Personalized customized sleep data monitoring method suitable for nighttime molar protection

Abstract

The invention relates to the technical field of sleep data monitoring, in particular to a personalized customized sleep data monitoring method suitable for nighttime molar protection, and the method comprises the following steps: constructing a high-frequency electroencephalogram characteristic factor based on amplitude disturbance, a sleep awakening tendency degree and a sleep ingot irregular condition; all peak values and all valley values in the low-frequency electroencephalogram signals are extracted, differences between the adjacent peak values and valley values are compared, and low-frequency electroencephalogram abnormality is determined; determining low-frequency fluctuation periodicity based on the similarity between different sub-low-frequency electroencephalogram signals and the abnormal condition of each sub-low-frequency electroencephalogram signal; and integrating the low-frequency electroencephalogram abnormality, the low-frequency fluctuation periodicity and the high-frequency electroencephalogram characteristic factors to obtain sleep damage coefficients of all light sleep stages, and monitoring night molar sleep data of the to-be-monitored patient. The invention aims to improve the accuracy of monitoring the night sleep data of the patient by using the electroencephalogram signal.

Description

Technical Field

[0001] The present application relates to the technical field of sleep data monitoring, and in particular to a sleep data monitoring method suitable for personalized customization of nighttime bruxism protection. Background Art

[0002] Nighttime bruxism is a non-physiological tooth contact condition, which refers to the involuntary contraction of the chewing muscles during sleep, causing the upper and lower teeth to clench or grind against each other rhythmically and intermittently, thereby interrupting the normal physiological resting position of the mandible. It is a sleep disorder. Personalized sleep data monitoring for nighttime bruxism protection can further provide scientific data support for the complete pathological mechanism of sleep bruxism and help relevant personnel better provide patients with personalized treatment plans.

[0003] Nighttime bruxism often occurs during the patient's light sleep period during the non-rapid eye movement (NREM) phase, causing damage to the teeth, periodontal tissues and temporomandibular joints, and affecting the patient's EEG signals during sleep. When using electroencephalogram (EEG) to analyze the sleep status of the monitored subject, the existing technology usually only analyzes based on the frequency and energy changes of the EEG, and is unable to extract and analyze the EEG signal change characteristics caused by the bruxism muscle activity according to the nighttime bruxism status of the monitored subject, so that the corresponding disturbance of the muscle behavior caused by bruxism in the EEG signal is affected, resulting in a high analysis error when using EEG signals to monitor the sleep data of nighttime bruxism subjects. Summary of the invention

[0004] In order to solve the above technical problems, the present application provides a sleep data monitoring method suitable for personalized customization of nighttime bruxism protection to solve the existing problems.

[0005] The sleep data monitoring method of the present application suitable for personalized customization of nighttime bruxism protection adopts the following technical solutions: An embodiment of the present application provides a sleep data monitoring method suitable for personalized customization of nighttime bruxism protection, the method comprising the following steps: Obtain high-frequency EEG signals and low-frequency EEG signals of any light sleep stage during the nighttime sleep of the patient to be monitored, as well as all sleep spindle signals in the high-frequency EEG signals; By analyzing the discrete degree of the amplitude of high-frequency EEG signals in the frequency domain in each light sleep stage, as well as the difference between the amplitude of high-frequency EEG signals and low-frequency EEG signals in the frequency domain, the degree of sleep arousal tendency in each light sleep stage can be determined; Determine the temporal disorder of all sleep spindle signals in the high-frequency EEG signal, as well as the temporal differences between different sleep spindle signals, so as to determine the irregularity of sleep spindles in each light sleep stage; For each sleep spindle signal in each light sleep stage, the amplitude disturbance of each sleep spindle signal in each light sleep stage is determined by analyzing the difference between its amplitude at the maximum frequency and the minimum frequency in the frequency domain, and the difference between its individual amplitudes in the frequency domain and the average distribution of all amplitudes in the frequency domain of the low-frequency EEG signal; The high-frequency EEG characteristic factors of each light sleep stage are obtained by integrating the amplitude disturbance, the degree of sleep-wake tendency and the irregularity of sleep. Extract all peaks and valleys in low-frequency EEG signals, and compare the differences between adjacent peaks and valleys to determine the abnormalities of low-frequency EEG in each light sleep stage; By analyzing the abnormalities of low-frequency EEG signals in each light sleep stage, the signals are divided into multiple sub-low-frequency EEG signals. Based on the similarities between different sub-low-frequency EEG signals and the abnormalities of each sub-low-frequency EEG signal, the periodicity of low-frequency fluctuations in each light sleep stage is determined. By integrating the low-frequency EEG abnormalities, low-frequency fluctuation periodicity and high-frequency EEG characteristic factors, the sleep impairment coefficient of each light sleep stage is obtained, and the nocturnal bruxism sleep data of the monitored patients is monitored.

[0006] Preferably, the method for determining the degree of sleep awakening tendency in each light sleep stage is: Calculate the variance of all amplitudes of high-frequency EEG signals in the frequency domain at each light sleep stage; Calculate the cumulative sum of the differences between all amplitudes of the high-frequency EEG signals in the frequency domain and all amplitudes of the low-frequency EEG signals in the frequency domain in each light sleep stage; The product of the variance and the cumulative sum is used as the degree of sleep awakening tendency in each light sleep stage.

[0007] Preferably, the method for determining the irregularity of the sleep state in each light sleep stage is: Obtain the generation time and disappearance time of each sleep spindle signal, and calculate the information entropy of the generation time of all sleep spindle signals in each light sleep stage; The time interval between the generation time and the disappearance time of each sleep spindle signal in each light sleep stage is recorded as the time interval of each sleep spindle signal in each light sleep stage; Calculate the cumulative sum of the time interval differences between different sleep spindle signals in each light sleep stage, and record it as the sleep difference sum value of each light sleep stage; The product of the sleep difference and value of all sleep spindle signals in each light sleep stage and the information entropy is taken as the sleep spindle irregularity status of each light sleep stage.

[0008] Preferably, the method for determining the amplitude disturbance of each sleep signal in each light sleep stage is: Calculate the difference between the amplitude of each sleep spindle signal at the maximum frequency and the amplitude at the minimum frequency in the frequency domain, and record it as the amplitude difference of each sleep spindle signal; Calculate the mean of all amplitudes in the frequency domain of low-frequency EEG signals in each light sleep stage, and record it as the mean low-frequency amplitude of each light sleep stage; Calculate the cumulative sum of the differences between all amplitudes of each sleep spindle signal in the frequency domain and the mean of all amplitudes of the low-frequency EEG signal in the frequency domain in each light sleep stage, and record it as the difference sum value of each sleep spindle signal in each light sleep stage; The inverse of the product of the amplitude difference of each sleep spindle signal in each light sleep stage and the difference sum value is taken as the amplitude disturbance susceptibility of each sleep spindle signal in each light sleep stage.

