An online seizure adaptive prediction method and system

Abstract

The application discloses an online self-adaptive seizure prediction method and system, relates to the technical field of medical signal processing and artificial intelligence, and comprises the following steps: acquiring multi-channel electroencephalogram signals of a target user; constructing an online self-adaptive seizure prediction model according to spatially constrained independent component analysis, brain function network and a transfer learning mechanism; inputting the multi-channel electroencephalogram signals into the online self-adaptive seizure prediction model, identifying and predicting a pre-seizure state, and obtaining a prediction result; and performing online early warning judgment according to the prediction result, and triggering an alarm if early warning conditions are met. The application solves the problem of unstable early warning effect caused by the variability of electroencephalogram data of epilepsy, realizes online seizure prediction with clinical accuracy and rapidness, and provides a basis for the treatment of intractable epilepsy and the research on the seizure mechanism of epilepsy.

Description

Technical Field

[0001] This invention relates to the fields of medical signal processing and artificial intelligence technology, and in particular to an online adaptive prediction method and system for epileptic seizures. Background Technology

[0002] Epilepsy is a clinical syndrome caused by multiple etiologies, characterized by paroxysmal, transient, repetitive, and stereotyped central nervous system dysfunction caused by highly synchronized abnormal discharges of neurons in the brain. It is the second most intractable disease after cerebrovascular diseases. Moreover, refractory epilepsy is extremely harmful, and there are no effective treatments. Long-term and frequent seizures can lead to brain damage and, in severe cases, even threaten life. Therefore, the prediction of epileptic seizures is crucial.

[0003] In recent years, research on epilepsy seizure prediction has made some progress, laying a theoretical foundation for its clinical application. However, it still falls short of clinical application. Currently, most epilepsy seizure prediction algorithms are based on publicly available online datasets, which suffer from insufficient sample size, lack of generalizability, and inadequate consideration of epilepsy heterogeneity. Therefore, the performance of these algorithms is not convincing when compared to large clinical datasets. The International League Against Epilepsy considers epilepsy a disease of disordered brain network connectivity, emphasizing the role of brain networks in epilepsy research. Furthermore, existing multivariate-based epilepsy seizure prediction methods largely fail to consider spatial synchronization and connectivity information within the brain, resulting in a weak connection to the pathogenesis of epilepsy.

[0004] Therefore, there is an urgent need for an online adaptive prediction method and system for epileptic seizures. Based on the accumulation of relevant epileptic EEG signal research, this method should focus on the difficult problem of epileptic seizure prediction and achieve accurate and rapid online epileptic seizure prediction. Summary of the Invention

[0005] The purpose of this invention is to provide an online adaptive prediction method and system for epileptic seizures, which solves the problem of unstable early warning effect caused by the variability of epileptic EEG data, thereby achieving accurate and rapid online prediction of epileptic seizures.

[0006] To achieve the above objectives, the present invention provides an online adaptive prediction method for epileptic seizures, comprising the following steps:

[0007] S1. Acquire multi-channel EEG signals from the target user;

[0008] S2. Based on fusion spatial constraint independent component analysis, brain functional networks, and transfer learning mechanisms, an online adaptive prediction model for epileptic seizures is constructed.

[0009] S3. Input multi-channel EEG signals into the online adaptive prediction model for epileptic seizures to identify and predict the pre-seizure state and obtain prediction results.

[0010] S4. Based on the prediction results, make an online early warning judgment. If the early warning conditions are met, trigger an alarm.

[0011] Preferably, the specific content of constructing an online adaptive prediction model for epileptic seizures in S2 based on fusion spatial constraint independent component analysis, brain functional networks, and transfer learning mechanisms includes:

[0012] Acquire users' multi-channel historical EEG signals and epileptic seizure data;

[0013] Training and testing sets were constructed based on multi-channel historical EEG signals and epileptic seizure data.

[0014] Based on fusion spatial constraint independent component analysis, an EEG noise estimation module is constructed;

[0015] Based on brain functional networks, a brain electroencephalogram (EEG) feature information mapping module was constructed.

[0016] Based on the transfer learning mechanism, an EEG information prediction module was constructed;

[0017] By connecting the EEG noise estimation module, the EEG feature information mapping module, and the EEG information prediction module, an initial model for the online adaptive prediction model of epileptic seizures is obtained.

