Temporal lobe epilepsy personalized prediction system and device based on multi-site frequency-crossing coupling

Abstract

The invention discloses a personalized prediction system and device for temporal lobe epilepsy based on multi-site frequency-crossing coupling, and the system comprises a signal collection module which is used for synchronously collecting electroencephalogram signals at a plurality of sites of the brain; the signal preprocessing module is used for preprocessing the collected electroencephalogram signals, drawing a cross-frequency coupling common modulation graph and segmenting frequency bands; the feature extraction module is used for extracting phase amplitude coupling features including average coupling center phase frequency, amplitude frequency and average coupling center intensity; the prediction module is used for inputting the phase amplitude coupling characteristics into a pre-trained machine learning model and outputting a prediction result of the temporal lobe epilepsy; and the early warning module is used for generating an early warning signal according to the prediction result. According to the method, a traditional fixed frequency band analysis normal form is broken through, and phase / amplitude migration characteristics are captured through energy distribution weighting; constructing a personalized adaptation mechanism to identify epilepsy atypical cross-frequency coupling features; and the interpretability of a neural oscillation desynchronization mechanism is higher.

Description

Technical Field

[0001] The present invention belongs to the field of electroencephalogram signal processing, and in particular relates to a personalized prediction system and device for temporal lobe epilepsy based on multi-site cross-frequency coupling. Background Art

[0002] Epilepsy is a chronic neurological disease caused by abnormal discharges of brain neurons, and its clinical manifestations are recurrent neurological dysfunction. The disease not only leads to a progressive decline in patients' cognitive function, but may also induce psychological disorders such as anxiety and depression, posing a serious threat to patients' quality of life. Current clinical treatment mainly relies on anti-epileptic drugs to control the disease, but about 30%-40% of patients will develop drug tolerance and develop into drug-resistant epilepsy (DRE). For DRE patients, surgical resection of the epileptogenic focus is an important treatment method. However, due to factors such as the localization accuracy of the epileptogenic focus, surgical risks and postoperative complications, a considerable proportion of patients still cannot receive effective treatment through surgery. In this context, the research value of epilepsy seizure prediction technology is becoming increasingly prominent. By accurately predicting the start time of epileptic seizures, patients can take protective measures in advance (such as adjusting body position and starting neuroregulatory devices), thereby significantly reducing the risk of accidental injuries during the seizure period and alleviating the psychological burden caused by the uncertainty of the seizure. More importantly, reliable prediction results can provide a basis for clinicians to optimize treatment plans, avoid excessive medical intervention, and improve the overall level of disease management.

[0003] Existing epilepsy prediction methods mainly rely on the analysis of electroencephalogram (EEG) signals. Methods such as linear analysis, nonlinear dynamic analysis, and time-frequency analysis have been widely used in epilepsy prediction research. Traditional epilepsy prediction methods are usually difficult to capture the dynamic coupling relationship between different frequency bands in the brain. In particular, the interaction between neural oscillations in different frequency bands often changes significantly before an epileptic seizure, and traditional features often cannot effectively capture this change. For this reason, research based on phase amplitude coupling (PAC) features has gradually attracted attention in recent years. PAC features can effectively reveal the coupling relationship between neural oscillations in different frequency bands, thereby providing new ideas for the early prediction of epileptic seizures.

[0004] Compared with traditional EEG features, the PAC technology system has multiple advantages: First, in terms of feature sensitivity, PAC can dynamically reflect the evolution of coupling strength in different frequency bands, especially within tens of minutes before an epileptic seizure. The significant enhancement of the frequency band coupling relationship, this time-varying characteristic is easily overlooked by conventional power spectrum analysis; secondly, in terms of time resolution, the characteristic values ​​calculated by PAC based on instantaneous phase and amplitude can be updated in seconds, which is conducive to capturing subtle electrophysiological changes before the attack; secondly, in terms of anti-interference ability, PAC features are highly robust to common electromyographic artifacts and motion noise in EEG signals, which is crucial to improving the practicality of bedside monitoring equipment. In contrast, traditional time domain features (such as signal slope and peak-to-peak value) are easily affected by noise interference to produce feature drift, which directly affects the stability of the prediction model.