[0009] Preferably, the expression of the high-frequency EEG characteristic factor of each light sleep stage is: ; In the formula, Represents the high-frequency EEG characteristic factor of the i-th light sleep stage; Indicates the degree of sleep awakening tendency in the i-th light sleep stage; Indicates the irregularity of sleep in the i-th light sleep stage; represents the amplitude disturbance of the p-th sleep spindle signal in the i-th light sleep stage; represents the total number of sleep spindle signals in the i-th light sleep stage; norm() represents the normalization function.

[0010] Preferably, the method for determining the abnormality of low-frequency EEG in each light sleep stage is: Calculate the difference between each adjacent peak and valley value in the low-frequency EEG signal of each light sleep stage, and use the interquartile range of the difference between all adjacent peaks and valley values ​​as the peak-valley difference value of each light sleep stage; Abnormality of low-frequency EEG in the i-th light sleep stage The expression is: ; In the formula, represents the peak-to-valley difference value of the i-th light sleep stage; Represents the difference between the yth peak and its adjacent valley in the low-frequency EEG signal of the i-th light sleep stage; represents the number of all peaks in the low-frequency EEG signal of the i-th light sleep stage; exp( ) represents an exponential function with a natural constant as the base.

[0011] Preferably, the method analyzes the abnormal conditions of low-frequency EEG signals in each light sleep stage and divides them into a plurality of sub-low-frequency EEG signals, including: The low-frequency EEG signals in each light sleep stage are used as the input of the anomaly detection algorithm, all abnormal points are output, and the low-frequency EEG signals between every two adjacent abnormal points and the two between them are recorded as sub-low-frequency EEG signals.

[0012] Preferably, the expression of the periodicity of the low-frequency fluctuations in each light sleep stage is: ; In the formula, represents the periodicity of low-frequency fluctuations in the i-th light sleep stage; represents the similarity between the mth sub-low-frequency EEG signal and the nth sub-low-frequency EEG signal in the i-th light sleep stage; It represents the difference between the abnormal probabilities of two abnormal points in the m sub-low-frequency EEG signals in the i-th light sleep stage obtained by the abnormality detection algorithm; represents the number of all sub-low-frequency EEG signals in the i-th light sleep stage; Indicates a preset constant greater than 0.

[0013] Preferably, the method for determining the sleep impairment coefficient of each light sleep stage is: Calculate the ratio of low-frequency EEG abnormality to low-frequency fluctuation periodicity in each light sleep stage, and record it as the low-frequency EEG characteristic factor in each light sleep stage; The sleep impairment coefficient of each light sleep stage is the normalized value of the product of the high-frequency EEG characteristic factor and the low-frequency EEG characteristic factor of each light sleep stage.

[0014] Preferably, the monitoring of the nocturnal bruxism sleep data of the patient to be monitored includes: The sleep impairment coefficients of multiple patients in each light sleep stage are obtained, and the sleep impairment coefficients of all light sleep stages of each patient are combined into a sleep feature vector, the sleep feature vectors of all patients are marked according to the sleep quality standard, all the marked sleep feature vectors are used as the input of the support vector machine, and the trained support vector machine model is output, the sleep feature vectors of the patient to be monitored are used as the input of the support vector machine model, and the sleep quality classification result of the patient to be monitored is output.

[0015] This application has at least the following beneficial effects: The present application constructs the degree of sleep-wake tendency by analyzing the change amplitude of high-frequency EEG signals and the difference between high-frequency EEG signals and low-frequency EEG signals, which is helpful to evaluate the stability of brain activity in the light sleep stage, judge whether the patient has significant high-frequency EEG signal behavior due to teeth grinding behavior, and reflect the degree of sleep-wake tendency of the patient; secondly, by analyzing the difference in the temporal distribution of sleep spindle signals in high-frequency EEG signals, the irregularity of sleep spindles is constructed, which is helpful to judge the stability of the duration of the patient's sleep spindle signals and reflect the degree of damage to the characteristics of the sleep spindle signals; then, by analyzing the difference between the sleep spindle signals and the low-frequency EEG signals, the amplitude disturbance is constructed, which is helpful to accurately judge whether the sleep spindle signals are affected by teeth grinding behavior, reduce the error of using EEG signals to monitor the sleep data of subjects who grind their teeth at night, and improve the accuracy of monitoring; further, by comprehensively considering the amplitude disturbance, the degree of sleep-wake tendency and the irregularity of sleep spindles, the high-frequency EEG characteristic factors of each light sleep stage are obtained, which reflects the characteristics of the change of high-frequency EEG signals caused by teeth grinding in patients during the light sleep period; The present application determines the abnormality of low-frequency EEG by analyzing the fluctuation of low-frequency EEG signals, which is helpful to judge the fluctuation size of low-frequency EEG signals during the patient's light sleep period and reflect the degree of disturbance of the patient's sleep state; further, by analyzing the abnormality and periodic changes of low-frequency EEG signals, the periodicity of low-frequency fluctuation is constructed, which is helpful to distinguish the influence of nocturnal bruxism and nocturnal snoring on the changes of low-frequency EEG signals during the patient's light sleep period, eliminates the interference of nocturnal bruxism monitoring caused by nocturnal snoring, and improves the accuracy of monitoring the patient's night sleep data using EEG signals; the present application constructs a sleep impairment coefficient by comprehensively considering low-frequency EEG abnormalities, low-frequency fluctuation periodicity and high-frequency EEG characteristic factors, which is used to monitor the patient's sleep quality. According to the patient's sleep quality impairment, the risk of damage to the patient's teeth and sleep quality is reduced, thereby improving the accuracy of monitoring the patient's night sleep data using EEG signals. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present application or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0017] Figure 1 A flowchart of the steps of a sleep data monitoring method for personalized nighttime bruxism protection provided in one embodiment of the present application; Figure 2 A schematic diagram of a sleep impairment coefficient extraction process provided by an embodiment of the present application; Figure 3A schematic diagram of a sleep quality assessment process for a patient to be monitored provided in one embodiment of the present application. DETAILED DESCRIPTION

[0018] In order to further explain the technical means and effects adopted by the present application to achieve the predetermined invention purpose, the specific implementation method, structure, characteristics and effects of the sleep data monitoring method suitable for personalized nighttime bruxism protection proposed in accordance with the present application are described in detail below in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0019] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs.

[0020] The specific scheme of the sleep data monitoring method suitable for personalized customization of nighttime bruxism protection provided by the present application is described in detail below with reference to the accompanying drawings.

