A method for generating electrical stimulation signals based on brain-computer interface and related equipment
A brain-computer interface method processes EEG, ECG, and heart sound signals to generate personalized electrical stimulation signals for autonomic neural regulation, addressing the limitations of current treatments and enhancing treatment efficacy and comfort.
Patent Information
- Application Number
- CN202411275729.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-12
- Publication Date
- 2024-12-03
- Estimated Expiration
- 2044-09-12
AI Technical Summary
Current treatments for autonomic dysfunction, such as drug therapy and invasive procedures like stellate ganglion block, are subjective, invasive, and inconvenient, while non-invasive brain-computer interfaces are not suitable for daily use due to discomfort and tissue damage.
A method and device using a brain-computer interface to generate electrical stimulation signals by analyzing EEG, ECG, and heart sound signals, processing them through a neural regulation model to regulate autonomic nervous activity, incorporating denoising, normalization, feature extraction, and fusion techniques to generate personalized autonomic neural regulation signals.
Provides a non-invasive, personalized, and effective method for autonomic neural regulation, reducing discomfort and improving treatment efficacy by using a brain-computer interface to analyze and generate electrical stimulation signals based on real-time physiological data.
Smart Images

Figure CN118797560B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of data processing technology, and in particular to a method for generating electrical stimulation signals based on a brain-computer interface and related equipment. Background Art
[0002] Autonomic dysfunction, especially overactivation of the sympathetic nerves, can lead to functional and organic diseases in various organ systems of the human body. Current treatments include drug therapy such as beta-blockers, and non-drug treatments include stellate ganglion block and auricular vagus nerve stimulation. However, traditional drug therapy has many limitations and disadvantages. For example, when clinicians use beta-blockers, they need to adjust the drug dosage mainly according to the patient's heart rate, which is subjective and dependent. After adjusting the drug dosage, the speed of onset affects the patient's treatment effect. When the drug is overdosed, the speed of drug metabolism affects the patient's safety.
[0003] Current non-drug treatment measures are also very limited. For example, the stellate ganglion block currently being performed clinically requires invasive surgery on the neck, which is difficult for patients to accept and requires repeated visits to the hospital, so its clinical development is limited.
[0004] In recent years, with the rapid development of artificial intelligence, brain-computer interfaces (BCI), organoids and other technical fields, the theories of EEG signals and the functional localization of corresponding brain areas have been gradually clarified, which has also brought new hope for the quantitative and qualitative evaluation of the function of the non-invasive autonomic nervous system. Most traditional brain-computer interfaces are non-invasive cap-based, commercial headband-based products or invasive microneedle-based methods, but these methods are not suitable for daily use due to inconvenience, inflammation and even irreversible damage to human tissues.
[0005] It should be noted that the information disclosed in the above background technology section is only used to enhance the understanding of the background of the present disclosure, and therefore may include information that does not constitute the prior art known to ordinary technicians in the field. Summary of the invention
[0006] The purpose of this application is to provide a method and related equipment for generating electrical stimulation signals based on a brain-computer interface, which at least overcomes the problems existing in the prior art to a certain extent, and generates multimodal data of the target user by comprehensively analyzing the physiological data, medical use information, historical physiological parameter information and real-time somatosensory information of the target user, including EEG signals, ECG signals and heart sound signal data. After feature extraction and fusion processing, these data can form physiological characteristic risk influencing factors. Using the target neural regulation early warning model, these multimodal data and the user's real-time somatosensory information are deeply analyzed to generate autonomic neural activity signal data. Finally, the model processes these signal data to generate electrical stimulation signals to achieve the purpose of autonomic neural regulation of the target user.
[0007] Other features and advantages of the present application will become apparent from the following detailed description, or may be learned in part by the practice of the present invention.
[0008] According to one aspect of the present application, a method for generating an electrical stimulation signal based on a brain-computer interface is provided, comprising: obtaining physiological data of a target user, medical use information of the target user, historical physiological parameter information of the target user, and real-time somatosensory information of the target user; preprocessing the physiological data of the target user to generate multimodal data of the target user, wherein the multimodal data of the target user includes EEG signal data of the target user, ECG signal data of the target user, and heart sound signal data of the target user; extracting features of the multimodal data of the target user based on the medical use information of the target user and the historical physiological parameter information of the target user to generate physiological feature risk influencing factors; fusing the physiological feature risk influencing factors to generate target multimodal data; obtaining a target neural regulation early warning model; processing the target multimodal data and the real-time somatosensory information of the target user based on the target neural regulation early warning model to generate autonomic nervous activity signal data of the target user; processing the autonomic nervous activity signal data of the target user based on the target neural regulation early warning model to generate an electrical stimulation signal, wherein the electrical stimulation signal is used to perform autonomic nervous regulation on the target user.
[0009] In one embodiment of the present application, the physiological data of the target user is preprocessed to generate multimodal data of the target user, including: denoising the physiological data of the target user to generate physiological attribute information of the target user, wherein the physiological attribute information of the target user is physiological data with a signal-to-noise ratio higher than a preset threshold; normalizing the physiological attribute information of the target user to generate multimodal data of the target user.
[0010] In one embodiment of the present application, feature extraction is performed on the multimodal data of the target user based on the medical use information of the target user and the historical physiological parameter information of the target user to generate a physiological feature risk influencing factor, including: feature extraction is performed on the multimodal data of the target user to generate physiological feature classification information, wherein the physiological feature classification information includes physiological feature data corresponding to several types; the physiological feature classification information is processed based on the medical use information of the target user to generate target physiological feature data; the target physiological feature data is processed based on the historical physiological parameter information of the target user to generate a physiological feature risk influencing factor.
[0011] In one embodiment of the present application, feature extraction processing is performed on the multimodal data of the target user to generate physiological feature classification information, including: processing the multimodal data of the target user to generate biological sign attribute information of the target user; performing feature extraction processing on the biological sign attribute information of the target user to generate a number of feature sample types; performing feature screening processing on the several feature sample types to generate real-time feature sample data; and generating physiological feature classification information based on the several real-time feature sample data.
[0012] In one embodiment of the present application, the physiological characteristic risk influencing factors are fused to generate target multimodal data, including: fusing the physiological characteristic risk influencing factors based on a preset long short-term memory network model to generate an initial physiological characteristic vector, wherein the initial physiological characteristic vector is composed of physiological characteristic risk influencing factors of several modalities; processing the initial physiological characteristic vector based on a preset weight feature processing rule to generate a target physiological characteristic vector, wherein the target physiological characteristic vector is obtained by adding physiological characteristic risk influencing factors of different modalities based on target weights; processing the target physiological characteristic vector based on a preset long short-term memory network model to generate a prediction result of multimodal data; processing the prediction result of the multimodal data to generate target multimodal data; the method also includes a calculation formula for obtaining the target physiological characteristic vector, and the calculation formula is: ;in, is the target physiological feature vector, is the weight of the ith feature, is the risk influencing factor of the ith feature.
