Sleep respiration diagnosis method and system based on simple brain electrode
Multimodal sleep data is obtained through simple brain electrodes and PSG devices, preprocessing and feature extraction are performed, and the problems of complex equipment operation and low diagnostic accuracy in the prior art are solved, and high-precision sleep breathing diagnosis is achieved.
Patent Information
- Application Number
- CN202411351975.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-26
- Publication Date
- 2024-12-27
AI Technical Summary
The existing polysomnography monitors are complex in operation, limiting their application in the home, and traditional sleep breathing diagnosis uses single-modal data, reducing the accuracy of diagnostic results.
EEG data is obtained through simple brain electrodes, sleep breathing data is obtained in combination with PSG devices, time sequence data processing and preprocessing is performed, and electrocardiogram, electroencephalogram, blood oxygen and chest and abdominal motion characteristics data are extracted using multimodal feature extraction model, and the data are carried out. Normalization and parallel fusion, sleep breathing diagnosis is performed through classifiers.
The equipment wearing operation is simplified, the accuracy of sleep breathing diagnosis is improved, and the accuracy of diagnosis results is enhanced through joint analysis of multimodal data.
Smart Images
Figure CN119184622A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of sleep monitoring, and in particular relates to a sleep breathing diagnosis method and system based on simple brain electrodes. Background Art
[0002] Sleep apnea syndrome is a common sleep disorder characterized by repeated apnea and hypopnea during sleep, leading to nocturnal hypoxia and decreased sleep quality. Currently, polysomnography is the most authoritative method for diagnosing sleep apnea syndrome.
[0003] However, current polysomnography equipment is complex to operate and usually needs to be performed in hospitals or professional sleep centers, which limits its application at home. Traditional sleep apnea diagnosis uses single-modality data for sleep apnea diagnosis, which reduces the accuracy of the diagnostic results. Summary of the invention
[0004] In order to solve the above problems existing in the prior art, the present invention provides a sleep respiratory disease diagnosis method and system based on simple brain electrodes.
[0005] The purpose of the present invention can be achieved through the following technical solutions:
[0006] S1: obtaining EEG data through simple brain electrodes, obtaining sleep breathing data through PSG equipment, adding timestamps to the sleep breathing data and the EEG data to obtain time-series sleep breathing data and time-series EEG data, wherein the time-series sleep breathing data includes: time-series ECG data, time-series blood oxygen saturation data, and time-series chest and abdominal movement data;
[0007] S2: Calculate the ECG preprocessing data through the ECG preprocessing model according to the time-series ECG data, calculate the EEG preprocessing data through the EEG preprocessing model according to the time-series EEG data, preset a blood oxygen saturation threshold, remove the time-series blood oxygen saturation data according to the blood oxygen saturation threshold to obtain blood oxygen saturation preprocessing data, and calculate the chest and abdominal motion preprocessing data through the chest and abdominal motion preprocessing model according to the time-series chest and abdominal motion data;
[0008] S3: Preset a unit time length, and perform segmentation processing on the ECG preprocessing data, the EEG preprocessing data, the blood oxygen saturation preprocessing data, and the chest and abdominal movement preprocessing data according to the unit time length to obtain an ECG segmentation sequence, an EEG segmentation sequence, a blood oxygen saturation segmentation sequence, and a chest and abdominal movement segmentation sequence;
[0009] S4: obtaining ECG feature data by using an ECG feature extraction model according to the ECG segmented sequence, obtaining EEG feature data by using an EEG feature extraction model according to the EEG segmented sequence, obtaining blood oxygen feature data by using a blood oxygen feature extraction model according to the blood oxygen saturation segmented sequence, and obtaining chest and abdominal motion feature data by using a motion feature extraction model according to the chest and abdominal motion segmented sequence;
[0010] S5: Normalize the ECG characteristic data, the EEG characteristic data, the blood oxygen characteristic data, and the chest and abdominal movement characteristic data to obtain ECG normalized data, EEG normalized data, blood oxygen normalized data, and chest and abdominal movement normalized data; perform parallel fusion on the ECG normalized data, the EEG normalized data, the blood oxygen normalized data, and the chest and abdominal movement normalized data to obtain a splicing vector; and calculate the sleep breathing diagnostic data through a classifier based on the splicing vector.
