KR101375673B1 - Method for warning of epileptic seizure using excitatory-inhibitory model based on the chaos neuron and electronic device supporting the same - Google Patents

Abstract

The present invention relates to a method for warning of an epileptic seizure and a terminal supporting the same, and the objective of the present invention is to enable the accurate and stable prediction and warning of an epileptic seizure using an excitatory/inhibitory population model based on chaos neurons. A method for warning an epileptic seizure disclosed in the present invention comprises the steps of: applying epileptic brain wave data to a chaos neuron-based excitatory-inhibitory (E-I) model and calculating an optimum parameter value for a connection coefficient; calculating a parameter in the E-I model for collected brain wave signals and comparing the calculated parameter with the optimum parameter value; and predicting an epileptic seizure based on the comparison result. The structure of a terminal supporting the method is also disclosed in the present invention. [Reference numerals] (110) Brain wave signal collection unit; (120) Input unit; (130) Alarm unit; (140) Display unit; (150) Storage unit; (151) Epileptic brain wave data; (160) System control unit

Description

{Method For warning of Epileptic seizure using Excitatory-Inhibitory model based on the Chaos Neuron and Electronic Device supporting the same}

The present invention relates to an epileptic seizure alert, and more particularly, to an epilepsy seizure alert method and a terminal supporting the seizure seizure prediction and warning using a chaotic neuron-based excitatory / inhibitory population model. It is about.

The number of patients with epilepsy is known to be more than 2 million in the United States, and 80% of these patients are reported to be able to control seizures by medicine and surgery. However, the remaining 25-30% of patients continue to have seizures. In the UK, there are 600,000 people with epilepsy. About 500 of these people are injured by sudden seizures and are killed by such injuries. Although there are no accurate statistics in Korea, it is estimated that there are about 300,000 to 400,000 people with liver disease.

An epileptic seizure causes a change in behavior or consciousness, such as a loss of consciousness or shaking arms and legs for a few seconds or minutes during a seizure. Epilepsy is a disorder of the brain that is usually abnormally active. Epilepsy involves seizures such as feelings of low feelings, emotions, or behaviors, or sometimes a race, spasms, or loss of consciousness. In some patients with epilepsy, seizures can occur only occasionally. But for others, it can happen more than 100 times a day. These seizures vary in number of seizures according to individual differences, and there is a difference in the risk of seizures. If a person with epileptic seizures has diseases such as hypoxia (chronic obstructive pulmonary disease, severe asthma), meningitis (meningitis), encephalitis, or brain tumor, the risk of seizures is higher.

Electroencephalogram (EEG) is a measurement of current flow in a living body caused by synchronized activity of nerve cells occurring on the surface of the brain cortex using an electrode attached to the scalp. These EEG can be non-invasive estimation of the activity information of the cranial nerve cells through the analysis of EEG as an electrical signal. Brain function research using EEG has various advantages compared to other brain function mapping methods because it has a simple measurement method, the least environmental constraints, low system configuration, and excellent time resolution. It is used as the most basic research tool in scientific research.

Epilepsy data analysis using the above-described brain waves is essential for epilepsy diagnosis, seizure detection and prediction. The main purpose of this epilepsy EEG analysis is to accurately diagnose epilepsy using EEG data and to reduce the risk of seizures in epilepsy patients. For this purpose, Tensor analysis or Dynamics System operation was applied for analysis. However, the aforementioned methods are for diagnosing diseases through data classification and performing data learning for application. In this case, the methods require a large amount of learning time and amount of data for data learning. In addition, when a new epilepsy pattern occurs, the above methods have a problem in that it is difficult to detect a new seizure pattern because only the patterns similar to the learned seizure patterns are detected.

Accordingly, an object of the present invention is to provide a warning model for predicting and diagnosing epileptic seizures based on pre-stored epileptic EEG data and analyzing the specificity of EEG data. It is to provide a supporting terminal.

In particular, the present invention provides an epileptic seizure warning method and terminal capable of more appropriately performing epileptic seizure prediction and providing an alarm by estimating epileptic seizures using a chaotic neuron model composed of an excitatory and an inhibitory subpopulation. To provide.

