Intelligent scoring method and system for meditation effect based on big data analysis

Abstract

The invention provides a meditation effect intelligent scoring method and system based on big data analysis, and relates to the technical field of data processing.The method comprises the steps that 1, brain wave signals, heart rate variability, skin conductance original waveforms and respiratory rhythm signals of a user during meditation are collected in real time through a wearable device, and the brain wave signals, the heart rate variability, the skin conductance original waveforms and the respiratory rhythm signals are sent to the wearable device; environment noise decibel values, illumination intensity and environment temperature parameters are obtained through an environment sensor, and a multi-dimensional original data set fusing the physiological signals and the environment parameters is generated; 2, according to the resting state physiological parameters in the multi-dimensional original data set, Gaussian mixture probability distribution is adopted, personalized reference features are generated, and the personalized reference features comprise mean vectors and covariance matrixes of brain wave signals, heart rate variability, skin conductance and respiratory rhythm parameters; according to the invention, through multi-source data acquisition and fusion, personalized reference construction, multi-dimensional physiological and environmental index analysis and dynamic score adjustment, quantitative evaluation of a meditation effect, which is comprehensive, accurate and real-time and fits individuals, is realized.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to a meditation effect intelligent scoring method and system based on big data analysis. Background Art

[0002] Traditional evaluation of meditation effectiveness relies primarily on subjective user descriptions or professional judgment. For example, after completing a meditation session, users may simply describe their experience as "feeling very relaxed" or "improved concentration," but these descriptions lack a specific measurement scale. Different users experience and express "relaxation" in vastly different ways. Some identify relaxation as steady breathing, while others consider inner peace and a clear mind. Professional judgments also vary widely. An experienced meditation instructor may determine a participant's meditation is effective based on observations of their posture and breathing rate, while another instructor, based on different criteria, may reach a different conclusion. This approach is subject to significant subjectivity and individual variability, lacking an objective, accurate, and quantitative evaluation standard. Without a standardized quantitative metric, it's difficult to scientifically and systematically analyze and compare the effects of meditation across users and over different time periods. It also makes it impossible to provide users with targeted improvement recommendations, such as "meditation is more effective when the proportion of alpha waves in brain waves reaches XX%."

[0003] While existing methods for evaluating meditation effectiveness attempt to utilize physiological signals for analysis, most focus solely on a single or a few physiological indicators, failing to fully reflect the comprehensive changes in the human body's physiological state during meditation. For example, some evaluation methods rely solely on heart rate variability to determine meditation effectiveness. If a user's heart rate decreases and heart rate variability increases during meditation, they conclude that the meditation effect is good. However, ignoring other important physiological signals such as brain waves and skin conductance can lead to misjudgments. If the user's brain waves, which are associated with relaxation, have a very low proportion of theta and alpha waves, they have not actually entered a deep meditation state.

[0004] Furthermore, some methods fail to account for individual physiological differences and employ uniform evaluation criteria, making it difficult to accurately assess the meditation effects of different users. For example, people of different ages and physical conditions naturally have varying baseline heart rates and brainwave characteristics. Using a uniform standard such as "a heart rate below 70 beats per minute indicates effective meditation" is clearly unfair to users with higher baseline heart rates, making it impossible to accurately assess their true meditation effects. Summary of the Invention

[0005] The technical problem to be solved by the present invention is to provide a meditation effect intelligent scoring method and system based on big data analysis, so as to achieve an objective, accurate and quantitative evaluation of the meditation effect and provide users with scientific meditation effect feedback.

[0006] In order to solve the above technical problems, the technical solutions of the present invention are as follows:

[0007] In a first aspect, a method for intelligently scoring meditation effects based on big data analysis is provided, the method comprising:

[0008] Step 1: Use wearable devices to collect real-time brainwave signals, heart rate variability, skin conductance raw waveforms, and respiratory rhythm signals during meditation. Use environmental sensors to obtain ambient noise decibel values, light intensity, and ambient temperature parameters to generate a multidimensional raw data set that integrates physiological signals and environmental parameters.

[0009] Step 2: Based on the resting-state physiological parameters in the multidimensional original dataset, a Gaussian mixture probability distribution is used to generate personalized baseline features, including the mean vector and covariance matrix of EEG signals, heart rate variability, skin conductance, and respiratory rhythm parameters;

[0010] Step 3: Decompose the EEG signal in the frequency domain to extract the energy proportions of theta and alpha waves, respectively. Calculate the chaos sensitivity index of the heart rate variability signal and the phase synchronization parameter of the original skin conductance waveform to form a real-time physiological feature vector.

[0011] Step 4: Calculate the degree of neural relaxation, autonomic nervous balance, and environmental interference based on the real-time physiological feature vector and environmental parameters. Couple the degree of neural relaxation, autonomic nervous balance, and environmental interference and convert the calculation results into a meditation effect score using a piecewise linear mapping algorithm.

[0012] Step 5: Compare the environmental interference level with a preset reference threshold. When the environmental interference level is greater than or equal to the reference threshold, activate the segmented attenuation mechanism.

[0013] Secondly, an intelligent meditation effect scoring system based on big data analysis includes:

[0014] The data acquisition module is used to collect brainwave signals, heart rate variability, skin conductance raw waveforms, and respiratory rhythm signals of users during meditation in real time through wearable devices. It also obtains ambient noise decibel values, light intensity, and ambient temperature parameters through environmental sensors to generate a multidimensional raw data set that integrates physiological signals and environmental parameters.

[0015] The benchmark generation module is used to generate personalized benchmark features based on the resting state physiological parameters in the multi-dimensional original data set using Gaussian mixture probability distribution, including the mean vector and covariance matrix of brain wave signals, heart rate variability, skin conductance, and respiratory rhythm parameters;

[0016] The feature extraction module is used to perform frequency domain decomposition on the EEG signal, extract the energy proportions of theta and alpha waves respectively, calculate the chaos sensitivity index of the heart rate variability signal and the phase synchronization parameter of the original skin conductance waveform, and form a real-time physiological feature vector;

[0017] The scoring calculation module is used to calculate the degree of neural relaxation, autonomic nervous balance and environmental interference based on the real-time physiological feature vector and environmental parameters, couple the degree of neural relaxation, autonomic nervous balance and environmental interference, and convert the calculation results into a meditation effect score through a piecewise linear mapping algorithm;

[0018] The attenuation processing module is used to compare the environmental interference level with a preset reference threshold. When the environmental interference level is greater than or equal to the reference threshold, the segmented attenuation mechanism is activated.

[0019] According to a third aspect, a computing device includes:

[0020] one or more processors;

[0021] The storage device is used to store one or more programs, and when the one or more programs are executed by the one or more processors, the one or more processors implement the method.

[0022] In a fourth aspect, a computer-readable storage medium stores a program, which implements the method when executed by a processor.

[0023] The above solution of the present invention includes at least the following beneficial effects:

[0024] Wearable devices collect multiple physiological signals, such as brainwaves and heart rate variability, while environmental sensors capture environmental parameters such as noise, light, and temperature, generating a multidimensional raw data set. This fusion of multi-source data comprehensively reflects the user's meditative state from both physiological and environmental perspectives, improving data integrity and accuracy compared to single data sources. Based on the user's resting physiological parameters, a Gaussian mixture probability distribution is used to generate personalized baseline features, accounting for individual physiological differences. Since different users have different baseline physiological states, establishing personalized baselines avoids assessment biases caused by the use of uniform standards and improves the accuracy and pertinence of assessments.

[0025] The brainwave signal is decomposed in the frequency domain to extract the energy proportions of theta and alpha waves. Combined with the chaos sensitivity index of the heart rate variability signal and the phase synchronization parameters of the raw skin conductance waveform, the system analyzes multiple physiological aspects, including brain activity, cardiovascular regulation, and nervous system response. This provides a deeper understanding of the physiological changes in the human body during meditation, making the assessment more comprehensive and professional. Environmental parameters are incorporated into the assessment system to calculate the degree of environmental interference, fully considering the impact of environmental factors on meditation effectiveness. Environmental factors such as noise, lighting, and temperature can directly affect the user's meditation experience and effectiveness. By quantifying environmental interference, the system can more realistically reflect the actual meditation state and provide users with more objective assessment results.

[0026] This generates a real-time physiological feature vector that reflects changes in the user's physiological state during meditation. The user's physiological state changes dynamically during meditation. Real-time monitoring and analysis capture these changes, providing immediate feedback to help users adjust their meditation state. Based on the real-time physiological feature vector and environmental parameters, neural relaxation, autonomic nervous balance, and environmental disturbance are dynamically calculated and coupled. Furthermore, when the environmental disturbance reaches a certain threshold, a segmented attenuation mechanism is activated, allowing the evaluation results to dynamically adjust as the user's physiological state and environment change, ensuring the timeliness and accuracy of the evaluation. The calculated results are converted into a meditation effect score using a piecewise linear mapping algorithm, transforming complex physiological and environmental factors into a specific numerical score, making the evaluation results more intuitive and easy to understand. This quantitative approach provides users with a clear standard to measure their own meditation effectiveness, facilitating self-assessment and comparison. The detailed evaluation dimensions and quantitative scores provide users with specific feedback, helping them understand their strengths and weaknesses during meditation. For example, users can adjust meditation methods and environment in a targeted manner based on the scores of indicators such as neural relaxation and autonomic nervous balance, thereby improving meditation effects and achieving personalized meditation training and improvement. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 This is a flow chart of a method for intelligently scoring meditation effects based on big data analysis, provided by an embodiment of the present invention.

[0028] Figure 2 This is a schematic diagram of an intelligent scoring system for meditation effects based on big data analysis provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0029] Exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although exemplary embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. Rather, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art.

[0030] like Figure 1 As shown, an embodiment of the present invention proposes a meditation effect intelligent scoring method based on big data analysis, the method comprising the following steps:

[0031] Step 1: Use wearable devices to collect real-time brainwave signals, heart rate variability, skin conductance raw waveforms, and respiratory rhythm signals during meditation. Use environmental sensors to obtain ambient noise decibel values, light intensity, and ambient temperature parameters to generate a multidimensional raw data set that integrates physiological signals and environmental parameters.

