Ear vagus nerve stimulation method based on electroencephalogram feedback

Abstract

The invention discloses a transaural vagus nerve stimulation method based on electroencephalogram feedback, and relates to the technical field of nerve regulation and control. According to the method, electroencephalogram signals of a specific brain region of a user are collected in real time through a flexible integrated electrode, time-frequency characteristics (such as delta, theta and alpha wave power) and nonlinear characteristics (such as entropy) are extracted after preprocessing, the current sleep stage or fatigue degree of the user is identified by using a CNN + LSTM model, taVNS stimulation parameters (frequency and intensity) are dynamically generated in combination with individual baseline electroencephalogram characteristics, and the user can be subjected to sleep state and fatigue state simulation. Stimulation is applied through an auricle stimulation electrode, and meanwhile, neural state changes are monitored in real time to iteratively optimize parameters. The problems that an existing taVNS target spot is not accurate, the stimulation mode is unknown, and the individual difference is large are solved through electroencephalogram feedback, safe, portable and personalized sleep and fatigue intervention is achieved, and remarkable clinical and application values are achieved.

Description

TECHNICAL FIELD

[0001] The application provides a transaural vagus nerve stimulation method, in particular to a transaural vagus nerve stimulation method based on electroencephalogram feedback. BACKGROUND

[0002] Sleep disorders and chronic fatigue have become important problems affecting the health of the population, and existing intervention techniques have significant limitations. Although drug treatment has a quick effect, it has the risk of addiction, liver and kidney function damage and other side effects; psychological intervention (such as cognitive behavioral therapy CBT-I) requires long-term guidance by professional personnel, and the popularization cost is high, which is difficult to cover a wide range of people. Among the neural regulation techniques, transcranial magnetic stimulation (TMS) and transcranial direct current stimulation (tDCS) are dependent on professional settings due to their large size and complex operation, and cannot meet the portability needs of home or mobile scenarios.

[0003] As a new non-invasive technology, transcutaneous auricular vagus nerve stimulation (taVNS) indirectly affects the central nervous system (such as the solitary nucleus, the locus coeruleus, the hippocampus, etc.) by stimulating the auricular branch of the vagus nerve, has the advantages of low risk, small side effects, and portable equipment, and has shown potential in chronic insomnia and fatigue management. However, there are three major bottlenecks in existing taVNS technology: first, the target is not accurate, the projection relationship between the auricular branch of the vagus nerve and the central brain area is complex, and existing research has only confirmed that it can activate the autonomic nervous system, but it cannot clearly define the specific division of labor of each brain area in sleep / cognitive regulation, resulting in the need for a "trial and error" mode in clinical application; second, the stimulation mode is not clear, the matching relationship between stimulation parameters (frequency, intensity, etc.) and neural response lacks systematic verification, for example, the neural regulation mechanism of low frequency (1-5Hz) and high frequency (20-100Hz) stimulation is not clear, and the parameter setting is mostly empirical; third, there are large individual differences, genetic background, brain network structure and other physiological characteristics significantly affect taVNS sensitivity, and existing technology cannot develop personalized programs according to individual brain function differences.

[0004] In addition, the structural design of existing taVNS devices has limitations: most devices only contain stimulation modules with fixed parameters, lack real-time neural activity monitoring functions, and cannot dynamically adjust stimulation strategies according to user status; although some devices integrate heart rate, blood oxygen and other physiological signal acquisition, electroencephalogram (EEG) signals as indicators directly reflecting central nervous activity are not included, making it difficult to accurately capture the real-time response of the cerebral cortex and subcortical nuclei, and making it difficult to achieve a "monitoring-regulation" closed loop. SUMMARY

[0005] To solve the problems of unclear target, rigid stimulation mode, poor individual adaptability of taVNS and the defects of safety and popularity of traditional intervention means in the prior art, the application provides a transauricular vagus nerve stimulation method based on electroencephalogram feedback, which realizes precise and personalized neural regulation by real-time acquisition and analysis of electroencephalogram signals and dynamic optimization of stimulation parameters.

