An electroencephalogram synchronous electroacupuncture stimulation calibration method and system based on embedded AI

Abstract

The application relates to the technical field of medical informatics and intelligent treatment control, and discloses an electroencephalogram synchronous electroacupuncture stimulation calibration method and system based on embedded AI. The embedded AI module is used for collecting and intelligently analyzing brain electrical signals of a patient in real time, generating brain state characteristics, dynamically adjusting electroacupuncture stimulation parameters in combination with subjective feedback of the patient, and forming a closed-loop feedback system. The system carries out individualized baseline calibration before treatment, improves the accuracy of brain state recognition, continuously monitors the brain electrical signals during treatment, dynamically updates the brain state characteristics, adjusts the electroacupuncture parameters according to the brain state characteristics, and ensures the individualization and dynamic optimization of the treatment. The system also has long-term learning ability, continuously accumulates treatment experience, optimizes the association rules, and improves the accuracy and efficiency of the treatment. The application significantly improves the pertinence, response speed and stimulation calibration accuracy of electroacupuncture treatment, and provides a new solution for individualized medical treatment.

Description

Technical Field

[0001] This invention belongs to the field of medical informatics and intelligent treatment control technology, and relates to a method and system for calibrating brain-electroacupuncture synchronous electroacupuncture stimulation based on embedded AI. Background Technology

[0002] Electroacupuncture is a physical therapy method developed from traditional acupuncture and combined with modern electrophysiology. It applies weak pulsed currents to acupuncture needles inserted into specific acupoints to unblock meridians, harmonize qi and blood, and treat diseases. Meanwhile, electroencephalography (EEG), as a non-invasive, real-time electrophysiological signal that reflects the activity of neurons in the cerebral cortex, has been widely used in medical diagnosis, neuroscience research, and human-computer interaction. In particular, the analysis of EEG signals using information technology to assess brain function has become an important branch of health informatics.

[0003] In existing technologies, electroacupuncture devices typically employ an open-loop control mode with preset parameters. Operators set parameters such as stimulation frequency, waveform, and intensity before treatment based on the patient's condition and clinical experience, and these parameters remain constant throughout the treatment. The selection of acupoints also largely relies on standardized treatment protocols or the operator's personal judgment. While some studies have attempted to simultaneously monitor physiological signals such as electroencephalograms (EEGs) during electroacupuncture treatment, this monitoring data is primarily used for post-treatment efficacy evaluation or scientific research and does not achieve real-time linkage with parameter adjustments during treatment. Furthermore, it lacks a precise calibration mechanism for stimulation parameters.

[0004] Existing technologies have significant shortcomings in practical applications. First, fixed stimulation parameters cannot adapt to individual differences among patients, nor to the dynamic physiological changes of the same patient during treatment, resulting in low personalization and an inability to finely calibrate individual responses. Second, the lack of objective feedback on the immediate effects of treatment interventions prevents the use of AI algorithms to assess and calibrate the specific impact of stimulation protocols on the patient's nervous system in real time, leading to a degree of uncertainty in the treatment process. Finally, neglecting physical factors such as changes in tissue electrical properties may result in unstable effective stimulation doses applied to tissues, affecting the repeatability and consistency of treatment effects and hindering precise output calibration. Summary of the Invention

[0005] In view of this, in order to solve the problems mentioned in the background technology, a method and system for EEG-synchronized electroacupuncture stimulation calibration based on embedded AI is proposed.

[0006] The objective of this invention can be achieved through the following technical solution: The first aspect of this invention provides a method for calibrating brain-electroacupuncture stimulation based on embedded AI, comprising: S1, acquiring real-time brain-electroacupuncture signal data of a patient, and analyzing the real-time brain-electroacupuncture signal data to generate brain state characteristics.

[0007] S2. Obtain the association rules used to establish the relationship between brain state and electroacupuncture stimulation, and apply the association rules to brain state features to generate initial electroacupuncture parameters.

[0008] S3. Based on the initial electroacupuncture parameters, control the electroacupuncture device to apply electroacupuncture stimulation, and collect the stimulation state data of the electroacupuncture device during the stimulation process.

[0009] S4. While applying electroacupuncture stimulation, continuously collect the patient's real-time EEG signal data, and combine it with stimulation state data to dynamically update the brain state characteristics in order to form a closed-loop feedback signal.

[0010] S5. Based on the closed-loop feedback signal, adjust the initial electroacupuncture parameters to generate the final electroacupuncture parameters, which are then used for subsequent electroacupuncture stimulation control.

[0011] The second aspect of the present invention provides an EEG-synchronized electroacupuncture stimulation calibration system based on embedded AI, comprising: a brain state feature generation module, which acquires real-time EEG signal data of patients and analyzes the real-time EEG signal data to generate brain state features.

[0012] The initial electroacupuncture parameter generation module obtains the association rules used to establish the relationship between brain state and electroacupuncture stimulation, and applies the association rules to brain state features to generate initial electroacupuncture parameters.

[0013] The electroacupuncture stimulation application module controls the electroacupuncture device to apply electroacupuncture stimulation according to the initial electroacupuncture parameters, and collects the stimulation status data of the electroacupuncture device during the stimulation process.