[0021] An embodiment of the present application provides a sleep data monitoring method suitable for personalized customization of nighttime bruxism protection. Specifically, the following sleep data monitoring method suitable for personalized customization of nighttime bruxism protection is provided. Figure 1 , the method comprises the following steps: Step S1: Acquire the high-frequency EEG signals and low-frequency EEG signals of any light sleep stage during the patient's sleep at night, as well as all sleep induration signals in the high-frequency EEG signals.

[0022] The EEG signals of the patient during sleep at night are obtained through smart wearable devices, and the EEG signals of the patient during sleep at night are amplified and denoised through signal amplifiers and adaptive filters.

[0023] The EEG signals of the falling asleep, light sleep, deep sleep, deep sleep and rapid eye movement (REM) periods in the non-rapid eye movement (NREM) period of a large number of patients with nocturnal bruxism were obtained through a polysomnography (PSG) system, and all the EEG signals were used as the input of a deep convolutional neural network (CNN), and the deep convolutional neural network was trained to obtain a trained neural network model; further, the EEG signals of the patients to be monitored during their night sleep were input into the trained neural network model, and the EEG signals of the patients to be monitored during the falling asleep, light sleep, deep sleep, deep sleep and rapid eye movement periods in the non-rapid eye movement (NREM) period of the night sleep were output, that is, the total EEG signals during the night sleep process were divided into EEG signals of different periods. When training the neural network model, cross entropy loss was used as the loss function, and Adam was used as the optimization algorithm.

[0024] Among them, the signal amplifier, adaptive filter and deep convolutional neural network are all well-known technologies, and their specific principles will not be repeated here.

[0025] Normally, nighttime sleep includes 4 to 5 sleep cycles, and the duration of NREM light sleep in each sleep cycle is 10 to 25 minutes. In this embodiment, the NREM light sleep EEG signal of every 20 minutes during the patient's nighttime sleep is regarded as a light sleep stage. Since nighttime sleep includes multiple sleep cycles, it also includes many light sleep stages. The high-frequency EEG signals (12 to 30 Hz) and low-frequency EEG signals (4 to 8 Hz) of the β wave in each light sleep stage of the patient to be monitored are obtained through high-pass filters and low-pass filters respectively. The high-frequency EEG signals of the β wave and the low-frequency EEG signals of the θ wave in each light sleep stage of the patient to be monitored are respectively used as the input of the fast Fourier transform (FFT) algorithm to obtain the representation of the high-frequency EEG waves in each light sleep stage in the frequency domain.

[0026] Among them, the principles of high-pass filter, low-pass filter and fast Fourier transform are all well-known technologies, and their respective principle processes are not described in detail again.

[0027] Step S2: The degree of sleep awakening tendency in each light sleep stage is determined by analyzing the discrete degree of the amplitude of the high-frequency EEG signal in the frequency domain and the difference between the amplitude of the high-frequency EEG signal and the low-frequency EEG signal in the frequency domain.

[0028] During the NREM light sleep period of the patient's sleep at night, the patient officially begins to sleep. At this time, the brain waves gradually become irregular, with frequency and amplitude fluctuating, and occasionally there are high-frequency, large-amplitude brain waves of sleep spindles and low-frequency, large-amplitude brain waves of K nodes; and the patient's nighttime teeth grinding will cause the chewing muscles to engage in tense activity, increase the patient's alertness and muscle tension, and trigger EEG activity in the motor area of ​​the cerebral cortex, which is specifically manifested in the patient's EEG signals during the light sleep period as follows: the difference between the amplitude of the high-frequency EEG signal and the amplitude of the low-frequency EEG signal becomes more and more significant, and the greater the difference between the amplitudes of the high-frequency EEG signal at different frequencies, at the same time, the patient is very likely to have a brief awakening during the teeth grinding process, resulting in an increase in local low-frequency fluctuations in the EEG signal.

[0029] Therefore, by analyzing the discrete degree of the amplitude of the high-frequency EEG signals in the frequency domain in each light sleep stage, as well as the difference between the amplitude of the high-frequency EEG signals and the low-frequency EEG signals in the frequency domain, the degree of sleep awakening tendency in each light sleep stage can be determined to judge whether the patient has bruxism. Specifically: Calculate the variance of all amplitudes of high-frequency EEG signals in the frequency domain at each light sleep stage; Furthermore, the cumulative sum of the differences between all amplitudes of the high-frequency EEG signals in the frequency domain and all amplitudes of the low-frequency EEG signals in the frequency domain in each light sleep stage is calculated.

[0030] Furthermore, the product of the variance and the cumulative sum is used as the degree of sleep awakening tendency in each light sleep stage.

[0031] According to the degree of sleep awakening tendency in each light sleep stage, it can be understood that when the amplitude variance of all frequency components in the high-frequency brain wave signal in the light sleep stage in the frequency domain is larger, it means that the instability of the patient's brain activity during the light sleep stage is stronger, and the brain experiences more frequent awakening and falling asleep during the light sleep stage; when the cumulative difference between the frequency amplitude of all high-frequency brain wave signals and the frequency amplitude of low-frequency brain wave signals in the light sleep stage is larger, it means that the high-frequency EEG signal behavior caused by the patient's teeth grinding behavior during the light sleep stage is more significant, and the degree of sleep awakening tendency caused by the patient's teeth grinding behavior at night during the light sleep stage is greater; On the contrary, when the amplitude variance of all frequency components in the high-frequency brain wave signal in the light sleep stage in the frequency domain is smaller, it means that the stability of the patient's brain activity during the light sleep stage is stronger, and the possibility of the brain experiencing awakening and falling asleep during the light sleep stage is smaller; when the cumulative difference between the frequency amplitude of all high-frequency brain wave signals and the frequency amplitude of low-frequency brain wave signals in the light sleep stage is smaller, it means that the possibility of the patient generating high-frequency EEG signal behavior due to teeth grinding behavior during the light sleep stage is smaller, and the patient's sleep awakening tendency caused by the patient's night teeth grinding behavior during the light sleep stage is smaller.

[0032] Step S3: Determine the temporal disorder degree of all sleep spindle signals in the high-frequency EEG signal, as well as the temporal difference between different sleep spindle signals, so as to determine the irregularity of sleep spindles in each light sleep stage.

[0033] Sleep spindles are a characteristic type of EEG activity that occurs during the NREM light sleep stage. They are spindle-shaped EEG waves with a higher frequency (12-16Hz) and a larger amplitude. They are a sign that sleep has entered the light sleep stage. Patients' nighttime bruxism is usually accompanied by strong muscle tension and is closely related to mental tension. Nighttime bruxism causes frequent contraction of the patient's masticatory muscles, reduces sleep stability, and may reduce or destruct the sleep spindle characteristics in the EEG signal. Specifically, the instability of the sleep spindle signal in the light sleep stage is enhanced, the amplitude of the sleep spindle signal is reduced, and the difference with the low-frequency EEG signal is small.

