Electroencephalogram signal detection method and epilepsy detection system based on abnormal discharge waveform

Abstract

The invention discloses an electroencephalogram signal detection method based on abnormal discharge waveforms and an epilepsy detection system. The method comprises the following steps: obtaining a plurality of abnormal discharge waveforms; performing morphological clustering to obtain abnormal discharge templates, namely, classifying a plurality of abnormal discharge templates according to correlation coefficients between cluster centroids of all abnormal discharge waveforms; screening abnormal matching templates from the plurality of abnormal discharge templates, namely, detecting distribution according to abnormal discharge, and screening out a plurality of abnormal discharge templates as the abnormal matching templates; and detecting an online electroencephalogram signal according to the abnormal matching template. A plurality of abnormal discharge waveforms are obtained, and then morphological clustering is performed on the abnormal discharge waveforms to obtain an abnormal discharge template, so that frequent modification in a feature extraction process can be avoided, and the generalization ability of clustering is improved; and according to the abnormal discharge detection distribution, the optimal n abnormal discharge templates are screened as the abnormal matching templates, so that the detection accuracy can be further improved.

Description

Technical Field

[0001] The present invention relates to the technical field of electroencephalogram (EEG) signal detection, and in particular to an EEG signal detection method based on abnormal discharge waveform and an epilepsy detection system. Background Art

[0002] Epilepsy is a chronic brain disease characterized by repeated epileptic seizures. It is caused by abnormal discharges of brain neurons, and the onset of the disease is characterized by repetitiveness and transience. EEG signals contain a wealth of information about brain activity. High-frequency oscillation signals are EEG activities under high gamma rhythms (>80Hz). High-frequency oscillation signals are closely related to epilepsy and are a reliable biomarker for epilepsy. They can reflect the severity of epilepsy and are helpful in evaluating the efficacy of epilepsy treatment, judging epilepsy susceptibility, and predicting epileptic seizures. Therefore, the detection of EEG signals can predict the onset of epilepsy to a certain extent.

[0003] In related technologies, EEG signal detection methods such as nonlinear dynamic models, EEG complexity, and abnormal discharge (spike) frequency are used to predict epileptic seizures. During the training process, such methods use all EEG data including physiological spikes and pathological spikes to generate features. However, due to the large differences in EEG discharge waveforms of different detection objects at different times, pathological spikes related to seizures cannot be accurately retained, resulting in insufficient accuracy of input features and low accuracy in identifying epileptic seizures. Summary of the invention

[0004] In view of the deficiencies in the prior art, the present invention provides an EEG signal detection method based on abnormal discharge waveform and an epilepsy detection system.

[0005] The technical solution adopted by the present invention to solve its technical problem is:

[0006] In a first aspect, the present invention provides an electroencephalogram signal detection method, comprising:

[0007] Obtaining multiple abnormal discharge waveforms;

[0008] The abnormal discharge template is obtained by morphological clustering, that is, multiple abnormal discharge templates are classified according to the correlation coefficient between the cluster centroids of each abnormal discharge waveform;

[0009] Screening abnormal matching templates from multiple abnormal discharge templates, that is, screening out several abnormal discharge templates as abnormal matching templates according to abnormal discharge detection distribution;

[0010] Online EEG signal detection based on anomaly matching templates.

[0011] In one example, classifying a plurality of abnormal discharge templates includes:

[0012] Obtaining existing clusters based on abnormal discharge waveforms;

[0013] Perform sliding correlation between the abnormal discharge waveform after the existing cluster and the cluster centroid of the existing cluster, record the sliding correlation coefficient during the sliding process, and use the maximum sliding correlation coefficient as the correlation coefficient between the abnormal discharge waveform and the cluster centroid; if the correlation coefficient is greater than the set threshold, the abnormal discharge waveform is included in the existing cluster; otherwise, the abnormal discharge waveform is used as a new cluster;

[0014] Iterate until all abnormal discharge waveforms are classified.

[0015] In one example, when multiple correlation coefficients between an abnormal discharge waveform and a cluster centroid are greater than a set threshold, the abnormal discharge waveform is classified into an existing cluster with the largest correlation coefficient.