[0006] The technical scheme adopted by the application to solve the technical problems is as follows: a transauricular vagus nerve stimulation method based on electroencephalogram feedback, comprising the following steps:

[0007] Electroencephalogram signal acquisition: the electroencephalogram (EEG) signals of a user are collected in real time by an electrode module arranged at a specific brain region (at least one of the frontal region, the parietal region and the occipital region) of the user's head. The electrode module is made of a flexible conductive material and integrates the functions of electroencephalogram acquisition and taVNS stimulation, optimizes the wearing structure based on 3D human modeling, ensures the head and neck fit, and reduces motion artifacts and electromagnetic interference.

[0008] Electroencephalogram signal processing and feature extraction: the collected electroencephalogram signals are preprocessed, the electromyographic interference is removed by adaptive filtering, and the electrooculogram artifacts are removed by independent component analysis (ICA); at least one of the power or power ratio of the delta wave, theta wave, alpha wave, beta wave and gamma wave is extracted by time-frequency analysis (short-time Fourier transform or wavelet transform), and at least one of the nonlinear indicators such as sample entropy, approximate entropy and coherence is extracted by nonlinear analysis (entropy calculation based on sliding time window) as the feature reflecting the current neural state of the user.

[0009] Neural state recognition: the extracted electroencephalogram features are input into a preset "CNN+LSTM" deep fusion model. The CNN is used to extract the local spatial features of the electroencephalogram time-frequency spectrum (such as frequency band abnormalities in a specific brain region), and the LSTM is used to capture the long-term time sequence dependence relationship of brain rhythm (such as sleep stage transition rules); the model outputs the current neural state of the user, including the sleep stage (awake, light sleep, deep sleep, rapid eye movement sleep) or the fatigue degree (alertness, mild fatigue, severe fatigue). The model is obtained by training of multiple source electroencephalogram samples, and the training samples contain labeled data corresponding to different sleep stages or fatigue degrees.

[0010] Stimulus parameter dynamic adjustment: According to the identified neural state, combined with the individual baseline EEG characteristics of the user (brain network controllability index in resting state, microstate transition characteristics), the taVNS target stimulation parameters (frequency, intensity) are generated based on the preset parameter mapping rules. For example: when the identification result is clear and mild fatigue, the target stimulation frequency is 20-100Hz; when the identification result is difficulty falling asleep or shallow sleep stage, the target stimulation frequency is 1-5Hz; when the identification result is deep sleep stage and the δ wave power is abnormally low, reduce the stimulation intensity to 50%-80% of the basic intensity.

[0011] Transaural vagus nerve stimulation: Through the stimulation electrode set in the vagus nerve distribution area of the user's auricle (its position is determined based on 3D human auricle modeling, which is the optimal stimulation site of the vagus nerve ear branch distribution, which is determined by pre-acquisition of stimulation response electroencephalogram signals at different auricle positions), the taVNS stimulation is applied according to the generated target stimulation parameters.

[0012] Feedback iteration optimization: During the stimulation process, steps 1-3 are repeated to monitor the changes in the user's neural state in real time; if it is monitored that the neural state does not reach the preset target (such as not entering the shallow sleep stage for 3 minutes), the target stimulation parameters are adjusted by re-executing step 4 until the neural state reaches the preset target or the stimulation period ends, forming a closed-loop control of "collection-analysis-decision-stimulation-feedback".

[0013] The one or more technical solutions provided in the embodiments of the present application have at least the following technical effects or advantages compared with the prior art:

[0014] A. Breakthrough in the core bottleneck of taVNS: Real-time capture of central nervous activity (such as brain region frequency band power, network microstate) through electroencephalogram signals, clear response characteristics of target brain area, solve the problem of "inaccurate target"; based on the dynamic mapping of EEG characteristics and stimulation parameters, the association between neural response and stimulation mode is established, and the "unknown stimulation mode" defect is overcome; combined with individual baseline EEG characteristics to optimize parameters, adapt to genetic and physiological differences, and improve individuality.