[0032] Step 2: Based on the resting-state physiological parameters in the multidimensional original dataset, a Gaussian mixture probability distribution is used to generate personalized baseline features, including the mean vector and covariance matrix of EEG signals, heart rate variability, skin conductance, and respiratory rhythm parameters;

[0033] Step 3: Decompose the EEG signal in the frequency domain to extract the energy proportions of theta and alpha waves, respectively. Calculate the chaos sensitivity index of the heart rate variability signal and the phase synchronization parameter of the original skin conductance waveform to form a real-time physiological feature vector.

[0034] Step 4: Calculate the degree of neural relaxation, autonomic nervous balance, and environmental interference based on the real-time physiological feature vector and environmental parameters. Couple the degree of neural relaxation, autonomic nervous balance, and environmental interference and convert the calculation results into a meditation effect score using a piecewise linear mapping algorithm.

[0035] Step 5: Compare the environmental interference level with a preset reference threshold. When the environmental interference level is greater than or equal to the reference threshold, activate the segmented attenuation mechanism.

[0036] In this embodiment of the present invention, a wearable device collects multi-dimensional physiological signals in real time, including brainwave signals, heart rate variability, raw skin conductance waveforms, and respiratory rhythm signals. Environmental sensors are also used to obtain ambient noise decibel levels, light intensity, and ambient temperature parameters to generate a multi-dimensional raw data set. This overcomes the limitations of traditional evaluation methods that rely on subjective perception or a single metric. Instead, it uses objective data to comprehensively reflect the physiological state of the human body during meditation and the influence of environmental factors. For example, brainwave signals can reflect brain activity, and raw skin conductance waveforms can reflect the level of activation of the autonomic nervous system. Using resting-state physiological parameters from the multi-dimensional raw data set, a Gaussian mixture probability distribution is used to generate personalized baseline features. This takes into account the differences in physiological characteristics of different users, eliminating the need for a uniform standard, making the assessment more tailored to individual circumstances. For example, users of different ages and physical conditions have different baseline physiological indicators. Personalized baseline features can accurately measure the degree to which each user's meditation effect improves relative to their own, improving the accuracy and reliability of the assessment.

[0037] By performing frequency domain decomposition on EEG signals, extracting the energy contributions of theta and alpha waves, and calculating the chaos sensitivity index of the heart rate variability signal and the phase synchronization parameters of the raw skin conductance waveform, the method explores physiological signal characteristics from multiple perspectives. Compared to methods that focus solely on a single or a few indicators, this method provides a deeper and more comprehensive understanding of changes in the human physiological state during meditation. For example, theta and alpha waves are closely associated with states such as relaxation and concentration, and analyzing their energy contributions can help determine the mental state during meditation. Neural relaxation, autonomic balance, and environmental disturbance are calculated separately and coupled together. These three factors are then converted into a meditation effectiveness score using a piecewise linear mapping algorithm. This method comprehensively considers physiological state and environmental factors to quantify meditation effectiveness, avoiding the one-sidedness of single-factor assessments. Furthermore, the piecewise linear mapping algorithm makes the scoring more realistic, providing users with intuitive and scientifically accurate quantification of meditation effectiveness.

[0038] The environmental interference level is compared to a preset baseline threshold. A segmented attenuation mechanism is activated when the environmental interference level exceeds the baseline threshold. Taking into account the impact of the environment on meditation effectiveness, the score is dynamically adjusted based on the degree of environmental interference, ensuring that the evaluation results more accurately reflect the actual meditation effect. For example, in a noisy environment, even if physiological state is good, the score will be correspondingly lowered due to the high environmental interference. This reminds users of the importance of environmental factors in meditation effectiveness and provides a reference for users to choose an appropriate meditation environment.

[0039] In a preferred embodiment of the present invention, step 1 above involves collecting brainwave signals, heart rate variability, skin conductance raw waveforms, and respiratory rhythm signals of the user during meditation in real time through a wearable device, and obtaining ambient noise decibel values, light intensity, and ambient temperature parameters through environmental sensors to generate a multidimensional raw data set that integrates physiological signals and environmental parameters. This may include:

[0040] In an embodiment of the present invention, a headband-type EEG monitoring device is selected. This type of device is equipped with multiple electrodes and can accurately collect EEG signals from the scalp surface. For example, the NeuroSky MindWave series uses dry electrode technology, is comfortable to wear, and is suitable for long-term EEG signal collection. Smart bracelets are selected, such as Fitbit bracelets on the market. They can monitor heart rate in real time through optical or electrode methods and calculate parameters related to heart rate variability. A skin conductance sensor is used, the principle of which is to reflect the skin conductance by measuring the change in resistance on the skin surface. Some professional biofeedback devices have integrated skin conductance sensors that can accurately collect the original waveform of skin conductance. A chest-strap respiratory sensor can obtain respiratory rhythm signals by detecting the rise and fall of the chest, while a nasal airflow sensor collects signals by monitoring the changes in the inflow and outflow of nasal airflow.

[0041] Noise decibel sensors, such as B&K's, can accurately measure ambient noise decibels. Light intensity sensors, such as the BH1750, convert light intensity into an electrical signal, enabling real-time monitoring of ambient light intensity. The DS18B20 ambient temperature sensor is a digital temperature sensor with high precision and strong anti-interference capabilities, capable of accurately measuring ambient temperature. Connect the EEG acquisition device, heart rate variability acquisition device, skin conductance acquisition device, and respiratory rhythm acquisition device to a data acquisition terminal (such as a smartphone, tablet, or dedicated data acquisition device). Connection methods can be Bluetooth or Wi-Fi, depending on device support. For example, the NeuroSky MindWave EEG cap can be paired with a smartphone via Bluetooth.

[0042] Configure the wearable device parameters, such as sampling frequency and gain, to ensure the collected signal quality meets requirements. Parameter setting methods may vary depending on the device, so refer to the device's manual for details. Connect the noise decibel sensor, light intensity sensor, and ambient temperature sensor to the data acquisition terminal, either using Bluetooth or Wi-Fi. Calibrate and configure the environmental sensors to ensure measurement accuracy. For example, the noise decibel sensor requires calibration before use to eliminate measurement errors. Before the user begins meditation, launch the data acquisition program. This data acquisition program can be dedicated software installed on the data acquisition terminal. The program simultaneously collects EEG signals, heart rate variability, skin conductance raw waveforms, and respiratory rhythm signals at the pre-set sampling frequency, as well as ambient noise decibel levels, light intensity, and ambient temperature parameters. The collected physiological signals and environmental parameter data are transmitted to the data acquisition terminal in real time via a connection. For example, when transmitting via Bluetooth, the device sends the collected data to the terminal in a specific Bluetooth protocol format. After receiving the data, the data acquisition terminal will perform preliminary processing and storage on the data, such as arranging the data in time series and storing it as a file in a specific format (such as CSV, JSON, etc.).

[0043] On the data acquisition terminal, the collected physiological signals and environmental parameter data are integrated. For example, EEG signals, heart rate variability, raw skin conductance waveforms, respiratory rhythm signals, and ambient noise decibel levels, light intensity, and ambient temperature parameters at the same time point are combined to form a single data record. As data is continuously collected and integrated, all data records are aggregated to generate a multidimensional raw data set that integrates physiological signals and environmental parameters.

[0044] In a preferred embodiment of the present invention, the above step 2, based on the resting state physiological parameters in the multidimensional original data set, uses Gaussian mixture probability distribution to generate personalized baseline features, including mean vectors and covariance matrices of brain wave signals, heart rate variability, skin conductance, and respiratory rhythm parameters, which may include:

[0045] Step 200 , determining a time interval during which the user is in a resting state from the multidimensional raw data set, and extracting physiological parameter data from the resting time interval, including brain wave signals, heart rate variability, skin conductance, and respiratory rhythm parameters;

[0046] Step 201: For each physiological parameter, a Gaussian mixture model is constructed to calculate the mean vector and covariance matrix of each physiological parameter;

[0047] Step 202 : Combining the mean vector and covariance matrix of each physiological parameter to form a personalized baseline feature.

[0048] In an embodiment of the present invention, a multidimensional raw data set is preliminarily processed, including removing outliers and noise data. For example, for EEG signals, if a sudden spike or a value outside the normal range appears, it can be identified as an outlier and removed using statistical methods (such as the 3σ principle); for heart rate variability data, noise can be removed using methods such as sliding average filtering. The user's behavior records or additional sensor information (such as button operation records on the device, user manual markings, etc.) are combined to assist in determining the resting state. Simultaneously, the changing trends of physiological parameters are analyzed. For example, when heart rate variability tends to be stable, the respiratory rhythm is gentle, and the EEG signal fluctuates slightly, it can be preliminarily considered that the user is in a resting state. Using a rule-based algorithm, such as setting a fluctuation range threshold for heart rate variability and a frequency threshold for the respiratory rhythm, when the physiological parameters continuously meet these threshold conditions for a period of time, the time period is determined to be a resting state time interval. Within the determined resting state time interval, EEG signals, heart rate variability, skin conductance, and respiratory rhythm parameters are extracted from the multidimensional raw data set. For brain wave signals, the numerical value of each sampling point is extracted in chronological order; for heart rate variability, relevant data such as the interval between adjacent heartbeats are extracted; for skin conductance, its original waveform data is extracted; for respiratory rhythm, parameter data such as respiratory frequency and respiratory depth are extracted.

[0049] Step 201: After acquiring physiological parameter data such as brain wave signals, heart rate variability, skin conductance, and respiratory rhythm, since the numerical ranges of different physiological parameters vary significantly (for example, the voltage amplitude of brain wave signals may be in the microvolt level, while heart rate variability data is measured in milliseconds), standardization is required to eliminate dimensionality effects and unify the data scale. First, the mean value of each type of physiological parameter data is calculated. Taking heart rate variability data as an example, the mean of the heart rate variability data set is obtained by adding the values of all adjacent heartbeat intervals in the data set and dividing it by the total number of data points. Next, the difference between each data point and the mean is calculated, these differences are squared, summed, and divided by the total number of data points. Finally, the square root of the result is taken to obtain the standard deviation of the data set. Once the mean and standard deviation are obtained, the mean is subtracted from each original data point and then divided by the standard deviation to complete the standardization. After this processing, the distribution of each type of physiological parameter data becomes a standard normal distribution with a mean of 0 and a standard deviation of 1.