[0014] The closed-loop feedback signal generation module continuously collects real-time EEG signal data from the patient while applying electroacupuncture stimulation, and dynamically updates the brain state characteristics by combining the stimulation state data to form a closed-loop feedback signal.

[0015] The final electroacupuncture parameter generation module adjusts the initial electroacupuncture parameters based on the closed-loop feedback signal to generate the final electroacupuncture parameters, which are then used for subsequent electroacupuncture stimulation control.

[0016] Compared with the prior art, the embodiments of the present invention have at least the following advantages or beneficial effects: (1) The present invention realizes the dynamic and personalized nature of electroacupuncture treatment by constructing a closed-loop feedback system based on real-time electroencephalogram (EEG) signals. The system can monitor the patient's brain state in real time and dynamically adjust the various parameters of electroacupuncture stimulation according to the instantaneous changes in this state, overcoming the limitation of the treatment parameters being unchanging in traditional methods, enabling the treatment plan to closely follow the patient's physiological rhythm for adaptive optimization, and significantly improving the targeting and response speed of the treatment.

[0017] This invention significantly improves the accuracy and reliability of brain state feature recognition by introducing an individualized baseline calibration mechanism. Before treatment begins, the system collects the patient's unique resting-state EEG data to calibrate the classification criteria. This ensures that subsequent state assessments no longer rely on universal population models, but are based on a precise understanding of the patient's individual neurophysiological characteristics, providing a solid and reliable data foundation for the entire personalized treatment process.

[0018] This invention integrates objective physiological indicators with patient subjective feedback, and combines brain function status with traditional Chinese medicine acupoint theory, making treatment decisions more comprehensive and scientific. By comprehensively considering EEG data and patient self-perception, an initial stimulation plan is generated, and the stimulation sites are dynamically selected based on the mapping relationship between brain state and acupoints. This method not only improves the comfort and humanization of treatment but also provides acupoint selection with a basis in modern neuroscience, enhancing the targeted nature of treatment.

[0019] This invention ensures treatment stability and system evolution by precisely controlling the stimulation process and learning from treatment experience over a long period. By monitoring tissue impedance in real time and compensating for the effective charge, the precise and constant physical stimulation dose is guaranteed. Simultaneously, the system can learn from successful experience data after each treatment to update its core association rules, enabling it to self-iterate and optimize. With increased use, the treatment strategy becomes more efficient and precise. Attached Figure Description

[0020] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0021] Figure 1 This is a schematic diagram of the method steps of the present invention.

[0022] Figure 2 This is a schematic diagram of the system structure connection of the present invention. Detailed Implementation

[0023] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0024] Please see Figure 1 The first aspect of the present invention provides a method for calibrating synchronous electroacupuncture stimulation based on embedded AI, comprising: S1, acquiring real-time EEG signal data of a patient, and performing AI intelligent analysis on the real-time EEG signal data to generate brain state characteristics.

[0025] In a specific embodiment of the present invention, the specific steps of AI intelligent analysis of real-time EEG signal data to generate brain state characteristics include: filtering the patient's EEG signal data to generate preprocessed EEG data.

[0026] It should be noted that the process of using AI to intelligently analyze real-time EEG signal data to generate brain state characteristics in this method first involves acquiring multi-channel real-time EEG signal data from electrodes at specific locations on the patient's scalp using an EEG acquisition device. This raw real-time EEG signal data cannot be used directly due to the presence of power frequency interference, electrooculography (EOG), and electromyography (EMG) artifacts; therefore, filtering processing is required. This processing employs a digital bandpass filter to remove irrelevant signals outside the frequency range, and a notch filter to eliminate power frequency interference at specific frequencies, thereby generating pre-processed EEG data with a high signal-to-noise ratio.

[0027] Frequency band energy parameters are extracted from the generated preprocessed EEG data to generate EEG feature vectors.

[0028] It should be noted that, next, in order to transform the preprocessed EEG data in the time domain into quantifiable features, it is necessary to extract its frequency band energy parameters. This step typically employs Fast Fourier Transform (FFT) spectral analysis to convert the preprocessed EEG data within a specific time window from the time domain to the frequency domain, obtaining its power spectral density. Subsequently, according to established neuroscience standards, the power spectrum is divided into different rhythmic frequency bands, such as δ (1-4Hz), θ (4-8Hz), α (8-13Hz), β (13-30Hz), etc. The energy parameters of each frequency band are obtained by integrating the power spectral density within that band's frequency range, as expressed by the formula: ,in, Represents a certain frequency band Energy parameters, For frequency The power spectral density at that location was calculated from preprocessed EEG data. and These are the start and cutoff frequencies of the frequency band. Combining the energy parameters of all target frequency bands forms a multi-dimensional EEG feature vector.

[0029] Obtain classification thresholds used to define different brain state types.

[0030] The brain state type is identified by comparing the EEG feature vector with the classification threshold, thereby generating brain state features.

