Breathing state determination method and device, computer equipment and storage medium

Abstract

The invention relates to the technical field of computers, and discloses a breathing state determination method and device, computer equipment and a storage medium, according to the breathing state determination method and device, electroencephalogram signals, mouth and nose signals and abdomen displacement signals of a subject during sleep are obtained, and the breathing state, including the breathing depth, frequency, duration and the like, of the subject is evaluated more comprehensively; the sleep period signals in the electroencephalogram signals are recognized, interference of other noise signals is reduced, and the recognition efficiency is improved; on the basis, the sleep airflow signal and the sleep displacement signal corresponding to the mouth-nose airflow signal and the abdomen displacement signal are respectively determined according to the sleep period signal, and the breathing state of the subject is determined according to the sleep period signal and the mouth-nose airflow signal, so that the accuracy of the breathing state is improved, and the corresponding breathing state detection cost is reduced.

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and in particular to a method and device for determining a respiratory state, a computer device and a storage medium. Background Art

[0002] The importance of sleep apnea and hypopnea syndrome is introduced. In the prior art, oral and nasal airflow signals are usually used to evaluate the breathing status of the subject, and the breathing status of the subject is determined by comparing the subject's breathing frequency with the normal breathing frequency. In the prior art, sleep polysomnography is used to perform sleep detection, but this method has the disadvantage of high cost. Therefore, a new method for detecting breathing status is urgently needed. Summary of the invention

[0003] In view of this, the present invention provides a method, apparatus, computer device and storage medium for determining a respiratory state to solve the problem of high cost of existing respiratory state detection.

[0004] In a first aspect, the present invention provides a method for determining a respiratory state, the method comprising:

[0005] Obtaining the EEG signals, oral and nasal airflow signals, and abdominal displacement signals of the subject during sleep;

[0006] Identify sleep phase signals in EEG signals;

[0007] Determine, according to the sleep period signal, a sleep airflow signal and a sleep displacement signal corresponding to the oral and nasal airflow signal and the abdominal displacement signal, respectively;

[0008] The subject's breathing state is determined based on the sleep displacement signal and the oral and nasal airflow signal.

[0009] Beneficial effects: The EEG signals, oral and nasal signals, and abdominal displacement signals of the subject during sleep are obtained to more comprehensively evaluate the subject's respiratory state, including the depth, frequency, duration, etc. of breathing; the sleep period signals in the EEG signals are identified, the interference of other noise signals is reduced, and the recognition efficiency is improved; on this basis, according to the sleep period signals, the sleep airflow signals and sleep displacement signals corresponding to the oral and nasal airflow signals and the abdominal displacement signals are respectively determined, and the subject's respiratory state is determined according to the sleep period signals and the oral and nasal airflow signals, which improves the accuracy of the respiratory state and reduces the corresponding respiratory state detection cost.

[0010] In an optional implementation, identifying a sleep phase signal in an EEG signal specifically includes:

[0011] Extract the main characteristic information of EEG signals;

[0012] Input the main feature information into a preset model, classify the main feature information, and determine the sleep period characteristics and the wakefulness period characteristics;

[0013] The sleep period signal is determined according to the sleep period characteristics, the wakefulness period characteristics and the EEG signal.

[0014] In an optional implementation, extracting main feature information of the EEG signal specifically includes:

[0015] Preprocessing the EEG signal to obtain a clean EEG signal;

[0016] Extracting sample entropy of clean EEG signals and determining characteristic information of clean EEG signals;

[0017] The principal component analysis method is used to reduce the dimension of feature information and obtain the main feature information.

[0018] In an optional implementation, determining the sleep airflow signal and the sleep displacement signal corresponding to the oral and nasal airflow signal and the abdominal displacement signal respectively according to the sleep period signal specifically includes:

[0019] Determining a sleep airflow signal corresponding to the oral and nasal airflow signal according to time information of the sleep period signal;

[0020] A sleep displacement signal of the abdomen displacement signal is determined according to the time information of the sleep period signal.

[0021] In an optional embodiment, determining the breathing state of the subject according to the sleep displacement signal and the oral and nasal airflow signal specifically includes:

[0022] When there is an oronasal airflow signal and no sleep displacement signal, the respiratory state is determined to be obstructive apnea;

[0023] When there is no oral or nasal airflow signal and no sleep displacement signal, the respiratory state is determined to be central apnea;

[0024] When there is no oronasal airflow signal and sleep displacement signal in the first preset time period, and there is an oronasal airflow signal and sleep displacement signal in the second preset time period, the respiratory state is mixed apnea, and the first preset time period is before the second preset time period.

[0025] In an optional embodiment, the oral and nasal airflow signal is acquired by a thermal sensor and / or a flow sensor.

[0026] In an optional implementation, the EEG signal is preprocessed using wavelet transform to obtain a clean EEG signal.

