Hypertension drug dynamic monitoring system and method based on brain-computer interface

Abstract

The invention provides a brain-computer interface-based hypertensive drug dynamic monitoring system and a brain-computer interface-based hypertensive drug dynamic monitoring method, designs a novel hypertensive drug dynamic monitoring mechanism, and analyzes the drug effect by mixing the electroencephalogram signal, the blood pressure signal and the drug taking time to form a multi-modal signal. A series of optimization settings continue to be matched from the aspect of details, so that through multi-modal fusion, dynamic modeling, personalized optimization and clinical closed loop, the system comprehensively surpasses the prior art in the aspects of monitoring dimension, decision depth and function integration, and through collaborative analysis of nerve-blood pressure-autonomic nerve, the accuracy of the system is improved. The normal form transformation of hypertension treatment from symptomatic administration to nerve-body fluid-blood vessel coordinated regulation is promoted, and the accuracy, individuation and safety of hypertension drug monitoring are remarkably improved.

Description

Technical Field

[0001] This application relates to the field of medical technology, specifically to a brain-computer interface-based system and method for dynamic monitoring of hypertension medication. Background Technology

[0002] Hypertension is a systemic condition characterized by elevated arterial pressure, which may be accompanied by functional or organic changes in organs such as the heart, blood vessels, brain, and kidneys. In severe cases, it can lead to diseases such as stroke, myocardial infarction, heart failure, aneurysm, and kidney failure.

[0003] Hypertension is a chronic disease that is difficult to cure completely. Currently, the common clinical treatment is to use antihypertensive drugs to control blood pressure within the normal range for a long time, thereby effectively preventing complications of hypertension. Therefore, the diagnosis and treatment of hypertension involves long-term blood pressure monitoring and drug therapy.

[0004] Under the above treatment methods, traditional methods of blood pressure monitoring include: 1) Ambulatory Blood Pressure Monitoring (ABPM) only records blood pressure results; 2) Smart pillboxes only remind you to take medication and record the time of taking it, but they cannot determine whether the medication is effective.

[0005] Clearly, for users who need to take medication, existing blood pressure monitoring methods have limitations in terms of monitoring performance, making it difficult to conveniently and accurately monitor users' hypertension medication. Summary of the Invention

[0006] This application provides a brain-computer interface-based dynamic monitoring system and method for hypertension drugs. It designs a novel dynamic monitoring mechanism for hypertension drugs, which analyzes drug effects by integrating signals from three aspects: electroencephalogram (EEG), blood pressure, and medication time, to form a multimodal signal. Furthermore, it includes a series of detailed optimization settings. Through multimodal fusion, dynamic modeling, personalized optimization, and clinical closed-loop, it comprehensively surpasses existing technologies in terms of monitoring dimensions, decision-making depth, and functional integration. Through the synergistic analysis of the nervous, blood pressure, and autonomic nervous systems, it promotes a paradigm shift in hypertension treatment from "symptomatic drug administration" to "neuro-humoral-vascular synergistic regulation," significantly improving the accuracy, personalization, and safety of hypertension drug monitoring.

[0007] In a first aspect, this application provides a brain-computer interface-based dynamic monitoring system for hypertension medication. The system includes an electroencephalogram (EEG) acquisition module, a blood pressure monitoring module, a medication administration time monitoring module, and a signal processing module. During the current user's hypertension medication administration monitoring period, the system includes the following processing: The brain-computer interface-based EEG acquisition module acquires the current user's brain signals and transmits them to the signal processing module. The blood pressure monitoring module collects the current user's blood pressure signal and transmits it to the signal processing module. The medication time monitoring module collects the current user's medication time and transmits it to the signal processing module; The signal processing module processes EEG signals, blood pressure signals, and medication time to obtain multimodal signals, and performs corresponding drug effect analysis based on the multimodal signals to obtain drug effect analysis results.

[0008] Secondly, this application provides a method for dynamic monitoring of hypertension medication based on a brain-computer interface. The method is applied to a brain-computer interface-based dynamic monitoring system for hypertension medication. The system includes an EEG acquisition module, a blood pressure monitoring module, a medication administration time monitoring module, and a signal processing module. During the current user's hypertension medication administration monitoring period, the method includes: The brain-computer interface-based EEG acquisition module acquires the current user's brain signals and transmits them to the signal processing module. The blood pressure monitoring module collects the current user's blood pressure signal and transmits it to the signal processing module. The medication time monitoring module collects the current user's medication time and transmits it to the signal processing module; The signal processing module processes EEG signals, blood pressure signals, and medication time to obtain multimodal signals, and performs corresponding drug effect analysis based on the multimodal signals to obtain drug effect analysis results.

[0009] Thirdly, this application provides a computer-readable storage medium storing a plurality of instructions adapted for loading by a processor to execute the method provided in the second aspect of this application.