[0031] It should be noted that, subsequently, the system acquires a set of pre-set classification thresholds, which are used to define different brain state types. The system compares the real-time calculated EEG feature vectors with these classification thresholds, and through logical judgment, identifies the patient's most likely current brain state type. This brain state type includes, but is not limited to, relaxation, tension, and anxiety. The identified brain state type, along with its corresponding EEG feature vector, constitutes the brain state features used for subsequent decision-making.

[0032] The above method transforms complex and continuous real-time EEG signal data into discrete or continuous brain state characteristics with clear physiological significance. The advantage of this technology lies in its ability to provide an objective, real-time, and quantitative assessment of a patient's neurological function, offering a direct and reliable basis for subsequent personalized stimulation parameter calibration.

[0033] In a specific embodiment of the present invention, before the step of obtaining the classification threshold for defining different brain state types, the method further includes: collecting baseline EEG signal data of the patient in a resting state before the start of treatment.

[0034] It should be noted that, to improve the individualized accuracy of brain state type identification, this method adds a personalized calibration step before applying the general classification threshold. This step is performed before the formal electroacupuncture treatment begins. First, the patient is guided into a quiet and comfortable environment and instructed to close their eyes, relax, and enter a resting state free from task-related or specific thought activities. In this state, EEG signal data is continuously collected for a period of time, such as three to five minutes, using an EEG acquisition device, serving as baseline EEG signal data characterizing the patient's unique neurophysiological background.

[0035] Baseline EEG signal data are analyzed to generate individualized baseline EEG characteristics for each patient.

[0036] It should be noted that, next, the system filters the baseline EEG signal data, then extracts the energy parameters and other features of each rhythm frequency band, and finally generates a set of quantified individualized baseline EEG features that represent the patient's unique EEG activity pattern in the resting state.

[0037] Based on individualized baseline EEG characteristics, classification thresholds are calibrated to generate personalized classification thresholds suitable for each patient.

[0038] It should be noted that, finally, based on this set of newly generated individualized baseline EEG features, the system's preset universal classification thresholds, derived from large-scale population statistics or standard models, are calibrated. The calibration process is implemented through an adjustment algorithm whose core idea is to shift or scale the universal thresholds towards the patient's individual baseline to accommodate individual differences. An exemplary calibration formula is as follows: ,in, It is the first of the personalized classification thresholds generated for this patient. Each feature threshold. It is the first one in the original universal classification threshold of the system. Each feature threshold. It is the first of the individualized baseline EEG characteristics of this patient that was just calculated. Each feature value. It is the population database on which the universal classification threshold is based, the first The average value of each feature. These are preset calibration coefficients, derived from specific experimental and clinical environments, used to control the magnitude of individualized adjustments. Using this formula, if a patient's baseline characteristics are higher than the population average, the corresponding classification threshold will be increased, and vice versa, thus generating a truly personalized set of classification thresholds applicable to the patient for real-time assessment of brain state during subsequent treatment.

[0039] The technical advantage of this method lies in its significant overcoming of the problem of inaccurate brain state classification caused by the vast differences in EEG signals between individuals. By introducing an individualized baseline calibration process before treatment, the classification threshold used to determine brain state type is no longer a one-size-fits-all standard, but rather tailored to each patient. This greatly improves the accuracy and reliability of brain state feature recognition, ensuring that subsequent electroacupuncture parameter selection and closed-loop feedback modulation based on brain state are grounded in a precise understanding of the patient's true neurophysiological state, thus providing a solid foundation for achieving truly personalized and precise treatment.

[0040] S2. Obtain the association rules used to establish the relationship between brain state and electroacupuncture stimulation, and apply the association rules to brain state features to generate initial electroacupuncture parameters.

[0041] In a specific embodiment of the present invention, the steps for applying association rules to brain state features to generate initial electroacupuncture parameters include: obtaining user feedback data input by the patient through a user interaction interface.

[0042] It's important to note that, firstly, the system acquires user feedback data input by the patient through a user interface located near the patient, such as a touchscreen or a terminal with physical buttons. This user interface presents the patient with standardized questions, such as using a visual analog scale or a numerical rating scale, asking the patient to rate their current level of discomfort, comfort, or treatment experience. The patient inputs a value or selects an option through interactive operation, and the system records this input as the user feedback data for that session.

[0043] By integrating user feedback data with brain state characteristics, a comprehensive state index is generated.

[0044] It should be noted that, next, in order to form an assessment criterion that simultaneously reflects objective physiological indicators and subjective psychological feelings, the system needs to fuse the acquired user feedback data with brain state characteristics obtained through EEG analysis. Since brain state characteristics are a multi-dimensional physiological vector, while user feedback data is usually a single scalar, direct computation lacks physical meaning. Therefore, the fusion process first requires standardizing both, mapping them to a unified dimensionless interval, such as 0 to 1. Subsequently, a comprehensive state index is generated through weighted summation. This process can be represented by the following formula: In this formula, This represents the final generated comprehensive status index. It is an objective state quantity obtained by standardizing the ratio of alpha wave to beta wave energy in brain state characteristics. It is a subjective state quantity obtained by standardizing user feedback data. and These are preset weighting coefficients, representing the importance of objective physiological signals and subjective user feedback in the comprehensive evaluation, and the sum of the two is 1.