[0027] In a second aspect, the present invention provides a respiratory state determination device, the device comprising:

[0028] A signal acquisition module is used to acquire the EEG signal, oral and nasal airflow signal, and abdominal displacement signal of the subject during sleep;

[0029] A sleep recognition module is used to recognize sleep phase signals in EEG signals;

[0030] A sleep signal determination module, used to determine a sleep airflow signal and a sleep displacement signal corresponding to the oral and nasal airflow signal and the abdominal displacement signal, respectively, according to the sleep period signal;

[0031] The respiratory state determination module is used to determine the respiratory state of the subject based on the sleep displacement signal and the oral and nasal airflow signal.

[0032] In a third aspect, the present invention provides a computer device, comprising: a memory and a processor, the memory and the processor being communicatively connected to each other, the memory storing computer instructions, and the processor executing the respiratory state determination method of the first aspect or any corresponding embodiment thereof by executing the computer instructions.

[0033] In a fourth aspect, the present invention provides a computer-readable storage medium having computer instructions stored thereon, the computer instructions being used to enable a computer to execute the respiratory state determination method of the first aspect or any corresponding embodiment thereof. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] In order to more clearly illustrate the specific implementation methods of the present invention or the technical solutions in the prior art, the drawings required for use in the specific implementation methods or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some implementation methods of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0035] Figure 1 is a flow chart of a method for determining a respiratory state according to an embodiment of the present invention;

[0036] Figure 2 is a flow chart of another method for determining a respiratory state according to an embodiment of the present invention;

[0037] Figure 3 is a schematic diagram of a method for determining a respiratory state according to an embodiment of the present invention;

[0038] Figure 4 is a schematic diagram of a method for determining a respiratory state according to an embodiment of the present invention;

[0039] Figure 5 is a schematic diagram of a method for determining a respiratory state according to an embodiment of the present invention;

[0040] Figure 6 is a schematic diagram of a method for determining a respiratory state according to an embodiment of the present invention;

[0041] Figure 7 is a schematic diagram of a method for determining a respiratory state according to an embodiment of the present invention;

[0042] Figure 8 is a schematic diagram of a method for determining a respiratory state according to an embodiment of the present invention;

[0043] Fig. 9 is a flow chart of a method for determining a respiratory state according to an embodiment of the present invention;

[0044] Fig.10 is a structural block diagram of a device for determining a respiratory state according to an embodiment of the present invention;

[0045] Fig.11 It is a schematic diagram of the hardware structure of a computer device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0046] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0047] The prevalence of sleep apnea and hypopnea syndrome in obese adults was investigated, and the results showed that the incidence rate was between 42% and 48% in men and between 8% and 38% in women. The data confirmed that the incidence rate was closely related to the patient's gender, and the survey showed that the incidence rate gradually increased with the patient's age. Therefore, it is of great significance to detect the disease as early as possible, diagnose the type of disease, and develop a dedicated treatment plan based on the patient's own condition and the severity of the disease.

[0048] In this embodiment, a method for determining a respiratory state is provided. Figure 1 is a flow chart of a method for determining a respiratory state according to an embodiment of the present invention. It should be noted that the steps shown in the flow chart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and, although a logical order is shown in the flow chart, in some cases, the steps shown or described can be executed in an order different from that here.

[0049] like Figure 1 As shown, the process of the respiratory state determination method includes the following steps:

[0050] Step S101, obtaining the EEG signal, oral and nasal airflow signal, and abdominal displacement signal of the subject during sleep.

[0051] Exemplarily, when the subject is sleeping, the EEG signal, oral and nasal airflow signal and abdominal displacement signal corresponding to the subject are obtained. Specifically, when obtaining the EEG signal, electrodes can be set on the subject's head so that the EEG signal of the subject during sleep can be obtained through the electrodes.

[0052] Breathing is the process of gas exchange between the human body and the external environment. The human body inhales oxygen and exhales carbon dioxide through breathing to provide energy for life activities. Oral and nasal airflow signals are often used to evaluate the breathing status of the subject. The normal breathing rate of the human body is about 12-20 times / minute, and the waveform is approximately a sine wave. The amplitude is similar and there will be no large fluctuations. If the amplitude or frequency of the oral and nasal airflow signal of the subject changes to a large extent, it can be judged that the subject is experiencing an abnormal breathing event.

[0053] Therefore, thermistor sensors and flow sensors can be used to collect the corresponding oral and nasal airflow signals of the subject. The principle of the thermistor sensor to measure oral and nasal airflow is: the exhaled gas comes from the lungs with a higher temperature, and the inhaled gas comes from the external environment with a lower temperature. The thermistor changes its resistance as the ambient temperature changes. By comparing the change in the resistance of the thermistor, the volume of the subject's inhaled and exhaled gas can be calculated. The principle of the flow sensor to measure oral and nasal airflow is: place the sensor in the subject's respiratory flow path. When the subject breathes, a gas pressure difference is generated on both sides of the sensor. The respiratory gas flow rate can be converted according to the following formula:

[0054]

[0055] Among them, q is the volume of the gas flowing through, Δp is the flow difference generated on both sides of the flow sensor, ρ is the density of the gas, and k is the coefficient, which is related to multiple parameters such as the sensor throttle diameter and expansion coefficient.