[0010] From the above, it can be concluded that this application has the following beneficial effects: With the goal of dynamic monitoring of hypertension medication users, this application designs a novel dynamic monitoring mechanism for hypertension drugs. It analyzes drug effects by integrating signals from three aspects: electroencephalogram (EEG), blood pressure, and medication time, to form a multimodal signal. Furthermore, it includes a series of optimization settings in detail. Through multimodal fusion, dynamic modeling, personalized optimization, and clinical closed loop, it comprehensively surpasses existing technologies in terms of monitoring dimensions, decision-making depth, and functional integration. By conducting synergistic analysis of the nervous, blood pressure, and autonomic nervous systems, it promotes a paradigm shift in hypertension treatment from "symptomatic drug administration" to "neuro-humoral-vascular synergistic regulation," significantly improving the accuracy, personalization, and safety of hypertension drug monitoring. Attached Figure Description

[0011] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0012] Figure 1 This is a schematic diagram of a system architecture for the brain-computer interface-based hypertension drug dynamic monitoring system of this application; Figure 2 This is a flowchart illustrating a brain-computer interface-based method for dynamic monitoring of hypertension medications according to this application. Detailed Implementation The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0013] The terms "first," "second," etc., used in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments described herein can be implemented in a sequence other than that illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or device that includes a series of steps or modules is not necessarily limited to those explicitly listed, but may include other steps or modules not explicitly listed or inherent to such processes, methods, products, or devices. The naming or numbering of steps appearing in this application does not imply that the steps in the method flow must be performed in the chronological / logical order indicated by the naming or numbering. The execution order of named or numbered process steps can be changed according to the desired technical purpose, as long as the same or similar technical effect is achieved.

[0014] The module division described in this application is a logical division. In practical applications, there may be other division methods. For example, multiple modules may be combined or integrated into another system, or some features may be ignored or not executed. In addition, the coupling or direct coupling or communication connection between modules shown or discussed may be through some interfaces, and the indirect coupling or communication connection between modules may be electrical or other similar forms, none of which are limited in this application. Furthermore, the modules or sub-modules described as separate components may or may not be physically separated, may or may not be physical modules, or may be distributed in multiple circuit modules. Some or all of the modules may be selected to achieve the purpose of the solution in this application according to actual needs.

[0015] First, refer to Figure 1 The diagram shown is a system architecture diagram of the brain-computer interface-based hypertension drug dynamic monitoring system of this application. The brain-computer interface-based hypertension drug dynamic monitoring system provided by this application for the dynamic monitoring target of hypertension drug users has four main modules in its system architecture design: an EEG acquisition module, a blood pressure monitoring module, a medication time monitoring module, and a signal processing module.

[0016] It is understandable that the four modules mentioned above are the modules involved in realizing the core functions of the solution in this application. In actual situations, other modules may also be involved, such as power supply modules and communication modules. Considering that these other modules are common system modules, they have not been explained in detail.

[0017] Next, based on the work content of the four major modules mentioned above, we will explain the general workflow of the system.

[0018] Specifically, the hypertension medication dynamic monitoring system can include the following processing during the current user's hypertension medication monitoring period: 1) The brainwave signal of the current user is acquired through a brain-computer interface-based EEG acquisition module and transmitted to the signal processing module; Understandably, the EEG acquisition module is responsible for acquiring EEG signals, specifically through the brain-computer interface, to acquire the EEG signals of the user (or patient) who needs to undergo dynamic monitoring of hypertension medication.

[0019] Brain-computer interface technology is mainly divided into three types: invasive, interventional, and non-invasive. Considering that this falls within the scope of existing technology, this application has not elaborated on it further.

[0020] This shows that this application plan is more suitable for users who are hospitalized or staying at home.

[0021] In terms of details, the EEG acquisition module can be based on a non-invasive flexible electrode headband to collect the current user's EEG signals.

[0022] As is easily understood, this application does not involve complex motion intention recognition targets. Therefore, a specific non-invasive brain-computer interface solution, such as a non-invasive flexible electrode headband, can be adopted, allowing users to wear the headband to collect EEG signals when monitoring is needed, assisting in the monitoring of hypertension medication, which has outstanding convenience.

[0023] On the other hand, the collected EEG signals may specifically include beta waves in the motor brain region and heart rate variability (HRV) signals related to autonomic nervous system regulation.

[0024] Understandably, beta waves in the motor brain region are active EEG signals, which are suitable for the application requirements of this application to assist in the monitoring of hypertension drugs. Similarly, heart rate variability signals containing information on the regulation of the cardiovascular system by neurohumoral factors are also suitable.

[0025] Among them, the heart rate variability signal can be obtained by wavelet transform analysis based on the original EEG signal.

[0026] After acquiring brainwave signals, the EEG acquisition module can transmit them to the signal processing module via pre-configured / real-time wired / wireless communication for further data processing. This is also true for the subsequent blood pressure monitoring module and medication time monitoring module.

[0027] As an example, signal transmission can be performed using Bluetooth communication, which has the advantages of low cost and stability.

[0028] 2) The blood pressure monitoring module collects the current user's blood pressure signal and transmits it to the signal processing module; Understandably, in the multimodal or multidimensional hypertension drug monitoring design involved in this application, in addition to the aforementioned electroencephalogram (EEG) signals, blood pressure signals are also involved, as well as the subsequent medication time.

[0029] In this context, in practical applications, the blood pressure monitoring module can select the appropriate device from the available different types of blood pressure monitoring equipment to perform specific blood pressure monitoring tasks, so as to collect the blood pressure signal required by the signal processing module in subsequent data processing.

[0030] It can also be seen from this that, in terms of details, the hypertension drug monitoring system based on brain-computer interface of this application can either directly integrate existing signal acquisition equipment, or call existing signal acquisition equipment outside the system in an external way, or further modify the existing signal acquisition equipment in terms of software and hardware and integrate it into the three major modules of this application. All of these are possible.