[0045] In one specific embodiment of the present invention, if the system design places greater emphasis on the accuracy and scientific validity of objective physiological signals, A typical value might be set to 0.7, while The value would be 0.3 to ensure that objective data dominates the assessment; conversely, if the system wishes to consider the patient's subjective feelings and comfort more, The value might be set to 0.6, accordingly. The value is 0.4 to reflect the importance of patient feedback. However, the specific typical value should be determined based on the goals of the clinical test and system design to ensure the comprehensiveness and accuracy of the overall assessment.

[0046] Based on comprehensive status indicators, association rules are applied to generate initial electroacupuncture parameters.

[0047] It should be noted that, finally, the system uses the calculated comprehensive state index as input and applies a pre-established association rule base. These association rules define the optimal electroacupuncture stimulation scheme corresponding to different comprehensive state index ranges. The system queries this rule base to find the rule that matches the current comprehensive state index value, extracts the electroacupuncture parameters specified by the rule, such as stimulation frequency, waveform, and intensity, and uses this set of parameters as the initial electroacupuncture parameters required for controlling the electroacupuncture device in this operation.

[0048] The technical advantage of this method lies in expanding the basis of treatment decisions from a single, purely objective physiological signal to a more comprehensive dimension that includes the patient's subjective experience. This combination of objectivity and subjectivity ensures that the generated initial electroacupuncture parameters not only respond to the patient's real-time neurophysiological changes but also fully consider the patient's individual feelings and tolerance. This not only improves the comfort and safety of treatment but also makes the initial setting of the treatment plan more closely aligned with the patient's individual needs, providing a more precise and humane starting point for the entire treatment process, thereby enhancing the overall acceptance and potential efficacy of the treatment.

[0049] S3. Based on the initial electroacupuncture parameters, control the electroacupuncture device to apply electroacupuncture stimulation, and collect the stimulation state data of the electroacupuncture device during the stimulation process.

[0050] In a specific embodiment of the present invention, the specific steps of controlling the electroacupuncture device to apply electroacupuncture stimulation according to the initial electroacupuncture parameters include: obtaining the acupoint mapping relationship between the brain state and acupoints.

[0051] It should be noted that after determining the initial electroacupuncture parameters, this method performs a series of steps to translate abstract treatment strategies into concrete physical operations in order to achieve precise control of the electroacupuncture device. First, the system needs to acquire the acupoint mapping relationship between brain states and acupoints pre-set in its internal knowledge base. This acupoint mapping relationship is a structured dataset that, based on traditional Chinese medicine meridian theory and modern neuroscience research, establishes a correspondence between different brain state characteristics, such as the strength combination of specific brain electrical rhythms, and one or a group of meridian acupoints with specific regulatory functions. This relationship is essentially a set of decision rules, providing a theoretical basis for the system to select the most suitable intervention target from acupoints throughout the body.

[0052] Based on brain state characteristics and by applying acupoint mapping relationships, the target stimulation sites are determined.

[0053] It should be noted that, next, the system uses the brain state characteristics obtained from the current analysis as query input and applies the aforementioned acupoint mapping relationship for matching. By searching the knowledge base, the system can identify the target stimulation site most associated with the current patient's brain state. This target stimulation site may be a single acupoint or a combination of acupoints with synergistic effects. This process enables dynamic and personalized selection of treatment targets, ensuring that electroacupuncture stimulation can accurately act on key nodes that regulate the current neurological functional state.

[0054] Control the electroacupuncture device to apply electroacupuncture stimulation to the target stimulation site.

[0055] It should be noted that, finally, once the target stimulation site is determined, the system combines the previously generated initial electroacupuncture parameters, such as stimulation frequency, intensity, and waveform, to generate specific control commands. These commands are sent to the electroacupuncture device, driving its corresponding output channels to apply precisely controlled electroacupuncture stimulation to the patient's body through the needle already inserted at the target stimulation site. For example, the command might explicitly instruct channel 1 to output stimulation to the needle located at the "Neiguan" acupoint with a sparse wave of 2 Hz, while channel 2 will output stimulation to the "Shenmen" acupoint with a dense wave of 50 Hz, thus completing the closed loop of the entire stimulation control.

[0056] The technical advantage of this method lies in its effective combination of objective quantitative indicators from modern neuroscience with the acupoint theory of traditional Chinese medicine, achieving a scientific and intelligent selection of treatment sites. By establishing a direct mapping between brain state and acupoints, electroacupuncture no longer relies on fixed acupoint prescriptions but can dynamically and precisely select stimulation targets based on the patient's real-time neurophysiological state. This adaptive acupoint selection strategy based on brain state greatly enhances the targeting and specificity of electroacupuncture.

[0057] In a specific embodiment of the present invention, after the step of controlling the electroacupuncture device to apply electroacupuncture stimulation to the target stimulation site, the method further includes: extracting tissue impedance parameters from the stimulation state data.