[0050] To find a Gaussian mixture model that describes the distribution of physiological parameter data, it is necessary to determine the number of Gaussian components. First, set a reasonable range of component numbers to try, for example, starting with one component and gradually increasing to ten. For each set number of components, a corresponding Gaussian mixture model is constructed to fit the standardized physiological parameter data. During the fitting process, the Akaike Information Criterion (AIC) is used to evaluate the quality of the model. The AIC comprehensively considers the model's fit to the data and its complexity. Specifically, the better the model fits the data and the fewer parameters it contains, the lower the AIC value. When trying models with different numbers of components, the AIC value corresponding to each model is calculated. For example, when trying to build a model with three Gaussian components, one AIC value is calculated based on the model's fit to the data; when building a model with four Gaussian components, another AIC value is obtained. Ultimately, the number of components corresponding to the model with the lowest AIC value is selected as the final Gaussian mixture model that describes the distribution of the physiological parameter data.

[0051] After determining the number of components in the Gaussian mixture model, the model is trained using the EM algorithm. During the initial training phase, the model parameters are randomly initialized, including the mean vector, covariance matrix, and component weights of each Gaussian component. Based on the model parameters currently initialized or obtained from the previous iteration, each data point is analyzed. For example, for a data point in the skin conductance data, the probability of that data point belonging to each Gaussian component is calculated based on the parameters of each Gaussian component (mean vector, covariance matrix, and component weight). This process is like calculating the "likelihood" of each data point belonging to each Gaussian component. Based on the probability of each data point belonging to each Gaussian component calculated in the E-step, the model parameters are re-estimated. Specifically, the weights of each Gaussian component are updated to better reflect the actual distribution of data points within that component; the mean vector of each Gaussian component is recalculated to better represent the central tendency of the data points within that component; and the covariance matrix is updated to better describe the dispersion and interrelationships of data points within that component. The E-step and M-step are repeated repeatedly, with each iteration bringing the model parameters closer to the true distribution characteristics of the data. When the preset maximum number of iterations is reached, the model parameters are considered converged, and the training process ends. After model training is complete, the mean vector and covariance matrix corresponding to each Gaussian component are obtained. If the model contains multiple Gaussian components, since each component has different importance in describing the data distribution (reflected by the component weight), a weighted average of the mean vectors and covariance matrices of each component is required based on the component weight.

[0052] In specific operations, the mean vector of each component is multiplied by its corresponding weight, and then these products are added together to obtain the final mean vector; a similar weighted calculation is also performed on the covariance matrix, and the difference between the component mean and the overall mean is taken into account, and finally a mean vector and covariance matrix that can fully reflect the distribution characteristics of the physiological parameter data are obtained.

[0053] In step 202, the mean vectors and covariance matrices of the EEG signals, heart rate variability, skin conductance, and respiratory rhythm parameters calculated in step 201 are organized to ensure consistent data format and dimensions. These mean vectors and covariance matrices are combined in a specific order, for example, by sequentially arranging all mean vectors to form a large vector and sequentially arranging all covariance matrices to form a matrix block. This ultimately forms a personalized baseline feature vector containing information about multiple physiological parameters.

[0054] Assume that before the user starts meditation training, he wears the device to collect physiological data and environmental data for 30 minutes to form a multidimensional original data set. Through analysis, it is found that the user is in a resting state for 5-10 minutes after the start of data collection (the user is sitting and resting at this time, without obvious physical activity and emotional fluctuations). The data from the 5th to 10th minute is extracted from the multidimensional original data set to obtain the brain wave signal in the resting state (such as collecting once per second, a total of 300 data points), heart rate variability data (adjacent heartbeat interval time data), skin conductance data (voltage value data of the original waveform) and respiratory rhythm data (respiratory frequency and depth data). Taking the heart rate variability data as an example, after standardization, the AIC criterion is used to determine that the Gaussian mixture model contains 2 components. The model is trained using the EM algorithm. After multiple iterations, the mean vectors of the two Gaussian components are [80,85] (assuming that they represent the average levels of two different heart rate variability states), and the covariance matrices are and After weighted averaging based on the component weights, the final mean vector and covariance matrix of heart rate variability are obtained. The same process is performed for EEG signals, skin conductance, and respiratory rhythm parameters. The mean vectors and covariance matrices of EEG signals, heart rate variability, skin conductance, and respiratory rhythm parameters are combined to form a personalized baseline feature that incorporates this information.

[0055] This method generates a personalized baseline signature based on the user's resting-state physiological parameters, fully accounting for individual differences. Because different users have varying physiological characteristics, traditional, unified benchmarks cannot accurately reflect their true individual conditions. This method, however, allows for a customized baseline for each user, making meditation effectiveness assessment more personalized and effective, improving both accuracy and effectiveness. The baseline signature is constructed based on resting-state data, where physiological parameters are relatively stable and less susceptible to external interference and temporary factors. Using this baseline for comparative analysis can reduce errors caused by data fluctuations and improve the stability and reliability of the baseline signature. By encompassing multiple physiological parameters, including EEG signals, heart rate variability, skin conductance, and respiratory rhythm, the constructed baseline signature comprehensively reflects the user's physiological state from multiple dimensions, providing rich information. Compared to benchmarks based on a single or limited set of physiological indicators, this method provides a more comprehensive assessment of physiological changes during meditation, providing more comprehensive data support for a deeper understanding of the effects of meditation on the body. Resting-state physiological parameters can change with changes in a user's physical condition, lifestyle, and other factors. By regularly updating the multidimensional raw dataset and regenerating personalized benchmark features, the evaluation benchmark can adapt to these changes, ensuring that the meditation effect evaluation is always timely and accurate, and continuously providing users with valuable feedback and guidance.

[0056] In a preferred embodiment of the present invention, the above step 3 of performing frequency domain decomposition on the EEG signal, extracting the energy proportions of theta waves and alpha waves respectively, and calculating the chaos sensitivity index of the heart rate variability signal and the phase synchronization parameter of the original skin conductance waveform to form a real-time physiological feature vector may include:

[0057] Step 300: converting the brainwave signal from the time domain to the frequency domain, and dividing the corresponding frequency bands in the frequency domain according to the frequency range of theta waves and alpha waves;

[0058] Step 301: Calculate the energy in each frequency band and obtain the frequency band energy ratio of theta waves and alpha waves through normalization processing to establish the brain wave characteristic baseline of the user in the meditation state;

[0059] Step 302: After establishing a baseline of the user's brainwave characteristics in the meditation state, obtain the current heart rate variability signal and use a dynamic time warping algorithm to compare the current heart rate variability signal with the personalized baseline characteristics to obtain a comparison result;

[0060] Step 303: Calculate the chaos sensitivity index of the heart rate variability signal based on the comparison result to reflect the degree of deviation between the current heart rate fluctuation and the personalized benchmark;

[0061] Step 304: After obtaining the EEG and HRV correlation features, the original skin conductance waveform and respiratory rhythm signal are obtained, and a phase lock value algorithm is used to analyze the phase synchronization between the original skin conductance waveform and the respiratory rhythm signal to determine the coordinated variation characteristics of the skin conductance signal and the respiratory rhythm.

[0062] Step 305 : The frequency band energy proportions of theta waves and alpha waves, the chaos sensitivity index of the heart rate variability signal, the original skin conductance waveform, and the respiratory rhythm signal are combined to form a real-time physiological feature vector including energy proportion, chaos sensitivity, and phase synchronization.

[0063] In the embodiments of the present invention, the EEG signal recorded in the time domain is a voltage value that varies over time, making it difficult to directly extract the characteristics of different frequency components from it. To better analyze the theta and alpha waves, it is necessary to convert them from the time domain to the frequency domain. This is accomplished with the help of algorithms such as the Fast Fourier Transform (FFT). Through these algorithms, the EEG signal, originally recorded in the time domain with time as the independent variable, can be converted into a signal with frequency as the independent variable, so that the distribution of different frequency components can be clearly seen. After obtaining the frequency domain signal, the corresponding frequency bands are accurately divided in the frequency domain signal based on the established frequency ranges of theta waves (frequency range of 4-7Hz) and alpha waves (frequency range of 8-13Hz). Just like in a spectrum graph, the areas where theta waves and alpha waves are located are marked according to the frequency range. The energy within each frequency band is calculated, and through normalization processing, the frequency band energy ratio of theta waves and alpha waves is obtained to establish a baseline for the EEG characteristics of the user in the meditative state.

[0064] Step 301, within the divided θ wave and α wave frequency bands, calculate the energy of the signal within each frequency band. Signal energy reflects the intensity of the frequency component in the entire brain wave signal. Energy can be calculated by, for example, squared summing the amplitudes of the signals within the frequency band. In order to eliminate the differences in brain wave signal intensity between individuals and the influence of changes in the overall intensity of the brain wave signal of the same user at different times, the calculated θ wave and α wave frequency band energies need to be normalized. Normalization is to divide the energy of each frequency band by the total energy of the entire brain wave signal to obtain the frequency band energy ratio of the θ wave and α wave. By performing such calculations and processing multiple times in a meditative state, a baseline of the user's brain wave characteristics in a meditative state can be established. This baseline can reflect the typical situation of the energy ratio of the θ wave and α wave when the user is meditating.

[0065] In step 302, a suitable heart rate monitoring device, such as a smart bracelet or a heart rate belt, is used to collect the user's current heart rate variability signal in real time. Heart rate variability refers to the change in the time interval between adjacent heartbeats, which reflects the regulatory function of the autonomic nervous system on the heart. The personalized baseline feature is generated in step 2 based on the user's resting physiological parameters, which contains relevant information about heart rate variability. The dynamic time warping (DTW) algorithm can handle the stretching and distortion problems of two time series signals on the time axis, and can find the final matching path between the two signals. The currently collected heart rate variability signal is dynamically time warped and compared with the heart rate variability data in the personalized baseline feature, and the comparison result is obtained by calculating indicators such as the distance between the two signals under the best matching path.

[0066] In step 303, based on the comparison results obtained in step 302, a specific algorithm is used to calculate the chaos sensitivity index of the heart rate variability signal. The chaos sensitivity index measures the complexity and irregularity of the heart rate variability signal, as well as the degree of difference between the current heart rate fluctuation and the personalized baseline. If the comparison results show that the current heart rate variability signal differs significantly from the baseline signal, the chaos sensitivity index value will also be correspondingly large, indicating that the current heart rate fluctuation has deviated from the personalized baseline.