[0015] B. Improve the safety and popularity of intervention: No drug intervention, avoid addiction and side effects; integrated device is portable and easy to use, no need for professional operation, solve the popularization problem of traditional psychological intervention and TMS / tDCS.

[0016] C. Realize closed-loop precise regulation: Through the whole process automation of "electroencephalogram monitoring-state recognition-parameter adjustment-stimulation feedback", dynamically adapt to the changes of user's neural state (such as the transition from wakefulness to sleep stage), significantly improve the efficiency of sleep regulation and fatigue improvement.

[0017] Additional advantages, objects, and features of the application will be apparent to those skilled in the art upon examination of the following detailed description, it being understood that each of the foregoing general statements are intended to be illustrative only and not limiting as to the present application. BRIEF DESCRIPTION OF DRAWINGS

[0018] Figure 1 The overall workflow diagram of the transauricular vagus nerve stimulation method based on electroencephalogram feedback of the present application;

[0019] Figure 2 The closed-loop control workflow diagram of the transauricular vagus nerve stimulation method based on electroencephalogram feedback of the present application

[0020] Figure 3 The neural state recognition algorithm model diagram of the transauricular vagus nerve stimulation method based on electroencephalogram feedback of the present application

[0021] Figure 4 The stimulation parameter dynamic mapping relationship diagram of the transauricular vagus nerve stimulation method based on electroencephalogram feedback of the present application. DETAILED DESCRIPTION

[0022] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0023] It should be noted that the terms "vertical", "horizontal", "up", "down", "left", "right" and similar expressions used herein are only for the purpose of illustration, and do not represent the only implementation.

[0024] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the present application belongs; the terms used herein in the specification only for the purpose of describing specific embodiments, and are not intended to limit the present application; the term "and / or" used herein includes any and all combinations of one or more related listed items.

[0025] As shown in Figures 1-4 The core components of the transauricular vagus nerve stimulation device based on electroencephalogram feedback of the present application include a flexible integrated electrode module, an electroencephalogram signal processing module, a neural state recognition module, a stimulation parameter generation module, a taVNS stimulation module, a feedback iterative control module, and an individual baseline electroencephalogram database. The connection and functional relationship of each component are as follows:

[0026] The signal output end of the flexible integrated electrode module is electrically connected to the signal input end of the EEG signal processing module, and is used to transmit the acquired raw EEG signals to the processing module; the stimulation signal input end of the electrode module is electrically connected to the signal output end of the taVNS stimulation module, and is used to receive stimulation parameters and apply stimulation to the auricular branch of the vagus nerve of the auricle.

[0027] The feature output terminal of the EEG signal processing module is communicatively connected to the feature input terminal of the neural state recognition module, which is used to transmit the preprocessed extracted EEG time-frequency features and nonlinear features to the recognition module;

[0028] The state output terminal of the neural state recognition module is communicatively connected to the state input terminal of the stimulus parameter generation module. At the same time, the stimulus parameter generation module is also communicatively connected to the individual baseline EEG database to generate target stimulus parameters by combining the current neural state with baseline features such as the controllability of the user's resting-state brain network and micro-state transitions.

[0029] The parameter output terminal of the stimulation parameter generation module is connected to the parameter input terminal of the taVNS stimulation module to transmit the target stimulation parameters (frequency, intensity) to the stimulation module.

[0030] The feedback iterative control module is connected to the EEG signal processing module, the neural state recognition module, and the stimulus parameter generation module respectively. It is used to receive the current neural state recognition results in real time, determine whether the preset intervention target has been achieved, and if not, trigger the stimulus parameter generation module to regenerate the parameters and control each module to repeatedly execute the "signal acquisition-processing-recognition-parameter adjustment" process to form a closed loop control.

[0031] In this implementation scheme, the combination of "flexible integrated electrode module + neural state recognition module" is used to directly capture the electrical activity characteristics of the cortex and subcortical nuclei (such as the thalamus and locus coeruleus) using EEG signals (such as the recovery of parietal alpha wave power after nap and abnormal microstate transition after SD). This clarifies the core brain regions that are related to sleep / fatigue regulation under different neural states, providing direct neural activity evidence for taVNS stimulation. For example, when "abnormally low delta wave power during deep sleep" is detected, it can be determined that the thalamus's regulatory function is insufficient, and the stimulation intensity can be reduced in a targeted manner to protect the deep sleep brain network, replacing the traditional "trial and error" target selection and improving the precision of regulation.