[0013] In one embodiment of the present application, a target neuromodulation warning model is obtained, including: processing the physiological characteristic risk influencing factor to generate identification information, wherein the identification information is used to characterize that the physiological characteristic risk influencing factor is in an abnormal state; obtaining an initial neuromodulation warning model that matches the identification information and a training sample set that matches the initial neuromodulation warning model; processing the training sample set based on preset processing rules to generate a training set and a verification set; processing the initial neuromodulation warning model based on the training set and the verification set to generate a target neuromodulation warning model.
[0014] In one embodiment of the present application, the target multimodal data and the real-time somatosensory information of the target user are processed based on the target neural regulation warning model to generate autonomic nervous activity signal data of the target user, including: processing the target multimodal data based on the target neural regulation warning model to generate skin conductance change information of the target user and skin temperature change information of the target user; processing the real-time somatosensory information of the target user based on the target neural regulation warning model to generate skin contraction information of the target user; processing the skin conductance change information and the skin temperature change information of the target user based on the skin contraction information of the target user to generate autonomic nervous activity signal data of the target user.
[0015] Another aspect of the present application is a device for generating electrical stimulation signals based on a brain-computer interface, characterized in that it includes: an acquisition module for acquiring physiological data of a target user, medical use information of the target user, historical physiological parameter information of the target user, and real-time somatosensory information of the target user; acquiring a target neural regulation early warning model; a processing module for preprocessing the physiological data of the target user to generate multimodal data of the target user, wherein the multimodal data of the target user includes EEG signal data of the target user, ECG signal data of the target user, and heart sound signal data of the target user; extracting features of the multimodal data of the target user based on the medical use information of the target user and the historical physiological parameter information of the target user to generate physiological feature risk influencing factors; fusing the physiological feature risk influencing factors to generate target multimodal data; processing the target multimodal data and the real-time somatosensory information of the target user based on the target neural regulation early warning model to generate autonomic nervous activity signal data of the target user; processing the autonomic nervous activity signal data of the target user based on the target neural regulation early warning model to generate an electrical stimulation signal, wherein the electrical stimulation signal is used to perform autonomic nervous regulation on the target user.
[0016] According to another aspect of the present application, an electronic device is provided, characterized in that it comprises: a first processor; and a memory for storing executable instructions of the first processor; wherein the first processor is configured to execute the above-mentioned method of generating electrical stimulation signals based on a brain-computer interface by executing the executable instructions.
[0017] According to another aspect of the present application, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a second processor, the method for generating an electrical stimulation signal based on a brain-computer interface is implemented.
[0018] According to another aspect of the present application, a computer program product is provided, comprising a computer program, wherein the computer program implements the above-mentioned method of generating electrical stimulation signals based on a brain-computer interface when executed by a third processor.
[0019] The present application provides a method and related equipment for generating electrical stimulation signals based on a brain-computer interface. The server generates multimodal data of the target user by comprehensively analyzing the target user's physiological data, medical use information, historical physiological parameter information and real-time somatosensory information. The data includes EEG signals, ECG signals and heart sound signal data. After feature extraction and fusion processing, these data can form physiological characteristic risk influencing factors. Using the target neural regulation early warning model, these multimodal data and the user's real-time somatosensory information are deeply analyzed to generate autonomic neural activity signal data. Finally, the model processes these signal data to generate electrical stimulation signals to achieve the purpose of autonomic neural regulation of the target user.
[0020] It is to be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present disclosure. BRIEF DESCRIPTION OF THE DRAWINGS
[0021] Figure 1 A flowchart showing a method for generating an electrical stimulation signal based on a brain-computer interface provided in an embodiment of the present application is shown;
[0022] Figure 2 A schematic diagram of the structure of a device for generating an electrical stimulation signal based on a brain-computer interface provided in one embodiment of the present application is shown;
[0023] Figure 3 A schematic diagram of the structure of an electronic device provided by an embodiment of the present application is shown;
[0024] Figure 4 A schematic diagram of a storage medium provided in an embodiment of the present application is shown. DETAILED DESCRIPTION
[0025] The preferred embodiments of the present invention are described below in conjunction with the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.
[0026] Combine the following Figure 1 To describe the method for generating electrical stimulation signals based on a brain-computer interface according to an exemplary embodiment of the present application. It should be noted that the following application scenarios are only shown to facilitate understanding of the spirit and principles of the present application, and the embodiments of the present application are not limited in this regard. On the contrary, the embodiments of the present application can be applied to any applicable scenario.
[0027] In one embodiment, the present application also proposes a method for generating electrical stimulation signals based on a brain-computer interface and related equipment. Figure 1 The flowchart of a method for generating an electrical stimulation signal based on a brain-computer interface according to an embodiment of the present application is schematically shown. Figure 1 As shown, the method is applied to a server, comprising:
[0028] S101, acquiring physiological data of a target user, medical use information of the target user, historical physiological parameter information of the target user, and real-time body sensory information of the target user.
[0029] In one embodiment, various physiological monitoring instruments are used, such as the wireless physiological analysis system PhysioLAB291, which can record ECG, EEG, EMG, EDA, blood volume and pulse, temperature, respiration, acceleration and other data. The medical use information of the target user can be determined according to their specific needs and symptoms. For example, the medical use information of the target user can be determined according to their specific needs and symptoms. The following are application scenarios for different symptoms: headache, lack of sleep, palpitations, fatigue, memory loss, etc. In addition, if the duration of onset of different symptoms and the time currently reserved for treating the symptoms are different, different means will be used to process them, which are customized according to the specific health status and needs of the user, aiming to provide personalized health management and disease prevention and treatment solutions.
[0030] The target user's historical physiological parameter information is crucial for personalized medicine and precision treatment. By analyzing these historical data, doctors can better understand the patient's health status, identify disease patterns, and develop the most appropriate treatment plan. The use of big data technology enables doctors to analyze the target user's historical physiological parameter information, including massive medical data such as genomic data, clinical test results, and medical imaging data, and design a more accurate and effective treatment plan for patients. By analyzing the patient's historical medical history, physiological data, and living habits, big data technology helps predict the patient's future possible diseases, optimize the allocation of medical resources, ensure that patients can obtain the most appropriate medical resources, and provide patients with personalized risk prediction, diagnosis, and treatment plans through the mining of personal genomes and other biological big data, and optimize the allocation of medical resources.
