Training method of taVNS combined with six-character table slow breathing

Abstract

The invention discloses a taVNS and six-character table combined slow respiration training method, which comprises the following steps of: synchronously acquiring electrocardio and electromyographic signals of a user, analyzing heart rate variability (HRV) and respiratory electromyographic characteristics in real time, and dynamically judging a respiratory phase; parameters and target spots of ear vagus nerve stimulation (taVNS) are adjusted in a self-adaptive mode based on a breathing phase-target region mapping table, and closed-loop control is formed in combination with six-character table breathing training. The defects that a traditional open-loop taVNS system is fixed in stimulation parameter and poor in adaptability are overcome.

Description

Technical Field

[0001] This invention relates to the field of bioelectrical signal detection, specifically to a training method combining taVNS with slow breathing using the Six Healing Sounds. Background Technology

[0002] Transcutaneous vagus nerve stimulation (taVNS) is an open-loop neuromodulation technique. Currently, it only stimulates the auricular branch of the vagus nerve in the concha / tragus region using fixed parameters to regulate the autonomic nervous system and related brain networks. It does not yet utilize biofeedback signals to control the stimulation process in real time, thus achieving closed-loop stimulation. TaVNS uses the auricular branch of the vagus nerve (Arnold's nerve) in the concha and tragus to: activate the nucleus tractus solitarius-dorsolateral vagal complex, enhancing parasympathetic tone; and modulate the prefrontal-limbic system circuit, improving anxiety, depression, and cognitive function. It is currently widely used to treat psychosomatic disorders such as migraines, insomnia, irritable bowel syndrome, functional dyspepsia, depression, and anxiety. These effects exhibit a dose-response relationship with stimulation intensity, frequency, pulse width, phase, and duration. Current clinical studies and animal experiments show that taVNS, at 0.5-2.0 mA, 20-30 Hz, pulse width 200-500 μs, and synchronized with the respiratory phase, can maximize parasympathetic activation and avoid sympathetic rebound. Open-loop fixed parameters are prone to "overstimulation" or "understimulation" due to individual differences. Real-time detection of heart rate, respiratory rate and depth, and emotional state changes enables more precise treatment (including stimulation frequency and dosage). Target phase sensitivity: Different areas of the auricle correspond to different nerve segments in the visceral reflex zones. The Six Healing Sounds (Liu Zi Qiong) is a breathing-based guided exercise; different sounds correspond to different acupoints, achieving a correspondence between auricular acupoint stimulation and respiratory stimulation of the internal organs, while maintaining the advantage of slow breathing. The ear stimulated by taVNS also corresponds to acupoints. In each stage of the Six Healing Sounds ("Xu-He-Hu-Xi-Chui-Xi"), the vagal afferent thresholds of the liver, heart, spleen, lungs, kidneys, and triple burner are different. Dynamically switching target points according to the respiratory phase can achieve "dual-synchronous treatment," increasing efficacy. Energy saving and safety: Closed-loop technology can reduce the average stimulation energy by 40-60%, reducing side effects such as skin irritation and dizziness, while extending battery life and making wearable technology possible. Summary of the Invention

[0003] To address the aforementioned shortcomings in the existing technology, this invention provides a training method combining taVNS with slow breathing using the Six Healing Sounds.

[0004] To achieve the above-mentioned objectives, the technical solution adopted by this invention is as follows: A training method combining TaVNS with the Six Healing Sounds for slow breathing includes the following steps: S1, synchronously collects the user's electrocardiogram and electromyography signals; S2, Calculate respiratory depth and amplitude parameters based on the electrocardiogram and electromyography signals; S3, determine the current respiratory phase based on the respiratory depth and amplitude parameters; S4, determine the auricular stimulation target area and stimulation parameters based on the respiratory phase lookup mapping table; S5, apply taVNS stimulation to the auricular stimulation target area; S6, the user performs the Six Healing Sounds breathing exercise according to the rhythm, and returns to step S1 to form a closed-loop control.

[0005] Furthermore, step S1 includes: S11 receives data streams from ECG and EMG sensors in real time via Bluetooth interface; S12, the data stream is stored in the communication thread data pool of the host computer for analysis and processing.

