A method for generating white noise soothing music similar to deep sleep brain waves

Abstract

The application discloses a white noise soothing music generation method of deep sleep-like brain waves; the method is as follows: 1, collecting the brain wave data of the target object in the sleep state. 2, performing staging processing on the brain wave data obtained in step 1. 3, performing frequency band separation on the extracted sleep stage to obtain brain wave data of one or more target frequency bands. 4, constructing a wave function of comprehensive rain sound; 5, establishing a mapping from a single cycle brain wave to a single raindrop sound waveform. 6, integrating the raindrop sounds generated by the single cycle brain wave signals of all channels and frequency bands together to form a continuous sleep-aiding rain sound frequency as the white noise soothing music. The application generates the matching raindrop sounds corresponding to the single cycle brain wave signals of different channels and different wave bands of the target object, and integrates all the obtained raindrop sounds to form a continuous rain sound frequency matched with the brain waves of the target object; the rain sound frequency is helpful for the sleep of the target object.

Description

Technical Field

[0001] This invention belongs to the fields of information technology and biomedical technology, and relates to a method for generating soothing white noise using EEG signals similar to deep sleep. Specifically, it relates to a method for generating white noise simulating the sound field of an outdoor rainy day while retaining the characteristics of deep sleep signals using multi-channel EEG signal waveforms. Background Technology

[0002] Electroencephalography (EEG) is a physiological and medical diagnostic method that uses multiple time-series recordings of electrical signals generated by the synchronized activity of clusters of brain neurons, collected through electrodes placed on the scalp. It effectively characterizes the activity state of the human brain and is an important diagnostic tool in clinical medicine (e.g., sleep monitoring, epilepsy diagnosis). After recording brain electrical activity using EEG, brain waves can be classified into different waveforms, called brain rhythms, based on the characteristic frequency range and spatial distribution of the signals. Typical waveforms include alpha waves (generally between 8 and 13 Hz, associated with motor imagery) and delta waves (generally between 0.5 and 4 Hz, the lowest frequency and largest amplitude brain rhythm). Different morphologies of brain electrical rhythms reflect different functional states of the brain. For example, when a person is awake, beta waves dominate the brainwaves, while an increase in the proportion of alpha waves indicates that the brain is entering a drowsy or sleepy state. When delta waves are particularly prominent, it indicates that the brain has entered deep sleep. Spikes, for instance, are a special type of abnormal brainwave waveform that often accompanies epileptic seizures and can therefore be used as a basis for the diagnosis and prediction of epilepsy. Methods of treating diseases by intervening in brain electrical signals have been applied in the medical field. For example, cranial electrotherapy stimulation (CES) uses microcurrents of a certain frequency applied to electrodes attached to the outside of the skull to stimulate the brain, and has been shown to have an enhancing effect on brainwaves, improving sleep quality by inducing alpha waves.

[0003] Deep sleep (N3 stage) is an artificially defined sleep stage that normally accounts for at least 20% of total sleep time. It is a crucial stage for the brain and body to rest and recover, and its duration and quality are important indicators of sleep quality. During deep sleep, EEG mainly shows brain waves with frequencies of 0.5 to 4 Hz, with low-frequency, high-amplitude delta waves with amplitudes of 75 to 200 mV dominating, generally exhibiting a stable and quiet trend.

[0004] White noise is a monotonous, repetitive humming sound with a power spectral density uniformly distributed across the entire frequency domain, covering the entire range of human hearing. It provides a monotonous auditory stimulus to the human ear. Natural sounds in daily life, such as rain, wind, flowing water, ocean sounds, and fan noise, are similar to white noise. It is a monotonous, predictable, and meaningless sound wave that is easily ignored, does not prolong sleep latency, and can be used as a monotonous stimulus to induce sleepiness. It can also induce and maintain sleep by masking environmental noise. The sleep-inducing effects of white noise and its masking effect on environmental noise have been extensively studied. Research shows that white noise can improve insomnia, help with sleep onset, and improve sleep quality by increasing delta rhythms in the brain and reducing bilateral hemisphere activity. It can also alleviate anxiety and reduce pain to some extent.

