Sleep staging method and device

Abstract

The present application is about a sleep staging method and device, the method comprising: obtaining a multi-channel EEG signal of a human body; inputting the multi-channel EEG signal into a sleep feature detection model component, and outputting a sleep feature signal in the multi-channel EEG signal and a signal type of the sleep feature signal through the sleep feature detection model component; and determining a sleep staging result of the human body according to the sleep feature signal and the signal type. Using the sleep staging method and device provided in each embodiment of the present application, a sleep detection model component with a high accuracy can be trained with less sample data.

Description

Technical Field

[0001] The present application relates to the technical field of sleep detection, and in particular to a sleep staging method and device. Background Art

[0002] With the development of economy and the improvement of living standards, Chinese people pay more attention to their health issues. Sleep, as an important part of everyone's life, plays an important role in people's physical and mental health. The "2018 China Sleep Quality Survey Report" surveyed 100,000 people, covering all provinces in the country. The survey results show that 16% of the respondents sleep less than 6 hours at night, which means they go to bed after 24:00 and get up before 6:00; 83.81% of the respondents are often troubled by sleep problems, of which 25.83% have difficulty falling asleep, 26.49% have shallow sleep, and 25.54% of the respondents are observed to have short breathing apnea.

[0003] It can be seen that sleep problems cannot be ignored. At the same time, the survey report also pointed out that some auxiliary treatment methods can alleviate sleep disorders and help people solve sleep problems. Sleep staging is an indispensable and important means in the process of studying sleep and sleep-related diseases, diagnosing and treating sleep disorders, and evaluating sleep quality. In the related technologies, there are endless algorithms for automatic sleep staging. The more commonly used technologies currently include machine learning methods. However, the machine learning models in the related technologies not only require more sample data during the training process, but the sleep staging results obtained are also not very accurate.

[0004] Therefore, a more accurate sleep staging method is urgently needed in the relevant technology. Summary of the invention

[0005] The purpose of the embodiments of the present application is to provide a sleep staging method and device, which can train a sleep detection model component with higher accuracy using less sample data.

[0006] The sleep staging method and device provided in the embodiment of the present application are implemented as follows:

[0007] A sleep staging method, the method comprising:

[0008] Obtain multi-channel EEG signals from the human body;

[0009] Inputting the multi-channel EEG signal into a sleep feature detection model component, and outputting a sleep feature signal in the multi-channel EEG signal and a signal type of the sleep feature signal through the sleep feature detection model component;

[0010] The sleep stage result of the human body is determined according to the sleep characteristic signal and the signal type.

[0011] A sleep staging device includes a processor and a memory for storing instructions executable by the processor, wherein when the processor executes the instructions:

[0012] Obtain multi-channel EEG signals from the human body;

[0013] Inputting the multi-channel EEG signal into a sleep feature detection model component, and outputting a sleep feature signal in the multi-channel EEG signal and a signal type of the sleep feature signal through the sleep feature detection model component;

[0014] The sleep stage result of the human body is determined according to the sleep characteristic signal and the signal type.

[0015] A sleep staging device comprises: an electroencephalogram (EEG) acquisition device and the sleep staging device, wherein the EEG acquisition device is used to acquire the multi-channel EEG signals of the human body.

[0016] A sleep staging device includes: an electroencephalogram acquisition device, a data sending device, and a data receiving device, wherein:

[0017] The EEG acquisition device is used to acquire multi-channel EEG signals of a human body;

[0018] The data sending device is used to send the multi-channel EEG signal to a server, and the server is configured to obtain a sleep characteristic signal and a signal type of the sleep characteristic signal in the multi-channel EEG signal by using a sleep characteristic detection model component, and determine a sleep staging result of the human body according to the sleep characteristic signal and the signal type;

[0019] The data receiving device is used to receive the sleep staging result of the human body.

[0020] A non-transitory computer-readable storage medium, when instructions in the storage medium are executed by a processor, enables the processor to perform the sleep staging method.

[0021] The sleep staging method and device provided by the present application can detect the sleep characteristic signals and their types in the multi-channel EEG signals of the human body by using the sleep detection model component, and determine the sleep staging result of the human body according to the sleep characteristic signals and their types. Among them, the sleep characteristic detection model component is obtained by training with multiple multi-channel EEG sample signals, and the multi-channel EEG sample signals are annotated with sleep characteristic signals of different signal types. On the one hand, the sleep staging result is closely related to the sleep characteristic signal. Therefore, by first obtaining the sleep characteristic signal in the multi-channel EEG sample signal using the model component, and then determining the sleep staging result according to the sleep characteristic signal, a relatively accurate result can be obtained; on the other hand, the sleep characteristic signal is annotated in the multi-channel EEG sample signal instead of the sleep staging stage. From the perspective of the difficulty of feature extraction, the amount of sample data required to extract the sleep characteristic signal is much less than that required to extract the sleep staging stage. Therefore, the sleep staging method provided by the embodiment of the present application can be used to train a sleep detection model component with a higher accuracy rate with less sample data. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0023] Figure 1 The diagram is a schematic diagram of a scenario according to an exemplary embodiment.

[0024] Figure 2 The figure is a flowchart of a method according to an exemplary embodiment.

[0025] Figure 3 The diagram is a schematic diagram of a scenario according to an exemplary embodiment.

[0026] Figure 4 The figure is a flowchart according to an exemplary embodiment.

[0027] Figure 5 The diagram is a schematic diagram of a scenario according to an exemplary embodiment.

[0028] Figure 6 is a waveform diagram shown according to an exemplary embodiment.