[0067] In step 304, a skin conductance sensor and a respiratory rhythm sensor are used to obtain a raw skin conductance waveform and a respiratory rhythm signal, respectively. Skin conductance reflects changes in sweat gland activity on the skin surface and is closely related to the activity of the autonomic nervous system; respiratory rhythm reflects changes in the frequency and depth of breathing. A phase lock value (PLV) algorithm is used to analyze the raw skin conductance waveform and the respiratory rhythm signal. The phase lock value algorithm can measure the degree of phase synchronization between the two signals. By calculating the phase lock value of the skin conductance signal and the respiratory rhythm signal, the phase synchronization between them can be determined. If the phase lock value is high, it indicates that the skin conductance signal and the respiratory rhythm signal have strong phase synchronization, that is, there is a characteristic of coordinated change between the two; otherwise, it indicates that the synchronization between the two is weak.

[0068] In step 305, the frequency band energy percentages of theta and alpha waves obtained in step 301, the chaos sensitivity index of the heart rate variability signal obtained in step 303, and the phase synchronization parameters of the original skin conductance waveform and respiratory rhythm signal obtained in step 304 are combined in a specific order to form a vector. This vector contains multiple physiological characteristic information and can comprehensively reflect the user's current physiological state.

[0069] Suppose a user is meditating, using devices such as an EEG cap, a smart wristband, and a skin conductance sensor to collect real-time physiological signals. After collecting the EEG signals, they are converted to the frequency domain using a fast Fourier transform (FFT), which delineates the frequency bands of theta waves (4-7 Hz) and alpha waves (8-13 Hz). The energy within the theta band is calculated to be 10 units, the energy within the alpha band to be 20 units, and the total energy of the entire EEG signal to be 50 units. After normalization, the energy proportion of the theta band is 10 ÷ 50 = 0.2, and the energy proportion of the alpha band is 20 ÷ 50 = 0.4. This establishes a baseline EEG signature for the user's meditation state. Simultaneously, the smart wristband collects the current heart rate variability signal. Using a dynamic time warping algorithm, this signal is compared with the heart rate variability data in the personalized baseline signature. It is found that the current heart rate variability signal and the baseline signal exhibit some time-scale and distortion, resulting in a significant matching distance between the two. Based on this comparison result, the chaos sensitivity index of the heart rate variability signal is calculated to be 0.8, indicating that the current heart rate fluctuation deviates significantly from the personalized benchmark.

[0070] The skin conductance sensor and respiration sensor collected raw skin conductance waveforms and respiratory rhythm signals, respectively. Using a phase lock value algorithm, analysis revealed a phase lock value of 0.6, indicating a certain degree of phase synchronization between the skin conductance and respiratory rhythm signals, demonstrating a coordinated variation. Finally, the θ wave band energy percentage of 0.2, the α wave band energy percentage of 0.4, the chaos sensitivity index of the heart rate variability signal of 0.8, and the phase synchronization parameter of 0.6 for skin conductance and respiratory rhythm were combined to form a real-time physiological feature vector [0.2, 0.4, 0.8, 0.6].

[0071] By integrating multiple physiological indicators, including brain waves, heart rate variability, skin conductance, and respiratory rhythm, the system comprehensively reflects the user's physiological state during meditation from different perspectives. While a single physiological indicator may only reflect a specific aspect, this multi-indicator approach can more accurately capture the user's overall physiological changes, providing a richer and more comprehensive basis for evaluating meditation effectiveness. When calculating the chaos sensitivity index for the heart rate variability signal, the current signal is compared with a personalized baseline feature, fully accounting for individual physiological differences. Different users have different physiological characteristics and baseline conditions. Using a personalized baseline allows for a more precise assessment of each user's heart rate fluctuations during meditation, improving the accuracy and specificity of the assessment. By analyzing the phase synchronization between the raw skin conductance waveform and the respiratory rhythm signal, it is possible to capture the synergistic changes between different physiological signals. The autonomic nervous system regulates multiple physiological systems, and different physiological signals may be intrinsically linked. Understanding this synergistic relationship helps us gain a deeper understanding of the comprehensive impact of meditation on the human physiological system, providing valuable information for further optimizing meditation training programs. The resulting real-time physiological feature vector can reflect the changes in the user's physiological state during meditation in real time. This allows for timely detection of abnormalities or changing trends in the user's physiological state during meditation, providing real-time feedback and guidance. For example, if the Chaos Sensitivity Index of the heart rate variability signal suddenly increases, it may indicate that the user's current meditation state is being disturbed and requires timely adjustment.

[0072] In a preferred embodiment of the present invention, after establishing the baseline of the user's brainwave characteristics in the meditation state in step 302, obtaining the current heart rate variability signal, and using the dynamic time warping algorithm to compare the current heart rate variability signal with the personalized baseline characteristics to obtain the comparison result may include:

[0073] Step 3020: After establishing the brainwave characteristic baseline, the user's current heart rate data during meditation is obtained in real time through a wearable heart rate monitoring device;

[0074] Step 3021: Process the current heart rate data, calculate the heart rate variability index, and obtain the current heart rate variability signal;

[0075] Step 3022: Use the dynamic time warping algorithm to compare the current heart rate variability signal with the heart rate variability benchmark data in the personalized benchmark feature. During the comparison process, calculate the distance between each matching point.

[0076] Step 3023: Determine the final matching path between the two signals based on the distance between each matching point, and accumulate the distance between each matching point to obtain a total comparison distance, that is, a comparison result.

[0077] In an embodiment of the present invention, after the establishment of the baseline of the user's brainwave characteristics in the meditation state is completed, the real-time monitoring function of the wearable heart rate monitoring device is turned on. Wearable heart rate monitoring devices such as smart bracelets, heart rate monitoring chest straps, etc., these devices contact the human skin through photoelectric sensors or electrodes, and continuously collect physiological signals generated by the heartbeat. Taking the smart bracelet as an example, its built-in photoelectric sensor will emit green light to illuminate the skin, and according to the changes in the degree of light absorption by the blood, it will capture the periodic changes in blood vessel volume when the heart beats, and then convert them into electrical signals. The device will perform preliminary processing on these electrical signals, and display and transmit the current heart rate data in real time in the form of digital or waveforms, to ensure that every heartbeat data can be recorded in a timely and accurate manner during the user's meditation process.

[0078] In step 3021, the acquired raw heart rate data often contains noise or outliers, so it is first preprocessed. Obvious erroneous data due to motion artifacts, poor device contact, and other factors are removed, such as sudden high or low heart rate values within a short period of time. Next, a filtering algorithm is used to smooth the data, removing high-frequency noise interference and making the heart rate data more stable and reliable. After preprocessing is complete, heart rate variability (HRV) metrics are calculated. HRV refers to the subtle differences between successive heartbeat cycles and is measured by calculating the variation in the intervals between adjacent heartbeats (RR intervals). During the calculation process, the time points corresponding to each heartbeat are accurately extracted, and the difference between two adjacent heartbeat time points is calculated to obtain a series of RR interval data. Based on this RR interval data, statistical methods are used to calculate various HRV metrics, such as the root mean square difference (RMSSD) of adjacent RR intervals and the standard deviation (SDNN) of RR intervals. These metrics reflect the characteristics of HRV from different perspectives. Combining these multiple metrics creates a HRV signal that comprehensively describes the current heart rate fluctuations.

[0079] In step 3022, the dynamic time warping (DTW) algorithm is used to solve the similarity measurement problem when two time series signals are not completely aligned on the time axis. The current heart rate variability signal and the heart rate variability benchmark data in the personalized benchmark feature are regarded as two time series. At the beginning of the comparison, the starting points of the two signals are used as the initial matching points, and the distance between the two initial points is calculated. The distance calculation can use methods such as Euclidean distance, that is, calculating the square root of the sum of the squares of the differences between the corresponding indicator values of the two points. Then, based on the current matching point, according to certain rules (such as allowing one unit of movement up, down, left, and right on the time axis), the next possible matching point is explored, and the distance between the new matching points is also calculated. During the exploration process, multiple possible matching paths will be formed. This process is repeated continuously, and the distance between each set of possible matching points is calculated until the end points of the two signals are matched, thereby obtaining the distance information between all possible matching points.

[0080] Step 3023: Based on the distances between all matching points calculated in step 3022, find a matching path that minimizes the total distance. This path is the final matching path between the two heart rate variability signals. During the specific search process, a dynamic programming method can be used to backtrack from the back to the front, and gradually determine the optimal matching path based on the minimum cumulative distance from each point to the end point. After determining the final matching path, the distances between all matching points on the path are accumulated, and the resulting sum is the total comparison distance between the two heart rate variability signals. This comparison distance intuitively reflects the degree of difference between the current heart rate variability signal and the heart rate variability baseline data in the personalized benchmark feature. The smaller the comparison distance, the more similar the current heart rate variability signal is to the baseline data; the larger the comparison distance, the greater the difference between the two.

[0081] Suppose a user meditates while wearing a smart bracelet. After establishing a baseline for EEG characteristics, the bracelet begins collecting real-time heart rate data. During a one-minute meditation session, the bracelet records heart rate data at second intervals, resulting in 60 heart rate values, such as [72, 73, 72, 74, 73, ...]. These raw heart rate data are processed by first removing an outlier value (85) due to arm shaking, and then using a low-pass filter to remove high-frequency noise. Next, the intervals between adjacent heartbeats are calculated. For example, the RR intervals are [800ms, 820ms, 810ms, 790ms, ...]. Heart rate variability metrics such as RMSSD (15ms) and SDNN (20ms) are calculated, forming the current heart rate variability signal [15, 20, ...]. The user's personalized baseline heart rate variability data, including the baseline data set [12, 18, ...], are compared using the dynamic time warping algorithm. During the comparison process, the distance between each set of possible matching points is calculated. For example, the Euclidean distance of the first set of matching points (15,12) is After continuous exploration and calculation, the final matching path is determined, and the distances of all matching points on the path are accumulated, resulting in a total comparison distance of 12. This indicates that the current user's heart rate variability signal during meditation is somewhat different from the personalized benchmark.

[0082] By comparing the current heart rate variability signal with personalized baseline characteristics, the system fully accounts for differences in resting state and physiological regulation among different users. Everyone's baseline heart rate and autonomic nervous system regulation function vary. Using a personalized baseline allows for a more accurate assessment of each user's heart rate fluctuations during meditation, avoiding the errors associated with using a uniform standard and ensuring that the assessment results are more accurate for each individual's actual state. Real-time acquisition and analysis of heart rate variability signals during meditation allows for timely capture of dynamic changes in heart rate. During meditation, the user's psychological state and environmental factors can all affect heart rate. This dynamic comparative analysis can identify trends and abnormal fluctuations in heart rate, providing timely feedback to help users adjust their meditation state or method, ultimately achieving optimal meditation results. Heart rate variability is a key indicator of autonomic nervous system function and is closely related to the body's level of relaxation and psychological state during meditation. Comparing the heart rate variability signal with baseline data, combined with other physiological indicators such as EEG, allows for a comprehensive and scientific assessment of meditation effectiveness from multiple dimensions, providing a deeper understanding of the physiological impact of meditation.