[0031] As a practical implementation, the blood pressure monitoring module here can specifically collect the current user's blood pressure signal (in the form of a pressure oscillation signal) based on the pressure sensor inside the blood pressure cuff. Understandably, the blood pressure cuff with integrated pressure sensor can be conveniently and comfortably worn on the user's arm while ensuring stable blood pressure signal collection.

[0032] On the other hand, blood pressure signals can specifically include systolic pressure (SP), diastolic blood pressure (DBP), and mean arterial pressure (ABP). The specific blood pressure status of these three aspects can also be used as a key reference indicator by the magnitude of the decrease.

[0033] Understandably, in the signal acquisition process at this point or before and after, preprocessing operations such as removing motion artifacts through appropriate filtering algorithms can be involved to improve signal quality.

[0034] 3) The medication time monitoring module collects the current user's medication time and transmits it to the signal processing module; Understandably, the signal acquisition goal of the medication time monitoring module is to provide an accurate time reference for subsequent multimodal signal fusion processing, which is usually represented by a timestamp.

[0035] In terms of specific implementation, the medication time monitoring module can identify the current user's medication time through image recognition processing based on the monitoring image, or it can collect the current user's medication time by monitoring the touch status of physical buttons or human voice reminders.

[0036] Specifically, as a practical implementation scheme, in this application, the medication time monitoring module can collect medication time based on the pressure sensor and accelerometer integrated into the hypertension drug bottle.

[0037] As can be seen, in accordance with the specific application of the present application, a drug bottle design for hypertension medication that integrates a pressure sensor and an accelerometer has been specially designed, thereby ensuring convenient, accurate and less susceptible to actual interference in the collection of medication time data.

[0038] Correspondingly, the medication time specifically includes the medication time, frequency, and action pattern obtained from the analysis of the relative stress motion signal (i.e., pressure signal and acceleration signal) of the bottle cap opening action.

[0039] As can be seen here, in terms of details, this application, in addition to determining the specific medication time (usually represented by a time range consisting of two time points before and after) through action feature matching, also involves further analysis of frequency and action patterns, thereby providing richer data references.

[0040] Among these, the easily understood and precisely captured medication time also helps to pinpoint the time span between the medication administration action and the current time point / segment.

[0041] Furthermore, it can be understood that the complex signal processing involved in the above three aspects, in practical applications, may involve the configuration of relevant processing algorithms in advance, or even the configuration of artificial intelligence (AI) models, as well as hardware processing such as microprocessors.

[0042] 4) Based on the EEG signal, blood pressure signal and medication time, the signal processing module processes the data to obtain multimodal signals, and performs corresponding drug effect analysis based on the multimodal signals to obtain the drug effect analysis results.

[0043] After the signals from the three modalities / aspects are collected through the first three modules, the information processing module can carry out centralized data processing and analysis to determine the specific drug effect on the current user.

[0044] The specific data processing and analysis process involves two major stages. The first stage can be understood as multimodal fusion processing to integrate and obtain multimodal signals. The second stage is to efficiently and accurately analyze the specific drug effects based on these multimodal signals.

[0045] In terms of details, based on EEG signals, blood pressure signals, and medication time, multimodal signals are obtained through processing, which may include: 1) Perform Fourier transform on EEG signals, blood pressure signals, and medication time to extract frequency features; Understandably, the Fourier Transform (FT) can convert the input signal from the time domain to the frequency domain, thereby extracting the corresponding frequency features, which can then be used as input parameters for subsequent medication action recognition.

[0046] 2) The CSP algorithm was used to extract β-wave suppression features from the EEG signals; CSP stands for Common Spatial Pattern.

[0047] The β-wave suppression characteristic corresponds to the signal characteristic of reduced power of high-frequency β waves (20-30Hz).

[0048] 3) Medication administration actions were identified using frequency features and β-wave suppression features, and action recognition results with a confidence level ≥ 0.85 were obtained; Understandably, this method uses frequency characteristics and beta wave suppression characteristics as inputs, and then uses corresponding processing algorithms to accurately capture the medication administration action, with the action recognition result having a confidence level of ≥0.85 as the valid recognition result.

[0049] As an example, the medication-taking action here can be recognized using a pre-configured dual-channel convolutional neural network (CNN) model.

[0050] 4) Based on the EEG signal, blood pressure signal, and medication time, and combined with the medication action recognition results, nanosecond-level time synchronization is performed to obtain a data packet containing timestamps, action features, and confidence levels, which serves as a multimodal signal. The nanosecond-level time synchronization uses the moment when the β wave begins to be suppressed as the timestamp t0 of the starting action.

[0051] Understandably, once the medication action recognition results, which accurately capture the timestamps, action features, and confidence levels involved in the medication action, are obtained, the medication action recognition results can be used as key reference factors to perform nanosecond-level / granular time synchronization processing on the previously obtained EEG signals, blood pressure signals, and medication time, in order to integrate them into a data packet containing timestamps, action features, and confidence levels, which serves as the subsequent input multimodal signal.

[0052] In the process of nanosecond-level time synchronization, the IEEE 1588 protocol, a precision clock synchronization protocol standard, can be followed.