[0029] Figure 7 is a waveform diagram shown according to an exemplary embodiment.

[0030] Figure 8 It is a block diagram of a device according to an exemplary embodiment.

[0031] Fig. 9is a block diagram of a device according to an exemplary embodiment DETAILED DESCRIPTION

[0032] Exemplary embodiments will be described in detail herein, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, the same numbers in different drawings represent the same or similar elements unless otherwise indicated. The implementations described in the following exemplary embodiments do not represent all implementations consistent with the present application. Instead, they are merely examples of devices and methods consistent with some aspects of the present application as detailed in the appended claims.

[0033] In order to facilitate those skilled in the art to understand the technical solution provided by the embodiments of the present application, the technical environment in which the technical solution is implemented is described below.

[0034] The more commonly used sleep detection methods in the related art mainly include machine learning methods. In the process of machine learning, a large amount of sample data is often required, and the sample data needs to be labeled. The most common labeling method in the related art is to mark the sleep stage of the human body in the EEG signal, and then train these sample data labeled with the sleep stage. However, in actual scenarios, the length of the human body's sleep stages varies, that is, the sleep characteristics of different people are different. Therefore, using the machine learning method in the related art to train a more accurate sleep staging model not only requires a large amount of sample data, but also the trained sleep staging model may not be accurate enough.

[0035] Based on actual technical needs similar to those described above, the sleep staging method provided in the present application can separate the sleep feature extraction and sleep staging in machine learning. This is because although the length of the human body's sleep stages varies, most of the brain wave characteristics in the same sleep stage are roughly similar. Therefore, by first determining the sleep characteristics and then dividing the sleep stages, a more accurate sleep staging result can be obtained.

[0036] The sleep staging method provided by each embodiment of the present application is described in detail below through two specific business scenarios.

[0037] Scene 1

[0038] like Figure 1 As shown, a user is suspected of having a sleep disorder. During the treatment in the hospital, the doctor needs to understand the user's sleep state, especially the duration of each sleep stage. Therefore, the sleep staging method provided in the embodiment of the present application can be used to detect the sleep staging results of the user's body 101. Figure 1As shown, the multi-channel EEG signals of the human body 101 can be collected by the EEG signal collection device 107. During the collection process, the detection electrodes 103a, 103b and 103c can be fixed at various positions on the head of the human body 101. After the interface 105 of the EEG signal collection device 107 is connected to the detection electrodes 103a, 103b and 103c, the EEG signal collection device 107 can collect the multi-channel EEG signals of the human body 101. Of course, the EEG signal collection device 107 is only exemplary, and can also include a helmet, a hat, bedding, a packaged portable detection electrode and other various devices.

[0039] like Figure 1 As shown, the EEG signal acquisition device 107 can also be coupled with a sleep staging device 109. Figure 1 This is only an exemplary coupling method. The EEG signal acquisition device 107 and the sleep staging device 109 can also be integrated into the same device, such as the sleep staging device 109 can be integrated inside the EEG signal acquisition device 107. This application does not limit this. The sleep staging device 109 can process the multi-channel EEG signals acquired by the EEG acquisition device 107 using the sleep staging method provided in any of the following embodiments of this application. Figure 2 FIG. 1 is a flow chart of a method for the sleep staging device 109 to process the multi-channel EEG signal. Figure 2 As shown, after acquiring the multi-channel EEG signal, the sleep staging device 109 can perform signal preprocessing such as removing baseline interference, removing power frequency interference, and removing non-EEG signals on the multi-channel EEG signal, and then input the signal into a sleep feature detection model component, which outputs a sleep feature signal and a signal type. The sleep feature detection model component can be trained using a plurality of multi-channel EEG sample signals. After acquiring the sleep feature signal and the signal type, the sleep staging result of the human body 101 can be determined based on the predicted sleep staging rules (such as the sleep staging rules prescribed by the American Sleep Association (AASM)), the sleep feature signal and the signal type. Figure 2 As shown, the sleep staging result may include the division of the sleep cycle in the time dimension. For example, the human body 101 is in the W stage from 23:00 to 1:00 and in the N1 sleep stage from 1:00 to 2:00.

[0040] Of course, after obtaining the sleep stage results, Figure 1 As shown, the sleep staging result can be displayed on the display 111, which can help doctors or patients understand specific test results or treatment results.

[0041] Scene 2

[0042] In another scenario, the collected multi-channel EEG signals can also be uploaded to the server for processing. This scenario is more suitable for portable scenarios. Users can use the portable EEG signal collection device 107 to collect multi-channel EEG signals at home, office or any other place. Figure 3 The portable EEG signal acquisition device 107 is only exemplary, and may also include a helmet, a hat, bedding, a packaged portable detection electrode, and other various devices. Figure 3 As shown, the EEG signal acquisition device 107 is coupled with the data transceiver 113. After the EEG signal acquisition device 107 acquires the multi-channel EEG signal of the human body 101, the data transceiver 113 can send the multi-channel EEG signal to the server 115. The server 115 can include any form of data processing server such as a cloud server and a distributed server. After receiving the multi-channel EEG signal, the server 115 can process the multi-channel EEG signal using the sleep staging method provided in each embodiment of the present application, and obtain the sleep staging result of the human body 101. Then, the server 115 can send the sleep staging result to the data transceiver 113. After receiving the sleep staging result, the data transceiver 113 can display the sleep staging result on the display 111, such as a client display such as a user's mobile phone or computer.