[0056] The abdominal displacement signal can be obtained by the following methods:

[0057] 1. Ultrasonic technology obtains abdominal displacement signals. Specifically, the ultrasonic sensor can be placed on the surface of the abdomen, and by receiving the reflected ultrasonic signal, the displacement change of the abdominal surface can be measured.

[0058] 2. Optical measurement method: Use a laser displacement sensor or camera to detect the displacement change of the abdominal surface. By projecting a laser beam or light onto the abdominal surface and measuring the change of the reflected beam or image, the displacement signal of the abdomen can be obtained.

[0059] 3. Mechanical measurement method: Use mechanical sensors, such as flexible resistors, inductive displacement sensors, etc., stick them on the surface of the abdomen, and obtain the displacement signal of the abdomen by measuring the change in the distance between the sensor and the surface of the abdomen.

[0060] 4. Bioelectric signal method: Obtain the displacement signal of the abdomen by measuring the electrical activity of the abdominal muscles. This method requires installing electrodes on the surface of the abdomen and recording the electrical signals of the muscles. By processing and analyzing these electrical signals, the displacement changes of the abdomen can be inferred.

[0061] Step S102, identifying sleep phase signals in the EEG signals;

[0062] For example, the EEG signal is the signal of the subject during the entire sleep process, which will inevitably introduce interference from physiological factors such as breathing, eye movement, and electromyography, as well as 50Hz industrial frequency interference, which needs to be filtered. In addition, the entire EEG signal also includes the signal before entering the sleep state (called the wakefulness period), so it is necessary to identify the corresponding sleep period signal among all the EEG signals.

[0063] Specifically, the following methods can be used to identify sleep phase signals in EEG signals:

[0064] 1) Manual interpretation: Professional doctors or technicians manually interpret EEG signals and judge the transition of sleep phases based on experience. This method requires manual participation and requires high professional knowledge and experience of the interpreter.

[0065] 2) Rule-based method: According to the characteristics of the sleep period, a series of rules are formulated, such as waveform characteristics, frequency characteristics, etc., and the EEG signals are matched according to the rules to identify the sleep period. This method is relatively simple, but there may be a certain misjudgment rate.

[0066] 3) Hybrid method: Combining model training and rule-based methods to improve the accuracy and stability of recognition. For example, you can first use a rule-based method for preliminary recognition, and then use the model for verification and adjustment.

[0067] It should be noted that different methods have their own advantages and disadvantages, and they need to be selected according to actual needs and conditions. At the same time, for accurate identification of sleep periods, factors such as individual differences and environmental interference need to be considered.

[0068] In a preferred embodiment, Figure 2 As shown, the above step S102 includes:

[0069] Step S1021, extracting main feature information of the EEG signal;

[0070] Exemplarily, the main feature information in the EEG signal is extracted, and the main features in the EEG signal can be determined to determine the sleep period signal. Specifically, the frequency feature, waveform feature, time domain feature, spatial feature and rhythm feature of the EEG signal can be identified.

[0071] In a preferred embodiment, the above step S1021 includes the following steps:

[0072] Step a1, preprocessing the EEG signal to obtain a clean EEG signal.

[0073] Exemplarily, the EEG signal is filtered and preprocessed to obtain a clean EEG signal, thereby reducing the interference of noise signals.

[0074] Specifically, wavelet transform can be used to preprocess EEG signals. Wavelet transform is currently widely used in various disciplines of signal processing, image compression, speech analysis and other fields. Because both the time window and the frequency window can be adaptively changed to produce variable resolution characteristics, it is often used in the field of sleep EEG analysis.

[0075] Wavelet transform is a signal analysis method improved on the basis of short-time Fourier transform. It meets the requirements of localization of time-frequency analysis. Its characteristics are that the observation window length is fixed and the window shape can be adjusted adaptively. In wavelet transform, an appropriate basic wavelet is selected first, which must meet the following conditions of compatibility:

[0076]

[0077] in, yes The Fourier transform of the basic wavelet ψ(t) is used to form a cluster of wavelet systems |ψ a,b (t)|,|ψ a,b (t)| is expressed as:

[0078]

[0079] b is the time shift, a is the scale factor. This cluster of wavelets is used as a basis to form a series of subspace models of different sizes, and then the signal to be analyzed is projected into each subspace. By decomposing it at different scales, the time-scale characteristics of the signal can be obtained.