[0018] The initial model is trained and tested based on the training and test sets, and the trained and tested initial model is used as an online adaptive prediction model for epileptic seizures.

[0019] Preferably, the specific content of constructing the EEG noise estimation module based on fusion spatial constraint independent component analysis includes:

[0020] The multi-channel historical EEG signals were filtered to obtain the filtered signal;

[0021] Based on the spatial constraint matrix and the fusion spatial constraint independent component analysis, the filtered signal is separated to obtain the filtered signal components;

[0022] Wavelet threshold denoising is performed on the filtered signal components to obtain the noise components;

[0023] Perform an inverse projection transformation on the noise components to obtain the noise estimate;

[0024] The noise estimate in the filtered signal is removed to obtain a clean EEG signal, and an EEG noise estimation module is obtained.

[0025] Preferably, the specific content of constructing the EEG feature information mapping module based on the brain functional network includes:

[0026] Frequency slice wavelet transform was used to extract the rhythm from the pure EEG signal to obtain the rhythmic EEG signal;

[0027] A multi-track autoregressive module was established using rhythmic EEG signals as the object, and the coefficients of the multi-track autoregressive module were solved by the least squares method.

[0028] Based on the partial orientation coherence method, a brain functional network diagram and connectivity matrix are constructed according to the coefficients of the multi-track autoregressive module. The brain functional network diagram and connectivity matrix are used as feature images to map EEG information, thus obtaining the EEG feature information mapping module.

[0029] Preferably, based on the transfer learning mechanism, the specific content of the EEG information prediction module includes:

[0030] Based on the feature images, a bidirectional long short-term memory network is constructed by applying inter-domain joint knowledge transfer to obtain the EEG information prediction module.

[0031] Preferably, the bidirectional long short-term memory network is a structural optimization based on the long short-term memory network. Specifically, it involves adding a connection from the input gate to the output gate in the long short-term memory network structure and merging the input gate and the forget gate into a single update gate.

[0032] Preferably, the online early warning judgment in S4 includes:

[0033] A preset warning threshold n and prediction count s are defined; where n is an integer greater than 1 and s is an integer greater than or equal to 0.

[0034] If s=0, it means that the online adaptive model for epileptic seizures did not detect EEG information of epileptic premonitory signals.

[0035] If s≥n, the prediction result is identified as a pre-seizure state, triggering an alarm;

[0036] If s < n, the prediction result is identified as the pre-seizure state, and prediction continues until s ≥ n, at which point an alarm is triggered.

[0037] This invention provides an online adaptive prediction system for epileptic seizures, comprising:

[0038] The signal acquisition module is used to acquire multi-channel EEG signals from the target user.

[0039] The model building module is used to construct an online adaptive prediction model for epileptic seizures based on fusion spatial constraint independent component analysis, brain functional networks, and transfer learning mechanisms.

[0040] The online prediction module is used to input multi-channel EEG signals into the online adaptive prediction model for epileptic seizures to identify and predict the pre-seizure state and obtain prediction results.

[0041] The online early warning module is used to make online early warning judgments based on the prediction results, and to trigger an alarm if the early warning conditions are met.

[0042] The present invention provides a computer device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the above-described online adaptive prediction method for epileptic seizures.

[0043] The present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-described online adaptive prediction of epileptic seizures.

[0044] In summary, the online adaptive prediction method and system for epileptic seizures provided by this invention offers the following advantages compared to traditional technologies: The online adaptive prediction model for epileptic seizures, constructed based on spatially constrained independent component analysis, brain functional networks, and transfer learning mechanisms, fully explores the changes in the topological structure of the brain functional network at different stages in different epilepsy patients. It addresses the problems of diverse EEG patterns during epileptic seizures, the presence of significant noise in EEG signals, and the complexity and variability of EEG data. Furthermore, by combining an online early warning mechanism, it solves the problem of unstable early warning effects caused by the variability of epileptic EEG data. Thus, it achieves accurate and rapid online prediction of epileptic seizures in clinical practice, providing a foundation for the treatment of refractory epilepsy and research on its seizure mechanisms.