[0006] Furthermore, step S2 includes an electrocardiogram (ECG) signal analysis sub-step and an electromyography (EMG) signal analysis sub-step, wherein, The electrocardiogram signal analysis sub-step includes: S21, the Pan-Tompkins algorithm is used to extract the peak point of the R wave in the electrocardiogram signal; S22, calculate the time interval between adjacent R-wave peaks, divide it by the sampling frequency to obtain the RR interval sequence, the formula is:

[0007] In the formula, For the i-th RR interval, t i and t i+1 These are the peak times of adjacent R-waves. The sampling frequency; S23, calculate heart rate variability characteristics based on the RR interval sequence, including mean RR interval MeanRR, mean heart rate MeanHR, RR interval standard deviation SDNN, root mean square RMSSD of the difference between adjacent RR intervals, and the proportion of adjacent RR interval differences greater than 50ms pNN50 and the proportion of adjacent RR interval differences greater than 20ms pNN20. S24, Perform Fourier transform on the RR interval sequence to calculate spectral characteristics, including very low frequency power (VLF), low frequency power (LF), high frequency power (HF), low frequency to high frequency power ratio (LF / HF), and normalized low frequency power (LF) and normalized high frequency power (HF).

[0008] The electromyographic signal analysis sub-step includes: S25 performs bandpass filtering (20-450Hz), power frequency notch filtering, full-wave rectification, and low-pass filtering (2-5Hz) on the electromyography signal to obtain the envelope signal; S26, on the envelope signal, an adaptive threshold and minimum interval peak detection method is used to extract burst peaks, and the time interval sequence PI between adjacent burst peaks is calculated using the following formula:

[0009] In the formula, This represents the interval of the j-th peak. S27, calculate rhythmic features based on the PI sequence, including average PIM, average activation rate, PI standard deviation SDPI, PI sequence RMSSD, and the proportion of PI difference greater than 50ms and greater than 20ms. S28, within the sliding window, statistical amplitude characteristics, including root mean square (RMS), mean absolute value (MAV), integrated electromyography (IEMG) value, and peak-to-valley difference; S29. Perform a Fourier transform on the bandpass-filtered electromyography signal to calculate the power spectrum characteristics, including the median frequency (MNF), average power frequency (MPF), low-frequency power, high-frequency power, high-frequency power ratio, and spectral centroid. Furthermore, determining the current breathing phase in step S3 includes: S31, the local maximum method is used to extract the peak point of the respiratory wave; S32, the waveform is divided by each respiratory wave peak point to distinguish between the exhalation and inhalation phases; S33. Take the three most recent respiratory waveforms and calculate the average expiratory duration and average inspiratory duration as the basis for respiratory phase determination.

[0010] Furthermore, the mapping table mentioned in step S4 is a respiratory phase-auricular target area mapping table, configured as follows: When the user pronounces the sound "shh," it corresponds to the liver aspect, stimulating the liver area ear acupoints; When the user pronounces the sound "ha", it corresponds to the heart phase and stimulates the acupoints in the heart area of ​​the ear; When a user pronounces the sounds "hu," "yan," "chui," and "xi," these sounds correspond to the lung, spleen, kidney, and triple burner, respectively, stimulating the corresponding ear acupoints.

[0011] Furthermore, the application of taVNS stimulation in step S5 includes: S51, based on the stimulation parameters determined in step S4, a control command is sent to the otoelectric controller via the host computer; S52, the otoelectric controller adjusts the stimulation intensity, frequency, and target location according to the command; S53, stimulation is activated and deactivated in sync with the respiratory phase to maximize vagal nerve activation.

[0012] Furthermore, step S6 includes: S61 plays a video tutorial of the Six Healing Sounds on the host computer interface, providing rhythm guidance; S62, the user performs the six-sound mantra movements and breathing according to the video, the six-sound mantra including the six sounds "xu, he, hu, yan, chui, xi"; S63 monitors user performance quality in real time and assesses training effectiveness through the electromyography-heart rate coupling index.