[0005] While traditional brainwave intervention methods and white noise intervention methods each have their advantages, they also have certain limitations in terms of ease of use and practical effects, and are not currently widely used. Patent CN102999701B discloses an innovative bidirectional conversion method between brainwaves and musical scores, which can generate corresponding MIDI music with a certain mode from brainwave signals and store brainwave signals as music, but it is currently difficult to demonstrate the further practicality of this invention.

[0006] Brainwave entrainment (BWE) is a novel brainwave intervention technique that utilizes the human brain's tendency to synchronize with external stimuli. By using specific stimuli, it induces a brainwave response in the brain that matches the frequency of the external stimulus, thereby achieving the effect of brainwave intervention. The paper "A comprehensive review of the psychological effects of brainwave entrainment. Alternative Ther Health Med. 2008" summarizes the experimental results and findings of BWE in improving human physiological and psychological functions.

[0007] Brainwave music (or brainwave music) is a way of transmitting brainwave characteristics and information in a way that uses sounds that can be received, distinguished, or understood by the human ear. It can be used as a carrier for brainwave entrainment technology. Studies have shown that sounds or music carrying electroencephalographic information (EEGs) have a certain impact on the human brain, and this impact can be positive. For example, the paper "Exploration of the Effect and Mechanism of Brainwave Music Intervention on Fatigue (Wang Junze, Xu Ruijie, Huang Binxin, et al. Fudan Journal. 2023)" points out that EEG music in different states has different effects on the human brain. The paper "Exploration of the Sleep-Promoting Effect of Brainwave Music (Cheng Yibo, Long Siyu, Lu Jing. Abstracts of the 21st National Conference on Psychology. 2018)" found that slow-wave sleep EEG music and REM sleep EEG music significantly improved the effect of shortening the time to fall asleep and prolonging the time to deep sleep compared with ordinary white noise. The paper "Improvement Effect of Alpha EEG Music on Memory of Middle School Students (Li Jie, An Bo, Cui Wei, et al. Chinese Journal of Mental Health. 2012)" confirmed that alpha EEG music has a more significant effect on improving memory in middle school students than classical music. Summary of the Invention

[0008] To address the shortcomings of existing technologies, this invention proposes a method for generating soothing rain-sound white noise using EEG signals from deep sleep. This method includes generating raindrop sound signals using a physical model and generating multi-frequency continuous rain sound signals by mapping the single-wave characteristics and spatiotemporal distribution characteristics of EEG signals. This achieves the generation of simulated rainy day white noise while preserving the characteristics of deep sleep EEG, combining the beneficial effects of brainwave music and white noise on the human brain, and providing a new approach for external stimulation therapy of neurological diseases such as insomnia and epilepsy.

[0009] A method for generating soothing white noise music based on deep sleep EEG data includes the following steps:

[0010] Step 1: Collect EEG data from the target subject who is asleep.

[0011] Step 2: Perform staged processing on the EEG data obtained in Step 1.

[0012] Step 3: Perform frequency band separation on the extracted sleep stages to obtain brainwave data for one or more target frequency bands.

[0013] Step 4: Construct the wave function P(t) of the combined rain sound as follows:

[0014]

[0015] Among them, A I The fitting parameters for the sound of raindrops hitting the ground are: t is time; f is the time. I It is the frequency of a random distribution; A B β represents the fitting parameters for the sound bubbles. Bω is the damping constant; ω is the angular resonant frequency, and its expression is: γ is the specific heat of air; P0 is the hydrostatic pressure in the water surrounding the bubble; a0 is the bubble radius; ρ0 is the water droplet density.

[0016] Step 5: Establish a mapping from a single cycle of brain waves to a single raindrop sound waveform.

[0017] 5-1. Extract the single-cycle brainwave signal for each channel and frequency band obtained in step 3.