[0016] In one example, the sliding correlation includes:

[0017] The next abnormal discharge waveform and the abnormal discharge waveform corresponding to the cluster centroid are slid toward each other, and the step length of each sliding is one sampling interval; each time the sliding is performed, the sliding correlation coefficient between the overlapping part of the next abnormal discharge waveform and the abnormal discharge waveform corresponding to the cluster centroid and the cluster centroid is calculated.

[0018] In one example, a plurality of abnormal discharge waveforms are obtained based on energy features of historical EEG signals, wherein the energy features of the historical EEG signals include: global discharge coefficient, line length, and envelope.

[0019] In one example, selecting an abnormal matching template from a plurality of abnormal discharge templates according to the abnormal discharge detection distribution includes:

[0020] Eliminate abnormal discharge templates whose number of abnormal discharges detected before the onset moment is less than the set threshold;

[0021] Among the remaining abnormal discharge templates, the median distance between the abnormal discharge detection time and the onset time of each abnormal discharge template is calculated, the median distances are arranged in ascending order, and the first n abnormal discharge templates are selected as abnormal matching templates.

[0022] In one example, detecting online EEG signals according to an abnormal matching template includes:

[0023] The historical EEG signals are detected using the abnormal matching templates, and the abnormal discharge features identified by the abnormal matching templates are counted; the classifier is trained using the abnormal discharge features, and the trained classifier is used for real-time detection of EEG signals.

[0024] In one example, the abnormal discharge features include: the number of abnormal discharges identified by each abnormal matching template and its proportion in the total number.

[0025] In one example, the moment when the energy characteristic of the historical EEG signal exceeds a set threshold is marked as abnormal discharge; and the data in a period of time before and after the abnormal discharge moment is intercepted as the abnormal discharge waveform.

[0026] In a second aspect, the present invention provides an epilepsy detection system, comprising:

[0027] A processor, running the EEG signal detection method to obtain a detection result at the moment of epileptic seizure;

[0028] The classifier is used to output the detection result.

[0029] In a third aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of the EEG signal detection method.

[0030] The beneficial effect of the present invention is that the EEG signal detection method based on abnormal discharge waveform and the epilepsy detection system of the present invention can avoid frequent modification of the feature extraction process and improve the generalization ability of clustering by acquiring multiple abnormal discharge waveforms and then performing morphological clustering on the abnormal discharge waveforms to obtain abnormal discharge templates; and then select the optimal n abnormal discharge templates as abnormal matching templates according to the abnormal discharge detection distribution, which can further improve the detection accuracy. This method can be used not only for early warning detection before epileptic seizures, but also for evaluation of treatment effects. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] The present invention is further described below in conjunction with the accompanying drawings and embodiments.

[0032] Figure 1 It is a flow chart of the EEG signal detection method of the present invention.

[0033] Figure 2 It is a schematic diagram of abnormal discharge moment identification of the present invention.

[0034] Figure 3 It is a schematic diagram related to sliding of the present invention.

[0035] Figure 4 It is a schematic diagram showing that the abnormal discharge detection time distribution of the present invention has no obvious tendency.

[0036] Figure 5 This is a schematic diagram showing that the abnormal discharge detection time distribution of the present invention is concentrated near the onset time.

[0037] Figure 6It is the distribution of the proportion of discharges recognized by n abnormal matching templates of the present invention.

[0038] Figure 7 It is a schematic diagram of the confusion matrix of the classification results of the present invention.

[0039] Figure 8 It is a comparison result diagram of the number of abnormal detections before and after the treatment of the present invention. DETAILED DESCRIPTION

[0040] The present invention will now be described in further detail with reference to the accompanying drawings. These drawings are simplified schematic diagrams, which only illustrate the basic structure of the present invention in a schematic manner, and therefore only show the components related to the present invention.

[0041] In the description of the present invention, it should be understood that the terms "center", "longitudinal", "lateral", "length", "width", "thickness", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", "clockwise", "counterclockwise", "axial", "radial", "circumferential" and the like indicate positions or positional relationships based on the positions or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present invention. In addition, features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of the present invention, unless otherwise specified, "multiple" means two or more.