[0011] Specifically, the ECG preprocessing model obtains ECG filtered data by filtering power frequency interference and myoelectric interference through a Butterworth low-pass filter according to the time-series ECG data, and obtains the ECG preprocessing data by removing baseline drift through a median filter according to the ECG filtered data. The frequency expression formula of the Butterworth low-pass filter is:
[0012]
[0013] Among them, a is the normalized adjustment parameter, w is the frequency of the Butterworth low-pass filter, n is the filter order, and w p is the passband edge frequency.
[0014] Specifically, the EEG preprocessing model segments the time-series EEG data through a preset time window to obtain EEG segments, decomposes the EEG segments through complete empirical mode decomposition to obtain EEG decomposition data, selects the intrinsic mode components of the electrooculogram signals of the EEG decomposition data, identifies the electrooculogram peak position through a constant false alarm algorithm based on the intrinsic mode components of the electrooculogram signals, sets a small segment of the signal centered on the electrooculogram peak position in the intrinsic mode components of the electrooculogram signals to zero to obtain a zeroed signal, and obtains the EEG preprocessing data through inverse CEEMAN reconstruction based on the zeroed signal.
[0015] Specifically, the thoraco-abdominal motion preprocessing model is used to eliminate the large fluctuation data in the time-series thoraco-abdominal motion data caused by interference such as coughing and talking. First, the outliers are removed from the time-series thoraco-abdominal motion data through the box plot method to obtain the cleaned data, and then the thoraco-abdominal motion preprocessing data is obtained by smoothing and denoising the cleaned data through the SG filter.
[0016] Specifically, the ECG feature extraction model is used to extract feature data of each ECG segmented data in the ECG segmented sequence. The ECG feature extraction model performs R wave positioning according to the ECG segmented data through the PT algorithm to obtain an R wave signal, extracts the interval signal of the R wave signal according to the R wave time interval to obtain an RR interval signal, extracts the peak signal of the R wave signal, performs cubic spline interpolation processing on the peak signal to obtain an EDR signal, calculates the time domain feature according to the RR interval signal through the HRV linear analysis method, calculates the frequency domain feature according to the EDR signal through the HRV linear analysis method, and splices the time domain feature and the frequency domain feature to obtain the ECG feature data.
[0017] Specifically, the EEG feature extraction model is used to extract feature data of each EEG segmented data in the EEG segmented sequence, and obtain EEG time domain features and EEG frequency domain features through wavelet transformation calculation according to the EEG segmented data, and obtain EEG entropy features through time series nonlinear detection method according to the EEG segmented data, and the time domain features, the frequency domain features, and the EEG entropy features are combined to obtain a feature set, and the EEG feature data is obtained by feature selection through the RelifF algorithm according to the feature set. The calculation formula of the RelifF algorithm is:
[0018]
[0019] Where w(A) is the weight of feature A, E is a randomly selected sample, Si is the i-th closest sample of the same type to sample E, m is the number of iterations, k is the number of nearest neighbor samples, and D i (REM) is the i-th sample in the REM category that is closest to sample E, P(REM) is the proportion of the REM category in the data set, and diff(A,E,S i ) is the feature A in sample R and sample S i The difference measure between .
[0020] Specifically, the motion feature extraction model is used to extract feature data of each segment of thoracoabdominal motion segment data in the thoracoabdominal motion segment sequence. The motion feature extraction model determines the signal trough position according to the thoracoabdominal motion segment data by a moving average method, calibrates the thoracoabdominal motion segment data by a qualitative diagnosis calibration method to obtain the tidal volume, and performs parameter quantitative analysis according to the trough position and the tidal volume to obtain the thoracoabdominal motion feature data. The expression of the qualitative diagnosis calibration method is:
[0021] W T =M*(K*V RC +V AB ),
[0022] Among them, WT is the calibrated tidal volume, K is the relative scaling factor between the uncalibrated thoracic and abdominal signals, M is the scaling factor for scaling the relative tidal volume in arbitrary units to the tidal volume in absolute volume units, V RC is the chest tidal volume, V AB is the abdominal tidal volume.