[0083] In a preferred embodiment of the present invention, after obtaining the characteristics related to brain waves and heart rate variability in step 304, obtaining the original skin conductance waveform and the respiratory rhythm signal, analyzing the phase synchronization between the original skin conductance waveform and the respiratory rhythm signal using a phase lock value algorithm, and determining the coordinated change characteristics of the skin conductance signal and the respiratory rhythm, may include:

[0084] Step 3040: Convert the skin conductance signal and the respiratory rhythm signal from the time domain features to the phase domain, obtain the skin conductance signal and the respiratory rhythm signal, and calculate the phase difference between the skin conductance signal and the respiratory rhythm signal at the same time;

[0085] Step 3041: Set a time window based on the phase difference at each moment, perform statistical analysis on the phase difference within the corresponding time period in each time window, and convert the phase difference at each moment into a complex number.

[0086] Step 3042: Fusing the complex phase difference at each moment in the time window to obtain a phase lock value in the corresponding time window.

[0087] Step 3043, summarizing the phase locking values of each time window to obtain a summary result;

[0088] Step 3044: Determine the coordinated variation characteristics between the skin conductance signal and the respiratory rhythm signal based on the summary results and the preset evaluation criteria.

[0089] In this embodiment of the present invention, after acquiring EEG and HRV-related features, a skin conductance sensor and a respiratory rhythm sensor (such as a chest-strap respiratory sensor or a nasal airflow sensor) are used to collect raw skin conductance waveforms and respiratory rhythm signals in real time. Because the raw signals are time-domain signals, they need to be converted to the phase domain to analyze the phase relationship between the signals.

[0090] During the specific conversion process, mathematical transformation methods (such as the Hilbert transform) can be used. This transformation can convert the original time domain signal into a form containing phase information, thereby obtaining the instantaneous phase of the skin conductance signal and the respiratory rhythm signal at each moment. After obtaining the instantaneous phase, the phase difference between the two signals at the same moment is calculated. For example, the instantaneous phase angle of the skin conductance signal at a certain moment is first determined, and then the instantaneous phase angle of the respiratory rhythm signal at the same moment is determined. The phase difference at that moment is obtained by subtracting the two. By performing such calculations for each moment one by one, a series of values reflecting the phase difference between the two signals at different moments can be obtained.

[0091] In step 3041, an appropriate time window length is set, such as 10 seconds or 20 seconds, based on the signal characteristics and analysis requirements. Within each set time window, statistical analysis is performed on the phase difference data within that time period. The phase difference at each moment is converted into a complex number. This complex number representation more comprehensively preserves the phase difference information. This conversion process, based on the complex number representation principle of trigonometric functions, converts the phase difference angle into a complex number with real and imaginary parts, making the phase difference information at each moment more mathematically complete and easier to calculate.

[0092] Step 3042: Perform a fusion operation on the complex phase differences at all moments within a time window. During the fusion process, these complex phase differences are first summed and then their average value is calculated. Finally, the modulus of the average value (i.e., the absolute value of the complex number) is taken to obtain the phase lock value within the time window. The phase lock value is a value between 0 and 1 that reflects the degree of phase synchronization between the skin conductance signal and the respiratory rhythm signal within this time window. The closer the value is to 1, the better the phase synchronization of the two signals within the time window; the closer the value is to 0, the worse the synchronization.

[0093] In step 3043, over time, multiple time windows and their corresponding phase lock values are generated. By summarizing the phase lock values for these different time windows, the average, median, or other statistical measure of the phase lock values across all time windows can be calculated to comprehensively reflect the phase synchronization between the skin conductance signal and the respiratory rhythm signal over the entire time period. This summary can provide a macroscopic view of the changing trend of the phase relationship between the two signals over a longer period of time.

[0094] Step 3044 pre-determines evaluation criteria based on practical experience. For example, when the average value in the summary results is ≥ 0.8, the skin conductance signal and the respiratory rhythm signal are judged to exhibit highly synchronized, synergistic variation, indicating a good coordination between the body's autonomic nervous system and respiratory system during meditation. If the average value is between 0.5 and 0.8, the two signals are considered to have a certain degree of synchronization, with a significant synergistic variation. When the average value is between 0.2 and 0.5, the synchronization is poor, with no significant synergistic variation. If the average value is < 0.2, the two signals are almost out of synchronization, with a very poor synergistic variation. By comparing the summary results with the evaluation criteria, the specific synergistic variation characteristics between the skin conductance signal and the respiratory rhythm signal are ultimately determined.

[0095] Suppose that in a meditation process monitoring, data is collected in a 15-second time window. In the first 15-second time window, the skin conductance signal and respiratory rhythm signal are collected by the sensor, and converted into the phase domain, and the phase difference at each moment is calculated, such as the phase difference of 0.2 radians in the first second, 0.3 radians in the second second, and so on, and a total of 15 phase difference data are obtained. After converting these phase differences into complex form, they are summed, averaged, and modulo, and the phase lock value of the time window is 0.7. Continue monitoring, and in the second 15-second time window, the phase lock value is also calculated to be 0.6, and the third time window is 0.8. The phase lock values of these three time windows (0.7, 0.6, 0.8) are summarized and the average value is calculated as According to the preset evaluation criteria, 0.7 is between 0.5 and 0.8, which determines that during this meditation process, the skin conductance signal and the respiratory rhythm signal have a certain degree of synchronization, and the coordinated change characteristics are more obvious.

[0096] By analyzing the phase synchronization between skin conductance and respiratory rhythm signals, we can uncover the inherent connections between different physiological signals regulated by the autonomic nervous system. Skin conductance reflects sweat gland activity and sympathetic nervous system excitability, while respiratory rhythm is related to respiratory regulation by the respiratory center and the autonomic nervous system. Identifying the synergistic changes between these two signals helps understand the interplay between physiological systems during meditation. This synergistic change provides a new dimension for evaluating meditation effectiveness. Good synergistic changes between skin conductance and respiratory rhythm often indicate a state of relaxation and focus; conversely, it may indicate disruptions in meditation or poor results. Compared to single-signal analysis, this multi-signal phase synchronization analysis can more accurately assess a user's meditation state. Based on the identified synergistic changes, personalized feedback and guidance can be provided to the user. If the synergistic changes are poor, the user can be guided to adjust their breathing rhythm, relax their body and mind, and improve the coordination of the autonomic nervous system. If the synergistic changes are good, the user can be encouraged to maintain their current state, helping them practice meditation more effectively and achieve better results.

[0097] In a preferred embodiment of the present invention, the above step 4, based on the real-time physiological feature vector and environmental parameters, respectively calculates the degree of neural relaxation, the degree of autonomic nervous balance and the degree of environmental interference, and couples the degree of neural relaxation, the degree of autonomic nervous balance and the degree of environmental interference, and converts the calculation results into a meditation effect score value through a piecewise linear mapping algorithm, which may include:

[0098] Step 400: Dynamically adjust the baseline relaxation level based on the ratio of the alpha wave energy percentage to the theta wave energy percentage, and superimpose the deviation of the current prefrontal EEG coherence status compared to the baseline status to determine the contribution of neural relaxation to the comprehensive evaluation index. Indicators related to heart rate variability chaos sensitivity, skin conductance phase synchronization, and respiratory rhythm stability are obtained from the real-time physiological feature vector and normalized to obtain the contribution of autonomic neural balance to the comprehensive evaluation index. Real-time noise decibel value, light intensity, and ambient temperature are extracted from environmental parameters, and the contribution of environmental interference to the comprehensive evaluation index is determined based on the baseline threshold.

[0099] Step 401: The contribution values of the degree of neural relaxation, the degree of autonomic nervous balance, and the degree of environmental interference to the comprehensive evaluation index are integrated to obtain a comprehensive index for evaluating the effect of meditation.

[0100] Step 402: Analyze the value range of the comprehensive index and divide the comprehensive index into several intervals, each interval corresponding to a different meditation effect level;

[0101] Step 403: For each interval, determine the minimum and maximum values of the comprehensive indicator within the corresponding interval, and determine the corresponding minimum and maximum values of the meditation effect score. According to equal proportion, each value of the comprehensive indicator within the interval is mapped to the corresponding position of the meditation effect score.

[0102] In the embodiment of the present invention, in the real-time physiological characteristic vector, the α wave energy ratio α is first accurately located. r and theta wave energy ratio θ r Alpha waves occur when the brain is relaxed and awake, with a frequency range of 8-13Hz, while theta waves are associated with sleepiness and deep relaxation, with a frequency of 4-7Hz. Calculate the ratio of the two The higher the ratio, the more the brain tends to be in a state of wakefulness and relaxation, and it can be used as a core reference for basic relaxation. When , it means that the brain is in a relatively ideal state of relaxation and wakefulness. At the same time, the current frontal lobe EEG coherence status φ is obtained c , the frontal EEG coherence reflects the level of information interaction and collaborative work between different brain regions. c Compared with the pre-established baseline condition φ b Compare and calculate the deviation amplitude |φ c -φ b |. If |φ c -φ b A smaller value means that the current brain function state is close to the baseline level and neural activity is stable; otherwise, it means that brain function is fluctuating.

[0103] Will with |φ c -φ b | Add and multiply by weight w N , w N The value range is between 0.3-0.5, which is obtained through a large amount of data analysis.