[0031] The target user's real-time somatosensory information, especially the skin's reaction and temperature changes, are crucial for assessing their acceptance of the current treatment level. Among them, the change in skin temperature is a manifestation of vascular reaction, which is affected by the external temperature and is also controlled by the body's regulatory mechanism. When the external temperature drops, the subcutaneous capillaries contract to reduce heat dissipation, resulting in a decrease in skin temperature; conversely, heat dissipation increases and the skin temperature rises. The regulation of skin temperature is mainly achieved through the reaction of blood vessels and sweat glands, which is an important part of body temperature regulation. Changes in skin temperature can reflect the heat exchange between the human body and the environment, which is of great significance for maintaining a stable body temperature. By real-time monitoring of the target user's skin temperature and other physiological parameters, medical professionals can more accurately determine whether the patient can accept the current treatment level and adjust the treatment plan in a timely manner.
[0032] S102: Preprocess the physiological data of the target user to generate multimodal data of the target user.
[0033] In one embodiment, the physiological data of the target user is denoised to generate physiological attribute information of the target user, wherein the physiological attribute information of the target user is physiological data with a signal-to-noise ratio higher than a preset threshold, and the physiological attribute information of the target user is normalized to generate multimodal data of the target user, wherein the multimodal data of the target user includes but is not limited to the electroencephalogram signal data of the target user, the electrocardiogram signal data of the target user, and the heart sound signal data of the target user.
[0034] Specifically, various sensors are used to collect the user's physiological data, such as electrocardiogram, electroencephalogram, and electromyography signals. The collected physiological data is filtered, artifact-removed, and other operations are performed to reduce the noise components in the data. For example, in brain functional magnetic resonance imaging (fMRI) data, the physiological interference of heart and respiratory signals has a significant impact on data quality. It is necessary to adopt data-driven denoising methods or use filters (such as low-pass filtering, high-pass filtering, and band-pass filtering) to estimate and eliminate these noises, and extract useful physiological features from the denoised signals, such as heart rate variability and electroencephalogram waveform characteristics. The signal-to-noise ratio (SNR) of the denoised physiological data is evaluated to ensure data quality. The signal-to-noise ratio is a measure of the ratio of the useful signal strength to the background noise strength in the signal. A high signal-to-noise ratio means better signal quality. Physiological data with a signal-to-noise ratio higher than a preset threshold is regarded as valid data, and further analysis is performed to generate physiological attribute information of the target user.
[0035] Normalize the physiological attribute information so that data from different sources or different dimensions can be compared and merged on the same scale. Normalize the signal to eliminate the scale differences between different data sources and ensure the consistency and comparability of the input data. Normalization is to scale the data so that it falls into a specific range, such as the [0,1] interval, while standardization is to centralize and unify the scale of the data by subtracting the mean and dividing by the standard deviation. Generate multimodal data for the normalized physiological attribute information, which may include EEG signal data, ECG signal data, and heart sound signal data. The generation of multimodal data helps to comprehensively evaluate the user's health status from different perspectives. During the recording of in vivo electrophysiological signals, the signal quality is monitored in real time and adjusted according to feedback to ensure a high signal-to-noise ratio of the data. Through the above steps, the physiological data of the target user can be effectively processed, high-quality physiological attribute information can be extracted, and support can be provided for subsequent analysis and evaluation.
[0036] S103, performing feature extraction on the multimodal data of the target user based on the medical use information of the target user and the historical physiological parameter information of the target user to generate a physiological feature risk influencing factor.
[0037] In one implementation, feature extraction is performed on the multimodal data of the target user to generate physiological feature classification information, wherein the physiological feature classification information includes physiological feature data corresponding to several types, extracting features such as RR interval and heart rate variability from the electrocardiogram signal; extracting S1 and S2 sound peak values, audio spectrum features, etc. from the heart sound signal; and extracting α wave and β wave power spectrum features, etc. from the electroencephalogram signal.
[0038] Specifically, for feature extraction of EEG signals, the fast Fourier transform (FFT) can be used to convert the time domain signal into the frequency domain signal, and the main frequency components can be extracted. The following calculation formula can be used to convert the time domain signal into the frequency domain signal: ; Where x(n) is the time domain signal and X(f) is the frequency domain signal. Wavelet transform (WT) is used to perform multi-scale decomposition of EEG signals and extract features of different frequency bands. Wavelet transform can perform multi-scale decomposition of EEG signals, which is achieved through continuous wavelet transform (CWT) or discrete wavelet transform (DWT). The decomposed signal can be analyzed at different time and frequency resolutions, which helps to capture the non-stationary characteristics and instantaneous changes in EEG signals. A variety of features can be extracted from the wavelet transformed signal, such as energy, frequency distribution, and statistical characteristics of wavelet coefficients. These features reflect the activity intensity and distribution of EEG signals in different frequency bands, which are of great significance for understanding the working state of the brain and pattern recognition.
[0039] For feature extraction of ECG signals, the Pan-Tomkins algorithm can be used to detect QRS complexes, extract heart rate and heart rate variability (HRV), and calculate time domain indicators of heart rate variability, such as standard deviation (SDNN) and root mean square deviation (RMSSD). The corresponding calculation formulas are as follows: , .
[0040] For feature extraction of heart sound signals, short-time Fourier transform (STFT) can be used to extract the time-frequency features of heart sound signals, and to extract features such as the peak value, energy, and frequency component of heart sound for analysis of heart sound signals.
[0041] The physiological characteristic classification information is processed based on the medical use information of the target user to generate the target physiological characteristic data. The medical use information of the target user can be determined based on their specific needs and symptoms. For example, if the target user experiences headaches in the morning but needs to attend an important meeting in the afternoon, they may need a treatment plan that can quickly relieve the symptoms. In this case, the selected physiological characteristic data will focus on those characteristics that can respond quickly and support temporary relief of headaches, rather than data required for long-term treatment plans.
[0042] Specifically, if the goal is to completely cure headaches, a comprehensive set of physiological characteristic data may be required, including but not limited to blood pressure, heart rate, blood oxygen saturation, etc., which can be collected in real time through various medical monitoring devices. However, if the goal is temporary relief, it may only be necessary to focus on a few key physiological characteristic data directly related to headaches, such as heart rate variability or specific neural activity patterns.
[0043] The target physiological characteristic data is processed based on the historical physiological parameter information of the target user to generate physiological characteristic risk influencing factors, where the physiological characteristic risk influencing factors include but are not limited to heart rate variability, EEG spectrum characteristics, and PCG audio characteristics. Mainly by analyzing the individual's physiological data, identifying abnormal deviations between physiological parameters, so as to evaluate and locate health risks, and identifying which physiological parameters have changed beyond the normal range or the deviations between parameters are too large through statistical analysis or machine learning algorithms. For example, the Pan-Tompkins algorithm can be used to process the electrocardiogram signal, extract the heart rate and heart rate variability (HRV), and analyze the changing trends of these parameters. When abnormal changes in physiological parameters are detected, early warning information is issued in time to help users or doctors take preventive measures. In addition, the physiological parameter baselines of different individuals may be different, so the specific circumstances of the individual should be considered in risk assessment to provide personalized assessment results. The user's physiological characteristic data can be effectively processed to generate clinically significant risk influencing factors, providing a scientific basis for health management and disease prevention.