[0005] Although PAC technology has shown unique value in the field of epilepsy prediction, existing methods still have limitations: 1) The problem of single feature extraction dimension. Current studies generally use the maximum value of the modulation index (MI) as the core feature. Although this method can reflect the coupling strength of a specific frequency band combination, it cannot describe the dynamic migration law of the phase center frequency and the amplitude center frequency (such as the pre-ictal period). Frequency band phase Frequency band migration, The frequency band amplitude extends to the ripple frequency band). Strategies that rely on fixed thresholds or global statistics result in insufficient sensitivity to changes in network states before seizures; 2) Lack of individualized adaptation capabilities. Existing methods often preset single frequency band combinations (such as θ phase-γ amplitude coupling), ignoring atypical coupling patterns caused by different epilepsy types (such as frontal lobe epilepsy and temporal lobe epilepsy) or pathological mechanisms (such as structural lesions and gene mutations). For example, clinical data show that some DRE patients have Abnormal coupling of phase and ripple frequency band amplitude, while traditional preset frequency band analysis methods have a detection blind spot for such variation patterns; 3) Limited feature interpretability. As a mathematical statistic, the calculation results of the MI indicator lack a clear correlation with the neurophysiological mechanism, and it is difficult to explain the specific change pattern of the excitation-inhibition balance in the epileptogenic network, which seriously restricts clinicians' trust in the prediction results. In response to the above problems, this patent proposes three core technological breakthroughs: breaking through the traditional fixed frequency band analysis paradigm; constructing a personalized epilepsy prediction model to analyze the individualized PAC model and effectively capture frontal lobe epilepsy. -Ripple coupling, drug-resistant epilepsy -Ultra-high frequency coupling and other atypical cross-frequency interaction features; innovatively proposed average coupling strength features, and established pathological associations between feature values ​​and neural oscillation desynchronization mechanisms through phase / amplitude contribution weight fusion. This solution builds an epilepsy prediction feature system that is more adaptable to complex clinical scenarios through collaborative innovation in three dimensions: dynamic frequency band tracking, individualized pattern adaptation, and enhanced physiological interpretability, providing a new methodological framework for the development of precise early warning systems. Summary of the invention

[0006] In view of the deficiencies in the prior art, the present invention provides a personalized prediction system and device for temporal lobe epilepsy based on multi-site cross-frequency coupling, which solves the problems raised in the above background technology.

[0007] A personalized prediction system for temporal lobe epilepsy based on multi-site cross-frequency coupling, comprising: A signal acquisition module, used to synchronously acquire EEG signals at multiple sites in the brain; The signal preprocessing module is used to preprocess the collected EEG signals, draw the co-modulation diagram of cross-frequency coupling, and divide the frequency bands; A feature extraction module, used to extract phase amplitude coupling features, including average coupling center phase frequency, average coupling center amplitude frequency and average coupling strength; A prediction module is used to input the phase amplitude coupling features into a pre-trained machine learning model and output the prediction results of temporal lobe epilepsy; The early warning module is used to generate early warning signals according to the prediction results.

[0008] Furthermore, the collected EEG signals are preprocessed, including: Adopt notch filter to remove power frequency noise and mean filter processing; The continuous electrical signal after notch filtering, power frequency noise removal and mean filtering is divided into processing windows according to a fixed division time length, while ensuring that adjacent windows overlap for a certain overlapping time length, and the overlapping time length is less than the fixed division time length.