[0018] 5-2. For each single-cycle brainwave signal, the brainwave frequency f and A are extracted; and a bubble radius a0 of a single raindrop sound wave function P(t) is generated accordingly; the fitting parameters A of the raindrop impacting the ground sound are... I Frequency f I Fitting parameters A for vocal fry B as follows:

[0019]

[0020]

[0021] f I =k fI f+E fI

[0022] a0 = k a0 f+W a0

[0023] Where, k AI k AB k fI k a0 E AI E AB E fI E a0 All parameters were adjusted.

[0024] Step 6: Integrate the raindrop sounds generated by the single-cycle brainwave signals of all channels and frequency bands into a continuous sleep-aiding rain sound audio, which can be used as white noise soothing music.

[0025] Preferably, in step 1, scalp patch electrodes are used to acquire EEG signals; the acquired EEG data is preprocessed to filter out interference.

[0026] Preferably, in step 2, the EEG data is used to perform sleep staging.

[0027] Preferably, the N3 sleep stage fragment is extracted in step 2.

[0028] Preferably, in step 3, seven frequency bands are extracted: the delta band from 0.5 Hz to 4.5 Hz, the theta band from 4 Hz to 8 Hz, the alpha band from 8 Hz to 13 Hz, the σ band from 11 Hz to 16 Hz, the β1 band from 13 Hz to 22 Hz, the β2 band from 22 Hz to 30 Hz, and the λ band from 30 Hz to 40 Hz.

[0029] As a preferred method, the process for constructing the wave function P(t) of the combined rain sound is as follows:

[0030] (1) Establish the sound pressure p when raindrops hit the ground. I The mathematical model for (r, t) is as follows:

[0031]

[0032] Where ρ is the air density; c is the speed of sound in air; r is the distance to the sound source; k is the wave number of the sound source; θ is the sound propagation angle; i is the imaginary unit; D S (t) is the dipole intensity in the sound field.

[0033] Dipole intensity D in the sound field S The expression for (t) is as follows:

[0034]

[0035] Where ρ0 is the density of the water droplet; c0 is the speed of sound in the water droplet; V I f is the velocity of a raindrop hitting the ground. I It is the frequency of a random distribution; β I It is related to frequency f I A constant in a predetermined proportion;

[0036] The simplified result is the wave function P of the sound of raindrops hitting the ground. I (t):

[0037]

[0038] (2) Establish the sound pressure p of the bubbles generated by raindrops hitting the water surface. B The mathematical model for (r, t) is as follows:

[0039]

[0040] Where, β B D is the damping constant; B ω represents the initial dipole intensity; ω is the resonant angular frequency of the raindrop and the water surface.

[0041] Damping constant β B The expression for the angular resonance frequency ω is as follows:

[0042]

[0043]

[0044] Where a0 is the bubble radius; γ is the specific heat of air; and P0 is the hydrostatic pressure in the water surrounding the bubble. th Let be the thermal damping constant of the raindrop.

[0045] The simplified wave function P of the vocal fry is obtained. B (t):

[0046]

[0047] The integrated wave function of the rain sound is obtained as P(t) = P I (t)+P B (t).

[0048] The present invention has the following beneficial effects:

[0049] 1. This invention generates matching raindrop sounds based on single-cycle brainwave signals from different channels and bands of the target object, and integrates all the obtained raindrop sounds to form a continuous rain sound audio that matches the brainwaves of the target object; this rain sound audio helps the target object sleep.

[0050] 2. This invention proposes a method for extracting EEG features. By filtering the original EEG frequency bands, and using the zero-crossing method to separate individual EEG cycles, the entire EEG segment is transformed into discrete single-cycle waves, and only the raindrop sound corresponding to the single-cycle wave is obtained.

[0051] 3. This invention proposes a simple method for generating simulated raindrop sound signals. By establishing a physical sound field model and merging simplified functions, it achieves two types of raindrop sound waveform functions controlled by four parameters. Simultaneously, this invention proposes a method for mapping EEG features to raindrop features, enabling the generation of raindrop sounds with corresponding features based on features extracted from EEG.

[0052] 4. This invention proposes a method to convert a whole segment of deep sleep EEG into rain scene sound, generating rain sound white noise that is beneficial to human sleep while preserving EEG characteristics. Attached Figure Description

[0053] Figure 1 This is a schematic diagram of the process of the present invention.