[0044] In another embodiment, feature extraction is performed on the multimodal data of the target user to generate physiological feature classification information. For details, please refer to the following content: the multimodal data of the target user is processed to generate biological sign attribute information of the target user. The multimodal data of the target user is processed, which includes the collection and preprocessing of signals such as electroencephalogram (EEG), electromyography (EMG), and heart rate (ECG). For example, the EEG signal can be feature extracted through a deep learning network such as ResNets-50, and the EOG and EMG signals can be processed using Mel frequency cepstral coefficients (MFCC), wherein the biological sign attribute information of the target user is used to characterize the current physical state information of the target user, that is, what symptoms the target user currently has, the reaction and duration of the symptoms, etc.
[0045] The target user's biological sign attribute information is subjected to feature extraction processing to generate several feature sample types. Specifically, the corresponding physiological parameter information is selected based on the target user's physical state. When performing feature extraction processing, the corresponding feature extraction technology can be used for different types of physiological signals to generate feature sample types. Among them, for the feature extraction of EEG signals, the following methods can be used: time domain features: such as zero crossing rate, standard deviation, maximum / minimum value, skewness, kurtosis, etc.; frequency domain features: power spectral density (PSD), band power, frequency peak, etc. obtained by Fourier transform; time-frequency domain features: time-frequency representation obtained by wavelet transform, such as wavelet energy, wavelet entropy, etc.; nonlinear features: such as approximate entropy, sample entropy, Lyapunov exponent, etc. For feature extraction of electromyographic signals, the following methods can be used: time domain features: including mean, variance, root mean square, waveform factor, skewness, kurtosis, etc.; frequency domain features: frequency distribution features obtained by Fourier transform, such as mean frequency, median frequency, band energy, etc.; time-frequency domain features: time-varying features extracted using wavelet transform; nonlinear dynamic features: such as Hjorth parameter, correlation dimension, etc. For feature extraction of heart rate signals, the following methods can be used: time domain features: standard deviation of heart rate, RR interval, maximum / minimum heart rate, etc.; frequency domain features: heart rate variability (HRV) indicators, such as SDNN, RMSSD, power spectral density, etc.; time-frequency domain features: use wavelet transform to analyze the non-stationary characteristics of heart rate signals; morphological features: width, amplitude, area of QRS complex, etc. For feature extraction of skin electrical activity, the following methods can be used: amplitude, rise time, recovery time, etc.; event-related skin electrical activity (phasicEDA) and nonspecific skin electrical activity (tonicEDA). The feature extraction of each physiological signal depends on the characteristics of the signal and the purpose of analysis. After feature extraction, multiple feature sample types can be generated, each type contains a set of specific features that can reflect different aspects of the physiological signal. In practical applications, specific feature types may be selected for analysis to achieve the best performance according to the requirements of specific tasks.
[0046] Perform feature screening on several feature sample types to generate real-time feature sample data, and generate physiological feature classification information based on several real-time feature sample data. Feature screening is a key step in the feature extraction process of multimodal biological sign attribute information. It helps to identify the features that are most important to model performance from a large number of features, thereby generating real-time feature sample data and generating physiological feature classification information based on these data. Feature screening can be achieved through a variety of methods, including statistical-based methods (such as correlation analysis, mutual information method), model-based methods (such as using random forests to evaluate feature importance), and embedded methods (such as automatically learning the importance of features during neural network training). These methods can help applicants identify the features that have the greatest impact on predicting target variables.
[0047] After feature screening, the applicant can obtain a streamlined feature set. The generation of real-time feature sample data needs to consider the timeliness of the data to ensure that the features can reflect the user's current status in a timely manner. This process may involve the use of dimensionality reduction techniques such as principal component analysis (PCA) to remove redundant information and retain the most representative features. Using the screened real-time feature sample data, a classification model can be constructed to generate physiological feature classification information. This may involve the use of machine learning algorithms such as support vector machines (SVM), random forests, deep learning, etc. to classify and identify feature data. The classification results can provide a basis for further health assessment and disease diagnosis, effectively process multimodal biological sign attribute information, screen out key features, and generate real-time physiological feature classification information, providing a scientific basis for health monitoring and disease prevention.
[0048] S104, fusing the physiological characteristic risk influencing factors to generate target multimodal data.
[0049] In one embodiment, the physiological characteristic risk influencing factors are fused based on a preset long short-term memory network model to generate an initial physiological characteristic vector, wherein the initial physiological characteristic vector is composed of physiological characteristic risk influencing factors of several modalities. After obtaining the physiological characteristic risk influencing factors (such as electroencephalogram (EEG), electromyogram (EMG), etc.), the feature data of different modalities are combined into a feature vector through feature-level fusion, and then sent to the classifier. It should be noted that there may be complementary information between different modalities, and the rational use of this information can effectively improve the recognition effect. At the same time, the method of feature fusion can be a fusion of data level, feature level, decision level or model level. The specific method selected depends on the characteristics of the data and the requirements of the recognition task. Through these steps, multimodal data can be effectively processed and useful physiological characteristic classification information can be generated to support further health assessment and disease diagnosis.
[0050] The initial physiological feature vector is processed based on the preset weight feature processing rules to generate a target physiological feature vector, where the target physiological feature vector is obtained by adding the physiological feature risk influencing factors of different modalities based on the target weight. First, it is necessary to identify and determine which physiological features are risk influencing factors, such as heart rate, blood pressure, respiratory rate, etc., and assign a weight to each risk influencing factor. These weights reflect the importance of each factor in the overall risk assessment. The allocation of weights can be based on expert knowledge, statistical analysis, or the prediction results of machine learning models.
[0051] Since the dimensions and value ranges of different physiological characteristics may be different, the initial physiological characteristic vector needs to be standardized to ensure that the contribution of each characteristic is comparable during weighted summation. The risk impact factor of each physiological characteristic is multiplied by its corresponding weight, and then all the results are added together to obtain the target physiological characteristic vector. This process can be expressed by the following mathematical formula: ;in, is the target physiological feature vector, is the weight of the ith feature, is the risk influencing factor of the ith feature.
[0052] The generated target physiological feature vector should be able to reflect the individual's physiological state and health risks. This vector can be used for further analysis, such as classification, clustering, or as input to a machine learning model. In practical applications, it may be necessary to update the weight distribution in real time based on new data and feedback, and adjust the target physiological feature vector to adapt to possible changes and new physiological features. It can effectively integrate multimodal physiological data to generate a feature vector that comprehensively reflects the individual's health status, providing support for health risk assessment and disease diagnosis.