[0009] Furthermore, the process of drawing the co-modulation diagram of the cross-frequency coupling is as follows: For the electrical signal in each processing window, the instantaneous phase of low-frequency oscillation and the instantaneous amplitude of high-frequency oscillation are extracted by complex wavelet transform, and the MI is calculated using the modulation index calculation formula; the horizontal axis is the phase frequency, and the vertical axis is the amplitude frequency, and the co-modulation map (Co-modulation map) of the cross-frequency coupling of phase frequency-amplitude frequency between different frequency bands in the processing window is obtained; the co-modulation map is used to describe the phase-amplitude coupling strength of multi-site signals; Furthermore, the modulation index MI calculation formula is: ; in, The instantaneous phase of the low-frequency oscillation at time is defined as , the instantaneous amplitude of high-frequency oscillation is defined as ; The instantaneous phase frequency is , the instantaneous amplitude frequency is The modulation index at ; is an imaginary unit used to construct rotation vectors on the complex plane.

[0010] Furthermore, dividing the frequency band includes: The low frequency oscillation is divided into 4 frequency bands, namely Wave, Wave, The overall low frequency band, high frequency oscillation is divided into 4 frequency bands, namely low wave, high waves, ripple waves and overall high frequency band.

[0011] Furthermore, the calculation process of the phase amplitude coupling feature is: The four low-frequency bands and four high-frequency bands obtained by dividing the frequency bands are combined into 16 cross-frequency coupling modes, and the average coupling center phase frequency, average coupling center amplitude frequency and average coupling strength characteristics are further calculated; among which: Average coupling center phase frequency : ; Average coupling center amplitude frequency : ; Average coupling strength characteristics : ; in, and represent the row and column indices of the modulation index matrix, respectively, Indicates time hour Position in the matrix ( ) value; and They are the phase-frequency dimension and the amplitude-frequency dimension respectively.

[0012] The machine learning model is a support vector machine, which is used to classify the 16 combined feature inputs corresponding to all channels, verify the performance of screening the individual optimal prediction channel and frequency band combination, save the model parameters and output the prediction results of temporal lobe epilepsy.

[0013] Furthermore, when the 16 frequency bands are combined, the , The MCCV features are input into the SVM model to obtain model evaluation indicators including accuracy, sensitivity, specificity, precision and F1 score, determine the channel and frequency band combination with the best classification accuracy, and save the model parameters corresponding to the optimal frequency band combination and channel.

[0014] The personalized prediction system for temporal lobe epilepsy also includes a post-processing module for optimizing the prediction results. The optimization process of the post-processing module is as follows: Post-processing the output of the classifier in the machine learning model on the prospective study dataset using a k-of-n module to label preictal states; The hourly false alarm rate, prediction sensitivity and prediction lead time are calculated based on the post-processing results.

[0015] A computer device comprises a memory and a processor, wherein the memory stores a computer program, and wherein the processor implements the functions of the above-mentioned temporal lobe epilepsy prediction system when executing the computer program.

[0016] Furthermore, the system adopts an embedded processor and a low-power design, is compatible with the long-term monitoring requirements of implantable devices, and supports the integration of a closed-loop neural regulation system.

[0017] Compared with the prior art, the present invention has the following beneficial effects: 1. The personalized prediction system for temporal lobe epilepsy of the present invention combines the phase frequency and amplitude frequency coupling between different frequency bands of multiple sites to predict temporal lobe epilepsy, thereby improving the accuracy and efficiency of temporal lobe epilepsy prediction.

[0018] 2. Compared with the traditional multi-channel feature extraction method, the present invention uses single-channel EEG signals to extract PAC features, which can not only reduce the complexity of data acquisition, but also fully capture the cross-band coupling information of neural oscillations. The system of the present invention not only has a high temporal resolution and can capture subtle changes before an epileptic seizure, but also has a strong anti-noise ability and can adapt to noise interference in clinical environments. In addition, the PAC feature can provide a more biologically meaningful explanation of the neural mechanism, and can effectively achieve early prediction of epileptic seizures, and has a high clinical application value. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 4 is an implementation flow chart of the personalized prediction system for temporal lobe epilepsy in an embodiment of the present invention.

[0020] Figure 2 1 is a characteristic description of a single-channel phase amplitude co-modulation diagram in an embodiment of the present invention.

[0021] Figure 3 These are the SVM model training results corresponding to different sub-band coupling results in the embodiment of the present invention.