[0080] The principle of wavelet transform is as follows Figure 3 As shown in the figure, D represents the overview signal after decomposition, which mainly includes the high-frequency signal part, and A represents the detail signal, which mainly includes the low-frequency signal part. It is divided into two parts layer by layer in order. The decomposition object of the lower layer is the low-frequency part obtained by the decomposition of the upper layer, and then decomposed according to the same rule.

[0081] Wavelet transform can be divided into three categories according to the choice of its basic wavelet: classical wavelet, orthogonal wavelet and biorthogonal wavelet.

[0082] Commonly used classic wavelets include: Haar wavelet, Mexican Hat wavelet, Gaussian wavelet, etc.

[0083] The wavelet is defined as:

[0084]

[0085] The Haar wavelet waveform is as follows Figure 4 As shown in the figure, the characteristic of Haar wave is that it is tightly supported in the time domain. Its advantage is that it can remove the phase distortion in the system. Its disadvantage is that Haar wave is not a continuous wavelet and is subject to certain limitations in processing actual signal problems. Common orthogonal wavelets can be divided into the following categories: Daubechies wavelet, Coiflets wavelet, Meyer wavelet and symmetric wavelet, etc. Such wavelets are represented by adding the scaling function multiplied by different weights.

[0086] Daubechies wavelet, referred to as dbN wavelet, N is the order of the wavelet. The characteristic is that the larger the order N, the larger the vanishing moments will be, the more concentrated the energy distribution will be, and the more obvious the frequency band division effect will be. However, the longer the corresponding filter length is, the lower the support will be, and the amount of calculation will increase significantly, the algorithm will take longer, and the real-time performance will decrease.

[0087] Step a2, extracting sample entropy of the cleaned EEG signal to determine feature information of the cleaned EEG signal;

[0088] For example, according to Parseval's Theorem, the energy contained in the target signal is always equal to the sum of the energies contained in the components into which the signal is decomposed in a complete set of orthogonal functions. The energy of EEG signals varies greatly between individuals, but the energy ratio of each EEG frequency band can reflect the state of the brain of the same research subject at different times. Therefore, the energy ratio of each EEG frequency band can be used as a key feature of sleep staging.

[0089] The energy of each EEG frequency band is defined as follows:

[0090]

[0091] Among them, x i (t) is the signal corresponding to different frequency bands, E i is the energy of this frequency band.

[0092] The energy proportion of each EEG frequency band is as follows:

[0093]

[0094] EEG signal is a nonlinear, non-stationary, chaotic weak electrical signal. Therefore, some nonlinear features can be used to characterize the brain activity state of the subject, including: KS entropy (Kolmogorov-sina, KS), approximate entropy (Approximate Entropy, ApEn) and sample entropy (Sample Entropy, SampEn, etc.).

[0095] Approximate entropy and sample entropy are both methods to quantify the time complexity of a sequence. Their essence is to detect the probability of generating a new subsequence in the time series. The lower the entropy value, the higher the self-similarity of the signal. The higher the entropy value, the lower the similarity of the signal sequence, that is, the more chaotic and complex the sequence is. In the calculation process of approximate entropy, there is a step of calculating the self-matching value, which will produce errors. The improved sample entropy discards this step during calculation, thus eliminating this part of the error. Therefore, sample entropy is chosen as the key feature of sleep staging.

[0096] The specific steps for calculating sample entropy are as follows:

[0097] (1) Define an original sequence of N data points u = u{u(1),u(2),…,u(N)}, which are arranged in time order into a set of m-dimensional vectors: X m (1),X m (2),…,X m (N-m+1), where:

[0098] X m (i)=[u(i),u(i+1),…,u(i+m-1)],(1≤i≤N-m+1)

[0099] This vector represents the value of m consecutive sequences starting from the i-th point:

[0100] (2) Define two m-dimensional vectors X m (i) and X m (j) The distance between them is d[X m (i),X m (j)] to represent the difference of the corresponding elements of two vectors, where the maximum difference is the distance between the two vectors, that is:

[0101] d[X m (i),X m (j)] = max{u(i+k)-u(j+k)}

[0102] (3) Given a threshold value r (r>0), traverse all i values ​​and count d[X m(i),X m (j)] is less than the number of values ​​r m (i) Calculate the ratio of the number below the threshold to the total number Nm, expressed as:

[0103]

[0104] (4) Calculate C i m (r) The average value for all i:

[0105]

[0106] (5) Replace m with m+I and repeat steps (1)-(4) to construct C i m+1 (r),φ m+1 (r).

[0107] (6) The sample entropy is defined as follows:

[0108]

[0109] Represents the embedding dimension, r is the tolerance threshold, and SD is defined as the standard deviation of the original sequence. The value of SampEn is related to both m and r. The common value of m is 1 or 2. The value of r is related to SD, and the value range is usually . After comparing the results of multiple experiments, the embedding dimension and tolerance threshold were finally selected. The data length N selected for sample entropy calculation is required to be between 100 and 5000

[66] . This study selected the length N as 100, the step size as 10, and the filling method as minimum value filling. By calculating a series of sample entropy values ​​corresponding to different frequency bands of EEG after wavelet decomposition of 30 seconds of EEG signal, the mean and variance of all sample entropy are calculated to realize the feature extraction of sample entropy of EEG signal in different frequency bands.