[0023] A system for sleep breathing diagnosis based on simple brain electrodes using the sleep breathing diagnosis method based on simple brain electrodes, characterized by comprising: a data acquisition module, a preprocessing module, a segmentation module, a feature extraction module, and a data analysis module;
[0024] The data acquisition module is used to acquire EEG data through simple brain electrodes, acquire sleep breathing data through PSG equipment, add time stamps to the sleep breathing data and the EEG data to obtain time series sleep breathing data and time series EEG data, wherein the time series sleep breathing data includes: time series ECG data, time series blood oxygen saturation data, and time series chest and abdominal movement data;
[0025] The preprocessing module is used to calculate the ECG preprocessing data through the ECG preprocessing model according to the time-series ECG data, calculate the EEG preprocessing data through the EEG preprocessing model according to the time-series EEG data, preset a blood oxygen saturation threshold, remove the time-series blood oxygen saturation data according to the blood oxygen saturation threshold to obtain the blood oxygen saturation preprocessing data, and calculate the chest and abdominal motion preprocessing data through the chest and abdominal motion preprocessing model according to the time-series chest and abdominal motion data;
[0026] The segmentation module is used to preset a unit time length, and segment the ECG preprocessing data, the EEG preprocessing data, the blood oxygen saturation preprocessing data, and the chest and abdominal movement preprocessing data according to the unit time length to obtain an ECG segmentation sequence, an EEG segmentation sequence, a blood oxygen saturation segmentation sequence, and a chest and abdominal movement segmentation sequence;
[0027] The feature extraction module is used to calculate the ECG feature data through the ECG feature extraction model according to the ECG segmented sequence, calculate the EEG feature data through the EEG feature extraction model according to the EEG segmented sequence, calculate the blood oxygen feature data through the blood oxygen feature extraction model according to the blood oxygen saturation segmented sequence, and calculate the chest and abdominal motion feature data through the motion feature extraction model according to the chest and abdominal motion segmented sequence;
[0028] The data analysis module is used to normalize the ECG characteristic data, the EEG characteristic data, the blood oxygen characteristic data, and the chest and abdominal movement characteristic data to obtain ECG normalized data, EEG normalized data, blood oxygen normalized data, and chest and abdominal movement normalized data, and to perform parallel fusion of the ECG normalized data, the EEG normalized data, the blood oxygen normalized data, and the chest and abdominal movement normalized data to obtain a splicing vector, and to obtain sleep breathing diagnostic data through classifier calculation based on the splicing vector.
[0029] An electronic device comprises a memory, a processor and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the sleep breathing diagnosis method based on simple brain electrodes as described above is implemented.
[0030] A storage medium containing computer executable instructions, characterized in that the computer executable instructions are used to execute the sleep breathing diagnosis method based on simple brain electrodes as described above when executed by a computer processor.
[0031] The beneficial effects of the present invention are:
[0032] (1) The polysomnogram monitor is equipped with simple brain electrodes. The patient only needs to simply position and stick the active electrode. The ground electrode and the reference electrode can be naturally extended to the specified position by setting the angle of the connector. The length of the connector can be set to make local fine adjustments, and the three-lead brain electrodes can be stuck to the required position at one time, which simplifies the number and complexity of independent positioning and wearing operations for each electrode.