[0054] Figure 2 This is a schematic diagram illustrating the electroencephalogram (EEG) acquisition process in this invention.

[0055] Figure 3 This is a schematic diagram illustrating the single-cycle separation and feature extraction of a single-channel brainwave signal in this invention.

[0056] Figure 4 This invention generates a raindrop sound signal diagram.

[0057] Figure 5 This is a diagram of a continuous rain sound audio signal generated by the present invention based on multi-channel sleep EEG data of a healthy adult.

[0058] Figure 6 This is a diagram of the rain sound audio signal generated by separating the α and δ frequency bands during deep sleep from the frontal dual-channel EEG data of a healthy adult.

[0059] Figure 7 This is a diagram showing the results of a sleep aid experiment using generated white noise soothing music in this invention. Detailed Implementation

[0060] The following is a detailed description of each step of the present invention.

[0061] A method for generating soothing white noise music based on deep sleep EEG data includes the following steps:

[0062] Step 1: EEG signal acquisition:

[0063] Electroencephalogram (EEG) signals were acquired from sleep subjects using scalp patch electrodes. The electrode placement and arrangement followed the 10-20 International Lead Consortium (ICC) standard. Figure 2 This is an internationally recognized standard for EEG electrode placement, where the distance from the midpoint of the frontal pole to the root of the nose and the distance from the occipital point to the external occipital protuberance each account for 10% of the total length of this line. All other points are spaced 20% of the total length of this line, hence the name "10-20 system." After acquiring the raw EEG signals, preprocessing is required to filter out interference. This embodiment requires removing environmental noise and electrooculography (EOG) interference from the raw data, specifically using spatial filtering, independent principal component analysis (ICA), or the average reference electrode method (ARE). These three methods are existing technologies and will not be elaborated upon here. After the above operations, clear multi-electrode channel EEG time-series signal data will be obtained.

[0064] Step 2, EEG sleep staging:

[0065] A pre-trained graph neural network (GNN) was used to stage sleep in EEG signals and extract segments of deep sleep (stage N3). Graph Neural Networks (GNNs) are a type of neural network algorithm based on graph structures. Compared to traditional neural networks, they can better extract spatial and temporal features and relationships from data. Therefore, by constructing brain connectivity networks to extract the temporal dependencies between adjacent sleep stages, they can capture sleep transition rules and improve the model's classification performance. After establishing the GNN model, a publicly available sleep staging dataset was first used as the training set to train the model. Once the required test metrics were met (i.e., staging accuracy reached 90%), the parameterized model was saved and used as the sleep staging tool in this invention.

[0066] Step 3: Brainwave frequency band separation:

[0067] The deep sleep (i.e., N3 stage) extracted in step 2 was filtered within a specified frequency range using FIR bandpass filtering to separate the brainwave signals of each band.

[0068] The FIR (Finite Impulse Response) filter used in this embodiment is a linear time-invariant digital filter with a finite impulse response. It consists of a series of linearly correlated delay lines with different delay times and weights, forming low-pass, band-pass, and high-pass filtering methods, thus enabling arbitrary filter frequency responses. This invention implements the FIR filter algorithm through programming, calculating an integral value at each sampling point of the input signal. Since the filter coefficients are the weighted sum of a series of sampling points preceding that sampling point, the final output of the filter is determined.

[0069] In the brainwave frequency bands separated using FIR filters, the main parameters are the lowest and highest filtering frequencies. Specifically, from low to high, 0.5 Hz to 4.5 Hz is the delta band, 4 Hz to 8 Hz is the theta band, 8 Hz to 13 Hz is the alpha band, 11 Hz to 16 Hz is the σ band, 13 Hz to 22 Hz is the β1 band, 22 Hz to 30 Hz is the β2 band, and 30 Hz to 40 Hz is the λ band. After filtering the EEG data of each channel using the above-set filter algorithm, multidimensional EEG data for each frequency band will be obtained.