[0110] Step a3, using the principal component analysis method to reduce the dimension of the feature information to obtain the main feature information.

[0111] Exemplarily, the principal component analysis (PCA) algorithm is used for feature dimensionality reduction. PCA is a dimensionality reduction method for extracting the main characteristic components of data, which uses orthogonal transformation to transform a set of linearly related variables into new independent variables. The main idea is to reduce the data dimension and make the principal components independent of each other under the premise of maximizing the retention of data information, and finally sort the principal components according to their contribution to the data difference, so as to facilitate subsequent feature screening. The PCA algorithm calculation process is divided into the following four steps:

[0112] (1) Let the original eigenvector be X:{x1,x2,,…x m}, m represents the feature dimension, n is the number of sample sets, X ij represents the jth eigenvector of the i-th sample. ij After normalization, we get

[0113]

[0114]

[0115]

[0116] and j It represents the sample mean and standard deviation of the j-th eigenvector.

[0117] (2) Calculate the correlation coefficient matrix R = (r ij ) m×m :

[0118]

[0119] r ij is the correlation coefficient between the i-th and j-th eigenvectors, where r ii =1, r ij =r ji .

[0120] (3) Calculate the eigenvalues ​​of the correlation coefficient matrix λ1≥λ2≥…λ m ≥0 and eigenvectors u1,u2,…,u m , define u j =(u 1j ,u 2j ,…,u mj ) T , then the original eigenvectors form m new principal component vectors which can be expressed as:

[0121]

[0122] Among them, y m Represents the mth principal component vector.

[0123] (4) Calculate the eigenvalue λ j The information contribution rate and cumulative contribution rate of (j=1,2,…,m). Define the principal component y j Information contribution rate b j for:

[0124]

[0125] Define the principal components as y1,y2,…,y p The cumulative contribution rate αp for:

[0126]

[0127] When α p When it is close to 1, the first p new eigenvectors are selected as principal components to replace the original m eigenvectors, so that the comprehensive evaluation value Z is obtained by analyzing the p principal components as shown below:

[0128]

[0129] Step S1022, input the main feature information into a preset model, classify the main feature information, and determine the sleep period characteristics and the wakefulness period characteristics.

[0130] Exemplarily, the preset model may be a support vector machine (SVM), a neural network, a random forest, etc., which is used to classify the main feature information to obtain the wakefulness signal and the sleep signal. In the embodiment of the present invention, the support vector machine (SVM) is used as an example for introduction.

[0131] SVM is a machine learning algorithm based on statistical learning. It takes into account both model complexity and learning ability, and can achieve the global optimal solution in a small sample size. SVM is often used in the field of nonlinear signal processing and classification, especially in the sleep staging problem based on EEG signals.

[0132] Take the linearly separable binary classification problem as an example: Suppose there are two classification samples (x i ,y i ), where x i ∈R d ,y i ∈±1,i-1,2,…,N. d is the characteristic dimension of each sample, X i represents the i-th eigenvector, y i Indicates the label corresponding to the true category of the sample. Construct a hyperplane: g(x) = (w·x+b) = 0 to correctly classify N samples. The distance from each sample point to be classified to the above hyperplane is defined as the geometric interval, and the formula is as follows:

[0133]

[0134] Pick the minimum value from all geometric intervals:

[0135]

[0136] Define the positive and negative class samples that are closest to the hyperplane on the interval boundary as support vectors. The minimum value of the geometric interval obtained by solving the following formula is the distance from the support vector to the hyperplane. According to the machine learning optimization problem, the above conditions are equivalent to:

[0137]

[0138] Using a graph to represent the classification idea is as Figure 5 shown, where the solid line represents the decision boundary, the dashed line represents the interval boundary, and the points on the dashed line represent the support vectors.

[0139] Convert the above problem into a quadratic programming problem with inequality constraints, and introduce Lagrange multipliers to transform the problem into an unconstrained problem. According to the formula, the weight vector can be obtained:

[0140]

[0141] where α i (i = 1, 2, …, l) are the Lagrange multipliers. Select a component α * of α * j such that it satisfies the condition 0 < α * j < C, and calculate b * :

[0142]

[0143] Finally, the linear separable binary classification decision function is derived as:

[0144]

[0145] If the sample data belongs to a non-linear problem in the feature space, then transform it w into a linearly separable problem in a higher-dimensional feature space. When transforming, select an appropriate kernel function to map the inner product operation in the inseparable space to the inner product operation in the separable space. The original features are mapped to high

[0146] dimensional space to obtain updated new features. The schematic diagram of the sample mapping process from low dimension to high dimension is as Figure 6 shown. The derivation idea of non-linear classification is generally similar to that of linear separability. Finally, the non-linear separable binary classification decision function can be derived:

[0147]

[0148] Common kernel functions include linear kernel, polynomial kernel, radial basis kernel function, and Sigmoid kernel, etc.