In order to achieve the above object, the present invention applies the epileptic EEG data to the chaotic neuron-based Excitatory-Inhibitory (EI) model while calculating the optimal parameter value for the coupling coefficient, the parameters in the EI model for the collected EEG signals Disclosed is a configuration of an epileptic seizure warning method based on a chaotic neuron model, the method comprising calculating and comparing the optimal parameter value and predicting epileptic seizure according to the comparison result.

The method may further include an alarm generating process of generating a predefined alarm when a parameter corresponding to an epileptic seizure is detected as a result of the comparison, wherein the alarm generating process is based on detection of a parameter corresponding to the epileptic seizure. It may include the step of performing each step alarm generation for the precursor phenomena.

The calculating process includes calculating an optimal parameter by applying LMA (Levenberg-Marquardt Algorithm) to a range of parameters of the E-I model.

The present invention also provides an input unit for receiving interstitial EEG data, a system control unit for applying an interstitial brain wave data to an EI (Excitatory-Inhibitory) model based on a chaotic neuron, and calculating an optimal parameter value for a coupling coefficient, an EEG signal of a wearer The EEG signal collection unit for collecting the, wherein the system control unit calculates the parameters in the EI model for the collected EEG signal, and compares with the optimal parameter value chaos characterized in that predicting epileptic seizure according to the comparison result Disclosed is a configuration of a neuron model based epileptic seizure warning support terminal.

The terminal further includes an alarm unit that generates a predefined alarm when a parameter corresponding to an epileptic seizure is detected, wherein the alarm unit is configured to detect a plurality of precursor phenomena based on detection of parameters corresponding to the epileptic seizure. Perform each stage alarm.

Meanwhile, the system controller may calculate the optimum parameter by applying LMA (Levenberg-Marquardt Algorithm) to a range of parameters of the E-I model to calculate the optimal parameter.

As described above, according to the epilepsy seizure warning method and a terminal supporting the same, the present invention supports stable and timely epilepsy seizure prediction by predicting epileptic seizures based on specificity for epileptic seizures.

Accordingly, the present invention assists in securing the safety of the patient by minimizing the threat of epileptic seizures so as to prepare for an accident or a risk due to an epileptic seizure.

1 is a view schematically showing the configuration of an epileptic seizure alarm terminal according to an embodiment of the present invention.
2 is a more detailed view of the system control unit configuration of FIG. 1;
3 is a diagram for explaining a chaotic neuron-based EI model.
4 is a diagram simulating epilepsy brain wave data using an EI model.
5 is a view for explaining a parameter range for epileptic seizure warning in the EI model according to the present invention.
6 is a view for explaining the epileptic seizure warning method by the EI model parameters according to the present invention.
7 is a view for explaining the operation of the epileptic seizure warning terminal of the present invention.

In the following description, only parts necessary for understanding the embodiments of the present invention will be described, and the description of other parts will be omitted so as not to obscure the gist of the present invention.

The terms and words used in the present specification and claims should not be construed as limited to ordinary or dictionary meanings and the inventor is not limited to the meaning of the terms in order to describe his invention in the best way. It should be interpreted as meaning and concept consistent with the technical idea of the present invention. Therefore, the embodiments described in the present specification and the configurations shown in the drawings are merely preferred embodiments of the present invention, and are not intended to represent all of the technical ideas of the present invention, so that various equivalents And variations are possible.

Hereinafter, with reference to the accompanying drawings will be described in detail an embodiment of the present invention.

The present invention described below provides an E-I (excited and inhibitory population) model for detecting and proactively warning epileptic seizures from epileptic EEG time series data. In particular, the present invention is for predicting and guiding epileptic seizures using a chaotic neuron-based E-I (excited and inhibitory population) model. Although epilepsy EEG data with nonlinear characteristics are characterized by a precursor to seizures before epilepsy occurs, most patients do not feel or ignore this prognostic phenomenon because of the seizure that actually occurs after the precursor. Are often in danger. On the other hand, the conventional epileptic seizure detection and prediction techniques are mostly predicted through the change pattern of the past data, it was almost impossible to detect and predict the change pattern that did not appear in the past. Therefore, there is a need for a method for improving the detection and prediction rate of epileptic seizure data having complex characteristics, and the present invention provides a technical idea that meets such needs.

The present invention described below simulates a real signal using time series epilepsy brain wave data having complex features, and uses a parameter obtained in the simulation process to warn a patient or a specialist in advance about an epileptic seizure before epileptic seizures. To help.