[0042] In the description of the present invention, it should be noted that, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0043] like Figure 1 As shown, the EEG signal detection method of this embodiment includes: step S1, obtaining multiple abnormal discharge waveforms. Step S2, morphological clustering to obtain abnormal discharge templates, that is, classifying multiple abnormal discharge templates according to the correlation coefficient between the cluster centroids of each abnormal discharge waveform. Step S3, screening abnormal matching templates from multiple abnormal discharge templates, that is, screening out several abnormal discharge templates as abnormal matching templates according to the abnormal discharge detection distribution. Step S4, detecting online EEG signals according to the abnormal matching templates.

[0044] That is to say, this embodiment first extracts abnormal discharge waveforms of different forms in the EEG signal, and then performs morphological clustering on the abnormal discharge waveforms (i.e., classifying abnormal discharge waveforms with similar / identical forms into the same category) to obtain different abnormal discharge templates. Several better ones are then selected from the multiple abnormal discharge templates as abnormal matching templates for the detection of EEG signals. This embodiment obtains abnormal matching templates based on morphological clustering, which can avoid frequent modifications of the feature extraction process and help reduce the amount of calculation in the detection process; by performing morphological clustering based on abnormal discharge waveforms and constructing abnormal matching templates to detect EEG signals, the accuracy of abnormal detection can be improved.

[0045] Specifically, the abnormal discharge waveform can be obtained based on the energy characteristics of historical EEG signals. For example, the energy characteristics of historical EEG signals include: global discharge coefficient, line length, and envelope. It should be noted that the energy characteristics of EEG signals change over time, and each moment has a corresponding energy characteristic. The moment when the energy characteristic exceeds the set threshold is marked as abnormal discharge, and the data in a period of time before and after the abnormal discharge moment is intercepted as the abnormal discharge waveform. Taking the global discharge coefficient as an example, Figure 2 As shown, a threshold of the global discharge coefficient is set (this threshold = the mean value of the global discharge coefficient + 3*variance), and the moment when the global discharge coefficient is higher than the threshold is marked as abnormal discharge. The signal segment is intercepted with 1s before the sampling moment exceeding the threshold as the starting point and 1s after the sampling moment as the end point as an abnormal discharge waveform. In this way, multiple abnormal discharge waveforms can be obtained.

[0046] This embodiment obtains an abnormal discharge template through morphological clustering, specifically including: obtaining an existing cluster based on an abnormal discharge waveform, sliding correlation between the abnormal discharge waveform after the existing cluster and the cluster centroid of the existing cluster, recording the sliding correlation coefficient during the sliding process, and taking the maximum sliding correlation coefficient as the correlation coefficient between the abnormal discharge waveform and the cluster centroid. If the correlation coefficient is greater than the set threshold, the abnormal discharge waveform is classified into the existing cluster; otherwise, the abnormal discharge waveform is used as a new cluster. Iterate until all abnormal discharge waveforms are classified. When there are multiple correlation coefficients between the abnormal discharge waveform and the cluster centroid that are greater than the set threshold, the abnormal discharge waveform is classified into the existing cluster with the largest correlation coefficient. Among them, the sliding correlation includes: sliding the next abnormal discharge waveform and the abnormal discharge waveform corresponding to the cluster centroid towards each other, and the step length of each sliding is one sampling interval; each time the sliding is performed, the sliding correlation coefficient between the overlapping part of the next abnormal discharge waveform and the abnormal discharge waveform corresponding to the cluster centroid and the cluster centroid is calculated.