[0045] The technical method of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0046] Figure 1 This is a flowchart of an online adaptive prediction method for epileptic seizures according to the present invention;

[0047] Figure 2 This is a structural diagram of the EEG noise estimation module in this invention;

[0048] Figure 3 This is an architecture diagram of the EEG feature information mapping module in this invention;

[0049] Figure 4 This is a structural diagram of the EEG information prediction module in this invention;

[0050] Figure 5 This is a structural diagram of the online early warning mechanism in this invention;

[0051] Figure 6A block diagram of an online adaptive prediction system for epileptic seizures according to the present invention. Detailed Implementation

[0052] The technical method of the present invention will be further described below with reference to the accompanying drawings and embodiments. It should be noted that, unless otherwise specifically stated, the relative arrangement, numerical expressions, and values ​​of the components and steps described in these embodiments do not limit the scope of this application.

[0053] The following description of at least one exemplary embodiment is merely illustrative and is in no way intended to limit the scope of this application and its application or use.

[0054] Techniques, systems, and equipment known to those skilled in the art may not be discussed in detail, but where appropriate, they should be considered part of the instruction manual.

[0055] In all the examples shown and discussed herein, any specific values ​​should be interpreted as merely exemplary and not as limitations. Therefore, other examples of exemplary embodiments may have different values.

[0056] Unless otherwise defined, the technical or scientific terms used in this invention shall have the ordinary meaning as understood by one of ordinary skill in the art to which this invention pertains.

[0057] like Figure 1 As shown, an online adaptive prediction method for epileptic seizures includes:

[0058] Step S1: Acquire the multi-channel EEG signal of the target user. The multi-channel EEG signal is the current multi-channel EEG signal data of the target user. Due to the complexity of the epileptic seizure mechanism, epileptic EEGs exhibit diversity. This diversity is reflected in the significant individual differences in EEG characteristics during seizures among different epileptic patients. Furthermore, EEG signals contain a large amount of noise. Traditional epileptic EEG pattern recognition methods often perform feature analysis based on the form and characteristics of a portion of the signal and apply the same features to all data, resulting in highly unstable performance across different individuals. This approach is only suitable for situations with strong signal pattern consistency, while clinical patients vary greatly, and EEG data is often complex and variable, representing typical non-consistent data. Therefore, traditional methods struggle to achieve good results. To address this, this invention constructs an online adaptive prediction model for epileptic seizures.

[0059] Step S2: Based on fusion spatial constraint independent component analysis, brain functional networks, and transfer learning mechanisms, an online adaptive prediction model for epileptic seizures is constructed. An epileptic EEG signal decoding method with adaptive learning capabilities is crucial for achieving clinical epilepsy early warning. Therefore, this invention addresses the complex and noise-sensitive nature of epileptic EEG patterns by designing algorithms from three aspects: signal denoising, feature parsing, and state prediction, thus constructing an online adaptive prediction model for epileptic seizures.

[0060] Furthermore, step S2 can be replaced by steps S201 to S207, as follows:

[0061] Step S201: Obtain the user's multi-channel historical EEG signals and epileptic seizure data. The multi-channel historical EEG signals are the user's historical multi-channel EEG signal data.

[0062] Step S202: Construct training and testing sets based on multi-channel historical EEG signals and epileptic seizure data.

[0063] Step S203: Construct an EEG noise estimation module based on the fusion of Spatially Constrained Independent Component Analysis (SCICA). This invention constructs an EEG noise estimation module based on fusion of SICA by first constructing a constraint matrix using prior knowledge. A spatial constraint matrix is ​​added to the traditional Fast Independent Component Analysis (SCICA) algorithm to reduce computation and enhance the robustness of the method. Then, wavelet thresholding is applied in reverse to denoise the module, and the advantages of SICA and traditional filters are combined to establish the noise estimation module. The module structure is as follows: Figure 2 As shown. Figure 2 In Let be a mixture matrix, where, The mixture matrix contains unknown prior components. SC-I CS For spatially constrained independent components. This is the final denoised EEG signal. X ( t () represents the input multi-channel historical EEG signal. These are the spatially constrained independent components obtained after transformation.

[0064] Specifically, step S203 includes:

[0065] The raw electroencephalogram (EEG) signal, i.e., the multi-channel historical EEG signal, is filtered to obtain a filtered signal. Specifically, the multi-channel historical EEG signal is passed through a bandpass filter, such as a Butterworth bandpass filter, to obtain a filtered signal with a frequency band of 0.5Hz-40Hz. y ( t );

[0066] Based on the spatial constraint matrix and SCICA, the filtered signal is separated to obtain the filtered signal components.