1 is a block diagram schematically showing the configuration of the epileptic seizure warning terminal 100 according to an embodiment of the present invention, Figure 2 is a view showing in more detail the configuration of the system control unit 160 of the epileptic seizure warning terminal 100 configuration to be.

First, referring to FIG. 1, an epileptic seizure warning terminal 100 of the present invention includes an EEG signal collecting unit 110, an input unit 120, an alarm unit 130, a display unit 140, a storage unit 150, and a system control unit. It includes the configuration of 160.

The epileptic seizure warning terminal 100 of the present invention including such a configuration simplifies the biological neurotransmission process and supports detecting and predicting epileptic seizures using a chaotic neuron model, which is a mathematical interpretation model, as a kind of neuron network. do. In particular, the epileptic seizure warning terminal 100 of the present invention enables diagnosis and prediction of epileptic seizure timing based on an Excitatory subpopulation-Inhibitory subpopulation (E-I) model composed of an excitatory / inhibitory population through selection of an optimal E-I model parameter value. At this time, the epileptic seizure warning terminal 100 of the present invention estimates an optimal parameter value from the parameters of the chaos-based E-I model using LMA (Levenberg-Marquardt Algorithm) in order to consider the complex characteristics of time-series epilepsy brain wave data. In addition, the epileptic seizure warning terminal 100 of the present invention may support excellent diagnosis and prediction of epileptic seizures by applying the estimated optimal parameter to a method for distinguishing a general signal from an epileptic seizure signal.

The EEG signal collecting unit 110 is a configuration for collecting the EEG signal of the wearer. The EEG signal collection unit 110 may be configured to include a sensor attached to the scalp of the epilepsy patient to collect electrical signals corresponding to the EEG, and a power supply unit for supplying power to the sensor. The EEG signal collection unit 110 may be provided in various accessories such as a headband or a hairband, earrings that can be continuously attached to the scalp of the wearer. In addition, the EEG signal collection unit 110 provides the collected EEG to the system control unit 160. The epileptic seizure warning terminal 100 of the present invention may be configured to include a probe unit and a control module connected to the probe unit to determine whether an epileptic seizure state is detected by analyzing an EEG signal provided by the probe unit. Accordingly, the EEG signal collection unit 110 may be provided in a configuration corresponding to the probe unit, the configuration described below may be provided in the form included in the control module. That is, the control module includes an alarm unit 130, a storage unit 150, and a system control unit 160, and may include an input unit 120 and a display unit 140 according to a designer's method.

The input unit 120 is configured to generate an input signal for operating the epileptic seizure warning terminal 100. For example, the input unit 120 may include a power button for controlling the operation of the epileptic seizure warning terminal 100, an adjustment button for adjusting the volume of the alarm unit 130, and the like. Meanwhile, when the configuration of the display unit 140 to be described below is provided in the form of a touch screen having an input function, the display unit 140 may be included in the configuration of the input unit 120. In particular, the input unit 120 supports time series epilepsy brain wave data 151 to be input from the outside. The time series epilepsy brain wave data 151 input through the input unit 120 may be temporarily stored in the storage unit 150, and the system controller 160 may operate the same to calculate an optimal parameter for distinguishing an epileptic seizure. .

The alarm unit 130 includes at least one information output means such as a lamp, a vibration, a speaker, and the like. The alarm unit 130 outputs predefined information when an EEG signal collection corresponding to an epileptic seizure occurs under the control of the system controller 160. For example, the alarm unit 130 may perform at least one of outputs such as blinking a lamp, a vibration output of a predetermined pattern, a predefined guide sound, and the like. The alarm unit 130 may continuously output the corresponding alarm while the epilepsy patient continues the epilepsy state, and may support to automatically stop the alarm output under the control of the system control unit 160 when the epileptic seizure is released. have. Accordingly, the alarm unit 130 may support the neighboring neighbors to recognize the situation in which the epilepsy patient is likely to have epileptic seizures in a specific environment or the situation in which the epilepsy is seized to maintain the epileptic seizure state and take appropriate measures. .

The display unit 140 is configured to output various information necessary for operating the epileptic seizure warning terminal 100. The display unit 140 may indicate information indicating the power supply status of the epileptic seizure warning terminal 100, information on the EEG signals collected in real time or at regular intervals, and whether the collected EEG signals are normal normal signals or epileptic seizure precursor signals. The classified information and the information for guiding the alarm output status can be output. The display unit 140 may display various virtual key buttons for manipulating the epileptic seizure warning terminal 100, and may be configured to have a touch function for selecting the corresponding virtual key button. That is, the display unit 140 may be configured as a touch screen.