[0053] In further detailed operations, for the standard medication action recognition features involved in the medication action recognition in the 3) processing stage, the preliminary processing may include the following: 3.1) Data Collection and Preprocessing During the data collection phase, beta waves in the corresponding motor brain regions were collected when the sample users performed medication-taking and non-medication actions. In the preprocessing stage, bandpass filtering was performed on the β waves of the motor brain regions of the samples to focus on the β wave frequency band, and then independent component analysis was used to remove electrooculography artifacts and electromyography artifacts. 3.2) Data partitioning and covariance matrix calculation In data segmentation, the beta waves in the motor brain regions of the samples were divided into drug-taking action signals X1 (i.e. EEG signals related to drug-taking actions) and non-drug-taking action signals X2 (i.e. EEG signals related to non-drug-taking actions). In the covariance matrix calculation stage, the covariance matrix C1 of the drug administration action signal X1 and the covariance matrix C2 of the non-drug administration action signal X2 are calculated, and then the total covariance matrix C is calculated, C=C1+C2. Taking the covariance matrix C1 as an example, first calculate the corresponding outer product X1i for each sample X1i in the drug administration action signal X1. T Then, the average of these 100 outer products is taken to obtain the covariance matrix C1.

[0054] 3.3) Whitening transformation In the eigenvalue decomposition stage, in the whitening transformation stage, the total covariance matrix C is decomposed into eigenvector matrix U and eigenvalue matrix Λ. In the whitening matrix calculation stage, the whitening transformation matrix P is calculated based on the eigenvalue matrix Λ, where P = Λ. -1 / 2 U T The covariance matrices C1 and C2 are whitened using the whitening transformation matrix P, respectively, to obtain whitening transformation results S1 and S2, where S1 = PC1P. T S2=PC2P T ; 3.4) Solving the generalized eigenvalue problem In the stage of solving the generalized eigenvalue problem, the generalized eigenvalue problem S1W=λS2W is solved to obtain the generalized eigenvalue λ and the generalized eigenvector W, and the generalized eigenvector W is sorted in descending order according to the corresponding eigenvalues. 3.5) Selecting Feature Vectors In the feature vector selection stage, the feature vectors corresponding to the largest and smallest eigenvalues ​​are selected to form the projection matrix W. The feature vectors can highlight the difference between the EEG signals of medication-taking actions and non-medication-taking actions. 3.6) Feature Extraction In the feature extraction stage, the medication action signal X1 is projected through the projection matrix W to obtain the feature vector Z, Z=W. T X1, then calculate the variance of each column of the feature vector Z to obtain the standard medication action features (that is, the variance values ​​obtained are the final extracted medication action features).

[0055] The extracted standard medication-taking action features can be used as a reference for subsequent processing algorithms to assist in accurate medication-taking action recognition.

[0056] It is worth noting that, as mentioned above, the 3) processing step can be performed by a dual-channel CNN model to identify the medication action. If a dual-channel CNN or other AI model is involved, the standard medication action features obtained can be integrated into the actual prediction process of the AI ​​model. The standard medication action features can be used as a reference template to perform similarity calculation with the action features collected in real time, which can help determine whether the current action conforms to the feature pattern of a valid medication action. Alternatively, it can be integrated into the model training process as "ground truth labels" (i.e. labeled real medication actions) for training samples, thereby assisting model application, optimizing model parameters, and improving the model's accuracy in recognizing medication actions.

[0057] On the other hand, in further detailed operations, the preceding 2) processing step, namely, extracting β-wave suppression features from the EEG signal using the CSP algorithm, can also achieve the β-wave suppression feature extraction target through the process of "CSP spatial filtering + power change analysis," which specifically may include: 2.1) Calculate the rate of change of β-wave power in the motion cortex. The rate of change of motor cortical β-wave power in EEG signals is calculated using the following formula: R = (P1 - P2) / P2, Where R represents the rate of change of β-wave power in the motor brain region, P1 represents the β-wave power in the motor brain region (e.g., 13-30Hz) during the drug administration action, and P2 represents the β-wave power in the motor brain region as a resting state. P2 can be obtained by power spectrum analysis of the energy of β-band signals (e.g., 13-30Hz). 2.2) Enhancing signal spatial discrimination using CSP EEG signals with R>0.2 are input into the CSP algorithm (a spatial filter) to further enhance the feature discrimination of the two types of signals in the spatial domain: the drug administration action and the resting state, and to obtain the enhanced signal. It is understandable that R>0.2 indicates that there is a significant change in the power of β waves in the EEG signal related to movement. At this time, it can be preliminarily determined that the corresponding EEG signal has the characteristic tendency of the corresponding drug-taking action, indicating that the EEG signal shows a significant tendency of β wave power change matching the drug-taking action.

[0058] In this case, EEG signals with R>0.2 can be initially screened out as input for subsequent data processing. The CSP algorithm can then be used to effectively enhance these signals, making it easier and more accurate to extract β-wave inhibition features.

[0059] 2.3) β-wave suppression feature extraction For the enhanced signal, extract the β-wave suppression feature.

[0060] Understandably, once a CSP-enhanced EEG signal with R>0.2 is obtained, the specific feature extraction work can be carried out according to the specific feature extraction algorithm configured for the β-wave inhibition feature, so as to obtain the β-wave inhibition feature as input for drug administration action recognition in the 3) processing stage.

[0061] After integrating high-quality multimodal signals as input, it helps to carry out accurate drug effect analysis and obtain corresponding drug effect analysis results.

[0062] The drug efficacy analysis here can be achieved through corresponding processing algorithms or AI models.