[0067] Wavelet threshold denoising is performed on the filtered signal components to obtain the noise components. .

[0068] Perform an inverse projection transform on the noise components to obtain the noise estimate. ,in, The mixture matrix is ​​a known prior component.

[0069] Remove noise estimates from filtered signals We obtain a clean EEG signal, i.e., a denoised EEG, and obtain an EEG noise estimation module.

[0070] This invention, based on traditional EEG constraint conditions, accurately estimates the noise state to obtain relatively pure EEG data, thereby suppressing and weakening the interference of non-Gaussian noise and artifacts on EEG signal analysis, laying a solid foundation for the revelation and extraction of effective feature information in subsequent EEG signal analysis.

[0071] Step S204: Construct an EEG feature information mapping module based on the brain functional network. To increase the acquisition of effective information from the EEG of different patients, this invention constructs an EEG feature information mapping module. The brain functional network can intuitively reflect the changes in the coupling dependence between channels before and after an epileptic seizure. Constructing the EEG feature information mapping module involves first extracting the rhythm of the pure EEG signal using Frequency-slicing Wavelet Transform (FSWT), and then constructing the brain functional network based on Partial Directed Coherence (PDC). The construction of the brain functional network first establishes a Multivariate Autoregressive (MVAR) module, uses the least squares method to solve for the coefficients and error coefficients of the MVAR module, then transforms it to the frequency domain using Fourier transform, uses PDC to calculate the causal relationship of information transmission between multi-channel EEG signals, and draws the brain functional network diagram and connectivity matrix as feature images to map EEG information. The architecture of the EEG feature information mapping module is as follows: Figure 3 As shown.

[0072] Specifically, step S204 includes:

[0073] Frequency slice wavelet transform was used to extract rhythms from the clean EEG signal, resulting in rhythmic EEG signals. The expression for the frequency slice wavelet transform is as follows:

[0074] ;

[0075] in, t For time parameters, ω For the center frequency, k For the index of the frequency slice, π Pi W f ( t , ω , k () represents the time-frequency features after wavelet transform of the frequency slice. u For the integral variable in the frequency domain, e iut It is a complex exponential function. For the Fourier transform of the signal, It is the conjugate Fourier transform of the mother wavelet.

[0076] The expression for the selected slicing function is:

[0077] ;

[0078] in, This is a rectangle slicing function. ω 0 represents the cutoff frequency of the rectangular slicing function.

[0079] ;

[0080] Where sin is the sine function.

[0081] .

[0082] .

[0083] A multi-track autoregressive module was established using rhythmic EEG signals as the object, and its coefficients were solved using the least squares method. The expression for the multi-track autoregressive module is as follows:

[0084] ;

[0085] in, n For a point in time, x 1( n ) is the first EEG channel in time n The signal value, xN ( n ) is the first N Each brainwave channel in time n The signal value, r The lag order is... r =1,2,…, p , p It is a positive integer. A r For the coefficients of the multi-track autoregressive module, w 1( n ) represents the residual term of the first channel. w N ( n ) is the first N The residual terms of each channel.

[0086] The expression for the coefficients of the multi-track autoregressive module is:

[0087] ;

[0088] in, a 11 ( r The first brainwave pathway is in lag. r At any given moment, the coefficient of influence on its own current signal. a 1N ( r ) is the first N Individual brainwave channels are lagging r At what time does the influence coefficient on the current signal of the first channel... a N1 ( r The first brainwave pathway is in lag. r At that moment, for the first N The influence coefficient of the current signal in each channel a NN ( r ) is the first N Individual brainwave channels are lagging r The influence coefficient of a given moment on its current signal.

[0089] Based on the partial-directed coherence method, a brain functional network map and connectivity matrix are constructed according to the coefficients of the multi-track autoregressive module. These brain functional network map and connectivity matrix are then used as feature images to map EEG information, resulting in EEG feature information mapping modules. Specifically, based on the partial-directed coherence method, a connectivity matrix is ​​constructed according to the coefficients of the multi-track autoregressive module. A threshold is selected to binarize the connectivity matrix, and then the brain functional network map is drawn at selected electrode locations, ultimately yielding the brain functional network map and connectivity matrix.