[0047] The following example illustrates the process of morphological clustering. Assume that 10 abnormal discharge waveforms are obtained, which are denoted as abnormal discharge waveforms w 1 ~w 10, randomly select abnormal discharge waveform w 1 For the existing cluster C 1 , at this time, the existing cluster C 1 The cluster centroid is the abnormal discharge waveform w 1 When calculating the sliding correlation, first calculate the abnormal discharge waveform w 2 With the existing cluster C 1 Abnormal discharge waveform w 1 Slide towards each other, the step length of each slide is one sampling interval, and each slide has an abnormal discharge waveform w 2 With abnormal discharge waveform w 1 There will be overlap between Figure 3 As shown), calculate the abnormal discharge waveform w 2 The corresponding overlapping part and the existing cluster C 1 It can be understood that a sliding correlation coefficient can be calculated for each sliding. Therefore, when calculating the sliding correlation, the abnormal discharge waveform w 2 With the existing cluster C 1 Multiple sliding correlation coefficients can be obtained between the cluster centroids. Among these multiple sliding correlations, the largest sliding correlation coefficient is selected as the abnormal discharge waveform w 2 With the existing cluster C 1 If the correlation coefficient is greater than a set threshold (e.g., 0.9), the abnormal discharge waveform w 2 Into the existing cluster C 1 Otherwise, the abnormal discharge waveform w 2 As a new cluster C 2 Assuming the abnormal discharge waveform w 2 is the new cluster C 2 The new cluster C 2 The cluster centroid is the abnormal discharge waveform w 2 Then the abnormal discharge waveform w 3 Respectively with the existing cluster C 1 , C 2 The cluster centroid of the sliding correlation is calculated, assuming that the abnormal discharge waveform w 3 With the existing cluster C 1 The correlation coefficient between the cluster centroids is 0.95, which is consistent with the existing cluster C 2 The correlation coefficient between the cluster centroids is 0.7, then the abnormal discharge waveform w 3 Into the existing cluster C 1 The remaining abnormal discharge waveform w 4 ~w 10Repeat the above process with the cluster centroid of the existing cluster to calculate the sliding correlation and complete the classification. When there are multiple correlation coefficients between the abnormal discharge waveform and the cluster centroid that are greater than the set threshold, the abnormal discharge waveform is classified into the existing cluster with the largest correlation coefficient. For example, the abnormal discharge waveform w 10 The correlation coefficients between the cluster centroids of the existing clusters are 0.91, 0.94, and 0.99 respectively. 10 It is classified into the existing cluster corresponding to 0.99. It should be noted that when an abnormal discharge waveform is added to the existing cluster, the cluster centroid is updated to the data average of all abnormal discharge waveforms in the cluster. The abnormal discharge waveforms of epileptic EEG signals vary in shape. By sliding the correlation coefficient, the discharges with the most similar waveforms can be directly clustered into one category, and the alignment of different discharges can be completed at the same time.

[0048] After classifying multiple abnormal discharge waveforms, a total of m abnormal discharge templates can be obtained. Different abnormal discharge templates will have different results for abnormal identification, so the m abnormal discharge templates need to be screened. The correct template will be able to detect more intensive abnormal discharges at locations closer to the epileptic seizure moment, and fewer abnormal discharges at locations farther away from the epileptic seizure moment. Therefore, abnormal matching templates can be screened from multiple abnormal discharge templates based on the abnormal discharge detection distribution, specifically including: eliminating abnormal discharge templates in which the number of abnormal discharge detections before the onset moment is less than the set threshold; in the remaining abnormal discharge templates, calculating the median distance between the abnormal discharge detection moment and the onset moment of each abnormal discharge template, arranging the median distances in order from small to large, and selecting the first n abnormal discharge templates as abnormal matching templates.

[0049] like Figure 4 As shown in Figure 1, using the abnormal discharge template, within 1800 seconds before the epileptic seizure, the distribution of abnormal discharge detection time has no obvious trend. Figure 5 As shown, using abnormal discharge template 2, within 1800s before the epileptic seizure, the detection time of abnormal discharges is basically concentrated in the position close to the seizure time. When performing template screening, firstly, the abnormal discharge templates whose number of abnormal discharge detections before the seizure time is less than the set threshold (for example, 20) can be eliminated. Then, among the remaining abnormal discharge templates, the median distance between the abnormal discharge time and the seizure time detected by each abnormal discharge template is calculated, and the median distance is arranged in order from small to large, and the first n (for example, n can be 5) abnormal discharge templates are selected as abnormal matching templates. It can be understood that the smaller the median distance, the closer the abnormal discharge detection time is to the seizure time. Thus, the optimal n abnormal discharge templates can be screened out for online EEG signal detection.