[0070] Step 4: Digital raindrop sound synthesis:

[0071] 4-1. Establish a raindrop sound function model

[0072] Based on physical principles, establish the sound pressure p when a raindrop hits the ground. I Mathematical model of (r, t):

[0073]

[0074] Where ρ is the air density; c is the speed of sound in the air; r is the distance from the sound source (specifically, the distance from the listener to the sound source); t is time; k is the wave number of the sound source; θ is the angle of sound propagation (specifically, the angle between the line connecting the sound source to the listener's ear and the normal to the ground); and i is the imaginary unit.

[0075] D S (t) is the dipole intensity in the sound field, which is derived from the sound field geometric model, as shown in the following formula:

[0076]

[0077] Where ρ0 is the density of the water droplet; c0 is the speed of sound in the water droplet; V I f is the velocity of a raindrop hitting the ground. I It is the frequency of a random distribution; β I It is related to frequency f I A constant in a certain proportion; f I ∈[1kHz, 16kHz], β I ∈[0.1f I 10f I ];β I with f I It is possible to integrate using the fitting coefficients.

[0078] In summary, the simplified result yields the wave function P of the sound of raindrops hitting the ground. I (t):

[0079]

[0080] Among them, parameters Its effect on loudness is linear and can be ignored.

[0081] Similarly, establish the sound pressure p of the bubbles generated by raindrops hitting the water surface. B The mathematical model for (r, t) is as follows:

[0082]

[0083] Where, β B D is the damping constant; B ω represents the initial dipole intensity; ω is the resonant angular frequency of the raindrop and the water surface.

[0084] Damping constant β B The angular resonance frequency ω can be further expressed as:

[0085]

[0086]

[0087] Where a0 is the bubble radius; γ is the specific heat of air; P0 is the hydrostatic pressure in the water surrounding the bubble, which is approximately standard atmospheric pressure in this embodiment. th Let be the thermal damping constant of the raindrop.

[0088] Similarly, the sound pressure level of bubbly sounds p B The mathematical model of (r, t) can be simplified to obtain the wave function P of the sound bubble. B (t):

[0089]

[0090] Among them, parameters It only affects loudness and can be ignored.

[0091] In this embodiment, a normal rainfall environment is simulated, so the parameters can be taken as follows:

[0092] G th =1.6×10 6 s / m

[0093] c = 343 m / s, c0 = 1497 m / s

[0094] ρ = 1.29 kg / m 3 ρ0=1000kg / m 3

[0095] P0 = 101.325 kPa

[0096] γ = 1.4

[0097]

[0098]

[0099] Substituting equations (1) and (2) into the equations, we can obtain the wave function P(t) of the combined rain sound as follows:

[0100]

[0101] Among them, f I ∈[1kHz, 16kHz], a0∈[1.6×10 -4 m, 4.7 × 10 -4 m]

[0102] Therefore, this embodiment establishes a system containing f I ,a0,A I / B There are two types of raindrop sound functions with four parameters each.

[0103] 4-2. Establish a mapping from a single cycle of brain waves to a single raindrop sound waveform.

[0104] The brainwave signal of a single cycle is extracted using the zero-crossing method. The extraction process is as follows: with time as the horizontal axis and brainwave amplitude as the vertical axis, the brainwave of a single cycle passes through the first zero point, a positive half-cycle, the second zero point, a negative half-cycle, and finally ends at the third zero point. Among them, half of the intensity difference between the highest point of the positive half-cycle and the lowest point of the negative half-cycle is called the brainwave amplitude A. The reciprocal of the time from the beginning to the end of the entire cycle is called the brainwave frequency f. The spatiotemporal characteristics of the positive and negative half-cycles are called the brainwave waveform.