[0034] Therefore, by analyzing the temporal disorder of the sleep spindle signals and the temporal differences between different sleep spindle signals, the irregularity of the sleep spindles in each light sleep stage can be determined, so as to more accurately monitor whether the patient has teeth grinding during the light sleep period. Specifically: The high-frequency EEG signals of the patient in each light sleep stage are used as the input of the deep convolutional neural network, and the sleep signal in the high-frequency EEG signals is output. The cross entropy is used as the loss function and Adam is used as the optimization algorithm.

[0035] Further, the generation time and disappearance time of each sleep spindle signal in each light sleep stage are obtained, and the information entropy of the generation time of all sleep spindle signals in each light sleep stage is calculated; Among them, the calculation process of information entropy is a well-known technology, and its specific calculation principles and steps are not repeated here.

[0036] Further, the time interval between the generation time and the disappearance time of each sleep spindle signal in each light sleep stage is recorded as the time interval of each sleep spindle signal in each light sleep stage; Further, the cumulative sum of the time interval differences between different sleep spindle signals in each light sleep stage is calculated and recorded as the sleep difference sum value of each light sleep stage; It should be noted that there are many methods for measuring the differences between data. In this embodiment, the difference between the time intervals of different sleep spindle signals in each light sleep stage is measured by calculating the absolute value of the difference between the time intervals of different sleep spindle signals. The implementer may also use other methods for measuring the differences between data, such as the square or ratio of the difference. This embodiment does not impose any special restrictions on the selection of the method for measuring the differences between data.

[0037] Furthermore, the product of the sleep difference and value of all sleep spindle signals in each light sleep stage and the information entropy is taken as the sleep spindle irregularity condition of each light sleep stage.

[0038] According to the irregularity of sleep spindles in each light sleep stage, it can be understood that when the information entropy of the sleep spindle signal generation time in the light sleep stage is larger, it means that the patient's sleep entering the light sleep period in the light sleep stage is more irregular, and the patient's sleep spindle signal instability during NREM light sleep due to nighttime bruxism is stronger; when the cumulative difference between the time intervals of all sleep spindle signals in the light sleep stage and all other sleep spindle signals is larger, it means that the duration of the sleep spindle signal generated by the patient in the process of entering the light sleep stage is more unstable, the characteristics of the sleep spindle signal generated during the light sleep stage due to nighttime bruxism is more obviously damaged, and the irregularity of the sleep spindle is greater; On the contrary, when the information entropy of the sleep spindle signal in the light sleep stage is smaller, it means that the process of the patient entering the light sleep period in the light sleep stage is more regular, and the possibility of the patient grinding his teeth at night is smaller; when the cumulative difference between the time intervals of all sleep spindle signals in the light sleep stage and all other sleep spindle signals is smaller, the possibility of the patient grinding his teeth at night is smaller, and the final irregular sleep spindle condition is smaller.

[0039] Step S4: For each sleep signal in each light sleep stage, the amplitude susceptibility of each sleep signal in each light sleep stage is determined by analyzing the difference between its amplitude at the maximum frequency and the minimum frequency in the frequency domain, and the difference between its individual amplitudes in the frequency domain and the average distribution of all amplitudes in the frequency domain of low-frequency EEG signals.

[0040] Teeth grinding behavior is greatly affected by emotional factors such as stress and anxiety. When patients with nocturnal bruxism enter the light sleep stage, teeth grinding will not only aggravate the irregularity of the sleep spindle signals generated during the light sleep stage and the degree of disruption of their duration, but the influence of patients' emotional factors will also cause the amplitude of the sleep spindle signals in the light sleep stage to be lower, that is, the amplitude of the sleep spindle signals is smaller, making them more difficult to identify.

[0041] Therefore, by analyzing the difference between the amplitude of the sleep spindle signal at the maximum frequency and the minimum frequency in the frequency domain, as well as the difference between its individual amplitudes in the frequency domain and the average distribution of all amplitudes in the frequency domain of the low-frequency EEG signal, the amplitude disturbance of each sleep spindle signal in each light sleep stage is determined, so as to further judge the patient's teeth grinding during sleep, specifically: Calculate the difference between the amplitude of each sleep spindle signal at the maximum frequency and the amplitude at the minimum frequency in the frequency domain, and record it as the amplitude difference of each sleep spindle signal; Furthermore, the mean of all amplitudes in the frequency domain of the low-frequency EEG signal in each light sleep stage is calculated and recorded as the mean low-frequency amplitude of each light sleep stage; Further, the cumulative sum of the differences between all amplitudes of each sleep spindle signal in the frequency domain and the mean value of all amplitudes of the low-frequency EEG signal in the frequency domain in each light sleep stage is calculated, and recorded as the difference sum value of each sleep spindle signal in each light sleep stage; Furthermore, the inverse of the product of the amplitude difference of each sleep spindle signal in each light sleep stage and the difference sum value is taken as the amplitude disturbance susceptibility of each sleep spindle signal in each light sleep stage.

[0042] It should be noted that there are many methods to measure the difference between data. In this embodiment, the difference between the corresponding amplitudes at the maximum and minimum frequencies is measured by calculating the absolute value of the difference between the amplitude of each sleep signal at the maximum frequency in the frequency domain and the amplitude at the minimum frequency in the frequency domain; the difference between the amplitude of the sleep signal in the frequency domain and the low-frequency signal is measured by calculating the absolute value of the difference between all amplitudes of each sleep signal in the frequency domain in each light sleep stage and the average of all amplitudes of the low-frequency EEG signal in the frequency domain. The implementer may also adopt other methods of measuring the difference between data, such as the square or ratio of the difference, according to the specific situation. Regarding the selection of the method of measuring the difference between data, this embodiment is not free from special restrictions.

[0043] According to the amplitude disturbance of each sleep signal in each light sleep stage, it can be understood that when the difference between the amplitude at the maximum frequency and the amplitude at the minimum frequency in the sleep signal in the light sleep stage is smaller, it means that the patient's teeth grinding condition has caused more serious interference to the sleep signal during the light sleep period, and the smaller the amplitude difference at all frequencies of the sleep signal, the more difficult it is to identify; when the cumulative difference between the frequency amplitude of all signal components in the sleep signal in the light sleep stage and the mean amplitude of all frequencies of the low-frequency EEG signal is smaller, it means that the high-frequency sleep signal generated by the patient during the light sleep period is lighter, the sleep signal generated during the light sleep period is more disturbed by teeth grinding, and the amplitude disturbance of the sleep signal is more strongly affected; On the contrary, when the difference between the amplitude at the maximum frequency and the amplitude at the minimum frequency in the sleep spindle signal in the light sleep stage is greater, it means that the interference of the patient's teeth grinding condition on the sleep spindle signal during the light sleep period is less, that is, the probability of teeth grinding is smaller; when the cumulative difference between the frequency amplitude of all signal components in the sleep spindle signal in the light sleep stage and the mean amplitude of all frequencies of the low-frequency EEG signal is greater, it means that the high-frequency sleep spindle signal generated by the patient in the light sleep period is more serious and the amplitude of the sleep spindle signal is less disturbed.