The storage unit 150 is a component for storing a program and various data for performing an epileptic seizure warning. In particular, the storage unit 150 may store various epilepsy brain wave data 151 having feature points for epileptic seizures. The storage unit 150 provides the epilepsy brain wave data 151 to the system controller 160 in response to the request of the system controller 160. The system controller 160 can then determine whether the epilepsy brain wave data 151 is in an epilepsy state by comparing it with the currently generated brain wave signal. The epilepsy brain wave data 151 may be input through the input unit 120 as described above. Alternatively, it may be stored and provided in advance when the epileptic seizure warning terminal 100 is manufactured.

The system controller 160 is configured to transmit various control signals for controlling the operation of the epileptic seizure warning terminal 100 and to process data. In particular, the system controller 160 calculates the chaotic neuron-based E-I model parameters for the epilepsy brain wave data 151 and estimates the optimal parameters to be used to distinguish epileptic seizures. To this end, the system controller 160 generates a chaotic neuron-based EI model for the epileptic EEG data 151 and calculates an optimal parameter value among the parameters applied to the generated EI model, as shown in FIG. 2. The optimal parameter selector 161 may include a chaos neuron-based EI model operator 163 that separates the currently collected EEG signals as a general signal and an epileptic seizure signal and controls an alarm alarm. The system controller 160 having such a configuration reproduces the potential activity of the neuron most similarly, implements an excitatory-suppressive neuron population model, and performs optimal parameter estimation and application corresponding to the coupling coefficient values in the EI model. When the state of the EEG signal collector 110 wearer is predicted as a potential step of the epileptic seizure state, it supports to output an alarm for this through the alarm unit 130.

,(여기서

)와 억제성 뉴런의 부분 모집단의 수

,(여기서

)는 시간 t에서 수학식 1과 수학식 2와 같이 나타낼 수 있다. The chaotic neuron model, on the other hand, is structured as a synaptic-excited excitatory and inhibitory population in a chaotic neural network consisting of four synapses and two chaotic neurons. In particular, one chaotic neuron model consists of excitatory and inhibitory subpopulations and the number of excitatory neuron subpopulations.

,(here

) And the number of subpopulations of inhibitory neurons

,(here

) Can be expressed as Equation 1 and Equation 2 at time t.

[Equation 1]

&Quot; (2) "

는 비선형 데이터의 시그모드 함수이다.

는 흥분성 부분 모집단을 시간 t로 미분한 것으로 흥분성 부분 모집단의 전위활동을 나타내며,

는 억제성 부분 모집단을 시간 t로 미분한 값으로 억제성 전위 활동을 나타낸다. 상술한 카오스 뉴런 집단은 4개의 시냅스에 대한 결합 계수

를 통해 연결되며, 결합 계수

는 시냅스의 연결 강도(가중치)를 의미한다. 또한 카오스 뉴런 집단은 흥분성 부분 모집단의 외부 입력 값

를 포함한다. 여기서, 결합 계수

의 시냅스 연결이 끊어지면, 모든 활동 전위 값은 기하급수적으로 0에 접근된다.

와

는 임계 문턱 값으로 충분히 큰 값으로 결정된다. 카오스 뉴런 모델에서 두 개의 부분 모집단을 연결하는 4개의 시냅스 연결 계수들은

,

,

,

으로 표현 가능하며, 모든 쌍들은 0이 아닌 값을 갖는다. In Equation 1, t means a data measurement time point,

Is a sig mode function of the nonlinear data.

Is the derivative of the excitatory subpopulation by time t, representing the potential activity of the excitatory subpopulation,

Denotes inhibitory translocation activity with the derivative of the inhibitory subpopulation by time t. The chaotic neuron population described above has binding coefficients for four synapses.

Connected through, the coupling coefficient

Denotes the strength (weight) of the synapse. In addition, chaotic neuron populations have external input values of excitatory subpopulations.

. Where coupling factor

When a synaptic connection is lost, all action potential values approach exponentially zero.

Wow

Is determined as a sufficiently large value as the threshold threshold. Four synaptic coupling coefficients linking two subpopulations in the chaotic neuron model

,

,

,

All pairs have nonzero values.