[0105] In this field, the characteristic used to evaluate the intensity of a segment of brain electrical activity is not the amplitude of the brain electrical signal, but rather the power spectral density P of the brain waves, and P∝A. 2 The amplitude A is linearly related to the sound intensity index, so mapping the amplitude of the EEG to the amplitude of the sound in the form of a quadratic function can preserve its characteristics. Although the frequency of the EEG is lower than the frequency range of the sound that the human ear can perceive, the distribution of its frequency bands is within a certain range (i.e., 0.5 Hz to 40 Hz), so a mapping from brain waves to rain sounds can be established. In the raindrop sound function established in step 4-1 (see Equation (3)), the parameter that determines the function graph is parameter A. I Parameter A B Frequency f I Bubble radius a0; where A I A B These determine the amplitudes of the impact sound and the bubble sound, respectively; f I a0 and e respectively affect the amplitude in an exponential manner, and also determine the frequencies of the impact sound and the bubble sound; while the relationship between the frequency and amplitude of EEG signals follows a power-law distribution, that is, the frequency also affects the amplitude in an exponential manner. Therefore, linearly mapping the EEG frequency to the frequency of raindrop sound can scientifically preserve its characteristics. On the other hand, statistical analysis of real rain sounds shows that when the falling raindrops are large and close to the listener, the impact sound P I Compared to bubble sound P B When the raindrops are large, the sound is similar to the distinct sound of heavy rain; conversely, when the raindrops are small and far away, the sound is similar to the background sound of light rain. Therefore, to reproduce the realistic sound of rain, this invention sets different coefficients for the mapping of the two parameters.

[0106] In summary, the feature mapping method established in this embodiment is as follows:

[0107]

[0108]

[0109] fI =k fI f+E fI

[0110] a0 = k a0 f+E a0

[0111] Where, k AI k AB k fI k a0 E AI E AB E fI W a0 is an adjustable constant; f and A are the brainwave frequency and amplitude of the target object, respectively.

[0112] In summary, this step establishes a system from feature extraction to feature mapping, scientifically mapping single-cycle brainwaves into the sound waveform of a single raindrop. The duration of the sound of a single raindrop hitting the ground is 0.01–0.05 seconds.

[0113] 4-3. Convert the entire brainwave sequence into a complete rain sound audio segment.

[0114] After extracting a single cycle using the zero-crossing method, the interval between the start time of one cycle and the start time of the next cycle is defined as the duration of two brainwave cycles, and the interval of the raindrop sound is set based on this.

[0115] A single frequency band of brainwaves in a single channel can be mapped to a continuous raindrop sound. Multiple frequency bands in a single channel can be synthesized into rain sound audio with similar loudness and diverse frequencies, and multiple channels can be synthesized into rain sound audio with different loudness. Thus, it is possible to generate rain sound audio with rich loudness, frequency, and components and a sense of space based on multi-channel EEG signals, realizing a scientific method for generating rain sound white noise from EEG signals.

[0116] Based on the above method, each channel of the target object's EEG signal generates a raindrop sound for each frequency band and each period; thus, different channels and different EEG signals form overlapping rain sound audio on the time axis.

[0117] The differences in rain sound audio among different target objects stem from the differences in the frequency, intensity, and spatiotemporal distribution of their sampled brainwave signals. Rain sound audio generated by the target object's deep sleep EEG carries physiological information about their deep sleep, and the human brain has a tendency to synchronize with external stimuli. Therefore, this music has a positive guiding effect on the target to enter a stable deep sleep state.

[0118] To verify the positive effect of the music generated in this embodiment on human sleep, a sleep-aid experiment was conducted using the soothing rain sound music generated by this invention. The experimental results are as follows: Figure 7As shown in the figure, the Pittsburgh Sleep Quality Index was used to quantify sleep quality. The survey experiment selected 31 volunteers aged 20-25. Each volunteer assessed their sleep quality after falling asleep normally, using ordinary rain-sound white noise for sleep aid, and using rain-sound white noise modulated by the present invention to induce deep sleep EEG. The Pittsburgh Sleep Quality Index scores were obtained (higher scores indicate poorer sleep quality). As can be seen from the figure, the average scores for falling asleep normally and using ordinary white noise for sleep aid, and ordinary white noise and music modulated by the present invention for sleep aid were 5.71, 5.53, and 5.41, respectively, showing a decreasing trend. These scores passed the lenient significance test (α=0.1), thus verifying the sleep-aiding effect of this embodiment.