[0044] Step S5: Comprehensively consider the amplitude disturbance, the degree of sleep-wake tendency and the irregularity of sleep, and obtain the high-frequency EEG characteristic factors of each light sleep stage.

[0045] As a sleep disorder, the EEG signals of patients with nocturnal bruxism during light sleep not only show unstable brain activity and a greater tendency to wake from sleep, but also interfere with and destroy the characteristic sleep spindle signals during light sleep. In order to characterize the characteristics of high-frequency EEG signal changes caused by bruxism in patients during light sleep, this embodiment determines the high-frequency EEG characteristic factors of each light sleep stage by analyzing the amplitude disturbance of all sleep spindle signals in each light sleep stage, the degree of sleep wakefulness tendency and the irregularity of the sleep spindles. The specific process is as follows: High-frequency EEG characteristic factors in the i-th light sleep stage The expression is: ; In the formula, Indicates the degree of sleep awakening tendency in the i-th light sleep stage; Indicates the irregularity of sleep in the i-th light sleep stage; represents the amplitude disturbance of the p-th sleep spindle signal in the i-th light sleep stage; represents the total number of sleep spindle signals in the i-th light sleep stage; norm() represents the normalization function.

[0046] According to the high-frequency EEG characteristic factors of each light sleep stage, it can be understood that when the degree of sleep awakening tendency in the light sleep stage is greater, it means that the high-frequency EEG signal behavior generated by the patient's teeth grinding behavior during the light sleep stage is more significant, and the patient experiences awakening and falling asleep more frequently during the light sleep stage; when the sleep spindle irregularity in the light sleep stage is greater, and the amplitude interference of all sleep spindle signals in the light sleep stage is stronger, it means that the patient's sleep spindle signal interference caused by teeth grinding during the light sleep stage is more obvious, the amplitude of the sleep spindle signal decreases more seriously, and it is more difficult to identify, and the high-frequency EEG characteristic factor is larger; On the contrary, the smaller the degree of sleep awakening tendency in the light sleep stage, the less significant the high-frequency EEG signal behavior caused by the patient's teeth grinding behavior during the light sleep stage; when the irregularity of sleep spindles in the light sleep stage is smaller, and the amplitude interference of all sleep spindle signals in the light sleep stage is weaker, it means that the interference phenomenon of the patient's sleep spindle signals caused by teeth grinding during the light sleep stage is less obvious, the phenomenon of the decrease in the amplitude of the sleep spindle signals is milder, and the high-frequency EEG characteristic factor is smaller.

[0047] Step S6: extract all peaks and all valleys in the low-frequency EEG signal, and compare the differences between adjacent peaks and valleys to determine the abnormality of the low-frequency EEG in each light sleep stage.

[0048] As a sleep disorder, during the NREM light sleep period of the patient's sleep at night, the tense activity of the chewing muscles caused by nocturnal bruxism will not only cause changes in the patient's high-frequency EEG signals, but also disturb the patient's sleep state during the light sleep period, making it difficult for the brain to maintain deep and stable slow-wave sleep, and the fluctuation of the θ wave low-frequency EEG signals (4~8Hz) will increase and the amplitude will decrease.

[0049] Based on the above analysis, the peak values ​​and valley values ​​in the low-frequency EEG signals of each light sleep stage are extracted. There are many commonly used peak extraction algorithms. In this embodiment, the AMPD (Automatic multiscale-based peak detection) peak detection algorithm is used to obtain all the peak values ​​and valley values ​​of the low-frequency EEG signals. The implementer may also use other peak extraction algorithms such as the derivative method. Regarding the selection of the peak extraction algorithm, this embodiment does not impose any special restrictions. The AMPD (Automatic multiscale-based peak detection) peak detection algorithm is a well-known technology, and its specific principle and process of extracting the peak value will not be repeated.

[0050] Furthermore, the low-frequency EEG abnormality is calculated for any light sleep stage to characterize the changes in low-frequency EEG signals due to bruxism, specifically: Calculate the difference between each adjacent peak and valley value in the low-frequency EEG signal of each light sleep stage, and use the interquartile range of the difference between all adjacent peaks and valley values ​​as the peak-valley difference value of each light sleep stage; Abnormality of low-frequency EEG in the i-th light sleep stage The expression is: ; In the formula, represents the peak-to-valley difference value of the i-th light sleep stage; Represents the difference between the yth peak and its adjacent valley in the low-frequency EEG signal of the i-th light sleep stage; represents the number of all peaks in the low-frequency EEG signal of the i-th light sleep stage; exp( ) represents an exponential function with a natural constant as the base.

[0051] The calculation process of the interquartile range is a well-known technique, and the specific calculation principle and steps will not be repeated here.

[0052] According to the abnormality of low-frequency EEG in each light sleep stage, it can be understood that when the interquartile range of the amplitude difference between adjacent peaks and troughs of low-frequency EEG signals in the frequency domain in the light sleep stage is larger, it means that the fluctuation amplitude change of the low-frequency EEG signals of the patient in the light sleep stage is more unstable, and the fluctuation of the low-frequency EEG signals in the light sleep stage of the patient due to nighttime bruxism is larger; when the cumulative difference between the amplitudes of adjacent peaks and troughs of each low-frequency EEG signal in the light sleep stage is larger, it means that the amplitude reduction of the low-frequency EEG signals caused by nighttime bruxism in the light sleep stage of the patient is more significant, and the abnormality of low-frequency EEG in the light sleep stage is greater; On the contrary, when the interquartile range of the amplitude difference between adjacent peaks and troughs of low-frequency EEG signals in the frequency domain in the light sleep stage is smaller, it means that the fluctuation amplitude change of the patient's low-frequency EEG signals in the light sleep stage is more stable, and the volatility of the patient's low-frequency EEG signals in the light sleep stage due to nocturnal bruxism is smaller; when the cumulative difference between the amplitudes of adjacent peaks and troughs of each low-frequency EEG signal in the light sleep stage is smaller, it means that the reduction in the amplitude of the low-frequency EEG signals caused by nocturnal bruxism in the patient's light sleep stage is less obvious, and the low-frequency EEG abnormality in the light sleep stage is smaller.

