of the cluster c, vkRefers to the center of cluster k.
S503: according to the obtained four-dimensional feature vector ViDegree of membership mu ofiCThe feature vector ViAnd classifying the cluster into the class to which the maximum membership degree belongs to obtain a clustering result.
In step S104, the median and the quartile range of each of the four features in each of the clustered classes are obtained using the box plot, and the obtained median and quartile range are analyzed to determine the waveform state of the class varying with time, thereby detecting the high-frequency oscillation rhythm.
Drawings
The invention will be further described with reference to the accompanying drawings and examples, in which:
FIG. 1 is a flow chart of a method for detecting a high-frequency oscillation rhythm based on an intelligent algorithm optimized fuzzy clustering in the embodiment of the present invention;
FIG. 2 is a graph showing the average singular value characteristic in two states of the high-frequency oscillation rhythm and the non-high-frequency oscillation rhythm according to the embodiment of the present invention;
FIG. 3 is a graph showing the line length characteristics in two states of a high-frequency oscillatory rhythm and a non-high-frequency oscillatory rhythm according to the embodiment of the present invention;
FIG. 4 is a graph showing the power ratio characteristics in two states of a high-frequency oscillatory rhythm and a non-high-frequency oscillatory rhythm in the embodiment of the present invention;
FIG. 5 is a graph showing the characteristic of the centroid of the frequency spectrum in two states of the high-frequency oscillation rhythm and the non-high-frequency oscillation rhythm according to the embodiment of the present invention.
Detailed Description
For a more clear understanding of the technical features, objects and effects of the present invention, embodiments of the present invention will now be described in detail with reference to the accompanying drawings.
The embodiment of the invention provides a high-frequency oscillation rhythm detection method based on intelligent algorithm optimized fuzzy clustering.
Referring to fig. 1, fig. 1 is a flowchart of a method for detecting a high-frequency oscillation rhythm based on an intelligent algorithm optimized fuzzy clustering in an embodiment of the present invention, which specifically includes the following steps:
s101: acquiring an epilepsia electroencephalogram signal time sequence of an epileptic patient, and calculating four characteristics of the electroencephalogram signal time sequence at different moments: mean singular value MSV, line length flPower ratio R and spectral centroid fcFurther form a plurality of feature vectors Vi(ii) a Each feature vector consists of four features at a moment;
s102: optimizing a fuzzy c-means clustering algorithm by adopting a simulated annealing algorithm and combining a genetic algorithm, and according to the plurality of eigenvectors ViObtaining an optimized clustering center Vc；
S103: according to the optimized clustering center vcObtaining a clustering result;
s104: according to the clustering result, a boxplot method is adopted to carry out statistical analysis on four characteristics in each clustered class, and the high-frequency oscillation rhythm of the epileptic brain electrical signal is detected; the four characteristics in each class refer to the average singular value MSV and the line length f of each class after the clustering is obtainedlPower ratio R and spectral centroid fc。
Further, in step S101, the specific calculation steps of the average singular value are as follows:
s201: obtaining a time-frequency information matrix S (tau, f) of the epileptic brain electrical signal time sequence x (t) by using S transformation, wherein the calculation formula is shown as formula (1):
in the above formula, f is frequency, j is an imaginary unit, and τ is a position parameter in time domain t;
s202: performing singular value decomposition on the obtained time-frequency information matrix S (tau, f) to obtain singular values, wherein a calculation formula is shown as a formula (2):
S(τ，f)＝UΓVT (2)
wherein S (tau, f) is M multiplied by N, and U and V are orthogonal matrixes of M multiplied by M, N multiplied by N respectively; Γ has a size of M × N and is in the form ofΛr×r＝diag(σ1，σ2，...σr)，σiThe singular values are called as time frequency information matrix S (tau, f), and r is the total number of the singular values;
s203: determining the effective values and the number of the obtained singular values by using a k-medoids algorithm;
s204: and calculating the average value of the effective singular values to obtain an average singular value MSV.
Further, in step S203, the effective number of the obtained singular values is determined by using a k-medoids algorithm, which specifically includes the following steps:
s301: singular value σ obtained from S202i(i ═ 1, 2.., r), o is selected1＝σ1And o2∈{σ2，σ3，...σrTaking the k-medoids as the center of the k-medoids algorithm;
s302: respectively calculating the remaining singular values to the center o according to the formula (3)1And o2Is a distance of
In the formula (3), sj∈{σ2，σ3，...，σrAnd sj≠o2(ii) a If d is1(sj，o1)≤d2(sj，o2) A 1 is tojIs assigned to1Centered cluster C1Performing the following steps; otherwise will sjIs assigned to2Centered cluster C2Performing the following steps;
s303: from cluster C2In (1), a value o is randomly selected2', as a clustering center, cluster C1Center in (A) is o1′＝o1(ii) a Calculating new o according to the method in S3021' and o2' Cluster C1' and C2'; and the following values were calculated:
wherein, if J is less than 0, use o2' as a New Cluster center instead of o2(ii) a Otherwise, the original center o is retained2；
S304: looping steps S302 and S303 until the newly generated cluster center does not change any more; containing sigma1The cluster is needed, and the obtained cluster number is q (q is more than or equal to 1 and less than or equal to r-1), namely q effective singular values sigmak(k＝1，2...，q)。
Further, in step S204, the calculation formula of the average value of the effective singular values is shown as formula (3):
in step S101, the line length flIs represented by equation (6):
in the formula (6), LlIs x (k) length, and x (k) is amplitude of epileptic brain electrical signal time sequence;
the calculation formula of the power ratio R is shown in formula (7):
in the formula (7), P[80-200]Is the power, P, of the suspected high frequency oscillation rhythm in ripple (ripple, 80-200 Hz) bandwidth[250-500]The power of the suspected high-frequency oscillation rhythm is within a fast ripple (250-500 Hz) bandwidth; the ripple bandwidth is 80-200 Hz; the fast ripple bandwidth is 250-500 Hz;
the frequency spectrum centroid fcIs represented by equation (8):
in equation (8), T is the sampling period, L is the window length, M (k) is the multi-window power spectral density estimate, andw (L) is a Hamming window of length L.
Further, in step S102, the specific steps of optimizing the fuzzy c-means clustering algorithm by using the simulated annealing algorithm in the intelligent algorithm in combination with the genetic algorithm are as follows:
s401: initializing control parameters: size of population individual NpMaximum number of evolutionary events NmaxCross probability PcProbability of variation PmAnnealing initiation temperature T0Coefficient of cooling tkTermination temperature Tend；
S402: randomly generating NpGroup initial clustering center vcC, C is the number of clusters, and an initial population Chrom is generated, and a membership value μ of the ith eigenvector belonging to the C-th class is calculated according to formula (9)ic：
In the formula (9), muicThe conditions are satisfied:c is the number of clusters and p is a weighted fuzzy parameter, usually taking the value 2, ViIs the extracted four-dimensional feature vector and is composed of the extracted average singular value, the line length, the power ratio, and the spectrum centroid, i ═ 1,2f，NfThe length of a preset four-dimensional feature vector; calculating to obtain a membership value muicThen, the j (j ═ 1, 2., N) is calculated according to the formula (8)p) Fitness f of group individualj：
S403: setting a cycle count variable gen as 0;
s404: performing selective crossover and mutation genetic operations on the population Chrom, and carrying out selection crossover and mutation genetic operations on new NpCalculating clustering center, membership degree and fitness value f 'of individual'j(ii) a If'j＞fjReplacing the new individual with the new individual and going to step S405; otherwise, with probabilityReceiving new individuals and discarding old individuals;
s405: if gen < NmaxIf so, go to S404; otherwise, go to S406;
s406: if Ti＜TendIf yes, the algorithm successfully returns the global optimal solution to the step S407; otherwise, executing a cooling operation Ti+1＝tkTiGo to S403;
s407: the resulting clustering center vcTo optimize the resulting parameters.
Further, in step S103, according to the optimized parameters, the specific steps of obtaining the clustering result are as follows:
s501: obtaining initialization parameters based on a fuzzy c-means clustering algorithm by adopting a simulated annealing algorithm and combining genetic algorithm optimization;
s502: calculating a four-dimensional feature vector by equation (11) based on the initialization parameterViDegree of membership μ to class C (C ═ 1,2, …, C)iC：
In the above formula, vcRefers to the center of the cluster c, vkRefers to the center of cluster k.
S503: according to the obtained four-dimensional feature vector ViDegree of membership mu ofiCThe feature vector ViAnd classifying the cluster into the class to which the maximum membership degree belongs to obtain a clustering result.