[0058] It should be noted that after the electroacupuncture device applies electroacupuncture stimulation to the target stimulation site according to the initial electroacupuncture parameters, this method immediately activates a local real-time feedback adjustment subsystem. This subsystem first acquires real-time stimulation state data from the sensing unit inside the electroacupuncture device, which includes the instantaneous voltage and current values ​​flowing through the needle. By performing high-speed sampling and calculation on these data, the system can extract tissue impedance parameters in real time. The tissue impedance parameters reflect the electrical characteristics of the interface between the electroacupuncture needle and human tissue, and are calculated based on Ohm's law by dividing the instantaneous voltage value by the corresponding instantaneous current value.

[0059] The output waveform and pulse duration of electroacupuncture stimulation are dynamically adjusted based on changes in the extracted tissue impedance parameters.

[0060] It should be noted that, because the human body is a dynamic physiological system, the tissue impedance parameters at the target stimulation site will fluctuate during electroacupuncture stimulation due to factors such as changes in blood flow and ion concentration migration. This fluctuation will cause changes in the effective charge actually applied to nerve tissue under constant voltage or constant current output mode, thus affecting the consistency of treatment effects. To address this issue, the system dynamically adjusts the output waveform and pulse duration of electroacupuncture stimulation based on the real-time changes in the monitored tissue impedance parameters. Its core objective is to maintain the effective charge delivered by each stimulation pulse at a constant target value. The formula for calculating the effective charge can be expressed as: ,in, It is the effective charge of a single pulse. It is the average current intensity during the stimulation pulse. This refers to the duration of the pulse. When the system detects an increase in tissue impedance parameters leading to a surge in pulse current... During descent, the control algorithm automatically increases the pulse duration. Conversely, the same applies to ensure that the product of the two is the effective charge. It always approaches the preset value.

[0061] The adjusted stimulation control command is generated and sent to the electroacupuncture device.

[0062] It should be noted that, based on this adjustment algorithm, the system generates adjusted stimulation control commands in real time. These commands include new parameters such as precisely calculated pulse duration or waveform amplitude. The commands are immediately sent to the electroacupuncture device via the communication interface, where they are executed by the device's microcontroller, thereby achieving output compensation and correction in the next stimulation pulse cycle.

[0063] The technical advantage of this method lies in its ability to achieve precise and stable control of the physical stimulation dose applied to acupoints through a high-speed, local closed-loop control. This surpasses the open-loop mode of traditional electroacupuncture devices, which only control the output voltage or current, effectively avoiding the uncertainty of treatment dosage caused by dynamic changes in individual tissue impedance. By ensuring that the effective charge applied to the target stimulation site remains stable, it provides a highly consistent and quantifiable stimulation input for subsequent analysis of EEG responses and evaluation of treatment effects, greatly improving the accuracy and reliability of the entire personalized treatment system.

[0064] S4. While applying electroacupuncture stimulation, continuously collect the patient's real-time EEG signal data, and combine it with stimulation state data to dynamically update the brain state characteristics in order to form a closed-loop feedback signal.

[0065] In a specific embodiment of the present invention, the specific steps of dynamically updating the brain state characteristics to form a closed-loop feedback signal include: analyzing real-time EEG signal data continuously collected during the application of electroacupuncture stimulation to calculate its changing trend and generate EEG response characteristics.

[0066] It should be noted that, to achieve closed-loop control of the treatment process, this method dynamically updates brain state characteristics while applying electroacupuncture stimulation to form a closed-loop feedback signal. This process first requires time-series analysis of the real-time EEG signal data continuously collected during electroacupuncture stimulation to calculate its trend relative to before stimulation. Specifically, the system compares the EEG feature vector analyzed within the current time window with the EEG feature vector in a baseline state, which can be the state at the beginning of treatment. By calculating the ratio between the two at key frequency band energy parameters, the intensity and direction of the brain's response to stimulation are quantified, thereby generating EEG response features containing trend information. The response intensity is represented by the absolute value of the ratio of each key frequency band energy parameter, reflecting the magnitude of the change in brain activity caused by the stimulation. The response direction is determined by the sign of the ratio; a positive value indicates that the stimulation enhances brain activity in that key frequency band, while a negative value indicates inhibition, thus comprehensively quantifying the dynamic response characteristics of the brain to electroacupuncture stimulation.

[0067] By correlating EEG response characteristics with stimulus state data, the immediate effects of treatment interventions can be assessed, and effect evaluation indicators can be generated.

[0068] It should be noted that, next, to evaluate the effectiveness of the current electroacupuncture stimulation protocol, the system will perform correlation analysis between this EEG response feature and the stimulation state data of the electroacupuncture device. The stimulation state data provides the input information for the current stimulation, such as frequency, intensity, and acupoints, while the EEG response feature represents the brain's output information. The purpose of the correlation analysis is to determine whether this output meets the expected treatment goal. For example, if the treatment goal is sedation and tranquilization, the correlation rule expects to observe an increase in alpha wave energy. If the EEG response feature does indeed show a significant increase in alpha wave energy, the intervention is considered effective. The system uses an evaluation function to transform this matching degree into a quantitative effect evaluation index. This effect evaluation index can be a continuous value, with positive values ​​representing a positive response and negative values ​​representing an adverse reaction; its absolute value represents the intensity of the response. This evaluation function can be expressed as: ,in, These are the performance evaluation metrics for the generated results. It is the changing trend obtained from real-time EEG signal data analysis, namely the EEG response characteristics. This is the current electroacupuncture stimulation parameter obtained from the stimulation state data. (Function) This represents a pre-defined set of rules, which is based on stimulus parameters. The actual EEG response is evaluated based on the corresponding expected therapeutic effect. Does it meet expectations or deviate from expectations, and output a quantitative score accordingly? .