[0053] Step S7: By analyzing the abnormal conditions of low-frequency EEG signals in each light sleep stage, the signals are divided into multiple sub-low-frequency EEG signals, and based on the similarities between different sub-low-frequency EEG signals and the abnormal conditions of each sub-low-frequency EEG signal, the periodicity of low-frequency fluctuations in each light sleep stage is determined.

[0054] The following disadvantages exist in evaluating the increase in fluctuations and decrease in amplitude of low-frequency EEG signals during the patient's light sleep period due to nighttime teeth grinding only through low-frequency EEG abnormalities, namely, the patient's snoring at night can cause airway obstruction and cause awakening reactions during sleep, which can also cause the patient's low-frequency EEG signal fluctuations to increase and the amplitude to decrease during light sleep. However, snoring at night is often accompanied by periodic fluctuations, while teeth grinding is usually an irregular behavior, that is, the periodicity of the changes in the patient's low-frequency EEG signals during light sleep caused by nighttime teeth grinding is lower than that caused by snoring at night.

[0055] Therefore, based on the above analysis, the low-frequency EEG signals of each light sleep stage are used as the input of the anomaly detection algorithm, all abnormal points and their respective abnormal probabilities are output, and the low-frequency EEG signals between every two adjacent abnormal points and the two between them are recorded as sub-low-frequency EEG signals.

[0056] It should be noted that there are many commonly used anomaly detection algorithms. In this embodiment, the Bayesian change point detection algorithm is used to obtain the various abnormal points in the low-frequency EEG signal and the abnormal probability of the abnormal points. The implementer may also use other anomaly detection methods such as the LOF anomaly detection algorithm. Regarding the selection of the anomaly detection algorithm, this embodiment does not make any special restrictions.

[0057] Among them, the Bayesian substation detection algorithm is a well-known technology, and its specific principle will not be repeated here.

[0058] In particular, the low-frequency EEG signals before the first abnormal point and the low-frequency EEG signals after the last abnormal point are not analyzed.

[0059] Furthermore, based on the similarity between different sub-low-frequency EEG signals and the abnormality of each sub-low-frequency EEG signal, the periodicity of low-frequency fluctuations in each light sleep stage is determined, specifically: The periodicity of low-frequency fluctuations in the i-th light sleep stage The expression is: ; In the formula, represents the similarity between the mth sub-low-frequency EEG signal and the nth sub-low-frequency EEG signal in the i-th light sleep stage; It represents the difference between the abnormal probabilities of two abnormal points in the m sub-low-frequency EEG signals in the i-th light sleep stage obtained by the abnormality detection algorithm; represents the number of all sub-low-frequency EEG signals in the i-th light sleep stage; Indicates a preset constant greater than 0, which is used to prevent the denominator from being 0. The value of is 0.01. Under the premise of ensuring that the denominator is not 0 and does not excessively affect the calculation result, the implementer can also set it according to the specific situation. This embodiment does not impose any special restrictions.

[0060] It should be noted that there are many methods for measuring the similarity between two signals. In this embodiment, the Jaccard similarity coefficient between the two sub-low-frequency EEG signals is calculated to measure the similarity between the two sub-low-frequency EEG signals. The implementer may also use other methods such as cosine similarity to measure the similarity between signals. This embodiment does not impose any special restrictions on the selection of methods for measuring the similarity between signals.

[0061] The calculation process of the Jaccard similarity coefficient is a well-known technique, and the specific calculation steps are not repeated here.

[0062] According to the periodicity of low-frequency fluctuations in each light sleep stage, it can be understood that when the difference in the abnormal probability between the two abnormal points corresponding to the sub-low-frequency EEG signals in the light sleep stage is smaller, it means that the abnormal condition of the low-frequency EEG signals in the light sleep stage of the patient is more prominent, and the changes in the low-frequency EEG signals in the light sleep stage of the patient are more likely to be caused by periodic snoring at night rather than irregular nighttime bruxism; when the cumulative result of the similarity between each sub-low-frequency EEG signal and the other sub-low-frequency EEG signals in the light sleep stage is larger, it means that the difference between each sub-low-frequency EEG signal in the light sleep stage of the patient is smaller, the synchronization of the changes in the low-frequency EEG signals caused by snoring at night during the sleep process of the patient is more significant, and the periodicity of the low-frequency fluctuations in the light sleep stage of the patient is larger, that is, the stronger the periodicity of the changes in the low-frequency EEG signals in the light sleep stage of the patient is, the more likely the changes in the low-frequency EEG signals in the light sleep stage of the patient are caused by periodic snoring at night rather than irregular nighttime bruxism; On the contrary, if the abnormal probability difference between the two abnormal points corresponding to the sub-low-frequency EEG signals in the light sleep stage is smaller, the greater the cumulative similarity between each sub-low-frequency EEG signal and the other sub-low-frequency EEG signals in the light sleep stage, the smaller the periodicity of the low-frequency fluctuation, and the weaker the periodicity of the low-frequency EEG signal changes in the patient's light sleep period, the more likely the patient will have irregular nocturnal bruxism behavior during sleep at night.

[0063] Preferably, the sleep impairment coefficient extraction process diagram provided in this embodiment is as follows: Figure 2 shown.

[0064] Step S8: The sleep impairment coefficient of each light sleep stage is obtained by combining the low-frequency EEG abnormality, low-frequency fluctuation periodicity and high-frequency EEG characteristic factors.

[0065] The abnormalities of low-frequency EEG and the periodicity of low-frequency fluctuations in the patient's light sleep EEG signals can be used to accurately assess the increase in low-frequency EEG signal fluctuations and decrease in amplitude caused by nocturnal bruxism, thereby reflecting the changing characteristics of the patient's light sleep low-frequency EEG signals affected by bruxism.

[0066] Based on the above analysis, the low-frequency EEG characteristic factor of any light sleep stage can be calculated as follows: The ratio of the low-frequency EEG abnormality to the low-frequency fluctuation periodicity in each light sleep stage was calculated and recorded as the low-frequency EEG characteristic factor in each light sleep stage.

[0067] According to the low-frequency EEG characteristic factor, it can be understood that the higher the low-frequency EEG abnormality in the light sleep stage, the smaller the periodicity of the low-frequency fluctuation is, which means that the more obvious the increase in low-frequency EEG signal fluctuation and the decrease in amplitude due to nocturnal bruxism during the patient's light sleep period, the more serious the disturbance of the patient's sleep quality caused by nocturnal bruxism during the patient's light sleep period, and the larger the low-frequency EEG characteristic factor; conversely, the lower the low-frequency EEG abnormality in the light sleep stage, the larger the periodicity of the low-frequency fluctuation is, which means that the less obvious the increase in low-frequency EEG signal fluctuation and the decrease in amplitude due to nocturnal bruxism during the patient's light sleep period, and the smaller the low-frequency EEG characteristic factor.