[0013] Furthermore, it also includes step S7: S71, the treatment effect evaluation model is invoked at intervals of one minute, and the model is implemented using a support vector machine; S72, the input features of the model include heart rate variability features, electromyographic features, and progress of practicing the Six Healing Sounds; S73, the model outputs the probability of the optimal strategy for controlling the otoelectric effect, and the strategy with the highest probability is selected to dynamically adjust the stimulation parameters.

[0014] The present invention has the following beneficial effects: This invention overcomes the problems of insufficient sensitivity and susceptibility to respiratory rhythm interference caused by using HRV as a single feedback indicator in existing technologies. By simultaneously introducing electromyography (EMG) to monitor respiratory muscle exertion and rhythm, it forms a dual peripheral feedback with HRV, thereby improving the accuracy and stability of closed-loop control.

[0015] This invention overcomes the shortcomings of existing taVNS systems, such as lack of active breathing-motor intervention and inability to amplify vagal tone. It uses the "Six Healing Sounds" breathing-pronunciation-limb guidance as a user-controllable active input, forming a multi-loop closed loop of "motor + breathing → EMG / HRV → taVNS → motor + breathing".

[0016] This invention overcomes the problem of fixed stimulation targets and inability to match them in real time with the breathing-voicing-visceral reflex zones—by establishing a "breathing phase-auricular target area mapping table" and dynamically switching the position of stimulation electrodes according to the six-syllable mantra pronunciation stage, thus achieving individualized and precise target areas.

[0017] This invention overcomes the shortcomings of existing systems, such as poor adaptability to patients with abnormal breathing depth and rhythm, and large fluctuations in efficacy. Based on this, it adaptively adjusts the stimulation parameters so that the system can maintain a constant and predictable intervention effect for users with different breathing modes. Attached Figure Description

[0018] Figure 1 This is a schematic diagram of the process of the present invention. Detailed Implementation

[0019] The specific embodiments of the present invention are described below to enable those skilled in the art to understand the present invention. However, it should be understood that the present invention is not limited to the scope of the specific embodiments. For those skilled in the art, various changes are obvious as long as they are within the spirit and scope of the present invention as defined and determined by the appended claims. All inventions utilizing the concept of the present invention are protected.

[0020] A training method combining taVNS with the Six Healing Sounds for slow breathing, as shown in the figure, includes the following steps: S11 receives data streams from ECG and EMG sensors in real time via Bluetooth interface; In this embodiment, electromyography (EMG, including the diaphragm, rectus abdominis, and sternocleidomastoid muscles) is introduced to monitor respiratory effort in real time, which, together with HRV, constitutes a "dual peripheral feedback," specifically including the following steps: Specifically, it includes the following sub-steps: S12, the data stream is stored in the communication thread data pool of the host computer for analysis and processing.

[0021] The host computer (such as an embedded system or mobile device) receives data streams from the ECG and EMG sensors in real time via a Bluetooth Low Energy (BLE) interface. The ECG sensor uses a standard lead configuration with a sampling frequency of 250-1000 Hz to ensure R-wave detection accuracy. The EMG sensor focuses on respiratory-related muscle groups (such as the diaphragm, rectus abdominis, and sternocleidomastoid muscles) with a sampling frequency of 500-2000 Hz to capture high-frequency electromyographic activity. Data is streamed into a multi-threaded architecture and stored in a communication data pool; the buffer size is dynamically adjusted according to real-time requirements to prevent data loss.

[0022] Because ECG and EMG signals may experience transmission delays, hardware timestamps or software interpolation methods are used to achieve signal synchronization. Specifically, a millisecond-level timestamp is added to each data packet, and a delay compensation value is calculated using a cross-correlation algorithm to ensure that the ECG R wave and the EMG burst peak are aligned on the time axis. This step provides a temporal consistency basis for subsequent joint analysis.

[0023] S2, Calculate respiratory depth and amplitude parameters based on the electrocardiogram and electromyography signals; In this embodiment, the host computer analyzes and processes the data, obtaining different otoscopic control strategies based on medical knowledge, and sends control commands to the otoscopic controller. The control strategy selection algorithm is a treatment effect evaluation model, implemented using an SVM model. The model outputs the probability of the optimal otoscopic control strategy, and the strategy with the highest probability is selected for treatment. The evaluation model is called every minute to adjust the treatment strategy in real time. The model input includes feature values ​​from ECG and EMG analysis, as well as the progress of practicing the Six Healing Sounds.