[0043] Through the above scenarios, users can learn about their personal sleep staging results without leaving home. When encountering problems such as sleep disorders, they can send the obtained sleep staging results to their doctors to facilitate the doctors in determining the diagnosis results.

[0044] The sleep staging method described in this application is described in detail below with reference to the accompanying drawings. Figure 4 : It is a method flow chart of an embodiment of the sleep staging method provided by the present application. Although the present application provides method operation steps as shown in the following embodiments or drawings, more or fewer operation steps may be included in the method based on routine or no creative labor. In the steps where there is no necessary causal relationship logically, the execution order of these steps is not limited to the execution order provided in the embodiments of the present application. When the method is executed in an actual sleep staging process or device, it can be executed in the method sequence shown in the embodiments or drawings or in parallel (for example, in a parallel processor or multi-threaded processing environment).

[0045] Specifically, an embodiment of the sleep staging method provided in the present application is as follows: Figure 4 As shown, the method may include:

[0046] S401: Acquire multi-channel EEG signals of a human body.

[0047] S403: Input the multi-channel EEG signal into a sleep feature detection model component, and output the sleep feature signal in the multi-channel EEG signal and the signal type of the sleep feature signal through the sleep feature detection model component.

[0048] S405: Determine the sleep stage result of the human body according to the sleep characteristic signal and the signal type.

[0049] In the embodiment of the present application, an EEG acquisition device can be used to acquire multi-channel EEG signals of a human body. The EEG acquisition device may include an instrument capable of acquiring EEG (electroencephalogram). The number of leads of the EEG acquisition devices provided by different manufacturers is different, for example, it may include 8-channel, 24-channel, 64-channel and other lead interfaces, and the present application does not limit this.

[0050] In one example, the multi-channel EEG signal can be obtained by connecting multiple electrodes to various positions of the head according to the electrode placement method of the international 10-20 system. The international 10-20 system involves 19 recording electrodes: Fp1, Fp2, F7, F3, Fz, F4, F8, T3, C3, Cz, C4, T4, T5, P3, Pz, P4, T6, 01, 02. Figure 5 As shown, the placement order of the 19 electrodes may include:

[0051] First, two baselines were determined on the scalp surface. One was the front-to-back line from the nasion to the external occipital protuberance, which was 100%, and the other was the left-right line between the bilateral preauricular concavities, which was 100%. The intersection of the two baselines on the top of the head was the location of the Cz electrode. FPz (frontal pole midline) was 10% behind the nasion. Every 20% from FPz was the location of an electrode, which was Fz (frontal midline), Cz (central midline), Pz (top midline) and Oz (occipital midline). The distance between Oz and the external occipital protuberance was 10%. The T3 (left middle temporal) electrode was located 10% from the left preauricular concavity on the line connecting the bilateral preauricular concavities. After that, an electrode was placed every 20% to the right, which was C3 (left center), Cz (central midline), C4 (right center) and T4 (right middle temporal). The distance between T4 and the right preauricular concavity was 10%. The line from FPz through T3 to Oz is the left temporal line, 10% to the left from FPz is FP1 (left frontal pole), and an electrode is placed every 20% from FP1 along the left lateral to the back, which are F7 (left anterior temporal), T3 (left middle temporal), T5 (left posterior temporal) and O1 (left occipital), among which T3 is the intersection of this line and the line connecting the two preauricular concavities, and O1 is 10% away from Oz. The line from FP2 along the right lateral to the back corresponds to this, and from front to back, they are FP2 (right frontal pole), F8 (right anterior temporal), T4 (right middle temporal), T6 (right posterior temporal) and O2 (right occipital). A line was drawn from FP1 to O1 and from FP2 to O2, which are the left and right parasagittal lines. From FP1 and FP2, every 20% backward is an electrode site. The left side is F3 (left frontal), C3 (left central), P3 (left parietal) and O1 (left occipital), and the right side is F4 (right frontal), C4 (right central), P4 (right parietal) and O2 (right occipital). In the 10-20 system, FPz and Oz are not included in the 19 recording sites.

[0052] Of course, in the embodiment of the present application, the collection position of the multi-channel EEG signal is not limited to the 19 collection positions specified by the international 10-20 system, but may also include any one or more of the 19 collection positions, or any other position of the human brain. The present application does not impose any limitation on this.

[0053] In an embodiment of the present application, after the multi-channel EEG signal is collected, the multi-channel EEG signal can be input into a sleep feature detection model component, and the sleep feature detection model component outputs the sleep feature signal in the multi-channel EEG signal and the signal type of the sleep feature signal. In one embodiment of the present application, the sleep feature detection model component can be trained using a plurality of multi-channel EEG sample signals, and the multi-channel EEG sample signals are annotated with sleep feature signals of different signal types. Among them, the multi-channel EEG sample signals can be obtained from real medical records of the hospital. For example, medical records of multiple patients with sleep disorders and who have undergone EEG testing can be collected from multiple tertiary hospitals. Of course, more realistic multi-channel EEG sample signals can also be obtained through other channels, and the present application does not limit this. Before training the sleep feature detection model component, sleep feature signals can be annotated in these multi-channel EEG sample signals. The sleep feature signals may include key signals in the EEG signal that are closely related to sleep staging. In some examples, the sleep feature signals may include at least one of the following signals: top waves (such as Figure 6 ), sleep spindles (as shown in Figure 7 As shown), K complex wave, sleep slow wave, alpha wave, sawtooth wave. According to the research of the American Academy of Sleep Medicine (AASM), the sleep cycle of adults can be divided into five stages: W stage, N1 stage sleep, N2 stage sleep, N3 stage sleep, and R stage sleep. The sleep characteristic signals involved in each sleep cycle are: the sleep characteristic signal of W stage is alpha wave, the sleep characteristic signal of N1 stage sleep is top wave, the sleep characteristic signal of N2 stage sleep is spindle wave and K complex wave, the sleep characteristic signal of N3 stage sleep is sleep slow wave, and the sleep characteristic signal of R stage sleep is sawtooth wave. Of course, with the deepening of medical research, the way of dividing sleep stages may also be different, and the sleep characteristic signals of each stage may also be different, which is not limited in this application. In addition, the sleep staging method is not limited to the staging method provided by AASM, and can be any sleep staging method provided by an authoritative organization, which is not limited in this application.