The chaotic neuron model described above can be simplified to the following equations (3) and (4) by the average value of the subpopulations of excitability and suppression.

&Quot; (3) "

&Quot; (4) "

In the present invention, the epileptic seizure is predicted and alarmed based on epileptic seizure data having nonlinear characteristics. Therefore, the basic chaotic neuron model is simplified as shown in FIG. The simplified model is provided with excitatory-suppressive neuron populations, which can be re-expressed in equations (5) and (6).

&Quot; (5) "

&Quot; (6) "

On the other hand, in the excitatory-suppressive neuron population, the value of the linkage coefficient of the subpopulation has an important effect on the action potential measurement. Therefore, the epileptic seizure warning terminal 100 of the present invention uses Levenberg-Marquardt Algorithm (LMA) to find the optimal subpopulation connection coefficient. Levenberg-Marquardt Algorithm (LMA) is the method proposed by Levenberg and Marquardt and is also called Gauss-Newton method. The LMA method, which is one of the optimization methods of nonlinear least-squares problems, is one of the effective methods for finding solutions in nonlinear data. The iterative LMA method is a method of finding a solution while repeatedly reducing the error by appropriately obtaining a step size from an initial estimate of Equation 7 described below.

[Equation 7]

는 초기 추정치 벡터,

은 스텝크기이며, 이때, 스텝 크기

을 구하는 방법은 아래의 수학식 8을 이용한다. 초기 추정치 벡터의 경우 각 환자별 간질 발작의 이력에 따라 결정되거나, 임의적이며 평균적으로 산정될 수 있다.In Equation (7)

Is the initial estimate vector,

Is the step size, where step size

To obtain the method using Equation 8 below. The initial estimate vector can be determined according to the history of epileptic seizures in each patient or can be estimated arbitrarily and on average.

&Quot; (8) "

는 Jacobian 행렬로

원소는

이며,

(추정치 수),

(측정치 수)이며,

는 에러함수이다. Levenberg-Marquardt Algorithm는 수학식 9와 같다. In Equation (8)

Is the Jacobian matrix

Element

Lt;

(Estimated number),

(Number of measurements),

Is an error function. Levenberg-Marquardt Algorithm is represented by Equation 9.

&Quot; (9) "

는 damping 파라미터,

은 스텝크기,

는 단위행렬이다.

가 무한대로 증가함에 따라,

는 영 벡터(a vector of zeros) 혹은 급격한 하강 방향으로 향한다. 즉

가 충분히 큰 경우 수학식 10과 같은 식이 성립한다. In Equation (9)

Is the damping parameter,

Silver step size,

Is the unit matrix.

As increases to infinity,

Is directed in the direction of a vector of zeros or in a sharp downward direction. In other words

If is sufficiently large, an equation such as Equation 10 holds.

&Quot; (10) "

의 크기를 제어하는 방법이다. 그러므로

의 값이 큰 경우에는 추정치가 해로부터 멀리 떨어진 경우에 해당되며, 수렴 시간은 길어지지만 적절히 작은 스텝으로 해에 대한 정확한 그레디언트 방향을 찾는다. 반대로

값이 아주 작은 경우에는 수학식 11과 같이 추정치가 해와 근접했을 때 반복의 마지막 단계에서 적합한 해를 효과적으로 찾을 수 있게 된다. The main difficult part of the application of the Levenberg-Marquardt method is

Is how to control the size. therefore

If the value of is large, then the estimate is far from the solution, and the convergence time will be longer, but in a small step, find the correct gradient direction for the solution. Contrary

If the value is very small, we can effectively find a suitable solution at the end of the iteration when the estimate is close to the solution (Equation 11).

&Quot; (11) "

값은 수학식 12를 사용하여 대각 원소의 평균값과 에러 값의 제곱에 비례하도록 업데이트를 한다. From Levenberg-Marquardt

The value is updated to be proportional to the square of the mean value and the error value of the diagonal element using Equation 12.

&Quot; (12) "

는 상수,

은 반복에서 추정치에 따른 계산치와 측정치의 차이이며,

은 대각원소의 개수이다. 따라서 에러 값이 커지면

값을 키워서 최속 강하법과 유사한 형태로 스텝 크기를 줄여서 해의 방향 쪽으로 값을 조절한다. In Equation (12)

Is a constant,

Is the difference between the calculated and measured values for the estimates in the iteration,

Is the number of diagonal elements. So if the error value increases

Adjust the value toward the solution direction by increasing the value and reducing the step size in a similar way to the fastest descent.