[0104] Specifically, a multimodal dataset containing more than 1,000 samples was collected, covering people of different ages (18-65 years old), genders, and meditation experience (novice to practitioners with more than 5 years of experience). In a standard meditation experimental environment, physiological data such as EEG signals, heart rate variability, and skin conductance, as well as environmental parameters such as ambient noise and light intensity, were collected simultaneously. Based on this dataset, an exhaustive method was used to identify the w NDiscrete scanning was performed, taking values from 0 to 1 with a step size of 0.01, and 101 evaluation models with different weight combinations were constructed. Each model calculated the meditation effect score for all samples and generated a data set containing 101 groups of scoring results. An evaluation system was established with the subjective score of meditation (1-10 points) as the gold standard, using the root mean square error (RMSE) and mean absolute error (MAE) as quantitative indicators. Cross-validation was performed on the 101 groups of models, and the data set was divided into 10 subsets. 9 subsets were used to train the model each time, and the remaining 1 subset was tested. This was repeated 10 times to ensure the robustness of the results. It was found through calculation that when w N When w<0.3, the RMSE between the model prediction score and the subjective score is greater than 1.2, and the MAE is greater than 0.9, indicating that the weight of neural relaxation is too low, resulting in the evaluation result deviating from the actual result. N When w is greater than 0.5, the evaluation results of the model on different sample sets fluctuate violently, with the standard deviation exceeding 0.8, indicating that the excessive weight causes the evaluation system to lose the balance of consideration for other physiological indicators. N In the range of 0.3-0.5, RMSE stabilizes at 0.7-0.9, MAE remains at 0.5-0.7, and the model generalization performance is the best.

[0105] To further determine the final value, a genetic algorithm was used to conduct a refined search in the range of 0.3-0.5. The population size was set to 50, the number of iterations was 100, and the RMSE after cross-validation was used as the fitness function. After multiple rounds of evolution, the algorithm converged to w N = 0.4, the RMSE of the model on the independent test set reached the lowest value of 0.68, MAE was 0.49, and it maintained stable performance in subsets with different feature distributions. N The value is 0.4. The contribution value of nerve relaxation to the comprehensive evaluation index is obtained

[0106] Extracting Heart Rate Variability Chaotic Sensitivity S from Real-time Physiological Feature Vectors hrv , skin conductance phase synchronization S sc and respiratory rhythm stability S rrs Related indicators. In order to eliminate the influence of dimension and facilitate comparison and calculation, each indicator is normalized. hrv For example, the normalization formula is in and It is the minimum and maximum value obtained based on statistics of a large number of people. For example, in a certain database, If the current S hrv =30, then the normalized value is Similarly, skin conductance phase synchronization S scand respiratory rhythm stability S rrs Normalize and then multiply by their respective weights ω1, ω2, and ω3.

[0107] First, in order to determine the reasonable value range of the weights, a set of evaluation models containing multiple groups of different weight combinations is constructed. From a large amount of actual sample data, the chaotic sensitivity S of heart rate variability of different individuals in various states (such as resting, exercising, meditation, etc.) is collected. hrv , skin conductance phase synchronization S sc and respiratory rhythm stability S rrs Data. For example, in a data set containing 1,000 samples, the numerical changes of these three indicators during the meditation process of each sample are recorded in detail. Using these sample data, the evaluation model with different weight combinations is trained and tested. During the test, evaluation indicators such as mean square error (MSE) and mean absolute error (MAE) are used to measure the closeness between the autonomic balance predicted by the model and the actual situation. For the chaos sensitivity of heart rate variability S hrv By continuously adjusting the value of ω1, the model's evaluation performance was observed. When ω1 was less than 0.2, the model's sensitivity to chaotic heart rate variability (HRV) when predicting autonomic balance was insufficient, resulting in significant deviations from actual results and high mean square error and mean absolute error. When ω1 was greater than 0.3, the model overemphasized the influence of chaotic heart rate variability (HRV), weakening the role of other indicators and also increasing the error in the evaluation results. After extensive experiments and analysis, it was determined that the model can more accurately assess autonomic balance when the ω1 value range is 0.2-0.3.

[0108] For skin conductance phase synchronization S sc A similar analysis was performed for the weight ω2. Because skin conductance phase synchronization is closely related to autonomic nervous system activity and is highly sensitive in reflecting autonomic balance, testing the model with different ω2 values revealed that when ω2 is less than 0.3, the model fails to fully reflect the important influence of skin conductance phase synchronization on autonomic balance. When ω2 is greater than 0.4, the weight of skin conductance phase synchronization is too high, masking the role of other indicators and leading to inaccurate evaluation results. Therefore, it was determined that the model's evaluation performance is optimal when the ω2 value range is 0.3-0.4.

[0109] Similarly, for respiratory rhythm stability S rrsThe weight ω3 of the model was determined. After a series of experiments and analyses, it was found that when the value of ω3 is less than 0.2, the contribution of respiratory rhythm stability to autonomic balance is not fully reflected in the model; when the value of ω3 is greater than 0.3, the weight of respiratory rhythm stability is too large, affecting the overall balance of the model. It was finally determined that when the value range of ω3 is 0.2-0.3, the model can reasonably evaluate the impact of respiratory rhythm stability on autonomic balance. After determining the value range, in order to balance the impact of each indicator on autonomic balance, ω1=0.25, ω2=0.35, and ω3=0.25 are taken to balance the impact of each indicator on autonomic balance. Adding the results gives Finally multiply by the weight w A .

[0110] The data on chaotic sensitivity of heart rate variability, skin conductance phase synchronization and respiratory rhythm stability of different people in various states were collected, covering individuals of different ages, genders, health conditions and meditation experience. In the experiment, different w values from 0.1 to 0.9 were tried for the calculation of autonomic balance. A The comprehensive evaluation index is calculated by combining other evaluation indicators. A When the value is less than 0.3, it is found that the proportion of autonomic nervous balance in the comprehensive evaluation is too low, resulting in the comprehensive evaluation index being unable to fully reflect the reflection of the autonomic nervous state by indicators such as heart rate variability, skin conductance and respiratory rhythm, causing a large deviation between the evaluation results and the actual autonomic nervous balance state. For example, in some samples, even if these physiological indicators show that the autonomic nervous system is in a good balance state, the comprehensive evaluation results fail to accurately reflect this situation due to the low weight. When w A When the value is greater than 0.4, the weight of the autonomic balance is too high, which will over-amplify the influence of this part of the index, making the comprehensive evaluation index too much affected by the autonomic balance, and relatively weakening the role of other important factors such as nerve relaxation and environmental interference, resulting in incomplete and inaccurate evaluation results. After repeated analysis and comparison of a large amount of experimental data, it was found that when w A When the value is between 0.3 and 0.4, the comprehensive evaluation index can most accurately reflect the actual state of autonomic nervous balance, and cooperate with other evaluation factors to make the entire evaluation system more scientific and reasonable. Within this range, 0.35 is taken as w A The contribution value of the autonomic nervous balance to the comprehensive evaluation index is obtained.

[0111] Extract real-time noise decibel value, light intensity and ambient temperature from environmental parameters, and record them as p1, p2, p3, etc. (i=1,2,...,n, n is the number of environmental parameters). Each environmental parameter has a corresponding benchmark threshold Noise floor threshold decibel, light intensity benchmark threshold Lux, ambient temperature reference threshold And the standard deviation σ obtained through a large amount of environmental data statistics i Calculate the degree of deviation of each environmental parameter from the baseline threshold For example, if the noise decibel value p1 = 50 decibels and its standard deviation σ1 = 5, then the degree of deviation is Take the square root of the sum of all deviations, Multiply it by the weight w E Collect relevant data under various environments, including different noise levels, light intensity, temperature changes and other environmental factors, as well as corresponding evaluation index data. E To observe its impact on the comprehensive evaluation index. E When the value is less than 0.1, it is found that the environmental interference degree accounts for too small a proportion in the comprehensive evaluation, resulting in the comprehensive evaluation index being insensitive to changes in environmental factors. Even if the environmental conditions change significantly, the impact on the comprehensive evaluation results is not significant, and the role of environmental interference in the overall evaluation cannot be accurately reflected. For example, in some scenarios with large environmental changes, the actual situation is that environmental interference has a greater impact on the system, but due to w E The comprehensive evaluation indicators do not reflect this change.

[0112] And when w E When the value is greater than 0.2, the weight of environmental interference is too high, which will make the comprehensive evaluation index too dependent on environmental factors and relatively ignore the contribution of other important factors, such as autonomic nervous balance and nervous relaxation. This will cause the comprehensive evaluation results to fluctuate too much when environmental factors change, and it will not be able to objectively and comprehensively reflect the overall situation. Through repeated experiments and analysis of a large number of data under different environmental conditions, it was found that when w E When the value is between 0.1 and 0.2, it can better balance the role of environmental interference and other factors in the comprehensive evaluation. Within this range, 0.15 is taken as w E The contribution value of environmental interference to the comprehensive evaluation index is obtained.

[0113] Step 401: Add the contribution values of the above-calculated nerve relaxation, autonomic nerve balance and environmental interference to the comprehensive evaluation index, that is, The final comprehensive index C used to evaluate the effect of meditation is obtained. This index integrates multiple dimensions of physiological and environmental factors to comprehensively reflect the current meditation state.

[0114] Step 402: Analyze the value range of the comprehensive indicator C and divide it into four intervals based on statistical analysis of a large amount of user meditation data and actual experience:

[0115] When C<0.3, the corresponding meditation effect level is "poor", indicating that the current meditation state is not good and is greatly disturbed by physiological or environmental factors.

[0116] When 0.3≤C<0.6, the corresponding meditation effect level is "average", which means that meditation has a certain effect, but there is still room for improvement.

[0117] When 0.6≤C<0.8, the corresponding meditation effect level is "good", indicating that the meditation state is relatively ideal and the physiological and environmental factors work together well.

[0118] When C≥0.8, the corresponding meditation effect level is "excellent", which means that a deep meditation state is reached and the body and mind are in a highly coordinated state.

[0119] Step 403: For each interval, determine the minimum and maximum values of the comprehensive index within the interval, as well as the corresponding minimum and maximum values of the meditation effect score.

[0120] In the interval 0.3≤C<0.6, the minimum value C1=0.3, the maximum value C2=0.6, the corresponding minimum meditation effect score=60, and the maximum meditation effect score=80.

[0121] According to the principle of equal proportion distribution, for each comprehensive indicator value C in the interval X , through the formula

[0122] This comprehensive assessment encompasses multiple aspects, including neural relaxation, autonomic balance, and environmental disturbances, encompassing brain activity, autonomic nervous system regulation, and external environmental factors. Compared to assessments based on a single or limited set of indicators, this method comprehensively reflects the physiological state of the body during meditation and the influence of the external environment, avoiding biased assessments and ensuring a more accurate and complete evaluation of meditation effectiveness. Individual differences are fully considered when determining neural relaxation and autonomic balance. By comparing with personalized baseline characteristics (such as the deviation of prefrontal EEG coherence from the baseline) and normalizing individual physiological indicators, it adapts to the physiological characteristics of different users, providing each user with a personalized assessment result, enhancing its relevance and effectiveness. Using scientific calculation methods and mathematical models, weighting is used to reflect the importance of different factors in meditation effectiveness. Parameters such as baseline thresholds and standard deviations are determined based on extensive statistical analysis. A piecewise linear mapping algorithm is used to convert the comprehensive indicators into a score, ensuring a scientifically grounded assessment process and the reliability and credibility of the results. After comprehensively calculating complex physiological and environmental factors, the system converts them into a specific meditation effect score, which is presented to users in an intuitive manner. Users can quickly understand their own meditation effects through the score, facilitating self-assessment and comparison. It also provides a clear reference for meditation guidance and training, helping users adjust their meditation strategies and improve their meditation results.