[0032] Through the combination of "neural state recognition module + stimulation parameter generation module + individual baseline database", the mapping relationship of "neural state-individual baseline-stimulation parameter" is established: on the one hand, the "CNN+LSTM" model trained based on multiple source EEG samples (CNN extracts spatial features of brain regions, and LSTM captures temporal regularity) can accurately distinguish different states such as "awake-mild fatigue", "difficulty falling asleep" and "deep sleep", avoiding recognition errors of single algorithm; on the other hand, combined with individual baseline characteristics (such as high resting state alpha wave power of a user, indicating that he is more sensitive to low frequency stimulation), the adaptive stimulation parameters (such as 1-3Hz instead of 5Hz stimulation when the user has difficulty falling asleep) are generated, breaking through the limitations of traditional taVNS "empirical parameter setting" and establishing scientific stimulation mode matching rules.

[0033] On the one hand, the introduction of individual baseline EEG database can store different users' genetic background associated brain characteristics (such as genetic polymorphism related brain network controllability difference) and physiological state characteristics (such as delta wave baseline power of different ages), so that the stimulation parameter generation module can adapt to individual differences and avoid "one-size-fits-all" stimulation scheme; on the other hand, the whole process automation "collection-analysis-decision-stimulation" closed loop design does not require professional operation, and users can independently complete wearing and intervention, while avoiding the defects of drug addiction and psychological intervention professional dependence, realizing "safe-portable-personalized" integrated intervention, and significantly improving the popularity of technology.

[0034] The present application (ear vagus nerve stimulation device based on EEG feedback) can adapt to three core scenes of home daily intervention, clinical diagnosis and treatment assistance, and mobile scene emergency regulation, and through connection with existing home sleep monitoring equipment, clinical neuroelectrophysiological equipment and mobile terminals, the problems of "isolated operation, data fragmentation and poor adaptability" of traditional devices are solved. The connection mode, use state and linkage details of the present application and the prior art are described in detail as follows:

[0035] I. Scene 1: Home daily sleep intervention (adapt to existing home sleep monitoring equipment)

[0036] 1. Prior art / device linkage object

[0037] Home sleep monitoring equipment (such as smart bracelet, sleep monitoring mattress) has heart rate, blood oxygen and body movement data collection function; household power supply socket (provides basic power supply); mobile phone APP (used for data display and parameter preset).

[0038] 2. Connection and installation details

[0039] Data connection: The EEG signal processing module of the device is connected to the smart bracelet and the mobile phone APP through Bluetooth 5.0 wireless connection, realizing the bidirectional synchronization of "EEG data-heart rate / body movement data"; the device host is connected to the household power socket through the Type-C interface, or uses the built-in lithium battery (8-hour endurance) to realize wireless operation.

[0040] Physical installation: The user wears the flexible integrated electrode module on the head and auricle before going to bed - the EEG acquisition end is attached to the frontal region (Fp1 / Fp2 sites), the parietal region (P3 / P4 sites), and is fixed by medical gel (to avoid falling off at night); the stimulation end is attached to the auricle and the concha (the dense area of the auricular branch of the vagus nerve), and is fixed by an elastic bandage (to adapt to different auricle sizes); the device host (including processing module, recognition module, and stimulation module) can be placed on the bedside table, or hung on the neck through a hanging rope (weight <100g, does not affect the sleeping posture).

[0041] 3. Usage state of each structure and linkage logic

[0042] Flexible integrated electrode module: Real-time acquisition of frontal and parietal EEG signals (sampling rate 250Hz, accuracy ±1μV) and auricle skin contact impedance (to ensure smooth stimulation circuit), transmission of raw EEG signals to the EEG signal processing module, and reception of low-frequency pulse signals (1-5Hz, used to assist sleep) from the taVNS stimulation module.