[0069] Based on the effect evaluation indicators, the brain state characteristics are updated and closed-loop feedback signals are generated.

[0070] It's important to note that, finally, the system updates the original brain state characteristics based on this effectiveness evaluation index. This doesn't modify the brain state characteristics themselves, but rather supplements or weights them to create a richer information package. For example, the effectiveness evaluation index can be added as a new dimension to the current brain state feature vector. This composite information, containing the real-time state and the intervention effectiveness evaluation in that state, constitutes a closed-loop feedback signal for subsequent parameter adjustments.

[0071] The technical advantage of this method lies in its ability to provide crucial decision-making support for the entire closed-loop control system by quantifying the immediate effects of electroacupuncture intervention in real time. It goes beyond simply monitoring the brain's state; it goes further to answer the core question of whether the current treatment is effective in improving that state. This deepening from state monitoring to effect evaluation enables the system to establish a dynamic causal relationship between stimulus input and neural response, providing clear direction and amplitude guidance for subsequent intelligent adjustments. This is a key step in moving from simple feedback to intelligent adaptive optimization.

[0072] S5. Based on the closed-loop feedback signal, adjust the initial electroacupuncture parameters to generate the final electroacupuncture parameters, which are then used for subsequent electroacupuncture stimulation control.

[0073] In a specific embodiment of the present invention, the specific steps of adjusting the initial electroacupuncture parameters to generate the final electroacupuncture parameters include: determining the adjustment direction and adjustment range of the parameters based on the closed-loop feedback signal, and generating the parameter adjustment amount.

[0074] It should be noted that this method adaptively adjusts the electroacupuncture parameters based on the closed-loop feedback signal generated in the previous step. First, the system's internal control algorithm module receives this closed-loop feedback signal and focuses on analyzing the effect evaluation index. This effect evaluation index quantifies the immediate effect of the current stimulation protocol. Based on the value of this effect evaluation index, combined with a preset optimization strategy, the control algorithm determines the direction and magnitude of parameter adjustment. For example, if the effect evaluation index is significantly positive, it indicates that the current stimulation is effective, and the system may decide to maintain the current parameters or make only minor adjustments; if the index is close to zero, it indicates that the stimulation is ineffective, and the system needs to perform an exploratory adjustment, such as increasing the stimulation intensity or changing the stimulation frequency; if the index is negative, it indicates that an adverse reaction has occurred, and the system must perform a reverse adjustment, such as reducing the intensity or switching to a stimulation waveform with an inhibitory effect.

[0075] The generated parameter adjustments are applied to the initial electroacupuncture parameters to generate the final electroacupuncture parameters.

[0076] It should be noted that, based on the above decisions, the control algorithm calculates a specific adjustment value for each adjustable electroacupuncture parameter, such as frequency, intensity, and pulse width. These values ​​collectively constitute a multidimensional parameter adjustment quantity. Applying this parameter adjustment quantity to the current initial electroacupuncture parameters generates the final electroacupuncture parameters for the next treatment cycle. This update process can be represented by the following vector formula: ,in, This is the final electroacupuncture parameter vector that will be generated soon, containing all the stimulation parameter values ​​required for the next cycle. This is the initial electrocautery parameter vector currently in use during this cycle. It is a parameter adjustment vector calculated by the control algorithm based on the closed-loop feedback signal, with each component corresponding to the adjustment value of an electrocautery parameter. This addition operation is performed in vector space, ensuring the independent adjustment and combination of each parameter.

[0077] The generated final electroacupuncture parameters are sent to the electroacupuncture device for use in the next cycle of electroacupuncture stimulation control.

[0078] It should be noted that after calculating the final electroacupuncture parameters, the system compiles them into new device control commands and sends them to the electroacupuncture device via the communication interface. Upon receiving these new commands, the electroacupuncture device stores them and uses them to execute the electroacupuncture stimulation control for the next cycle, thus completing a full closed-loop feedback regulation.

[0079] The technical effect achieved by this method is that it constructs the core driving engine of the entire personalized treatment system, enabling the system to learn autonomously and optimize in real time. By directly translating the evaluation results of treatment effects into adjustments to stimulation parameters, this method establishes a complete closed-loop pathway from brain response to stimulation regulation. This transforms electroacupuncture treatment from a static, pre-set procedure into a dynamic, adaptive process that continuously strives for optimal treatment results. It can automatically identify and maintain the most effective stimulation protocol for a specific patient during treatment, thereby maximizing and personalizing the treatment effect.

[0080] In a specific embodiment of the present invention, the association rules can be updated through the following steps: after the treatment is completed, the final electroacupuncture parameters and the final brain state characteristics of this treatment are obtained to form the empirical data of this treatment.