[0068] Furthermore, based on the comprehensive interference of the patient's high-frequency EEG signals during light sleep with nocturnal bruxism and the changing characteristics of the patient's low-frequency EEG signals, the impaired sleep quality caused by nocturnal bruxism during light sleep can be evaluated, which can be used to assist relevant personnel in personalizing the treatment plan and subsequent improvement of the patient's nocturnal bruxism based on the patient's impaired sleep quality.

[0069] Based on the above analysis, the sleep impairment coefficient of any light sleep stage can be determined by the high-frequency EEG characteristic factor and the low-frequency EEG characteristic factor of the light sleep stage. Specifically: The sleep impairment coefficient of each light sleep stage is the normalized value of the product of the high-frequency EEG characteristic factor and the low-frequency EEG characteristic factor of each light sleep stage.

[0070] According to the sleep impairment coefficient of each light sleep stage, it can be understood that when the patient's sleep quality during light sleep is more affected by nocturnal teeth grinding, the patient's tendency to awaken during light sleep is higher, the more significant the high-frequency EEG signal generation behavior during light sleep is more significant, the more significant the increase in low-frequency EEG signal fluctuations and decrease in amplitude caused by nocturnal teeth grinding, the greater the patient's sleep impairment coefficient; conversely, when the patient's sleep quality during light sleep is less affected by nocturnal teeth grinding, the patient's tendency to awaken during light sleep is lower, the high-frequency EEG signal generation behavior during light sleep is less significant, and the patient's sleep impairment coefficient is smaller.

[0071] Step S9: Based on the sleep impairment coefficient of each light sleep stage, the patient's nocturnal bruxism sleep data is monitored.

[0072] The higher the sleep impairment coefficient, the worse the patient's sleep quality is affected by nighttime bruxism, the greater the patient's tendency to wake up during light sleep, and the less effective sleep they get. Relevant personnel can evaluate the patient's sleep status based on the size or change characteristics of the sleep impairment coefficient at each stage of the patient's light sleep, specifically: Obtain the patient's EEG signal data under different standard sleep qualities, and obtain the sleep impairment coefficient of each light sleep stage during the patient's light sleep period through the above method, record the vector composed of the sleep impairment coefficients corresponding to all light sleep stages of the patient as a sleep feature vector, and mark the sleep feature vector according to the standard sleep quality, and the marking content is high-quality sleep quality, good sleep quality, and poor sleep quality. Use all the marked sleep feature vectors as the input of the SVM support vector machine to train the SVM classification model, and use the trained SVM model as the patient's nocturnal bruxism sleep monitoring classification model. Since the training of the SVM support vector machine is a well-known technology, the specific acquisition process will not be described in detail.

[0073] Furthermore, the sleep feature vector composed of the sleep impairment coefficients corresponding to all light sleep stages of the monitored patient during the light sleep period is used as the input of the trained SVM classification model to obtain the patient's sleep quality classification result.

[0074] Among them, the SVM support vector machine and the process of training the SVM model are well-known technologies, and the specific principle and process are not repeated here.

[0075] Preferably, the sleep quality assessment process diagram of the patient to be monitored provided in this embodiment is as follows: Figure 3 shown.

[0076] It should be noted that the method for obtaining the sleep impairment coefficient and the sleep characteristic vector of the multiple patients and the patient to be monitored in each light sleep stage is the same, and all follow the method for obtaining the sleep impairment coefficient and the sleep characteristic vector in steps S1-S9.

[0077] So far, this embodiment comprehensively analyzes the degree of high-frequency EEG signal generation caused by teeth grinding behavior in the patient's light sleep stage and the interference status of sleep spindle characteristic signals, and accurately evaluates the changes in high-frequency EEG signals in the patient's light sleep period caused by nighttime teeth grinding; obtains low-frequency EEG signal characteristic factors based on the characteristics of low-frequency EEG signal changes caused by nighttime teeth grinding in the patient's light sleep period, and distinguishes between nighttime snoring and nighttime teeth grinding that cause changes in low-frequency EEG signals in the patient's light sleep period, and further considers the increase in fluctuation and decrease in amplitude of low-frequency EEG signals in the patient's light sleep period; obtains the sleep impairment coefficient through high-frequency EEG characteristic factors and low-frequency EEG characteristic factors to evaluate the patient's sleep quality, and at the same time can help relevant personnel accurately grasp the patient's nighttime teeth grinding status according to the characteristics of light sleep EEG signals, and provide scientific data support for subsequent personalized customization of nighttime teeth grinding protection measures, which can improve the accuracy of nighttime sleep quality monitoring using EEG signals while reducing the risk of damage to patients' teeth and sleep quality.

[0078] It should be noted that the above sequence of the embodiments of the present application is for description only and does not represent the advantages and disadvantages of the embodiments. The above is a description of a specific embodiment of this specification. In addition, the processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0079] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.

[0080] The embodiments described above are only used to illustrate the technical solutions of the present application, rather than to limit them. Modifications to the technical solutions recorded in the aforementioned embodiments, or equivalent replacement of some of the technical features therein, do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present application, and should all be included in the protection scope of the present application.

Claims

1. A sleep data monitoring method suitable for personalized nighttime bruxism protection, characterized in that: The method comprises the following steps: Obtain high-frequency EEG signals and low-frequency EEG signals of any light sleep stage during the nighttime sleep of the patient to be monitored, as well as all sleep spindle signals in the high-frequency EEG signals; By analyzing the discrete degree of the amplitude of high-frequency EEG signals in the frequency domain in each light sleep stage, as well as the difference between the amplitude of high-frequency EEG signals and low-frequency EEG signals in the frequency domain, the degree of sleep arousal tendency in each light sleep stage can be determined; Determine the temporal disorder of all sleep spindle signals in the high-frequency EEG signal, as well as the temporal differences between different sleep spindle signals, so as to determine the irregularity of sleep spindles in each light sleep stage; For each sleep spindle signal in each light sleep stage, the amplitude disturbance of each sleep spindle signal in each light sleep stage is determined by analyzing the difference between its amplitude at the maximum frequency and the minimum frequency in the frequency domain, and the difference between its individual amplitudes in the frequency domain and the average distribution of all amplitudes in the frequency domain of the low-frequency EEG signal; The high-frequency EEG characteristic factors of each light sleep stage are obtained by integrating the amplitude disturbance, the degree of sleep-wake tendency and the irregularity of sleep. Extract all peaks and valleys in low-frequency EEG signals, and compare the differences between adjacent peaks and valleys to determine the abnormalities of low-frequency EEG in each light sleep stage; By analyzing the abnormalities of low-frequency EEG signals in each light sleep stage, the signals are divided into multiple sub-low-frequency EEG signals. Based on the similarities between different sub-low-frequency EEG signals and the abnormalities of each sub-low-frequency EEG signal, the periodicity of low-frequency fluctuations in each light sleep stage is determined. By integrating the low-frequency EEG abnormalities, low-frequency fluctuation periodicity and high-frequency EEG characteristic factors, the sleep impairment coefficient of each light sleep stage is obtained, and the nocturnal bruxism sleep data of the monitored patients is monitored.