[0043] EEG signal processing module: "Adaptive filtering (remove turning muscle EMG interference) + ICA independent component analysis (remove eye movement artifacts)" preprocessing of raw EEG signals, extraction of δ wave (0.5-4Hz) power and α wave (8-13Hz) power ratio, synchronous reception of heart rate variability (HRV) data transmitted by the smart bracelet, and fusion of "EEG features + physiological features" and sending to the neural state recognition module.

[0044] Neural state recognition module: combined with the user's baseline data (such as resting state δ wave power mean) preset by the mobile phone APP, the sleep state is judged by the "CNN+LSTM" model - if it is identified as "difficulty falling asleep (α wave power ratio >40%, HRV <50ms)", send "low-frequency stimulation required" instruction to the stimulation parameter generation module.

[0045] Stimulation parameter generation module: generates stimulation parameters of "frequency 2Hz, intensity 1.0mA", transmits to the taVNS stimulation module; at the same time, the parameters are synchronized to the mobile phone APP, and the user can manually fine-tune (±0.2mA intensity) through the APP.

[0046] Feedback iteration control module: receive the neural state recognition result every 5 minutes, if it is monitored that "the alpha wave power is reduced to below 20% for 3 minutes (entering light sleep)", trigger the stimulation parameter generation module to reduce the intensity to 0.6 mA; if it is identified as "deep sleep (delta wave power ratio > 50%)", suspend the stimulation to avoid disturbing the deep sleep period - this process does not require user operation, the whole process is completed by the closed-loop adjustment of the device and the smart bracelet, solving the defects of the traditional home taVNS device "fixed parameter stimulation, unable to adapt to sleep stage".

[0047] II. Scenario two: clinical insomnia / fatigue diagnosis and treatment assistance (adapt to existing clinical neuroelectrophysiological equipment)

[0048] 1. Existing technology / device linkage object

[0049] Clinical electroencephalogram (such as Nihon Kohden EEG system, with high-resolution electroencephalogram acquisition function), hospital electronic medical record system (store patient history and baseline electroencephalogram data), doctor workstation (for parameter setting and result analysis).

[0050] 2. Connection and installation details

[0051] Data connection: the electroencephalogram signal processing module of the device is connected to the clinical electroencephalogram through Ethernet, realizing real-time synchronization of electroencephalogram data (sampling rate is increased to 1000 Hz, meeting the accuracy of clinical diagnosis); the stimulation parameter generation module is connected to the doctor workstation and electronic medical record system through the hospital intranet, which can call the patient's past sleep monitoring report (such as PSQI score, past taVNS stimulation response record).

[0052] Physical installation: assisted by medical staff to wear flexible integrated electrode module, increase temporal (T3 / T4) and occipital (O1 / O2) electrode sites (cover key areas of whole brain) for electroencephalogram acquisition, improve signal acquisition stability through conductive paste; the stimulation end is calibrated to the "optimal target point of vagus nerve auricular branch" (based on patient auricle CT modeling data) through infrared positioning instrument, ensuring stimulation accuracy; the device host is fixed on the equipment cart beside the treatment bed, sharing the stabilized power supply with the electroencephalogram, avoiding voltage fluctuation affecting the signal.

[0053] 3. Usage state of each structure and linkage logic

[0054] Flexible integrated electrode module: acquire 16-channel electroencephalogram signals, transmit to the electroencephalogram signal processing module at the same time, and send to the clinical electroencephalogram for the doctor to compare the consistency of "device electroencephalogram data and standard EEG data", ensuring the accuracy of feature extraction.

[0055] Neural state recognition module: combined with patient baseline data in electronic medical record system (e.g. chronic insomnia patients have lower delta wave power baseline), output "sleep stage + brain network state" dual results (e.g. "light sleep, reduced top- thalamic connectivity"), sent to physician workstation - physicians can manually adjust stimulation strategy based on this (e.g. increase stimulation frequency to 3 Hz, enhance thalamic regulation).