[0081] It should be noted that this method incorporates a long-term learning and optimization mechanism to update the core association rules. This mechanism is activated after a single electroacupuncture treatment session is completely completed. The system first archives the final outcomes of this treatment, extracting the final electroacupuncture parameters that have proven effective at the end of the treatment, as well as the expected final brain state characteristics achieved by the patient under the influence of these parameters. These two sets of data are packaged into a single data pair, constituting empirical data reflecting the success of this treatment.

[0082] Obtain historical experience data accumulated during the course of treatment.

[0083] It should be noted that, next, the system will access its internal long-term knowledge base to retrieve historical experience data accumulated over all past treatments for this patient or other patients with similar symptoms. This historical experience data consists of a large number of previous successful experience data pairs, forming a massive dataset. The system will then incorporate the newly generated experience data from this treatment as a new data point and add it to this historical experience dataset, thereby achieving incremental accumulation of knowledge.

[0084] The empirical data from this treatment is integrated with historical empirical data, and the association rules are learned and updated.

[0085] It's important to note that, based on this, the system will initiate an offline learning and updating process to reanalyze the entire experience dataset containing the new data. This process can employ association rule mining algorithms or more complex machine learning models, such as decision trees, support vector machines, or neural network regression. Its goal is to relearn and refine the mapping relationship between brain states and optimal electroacupuncture parameters from a larger dataset. Through this learning process, existing association rules may be revised, for example, the stimulation frequency range corresponding to a certain state may be fine-tuned; new, more refined rules may also be discovered, such as the stimulation waveform being more critical than the stimulation frequency in a specific brain state. The updated association rules are then rewritten into the system's knowledge base for use in subsequent treatment sessions.

[0086] The technical advantage of this method lies in endowing the entire personalized treatment system with long-term evolutionary capabilities beyond a single treatment. Through review and learning after each treatment, the system can continuously optimize itself based on successful experiences, and its core knowledge base of association rules grows and evolves accordingly. This makes treatment strategies no longer static, but rather increasingly precise and tailored to the unique response patterns of each individual patient with each treatment session. This long-term learning mechanism ensures that the system can provide continuously optimized treatment plans, converging to optimal stimulation parameters more quickly in future treatments, thereby improving the long-term effectiveness and efficiency of treatment.

[0087] Reference Figure 2 The second aspect of the present invention provides a brainwave-synchronized electroacupuncture stimulation calibration system based on embedded AI, comprising: a brain state feature generation module, an initial electroacupuncture parameter generation module, an electroacupuncture stimulation application module, a closed-loop feedback signal formation module, and a final electroacupuncture parameter generation module.

[0088] The brain state feature generation module is connected to the initial electroacupuncture parameter generation module, the initial electroacupuncture parameter generation module is connected to the electroacupuncture stimulation application module, both the brain state feature generation module and the electroacupuncture stimulation application module are connected to the closed-loop feedback signal formation module, and both the initial electroacupuncture parameter generation module and the closed-loop feedback signal formation module are connected to the final electroacupuncture parameter generation module.

[0089] The brain state feature generation module acquires the patient's real-time EEG signal data and analyzes the real-time EEG signal data to generate brain state features.

[0090] The initial electroacupuncture parameter generation module obtains association rules for establishing the relationship between brain state and electroacupuncture stimulation, and applies the association rules to brain state features to generate initial electroacupuncture parameters.

[0091] The electroacupuncture stimulation application module controls the electroacupuncture device to apply electroacupuncture stimulation according to the initial electroacupuncture parameters, and collects the stimulation status data of the electroacupuncture device during the stimulation process.

[0092] The closed-loop feedback signal forming module continuously collects real-time EEG signal data from the patient while applying electroacupuncture stimulation, and dynamically updates the brain state characteristics by combining the stimulation state data to form a closed-loop feedback signal.

[0093] The final electroacupuncture parameter generation module adjusts the initial electroacupuncture parameters based on the closed-loop feedback signal to generate the final electroacupuncture parameters, which are then used for subsequent electroacupuncture stimulation control.

[0094] The above content is merely an example and illustration of the concept of the present invention. Those skilled in the art can make various modifications or additions to the specific embodiments described, or use similar methods to replace them, as long as they do not deviate from the concept of the invention or exceed the scope defined by the present invention, and all such modifications and additions should fall within the protection scope of the present invention.

Claims

1. A method for calibrating EEG-synchronized electroacupuncture stimulation based on embedded AI, characterized in that, include: S1. Acquire real-time EEG signal data from patients and perform AI intelligent analysis on the real-time EEG signal data to generate brain state characteristics; S2. Obtain the association rules used to establish the relationship between brain state and electroacupuncture stimulation, and apply the association rules to brain state features to generate initial electroacupuncture parameters. S3. Based on the initial electroacupuncture parameters, control the electroacupuncture device to apply electroacupuncture stimulation, and collect the stimulation state data of the electroacupuncture device during the stimulation process; S4. While applying electroacupuncture stimulation, continuously collect the patient's real-time EEG signal data, and combine it with stimulation state data to dynamically update the brain state characteristics in order to form a closed-loop feedback signal. S5. Based on the closed-loop feedback signal, adjust the initial electroacupuncture parameters to generate the final electroacupuncture parameters, which are then used for subsequent electroacupuncture stimulation control.