[0053] The target physiological feature vector is processed based on a preset long short-term memory network model to generate prediction results of multimodal data, and the prediction results of multimodal data are processed to generate target multimodal data. Long short-term memory network (LSTM) is a powerful recurrent neural network that can process sequence data, especially for problems with long-term dependencies. When processing the target physiological feature vector, LSTM can learn the time dependency in the sequence data and generate prediction results of multimodal data. The model can receive the target physiological feature vector as input. The construction of the model may include defining the number of neurons in the LSTM layer, selecting the activation function (such as tanh or ReLU), and determining the configuration of the output layer, and training the LSTM model using historical physiological data. During the training process, LSTM will learn the long-term and short-term dependencies between features, so that it can effectively predict multimodal data, where the target multimodal data is generated based on the symptoms that the target user currently needs to solve urgently.
[0054] S105, obtaining a target neuromodulation warning model.
[0055] In one embodiment, the physiological characteristic risk influencing factor is processed to generate identification information, wherein the identification information is used to characterize that the physiological characteristic risk influencing factor is in an abnormal state; an initial neuromodulation warning model matching the identification information and a training sample set matching the initial neuromodulation warning model are obtained; the training sample set is processed based on preset processing rules to generate a training set and a verification set; the initial neuromodulation warning model is processed based on the training set and the verification set to generate a target neuromodulation warning model.
[0056] In another embodiment, the training data set is processed based on preset processing rules to generate a training set and a verification set, specifically including: obtaining physiological characteristic risk influencing factors of other users that match the identification information; performing feature extraction on the physiological characteristic risk influencing factors of other users to generate a feature data set; obtaining any number of data features in the feature data set; generating adjacent features based on the distance between any number of data features and other numbers of data features of the same category, wherein the adjacent features include a preset number of any number of data features; determining a sampling ratio based on the number of each data feature in the training set; determining a sampling ratio based on the sampling ratio; sampling adjacent features based on the sampling ratio to generate a preset number of sampling features; generating multiple data groups based on any data feature and each sampling feature, wherein each data group includes a preset number of data samples, wherein each data group includes a preset number of data samples, wherein at least one data sample includes identification information; processing multiple data groups to generate a training set and a verification set, wherein both the training set and the verification set include a number of data features.
[0057] Specifically, in data analysis and machine learning, the sampling ratio and sampling rate are determined based on the number of data features in the training set. By counting the number of each data feature in the training set, the distribution of different features can be understood. The sampling ratio is determined according to the number of data features and the size of the data set. The sampling ratio can help us decide how many samples to extract from each feature. According to the sampling ratio, the specific sampling ratio can be calculated. The sampling ratio is the ratio of the actual number of samples to the total number. The adjacent features are sampled using the sampling ratio to generate a preset number of sampling features. This process ensures the representativeness and diversity of the samples. Through the sampling process, a set of sampling features is generated, which will be used for subsequent data analysis and model training. Multiple data sets are constructed based on any data feature and each sampling feature. Each data set contains a preset number of data samples. It is ensured that the samples in each data set are diverse and contain at least one sample with identification information, which helps the model learn the feature performance in different situations. These data sets are used for model training to evaluate the contribution of different features to the prediction target. The machine learning model is trained using the sampling feature set, and the performance of the model is evaluated through the validation set. According to the results of model training and validation, the sampling ratio and rate may need to be adjusted to optimize the prediction ability of the model. With this approach, large-scale data sets can be processed efficiently while maintaining data quality and model generalization capabilities, helping to identify potential security risks and abnormal behaviors.
[0058] S106: Process the target multimodal data and the real-time somatosensory information of the target user based on the target neural regulation warning model to generate autonomic neural activity signal data of the target user.
[0059] In one embodiment, the target multimodal data is processed based on the target neural regulation warning model to generate the target user's skin conductance change information and the target user's skin temperature change information. When the target multimodal data is processed based on the target neural regulation warning model, the target user's skin conductance change information and skin temperature change information are two key physiological parameters. Galvanic skin response (GSR) is a method for monitoring changes in skin conductance, which can reflect the activity state of the sympathetic nerves. When the human body experiences emotional changes or physiological arousal, sweat gland activity increases, resulting in measurable changes in skin conductance. Skin temperature monitoring is usually achieved through a contact temperature sensor, which can indirectly reflect the activity of the autonomic nerves, because changes in skin temperature mainly depend on the blood flow through the skin, and the amount of blood flow is affected by the excitability of the sympathetic nerves.
[0060] In practical applications, multimodal data fusion technology can effectively integrate information from different modalities and improve the performance of the model. It can project single-modal features into a shared semantic subspace in different ways to achieve feature fusion. These technologies can be applied to integrate multimodal physiological data such as skin conductance and skin temperature to generate more comprehensive physiological feature classification information. In terms of monitoring skin temperature changes, studies have shown that the increase in skin temperature can be used as an objective indicator to evaluate the effect of regional blockade, especially on the hairless skin at the distal end of the limbs. The increase in skin temperature after sympathetic nerve blockade is particularly obvious. It can effectively process and analyze physiological parameters such as skin conductance and skin temperature, and provide accurate physiological feature classification information for the neuromodulation early warning model.
[0061] The target user's real-time somatosensory information is processed based on the target nerve regulation warning model to generate the target user's skin contraction information. The target user's skin conductance change information and the target user's skin temperature change information are processed based on the target user's skin contraction information to generate the target user's autonomic nerve activity signal data. That is, since each user has certain differences in nerve response, for example, for the same symptom, some people may have obvious reactions at the beginning of the symptom, while some people may not react until the symptom continues to develop to a certain extent. At this time, the degree of the user's symptoms is judged by the target user's skin conductance change information and skin temperature change information, and the reaction of each user is judged by referring to the skin contraction information of each user, so as to generate autonomic nerve activity signal data that conforms to each user.
[0062] S107, processing the autonomic nerve activity signal data of the target user based on the target nerve regulation warning model to generate an electrical stimulation signal.
[0063] In one embodiment, in order to reduce the problem of the prior art directly setting the target voltage when starting the electrical stimulation device, the patient may feel strong stimulation without any warning, thereby affecting the patient's comfort. When this solution receives the autonomic nervous activity signal data of the target user, a single stimulation is composed of multiple pulse waveforms with a fixed frequency, wherein the voltage of the first pulse waveform in the first half rises rapidly from 0 to the patient's target voltage value. In the subsequent pulse waveforms, all are square wave pulses of the target voltage. The present invention can effectively improve the patient's adaptability to voltage and effectively alleviate the patient's discomfort caused by electric shock every time the stimulator starts pulse stimulation.
[0064] In addition, the present invention also includes a calculation formula for generating a comfortable voltage value, which is specifically shown as follows:
[0065] .
[0066] A is the target voltage value of the pulse, and B is the rise rate constant, which defaults to 100. The speed of voltage rise can be adjusted by changing the value of B.