In step S104, the median and the quartile range of each of the four features in each of the clustered classes are obtained using the box plot, and the obtained median and quartile range are analyzed to determine the waveform state of the class varying with time, thereby detecting the high-frequency oscillation rhythm.
Graphs of the four characteristics in two different states of the high-frequency oscillatory rhythm and the non-high-frequency oscillatory rhythm, respectively, are shown in fig. 2 to 5.
Referring to fig. 2, fig. 2 is a graph of the average singular value in both the high-frequency oscillatory rhythm and the non-high-frequency oscillatory rhythm, from which it can be seen that the value of the average singular value in the high-frequency oscillatory rhythm state is higher than that in the non-high-frequency oscillatory rhythm state;
referring to fig. 3, fig. 3 is a graph of line lengths in different states, where the value of the line length in the high frequency oscillation rhythm is significantly higher than that in the non-high frequency oscillation rhythm;
referring to fig. 4, fig. 4 is a graph of power ratio at different states, the value of the power ratio at high frequency oscillation rhythm being significantly higher than the value at non-high frequency oscillation rhythm;
referring to fig. 5, fig. 5 is a graph of the spectral centroid in different states, from which it can be seen that the value of the spectral centroid in the high-frequency oscillatory rhythm state is generally higher than that in the non-high-frequency oscillatory rhythm state. As can be seen from fig. 2 to 5, in the two states of the high-frequency oscillation rhythm and the non-high-frequency oscillation rhythm, the average singular value, the line length, the power ratio and the spectrum centroid are significantly different, and can be used for classifying the high-frequency oscillation rhythm.
In order to verify that the features extracted from the patient data can also be used to distinguish between high frequency oscillatory rhythms and non-high frequency oscillatory rhythms, the ManWhitney U test is used to analyze whether there is a significant difference in the 4 features extracted from different patient high frequency oscillatory rhythms and non-high frequency oscillatory rhythms. The absolute value of the statistic | Z | obtained by the mann whitney U test is shown in table 1:
table 1: manhutney U test under high-frequency oscillation rhythm and non-high-frequency oscillation rhythm states of 5 patients
As can be seen from Table 1, the absolute value | Z | of the statistic exceeds Z α/2(α is significance level, set to 0.05, and the value of Z α/2 is 1.96), falls into the negative domain, and rejects the original hypothesis, so that 4 features extracted under the high-frequency oscillation rhythm and non-high-frequency oscillation rhythm states of 5 patients have significant difference, namely the 4 features can be used for distinguishing the high-frequency oscillation rhythm from the epileptic brain electrical signal.
Clustering feature vectors extracted from patient epilepsia electroencephalogram signals by adopting a fuzzy clustering algorithm to obtain different categories; in order to give descriptive statistics for each category, according to a clustering result obtained by the high-frequency oscillation rhythm detection method based on the intelligent algorithm optimized fuzzy clustering, two indexes of median and quartile range are obtained by using a boxplot method, statistical analysis is carried out on different category characteristics, and the high-frequency oscillation rhythm of the epileptic brain electrical signals is detected. The different categories are indicated in table 2:
TABLE 2 different categories of indices
As can be seen from table 2: class 2 has the lowest power ratio median, the highest spectrum centroid median, and the spectrum centroid median of 260Hz, so class 2 is FRs; the average singular value, the line length and the power ratio median of the class 1 are the largest, and the median of the spectrum centroid is 141Hz, so the class 1 is ripples; the average singular value and line length are lower in category 3 compared to categories 1 and 2, so category 3 is an artifact; wherein Ripples and FRs (Ripples + FRs) are both high frequency oscillation rhythms.
The invention has the beneficial effects that: the detection speed of the high-frequency oscillation rhythm of the epileptic brain electrical signals is improved, and doctors are helped to diagnose epilepsy and excise epileptogenic focus.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents, improvements and the like that fall within the spirit and principle of the present invention are intended to be included therein.