2. The method for calibrating EEG-synchronized electroacupuncture stimulation based on embedded AI according to claim 1, characterized in that, The specific steps for analyzing real-time EEG signal data and generating brain state characteristics include: The patient's EEG signal data is filtered to generate preprocessed EEG data; Frequency band energy parameters are extracted from the generated preprocessed EEG data to generate EEG feature vectors; Obtain classification thresholds used to define different brain state types; The brain state type is identified by comparing the EEG feature vector with the classification threshold, thereby generating brain state features.

3. The method for calibrating EEG-synchronized electroacupuncture stimulation based on embedded AI according to claim 2, characterized in that, Before the step of obtaining the classification thresholds used to define different brain state types, the method further includes: Before treatment begins, baseline EEG signal data of the patient at rest are collected. Analyze baseline EEG signal data to generate individualized baseline EEG characteristics for each patient; Based on individualized baseline EEG characteristics, classification thresholds are calibrated to generate personalized classification thresholds suitable for each patient.

4. The method for calibrating EEG-synchronized electroacupuncture stimulation based on embedded AI according to claim 1, characterized in that, The specific steps for applying association rules to brain state characteristics to generate initial electroacupuncture parameters include: Obtain user feedback data input by patients through the user interface; By integrating user feedback data with brain state characteristics, a comprehensive state index is generated; Based on comprehensive status indicators, association rules are applied to generate initial electroacupuncture parameters.

5. The method for calibrating EEG-synchronized electroacupuncture stimulation based on embedded AI according to claim 1, characterized in that, The specific steps for controlling the electroacupuncture device to apply electroacupuncture stimulation based on the initial electroacupuncture parameters include: To obtain the mapping relationship between brain state and acupoints; Based on brain state characteristics and by applying acupoint mapping relationships, the target stimulation site is determined; Control the electroacupuncture device to apply electroacupuncture stimulation to the target stimulation site.

6. The method for calibrating EEG-synchronized electroacupuncture stimulation based on embedded AI according to claim 5, characterized in that, After the step of controlling the electroacupuncture device to apply electroacupuncture stimulation to the target stimulation site, the method further includes: Extracting tissue impedance parameters from stimulus state data; The output waveform and pulse duration of electroacupuncture stimulation are dynamically adjusted based on the changes in the extracted tissue impedance parameters. The adjusted stimulation control command is generated and sent to the electroacupuncture device.

7. The method for calibrating EEG-synchronized electroacupuncture stimulation based on embedded AI according to claim 1, characterized in that, The specific steps for dynamically updating brain state characteristics to form a closed-loop feedback signal include: Analyze real-time EEG signal data continuously acquired during the application of electroacupuncture stimulation to calculate its changing trends and generate EEG response characteristics; By correlating EEG response characteristics with stimulus state data, the immediate effects of treatment intervention can be evaluated, and effect evaluation indicators can be generated. Based on the effect evaluation indicators, the brain state characteristics are updated and closed-loop feedback signals are generated.

8. The method for calibrating EEG-synchronized electroacupuncture stimulation based on embedded AI according to claim 1, characterized in that, The specific steps for adjusting the initial electroacupuncture parameters to generate the final electroacupuncture parameters include: Based on the closed-loop feedback signal, the adjustment direction and magnitude of the parameters are determined, and the parameter adjustment amount is generated; The generated parameter adjustments are applied to the initial electroacupuncture parameters to generate the final electroacupuncture parameters; The generated final electroacupuncture parameters are sent to the electroacupuncture device for use in the next cycle of electroacupuncture stimulation control.

9. The method for calibrating EEG-synchronized electroacupuncture stimulation based on embedded AI according to claim 1, characterized in that, Association rules can be updated using the following steps: After the treatment, the final electroacupuncture parameters and final brain state characteristics of this treatment are obtained to form the empirical data of this treatment; Acquire historical experience data accumulated during past treatment processes; The empirical data from this treatment is integrated with historical empirical data, and the association rules are learned and updated.

10. A brainwave-synchronized electroacupuncture stimulation calibration system based on embedded AI, characterized in that, include: The brain state feature generation module acquires the patient's real-time EEG signal data and analyzes the real-time EEG signal data to generate brain state features. The initial electroacupuncture parameter generation module obtains the association rules used to establish the relationship between brain state and electroacupuncture stimulation, and applies the association rules to brain state features to generate initial electroacupuncture parameters. The electroacupuncture stimulation application module controls the electroacupuncture device to apply electroacupuncture stimulation according to the initial electroacupuncture parameters, and collects the stimulation status data of the electroacupuncture device during the stimulation process; The closed-loop feedback signal generation module continuously collects real-time EEG signal data from the patient while applying electroacupuncture stimulation, and dynamically updates the brain state characteristics in combination with stimulation state data to form a closed-loop feedback signal. The final electroacupuncture parameter generation module adjusts the initial electroacupuncture parameters based on the closed-loop feedback signal to generate the final electroacupuncture parameters, which are then used for subsequent electroacupuncture stimulation control.

Images