[0063] If AI models are involved, there is a drug effect analysis model. This drug effect analysis model performs corresponding drug effect analysis processing based on the multimodal signals input to the model. The multimodal signals input to the model are obtained from the signals collected by the EEG acquisition module, blood pressure monitoring module, and medication time monitoring module.

[0064] As an example, multimodal signals can be treated as time-series data and, through Long Short-Term Memory (LSTM) networks or their related variant networks, can be used to accurately predict the specific effects of drugs in the present and future from a temporal perspective.

[0065] Furthermore, it is understandable that the above-mentioned AI model setups typically involve early-stage model training in practical applications, allowing for subsequent use of the trained model. The specific model architecture, training scheme, and loss function used during training can be based on existing solutions, further optimized versions of existing solutions, or even novel self-developed solutions. All of these can be flexibly configured according to actual needs.

[0066] After accurately determining the drug efficacy analysis results, and considering a more refined user experience, this application solution can also automatically process corresponding drug dosage adjustment suggestions.

[0067] That is, the signal processing module can also be used for: Based on the drug efficacy analysis results, recommendations for adjusting drug dosage are generated.

[0068] Understandably, in practice, the dosage adjustment can be determined based on the current drug efficacy analysis results and the appropriate standard drug dosage, or other methods can be used to adjust the drug dosage. In this way, the corresponding drug dosage adjustment suggestions can be output through specific suggestion output methods.

[0069] In the specific drug dosage adjustment suggestion output stage, the options include local storage, off-site storage, result display, result push, prompts to complete suggestion generation, or further data processing. These can be adaptively adjusted according to actual needs.

[0070] Next, a preferred end-user application scheme of this application is presented in practical applications.

[0071] Specifically, the signal processing module can also be used for: 1) Patient end Provide feedback on the current user's medication administration action.

[0072] Understandably, the medication action feedback here is based on the medication action recognition results identified earlier. It can provide feedback on whether the system has recorded a normal medication event, whether the user has taken the medication within the planned normal medication time, and can also recommend standard medication methods, such as suggesting that the user take the medication with water or before / after meals.

[0073] In addition, the feedback on medication administration can be provided not only through the user equipment (UE) that the user has bound, but also through the signal acquisition modules of the three aspects mentioned above, such as the blood pressure cuff.

[0074] Specifically, UE can be a personal digital assistant (PDA), smartphone, tablet, smart bracelet, or other terminal device.

[0075] Specific feedback methods can include images, voice, vibration, indicator lights, etc.

[0076] 2) Doctor's side The treatment effect report is output to the doctor.

[0077] Understandably, based on relevant visualization interfaces (involving the application of displays (including touch screens), the visualization output of medical reports, including drug effect analysis results and / or drug dosage adjustment suggestions, helps doctors to more intuitively understand the current medication status of users, thereby enabling them to grasp and respond to the user's hypertension control status in a timely and accurate manner, ensuring a high-quality level of diagnosis and treatment.

[0078] The efficacy report can be output in specific report formats such as pdf, doc, txt, jpg, png, and ppt.

[0079] 3) Medicine box end Adjust the drug dosage of the smart pillbox.

[0080] It is understandable that some users are using smart pillboxes. In this regard, after analyzing and obtaining drug dosage adjustment suggestions adapted to the current user's situation, the smart pillbox can be further adapted for application.

[0081] Specifically, the main application of smart pillboxes is to remind users what medication to take at what time, as well as to monitor medication inventory. In this context, medication dosage adjustments can be made based on medication dosage recommendations, such as the time, amount, or type of medication dispensed from the pillbox. This is a further application that extends from the existing smart pillboxes.

[0082] Furthermore, it is understandable that the processing of the above three stages can be optionally configured in practice, and it is not necessary to configure them simultaneously. Moreover, the specific output device structures involved, such as UEs or smart pillboxes, can belong to the system's structural scope or be external devices.

[0083] Furthermore, in terms of details, the following are some application examples: 1) The drug efficacy evaluation value in the drug efficacy analysis results (which can be expressed as follows) can be calculated using the following formula: Y=sigmoid(m1F1+ m1F2+m2△P+m3△t), Where Y is the drug efficacy evaluation value, which takes the value of (0,1); m1, m2 and m3 are different weight coefficients; F1 is the β wave in the motor brain region; F2 is the heart rate variability signal; ΔP is the blood pressure decrease; and Δt is the time interval after medication.

[0084] The drug efficacy assessment value can then be combined with the corresponding threshold to determine whether a drug dosage adjustment needs to be triggered.

[0085] For example, it can be configured so that when Y≥0.8, the drug is deemed effective; otherwise, a dose adjustment is triggered.

[0086] Understandably, the drug dosage adjustment here can be applied to both smart pillboxes and the content presented in the treatment reports and other processing solutions output by doctors.

[0087] 2) The recommended dosage values ​​for drug dosage adjustments can be calculated using the following formula: D = n1F1 + n1F2 + n2△P + n3△t, Where n1, n2, and n3 are different weighting coefficients.

[0088] In practice, m1, m2, and m3 above, as well as n1, n2, and n3 here, can be manually configured or optimized using search algorithms such as Particle Swarm Optimization (PSO).

[0089] Understandably, the above settings are the practical solutions provided by this application based on specific quantitative formulas, and have good application value.