[0090] The expression for the partial orientation coherence function is:

[0091] ;

[0092] in, P xy ( f () is a partial orientation coherence function used to quantify EEG channels. x and y The intensity of causal information transmission at frequency f Frequency domain coefficient matrix A y ( f The conjugate transpose of ) A y ( f (electroencephalography channel) y The corresponding frequency domain coefficient matrix. Where:

[0093] ;

[0094] in, A xy ( f (electroencephalography channel) x and y In frequency f The frequency domain coefficient matrix below, f For frequency variables, I It is the identity matrix. e -πirf It is a complex exponential function.

[0095] Based on the rhythmic characteristics of electroencephalogram (EEG) signals, this invention evaluates the brain functional network of Granger causality using a multi-track autoregressive module. It fully explores the changes in the topological structure of the brain functional network of epilepsy patients at different stages under different coherence analysis strategies and constructs an EEG feature information mapping module with universal significance.

[0096] Step S205: Construct an EEG information prediction module based on the transfer learning mechanism. This invention addresses the individual differences among patients and the sparsity of epileptic states, and tackles the problem of different distributions between training and test data by constructing a deep EEG information prediction module based on transfer learning. Within the theoretical framework of deep learning, a bidirectional long short-term memory (BiLSTM) network is constructed using feature images (brain functional network diagrams and connectivity matrices) as input. This utilizes inter-domain joint knowledge transfer to fully leverage existing data information to classify target domain data, improving the algorithm's adaptability to different patient data. To enable the Long Short-Term Memory (LSTM) network to better learn long-term historical information, an optimization method for LSTM is proposed: increasing the connection from the input gate to the output gate and merging the input gate and forget gate into a single update gate. This improves the model's automatic learning ability of EEG features and enhances its robustness. The structure of the EEG information prediction module is as follows: Figure 4 As shown, Figure 4 middle F 1, F 2,..., F i-1 , F i For input layer data, L 1, L 2,..., L i-1 , L i This represents the hidden state of the LSTM layer. I 1, I 2,..., I i-1 , I i Attention-weighted features.

[0097] Specifically, step S205 includes:

[0098] Based on feature images, a bidirectional long short-term memory network is constructed using inter-domain joint knowledge transfer to obtain an EEG information prediction module. The bidirectional long short-term memory network is based on structural optimization of the long short-term memory network, specifically by adding a connection from the input gate to the output gate in the long short-term memory network structure, and merging the input gate and the forget gate into a single update gate.

[0099] The diversity and variability of EEG states in epilepsy patients lead to a mismatch between the source domain used for training the prediction model and the target domain used for testing, resulting in poor performance of the EEG information prediction module in the target domain. To improve the adaptability and learning ability of the EEG information prediction module to data from different patients, this invention constructs an EEG information prediction module based on transfer learning. Through domain-joint knowledge transfer and deep optimization networks, it automatically learns and summarizes the inherent patterns of feature maps, proposes an optimization method for the prediction network, expands the algorithm's adaptability to different scenarios, and improves its predictive performance, achieving adaptive prediction of epilepsy EEG signals from different patients.

[0100] Step S206: Connect the EEG noise estimation module, the EEG feature information mapping module, and the EEG information prediction module to obtain the initial model of the online adaptive prediction model for epileptic seizures.

[0101] Step S207: Train and test the initial model based on the training set and test set, and use the trained and tested initial model as an online adaptive prediction model for epileptic seizures.

[0102] Step S3: Input the multi-channel EEG signals into the online adaptive prediction model for epileptic seizures to identify and predict the pre-seizure state and obtain the prediction results.

[0103] Step S4: Based on the prediction results, perform online early warning judgment. If the early warning conditions are met, trigger an alarm. The specific content of the online early warning judgment includes:

[0104] The preset warning threshold n and the number of predictions s are defined. Where n is an integer greater than 1 and s is an integer greater than or equal to 0.

[0105] If s=0, it means that the online adaptive model for epileptic seizures did not detect EEG information of pre-epileptic aura.

[0106] If s≥n, the prediction result is identified as a pre-seizure state, triggering an alarm.

[0107] If s < n, the prediction result is identified as the pre-seizure state, and prediction continues until s ≥ n, at which point an alarm is triggered.