[0067] As a non-invasive brain-computer interface technology, the in-ear brain-computer interface is an easy-to-operate and most easily accepted method for patients. It can detect the patient's brain electrical activity. On this basis, by performing multimodal data fitting on EEG signals, autonomic nerve activity signals and ECG signals including heart rate variability analysis and other data reflecting the state of autonomic nervous function, it can more accurately reflect the activity of autonomic nervous function, thereby providing a more effective and safer detection indicator reference for ear vagus nerve stimulation.
[0068] At the same time, since the auricular branch of the vagus nerve passes through the auricular cartilage, it is distributed in the concha cavity, the posterior auricular muscle and the middle and upper part of the dorsum of the ear (external auditory canal); it also branches to the base of the helix and the triangular fossa, the antihelix and the middle part of the scaphoid. In-ear autonomic nerve regulation equipment can not only collect EEG signals through electrodes in the ear, but also activate the vagus nerve through electrical stimulation, laser stimulation, millimeter wave stimulation, ultrasonic stimulation and other methods connected to the electrodes in the concha, inhibit the over-activation of the sympathetic nerves, thereby regulating the rebalance of the autonomic nerves and restoring the body's neural-humoral-metabolic regulation homeostasis.
[0069] Therefore, this application collects multimodal data based on flexible electrodes, interactively integrates and analyzes the data through brain-computer interfaces and target neural regulation warning models, and then uses different stimulation methods to regulate the autonomic nervous system, which has great innovation and clinical application value.
[0070] By applying the above technical solution, the server obtains the physiological data of the target user, the medical use information of the target user, the historical physiological parameter information of the target user and the real-time somatosensory information of the target user; performs denoising processing on the physiological data of the target user to generate the physiological attribute information of the target user, wherein the physiological attribute information of the target user is the physiological data with a signal-to-noise ratio higher than a preset threshold; performs normalization processing on the physiological attribute information of the target user to generate the multimodal data of the target user, wherein the multimodal data of the target user includes the EEG signal data of the target user, the ECG signal data of the target user and the heart sound signal data of the target user; processes the multimodal data of the target user to generate the biological sign attribute information of the target user; performs feature extraction processing on the biological sign attribute information of the target user to generate several feature sample types; performs feature screening processing on several feature sample types to generate real-time feature sample data; generates physiological feature classification information based on several real-time feature sample data, wherein the physiological feature classification information includes physiological feature data corresponding to several types; processes the physiological feature classification information based on the medical use information of the target user to generate target physiological feature data.
[0071] The target physiological characteristic data is processed based on the historical physiological parameter information of the target user to generate the physiological characteristic risk influencing factor; the physiological characteristic risk influencing factor is fused based on the preset long short-term memory network model to generate an initial physiological characteristic vector, wherein the initial physiological characteristic vector is composed of the physiological characteristic risk influencing factors of several modalities; the initial physiological characteristic vector is processed based on the preset weight feature processing rule to generate a target physiological characteristic vector, wherein the target physiological characteristic vector is obtained by adding the physiological characteristic risk influencing factors of different modalities based on the target weight; the target physiological characteristic vector is processed based on the preset long short-term memory network model to generate the prediction result of multimodal data; the prediction result of multimodal data is processed to generate the target multimodal data; the physiological characteristic risk influencing factor is processed to generate identification information, wherein the identification information is used to characterize that the physiological characteristic risk influencing factor is in an abnormal state; an initial neuromodulation warning model matching the identification information and a training sample set matching the initial neuromodulation warning model are obtained; the training sample set is processed based on the preset processing rule to generate a training set and a validation set; the initial neuromodulation warning model is processed based on the training set and the validation set to generate a target neuromodulation warning model.
[0072] The target multimodal data is processed based on the target neural regulation early warning model to generate the target user's skin conductance change information and the target user's skin temperature change information; the target user's real-time somatosensory information is processed based on the target neural regulation early warning model to generate the target user's skin contraction information; the target user's skin conductance change information and the target user's skin temperature change information are processed based on the target user's skin contraction information to generate the target user's autonomic nerve activity signal data; the target user's autonomic nerve activity signal data is processed based on the target neural regulation early warning model to generate an electrical stimulation signal, wherein the electrical stimulation signal is used to perform autonomic nerve regulation on the target user. By comprehensively analyzing the target user's physiological data, medical use information, historical physiological parameter information and real-time somatosensory information, the target user's multimodal data is generated, including EEG signals, ECG signals and heart sound signal data, which can form physiological characteristic risk influencing factors after feature extraction and fusion processing. Using the target neural regulation early warning model, these multimodal data and the user's real-time somatosensory information are deeply analyzed to generate autonomic nerve activity signal data. Finally, the model processes these signal data to generate electrical stimulation signals to achieve the purpose of autonomic nerve regulation on the target user.
[0073] In one embodiment, if Figure 2 As shown, the present application also provides a device for generating an electrical stimulation signal based on a brain-computer interface, comprising:
[0074] The acquisition module 201 is used to acquire physiological data of the target user, medical use information of the target user, historical physiological parameter information of the target user and real-time somatosensory information of the target user; and acquire a target nerve regulation early warning model;
[0075] The processing module 202 is used to pre-process the physiological data of the target user to generate multimodal data of the target user, wherein the multimodal data of the target user includes the EEG signal data of the target user, the ECG signal data of the target user and the heart sound signal data of the target user; extract features of the multimodal data of the target user based on the medical use information of the target user and the historical physiological parameter information of the target user to generate physiological characteristic risk influencing factors; fuse the physiological characteristic risk influencing factors to generate target multimodal data; process the target multimodal data and the real-time somatosensory information of the target user based on the target neural regulation early warning model to generate autonomic nervous activity signal data of the target user; process the autonomic nervous activity signal data of the target user based on the target neural regulation early warning model to generate an electrical stimulation signal, wherein the electrical stimulation signal is used to perform autonomic nervous regulation on the target user.
[0076] In another embodiment of the present application, the processing module 202 is configured to pre-process the physiological data of the target user to generate multimodal data of the target user, including: denoising the physiological data of the target user to generate physiological attribute information of the target user, wherein the physiological attribute information of the target user is physiological data with a signal-to-noise ratio higher than a preset threshold; normalizing the physiological attribute information of the target user to generate multimodal data of the target user.
[0077] In another embodiment of the present application, the processing module 202 is configured to extract features from the multimodal data of the target user based on the medical use information of the target user and the historical physiological parameter information of the target user to generate a physiological feature risk influencing factor, including:
[0078] Extracting features from the multimodal data of the target user to generate physiological feature classification information, wherein the physiological feature classification information includes physiological feature data corresponding to several types;
[0079] Processing the physiological characteristic classification information based on the medical use information of the target user to generate target physiological characteristic data;
[0080] The target physiological characteristic data is processed based on the historical physiological parameter information of the target user to generate a physiological characteristic risk influencing factor.