[0090] In conclusion, regarding the above-mentioned solutions, this application designs a novel dynamic monitoring mechanism for hypertension medications, aiming at the dynamic monitoring of hypertension medication users. It analyzes drug effects by integrating EEG signals, blood pressure signals, and medication time to form a multimodal signal. Furthermore, it incorporates a series of detailed optimization settings. Through multimodal fusion, dynamic modeling, personalized optimization, and clinical closed-loop management, this mechanism comprehensively surpasses existing technologies in terms of monitoring dimensions, decision-making depth, and functional integration. Through the synergistic analysis of the nervous, blood pressure, and autonomic nervous systems, it promotes a paradigm shift in hypertension treatment from "symptomatic medication" to "neuro-humoral-vascular synergistic regulation," significantly improving the accuracy, personalization, and safety of hypertension medication monitoring.

[0091] Understandably, the above is an introduction to the brain-computer interface-based hypertension drug dynamic monitoring system of this application. Corresponding to the operation of the system, this application also provides a brain-computer interface-based hypertension drug dynamic monitoring method. Obviously, the brain-computer interface-based hypertension drug dynamic monitoring method is applied to the brain-computer interface-based hypertension drug dynamic monitoring system, which mainly includes an EEG acquisition module, a blood pressure monitoring module, a medication time monitoring module, and a signal processing module.

[0092] Based on this, during the current user's hypertension medication monitoring period, a brain-computer interface-based dynamic monitoring method for hypertension medication, such as... Figure 2 The flowchart shown is a schematic diagram of a brain-computer interface-based method for dynamic monitoring of hypertension medications according to this application, which may specifically include the following steps S201 to S204: Step S201: Collect the current user's brainwave signal through the brain-computer interface-based EEG acquisition module and transmit it to the signal processing module; Step S202: Collect the current user's blood pressure signal through the blood pressure monitoring module and transmit it to the signal processing module; Step S203: Collect the current user's medication time through the medication time monitoring module and transmit it to the signal processing module; In step S204, the signal processing module processes the EEG signal, blood pressure signal, and medication time to obtain a multimodal signal, and performs corresponding drug effect analysis based on the multimodal signal to obtain the drug effect analysis results.

[0093] In one exemplary embodiment, the EEG acquisition module is specifically based on a non-invasive flexible electrode headband to acquire the current user's EEG signals, which specifically include beta waves in the motor brain region and heart rate variability signals related to autonomic nervous system regulation. The blood pressure monitoring module specifically collects the current user's blood pressure signal based on the pressure sensor inside the blood pressure cuff. The blood pressure signal specifically includes systolic pressure, diastolic pressure, and mean arterial pressure. The medication time monitoring module is based on the pressure sensor and accelerometer integrated into the hypertension drug bottle to collect the medication time. The medication time specifically includes the medication time, frequency and action pattern obtained by parsing the relative stress motion signal of the bottle cap opening action.

[0094] In yet another exemplary embodiment, multimodal signals are obtained by processing the electroencephalogram (EEG) signals, blood pressure signals, and medication time, including: Fourier transform was performed on EEG signals, blood pressure signals, and medication time to extract frequency features; The CSP algorithm was used to extract β-wave suppression features from the EEG signals; Medication administration actions were identified using frequency features and β-wave suppression features, and action recognition results with a confidence level ≥ 0.85 were obtained. Based on EEG signals, blood pressure signals, and medication time, and combined with the medication action recognition results, nanosecond-level time synchronization is performed to obtain a data packet containing timestamps, action features, and confidence levels, which serves as a multimodal signal. The nanosecond-level time synchronization uses the moment when the β wave begins to be suppressed as the timestamp t0 of the starting action.

[0095] In yet another exemplary embodiment, the standard medication action recognition features involved in medication action recognition include the following processing: During the data collection phase, beta waves in the corresponding motor brain regions were collected when the sample users performed medication-taking and non-medication actions. In the preprocessing stage, bandpass filtering was performed on the β waves of the motor brain regions of the samples to focus on the β wave frequency band, and then independent component analysis was used to remove electrooculography artifacts and electromyography artifacts. During the data segmentation phase, the beta waves in the motor brain regions of the samples were divided into drug-induced action signals X1 and non-drug-induced action signals X2. In the covariance matrix calculation stage, the covariance matrix C1 of the drug administration action signal X1 and the covariance matrix C2 of the non-drug administration action signal X2 are calculated, and then the total covariance matrix C is calculated, C=C1+C2. In the eigenvalue decomposition stage, the total covariance matrix C is decomposed to obtain the eigenvector matrix U and the eigenvalue matrix Λ. In the whitening matrix calculation stage, the whitening transformation matrix P is calculated based on the eigenvalue matrix Λ, where P = Λ. -1 / 2 U T The covariance matrices C1 and C2 are whitened using the whitening transformation matrix P, respectively, to obtain whitening transformation results S1 and S2, where S1 = PC1P. T S2=PC2P T ; In the stage of solving the generalized eigenvalue problem, the generalized eigenvalue problem S1W=λS2W is solved to obtain the generalized eigenvalue λ and the generalized eigenvector W, and the generalized eigenvector W is sorted in descending order according to the corresponding eigenvalues. In the eigenvector selection stage, eigenvectors corresponding to the largest and smallest eigenvalues ​​are selected to form the projection matrix W; In the feature extraction stage, the medication action signal X1 is projected through the projection matrix W to obtain the feature vector Z, Z=W. T X1 is calculated, and then the variance of each column of the feature vector Z is calculated to obtain the standard medication action characteristics.