[0108] like Figure 5 As shown, multi-channel EEG signals are processed through a sliding window and EEG decoding to output prediction results. The prediction results are then used to determine whether they match the premonitory signs of an attack. If the prediction is determined to be a premonitory sign of an attack, an alarm is generated, allowing epilepsy to be effectively controlled before an attack occurs.

[0109] This invention is based on an online adaptive prediction model for epileptic seizures, combined with an online early warning mechanism for epileptic seizures. It can simulate clinical scenarios and demonstrates the effectiveness of this invention in early warning of epileptic seizures.

[0110] This invention provides an online adaptive prediction system for epileptic seizures, such as... Figure 6 As shown, it includes:

[0111] The signal acquisition module is used to acquire multi-channel EEG signals from the target user.

[0112] The model building module is used to construct an online adaptive prediction model for epileptic seizures based on fusion spatial constraint independent component analysis, brain functional networks, and transfer learning mechanisms.

[0113] The online prediction module is used to input multi-channel EEG signals into the online adaptive prediction model for epileptic seizures, identify and predict the pre-seizure state, and obtain prediction results.

[0114] The online early warning module is used to make online early warning judgments based on the prediction results, and to trigger an alarm if the early warning conditions are met.

[0115] The present invention provides a computer device comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the content of an online adaptive prediction method for epileptic seizures.

[0116] The present invention provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the content of an online adaptive prediction method for epileptic seizures.

[0117] Finally, it should be noted that the above embodiments are only used to illustrate the technical methods of the present invention and not to limit them. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the technical methods of the present invention, and these modifications or equivalent substitutions cannot cause the modified technical methods to deviate from the spirit and scope of the technical methods of the present invention.

Claims

1. An online adaptive prediction method for epileptic seizures, characterized in that, The online adaptive prediction method for epileptic seizures includes the following steps: S1. Acquire multi-channel EEG signals from the target user; S2. Based on fusion spatial constraint independent component analysis, brain functional networks, and transfer learning mechanisms, an online adaptive prediction model for epileptic seizures is constructed. S3. Input multi-channel EEG signals into the online adaptive prediction model for epileptic seizures to identify and predict the pre-seizure state and obtain prediction results. S4. Based on the prediction results, make an online early warning judgment, and trigger an alarm if the early warning conditions are met; The specific content of S2, which constructs an online adaptive prediction model for epileptic seizures based on fusion spatial constraint independent component analysis, brain functional networks, and transfer learning mechanisms, includes: Acquire users' multi-channel historical EEG signals and epileptic seizure data; Training and testing sets were constructed based on multi-channel historical EEG signals and epileptic seizure data. Based on fusion spatial constraint independent component analysis, an EEG noise estimation module is constructed; Based on brain functional networks, a brain electroencephalogram (EEG) feature information mapping module was constructed. Based on the transfer learning mechanism, an EEG information prediction module was constructed; By connecting the EEG noise estimation module, the EEG feature information mapping module, and the EEG information prediction module, an initial model for the online adaptive prediction model of epileptic seizures is obtained. The initial model is trained and tested based on the training set and the test set, and the trained and tested initial model is used as an online adaptive prediction model for epileptic seizures. Based on fusion spatially constrained independent component analysis, the specific content of the EEG noise estimation module includes: The multi-channel historical EEG signals were filtered to obtain the filtered signal; Based on the spatial constraint matrix and the fusion spatial constraint independent component analysis, the filtered signal is separated to obtain the filtered signal components; Wavelet threshold denoising is performed on the filtered signal components to obtain the noise components; Perform an inverse projection transformation on the noise components to obtain the noise estimate; The noise estimate in the filtered signal is removed to obtain a clean EEG signal, and an EEG noise estimation module is obtained. Based on brain functional networks, the specific content of constructing the EEG feature information mapping module includes: Frequency slice wavelet transform was used to extract the rhythm from the pure EEG signal to obtain the rhythmic EEG signal; A multi-track autoregressive module was established using rhythmic EEG signals as the object, and the coefficients of the multi-track autoregressive module were solved by the least squares method. Based on the partial orientation coherence method, a brain functional network diagram and connectivity matrix are constructed according to the multi-track autoregressive module coefficients. The brain functional network diagram and connectivity matrix are used as feature images for mapping EEG information to obtain the EEG feature information mapping module. Based on the transfer learning mechanism, the specific content of constructing the EEG information prediction module includes: Based on the feature images, a bidirectional long short-term memory network is constructed by applying inter-domain joint knowledge transfer to obtain the EEG information prediction module.