4 is a diagram illustrating a simulation of epilepsy EEG data based on the E-I model, Figure 5 is a view for explaining a parameter range for the epileptic seizure warning in the E-I model of the present invention. 6 is a view for explaining an epileptic seizure warning method according to a model parameter in the present invention.

The epileptic seizure warning terminal 100 of the present invention warns of epileptic seizures in advance using an optimal value of a parameter (synapse) connecting each population in an E-I model based on a chaotic neuron model. In this process, as shown in FIG. 4, when the epileptic seizure warning terminal 100 is provided with an EEG signal, the evaluation signal may be generated by performing a gray box analysis on the EEG signal.

The parameters of the model must be chosen correctly because the accuracy of the epileptic seizure warning is determined by detecting the optimal parameters in the E-I (excited and inhibitory population) model. That is, in the case of an optimal parameter for epileptic seizures, as shown in FIG. 5, an optimal value of parameter 1 (C1) has a value greater than 5, and parameter 2 (C2) has a value less than zero. In addition, parameter 3 (C3) has a value smaller than 10, and parameter 4 (C4) has a value within a range similar to -4.

In FIG. 6, an alarm point may be defined based on data on an epileptic seizure. That is, when an EEG signal or an evaluation signal evaluating an EEG signal is input as in "Input", the signal is classified by a certain window, and the divided values are divided into "Ictal Region" and "Interictal Region" as shown in the following table. Classify as classified. It can be assumed that substantial epileptic seizures have occurred since the 24th window in the entire window. Looking at the data shown, signal detection belonging to the "Ictal Region" occurs in the first fifth window. The epileptic seizure warning terminal 100 may output the first alarm through the alarm unit 130 when such signal detection occurs in the EEG signal detection process. In particular, the epileptic seizure alarm terminal 100 may output the corresponding alarm to provide that the alarm is an initial alarm. And when the EEG signal is detected values belonging to the "Ictal Region", the epileptic seizure warning terminal 100 may request a patient or a neighbor to the seizure preparation by guiding the alarm for additional epileptic seizures. In this process, the epileptic seizure information terminal 100 performs optimal parameter estimation on the coupling coefficients based on the chaotic time neuron EI model for the collected EEG signals, and checks the variation of the corresponding optimal parameter values for the epileptic seizure. Alarm You can perform an alarm. In particular, the epileptic seizure warning terminal 100 of the present invention may support to output an alarm for the possibility of epileptic seizure when at least one of the optimal parameters deviates from a predetermined predetermined range value as described above. It is possible to assign certain steps to each of a plurality of times and to support performing different alarm generations for each step.

7 is a view for explaining a method of operating an epileptic seizure warning terminal according to an embodiment of the present invention.

Referring to FIG. 7, the epilepsy seizure warning terminal operating method of the present invention first collects epilepsy brain wave data 151. The epilepsy brain wave data 151 may be collected during epileptic seizures of various epileptic patients. In operation S103, the system controller 160 may perform the EEG signal processing. In particular, the system controller 160 may perform gray box analysis on the received epilepsy brain wave data 151, and calculate a range of parameters corresponding to the coupling coefficients according to the chaotic neuron-based E-I model operation described above. Next, the system controller 160 applies the LMA in the calculated range of parameters to perform the optimal parameter selection in step S105. In other words, the epileptic seizure warning terminal 100 calculates a range of coupling coefficient values of the chaotic neuron-based EI model based on the epileptic EEG data 151 and performs optimal parameter selection based on the normal normal signal and epileptic seizure signal for EEG. Establish a reference point to identify.

Next, the EEG signal collection unit 110 corresponding to the probe unit of the epileptic seizure warning terminal 100 is worn at the point where the EEG signal of the patient can be detected and supports to collect the EEG signal in step S107. To this end, the EEG signal collection unit 110 may be disposed at a point where the EEG signal collection of the epilepsy patient is easily provided in a specific accessory form to easily support the EEG signal collection of the epilepsy patient.

When the EEG signal collection of the wearer is provided, the system controller 160 performs an E-I model operation on the collected EEG signal, and compares the parameter values corresponding to the corresponding connection coefficients with the predefined optimal parameter values.