[0096] In yet another exemplary embodiment, the CSP algorithm is used to extract beta wave suppression features from the EEG signal, including: The rate of change of motor cortical β-wave power in EEG signals is calculated using the following formula: R = (P1 - P2) / P2, Where R represents the rate of change of β-wave power in the motor brain region, P1 represents the β-wave power in the motor brain region during the drug administration action, and P2 represents the β-wave power in the motor brain region as a resting state. P2 can be obtained from the β-band signal energy through power spectrum analysis. EEG signals with R>0.2 are input into the CSP algorithm to further enhance the feature discrimination of the two types of signals in the spatial domain: drug administration action and resting state, and to obtain the enhanced signal. For the enhanced signal, extract the β-wave suppression feature.

[0097] In yet another exemplary embodiment, the method further includes: Based on the drug efficacy analysis results, recommendations for adjusting drug dosage are generated.

[0098] In yet another exemplary embodiment, the method further includes: Provide feedback on the current user's medication administration action; Output treatment reports to doctors; Adjust the drug dosage of the smart pillbox.

[0099] In yet another exemplary embodiment, the drug efficacy assessment value in the drug efficacy analysis results is calculated using the following formula: Y=sigmoid(m1F1+ m1F2+m2△P+m3△t), Where Y is the drug efficacy assessment value; m1, m2, and m3 are different weighting coefficients; F1 is the beta wave in the motor brain region; F2 is the heart rate variability signal; ΔP is the magnitude of blood pressure decrease; and Δt is the time interval after medication administration. The recommended dosage values ​​for drug dosage adjustments are calculated using the following formula: D = n1F1 + n1F2 + n2△P + n3△t, Where n1, n2, and n3 are different weighting coefficients.

[0100] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working process of the brain-computer interface-based dynamic monitoring method for hypertension medication described above can be found in, for example... Figure 1 The description of the brain-computer interface-based hypertension drug dynamic monitoring system in the corresponding embodiment will not be repeated here.

[0101] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be performed by instructions, or by instructions controlling related hardware. These instructions can be stored in a computer-readable storage medium and loaded and executed by a processor.

[0102] Therefore, this application provides a computer-readable storage medium storing a plurality of instructions that can be loaded by a processor to execute the present application. Figure 2 The steps of the brain-computer interface-based method for dynamic monitoring of hypertension medication in the corresponding embodiment can be referred to as follows for specific operations. Figure 2 The description of the brain-computer interface-based dynamic monitoring method for hypertension medication in the corresponding embodiments will not be repeated here.

[0103] The computer-readable storage medium may include: read-only memory (ROM), random access memory (RAM), disk or optical disk, etc.

[0104] Because of the instructions stored in the computer-readable storage medium, the present application can be executed as described above. Figure 2 The steps of the brain-computer interface-based method for dynamic monitoring of hypertension medication in the corresponding embodiments can therefore be implemented as described in this application. Figure 2 The beneficial effects of the brain-computer interface-based dynamic monitoring method for hypertension medication in the corresponding embodiments are detailed in the preceding description and will not be repeated here.

[0105] The above provides a detailed description of the brain-computer interface-based hypertension drug dynamic monitoring system, the brain-computer interface-based hypertension drug dynamic monitoring method, and the computer-readable storage medium provided in this application. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the core ideas of this application. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this application. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A brain-computer interface-based dynamic monitoring system for hypertension medication, characterized in that, The hypertension medication dynamic monitoring system includes an EEG acquisition module, a blood pressure monitoring module, a medication dosing time monitoring module, and a signal processing module. During the current user's hypertension medication dosing monitoring period, the hypertension medication dynamic monitoring system includes the following processing: The brain-computer interface-based EEG acquisition module acquires the current user's brain signals and transmits them to the signal processing module. The blood pressure monitoring module collects the current user's blood pressure signal and transmits it to the signal processing module. The medication time monitoring module collects the current user's medication time and transmits it to the signal processing module; The signal processing module processes the electroencephalogram (EEG) signal, the blood pressure signal, and the medication time to obtain a multimodal signal, and performs corresponding drug effect analysis based on the multimodal signal to obtain the drug effect analysis results.

2. The hypertension drug dynamic monitoring system according to claim 1, characterized in that, The EEG acquisition module is specifically based on a non-invasive flexible electrode headband to acquire the EEG signals of the current user. The EEG signals specifically include β waves in the motor brain region and heart rate variability signals related to autonomic nervous system regulation. The blood pressure monitoring module specifically collects the current user's blood pressure signal based on the pressure sensor inside the blood pressure cuff. The blood pressure signal specifically includes systolic pressure, diastolic pressure, and mean arterial pressure. The medication time monitoring module is specifically based on the pressure sensor and accelerometer integrated into the hypertension drug bottle to collect the medication time. The medication time specifically includes the medication time, frequency and action pattern obtained by parsing the relative stress motion signal of the bottle cap opening action.