[0149] Linear kernel: K(x i , x) = x i , x

[0150] Polynomial kernel: K(x i ,x)=[(x i ,x)+1] d

[0151] Radial basis kernel function:

[0152] Sigmoid kernel: K(x i ,x)=tanh[v(x i ,x)+c]

[0153] The above four kernel functions are all suitable for situations with less sample data and lower dimensions. When the number of samples is very large, radial basis kernel function and sigmoid kernel function are generally used. The radial basis kernel function has a better recognition effect for classification problems with higher sample data dimensions. It can map instance samples from the original input space to infinite dimensional space, greatly enhance the generalization ability of the model, and maintain a wider convergence domain.

[0154] The schematic diagram of the sleep state recognition algorithm based on EEG signals corresponding to the above method is as follows: Figure 7 shown.

[0155] Step S1023, determining a sleep period signal according to the sleep period characteristics, the wakefulness period characteristics and the EEG signal.

[0156] For example, after determining the sleep period characteristics and the wake period characteristics, the EEG signals corresponding to the sleep period and the wake period can be determined according to the sleep period characteristics and the wake period characteristics. The corresponding sleep period signal can be determined by removing the EEG signal corresponding to the wake period from the EEG signal.

[0157] Step S103, determining the sleep airflow signal and the sleep displacement signal corresponding to the oral and nasal airflow signal and the abdominal displacement signal respectively according to the sleep period signal.

[0158] Exemplarily, after obtaining an accurate sleep period signal, the sleep airflow signal and the sleep displacement signal in the corresponding time period are determined within the time interval corresponding to the oral and nasal airflow signal and the abdominal displacement signal according to the time information corresponding to the sleep period signal.

[0159] According to the sleep period signal, the sleep airflow signal and the sleep displacement signal corresponding to the oral and nasal airflow signal and the abdominal displacement signal are determined respectively, specifically including:

[0160] Determining a sleep airflow signal corresponding to the oral and nasal airflow signal according to time information of the sleep period signal;

[0161] A sleep displacement signal of the abdomen displacement signal is determined according to the time information of the sleep period signal.

[0162] Step S104, determining the breathing state of the subject according to the sleep displacement signal and the oral and nasal airflow signal.

[0163] Exemplarily, different respiratory states correspond to different oronasal airflow signals and sleep displacement signals. The onset characteristic of OSAS (obstructive apnea) is that the oronasal airflow stops, while the chest and abdominal breathing movements still exist; the onset characteristic of CSAS (central apnea) is that both the oronasal airflow and the chest and abdominal breathing movements stop; the onset characteristic of MSAS (mixed apnea) is that in a respiratory apnea process, the oronasal airflow and the chest and abdominal breathing movements stop in the first half, and the oronasal airflow stops in the second half, while the chest and abdominal breathing movements exist.

[0164] like Figure 8 The following are four typical respiratory events corresponding to the abdominal displacement signal. The box is the onset time window marked by the expert. The waveforms of different onset types are very different from those of normal ones. Based on the different abdominal movement characteristics, the signal can be used for SAHS detection and four-classification.

[0165] In a preferred embodiment, Fig. 9 As shown, the above step S104 specifically includes:

[0166] Step S1041, when there is an oral and nasal airflow signal and no sleep displacement signal, determining that the respiratory state is obstructive apnea;

[0167] Step S1042, when there is no oral and nasal airflow signal and no sleep displacement signal, determining that the respiratory state is central apnea;

[0168] Step S1043, when there is no oronasal airflow signal and sleep displacement signal in the first preset time period, and there is an oronasal airflow signal and sleep displacement signal in the second preset time period, the respiratory state is mixed apnea, and the first preset time period is before the second preset time period.

[0169] The respiratory state determination method provided in this embodiment obtains the EEG signal, oral and nasal signals and abdominal displacement signal of the subject during sleep, and more comprehensively evaluates the respiratory state of the subject, including the depth, frequency, duration and other aspects of breathing; identifies the sleep phase signal in the EEG signal, reduces the interference of other noise signals, and improves the recognition efficiency; on this basis, according to the sleep phase signal, determines the sleep airflow signal and the sleep displacement signal corresponding to the oral and nasal airflow signal and the abdominal displacement signal, respectively, and determines the respiratory state of the subject according to the sleep phase signal and the oral and nasal airflow signal, thereby improving the accuracy of the respiratory state and reducing the corresponding respiratory state detection cost.