[0022] Figure 4 This is an example of actual prediction results of the personalized prediction system for temporal lobe epilepsy in an embodiment of the present invention. DETAILED DESCRIPTION

[0023] The present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be pointed out that the embodiments described below are intended to facilitate the understanding of the present invention and do not have any limiting effect on the present invention.

[0024] In the process of studying the changes in phase-amplitude coupling from before to after an epileptic seizure, it was found that the phase-amplitude coupling characteristics of each site changed significantly from before to after an epileptic seizure. The study also found that the results of predicting epilepsy using phase-amplitude coupling characteristics of different frequency bands in the same individual were different, and there was a frequency band combination with the best ability to predict epilepsy. At the same time, the best frequency band combination was different between different individuals.

[0025] Based on this research result, the embodiment of the present invention provides a personalized prediction system for temporal lobe epilepsy based on multi-site cross-frequency coupling. The specific steps of the implementation process of the system are as follows: Figure 1 The detailed steps are as follows: Step 1: Acquire abnormal electrical signals synchronously recorded at multiple sites in the hippocampus and limbic system and prepare a data set.

[0026] In the embodiment, electrodes are implanted into multiple sites on the hippocampus and limbic system of the brain, where the limbic system refers to the structures around the hippocampus, such as the anterior nucleus of the thalamus, the amygdala, etc., which are clear to those skilled in the art. After the electrodes are implanted, the EEG signal acquisition device is started, and the EEG signals of multiple sites are continuously collected. There are abnormal discharge signals in the EEG signals, and the abnormal discharge signals refer to the EEG signals of temporal lobe epilepsy. The collected EEG signals are input into the temporal lobe epilepsy prediction system and received.

[0027] Step 2: Preprocess the data set and calculate features.

[0028] The purpose of preprocessing is to improve data quality and eliminate noise. Specifically, the received EEG signal is preprocessed to determine the parameter combination used for the prediction model, including: (a) The data set is filtered with a 50 Hz notch filter to remove power frequency noise and an 8-point mean filter is performed on each acquisition channel.

[0029] (b) For each recording site, a co-modulation map is plotted over continuous time with the horizontal axis representing the phase frequency and the vertical axis representing the amplitude frequency. This co-modulation map is used to describe the changes in the cross-frequency coupling characteristics between different frequency bands at each site.

[0030] The constructed co-modulation map is mainly used for cross-frequency analysis of continuous signals. Specifically, the construction process of the co-modulation map is: The collected continuous electrical signal is divided into processing windows according to a fixed division time length, while ensuring that adjacent windows overlap for a certain overlapping time length, and the overlapping time length is less than the fixed division time length. The electrical signal in each processing window is subjected to complex wavelet transform to extract the instantaneous phase of the low-frequency signal and the instantaneous amplitude of the high-frequency signal to calculate MI, and obtain a co-modulation diagram reflecting the phase coupling characteristics of the low-frequency oscillation and the amplitude coupling characteristics of the high-frequency oscillation in the processing window.

[0031] For example, the continuous EEG signal is divided into processing windows of 5s, and in order to improve the resolution of the co-modulation map in the continuous time, the adjacent processing windows overlap by 1s. In each window, the complex wavelet transform calculation formula is as follows: ; in, are complex coefficients, is the scale parameter, is the time displacement parameter, is the conjugate of the complex wavelet. The phase of the low-frequency oscillation is defined as , the amplitude of the high-frequency oscillation is defined as , the specific calculation formula is as follows: ; ; MI is a phase-amplitude coupling quantitative index based on information entropy. Its core is to reflect the phase modulation effect through the statistical characteristics of amplitude distribution. The calculation formula is as follows: ; After calculating the modulation index, the co-modulation diagram is obtained by plotting the phase frequency on the horizontal axis and the amplitude frequency on the vertical axis with a resolution of 0.1 Hz and 1 Hz.

[0032] (c) Calculate features based on the obtained phase-amplitude co-modulation map.