[0054] In an embodiment of the present application, after acquiring the multiple multi-channel EEG sample signals, the multi-channel EEG sample signals can be subjected to signal preprocessing. The signal preprocessing method may include at least one of the following: removing baseline drift, removing power frequency interference, and removing non-EEG signals. Among them, in the removal of baseline drift, the baseline drift belongs to a low-frequency non-stationary random signal. In an embodiment of the present application, the multi-resolution property based on the wavelet transform can be used to remove the baseline drift by using a wavelet decomposition and reconstruction method. In the process of removing power frequency interference, since the power frequency interference is relatively strong and is ubiquitous in a non-shielded environment, in an embodiment of the present application, the power frequency interference can be filtered out by a dedicated notch filter. For example, a 4th-order IIR notch filter with a notch frequency of 50Hz can be used for filtering. In the actual application environment, the sleep EEG signal contains the following EEG characteristic waves: α wave frequency range is 8-13Hz, β wave frequency range is 13-30Hz, δ wave frequency range is 1-4Hz, θ wave frequency range is 4-8Hz, sawtooth wave frequency range is 2-6Hz, spindle wave frequency range is 12-14Hz, β1 wave frequency range is 13-17Hz, β2 wave frequency range is 17-30Hz. For signal components outside the above frequency range, they may be generated by other electrophysiological activities of the human body and need to be filtered out. In an embodiment of the present application, a 10th-order IIR Butterworth bandpass filter can be used to filter out non-EEG signals. For example, in one example, the frequency of the bandpass filter can be set to 0.2-40Hz. Of course, in other embodiments, the signal preprocessing method for the multi-channel EEG sample signal can also include many other methods, which are not limited in this application.

[0055] In an embodiment of the present application, in the process of using a plurality of multi-channel EEG sample signals to train the sleep feature detection model component, the sleep feature signal can be intercepted from the multi-channel EEG sample signal. In one embodiment, the intercepted window width can be fixed. When the sampling rate of the multi-channel EEG sample signal is 256 Hz (i.e., 256 samples per second), the intercepted window width can be set to 256. Of course, in other embodiments, the width can also be set to 250, 300, etc., and the present application does not limit this.

[0056] In an embodiment of the present application, after intercepting the sleep characteristic signal, the sleep characteristic detection model component can be constructed, and the sleep characteristic detection model component is provided with training parameters. Then, the sleep characteristic signals can be respectively input into the sleep characteristic detection model component to generate prediction results. In one embodiment, the prediction results can include the probability that the input multi-channel EEG signals are various sleep characteristic signals. Finally, based on the difference between the prediction result and the type of the sleep characteristic signal, the training parameters can be iteratively adjusted until the difference meets the preset requirements.

[0057] In the above embodiments, the sleep characteristic signal is intercepted from the multi-channel EEG sample signal and then learned, which can greatly reduce the amount of data required for training the sleep characteristic detection model component. At the same time, the trained sleep characteristic detection model component can also accurately determine whether the input multi-channel EEG sample signal contains a sleep characteristic signal and the type corresponding to the sleep characteristic signal.

[0058] In an embodiment of the present application, the sleep feature detection model component may include a model component trained using a machine learning method. The machine learning method may also include a K-nearest neighbor algorithm, a perceptron algorithm, a decision tree, a support vector machine, a logistic regression, a maximum entropy, etc., and correspondingly, the generated model components may be naive Bayes, hidden Markov, etc. Of course, in other embodiments, the machine learning method may also include a deep learning method, a reinforcement learning method, etc., and the generated model components may include a convolutional neural network model component (CNN), a recurrent neural network model component (RNN), LeNet, ResNet, a long short-term memory network model component (LSTM), a bidirectional long short-term memory network model component (Bi-LSTM), etc., and the present application does not limit this.

[0059] In practical applications, the acquired multi-channel EEG signals are often signals that last for a long time, especially for sleep stage detection, which may take about 8 hours. If all EEG signals are input into the sleep feature detection model component for detection at one time, the data processing volume of the sleep feature detection model component may be too large, and the corresponding model scale should also be relatively large. Based on this, the acquired multi-channel EEG signals can be processed one by one according to the sliding window. In one embodiment of the present application, a sliding window with a preset length can be used to slide on the multi-channel EEG signal according to a preset step length, and the multi-channel EEG signal in the sliding window is input into the sleep feature detection model component at each sliding, and the sleep feature signal in the multi-channel EEG signal in the sliding window is output through the sleep feature detection model component. In one example, the length of the sliding window can be greater than 1 / 4 of the sampling rate and less than four times the sampling rate, and the moving step length of the sliding window can also be set to be less than the length of a single sliding window. Of course, in other embodiments, the length and step length of the sliding window can also have other settings, and this application is not limited here. It should be noted that the length of the sliding window needs to be consistent with the length of the window for extracting the sleep characteristic signal, that is, if the length of the window for extracting the sleep characteristic signal in the training stage is 256, then the length of the sliding window is also 256.