Claims (7)

1. A high-frequency oscillation rhythm detection method based on intelligent algorithm optimization fuzzy clustering is characterized in that: the method comprises the following steps:
s101: acquiring an epilepsia electroencephalogram signal time sequence of an epileptic patient, and calculating four characteristics of the electroencephalogram signal time sequence at different moments: mean singular value MSV, line length flPower ratio R and spectral centroid fcFurther form a plurality of feature vectors Vi(ii) a Each feature vector consists of four features at a moment;
s102: optimizing a fuzzy c-means clustering algorithm by adopting a simulated annealing algorithm and combining a genetic algorithm, and according to the plurality of eigenvectors ViObtaining an optimized clustering center vc；
S103: according to the optimized clustering center vcObtaining a clustering result;
s104: according to the clustering result, a boxplot method is adopted to carry out statistical analysis on four characteristics in each clustered class, and the high-frequency oscillation rhythm of the epileptic brain electrical signal is detected; the four characteristics in each class refer to the average singular value MSV and the line length f of each class after the clustering is obtainedlPower ratio R and spectral centroid fc。
2. The method for detecting the high-frequency oscillation rhythm based on the intelligent algorithm optimized fuzzy clustering, as claimed in claim 1, wherein: in step S101, the specific calculation steps of the average singular value MSV are as follows:
s201: obtaining a time-frequency information matrix S (tau, f) of the epileptic brain electrical signal time sequence x (t) by using S transformation, wherein the calculation formula is shown as formula (1):
in the above formula, f is frequency, j is an imaginary unit, and τ is a position parameter in time domain t;
s202: performing singular value decomposition on the obtained time-frequency information matrix S (tau, f) to obtain singular values, wherein a calculation formula is shown as a formula (2):
S(τ,f)＝UΓVT (2)
wherein S (tau, f) is M multiplied by N, and U and V are orthogonal matrixes of M multiplied by M, N multiplied by N respectively; Γ has a size of M × N and is in the form ofΛr×r＝diag(σ1,σ2,…σr)，σiThe singular values are called as time frequency information matrix S (tau, f), and r is the total number of the singular values;
s203: determining the effective values and the number of the obtained singular values by using a k-medoids algorithm;
s204: and calculating the average value of the effective singular values to obtain an average singular value MSV.
3. As claimed in claim 2, in step S203, the effective number of singular values obtained is determined by using a k-medoids algorithm, which comprises the following specific steps:
s301: singular value σ obtained from S202i(i ═ 1, 2.., r), o is selected1＝σ1And o2∈{σ2,σ3,…σrTaking the k-medoids as the center of the k-medoids algorithm;
s302: respectively calculating the remaining singular values to the center o according to the formula (3)1And o2Is a distance of
In the formula (3), sj∈{σ2,σ3,…,σrAnd sj≠o2(ii) a If d is1(sj,o1)≤d2(sj,o2) A 1 is tojIs assigned to1Centered cluster C1Performing the following steps; otherwise will sjIs assigned to2Centered cluster C2Performing the following steps;
s303: from cluster C2In (1), a value o is randomly selected2', as a clustering center, cluster C1Center in (A) is o1′＝o1(ii) a Calculating new o according to the method in S3021' and o2' Cluster C1' and C2'; and the following values were calculated:
wherein, if J is less than 0, use o2' as a New Cluster center instead of o2(ii) a Otherwise, the original center o is retained2；
S304: looping steps S302 and S303 until the newly generated cluster center does not change any more; containing sigma1The cluster is needed, and the obtained cluster number is q (q is more than or equal to 1 and less than or equal to r-1), namely q effective singular values sigmak(k＝1,2...,q)。
4. As claimed in claim 3, in step S204, the calculation formula of the average value of the valid singular values is as shown in equation (3):
in step S101, the line length flIs represented by equation (6):
in the formula (6), LlIs x (k) length, and x (k) is amplitude of epileptic brain electrical signal time sequence;
the calculation formula of the power ratio R is shown in formula (7):