Claims

1. A method for generating soothing white noise music based on deep sleep EEG patterns, characterized in that: Includes the following steps: Step 1: Collect EEG data from the target subject while they are asleep; Step 2: Perform staged processing on the EEG data obtained in Step 1 and extract the N3 sleep stage segments; Step 3: Perform frequency band separation on the extracted sleep stages to obtain brainwave data in seven target frequency bands, namely the delta band from 0.5 Hz to 4.5 Hz, the theta band from 4 Hz to 8 Hz, the alpha band from 8 Hz to 13 Hz, the σ band from 11 Hz to 16 Hz, the β1 band from 13 Hz to 22 Hz, the β2 band from 22 Hz to 30 Hz, and the gamma band from 30 Hz to 40 Hz. Step 4: Construct the wave function of the integrated rain sound as follows: in, These are the fitting parameters for the sound of raindrops hitting the ground; For time; It is a frequency of random distribution; These are the fitting parameters for the sound bubbles; Here is the damping constant; Let be the angular resonant frequency, and its expression is: ; The specific heat of air; The static pressure in the water surrounding the bubble; Where is the bubble radius; The density of the water droplets; Step 5: Establish a mapping from a single cycle of brainwaves to a single raindrop sound waveform; Step 5-1. Extract the single-cycle brainwave signal of each channel and frequency band obtained in Step 3 using the zero-crossing method; half of the intensity difference between the highest point of the positive half-cycle and the lowest point of the negative half-cycle in the single-cycle brainwave signal is called the amplitude A of the brainwave; the reciprocal of the time from the beginning to the end of the entire cycle is called the frequency f of the brainwave. Step 5-2. For each single-cycle brainwave signal, extract the brainwave frequency f and amplitude A; and generate a corresponding single raindrop sound wave function. bubble radius Fitting parameters for the sound of raindrops hitting the ground ;frequency Fitting parameters for vocal fry as follows: in, , , , , , , , All parameters were adjusted; Step 6: Integrate the raindrop sounds generated by the single-cycle brainwave signals of all channels and frequency bands into a continuous sleep-aiding rain sound audio, which can be used as white noise soothing music. After extracting a single cycle using the zero-crossing method, the interval between the start time of one cycle and the start time of the next cycle is defined as the duration of two brainwave cycles, and the interval of the raindrop sound is set accordingly. The single frequency band brainwave of a single channel is mapped to a continuous raindrop sound, and multiple channels are synthesized into rain sound audio with different loudness.

2. The method for generating soothing white noise music based on deep sleep EEG as described in claim 1, characterized in that: In step 1, scalp patch electrodes are used to acquire EEG signals; the acquired EEG data are preprocessed to filter out interference.

3. The method for generating white noise soothing music based on deep sleep EEG as described in claim 1, characterized in that: In step 2, the EEG data is used to perform sleep staging.

4. The method for generating white noise soothing music based on deep sleep EEG as described in claim 1, characterized in that: Wave function of combined rain sound The construction process is as follows: (1) Establish the sound pressure when raindrops hit the ground The mathematical model is as follows: Where ρ is the air density; c is the speed of sound in the air; r is the distance to the sound source; k is the wave number; and θ is the sound propagation angle. The imaginary unit; It is the dipole intensity in the sound field; Dipole intensity in the sound field The expression is as follows: in, The density of the water droplets; The speed of sound in a water droplet; The speed at which raindrops hit the ground; It is a frequency of random distribution; It is related to frequency A constant in a predetermined proportion; The simplified result is the wave function of the sound of raindrops hitting the ground. : (2) Establish the sound pressure of air bubbles generated by raindrops hitting the water surface. The mathematical model is as follows: in, Here is the damping constant; The initial dipole strength; The resonant angular frequency of the raindrop and the water surface; Damping constant and angular resonance frequency The expression is as follows: in, Where is the bubble radius; The specific heat of air; The static pressure in the water surrounding the bubble; Let be the thermal damping constant of the raindrop; The wave function of the bubbly sound is obtained by simplification. : （2） The wave function of the integrated rain sound is obtained by integration. .

Images