2. The sleep data monitoring method suitable for personalized customization of nighttime bruxism protection according to claim 1, characterized in that: The method for determining the degree of sleep awakening tendency in each light sleep stage is as follows: Calculate the variance of all amplitudes of high-frequency EEG signals in the frequency domain at each light sleep stage; Calculate the cumulative sum of the differences between all amplitudes of the high-frequency EEG signals in the frequency domain and all amplitudes of the low-frequency EEG signals in the frequency domain in each light sleep stage; The product of the variance and the cumulative sum is used as the degree of sleep awakening tendency in each light sleep stage.

3. The sleep data monitoring method suitable for personalized customization of nighttime bruxism protection according to claim 1, characterized in that: The method for determining the irregular sleep state of each light sleep stage is as follows: Obtain the generation time and disappearance time of each sleep spindle signal, and calculate the information entropy of the generation time of all sleep spindle signals in each light sleep stage; The time interval between the generation time and the disappearance time of each sleep spindle signal in each light sleep stage is recorded as the time interval of each sleep spindle signal in each light sleep stage; Calculate the cumulative sum of the time interval differences between different sleep spindle signals in each light sleep stage, and record it as the sleep difference sum value of each light sleep stage; The product of the sleep difference and value of all sleep spindle signals in each light sleep stage and the information entropy is taken as the sleep spindle irregularity status of each light sleep stage.

4. The sleep data monitoring method suitable for personalized customization of nighttime bruxism protection according to claim 1, characterized in that: The method for determining the amplitude disturbance of each sleep signal in each light sleep stage is as follows: Calculate the difference between the amplitude of each sleep spindle signal at the maximum frequency and the amplitude at the minimum frequency in the frequency domain, and record it as the amplitude difference of each sleep spindle signal; Calculate the mean of all amplitudes in the frequency domain of low-frequency EEG signals in each light sleep stage, and record it as the mean low-frequency amplitude of each light sleep stage; Calculate the cumulative sum of the differences between all amplitudes of each sleep spindle signal in the frequency domain and the mean of all amplitudes of the low-frequency EEG signal in the frequency domain in each light sleep stage, and record it as the difference sum value of each sleep spindle signal in each light sleep stage; The inverse of the product of the amplitude difference of each sleep spindle signal in each light sleep stage and the difference sum value is taken as the amplitude disturbance susceptibility of each sleep spindle signal in each light sleep stage.

5. The sleep data monitoring method suitable for personalized customization of nighttime bruxism protection according to claim 1, characterized in that: The expression of the high-frequency EEG characteristic factor of each light sleep stage is: ; In the formula, Represents the high-frequency EEG characteristic factor of the i-th light sleep stage; Indicates the degree of sleep awakening tendency in the i-th light sleep stage; Indicates the irregularity of sleep in the i-th light sleep stage; represents the amplitude disturbance of the p-th sleep spindle signal in the i-th light sleep stage; represents the total number of sleep spindle signals in the i-th light sleep stage; norm() represents the normalization function.

6. The sleep data monitoring method suitable for personalized customization of nighttime bruxism protection according to claim 1, characterized in that: The method for determining the abnormality of low-frequency EEG in each light sleep stage is as follows: Calculate the difference between each adjacent peak and valley value in the low-frequency EEG signal of each light sleep stage, and use the interquartile range of the difference between all adjacent peaks and valley values ​​as the peak-valley difference value of each light sleep stage; Abnormality of low-frequency EEG in the i-th light sleep stage The expression is: ; In the formula, represents the peak-to-valley difference value of the i-th light sleep stage; Represents the difference between the yth peak and its adjacent valley in the low-frequency EEG signal of the i-th light sleep stage; represents the number of all peaks in the low-frequency EEG signal of the i-th light sleep stage; exp( ) represents an exponential function with a natural constant as the base.

7. The sleep data monitoring method suitable for personalized customization of nighttime bruxism protection according to claim 1, characterized in that: By analyzing the abnormal conditions of low-frequency EEG signals in each light sleep stage, the signals are divided into multiple sub-low-frequency EEG signals, including: The low-frequency EEG signals in each light sleep stage are used as the input of the anomaly detection algorithm, all abnormal points are output, and the low-frequency EEG signals between every two adjacent abnormal points and the two between them are recorded as sub-low-frequency EEG signals.

8. The sleep data monitoring method suitable for personalized customization of nighttime bruxism protection according to claim 7, characterized in that: The expression of the periodicity of low-frequency fluctuations in each light sleep stage is: ; In the formula, represents the periodicity of low-frequency fluctuations in the i-th light sleep stage; represents the similarity between the mth sub-low-frequency EEG signal and the nth sub-low-frequency EEG signal in the i-th light sleep stage; It represents the difference between the abnormal probabilities of two abnormal points in the m sub-low-frequency EEG signals in the i-th light sleep stage obtained by the abnormality detection algorithm; represents the number of all sub-low-frequency EEG signals in the i-th light sleep stage; Indicates a preset constant greater than 0.

9. The sleep data monitoring method suitable for personalized customization of nighttime bruxism protection according to claim 1, characterized in that: The method for determining the sleep impairment coefficient of each light sleep stage is: Calculate the ratio of low-frequency EEG abnormality to low-frequency fluctuation periodicity in each light sleep stage, and record it as the low-frequency EEG characteristic factor in each light sleep stage; The sleep impairment coefficient of each light sleep stage is the normalized value of the product of the high-frequency EEG characteristic factor and the low-frequency EEG characteristic factor of each light sleep stage.

10. The sleep data monitoring method suitable for personalized customization of nighttime bruxism protection according to claim 1, characterized in that: The monitoring of the nighttime teeth grinding sleep data of the patient to be monitored includes: The sleep impairment coefficients of multiple patients in each light sleep stage are obtained, and the sleep impairment coefficients of all light sleep stages of each patient are combined into a sleep feature vector, the sleep feature vectors of all patients are marked according to the sleep quality standard, all the marked sleep feature vectors are used as the input of the support vector machine, and the trained support vector machine model is output, the sleep feature vectors of the patient to be monitored are used as the input of the support vector machine model, and the sleep quality classification result of the patient to be monitored is output.

Images