in the formula (7), P[80-200]Is the power, P, of the suspected high frequency oscillation rhythm in ripple (ripple, 80-200 Hz) bandwidth[250-500]The power of the suspected high-frequency oscillation rhythm is within a fast ripple (250-500 Hz) bandwidth; the ripple bandwidth is 80-200 Hz; the fast ripple bandwidth is 250-500 Hz;
the frequency spectrum centroid fcIs represented by equation (8):
in equation (8), T is the sampling period, L is the window length, M (k) is the multi-window power spectral density estimate, andw (L) is a Hamming window of length L.
5. The method for detecting the high-frequency oscillation rhythm based on the intelligent algorithm optimized fuzzy clustering, as claimed in claim 1, wherein: in step S102, the specific steps of optimizing the fuzzy c-means clustering algorithm by using the simulated annealing algorithm in the intelligent algorithm in combination with the genetic algorithm are as follows:
s401: initializing control parameters: size of population individual NpMaximum number of evolutionary events NmaxCross probability PcProbability of variation PmAnnealing initiation temperature T0Coefficient of coolingtkTermination temperature Tend；
S402: randomly generating NpGroup initial clustering center vcC is 1,2, …, C is the number of clusters, and generates an initial population Chrom, and calculates a membership value mu of the ith eigenvector belonging to the C-th class according to the formula (9)ic：
In the formula (9), muicThe conditions are satisfied:c is the number of clusters and p is a weighted fuzzy parameter, usually taking the value 2, ViIs the extracted four-dimensional feature vector and is composed of the extracted average singular value, the line length, the power ratio and the spectrum centroid, i ═ 1,2, …, Nf，NfThe length of a preset four-dimensional feature vector; calculating to obtain a membership value muicThen, the j (j ═ 1, 2., N) is calculated according to the formula (8)p) Fitness f of group individualj：
S403: setting a cycle count variable gen as 0;
s404: performing selective crossover and mutation genetic operations on the population Chrom, and carrying out selection crossover and mutation genetic operations on new NpCalculating clustering center, membership degree and fitness value f 'of individual'j(ii) a If'j＞fjReplacing the new individual with the new individual and going to step S405; otherwise, with probabilityReceiving new individuals and discarding old individuals;
s405: if gen < NmaxIf so, go to S404; otherwise, go to S406;
s406: if Ti＜TendIf yes, the algorithm successfully returns the global optimal solution to the step S407; otherwise, executing a cooling operation Ti+1＝tkTiGo to S403;
s407: the resulting clustering center vcTo optimize the resulting parameters.
6. The method for detecting the high-frequency oscillation rhythm based on the intelligent algorithm optimized fuzzy clustering, as claimed in claim 5, wherein: in step S103, according to the optimization parameters, the specific steps of obtaining the clustering result are as follows:
s501: obtaining initialization parameters based on a fuzzy c-means clustering algorithm by adopting a simulated annealing algorithm and combining genetic algorithm optimization;
s502: calculating a four-dimensional feature vector V by equation (11) based on the initialization parametersiDegree of membership μ to class C (C ═ 1,2, …, C)iC：
In the above formula, vcRefers to the center of the cluster c, vkRefers to the center of cluster k.
S503: according to the obtained four-dimensional feature vector ViDegree of membership mu ofiCThe feature vector ViAnd classifying the cluster into the class to which the maximum membership degree belongs to obtain a clustering result.
7. The method for detecting the high-frequency oscillation rhythm based on the intelligent algorithm optimized fuzzy clustering, as claimed in claim 1, wherein: in step S104, the median and the quartile range of each of the four features in each of the clustered classes are obtained using the box plot, and the obtained median and quartile range are analyzed to determine the waveform state of the class varying with time, thereby detecting the high-frequency oscillation rhythm.
CN201910747243.4A 2019-08-14 2019-08-14 High-frequency oscillation rhythm detection method based on intelligent algorithm optimized fuzzy clustering Active CN110584600B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201910747243.4A CN110584600B (en) 2019-08-14 2019-08-14 High-frequency oscillation rhythm detection method based on intelligent algorithm optimized fuzzy clustering