[0033] Specifically, the calculation formula for the phase amplitude coupling characteristics within the window is as follows: Average coupling center phase frequency : ; Average coupling center amplitude frequency : ; Average coupling strength characteristics : ; in, and Represent the row and column indices of the modulation index matrix respectively. Indicates time hour Position in the matrix ( ) value, and are the phase-frequency dimension and the amplitude-frequency dimension respectively, is the average coupling center phase frequency, is the average coupling center amplitude frequency, is the average coupling strength.

[0034] Step 3: Divide the EEG signal into different periods, build a prediction model, and perform channel selection and frequency band optimization to determine the model parameters based on the model classification results of the phase-amplitude coupling characteristics obtained from the multi-site phase-amplitude co-modulation map.

[0035] The start and end time of the epileptic EEG signals in the collected data set are annotated, and the inter-epileptic period and pre-epileptic period are mainly extracted based on the start and end time of the epileptic seizure.

[0036] In the embodiment, the pre-epileptic period is defined as the time range from 180 seconds before the epileptic onset to the epileptic onset; correspondingly, the inter-epileptic period is the time range from the end of the epileptic onset to 180 seconds before the next epileptic onset.

[0037] Calculation data set EEG signal different recording sites and 4 low-frequency oscillation bands ( Wave, Wave, wave and overall low frequency band) and 4 high frequency oscillation bands (low wave, high The phase amplitude coupling features were obtained under 16 frequency band combinations (e.g., epilepsy, ripple waves, and overall high frequency bands). The calculated features were divided into training sets and test sets, and put into the SVM model for training; the model parameters after training were saved, and the model performance was tested using the test set to obtain an epilepsy prediction model based on cross-frequency coupling.

[0038] Specifically, training and testing an epilepsy prediction model based on multi-site phase amplitude coupling features includes: The dataset was randomly divided into 10 subsets, 9 subsets were used to train the model in each round, and the remaining 1 subset was used to test the model, which was repeated 10 times in total. The classification results were evaluated through the classification confusion matrix of the model, and the five model performance indicators of accuracy, sensitivity, specificity, precision, and F1 score were calculated.

[0039] In the embodiment, 10 channels and 16 frequency bands are combined respectively. , and The features are put into SVM model training. After ten-fold cross-validation, five model performance indicators are obtained. The calculation method of model performance indicators is as follows: ; ; ; ; ; in, True positive samples are pre-epileptic samples classified as pre-epileptic samples. True negative samples are epileptic interictal samples classified as interictal interictals. For false positive samples, False negative samples.

[0040] In the SVM classification results, the classification performance indicators of different channels and different frequency band combinations are different. The channel and frequency band combination with the highest classification accuracy is selected as the optimal parameter combination of the model. The model parameters of the optimal parameter combination are saved and used in the subsequent personalized epilepsy prediction system. Different from the traditional single-band or fixed-band combination, the present invention proposes a dynamic screening mechanism for 16 cross-band combinations, and for the first time combines 4 low-frequency ( / / / 1-12Hz) and 4 high frequencies (low / high / ripple / 30-200Hz) was introduced into the prediction of temporal lobe epilepsy, revealing the nonlinear correlation between different frequency bands.

[0041] In step 4, using the forward-looking data set, the features of the preprocessed signal are calculated and input into the pre-trained SVM model for classification.

[0042] Specifically, post-processing is a crucial step in the epilepsy prediction process. The main purpose is to optimize and adjust the original prediction results output by the model to improve the accuracy, reliability and practicality of the prediction. It usually includes data smoothing, threshold selection and other processing methods to reduce false positives and false negatives. The "k-of-n" module proposed by Parvez et al. is used to post-process the labels obtained by the real-time classification of the classifier. The label 0 represents the interictal state and 1 represents the preictal state. If there are more than or equal to k output samples with a label of 1 in n consecutive windows, it is marked as the preictal state and an early warning is issued, otherwise it is marked as the interictal state without an early warning.

[0043] In this embodiment, the "8-of-10" module is used to post-process the labels output by the model based on the prospective dataset to obtain the hourly false alarm rate (FPR / h), the true predicted percentage (TPP) and the prediction horizon (PH). The formulas for calculating the hourly false alarm rate and the success prediction rate are as follows: ; .