Thereafter, the system controller 160 checks whether there is a parameter value corresponding to the epileptic seizure precursor in step S111. In this step, the system controller 160 determines that the parameter values in the chaotic neuron-based EI model of the collected EEG signals are an epileptic seizure precursor phenomenon, and in step S113, The alarm unit 130 is controlled to perform an alarm. In this case, the system controller 160 may support to output a signal divided by time series such as an initial alert, a mid-term alert, or a seizure alert for an epileptic seizure.

On the other hand, if there is no parameter corresponding to the epileptic seizure precursor phenomenon in step S111, the system control unit 160 checks whether there is an input signal for terminating the operation of the epileptic seizure warning terminal 100, and separately. If there is no end input of S107, the process may be branched to step S107 and the next process may be performed again.

As described above, feature points are extracted from epilepsy EEG data according to an epilepsy seizure warning method and a terminal supporting the seizure, and based on the analysis, the EEG signal of the operator is predicted to predict an accurate epileptic seizure and alarm To help.

The present invention can increase the clinical utility by distinguishing the type of epileptic seizures and developing a treatment or treatment to stop the seizure, thereby enabling the creation of socioeconomic benefits. In addition, since the present invention can be provided as a miniaturized system that can automatically detect and predict epileptic seizures automatically, economic and industrial development and international competitiveness can be strengthened according to the production and sales of the system, and contributing to the improvement of public welfare. Do. In addition, as an application field of the present invention, a new paradigm can be accelerated by providing a new paradigm such as the intelligent growth robot industry, the IT industry, and the game entertainment industry through human brain analysis.

It should be noted that the embodiments disclosed in the present specification and drawings are only illustrative of specific examples for the purpose of understanding, and are not intended to limit the scope of the present invention. It will be apparent to those skilled in the art that other modifications based on the technical idea of the present invention are possible in addition to the embodiments disclosed herein.

100: epileptic seizure alarm terminal
110: EEG signal collection unit
120: Input unit
130: alarm unit
140: display unit
150:
160: system control unit

Claims (8)

Calculating the optimal parameter value for the coupling coefficient while applying the epileptic EEG data to the chaotic neuron-based Excitatory-Inhibitory (EI) model;
Calculating parameters in the EI model for the collected EEG signals and comparing them with optimal parameter values;
Predicting epileptic seizures according to the comparison result;
Epilepsy seizure warning method based on a chaotic neuron model comprising a. The method of claim 1,
An alarm generation process of generating a predefined alarm when a parameter corresponding to an epileptic seizure is detected as a result of the comparison;
Epilepsy seizure warning method based on a chaotic neuron model, further comprising a. 3. The method of claim 2,
The alarm generation process
Performing each stage of alarm generation for a plurality of precursor phenomena based on detection of parameters corresponding to the epileptic seizure;
Epilepsy seizure warning method based on a chaotic neuron model comprising a. The method of claim 1,
The calculation process
Calculating an optimal parameter by applying LMA (Levenberg-Marquardt Algorithm) to a range of parameters of the EI model;
Epilepsy seizure warning method based on a chaotic neuron model comprising a. An input unit for receiving interstitial brain wave data;
A system controller which applies the epilepsy brain wave data to a chaotic neuron-based Excitatory-Inhibitory (EI) model and calculates an optimal parameter value for a coupling coefficient;
Includes; EEG signal collection unit for collecting the EEG signal of the wearer,
The system control unit
The epilepsy seizure warning support terminal of the chaotic neuron model, characterized in that for calculating the parameters in the EI model for the collected EEG signal to compare with the optimal parameter value and predict epileptic seizure according to the comparison result. 6. The method of claim 5,
An alarm unit for generating a predefined alarm when a parameter corresponding to an epileptic seizure is detected as a result of the comparison;
Epilepsy seizure warning support terminal based on the chaotic neuron model, further comprising a. The method according to claim 6,
The alarm unit
An epilepsy seizure warning support terminal based on a chaotic neuron model, characterized in that for generating alarms for each stage based on detection of parameters corresponding to the epileptic seizure. 6. The method of claim 5,
The system control unit
An apparatus for supporting epileptic seizure warning based on a chaotic neuron model, wherein the optimal parameter is calculated by applying LMA (Levenberg-Marquardt Algorithm) to a range of parameters of the EI model to calculate the optimal parameter.

Images