[0024] The analysis and processing methods are divided into two parts.

[0025] Electrocardiogram: 1. The Pan–Tompkins algorithm is used to extract the peak value of the R-wave signal, and the implementation method is a Python library function.

[0026] 2. Calculate the time interval between adjacent R peaks, divide by the sampling frequency, and obtain the rr interval data in seconds. Calculate MeanRR, MeanHR (mean RR / heart rate), SDNN (standard deviation of RR), RMSSD (root mean square of the difference between adjacent RR values), and pNN50 / pNN20 (the proportion of adjacent RR differences > 50 / 20 ms). 3. Use Fourier transform to obtain the spectral data of the rr interval. Obtain the VLF / LF / HF power, LF / HF (cross / sub-cross inductance balance), normalized LF, and HF (normalized power) characteristics.

[0027] Electromyography: 1. A stable envelope is obtained by using "bandpass filtering (20–450 Hz) + power frequency notch filtering (50 / 60 Hz) + full-wave rectification + low-pass filtering (2–5 Hz)". On the envelope, burst peaks / activation peaks are extracted by peak detection with adaptive threshold and minimum interval. The implementation method is Python library functions.

[0028] 2. Calculate the difference between adjacent burst peak time points, divide by the sampling frequency to obtain the PI (peak interval) sequence in seconds; calculate rhythmic features such as MeanPI, MeanAR (=1 / MeanPI), SDPI, RMSSD_PI, pNN50_PI / pNN20_PI, and statistically analyze amplitude features such as RMS, MAV, IEMG, and peak-valley difference within a fixed sliding window to characterize the force intensity.

[0029] 3. Use the Fourier transform method to obtain the EMG power spectrum after bandpass, and calculate fatigue-related characteristics such as MNF (median frequency), MPF (average power frequency), Low (20–150 Hz) / High (150–350 Hz) band power, High / Low ratio, and spectral centroid.

[0030] Specifically: The Pan-Tompkins algorithm is used to detect the peak value of the R wave in ECG signals in real time. This algorithm includes bandpass filtering (5-15 Hz), differentiation, squaring, and sliding window integration, with adaptive threshold updates to suppress noise. It can be implemented using Python libraries (such as Biosppy) or embedded C code, ensuring efficient operation on resource-constrained devices. The interval between adjacent R-wave peaks (RR interval) is calculated based on the detected R-wave peak time points. The formula is:

[0031] In the formula, For the i-th RR interval, t i and t i+1 These are the peak times of adjacent R-waves. The sampling frequency is typically set to a 5-10 minute window to balance real-time performance and stability.

[0032] Calculating multiple features based on RR interval sequences: Mean RR interval (MeanRR): This reflects the average heart rate; Mean Heart Rate (MeanHR): (Unit: bpm); RR interval standard deviation (SDNN): measures overall variability, and the formula is:

[0033] The root mean square (RMSSD) of the difference between adjacent RR intervals: characterizes parasympathetic activity, and the formula is:

[0034] pNN50 and pNN20: Calculate the proportions of adjacent RR differences greater than 50ms and 20ms, respectively. The formula is as follows:

[0035] After resampling (e.g., 4 Hz) and detrending the RR interval sequence, the power spectral density is calculated using Fast Fourier Transform (FFT) or Lomb-Scargle periodograms. Features include: Extremely low frequency power (VLF, 0.003-0.04 Hz): related to body temperature regulation; Low-frequency power (LF, 0.04-0.15 Hz): reflects the balance between sympathetic and parasympathetic induction; High-frequency power (HF, 0.15-0.4 Hz): characterizes respiratory sinus arrhythmia and is directly related to vagal nerve activity; LF / HF ratio: assesses autonomic balance. The electromyographic signal is first bandpass filtered (Butterworth filter, order 4) from 20-450 Hz to remove baseline drift and high-frequency noise, and then a 50 / 60 Hz power frequency notch filter (e.g., an adaptive filter) is applied to suppress power supply interference. Subsequently, full-wave rectification and a 2-5 Hz low-pass filter (first-order Butterworth filter) are performed to obtain the envelope signal, which approximately reflects the muscle activation intensity. On the envelope signal, an adaptive thresholding method (e.g., dynamic threshold = mean + 3 × standard deviation) is used to detect burst peaks, and a minimum interval (e.g., 0.5 seconds) is set to avoid spurious peaks. The time interval between adjacent burst peaks (PI sequence) is calculated using the following formula:

[0036] Its rhythmic characteristics include mean PI (MeanPI), mean activation rate (MeanAR, where MeanAR=1 / MeanPI), standard deviation of PI (SDPI), RMSSD of PI sequence (RMSSD_PI), and pNN50_PI and pNN20_PI (analogous to HRV calculation). Amplitude characteristics include root mean square (RMS), mean absolute value (MAV), integrated electromyography (IEMG) value, and peak-to-trough difference of the statistical envelope signal within a sliding window (e.g., 10 seconds), quantifying respiratory effort.

[0037] Based on the above signals, perform FFT on the original bandpass filtered electromyography signal to calculate: Median frequency (MNF): the midpoint of the power spectrum, reflecting muscle fatigue; Average power frequency (MPF): power-weighted average frequency; High-low frequency power ratio: The ratio of power in the High (150-350 Hz) to Low (20-150 Hz) frequency bands, used to assess fatigue status; Spectral centroid: further characterizes the spectral distribution.

[0038] S3, determine the current respiratory phase based on the respiratory depth and amplitude parameters; In this embodiment, the specific steps include the following: S31, the local maximum method is used to extract the peak point of the respiratory wave; Specifically, the local maximum method is used to extract peak points from the preprocessed respiratory signal (which can be obtained from an ECG-derived or independent respiratory sensor). The algorithm sets a minimum peak height (e.g., 20% of the signal amplitude) and a minimum peak interval (e.g., 2 seconds) to reduce false detections.

[0039] S32, the waveform is divided by each respiratory wave peak point to distinguish between the exhalation and inhalation phases; The waveform is divided into an inspiratory phase (before the peak) and an expiratory phase (after the peak) using each respiratory wave peak as a dividing point. The inflection point is determined with the help of the zero-crossing point of the first derivative.

[0040] S33. Take the three most recent respiratory waveforms and calculate the average expiratory duration and average inspiratory duration as the basis for respiratory phase determination.

[0041] Take the three most recent respiratory waves and calculate the average expiratory duration ( T exh ) and average inhalation time ( T inh The phase determination logic is as follows: if the current time is within the exhalation period, it is marked as "exhalation phase"; otherwise, it is marked as "inhalation phase". This parameter provides input for S4.

[0042] S4, determine the auricular stimulation target area and stimulation parameters based on the respiratory phase lookup mapping table; In this embodiment, the mapping table is a pre-configured respiratory phase-auricular target area mapping table. For example, when the respiratory phase corresponds to the "shh" sound, the target area is the liver auricular point, and when the "ha" sound corresponds to the heart auricular point. The stimulation parameters include intensity (0.5-5 mA) and frequency (1-30 Hz), which are obtained directly from the table lookup.

[0043] The mapping table data structure uses a hash table or relational database to store the correspondence between respiratory phases and auricular target areas. The table structure contains the following fields: When the user pronounces the sound "shh," it corresponds to the liver aspect, stimulating the liver area ear acupoints; When the user pronounces the sound "ha", it corresponds to the heart phase and stimulates the acupoints in the heart area of ​​the ear; When a user pronounces the sounds "hu," "yan," "chui," and "xi," these sounds correspond to the lung, spleen, kidney, and triple burner, respectively, stimulating the corresponding ear acupoints.

[0044] Correspondingly, the respiratory phase coding is as follows: 1 represents the "shh" sound phase, 2 represents the "ha" sound phase, and so on. The system receives the respiratory phase judgment results from S3 in real time and calculates the optimal target area using the following formula:

[0045] S5, apply taVNS stimulation to the auricular stimulation target area; In this embodiment, the determined stimulation parameters are converted into physical stimulation signals. The stimulation signals are generated by a programmable waveform generator using biphasic rectangular pulses. Key parameters include pulse width (adjustable from 100-500 μs), phase interval (50-200 μs), and rise / fall time (≤10 μs). The pulse waveform is generated using mathematical formulas.