[0044] An embodiment of the present invention further provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor implements the functions of the temporal lobe epilepsy prediction system in the above embodiment when executing the computer program.

[0045] In order to verify the effect of the present invention, the temporal lobe epilepsy prediction system of the present invention was tested through the following experiments.

[0046] This experiment used 6 Sprague-Dawley rats with lithium chloride-pilocarpine abdominal injection. After inducing the rats to a persistent epileptic state, 45μm nickel-chromium alloy electrodes were implanted at 10 sites on the bilateral hippocampus (CA1, CA3, DG, SUB) and the left and right amygdala (AMD) for uninterrupted monitoring for 6 to 8 weeks. The signal sampling frequency was 1000Hz and the resolution was 16 bits.

[0047] Preprocessing stage: The collected EEG signals are time-labeled and divided into the inter-epileptic period, pre-epileptic period and epileptic attack period. Each channel is subjected to 50Hz power frequency notch denoising and 8-point mean filtering. Figure 2 The EEG signal of channel 5 (R-CA3) containing an epileptic seizure after filtering is shown. The data set is divided into continuous windows with a window length of 5s and an overlap time of 1s. The instantaneous phase of the low-frequency band and the instantaneous amplitude of the high-frequency band are extracted using complex wavelet transform in each window, and the degree of phase-amplitude coupling in each window is calculated using the MI calculation formula to obtain the co-modulation map (such as Figure 2 The co-modulation map is divided into sub-bands of delta wave (1-3Hz), theta wave (3-8Hz), alpha wave (8-12Hz) and the overall low frequency band (1-12Hz) and low gamma wave (30-50Hz), high gamma wave (50-120Hz), ripple wave (120-200Hz) and the overall high frequency band (30-200Hz). Then through , and The characteristic calculation formula can obtain the characteristic values ​​corresponding to 16 frequency band combinations. Figure 2This is the co-modulation diagram of the phase frequency range 1-12Hz and amplitude frequency range 120-200Hz of channel 5 in experimental example 1. The average coupling strength of the co-modulation diagram is MCCV. The five-pointed star in the figure represents the average coupling center of the co-modulation diagram, and its corresponding horizontal and vertical coordinates are and .

[0048] The phase amplitude coupling characteristics of the 1-12 / 30-200Hz frequency band combination of 10 channels obtained after preprocessing , and Using the SVM model for classification, Table 1 shows the classification results of 10 channels under the above experimental conditions. For Experimental Example 1, the channel with the highest accuracy (R-CA3) is selected to calculate the 16 frequency band combinations of this channel. , and feature. Figure 3 The classification results of 16 frequency band combinations obtained by using the SVM classifier for this channel are shown. For Experimental Example 1, the frequency band combination with the highest classification accuracy is 1-12Hz / Ripple. Based on the above experimental results, the parameter combination of the 1-12Hz / Ripple frequency band combination using R-CA3 is saved for the online epilepsy prediction system of Experimental Example 1.

[0049] Table 1

[0050] On the prospective dataset, each experimental case is predicted using the personalized prediction model saved after pre-training, and the prediction results are post-processed by the “8-of-10” module.

[0051] Through this experimental verification, the specific prediction results of each experimental example are shown in Table 2. The prediction system of the present invention can achieve the highest successful prediction of 87.5% of epileptic seizures, an average of 5.68±2.81 false alarms per hour, and an average of 63.47±14.36s in advance to predict epileptic seizures. Figure 4 A schematic diagram of the prediction model of the present invention successfully predicting an epileptic seizure result at one time is given.