[0081] In another embodiment of the present application, the processing module 202 is configured to perform feature extraction processing on the multimodal data of the target user to generate physiological feature classification information, including:
[0082] Processing the multimodal data of the target user to generate biological attribute information of the target user;
[0083] Performing feature extraction processing on the biological attribute information of the target user to generate a plurality of feature sample types;
[0084] Performing feature screening processing on the several feature sample types to generate real-time feature sample data;
[0085] Generate physiological feature classification information based on a number of real-time feature sample data.
[0086] In another embodiment of the present application, the processing module 202 is configured to perform fusion processing on the physiological characteristic risk influencing factors to generate target multimodal data, including:
[0087] Based on a preset long short-term memory network model, the physiological characteristic risk influencing factors are fused to generate an initial physiological characteristic vector, wherein the initial physiological characteristic vector is formed by splicing the physiological characteristic risk influencing factors of several modalities;
[0088] The initial physiological feature vector is processed based on a preset weight feature processing rule to generate a target physiological feature vector, wherein the target physiological feature vector is obtained by adding the physiological feature risk influencing factors of different modalities based on the target weight;
[0089] Processing the target physiological feature vector based on a preset long short-term memory network model to generate a prediction result of multimodal data;
[0090] The prediction results of the multimodal data are processed to generate target multimodal data.
[0091] In another embodiment of the present application, the processing module 202 is configured to obtain a target nerve regulation early warning model, including:
[0092] Processing the physiological characteristic risk influencing factor to generate identification information, wherein the identification information is used to indicate that the physiological characteristic risk influencing factor is in an abnormal state;
[0093] Acquire an initial neural regulation early warning model that matches the identification information and a training sample set that matches the initial neural regulation early warning model;
[0094] Process the training sample set based on preset processing rules to generate a training set and a validation set;
[0095] The initial neuromodulation early warning model is processed based on the training set and the validation set to generate a target neuromodulation early warning model.
[0096] In another embodiment of the present application, the processing module 202 is configured to process the target multimodal data and the real-time somatosensory information of the target user based on the target neural regulation early warning model to generate autonomic nerve activity signal data of the target user, including:
[0097] Processing the target multimodal data based on the target neural regulation early warning model to generate skin conductance change information and skin temperature change information of the target user;
[0098] Processing the real-time somatosensory information of the target user based on the target nerve regulation early warning model to generate skin contraction information of the target user;
[0099] The skin conductance change information of the target user and the skin temperature change information of the target user are processed based on the skin contraction information of the target user to generate autonomic nerve activity signal data of the target user.
[0100] The present application embodiment provides an electronic device, such as Figure 3 As shown, the electronic device 3 includes a first processor 300, a memory 301, a bus 302 and a communication interface 303. The first processor 300, the communication interface 303 and the memory 301 are connected via the bus 302; the memory 301 stores a computer program that can be run on the first processor 300, and when the first processor 300 runs the computer program, the method for generating electrical stimulation signals based on a brain-computer interface provided in any of the aforementioned embodiments of the present application is executed.
[0101] The memory 301 may include a high-speed random access memory (RAM), and may also include a non-volatile memory, such as at least one disk memory. The communication connection between the system network element and at least one other network element is realized through at least one communication interface 303 (which may be wired or wireless), and the Internet, wide area network, local area network, metropolitan area network, etc. may be used.
[0102] The bus 302 may be an ISA bus, a PCI bus, or an EISA bus, etc. The bus may be divided into an address bus, a data bus, a control bus, etc. Among them, the memory 301 is used to store programs, and the first processor 300 executes the program after receiving the execution instruction. The method for generating an electrical stimulation signal based on a brain-computer interface disclosed in any implementation of the embodiment of the present application may be applied to the first processor 300, or implemented by the first processor 300.
[0103] The first processor 300 may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by the hardware integrated logic circuit or software instructions in the first processor 300. The above-mentioned first processor 300 can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a readily available programmable gate array (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. The methods, steps and logic block diagrams disclosed in the embodiments of the present application can be implemented or executed. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor, etc. The steps of the method disclosed in the embodiments of the present application can be embodied as being executed by a hardware decoding processor, or by a combination of hardware and software modules in a decoding processor. The software module can be located in a mature storage medium in the field such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory or an electrically erasable programmable memory, a register, etc. The storage medium is located in the memory 301, and the first processor 300 reads the information in the memory 301 and completes the steps of the above method in combination with its hardware.
[0104] The electronic device provided in the above-mentioned embodiments of the present application and the method for generating electrical stimulation signals based on brain-computer interface provided in the embodiments of the present application are based on the same inventive concept and have the same beneficial effects as the methods adopted, run or implemented by the application programs stored therein.
[0105] The present application embodiment provides a computer-readable storage medium, such as Figure 4 As shown, the computer-readable storage medium 401 stores a computer program, and when the computer program is read and executed by the second processor 402, the method for generating an electrical stimulation signal based on a brain-computer interface as described above is implemented.
[0106] The technical solution of the embodiment of the present application, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for enabling an electronic device (which may be an air conditioner, a refrigeration device, a personal computer, a server, or a network device, etc.) or a processor to execute all or part of the steps of the method described in the embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a ROM, a RAM, a magnetic disk, or an optical disk.
[0107] The computer-readable storage medium provided in the above-mentioned embodiments of the present application and the method for generating electrical stimulation signals based on a brain-computer interface provided in the embodiments of the present application are based on the same inventive concept and have the same beneficial effects as the methods adopted, run or implemented by the application programs stored therein.
[0108] An embodiment of the present application provides a computer program product, including a computer program, wherein the computer program is executed by a third processor to implement the method described above.
[0109] The computer program product provided by the above-mentioned embodiments of the present application and the method for generating electrical stimulation signals based on brain-computer interface provided by the embodiments of the present application are based on the same inventive concept and have the same beneficial effects as the methods adopted, run or implemented by the application programs stored therein.
[0110] It should be noted that, in this application, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or still includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise one..." do not exclude the presence of other identical elements in the process, method, article or device including the elements.
[0111] Each embodiment in the present application is described in a related manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the method for evaluating the generation of electrical stimulation signals based on brain-computer interface, electronic device, electronic device, and readable storage medium embodiment, since they are basically similar to the above-mentioned embodiment of the method for generating electrical stimulation signals based on brain-computer interface, the description is relatively simple, and the relevant parts can be referred to the partial description of the embodiment of the method for generating electrical stimulation signals based on brain-computer interface.
[0112] Although the present application is disclosed as above, the present application is not limited thereto. Any person skilled in the art can make various changes and modifications without departing from the spirit and scope of the present application, so the protection scope of the present application shall be subject to the scope defined by the claims.