[0056] Stimulation parameter generation module: receive custom parameters from physician workstation (e.g. "frequency 4 Hz, intensity 1.5 mA, stimulation period 20 minutes") or generate parameters based on default clinical protocols, transmit to taVNS stimulation module; at the same time, real-time feedback of brain electrical response data during stimulation (e.g. post-stimulation delta wave power change rate) to the electronic medical record system, forming diagnosis and treatment records.

[0057] Feedback iterative control module: if the electroencephalograph detects "abnormal brain activity (e.g. sharp waves) after stimulation", the stimulation module will be suspended immediately and an alarm will be sent to the physician workstation - this linkage design solves the problem of "lack of real-time electroencephalogram monitoring and safety hazards" in traditional clinical taVNS devices, while achieving "diagnosis and treatment data - stimulation scheme - medical record" integrated management, improving clinical efficiency.

[0058] Three, scenario three: mobile scene fatigue emergency adjustment (adapt to existing mobile terminal equipment)

[0059] 1. Existing technology / device linkage object

[0060] Mobile phone APP (with fatigue state reminder function), car Bluetooth system (for driving scene linkage), portable power bank (provides mobile power supply).

[0061] 2. Connection and installation details

[0062] Data connection: the device connects the mobile phone APP and the car Bluetooth system through Bluetooth Low Energy (BLE), the mobile phone APP can read the neural state recognition results of the device, and the car system can receive the "fatigue warning" signal of the device; the device is connected to the power bank through the Type-C interface (5V / 2A power supply), supporting 4 hours of continuous use.

[0063] Physical installation: the flexible integrated electrode module adopts "lightweight ear-hanging design" - the EEG acquisition end is simplified to a single electrode on the frontal region (attached to the Fz site), fixed on the head through the ear-hanging; the stimulation end is integrated on the inner side of the ear-hanging, naturally attached to the triangular fossa of the auricle; the device host is integrated at the end of the ear-hanging (weight < 30g), which does not affect the user's walking, driving and other actions.

[0064] 3. Usage state of each structure and linkage logic

[0065] Flexible integrated electrode module: In the driving / office scene, the real-time acquisition of the simplified electroencephalogram signal (focusing on alpha wave, theta wave ratio) is transmitted to the electroencephalogram signal processing module; the stimulation end outputs a high-frequency pulse signal (50-80Hz, used to improve alertness) after receiving the parameters, and the stimulation intensity is controlled at 0.8-1.2mA (to avoid interfering with the user's attention).

[0066] Electroencephalogram signal processing module: quickly pre-process the electroencephalogram signal (processing delay <1 second), extract the theta / alpha wave ratio (theta / alpha>0.6 when tired), and send to the neural state recognition module.

[0067] Neural state recognition module: if it is identified as'mild fatigue (theta / alpha=0.6-0.8)', send 'fatigue reminder' to the mobile phone APP, and send 'high-frequency low-intensity stimulation' instruction to the stimulation parameter generation module; if it is identified as'severe fatigue (theta / alpha>0.8)', send 'warning signal' to the vehicle-mounted Bluetooth system synchronously, and trigger the vehicle-mounted screen flicker reminder.

[0068] Stimulation parameter generation module: generate 'frequency 60Hz, intensity 1.0mA' stimulation parameters, transmit to the taVNS stimulation module, pause for 10 seconds every 30 seconds of stimulation (to avoid discomfort caused by continuous stimulation); the mobile phone APP displays 'current fatigue degree (percentage)' and'stimulation remaining time' in real time, and the user can manually stop the stimulation.

[0069] Feedback iterative control module: update the neural state every 1 minute, if the theta / alpha ratio decreases to below 0.5 (alertness recovery), automatically reduce the stimulation intensity to 0.6mA; if the theta / alpha ratio does not decrease for 5 minutes, prompt'suggestion to rest' through the mobile phone APP - this design solves the problem of no precise fatigue intervention device in the traditional mobile scene, realizes the linkage of'real-time monitoring-emergency regulation-safety warning', and is suitable for driving, commuting and other high-frequency mobile scenes.