[0050] Detecting online EEG signals based on abnormal matching templates includes: detecting historical EEG signals using abnormal matching templates, counting abnormal discharge features identified by each abnormal matching template; training a classifier using abnormal discharge features, and using the trained classifier for real-time detection of EEG signals. Abnormal discharge features include: the number of abnormal discharges identified by each abnormal matching template and its proportion in the total number. When detecting historical EEG signals using abnormal matching templates, the historical EEG signals are divided into multiple signal segments with a length of 10 minutes, and each signal segment is identified using n abnormal matching templates, and the number of abnormal discharges detected by the n abnormal matching templates in each signal segment is recorded. For example, taking 5 abnormal matching templates as an example, in each signal segment, the number of abnormal discharges detected by the 5 abnormal matching templates is recorded as a, b, c, d, and e respectively, then the proportion of abnormal discharges detected by each abnormal matching template is: a / (a+b+c+d+e), b / (a+b+c+d+e), c / (a+b+c+d+e), d / (a+b+c+d+e), and e / (a+b+c+d+e). Figure 6 As shown in the figure, at a location far from the epileptic seizure moment, the distribution of the proportion of abnormal discharges detected by the five abnormal matching templates varies greatly (different colors represent the proportion of detection by different templates). At a location close to the epileptic seizure moment, the distribution of the proportion of abnormal discharges detected by the five abnormal matching templates is relatively stable, with the distribution pattern that templates one and five account for a small proportion, and templates two, three, and four account for a large proportion.

[0051] After obtaining the abnormal discharge features, the classifier can be trained using the abnormal discharge features. A signal segment is captured every 10 minutes, and the number and proportion of abnormal discharges detected by a set of abnormal matching templates in the signal segment are obtained as training features. The hour before each attack is identified as the pre-attack period, and the corresponding abnormal discharge features are extracted as pre-attack training data. The data within 1 day before and after the attack is marked as uncertain data. The data after removing the pre-attack data and uncertain data is used as interictal data, and the corresponding abnormal discharge features are extracted as interictal training data. Figure 7 As shown, the naive Bayes method is used to train the early recognition classifier, and the detection accuracy of the trained early recognition classifier can reach 98.8%.

[0052] Abnormal discharges occur before and after an epileptic seizure, indicating that the patient is in a state prone to seizures. Electrical stimulation therapy can help patients get out of this state, but the short-term effect of treatment is still difficult to determine. By identifying abnormal discharges and counting the number of abnormal discharges identified before and after treatment, it is possible to confirm whether electrical stimulation therapy has reduced the occurrence of abnormal states, thereby determining the treatment effect. Figure 8As shown, using the stable time within a few days after the patient's attack, the data of 5 hours before and after treatment, the abnormal state is detected by the early recognition classifier, and the number of detections is compared. As can be seen from the figure, the average number of abnormal detections before treatment is 0.75 times, and the average number of abnormal detections after treatment is 0.2 times, which is significantly reduced, proving the effectiveness of the treatment plan. In other words, the detection method of this embodiment can not only be used for early warning of epilepsy, but also for evaluation of treatment effects.

[0053] The detection method provided in this embodiment obtains abnormal discharge waveforms, performs morphological clustering on the abnormal discharge waveforms to obtain different abnormal discharge templates, and then selects the best n abnormal matching templates from them for detecting abnormal discharges. This can avoid frequent modifications of the feature extraction process on the one hand, which is beneficial to reducing the amount of calculation; on the other hand, it is beneficial to improve the generalization ability of clustering and improve the accuracy of the detection results.

[0054] The present invention also provides an epilepsy detection system, including: a processor and a classifier, the processor is used for an electroencephalogram signal detection method to obtain a detection result at the time of epilepsy attack; the classifier is used to output the detection result. The processor may include one or more single-core or multi-core processors, and the processor may also be a general-purpose processor, a digital signal processor (Digital Signal Processor, DSP), an application-specific integrated circuit (Application Specific Integrated Circuit, ASIC), a field-programmable gate array (Field-Programmable Gate Array, FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc. In addition, the processor may be implemented by an integrated circuit chip. The classifier may be obtained through model training.

[0055] The present invention also provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the EEG signal detection method are implemented. The storage medium includes: a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and other media that can store program codes.

[0056] In summary, the EEG signal detection method and epilepsy detection system based on abnormal discharge waveforms of the present invention can avoid frequent modifications in the feature extraction process and improve the generalization ability of clustering by acquiring multiple abnormal discharge waveforms and then performing morphological clustering on the abnormal discharge waveforms to obtain abnormal discharge templates; and then select the optimal n abnormal discharge templates as abnormal matching templates according to the abnormal discharge detection distribution, which can further improve the detection accuracy. This method can not only be used for early warning detection before epileptic seizures, but also for evaluating treatment effects.