[0170] In this embodiment, a breathing state determination device is also provided, which is used to implement the above-mentioned embodiments and preferred implementation modes, and the descriptions that have been made will not be repeated. As used below, the term "module" can implement a combination of software and / or hardware of a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, the implementation of hardware, or a combination of software and hardware, is also possible and conceivable.

[0171] This embodiment provides a breathing state determination device, such as Fig.10 The device comprises:

[0172] The signal acquisition module 1001 is used to acquire the EEG signal, oral and nasal airflow signal, and abdominal displacement signal of the subject during sleep;

[0173] A sleep identification module 1002 is used to identify sleep phase signals in EEG signals;

[0174] A sleep signal determination module 1003 is used to determine a sleep airflow signal and a sleep displacement signal corresponding to the oral and nasal airflow signal and the abdominal displacement signal, respectively, according to the sleep period signal;

[0175] The respiratory state determination module 1004 is used to determine the respiratory state of the subject according to the sleep displacement signal and the oral and nasal airflow signal.

[0176] In some optional implementations, identifying a sleep module specifically includes:

[0177] Extracting main feature units, used to extract main feature information of EEG signals;

[0178] A classification unit, used to input the main feature information into a preset model, classify the main feature information, and determine the sleep period characteristics and the wakefulness period characteristics;

[0179] The signal determination unit is used to determine the sleep period signal according to the sleep period characteristics, the wakefulness period characteristics and the EEG signal.

[0180] In some optional implementations, extracting the main characteristic units specifically includes:

[0181] A preprocessing subunit, used for preprocessing the EEG signal to obtain a clean EEG signal;

[0182] A feature extraction subunit is used to extract sample entropy of the cleaned EEG signal and determine feature information of the cleaned EEG signal;

[0183] The dimension reduction subunit is used to reduce the dimension of feature information using the principal component analysis method to obtain the main feature information.

[0184] In some optional implementations, determining the sleep signal module specifically includes:

[0185] an airflow determination unit, for determining a sleep airflow signal corresponding to the oral and nasal airflow signal according to time information of the sleep period signal;

[0186] The displacement determination unit is used to determine the sleep displacement signal of the abdomen displacement signal according to the time information of the sleep period signal.

[0187] In some optional implementations, the respiratory state determination module specifically includes:

[0188] A first state determination unit, configured to determine that the respiratory state is obstructive apnea when there is an oral and nasal airflow signal and no sleep displacement signal;

[0189] A second state determination unit, configured to determine that the respiratory state is central apnea when there is no oral or nasal airflow signal and no sleep displacement signal;

[0190] The third state determination unit is used to determine that the respiratory state is mixed apnea when there is no oronasal airflow signal and sleep displacement signal in a first preset time period, and there is an oronasal airflow signal and sleep displacement signal in a second preset time period, and the first preset time period is before the second preset time period.

[0191] In some optional embodiments, the oral and nasal airflow signal is acquired by a thermal sensor and / or a flow sensor.

[0192] In some optional implementations, the EEG signal is preprocessed using wavelet transform to obtain a clean EEG signal.

[0193] The further functional description of each of the above modules and units is the same as that of the above corresponding embodiments and will not be repeated here.

[0194] The respiratory state determination device in this embodiment is presented in the form of a functional unit, where the unit refers to an ASIC (Application Specific Integrated Circuit) circuit, a processor and memory that executes one or more software or fixed programs, and / or other devices that can provide the above functions.

[0195] The embodiment of the present invention also provides a computer device having the above Fig.10 A breathing state determination device is shown.

[0196] See also Fig.11 , Fig.11 is a schematic diagram of the structure of a computer device provided by an optional embodiment of the present invention, such as Fig.11As shown, the computer device includes: one or more processors 10, a memory 20, and interfaces for connecting various components, including high-speed interfaces and low-speed interfaces. Various components are connected to each other using different buses for communication, and can be installed on a common mainboard or installed in other ways as needed. The processor can process the instructions executed in the computer device, including instructions stored in or on the memory to display the graphical information of the GUI on an external input / output device (such as, a display device coupled to the interface). In some optional embodiments, if necessary, multiple processors and / or multiple buses can be used together with multiple memories and multiple memories. Similarly, multiple computer devices can be connected, and each device provides some necessary operations (for example, as a server array, a group of blade servers, or a multi-processor system). Fig.11 A processor 10 is taken as an example.

[0197] The processor 10 may be a central processing unit, a network processor or a combination thereof. The processor 10 may further include a hardware chip. The hardware chip may be a dedicated integrated circuit, a programmable logic device or a combination thereof. The programmable logic device may be a complex programmable logic device, a field programmable gate array, a general purpose array logic or any combination thereof.

[0198] The memory 20 stores instructions executable by at least one processor 10, so that at least one processor 10 executes the method shown in the above embodiment.