[0060] In practical applications, EEG signals are very weak bioelectric signals that need to be amplified millions of times before they can be recorded from the scalp. In this process, various non-brain source signals may also enter the amplifier and mix into the EEG signals. These signals are pseudo-difference signals. The identification of pseudo-difference signals is crucial to the accuracy of EEG judgment, especially with the widespread application of long-term EEG monitoring technology in clinical practice. The pseudo-difference in EEG recordings is inevitably increased in large quantities, the source is more complex, and the identification is more difficult. Simple bandpass filtering has been difficult to filter out pseudo-difference signals. The sources of pseudo-difference signals mainly include physiological activities, interference from EEG instruments and electrodes, environmental electromagnetic interference, patient's own movement, etc. Based on this, in one embodiment of the present application, pseudo-difference signals can be identified from the multi-channel EEG signals and the pseudo-difference signals can be removed. In an embodiment of the present application, the pseudo-difference detection model component can be used to detect the pseudo-difference signals in the multi-channel EEG signals. The pseudo-difference detection model component can also be trained using machine learning. Similar to the method of training the sleep feature detection model component, the pseudo-difference detection model component can also be trained using multiple multi-channel EEG sample signals. In this process, similarly, an artifact signal can be intercepted from the sample signal, such as an electrocardiogram artifact, a motion artifact, other artifacts, etc. Then, the artifact signal and some non-artifact signals can be input into the artifact detection model component, and the artifact detection model component can be adjusted until the artifact detection model component meets the preset requirements.

[0061] In another embodiment of the present application, the sleep feature detection model component and the artifact detection model component can be combined into one model component, that is, the artifact signal and the sleep feature signal are input together into the sleep feature detection model component for training. In addition, artifact detection can be performed before or after sleep feature detection, and the present application does not limit this.

[0062] In addition, in the process of performing artifact detection, it is also possible to combine the ECG signal or the EMG signal for detection. This is because the order of magnitude of the ECG signal and the EMG signal is much higher than that of the EEG signal. By combining the EEG signal, the ECG signal, and the EMG signal, the artifact signal in the EEG signal can be accurately determined. Based on this, the ECG signal and / or the EMG signal of the human body can also be collected, and the multi-channel EEG signal, the ECG signal, and / or the EMG signal are input into the artifact detection model component, and the artifact signal in the multi-channel EEG signal is output through the artifact detection model component.

[0063] It should be noted that after determining the artifact signal and the sleep characteristic signal, the position where the artifact signal and the sleep characteristic signal appear simultaneously in the multi-channel EEG signal can also be obtained, and the mark of the sleep characteristic signal at the position can be cancelled.

[0064] In an embodiment of the present application, after determining the sleep characteristic signal in the polyencephalogram signal and the signal type of the sleep characteristic signal, the sleep staging result of the human body can be determined according to the signal type of the sleep characteristic signal. In one embodiment of the present application, the sleep staging result of the human body can be determined according to the sleep cycle division method of AASM. For example, α waves correspond to W stage, top waves correspond to N1 stage sleep, spindle waves and K complex waves correspond to N2 stage sleep, slow waves of sleep correspond to N3 stage sleep, sawtooth waves correspond to R stage sleep, and so on.

[0065] The sleep staging method provided by the present application can detect the sleep characteristic signals and their types in the multi-channel EEG signals of the human body by using the sleep detection model component, and determine the sleep staging result of the human body according to the sleep characteristic signals and their types. Among them, the sleep characteristic detection model component is obtained by training with multiple multi-channel EEG sample signals, and the multi-channel EEG sample signals are annotated with sleep characteristic signals of different signal types. On the one hand, the sleep staging result is closely related to the sleep characteristic signal. Therefore, the sleep characteristic signal in the multi-channel EEG sample signal is first obtained by using the model component, and then the sleep staging result is determined according to the sleep characteristic signal, so that a relatively accurate result can be obtained; on the other hand, the sleep characteristic signal is annotated in the multi-channel EEG sample signal, instead of the sleep staging stage. From the perspective of the difficulty of feature extraction, the amount of sample data required to extract the sleep characteristic signal is much less than that required to extract the sleep staging stage. Therefore, the sleep staging method provided by the embodiment of the present application can be trained to obtain a sleep detection model component with a higher accuracy rate with less sample data.

[0066] Corresponding to the above sleep staging method, such as Figure 8 As shown, the present application also provides a sleep staging device, including a processor and a memory for storing processor executable instructions, and when the processor executes the instructions, the following can be achieved:

[0067] Obtain multi-channel EEG signals from the human body;

[0068] Inputting the multi-channel EEG signal into a sleep feature detection model component, and outputting a sleep feature signal in the multi-channel EEG signal and a signal type of the sleep feature signal through the sleep feature detection model component;

[0069] The sleep stage result of the human body is determined according to the sleep characteristic signal and the signal type.

[0070] Optionally, in one embodiment of the present application, the sleep feature detection model component is configured to be trained in the following manner:

[0071] Acquire multiple multi-channel EEG sample signals;

[0072] Extracting a sleep characteristic training signal from the multi-channel EEG sample signal;

[0073] Constructing a sleep feature detection model component, wherein the sleep feature detection model component is provided with training parameters;

[0074] respectively inputting the sleep characteristic training signals into the sleep characteristic detection model components to generate prediction results;

[0075] Based on the difference between the prediction result and the type of the sleep characteristic training signal, the training parameter is iteratively adjusted until the difference meets a preset requirement.