[0070] Although the present application has been disclosed as above with reference to the preferred embodiments, it is not intended to limit the present application, and any person skilled in the art can make various modifications and modifications without departing from the spirit and scope of the present application, therefore the protection scope of the present application should be defined by the claims.

Claims

1. A transauricular vagus nerve stimulation method based on electroencephalographic feedback, characterized in that, Includes the following steps: (1) EEG signal acquisition: The user’s EEG signal is acquired in real time by an electrode module placed in a specific brain region of the user’s head. The EEG signal is used to reflect the user’s current sleep state or fatigue-related neural activity state. (2) EEG signal processing and feature extraction: The EEG signals collected in step (1) are preprocessed to remove noise, and then EEG features are extracted by time-frequency analysis and nonlinear analysis methods. The EEG features include the power or power ratio of at least one frequency band of δ wave, θ wave, α wave, β wave, γ wave, and at least one nonlinear index of sample entropy, approximate entropy, coherence. (3) Neurological state recognition: Input the EEG features extracted in step (2) into the preset state recognition model and output the user's current neurological state. The neurological state includes sleep stage (awake, light sleep, deep sleep, REM sleep) or fatigue level (alert, mild fatigue, severe fatigue). (4) Dynamic adjustment of stimulation parameters: Based on the neural state identified in step (3), target stimulation parameters for transcutaneous vagus nerve stimulation (taVNS) are generated based on preset parameter mapping rules. The target stimulation parameters include stimulation frequency and stimulation intensity. (5) Transauricular vagus nerve stimulation: TaVNS stimulation is applied to the user by means of stimulation electrodes placed in the distribution area of ​​the vagus nerve in the user's auricle, according to the target stimulation parameters generated in step (4).

2. The method according to claim 1, characterized in that, In step (1), the specific brain region includes at least one of the frontal region, parietal region, and occipital region. The electrode module is an integrated electrode made of flexible conductive material. The integrated electrode integrates EEG acquisition function and taVNS stimulation function.

3. The method according to claim 1, characterized in that, In step (2), the preprocessing includes using adaptive filtering to remove electromyographic interference and using independent component analysis (ICA) to remove electrooculography artifacts. The time-frequency analysis includes short-time Fourier transform or wavelet transform, and the nonlinear analysis includes entropy calculation based on a sliding time window.

4. The method according to claim 1, characterized in that, In step (3), the state recognition model is The "CNN+LSTM" deep fusion model uses CNN to extract spatial features of EEG time-frequency maps and LSTM to capture the temporal dependencies of EEG signals. The state recognition model is trained using multi-source EEG samples, which include labeled EEG data corresponding to different sleep stages or different levels of fatigue.

5. The method according to claim 1, characterized in that, In step (4), the parameter mapping rules include: when the identification result is awake and slightly fatigued, the target stimulation frequency is 20-100Hz; when the identification result is difficulty falling asleep or light sleep stage, the target stimulation frequency is 1-5Hz; when the identification result is deep sleep stage and the delta wave power is abnormally low, the stimulation intensity is reduced to 50%-80% of the base intensity.

6. The method according to claim 1, characterized in that, It also includes step (6) feedback iteration optimization: during the application of stimulation in step (5), repeat steps (1)-(3) to monitor the changes in the user's neural state in real time. If the neural state is not detected to reach the preset target (e.g., not entering the light sleep stage for 3 minutes), then step (4) is re-executed to adjust the target stimulation parameters until the neural state reaches the preset target or the stimulation cycle ends.

7. The method according to claim 1, characterized in that, In step (4), when generating the target stimulus parameters, the user's individual baseline EEG characteristics are also combined. The individual baseline EEG characteristics are the EEG characteristics collected by the user in a resting state, including brain network controllability indicators or microstate transition characteristics.

8. The method according to claim 1, characterized in that, In step (5), the position of the stimulation electrode is determined based on 3D human auricle modeling, and is aligned with the optimal stimulation site of the auricular branch of the vagus nerve. The optimal stimulation site is determined by pre-collecting stimulation response EEG signals from different auricle positions.

Images