[0052] Table 2

[0053] The embodiments described above provide a detailed description of the technical solutions and beneficial effects of the present invention. It should be understood that the above are only specific embodiments of the present invention and are not intended to limit the present invention. Any modifications, supplements and equivalent substitutions made within the scope of the principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A personalized prediction system for temporal lobe epilepsy based on multi-site cross-frequency coupling, characterized in that: include: A signal acquisition module, used to synchronously acquire EEG signals at multiple sites in the brain; The signal preprocessing module is used to synchronously preprocess the collected EEG signals, draw the co-modulation diagram of cross-frequency coupling, and divide the frequency bands; A feature extraction module, used to extract phase amplitude coupling features, including average coupling center phase frequency, average coupling center amplitude frequency and average coupling strength; A prediction module, which is used to input the phase amplitude coupling features into a pre-trained machine learning model and output the prediction results of epilepsy; The early warning module is used to generate early warning signals according to the prediction results.

2. The personalized prediction system for temporal lobe epilepsy based on multi-site cross-frequency coupling according to claim 1, characterized in that: Preprocess the collected EEG signals, including: Power frequency noise filtering, mean filtering and segmentation processing; The continuous EEG signal after notch filtering, power frequency noise removal and mean filtering is divided into processing windows according to a fixed time length, while ensuring that adjacent windows overlap for a certain time length, and the overlapping time length is less than the fixed division time length.

3. The personalized prediction system for temporal lobe epilepsy based on multi-site cross-frequency coupling according to claim 1, characterized in that: The process of drawing the co-modulation diagram of the cross-frequency coupling is as follows: The instantaneous phase of low-frequency oscillation and the instantaneous amplitude of high-frequency oscillation are extracted from the EEG signal in each processing window by using complex wavelet transform, and the modulation index is calculated using the modulation index calculation formula; the horizontal axis is the phase frequency and the vertical axis is the amplitude frequency, and the co-modulation diagram of the cross-frequency coupling of phase frequency-amplitude frequency between different frequency bands in the processing window is obtained; Wherein, the modulation index calculation formula is: ； in, The instantaneous phase of the low-frequency oscillation at time is defined as , the instantaneous amplitude of high-frequency oscillation is defined as ; The instantaneous phase frequency is , the instantaneous amplitude frequency is The modulation index at ; is an imaginary unit used to construct rotation vectors on the complex plane.

4. The personalized prediction system for temporal lobe epilepsy based on multi-site cross-frequency coupling according to claim 1, characterized in that: The divided frequency bands include: The low frequency oscillation is divided into 4 frequency bands, namely Wave, Wave, The overall low frequency band, high frequency oscillation is divided into 4 frequency bands, namely low wave, high waves, ripple waves and overall high frequency band.

5. The personalized prediction system for temporal lobe epilepsy based on multi-site cross-frequency coupling according to claim 1, characterized in that: The calculation process of the phase amplitude coupling characteristic is: The four low-frequency bands and four high-frequency bands obtained by dividing the frequency bands are combined into 16 cross-frequency coupling modes, and the average coupling center phase frequency, average coupling center amplitude frequency and average coupling strength characteristics are further calculated, where: Average coupling center phase frequency : ； Average coupling center amplitude frequency : ； Average coupling strength characteristics : ； in, and represent the row and column indices of the modulation index matrix, respectively, Indicates time hour Position in the matrix ( ) value; and They are the phase-frequency dimension and the amplitude-frequency dimension respectively.

6. The personalized prediction system for temporal lobe epilepsy based on multi-site cross-frequency coupling according to claim 1, characterized in that: The machine learning model is a support vector machine, which is used to classify the 16 combined feature inputs corresponding to all channels, verify the performance of screening the individual optimal prediction channel and frequency band combination, save the model parameters and output the prediction results of temporal lobe epilepsy.

7. The personalized prediction system for temporal lobe epilepsy based on multi-site cross-frequency coupling according to claim 1, characterized in that: The personalized prediction system for temporal lobe epilepsy also includes a post-processing module for optimizing the prediction results. The optimization process of the post-processing module is as follows: Post-process the output of the classifier in the machine learning model using a k-of-n module to label the pre-ictal state; According to the post-processing results, the hourly false alarm rate, prediction sensitivity and prediction lead time are obtained.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the functions of the personalized prediction system for temporal lobe epilepsy according to any one of claims 1 to 7 are realized.

Images