2. The online adaptive prediction method for epileptic seizures according to claim 1, characterized in that, The bidirectional long short-term memory network is based on the structural optimization of the long short-term memory network. Specifically, it adds a connection from the input gate to the output gate in the long short-term memory network structure and merges the input gate and the forget gate into an update gate.

3. The online adaptive prediction method for epileptic seizures according to claim 1, characterized in that, Online early warning judgments in S4 include: A preset warning threshold n and prediction count s are defined; where n is an integer greater than 1 and s is an integer greater than or equal to 0. If s=0, it means that the online adaptive model for epileptic seizures did not detect EEG information of epileptic premonitory signals. If s≥n, the prediction result is identified as a pre-seizure state, triggering an alarm; If s < n, the prediction result is identified as the pre-seizure state, and prediction continues until s ≥ n, at which point an alarm is triggered.

4. An online adaptive prediction system for epileptic seizures, characterized in that, include: The signal acquisition module is used to acquire multi-channel EEG signals from the target user. The model building module is used to construct an online adaptive prediction model for epileptic seizures based on fusion spatial constraint independent component analysis, brain functional networks, and transfer learning mechanisms. Based on the fusion of spatially constrained independent component analysis, brain functional networks, and transfer learning mechanisms, the specific content of constructing an online adaptive prediction model for epileptic seizures includes: Acquire users' multi-channel historical EEG signals and epileptic seizure data; Training and testing sets were constructed based on multi-channel historical EEG signals and epileptic seizure data. Based on fusion spatial constraint independent component analysis, an EEG noise estimation module is constructed; Based on brain functional networks, a brain electroencephalogram (EEG) feature information mapping module was constructed. Based on the transfer learning mechanism, an EEG information prediction module was constructed; By connecting the EEG noise estimation module, the EEG feature information mapping module, and the EEG information prediction module, an initial model for the online adaptive prediction model of epileptic seizures is obtained. The initial model is trained and tested based on the training set and the test set, and the trained and tested initial model is used as an online adaptive prediction model for epileptic seizures. Based on fusion spatially constrained independent component analysis, the specific content of the EEG noise estimation module includes: The multi-channel historical EEG signals were filtered to obtain the filtered signal; Based on the spatial constraint matrix and the fusion spatial constraint independent component analysis, the filtered signal is separated to obtain the filtered signal components; Wavelet threshold denoising is performed on the filtered signal components to obtain the noise components; Perform an inverse projection transformation on the noise components to obtain the noise estimate; The noise estimate in the filtered signal is removed to obtain a clean EEG signal, and an EEG noise estimation module is obtained. Based on brain functional networks, the specific content of constructing the EEG feature information mapping module includes: Frequency slice wavelet transform was used to extract the rhythm from the pure EEG signal to obtain the rhythmic EEG signal; A multi-track autoregressive module was established using rhythmic EEG signals as the object, and the coefficients of the multi-track autoregressive module were solved by the least squares method. Based on the partial orientation coherence method, a brain functional network diagram and connectivity matrix are constructed according to the multi-track autoregressive module coefficients. The brain functional network diagram and connectivity matrix are used as feature images for mapping EEG information to obtain the EEG feature information mapping module. Based on the transfer learning mechanism, the specific content of constructing the EEG information prediction module includes: Based on the feature images, a bidirectional long short-term memory network is constructed by applying inter-domain joint knowledge transfer to obtain the EEG information prediction module; The online prediction module is used to input multi-channel EEG signals into the online adaptive prediction model for epileptic seizures to identify and predict the pre-seizure state and obtain prediction results. The online early warning module is used to make online early warning judgments based on the prediction results, and to trigger an alarm if the early warning conditions are met.

5. A computer device, comprising: The memory, the processor, and the computer program stored in the memory and executable on the processor are characterized in that the processor executes the computer program to implement the content of the online adaptive prediction method for epileptic seizures according to any one of claims 1-3.

6. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the content of the online adaptive prediction method for epileptic seizures as described in any one of claims 1-3.

Images