[0199] The memory 20 may include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function; the data storage area may store data created according to the use of the computer device, etc. In addition, the memory 20 may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some optional embodiments, the memory 20 may optionally include a memory remotely arranged relative to the processor 10, and these remote memories may be connected to the computer device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0200] The memory 20 may include a volatile memory, such as a random access memory; the memory may also include a non-volatile memory, such as a flash memory, a hard disk or a solid state drive; the memory 20 may also include a combination of the above types of memory.

[0201] The computer device further comprises a communication interface 30 for the computer device to communicate with other devices or a communication network.

[0202] The embodiment of the present invention also provides a computer-readable storage medium. The method according to the embodiment of the present invention can be implemented in hardware, firmware, or can be implemented as a computer code that can be recorded in a storage medium, or can be implemented as a computer code that is originally stored in a remote storage medium or a non-temporary machine-readable storage medium and will be stored in a local storage medium through a network download, so that the method described herein can be stored in such software processing on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. Among them, the storage medium can be a magnetic disk, an optical disk, a read-only storage memory, a random access memory, a flash memory, a hard disk or a solid-state hard disk, etc.; further, the storage medium can also include a combination of the above types of memories. It can be understood that a computer, a processor, a microprocessor controller, or programmable hardware includes a storage component that can store or receive software or computer code. When the software or computer code is accessed and executed by a computer, a processor, or hardware, the method shown in the above embodiment is implemented.

[0203] Although the embodiments of the present invention have been described in conjunction with the accompanying drawings, those skilled in the art may make various modifications and variations without departing from the spirit and scope of the present invention, and such modifications and variations are all within the scope defined by the appended claims.

Claims

1. A method for determining a respiratory state, characterized in that: The method comprises: Obtaining the EEG signals, oral and nasal airflow signals, and abdominal displacement signals of the subject during sleep; Identifying a sleep phase signal in the EEG signal; Determining, according to the sleep period signal, a sleep airflow signal and a sleep displacement signal corresponding to the oral and nasal airflow signal and the abdominal displacement signal, respectively; The respiratory state of the subject is determined based on the sleep displacement signal and the oral and nasal airflow signal.

2. The method according to claim 1, characterized in that The step of identifying the sleep phase signal in the EEG signal specifically includes: Extracting main characteristic information of the EEG signal; Inputting the main characteristic information into a preset model, classifying the main characteristic information, and determining the sleep period characteristics and the wakefulness period characteristics; A sleep period signal is determined according to the sleep period characteristics, the wakefulness period characteristics and the electroencephalogram signal.

3. The method according to claim 2, characterized in that The extracting of main characteristic information of the EEG signal specifically includes: Preprocessing the EEG signal to obtain a clean EEG signal; Extracting sample entropy of the cleaned EEG signal to determine feature information of the cleaned EEG signal; The characteristic information is reduced in dimension using the principal component analysis method to obtain the main characteristic information.

4. The method according to claim 1, characterized in that: The step of determining the sleep airflow signal and the sleep displacement signal corresponding to the oral and nasal airflow signal and the abdominal displacement signal respectively according to the sleep period signal specifically includes: determining the sleep airflow signal corresponding to the oral and nasal airflow signal according to the time information of the sleep period signal; The sleep displacement signal of the abdomen displacement signal is determined according to the time information of the sleep period signal.

5. The method according to any one of claims 2 to 4, characterized in that: Determining the respiratory state of the subject according to the sleep displacement signal and the oral and nasal airflow signal specifically includes: When the oral and nasal airflow signal is present and the sleep displacement signal is absent, determining that the respiratory state is obstructive apnea; When the oral and nasal airflow signal does not exist and the sleep displacement signal does not exist, determining that the respiratory state is central apnea; When the oronasal airflow signal and the sleep displacement signal do not exist in a first preset time period, and the oronasal airflow signal and the sleep displacement signal exist in a second preset time period, the respiratory state is mixed apnea, and the first preset time period is before the second preset time period.

6. The method according to claim 5, characterized in that The oral and nasal airflow signal is obtained through a thermal sensor and / or a flow sensor.

7. The method according to claim 6, characterized in that The electroencephalogram signal is preprocessed by wavelet transform to obtain the cleaned electroencephalogram signal.

8. A breathing state determination device, characterized in that: The device comprises: A signal acquisition module is used to acquire the EEG signal, oral and nasal airflow signal, and abdominal displacement signal of the subject during sleep; A sleep identification module, used for identifying a sleep phase signal in the EEG signal; a sleep signal determination module, configured to determine, according to the sleep period signal, a sleep airflow signal and a sleep displacement signal corresponding to the oral and nasal airflow signal and the abdominal displacement signal, respectively; The respiratory state determination module is used to determine the respiratory state of the subject based on the sleep period signal and the oral and nasal airflow signal.

9. A computer device, characterized in that: include: A memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the respiratory state determination method according to any one of claims 1 to 7 by executing the computer instructions.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a computer to execute the respiratory state determination method according to any one of claims 1 to 7.

Images