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201910747243.4A CN110584600B (en) 2019-08-14 2019-08-14 High-frequency oscillation rhythm detection method based on intelligent algorithm optimized fuzzy clustering

Publications (2)

Publication Number Publication Date
CN110584600A true CN110584600A (en) 2019-12-20
CN110584600B CN110584600B (en) 2021-08-27

Family

ID=68854073

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201910747243.4A Active CN110584600B (en) 2019-08-14 2019-08-14 High-frequency oscillation rhythm detection method based on intelligent algorithm optimized fuzzy clustering

Country Status (1)

Country Link
CN (1) CN110584600B (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113499083A (en) * 2021-05-18 2021-10-15 生物岛实验室 Scalp electroencephalogram high-frequency oscillation automatic detection method
CN114159076A (en) * 2021-09-17 2022-03-11 天津大学 Electroencephalogram high-frequency oscillation signal automatic detection system for assisting epileptic focus positioning

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20120245481A1 (en) * 2011-02-18 2012-09-27 The Trustees Of The University Of Pennsylvania Method for automatic, unsupervised classification of high-frequency oscillations in physiological recordings
WO2013185775A1 (en) * 2012-06-12 2013-12-19 Ictalcare A/S Method and system of detecting seizures
CN107220708A (en) * 2017-06-23 2017-09-29 哈尔滨工业大学 Adaptive simulated annealing genetic algorithm for sleep cerebral electricity feature selecting by stages
CN108606781A (en) * 2018-05-23 2018-10-02 广东工业大学 A kind of epilepsy localization system based on the brain electricity higher-order of oscillation rhythm and pace of moving things
CN108875604A (en) * 2018-05-31 2018-11-23 中国地质大学（武汉） Higher-order of oscillation rhythm and pace of moving things detection method based on fuzzy clustering optimization gauss mixed model
CN108932476A (en) * 2018-05-31 2018-12-04 中国地质大学（武汉） Higher-order of oscillation rhythm and pace of moving things detection method based on quantization error optimization gauss mixed model

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20120245481A1 (en) * 2011-02-18 2012-09-27 The Trustees Of The University Of Pennsylvania Method for automatic, unsupervised classification of high-frequency oscillations in physiological recordings
WO2013185775A1 (en) * 2012-06-12 2013-12-19 Ictalcare A/S Method and system of detecting seizures
CN107220708A (en) * 2017-06-23 2017-09-29 哈尔滨工业大学 Adaptive simulated annealing genetic algorithm for sleep cerebral electricity feature selecting by stages
CN108606781A (en) * 2018-05-23 2018-10-02 广东工业大学 A kind of epilepsy localization system based on the brain electricity higher-order of oscillation rhythm and pace of moving things
CN108875604A (en) * 2018-05-31 2018-11-23 中国地质大学（武汉） Higher-order of oscillation rhythm and pace of moving things detection method based on fuzzy clustering optimization gauss mixed model
CN108932476A (en) * 2018-05-31 2018-12-04 中国地质大学（武汉） Higher-order of oscillation rhythm and pace of moving things detection method based on quantization error optimization gauss mixed model