[0076] Optionally, in one embodiment of the present application, the processor inputs the multi-channel EEG signal into a sleep feature detection model component during the implementation step, and the sleep feature detection model component outputs a sleep feature signal in the multi-channel EEG signal and a signal type of the sleep feature signal, including:

[0077] A sliding window with a preset length is used to slide on the multi-channel EEG signal according to a preset step length, and the multi-channel EEG signal in the sliding window is input into a sleep feature detection model component during each sliding. The sleep feature detection model component outputs the sleep feature signal in the multi-channel EEG signal in the sliding window.

[0078] Optionally, in one embodiment of the present application, the processor, before implementing the step of inputting the multi-channel EEG signal into a sleep feature detection model component and outputting the sleep feature signal in the multi-channel EEG signal and the signal type of the sleep feature signal through the sleep feature detection model component, further includes:

[0079] Collecting electrocardiographic signals and / or electromyographic signals of the human body;

[0080] The multi-channel EEG signal, the electrocardiogram signal and / or the electromyography signal are input into an artifact detection model component, and the artifact detection model component outputs an artifact signal in the multi-channel EEG signal.

[0081] Optionally, in one embodiment of the present application, after the processor implements the step of outputting the sleep characteristic signal in the multi-channel EEG signal and the signal type of the sleep characteristic signal through the sleep characteristic detection model component, the processor further includes:

[0082] The position where the artifact signal and the sleep characteristic signal appear simultaneously in the multi-channel electroencephalogram signal is obtained, and the determination of the sleep characteristic signal at the position is canceled.

[0083] Optionally, in one embodiment of the present application, the processor, before implementing the step of inputting the multi-channel EEG signal into a sleep feature detection model component and outputting the sleep feature signal in the multi-channel EEG signal and the signal type of the sleep feature signal through the sleep feature detection model component, further includes:

[0084] The multi-channel EEG signal is subjected to at least one of the following signal preprocessing: removing baseline interference, removing power frequency interference, and removing non-EEG signals.

[0085] Optionally, in one embodiment of the present application, the processor, when implementing the step of determining the sleep stage result of the human body according to the sleep characteristic signal and the signal type, includes:

[0086] Based on preset sleep staging rules, the sleep staging result of the human body is determined according to the sleep characteristic signal and the signal type.

[0087] Optionally, in one embodiment of the present application, the sleep characteristic signal includes at least one of the following signals: apical wave, sleep spindle wave, K complex wave, sleep slow wave, alpha wave, and sawtooth wave.

[0088] On the other hand, the present application also provides a sleep staging device, which may include: an EEG acquisition device and the sleep staging device described in any of the above embodiments, wherein the EEG acquisition device is used to acquire the multi-channel EEG signal of the human body. The sleep staging device may include a smart helmet, smart bedding, smart hat and other products in various forms, which are not limited in the present application.

[0089] like Fig. 9 As shown, the present application also provides a sleep staging device 900, including: an EEG acquisition device 901, a data sending device 903, and a data receiving device 905, wherein:

[0090] The EEG acquisition device 901 is used to acquire the multi-channel EEG signals of the human body;

[0091] The data sending device 903 is used to send the multi-channel EEG signal to a server, and the server is configured to obtain the sleep characteristic signal in the multi-channel EEG signal and the signal type of the sleep characteristic signal by using the sleep characteristic detection model component, and determine the sleep staging result of the human body according to the sleep characteristic signal and the signal type;

[0092] The data receiving device 905 is used to receive the sleep staging result of the human body.

[0093] Optionally, in one embodiment of the present application, the device 900 may further include:

[0094] The display device 907 is used to display the sleep staging result.

[0095] Through the embodiments of the present application, the sleep staging device 900 can send the collected multi-channel EEG signals of the human body to the server side for processing. The server side may include any form of data processing server such as a cloud server and a distributed server. After processing the multi-channel EEG signals according to the sleep staging method adopted in each embodiment of the present application, the server side generates a sleep staging result of the human body and returns the sleep staging result to the user side. Based on this, users can understand their personal sleep staging results without leaving home. When encountering problems such as sleep disorders, they can send the obtained sleep staging results to the doctor to facilitate the doctor to determine the diagnosis results.

[0096] On the other hand, the present application further provides a computer-readable storage medium having computer instructions stored thereon, which, when executed, implement the steps of the method described in any of the above embodiments.

[0097] The computer-readable storage medium may include a physical device for storing information, which is usually to digitize the information and then store it in a medium using electrical, magnetic or optical means. The computer-readable storage medium described in this embodiment may include: devices that store information using electrical energy, such as various memories, such as RAM, ROM, etc.; devices that store information using magnetic energy, such as hard disks, floppy disks, magnetic tapes, magnetic core memories, bubble memories, and USB flash drives; devices that store information optically, such as CDs or DVDs. Of course, there are other types of readable storage media, such as quantum memories, graphene memories, and so on.