[0123] In a preferred embodiment of the present invention, in step 5, the environmental interference level is compared with a preset reference threshold. When the environmental interference level is greater than or equal to the reference threshold, a segmented attenuation mechanism is activated. The segmented attenuation mechanism includes:

[0124] Mild attenuation mode: If the baseline threshold is less than or equal to the environmental interference level and less than 1.2 times the baseline threshold, a linear attenuation function is used to reduce the meditation effect score by a certain ratio, and the reduction ratio must be greater than or equal to the preset lower limit ratio.

[0125] Moderate decay mode: If 1.2 times the baseline threshold ≤ environmental interference < 1.5 times the baseline threshold, an exponential decay function is enabled to adjust the meditation effect score.

[0126] Severe attenuation mode: If the environmental interference level is ≥1.5 times the baseline threshold, cliff-like attenuation is initiated.

[0127] In the embodiment of the present invention, after the comprehensive evaluation index is calculated, the calculation result of the environmental interference degree is separated from the comprehensive evaluation index. The environmental interference degree is calculated by calculating the deviation of the real-time noise decibel value, light intensity, ambient temperature and other environmental parameters from their respective reference thresholds, and combining the corresponding weights to obtain the value. The calculation method is as follows: The preset benchmark threshold is a key parameter automatically generated by correlation analysis of environmental interference data and corresponding meditation effect scores. Specifically, more than 100,000 sets of environmental parameters (including noise decibels, light intensity, ambient temperature, etc.) from different users and their corresponding comprehensive meditation effect score data are collected. First, the environmental interference data are sorted in ascending order and divided into 100 equally spaced intervals, and the average meditation effect score of users in each interval is counted. The analysis found that when the environmental interference degree is in the range of 0-0.5, more than 85% of users' meditation effect scores remain above 70 points (out of 100), and the standard deviation of score fluctuations is less than 5; when the environmental interference degree exceeds 0.5, the average score begins to show a significant downward trend, and the standard deviation of score fluctuations increases to more than 15.

[0128] Based on the above data characteristics, we automatically identified a critical point of 0.5. Below this value, changes in environmental disturbance levels have a less than 3% impact on meditation effectiveness scores, and the score distribution shows a highly concentrated trend. Above this value, every 0.1 increase in environmental disturbance levels leads to an average decrease of 8-10 points in meditation effectiveness scores. Therefore, 0.5 was established as the baseline threshold, serving as a quantitative standard for determining whether the environment significantly affects meditation effectiveness.

[0129] Mild attenuation mode: When the environmental interference level meets the baseline threshold ≤ environmental interference level < 1.2 times the baseline threshold, the meditation enters the mild attenuation mode. Assuming the baseline threshold is 0.5, the environmental interference level is between 0.5 and 0.6. A linear attenuation function is used to reduce the meditation effect score. The linear attenuation function can be expressed as S n =S o ×(1-k), where S n is the meditation effect score after attenuation, S o is the score value before attenuation, and k is the attenuation ratio. The determination of the attenuation ratio k needs to ensure that it is ≥ the preset lower limit ratio. This lower limit ratio is also set based on a large amount of experimental data and experience, in order to avoid excessive attenuation affecting the rationality of the evaluation. For example, if the preset lower limit ratio is 0.1, if the current meditation effect score is 80, and k = 0.15 is calculated based on the relationship between the specific environmental interference and the benchmark threshold, then the score value S after attenuation is 0.15. n =80×(1-0.15)=68.

[0130] Moderate decay mode: When the environmental noise level is 1.2 times the baseline threshold ≤ the environmental noise level < 1.5 times the baseline threshold, the moderate decay mode is enabled. For example, the environmental noise level is between 0.6-0.75. The exponential decay function is used to adjust the meditation effect score. The exponential decay function can be expressed as S n =S o ×a b×(干扰度-1.2×基准阈值), where a and b are parameters determined based on experimental data. First, from the more than 100,000 sets of real meditation data collected in the past, more than 2,000 sets of valid data with environmental interference levels between 1.2 and 1.5 times the baseline threshold were screened out. These data contain accurately recorded environmental parameters (noise decibels, light intensity, temperature, etc.) and the meditation effect score at the corresponding moment. Interference data caused by equipment failure, sudden user actions, etc. have been eliminated through an outlier detection algorithm. A grid search strategy is used to exhaustively test parameters a and b. The value range of a is set to 0.1-0.9 with a step size of 0.1; the value range of b is set to 1-5 with a step size of 1, forming a total of 45 parameter combinations. For each set (a, b), the system substitutes the environmental interference levels in the more than 2,000 screened data sets into the exponential decay function to calculate the predicted meditation effect score.

[0131] The predicted scores were compared with the actual scores one by one, and the mean absolute error (MAE) was used as the evaluation indicator. MAE is calculated by taking the absolute value of the difference between the predicted score and the actual score of each sample, and then calculating the average of the absolute value differences of all samples. The closer the value is to 0, the lower the degree of deviation between the predicted score and the actual score, that is, the more accurate the function's portrayal of the relationship between environmental interference and meditation effect. After measuring 45 parameter combinations one by one, it was found that when a=0.8 and b=2, the MAE value reached a minimum of 3.1. Specifically, under this set of parameters, the error between the predicted score and the actual score of 87% of the samples is controlled within ±5 points, and the error distribution of the remaining 13% of the samples is random, with no obvious deviation. For example, when the environmental interference level is 1.3 times the baseline threshold and the original score is 70 points, the predicted score is 70×0.8 when substituted into the formula. 2 ×(1.3-1.2) =70×0.8 0.2 ≈63.7, which is highly consistent with the fluctuation trend of scores under the same interference level in actual data. Therefore, a = 0.8 and b = 2 were finally selected as the final parameter combination of the exponential decay function to ensure that the negative impact of environmental interference on meditation effect can be accurately quantified within this range.

[0132] Severe attenuation mode: When the environmental interference level is ≥1.5 times the baseline threshold, that is, the environmental interference level is ≥0.75, cliff-like attenuation is initiated. Cliff-like attenuation means a substantial and rapid reduction in the meditation effect score to highlight the destructive impact of severe environmental interference on the meditation effect. Set the meditation effect score to a lower fixed value, such as setting the score directly to 30; or calculate it according to a very large attenuation ratio, for example, set the attenuation ratio to 0.7. If the original score is 80, the attenuated score is 80×(1-0.7)=24. This method allows users to intuitively realize that the current environment has seriously interfered with the meditation effect, thereby prompting them to improve the meditation environment.

[0133] like Figure 2 As shown, an embodiment of the present invention further provides a meditation effect intelligent scoring system based on big data analysis, comprising:

[0134] The data acquisition module is used to collect brainwave signals, heart rate variability, skin conductance raw waveforms, and respiratory rhythm signals of users during meditation in real time through wearable devices. It also obtains ambient noise decibel values, light intensity, and ambient temperature parameters through environmental sensors to generate a multidimensional raw data set that integrates physiological signals and environmental parameters.

[0135] The benchmark generation module is used to generate personalized benchmark features based on the resting state physiological parameters in the multi-dimensional original data set using Gaussian mixture probability distribution, including the mean vector and covariance matrix of brain wave signals, heart rate variability, skin conductance, and respiratory rhythm parameters;

[0136] The feature extraction module is used to perform frequency domain decomposition on the EEG signal, extract the energy proportions of theta and alpha waves respectively, calculate the chaos sensitivity index of the heart rate variability signal and the phase synchronization parameter of the original skin conductance waveform, and form a real-time physiological feature vector;

[0137] The scoring calculation module is used to calculate the degree of neural relaxation, autonomic nervous balance and environmental interference based on the real-time physiological feature vector and environmental parameters, couple the degree of neural relaxation, autonomic nervous balance and environmental interference, and convert the calculation results into a meditation effect score through a piecewise linear mapping algorithm;

[0138] The attenuation processing module is used to compare the environmental interference level with a preset reference threshold. When the environmental interference level is greater than or equal to the reference threshold, the segmented attenuation mechanism is activated.

[0139] It should be noted that this system is a system corresponding to the above method, and all implementation methods in the above method embodiment are applicable to this embodiment and can achieve the same technical effects.

[0140] An embodiment of the present invention further provides a computing device comprising: a processor and a memory storing a computer program, wherein the computer program, when executed by the processor, performs the above-described method. All implementations in the above-described method embodiments are applicable to this embodiment and can achieve the same technical effects.

[0141] The embodiment of the present invention further provides a computer-readable storage medium storing instructions, which, when executed on a computer, causes the computer to execute the above-described method. All implementations in the above-described method embodiment are applicable to this embodiment and can achieve the same technical effects.

[0142] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present invention.

[0143] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0144] In the embodiments provided by the present invention, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the units is merely a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the mutual coupling or direct coupling or communication connection shown or discussed can be through some interface, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0145] The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0146] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0147] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes various media that can store program code, such as a USB flash drive, a mobile hard drive, ROM, RAM, a magnetic disk, or an optical disk.

[0148] In addition, it should be noted that, in the apparatus and method of the present invention, it is obvious that each component or each step can be decomposed and / or recombined. These decompositions and / or recombinations should be regarded as equivalent solutions of the present invention. Moreover, the steps of performing the above-mentioned series of processing can naturally be performed in chronological order according to the order of description, but it is not necessary to perform them in chronological order, and some steps can be performed in parallel or independently of each other. For those of ordinary skill in the art, it will be understood that all or any steps or components of the method and apparatus of the present invention can be implemented in any computing device (including processors, storage media, etc.) or a network of computing devices in hardware, firmware, software or a combination thereof, which can be achieved by those of ordinary skill in the art using basic programming skills after reading the description of the present invention.