[0046] Precise control is achieved, where A is the amplitude, τ is the pulse width, δ is the phase interval, and u(t) is the unit step function. Dynamic positioning of the stimulation target is realized through a multi-electrode array (e.g., an 8×8 matrix) and an analog switch matrix, achieving a switching time of ≤10ms and a positioning accuracy of ≤1mm. The system monitors the electrode-skin contact impedance in real time (requiring <50kΩ) and uses a closed-loop control algorithm to ensure a stimulation dose error of <5%. The actual dose calculation formula is...

[0047] S6, the user performs the Six Healing Sounds breathing exercise according to the rhythm, and returns to step S1 to form a closed-loop control.

[0048] Users perform breathing exercises using the six-character mantra guided by the video tutorial provided on the host computer interface. The system uses an inertial measurement unit (IMU) to capture the characteristics of the vocalization movements in real time, including the amplitude of joint angle changes, synchronization error with the standard beat, and the continuity of the rate of change of acceleration. Breathing-movement coordination is expressed by a formula.

[0049] Quantization requires a coordination coefficient ≥ 0.7 to ensure training quality. Closed-loop control uses a multivariable PID controller, and its control input is calculated as follows:

[0050] Where e(t) represents the deviation between the setpoint and the actual value. The system also integrates reinforcement learning algorithms to optimize the control strategy, based on the Q-learning formula.

[0051] The system adaptively adjusts stimulation parameters. A safety monitoring system monitors abnormal signals such as heart rate arrhythmia in real time, providing multi-dimensional visual, auditory, and tactile feedback, and recording a complete operation log. After completing the current cycle, the system automatically returns to step S1 to begin a new round of data acquisition and processing, forming a continuously optimized closed-loop training process.

[0052] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0053] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0054] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0055] Specific embodiments have been used to illustrate the principles and implementation methods of this invention. The descriptions of the embodiments above are only for the purpose of helping to understand the method and core ideas of this invention. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this invention. Therefore, the content of this specification should not be construed as a limitation of this invention.

[0056] Those skilled in the art will recognize that the embodiments described herein are intended to help the reader understand the principles of the invention, and should be understood that the scope of protection of the invention is not limited to such specific statements and embodiments. Those skilled in the art can make various other specific modifications and combinations based on the technical teachings disclosed in this invention without departing from the spirit of the invention, and these modifications and combinations are still within the scope of protection of this invention.

Claims

1. A training method combining TaVNS with the Six Healing Sounds for slow breathing, characterized in that... Includes the following steps: S1, synchronously collects the user's electrocardiogram and electromyography signals; S2, Calculate respiratory depth and amplitude parameters based on the electrocardiogram and electromyography signals; S3, determine the current respiratory phase based on the respiratory depth and amplitude parameters; S4, determine the auricular stimulation target area and stimulation parameters based on the respiratory phase lookup mapping table; S5, apply taVNS stimulation to the auricular stimulation target area; S6, the user performs the Six Healing Sounds breathing exercise according to the rhythm, and returns to step S1 to form a closed-loop control.

2. The training method of taVNS combined with the Six Healing Sounds slow breathing according to claim 1, characterized in that: Step S1 includes: S11 receives data streams from ECG and EMG sensors in real time via Bluetooth interface; S12, the data stream is stored in the communication thread data pool of the host computer for analysis and processing.