[0098] In the 1990s, improvements to a technology could be clearly distinguished as hardware improvements (for example, improvements to the circuit structure of diodes, transistors, switches, etc.) or software improvements (improvements to the method flow). However, with the development of technology, many improvements to the method flow today can be regarded as direct improvements to the hardware circuit structure. Designers almost always obtain the corresponding hardware circuit structure by programming the improved method flow into the hardware circuit. Therefore, it cannot be said that an improvement in a method flow cannot be implemented using a hardware entity module. For example, a programmable logic device (PLD) (such as a field programmable gate array (FPGA)) is such an integrated circuit whose logical function is determined by the user's programming of the device. Designers can "integrate" a digital system on a PLD by programming it themselves, without having to ask a chip manufacturer to design and produce a dedicated integrated circuit chip. Moreover, nowadays, instead of manually making integrated circuit chips, this kind of programming is mostly implemented by "logic compiler" software, which is similar to the software compiler used when developing and writing programs, and the original code before compilation must also be written in a specific programming language, which is called hardware description language (HDL). There is not only one HDL, but many kinds, such as ABEL (Advanced Boolean Expression Language), AHDL (Altera Hardware Description Language), Confluence, CUPL (Cornell University Programming Language), HDCal, JHDL (Java Hardware Description Language), Lava, Lola, MyHDL, PALASM, RHDL (Ruby Hardware Description Language), etc. The most commonly used ones are VHDL (Very-High-Speed ​​Integrated Circuit Hardware Description Language) and Verilog. Those skilled in the art should also know that it is only necessary to program the method flow slightly in the above-mentioned hardware description languages ​​and program it into the integrated circuit, and then it is easy to obtain the hardware circuit that implements the logic method flow.

[0099] The controller can be implemented in any appropriate manner, for example, the controller can take the form of a microprocessor or processor and a computer-readable medium storing a computer-readable program code (such as software or firmware) that can be executed by the (micro)processor, a logic gate, a switch, an application-specific integrated circuit (ASIC), a programmable logic controller, and an embedded microcontroller. Examples of controllers include, but are not limited to, the following microcontrollers: ARC625D, Atmel AT91SAM, Microchip PIC18F26K20, and Silicone Labs C8051F320. The memory controller can also be implemented as part of the control logic of the memory. Those skilled in the art also know that in addition to implementing the controller in a purely computer-readable program code manner, the controller can be implemented in the form of a logic gate, a switch, an application-specific integrated circuit, a programmable logic controller, and an embedded microcontroller by logically programming the method steps. Therefore, this controller can be considered as a hardware component, and the devices included therein for implementing various functions can also be regarded as structures within the hardware component. Or even, the devices for implementing various functions can be regarded as both software modules for implementing the method and structures within the hardware component.

[0100] The systems, devices, modules or units described in the above embodiments may be implemented by computer chips or entities, or by products with certain functions. A typical implementation device is a computer. Specifically, the computer may be, for example, a personal computer, a laptop computer, a cellular phone, a camera phone, a smart phone, a personal digital assistant, a media player, a navigation device, an email device, a game console, a tablet computer, a wearable device, or a combination of any of these devices.

[0101] For the convenience of description, the above device is described in various units according to their functions. Of course, when implementing the present application, the functions of each unit can be implemented in the same or multiple software and / or hardware.

[0102] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0103] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0104] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0105] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0106] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory. The memory may include non-permanent storage in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. Memory is an example of a computer-readable medium.

[0107] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

[0108] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0109] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.

[0110] The present application may be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. The present application may also be practiced in distributed computing environments where tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules may be located in local and remote computer storage media, including storage devices.

[0111] Each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the system embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0112] The above is only an embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the scope of the claims of the present application.

Claims

1. A sleep staging method, characterized in that: The method comprises: Obtain multi-channel EEG signals from the human body; Inputting the multi-channel EEG signal into a sleep feature detection model component, and outputting a sleep feature signal in the multi-channel EEG signal and a signal type of the sleep feature signal through the sleep feature detection model component; Determining a sleep stage result of the human body according to the sleep characteristic signal and the signal type; The step of inputting the multi-channel EEG signal into a sleep feature detection model component, and outputting a sleep feature signal in the multi-channel EEG signal and a signal type of the sleep feature signal through the sleep feature detection model component, includes: Using a sliding window with a preset length to slide on the multi-channel EEG signal according to a preset step length, and inputting the multi-channel EEG signal in the sliding window into a sleep feature detection model component at each sliding, and outputting a sleep feature signal in the multi-channel EEG signal in the sliding window through the sleep feature detection model component; The sleep feature detection model component is trained using a sleep feature training signal intercepted from the multi-channel EEG sample signal, and the window length of the sleep feature training signal intercepted is the same as the preset length of the sliding window.

2. The sleep staging method according to claim 1, characterized in that: The sleep feature detection model component is configured to be trained in the following manner: Acquire multiple multi-channel EEG sample signals; Extracting a sleep characteristic training signal from the multi-channel EEG sample signal; Constructing a sleep feature detection model component, wherein the sleep feature detection model component is provided with training parameters; respectively inputting the sleep characteristic training signals into the sleep characteristic detection model components to generate prediction results; Based on the difference between the prediction result and the type of the sleep characteristic training signal, the training parameter is iteratively adjusted until the difference meets a preset requirement.

3. The sleep staging method according to claim 1, characterized in that: Before inputting the multi-channel EEG signal into the sleep feature detection model component and the sleep feature detection model component outputting the sleep feature signal in the multi-channel EEG signal and the signal type of the sleep feature signal, the method further includes: Collecting electrocardiogram signals and / or electromyography signals of the human body; The multi-channel EEG signal, the electrocardiogram signal and / or the electromyography signal are input into an artifact detection model component, and the artifact detection model component outputs an artifact signal in the multi-channel EEG signal.

4. The sleep staging method according to claim 3, characterized in that: After the sleep characteristic detection model component outputs the sleep characteristic signal in the multi-channel EEG signal and the signal type of the sleep characteristic signal, the method further includes: The position where the artifact signal and the sleep characteristic signal appear simultaneously in the multi-channel electroencephalogram signal is obtained, and the determination of the sleep characteristic signal at the position is canceled.