[0057] Based on the above ideal embodiments of the present invention, the relevant staff can make various changes and modifications without departing from the technical concept of the present invention through the above description. The technical scope of the present invention is not limited to the contents of the specification, but must be determined according to the scope of the claims.

Claims

1. A method for detecting an electroencephalogram signal, characterized in that: include: Obtaining multiple abnormal discharge waveforms; The abnormal discharge template is obtained by morphological clustering, that is, multiple abnormal discharge templates are classified according to the correlation coefficient between the cluster centroids of each abnormal discharge waveform; Screening abnormal matching templates from multiple abnormal discharge templates, that is, screening out several abnormal discharge templates as abnormal matching templates according to abnormal discharge detection distribution; Online EEG signal detection based on anomaly matching templates.

2. The method for detecting electroencephalogram signals according to claim 1, wherein: The classification of multiple abnormal discharge templates includes: Obtaining existing clusters based on abnormal discharge waveforms; Perform sliding correlation between the abnormal discharge waveform after the existing cluster and the cluster centroid of the existing cluster, record the sliding correlation coefficient during the sliding process, and use the maximum sliding correlation coefficient as the correlation coefficient between the abnormal discharge waveform and the cluster centroid; if the correlation coefficient is greater than the set threshold, the abnormal discharge waveform is included in the existing cluster; otherwise, the abnormal discharge waveform is used as a new cluster; Iterate until all abnormal discharge waveforms are classified.

3. The method for detecting electroencephalogram signals according to claim 2, wherein: When multiple correlation coefficients between the abnormal discharge waveform and the cluster centroid are greater than a set threshold, the abnormal discharge waveform is classified into the existing cluster with the largest correlation coefficient.

4. The method for detecting electroencephalogram signals according to claim 2, wherein: The sliding related includes: The next abnormal discharge waveform and the abnormal discharge waveform corresponding to the cluster centroid are slid toward each other, and the step length of each sliding is one sampling interval; each time the sliding is performed, the sliding correlation coefficient between the overlapping part of the next abnormal discharge waveform and the abnormal discharge waveform corresponding to the cluster centroid and the cluster centroid is calculated.

5. The method for detecting electroencephalogram signals according to claim 1, wherein: Based on the energy characteristics of historical EEG signals, multiple abnormal discharge waveforms are obtained; among them, The energy characteristics of the historical EEG signal include: global discharge coefficient, line length, and envelope.

6. The method for detecting electroencephalogram signals according to claim 1, wherein: According to the abnormal discharge detection distribution, an abnormal matching template is selected from multiple abnormal discharge templates, including: Eliminate abnormal discharge templates whose number of abnormal discharges detected before the onset moment is less than the set threshold; Among the remaining abnormal discharge templates, the median distance between the abnormal discharge detection time and the onset time of each abnormal discharge template is calculated, the median distances are arranged in ascending order, and the first n abnormal discharge templates are selected as abnormal matching templates.

7. The method for detecting electroencephalogram signals according to claim 1, wherein: The detecting of online EEG signals according to the abnormal matching template comprises: The historical EEG signals are detected using the abnormal matching templates, and the abnormal discharge features identified by the abnormal matching templates are counted; the classifier is trained using the abnormal discharge features, and the trained classifier is used for real-time detection of EEG signals.

8. The method for detecting electroencephalogram signals according to claim 7, wherein: The abnormal discharge features include: the number of abnormal discharges identified by each abnormal matching template and its proportion in the total number.

9. The method for detecting electroencephalogram signals according to claim 1, wherein: The moment when the energy characteristics of historical EEG signals exceed the set threshold is marked as abnormal discharge; The data in a period of time before and after the abnormal discharge moment are intercepted as the abnormal discharge waveform.

10. An epilepsy detection system, characterized in that: include: A processor, executing the EEG signal detection method according to any one of claims 1 to 9, to obtain a detection result of the epileptic seizure moment; The classifier is used to output the detection result.

11. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the EEG signal detection method according to any one of claims 1 to 9 are implemented.

Images