[0149] Therefore, the purpose of the present invention can also be achieved by running a program or a group of programs on any computing device. The computing device can be a well-known general-purpose device. Therefore, the purpose of the present invention can also be achieved simply by providing a program product containing program code that implements the method or device. That is to say, such a program product also constitutes the present invention, and the storage medium storing such a program product also constitutes the present invention. Obviously, the storage medium can be any well-known storage medium or any storage medium developed in the future. It should also be pointed out that in the device and method of the present invention, it is obvious that each component or each step can be decomposed and / or recombined. These decompositions and / or recombinations should be regarded as equivalent schemes of the present invention. In addition, the steps of performing the above-mentioned series of processing can naturally be performed in chronological order according to the order of description, but do not necessarily need to be performed in chronological order. Certain steps can be performed in parallel or independently of each other.

[0150] The above is a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as within the scope of protection of the present invention.

Claims

1. A meditation effect intelligent scoring method based on big data analysis, characterized in that: The method comprises: Step 1: Use wearable devices to collect real-time brainwave signals, heart rate variability, skin conductance raw waveforms, and respiratory rhythm signals during meditation. Use environmental sensors to obtain ambient noise decibel values, light intensity, and ambient temperature parameters to generate a multidimensional raw data set that integrates physiological signals and environmental parameters. Step 2: Based on the resting-state physiological parameters in the multidimensional original dataset, a Gaussian mixture probability distribution is used to generate personalized baseline features, including the mean vector and covariance matrix of EEG signals, heart rate variability, skin conductance, and respiratory rhythm parameters; Step 3: Decompose the EEG signal in the frequency domain to extract the energy proportions of theta and alpha waves, respectively. Calculate the chaos sensitivity index of the heart rate variability signal and the phase synchronization parameter of the original skin conductance waveform to form a real-time physiological feature vector. Step 4: Calculate the degree of neural relaxation, autonomic nervous balance, and environmental interference based on the real-time physiological feature vector and environmental parameters. Couple the degree of neural relaxation, autonomic nervous balance, and environmental interference and convert the calculation results into a meditation effect score using a piecewise linear mapping algorithm. Step 5: Compare the environmental interference level with a preset reference threshold. When the environmental interference level is greater than or equal to the reference threshold, activate the segmented attenuation mechanism.

2. The meditation effect intelligent scoring method based on big data analysis according to claim 1 is characterized in that: Based on the resting state physiological parameters in the multidimensional original data set, a Gaussian mixture probability distribution is used to generate personalized baseline features, including the mean vector and covariance matrix of brain wave signals, heart rate variability, skin conductance, and respiratory rhythm parameters, including: Determine the time intervals during which the user is in a resting state from the multidimensional raw data set, and extract physiological parameter data from the resting time intervals, including brain wave signals, heart rate variability, skin conductance, and respiratory rhythm parameters; For each physiological parameter, a Gaussian mixture model is constructed to calculate the mean vector and covariance matrix of each physiological parameter; The mean vector and covariance matrix of each physiological parameter are combined to form a personalized baseline feature.

3. The meditation effect intelligent scoring method based on big data analysis according to claim 2 is characterized in that: Perform frequency domain decomposition on the EEG signal to extract the energy proportions of theta and alpha waves, respectively. Calculate the chaos sensitivity index of the heart rate variability signal and the phase synchronization parameter of the original skin conductance waveform to form a real-time physiological feature vector, including: Convert brain wave signals from the time domain to the frequency domain, and in the frequency domain, divide the corresponding frequency bands according to the frequency range of theta waves and alpha waves; Calculate the energy within each frequency band and, through normalization, obtain the frequency band energy ratios of theta and alpha waves to establish a baseline for the user's brainwave characteristics during meditation. After establishing a baseline of the user's brainwave characteristics in a meditative state, the current heart rate variability signal is obtained and compared with the personalized baseline characteristics using a dynamic time warping algorithm to obtain a comparison result. Based on the comparison results, the chaos sensitivity index of the heart rate variability signal is calculated to reflect the degree of deviation between the current heart rate fluctuation and the personalized benchmark; After obtaining the relevant features of brain waves and heart rate variability, the original skin conductance waveform and respiratory rhythm signal are obtained. A phase locking value algorithm is used to analyze the phase synchronization between the original skin conductance waveform and the respiratory rhythm signal to determine the coordinated change characteristics of the skin conductance signal and the respiratory rhythm. The frequency band energy proportions of theta and alpha waves, the chaos sensitivity index of the heart rate variability signal, the original skin conductance waveform and the respiratory rhythm signal are combined to form a real-time physiological feature vector that includes energy proportion, chaos sensitivity and phase synchronization.

4. The meditation effect intelligent scoring method based on big data analysis according to claim 3 is characterized in that: After establishing a baseline for the user's meditative EEG characteristics, the current heart rate variability signal is obtained. Using the dynamic time warping algorithm, the current heart rate variability signal is compared with the personalized baseline characteristics to obtain the comparison results, including: After establishing a baseline for brainwave characteristics, wearable heart rate monitoring devices are used to obtain the user's current heart rate data during meditation in real time; Process the current heart rate data, calculate the heart rate variability index, and obtain the current heart rate variability signal; Using the dynamic time warping algorithm, the current heart rate variability signal is compared with the heart rate variability baseline data in the personalized baseline feature. During the comparison process, the distance between each matching point is calculated. According to the distance between each matching point, the final matching path between the two signals is determined, and the distance between each matching point is accumulated to obtain the total comparison distance, that is, the comparison result.

5. The meditation effect intelligent scoring method based on big data analysis according to claim 4 is characterized in that: After obtaining the relevant features of brain waves and heart rate variability, the original skin conductance waveform and respiratory rhythm signal are obtained. A phase locking value algorithm is used to analyze the phase synchronization between the original skin conductance waveform and the respiratory rhythm signal, and determine the coordinated change characteristics of the skin conductance signal and the respiratory rhythm, including: Convert the skin conductance signal and respiratory rhythm signal from time domain features to phase domain, obtain the skin conductance signal and respiratory rhythm signal, and calculate the phase difference between the skin conductance signal and respiratory rhythm signal at the same time; According to the phase difference at each moment, a time window is set. Within each time window, the phase difference within the corresponding time period is statistically analyzed, and the phase difference at each moment is converted into a complex form. The complex phase difference at each moment in the time window is fused to obtain the phase locking value in the corresponding time window; Summarize the phase locking values of each time window to obtain a summary result; According to the summary results and the preset evaluation criteria, the coordinated change characteristics between the skin conductance signal and the respiratory rhythm signal are determined.

6. The meditation effect intelligent scoring method based on big data analysis according to claim 5 is characterized in that: Based on the real-time physiological feature vector and environmental parameters, the degree of neural relaxation, the degree of autonomic balance and the degree of environmental interference are calculated respectively. The degree of neural relaxation, the degree of autonomic balance and the degree of environmental interference are coupled and calculated. At the same time, the calculation results are converted into meditation effect scores through the piecewise linear mapping algorithm, including: The basic relaxation level is dynamically adjusted by the ratio of the alpha wave energy ratio to the theta wave energy ratio, and the deviation amplitude of the current prefrontal EEG coherence status compared to the baseline status is superimposed to determine the contribution of neural relaxation to the comprehensive evaluation index. Indicators related to heart rate variability chaos sensitivity, skin conductance phase synchronization, and respiratory rhythm stability are obtained from the real-time physiological feature vector and normalized to obtain the contribution of autonomic neural balance to the comprehensive evaluation index. The real-time noise decibel value, light intensity, and ambient temperature are extracted from the environmental parameters, and the contribution of environmental interference to the comprehensive evaluation index is determined based on the baseline threshold. The contribution values of neural relaxation, autonomic nervous balance, and environmental interference to the comprehensive evaluation index are integrated to obtain the final comprehensive index for evaluating meditation effects. Analyze the value range of the comprehensive index and divide it into several intervals, each of which corresponds to a different level of meditation effect; For each interval, determine the minimum and maximum values of the comprehensive indicator within the corresponding interval, and determine the corresponding minimum and maximum values of the meditation effect score. According to equal proportion, each value of the comprehensive indicator within the interval corresponds to the corresponding position of the meditation effect score.

7. The meditation effect intelligent scoring method based on big data analysis according to claim 6 is characterized in that: The segmented attenuation mechanism includes: Mild attenuation mode: If the baseline threshold is less than or equal to the environmental interference level and less than 1.2 times the baseline threshold, a linear attenuation function is used to reduce the meditation effect score by a certain ratio, and the reduction ratio must be greater than or equal to the preset lower limit ratio. Moderate decay mode: If 1.2 times the baseline threshold ≤ environmental interference < 1.5 times the baseline threshold, an exponential decay function is enabled to adjust the meditation effect score. Severe attenuation mode: If the environmental interference level is ≥1.5 times the baseline threshold, cliff-like attenuation is initiated.

8. An intelligent scoring system for meditation effects based on big data analysis, the system implementing the method according to any one of claims 1 to 7, characterized in that: include: The data acquisition module is used to collect brainwave signals, heart rate variability, skin conductance raw waveforms, and respiratory rhythm signals of users during meditation in real time through wearable devices. It also obtains ambient noise decibel values, light intensity, and ambient temperature parameters through environmental sensors to generate a multidimensional raw data set that integrates physiological signals and environmental parameters. The benchmark generation module is used to generate personalized benchmark features based on the resting state physiological parameters in the multi-dimensional original data set using Gaussian mixture probability distribution, including the mean vector and covariance matrix of brain wave signals, heart rate variability, skin conductance, and respiratory rhythm parameters; The feature extraction module is used to perform frequency domain decomposition on the EEG signal, extract the energy proportions of theta and alpha waves respectively, calculate the chaos sensitivity index of the heart rate variability signal and the phase synchronization parameter of the original skin conductance waveform, and form a real-time physiological feature vector; The scoring calculation module is used to calculate the degree of neural relaxation, autonomic nervous balance and environmental interference based on the real-time physiological feature vector and environmental parameters, couple the degree of neural relaxation, autonomic nervous balance and environmental interference, and convert the calculation results into a meditation effect score through a piecewise linear mapping algorithm; The attenuation processing module is used to compare the environmental interference level with a preset reference threshold. When the environmental interference level is greater than or equal to the reference threshold, the segmented attenuation mechanism is activated.

9. A computing device, characterized in that include: one or more processors; A storage device for storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a program, which, when executed by a processor, implements the method according to any one of claims 1 to 7.

Images