5. The sleep staging method according to claim 1, characterized in that: Before inputting the multi-channel EEG signal into the sleep feature detection model component and the sleep feature detection model component outputting the sleep feature signal in the multi-channel EEG signal and the signal type of the sleep feature signal, the method further includes: The multi-channel EEG signal is subjected to at least one of the following signal preprocessing: removing baseline interference, removing power frequency interference, and removing non-EEG signals.

6. The sleep staging method according to claim 1, characterized in that: Determining the sleep stage result of the human body according to the sleep characteristic signal and the signal type includes: Based on preset sleep staging rules, the sleep staging result of the human body is determined according to the sleep characteristic signal and the signal type.

7. The sleep staging method according to claim 1, characterized in that: The sleep characteristic signal includes at least one of the following signals: top wave, sleep spindle wave, K complex wave, sleep slow wave, alpha wave, and sawtooth wave.

8. A sleep staging device, characterized in that: The invention comprises a processor and a memory for storing processor-executable instructions, wherein when the processor executes the instructions, the following are achieved: Obtain multi-channel EEG signals from the human body; Inputting the multi-channel EEG signal into a sleep feature detection model component, and outputting a sleep feature signal in the multi-channel EEG signal and a signal type of the sleep feature signal through the sleep feature detection model component; Determining a sleep stage result of the human body according to the sleep characteristic signal and the signal type; The step of inputting the multi-channel EEG signal into a sleep feature detection model component, and outputting a sleep feature signal in the multi-channel EEG signal and a signal type of the sleep feature signal through the sleep feature detection model component, includes: Using a sliding window with a preset length to slide on the multi-channel EEG signal according to a preset step length, and inputting the multi-channel EEG signal in the sliding window into a sleep feature detection model component at each sliding, and outputting a sleep feature signal in the multi-channel EEG signal in the sliding window through the sleep feature detection model component; The sleep feature detection model component is trained using a sleep feature training signal intercepted from the multi-channel EEG sample signal, and the window length of the sleep feature training signal intercepted is the same as the preset length of the sliding window.

9. The sleep staging device according to claim 8, characterized in that: The sleep feature detection model component is configured to be trained in the following manner: Acquire multiple multi-channel EEG sample signals; Extracting a sleep characteristic training signal from the multi-channel EEG sample signal; Constructing a sleep feature detection model component, wherein the sleep feature detection model component is provided with training parameters; respectively inputting the sleep characteristic training signals into the sleep characteristic detection model components to generate prediction results; Based on the difference between the prediction result and the type of the sleep characteristic training signal, the training parameter is iteratively adjusted until the difference meets a preset requirement.

10. The sleep staging device according to claim 8, characterized in that: The processor, before implementing the step of inputting the multi-channel EEG signal into the sleep feature detection model component and the sleep feature detection model component outputting the sleep feature signal in the multi-channel EEG signal and the signal type of the sleep feature signal, further includes: Collecting electrocardiogram signals and / or electromyography signals of the human body; The multi-channel EEG signal, the electrocardiogram signal and / or the electromyography signal are input into an artifact detection model component, and the artifact detection model component outputs an artifact signal in the multi-channel EEG signal.

11. The sleep staging device according to claim 10, characterized in that: After the processor implements the step of outputting the sleep characteristic signal in the multi-channel EEG signal and the signal type of the sleep characteristic signal through the sleep characteristic detection model component, the processor further includes: The position where the artifact signal and the sleep characteristic signal appear simultaneously in the multi-channel electroencephalogram signal is obtained, and the determination of the sleep characteristic signal at the position is canceled.

12. The sleep staging device according to claim 8, characterized in that: The processor, before implementing the step of inputting the multi-channel EEG signal into the sleep feature detection model component and the sleep feature detection model component outputting the sleep feature signal in the multi-channel EEG signal and the signal type of the sleep feature signal, further includes: The multi-channel EEG signal is subjected to at least one of the following signal preprocessing: removing baseline interference, removing power frequency interference, and removing non-EEG signals.

13. The sleep staging device according to claim 8, characterized in that: The processor, when implementing the step of determining the sleep stage result of the human body according to the sleep characteristic signal and the signal type, includes: Based on preset sleep staging rules, the sleep staging result of the human body is determined according to the sleep characteristic signal and the signal type.

14. The sleep staging device according to claim 8, characterized in that: The sleep characteristic signal includes at least one of the following signals: top wave, sleep spindle wave, K complex wave, sleep slow wave, alpha wave, and sawtooth wave.

15. A sleep staging device, characterized in that: include: An electroencephalogram acquisition device and a sleep staging device according to any one of claims 8 to 14, wherein: The EEG acquisition device is used to acquire the multi-channel EEG signals of the human body.

16. A sleep staging device, characterized in that: include: EEG acquisition device, data sending device, data receiving device, wherein: The EEG acquisition device is used to acquire multi-channel EEG signals of a human body; The data sending device is used to send the multi-channel EEG signal to a server, and the server adopts the sleep staging method according to any one of claims 1 to 7; The data receiving device is used to receive the sleep staging result of the human body.

17. The sleep staging device according to claim 16, characterized in that: The device also includes: A display device is used to display the sleep staging result.

18. A non-transitory computer-readable storage medium, characterized in that: When the instructions in the storage medium are executed by a processor, the processor is enabled to execute the sleep staging method described in any one of claims 1 to 7.

Images