Claims
1. A method for generating electrical stimulation signals based on a brain-computer interface, It is characterized in that include: Acquire physiological data of the target user, medical use information of the target user, historical physiological parameter information of the target user, and real-time somatosensory information of the target user; Preprocessing the physiological data of the target user to generate multimodal data of the target user, wherein the multimodal data of the target user includes EEG signal data of the target user, ECG signal data of the target user, and heart sound signal data of the target user; Extracting features from the multimodal data of the target user based on the medical use information of the target user and the historical physiological parameter information of the target user to generate a physiological feature risk influencing factor; Performing fusion processing on the physiological characteristic risk influencing factors to generate target multimodal data; Acquire a targeted neuromodulation warning model; Processing the target multimodal data and the real-time somatosensory information of the target user based on the target neural regulation early warning model to generate autonomic nerve activity signal data of the target user; Processing the autonomic nerve activity signal data of the target user based on the target nerve regulation early warning model to generate an electrical stimulation signal, wherein the electrical stimulation signal is used to regulate the autonomic nerve of the target user; Based on the medical use information of the target user and the historical physiological parameter information of the target user, feature extraction is performed on the multimodal data of the target user to generate a physiological feature risk influencing factor, including: Extracting features from the multimodal data of the target user to generate physiological feature classification information, wherein the physiological feature classification information includes physiological feature data corresponding to several types; Processing the physiological characteristic classification information based on the medical use information of the target user to generate target physiological characteristic data; Processing the target physiological characteristic data based on the historical physiological parameter information of the target user to generate a physiological characteristic risk influencing factor; The target multimodal data and the real-time somatosensory information of the target user are processed based on the target neural regulation early warning model to generate autonomic nerve activity signal data of the target user, including: Processing the target multimodal data based on the target neural regulation early warning model to generate skin conductance change information and skin temperature change information of the target user; Processing the real-time somatosensory information of the target user based on the target nerve regulation early warning model to generate skin contraction information of the target user; The skin conductance change information of the target user and the skin temperature change information of the target user are processed based on the skin contraction information of the target user to generate autonomic nerve activity signal data of the target user.
2. The method according to claim 1, It is characterized in that The physiological data of the target user is preprocessed to generate multimodal data of the target user, including: Performing denoising processing on the physiological data of the target user to generate physiological attribute information of the target user, wherein the physiological attribute information of the target user is physiological data with a signal-to-noise ratio higher than a preset threshold; The physiological attribute information of the target user is normalized to generate multimodal data of the target user.
3. The method according to claim 1, It is characterized in that Performing feature extraction processing on the multimodal data of the target user to generate physiological feature classification information includes: Processing the multimodal data of the target user to generate biological attribute information of the target user; Performing feature extraction processing on the biological attribute information of the target user to generate a plurality of feature sample types; Performing feature screening processing on the several feature sample types to generate real-time feature sample data; Generate physiological feature classification information based on a number of real-time feature sample data.
4. The method according to claim 1, It is characterized in that The physiological characteristic risk influencing factors are fused to generate target multimodal data, including: Based on a preset long short-term memory network model, the physiological characteristic risk influencing factors are fused to generate an initial physiological characteristic vector, wherein the initial physiological characteristic vector is spliced by physiological characteristic risk influencing factors of several modalities; The initial physiological feature vector is processed based on a preset weight feature processing rule to generate a target physiological feature vector, wherein the target physiological feature vector is obtained by adding the physiological feature risk influencing factors of different modalities based on the target weights; Processing the target physiological feature vector based on a preset long short-term memory network model to generate a prediction result of multimodal data; Processing the prediction results of the multimodal data to generate target multimodal data; The method also includes a calculation formula for obtaining a target physiological characteristic vector, and the calculation formula is: ; in, is the target physiological feature vector, is the weight of the ith feature, is the risk influencing factor of the ith feature.
5. The method according to claim 1, It is characterized in that Obtain targeted neuromodulation early warning models, including: Processing the physiological characteristic risk influencing factor to generate identification information, wherein the identification information is used to indicate that the physiological characteristic risk influencing factor is in an abnormal state; Acquire an initial neural regulation early warning model that matches the identification information and a training sample set that matches the initial neural regulation early warning model; Process the training sample set based on preset processing rules to generate a training set and a validation set; The initial neuromodulation early warning model is processed based on the training set and the validation set to generate a target neuromodulation early warning model.
6. A device for generating electrical stimulation signals based on a brain-computer interface, It is characterized in that The device comprises: An acquisition module is used to acquire physiological data of a target user, medical use information of a target user, historical physiological parameter information of a target user, and real-time somatosensory information of a target user; and to acquire a target neural regulation early warning model; A processing module, used to pre-process the physiological data of the target user to generate multimodal data of the target user, wherein the multimodal data of the target user includes EEG signal data of the target user, ECG signal data of the target user and heart sound signal data of the target user; extract features of the multimodal data of the target user based on the medical use information of the target user and the historical physiological parameter information of the target user to generate physiological characteristic risk influencing factors; perform fusion processing on the physiological characteristic risk influencing factors to generate target multimodal data; process the target multimodal data and the real-time somatosensory information of the target user based on the target neural regulation early warning model to generate autonomic nerve activity signal data of the target user; process the autonomic nerve activity signal data of the target user based on the target neural regulation early warning model to generate an electrical stimulation signal, wherein the electrical stimulation signal is used to perform autonomic neural regulation on the target user; Extracting features from the multimodal data of the target user based on the medical use information of the target user and the historical physiological parameter information of the target user to generate a physiological feature risk influencing factor, including: extracting features from the multimodal data of the target user to generate physiological feature classification information, wherein the physiological feature classification information includes physiological feature data corresponding to several types; processing the physiological feature classification information based on the medical use information of the target user to generate target physiological feature data; processing the target physiological feature data based on the historical physiological parameter information of the target user to generate a physiological feature risk influencing factor; The target multimodal data and the real-time somatosensory information of the target user are processed based on the target neural regulation warning model to generate the autonomic nervous activity signal data of the target user, including: processing the target multimodal data based on the target neural regulation warning model to generate the skin conductance change information of the target user and the skin temperature change information of the target user; processing the real-time somatosensory information of the target user based on the target neural regulation warning model to generate the skin contraction information of the target user; processing the skin conductance change information and the skin temperature change information of the target user based on the skin contraction information of the target user to generate the autonomic nervous activity signal data of the target user.
7. An electronic device, It is characterized in that include: a first processor; and a memory for storing executable instructions of the first processor; Wherein, the first processor is configured to execute the method for generating an electrical stimulation signal based on a brain-computer interface as described in any one of claims 1 to 5 by executing the executable instructions.
8. A computer-readable storage medium having a computer program stored thereon, It is characterized in that When the computer program is executed by the second processor, the method for generating an electrical stimulation signal based on a brain-computer interface as described in any one of claims 1 to 5 is implemented.