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
O.L. SMART等: "Automatic detection of high frequency epileptiform oscillations from intracranial EEG recordings of patients with neocortical epilepsy", 《PROC. IEEE DENVER SECTION TECHNICAL, PROFESSIONAL AND STUDENT DEVELOPMENT WORKSHOP》 *
Z. FANG等: "Fast automatic localization of epileptic seizure onset zones using complex morlet wavelet transform-based singular value decomposition", 《PROC. 37TH CHINESE CONTROL CONF》 *
白莉媛等: "一种基于模拟退火和遗传算法的模糊聚类方法", 《计算机工程与应用》 *

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113499083A (en) * 2021-05-18 2021-10-15 生物岛实验室 Scalp electroencephalogram high-frequency oscillation automatic detection method
CN113499083B (en) * 2021-05-18 2023-10-20 生物岛实验室 Scalp electroencephalogram high-frequency oscillation automatic detection method
CN114159076A (en) * 2021-09-17 2022-03-11 天津大学 Electroencephalogram high-frequency oscillation signal automatic detection system for assisting epileptic focus positioning
CN114159076B (en) * 2021-09-17 2024-02-27 天津大学 Automatic detection system for electroencephalogram high-frequency oscillation signals for assisting epileptic focus positioning

Also Published As

Publication number Publication date
CN110584600B (en) 2021-08-27

Similar Documents

Publication Publication Date Title
George et al. Classification of epileptic EEG signals using PSO based artificial neural network and tunable-Q wavelet transform
US9326698B2 (en) Method for automatic, unsupervised classification of high-frequency oscillations in physiological recordings
Wang et al. Seizure prediction using directed transfer function and convolution neural network on intracranial EEG
Burrello et al. Hyperdimensional computing with local binary patterns: One-shot learning of seizure onset and identification of ictogenic brain regions using short-time iEEG recordings
CN108875604B (en) High-frequency oscillation rhythm detection method based on fuzzy clustering optimization Gaussian mixture model
Fraser et al. Automated biosignal quality analysis for electromyography using a one-class support vector machine
Ghassemi et al. Combined accelerometer and EMG analysis to differentiate essential tremor from Parkinson's disease
CN113786204B (en) Epileptic intracranial brain electrical signal early warning method based on deep convolution attention network
CN110876626A (en) Depression detection system based on optimal lead selection of multi-lead electroencephalogram
Sairamya et al. A new approach for automatic detection of focal EEG signals using wavelet packet decomposition and quad binary pattern method
CN110584598B (en) Time-frequency analysis based method for automatically separating peak activities in high-frequency oscillation rhythm
CN110584600B (en) High-frequency oscillation rhythm detection method based on intelligent algorithm optimized fuzzy clustering
WO2008057365A2 (en) Epileptic event detection systems
CN108932476B (en) High-frequency oscillation rhythm detection method based on quantization error optimization Gaussian mixture model
CN106264510B (en) Modeling method for screening surgical patients
CN112545513A (en) Music-induced electroencephalogram-based depression identification method
Quintero-Rincón et al. Multivariate Bayesian classification of epilepsy EEG signals
Bozkurt et al. Comparison of AR parametric methods with subspace-based methods for EMG signal classification using stand-alone and merged neural network models
Karthick et al. Could we have missed out the seizure onset: A study based on intracranial EEG
CN116807496B (en) Method, device, equipment and medium for positioning epileptic interval brain wave abnormal signals
US20050273286A1 (en) Methods for identifying brain nuclei from micro-electrode signals
Alipoor et al. Classification of EEG signals in four groups, including healthy subjects with open/closed eyes and epilepsy subjects with/without seizure by PSD estimate (using the multitaper method) and ANN
Cardona et al. NEUROZONE: On-line recognition of brain structures in stereotactic surgery-application to Parkinson's disease
CN113598792B (en) Epileptic electroencephalogram classification method based on supervised feature fusion algorithm
US20070167856A1 (en) Microelectrode recording analysis and visualization for improved target localization

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
GR01 Patent grant
GR01 Patent grant