3. The hypertension drug dynamic monitoring system according to claim 1 or 2, characterized in that, The multimodal signal is obtained by processing the electroencephalogram (EEG) signal, the blood pressure signal, and the medication administration time, including: Fourier transform was performed on the electroencephalogram (EEG) signal, the blood pressure signal, and the medication administration time to extract frequency features; The CSP algorithm was used to extract β-wave suppression features from the EEG signals. The medication administration action was identified using the frequency features and the β-wave suppression features, and an action recognition result with a confidence level ≥ 0.85 was obtained. Based on the EEG signal, the blood pressure signal, and the medication time, and combined with the medication action recognition result, nanosecond-level time synchronization is performed to obtain a data packet containing a timestamp, action features, and confidence level, which serves as the multimodal signal. The nanosecond-level time synchronization uses the moment when the β wave begins to be suppressed as the timestamp t0 of the starting action.

4. The hypertension drug dynamic monitoring system according to claim 3, characterized in that, The standard medication administration action recognition features involved in the medication administration action recognition include the following processing: During the data collection phase, beta waves in the corresponding motor brain regions were collected when the sample users performed medication-taking and non-medication actions. In the preprocessing stage, the β waves of the motor brain region of the sample are bandpass filtered to focus on the β wave frequency band, and then independent component analysis is used to remove electrooculography artifacts and electromyography artifacts. During the data segmentation phase, the β waves in the motor brain regions of the samples were divided into drug-induced action signals X1 and non-drug-induced action signals X2. In the covariance matrix calculation stage, the covariance matrix C1 of the medication action signal X1 and the covariance matrix C2 of the non-medication action signal X2 are calculated, and then the total covariance matrix C is calculated, C=C1+C2. In the eigenvalue decomposition stage, the total covariance matrix C is decomposed to obtain the eigenvector matrix U and the eigenvalue matrix Λ. In the whitening matrix calculation stage, the whitening transformation matrix P is calculated based on the eigenvalue matrix Λ, where P = Λ. -1 / 2 U T The covariance matrices C1 and C2 are subjected to whitening transformation using the whitening transformation matrix P, respectively, to obtain whitening transformation results S1 and S2, where S1 = PC1P. T S2=PC2P T ; In the stage of solving the generalized eigenvalue problem, the generalized eigenvalue problem S1W=λS2W is solved to obtain the generalized eigenvalue λ and the generalized eigenvector W, and the generalized eigenvector W is sorted in descending order according to the corresponding eigenvalues. In the eigenvector selection stage, eigenvectors corresponding to the largest and smallest eigenvalues ​​are selected to form the projection matrix W; In the feature extraction stage, the medication action signal X1 is projected through the projection matrix W to obtain the feature vector Z, Z=W. T X1, then calculate the variance of each column of the feature vector Z to obtain the standard medication action features.

5. The hypertension drug dynamic monitoring system according to claim 3, characterized in that, The extraction of β-wave suppression features from the EEG signal using the CSP algorithm includes: The rate of change of motor cortical β-wave power in the electroencephalogram (EEG) signal was calculated using the following formula: R = (P1 - P2) / P2, Where R represents the rate of change of β-wave power in the motor brain region, P1 represents the β-wave power in the motor brain region during the drug administration action, and P2 represents the β-wave power in the motor brain region as a resting state. P2 can be obtained from the β-band signal energy through power spectrum analysis. The EEG signal with R>0.2 is input into the CSP algorithm to further enhance the feature discrimination of the two types of signals in the spatial domain: the drug administration action and the resting state, and to obtain the enhanced signal. For the enhanced signal, β-wave suppression features are extracted.

6. The hypertension drug dynamic monitoring system according to claim 3, characterized in that, The signal processing module is also used for: Based on the drug efficacy analysis results, recommendations for adjusting drug dosage are generated.

7. The hypertension drug dynamic monitoring system according to claim 6, characterized in that, The signal processing module is also used for: Provide feedback on the medication administration action to the current user. Output treatment reports to doctors; Adjust the drug dosage of the smart pillbox.

8. The hypertension drug dynamic monitoring system according to claim 7, characterized in that, The drug efficacy evaluation value in the drug efficacy analysis results is calculated using the following formula: Y=sigmoid(m1F1+ m1F2+m2△P+m3△t), Wherein, Y is the drug efficacy evaluation value; m1, m2, and m3 are different weighting coefficients; F1 is the β wave in the motor brain region; F2 is the heart rate variability signal; ΔP is the magnitude of blood pressure decrease; and Δt is the time interval after medication administration. The recommended dosage values ​​for the drug dosage adjustment are calculated using the following formula: D = n1F1 + n1F2 + n2△P + n3△t, Where n1, n2, and n3 are different weighting coefficients.

9. A method for dynamic monitoring of hypertension medication based on brain-computer interface, characterized in that, The method is applied to a brain-computer interface-based dynamic monitoring system for hypertension medication, which includes an EEG acquisition module, a blood pressure monitoring module, a medication dosing time monitoring module, and a signal processing module. During the current user's hypertension medication dosing monitoring period, the method includes: The brain-computer interface-based EEG acquisition module acquires the current user's brain signals and transmits them to the signal processing module. The blood pressure monitoring module collects the current user's blood pressure signal and transmits it to the signal processing module. The medication time monitoring module collects the current user's medication time and transmits it to the signal processing module; The signal processing module processes the electroencephalogram (EEG) signal, the blood pressure signal, and the medication time to obtain a multimodal signal, and performs corresponding drug effect analysis based on the multimodal signal to obtain the drug effect analysis results.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a plurality of instructions adapted for loading by a processor to execute the method of claim 9.

Images