3. The training method of taVNS combined with the Six Healing Sounds slow breathing according to claim 1, characterized in that: Step S2 includes an electrocardiogram (ECG) signal analysis sub-step and an electromyography (EMG) signal analysis sub-step, wherein, The electrocardiogram signal analysis sub-step includes: S21, the Pan-Tompkins algorithm is used to extract the peak point of the R wave in the electrocardiogram signal; S22, calculate the time interval between adjacent R-wave peaks, divide it by the sampling frequency to obtain the RR interval sequence, the formula is: In the formula, For the i-th RR interval, t i and t i+1 These are the peak times of adjacent R-waves. The sampling frequency; S23, calculate heart rate variability characteristics based on the RR interval sequence, including mean RR interval MeanRR, mean heart rate MeanHR, RR interval standard deviation SDNN, root mean square RMSSD of the difference between adjacent RR intervals, and the proportion of adjacent RR interval differences greater than 50ms pNN50 and the proportion of adjacent RR interval differences greater than 20ms pNN20. S24, Perform Fourier transform on the RR interval sequence to calculate spectral characteristics, including very low frequency power VLF, low frequency power LF, high frequency power HF, low frequency to high frequency power ratio LF / HF, and normalized low frequency power Normalized LF and normalized high frequency power Normalized HF. The electromyographic signal analysis sub-step includes: S25 performs bandpass filtering (20-450Hz), power frequency notch filtering, full-wave rectification, and low-pass filtering (2-5Hz) on the electromyography signal to obtain the envelope signal; S26, on the envelope signal, an adaptive threshold and minimum interval peak detection method is used to extract burst peaks, and the time interval sequence PI between adjacent burst peaks is calculated using the following formula: In the formula, This represents the interval of the j-th peak. S27, calculate rhythmic features based on the PI sequence, including average PIM, average activation rate, PI standard deviation SDPI, PI sequence RMSSD, and the proportion of PI difference greater than 50ms and greater than 20ms. S28, within the sliding window, statistical amplitude characteristics, including root mean square (RMS), mean absolute value (MAV), integrated electromyography (IEMG) value, and peak-to-valley difference; S29. Perform Fourier transform on the bandpass filtered electromyography signal to calculate the power spectrum characteristics, including median frequency (MNF), average power frequency (MPF), low-frequency band power, high-frequency band power, high-low frequency power ratio, and spectral centroid.

4. The training method for slow breathing combining TaVNS with the Six Healing Sounds according to claim 1, characterized in that, Step S3, determining the current respiratory phase, includes: S31, the local maximum method is used to extract the peak point of the respiratory wave; S32, the waveform is divided by each respiratory wave peak point to distinguish between the exhalation and inhalation phases; S33. Take the three most recent respiratory waveforms and calculate the average expiratory duration and average inspiratory duration as the basis for respiratory phase determination.

5. The training method for slow breathing combining TaVNS with the Six Healing Sounds according to claim 1, characterized in that, The mapping table mentioned in step S4 is a respiratory phase-auricular target area mapping table, which is configured as follows: When the user pronounces the sound "shh," it corresponds to the liver aspect, stimulating the liver area ear acupoints; When the user pronounces the sound "ha", it corresponds to the heart phase and stimulates the acupoints in the heart area of ​​the ear; When a user pronounces the sounds "hu," "yan," "chui," and "xi," they correspond to the lung, spleen, kidney, and triple burner, respectively, stimulating the corresponding ear acupoints.

6. The training method for slow breathing combining TaVNS with the Six Healing Sounds according to claim 1, characterized in that, Step S5, which involves applying taVNS stimulation, includes: S51, based on the stimulation parameters determined in step S4, a control command is sent to the otoelectric controller via the host computer; S52, the otoelectric controller adjusts the stimulation intensity, frequency, and target location according to the command; S53, stimulation is activated and deactivated in sync with the respiratory phase to maximize vagal nerve activation.

7. The training method for slow breathing combining TaVNS with the Six Healing Sounds according to claim 1, characterized in that, Step S6 includes: S61 plays a video tutorial of the Six Healing Sounds on the host computer interface, providing rhythm guidance; S62, the user performs the six-sound mantra movements and breathing according to the video, the six-sound mantra including the six sounds "xu, he, hu, yan, chui, xi"; S63 monitors user performance quality in real time and assesses training effectiveness through the electromyography-heart rate coupling index.

8. The training method for slow breathing combining taVNS with the Six Healing Sounds according to claim 7, characterized in that, It also includes step S7: S71, the treatment effect evaluation model is invoked at intervals of one minute, and the model is implemented using a support vector machine; S72, the input features of the model include heart rate variability features, electromyographic features, and progress of practicing the Six Healing Sounds. S73, the model outputs the probability of the optimal strategy for controlling the otoelectric effect, and the strategy with the highest probability is selected to dynamically adjust the stimulation parameters.

Images