[0033] (2) By collecting ECG data, EEG data, chest and abdominal movement data, and blood oxygen saturation data, and extracting the features of the ECG data, EEG data, chest and abdominal movement data, and blood oxygen saturation data, the features of the ECG data, EEG data, chest and abdominal movement data, and blood oxygen saturation data are normalized and fused in parallel to obtain a spliced vector, and the spliced vector is transmitted to a classifier for classification processing to obtain sleep breathing diagnosis data. By performing joint analysis of multimodal data, the accuracy of the sleep breathing diagnosis data is improved.
[0034] (3) By collecting chest and abdominal movement data, preprocessing and extracting features from the chest and abdominal movement data to obtain chest and abdominal movement feature data, and jointly analyzing the chest and abdominal movement feature data, the accuracy of sleep breathing is greatly improved. BRIEF DESCRIPTION OF THE DRAWINGS
[0035] In order to facilitate understanding by those skilled in the art, the present invention is further described below with reference to the accompanying drawings.
[0036] Figure 1It is a flow chart of the sleep breathing disease diagnosis method based on simple brain electrodes of the present invention. DETAILED DESCRIPTION
[0037] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the specific implementation methods, structures, features and effects of the present invention are described in detail below in conjunction with the accompanying drawings and preferred embodiments.
[0038] See also Figure 1 , a sleep breathing diagnosis method and system based on simple brain electrodes;
[0039] S1: obtaining EEG data through simple brain electrodes, obtaining sleep breathing data through PSG equipment, adding timestamps to the sleep breathing data and the EEG data to obtain time-series sleep breathing data and time-series EEG data, wherein the time-series sleep breathing data includes: time-series ECG data, time-series blood oxygen saturation data, and time-series chest and abdominal movement data;
[0040] S2: Calculate the ECG preprocessing data through the ECG preprocessing model according to the time-series ECG data, calculate the EEG preprocessing data through the EEG preprocessing model according to the time-series EEG data, preset a blood oxygen saturation threshold, remove the time-series blood oxygen saturation data according to the blood oxygen saturation threshold to obtain blood oxygen saturation preprocessing data, and calculate the chest and abdominal motion preprocessing data through the chest and abdominal motion preprocessing model according to the time-series chest and abdominal motion data;
[0041] S3: Preset a unit time length, and perform segmentation processing on the ECG preprocessing data, the EEG preprocessing data, the blood oxygen saturation preprocessing data, and the chest and abdominal movement preprocessing data according to the unit time length to obtain an ECG segmentation sequence, an EEG segmentation sequence, a blood oxygen saturation segmentation sequence, and a chest and abdominal movement segmentation sequence;
[0042] S4: obtaining ECG feature data by using an ECG feature extraction model according to the ECG segmented sequence, obtaining EEG feature data by using an EEG feature extraction model according to the EEG segmented sequence, obtaining blood oxygen feature data by using a blood oxygen feature extraction model according to the blood oxygen saturation segmented sequence, and obtaining chest and abdominal motion feature data by using a motion feature extraction model according to the chest and abdominal motion segmented sequence;
[0043] S5: Normalize the ECG characteristic data, the EEG characteristic data, the blood oxygen characteristic data, and the chest and abdominal movement characteristic data to obtain ECG normalized data, EEG normalized data, blood oxygen normalized data, and chest and abdominal movement normalized data; perform parallel fusion on the ECG normalized data, the EEG normalized data, the blood oxygen normalized data, and the chest and abdominal movement normalized data to obtain a splicing vector; and calculate the sleep breathing diagnostic data through a classifier based on the splicing vector.
[0044] In this embodiment, the simple brain electrode is an integrated three-lead disposable conductive hydrogel brain electrode, which is composed of an active electrode, a ground electrode, and a reference electrode. Through the integrated design, the patient only needs to simply position and stick the active electrode. The ground electrode and the reference electrode can be naturally extended to the designated position by setting the angle of the connector, and local fine-tuning can be performed by setting the length of the connector. The three-lead brain electrodes can be stuck to the desired position at one time, which simplifies the number and complexity of independent positioning and wearing operations for each electrode.
[0045] Specifically, the ECG preprocessing model obtains ECG filtered data by filtering power frequency interference and myoelectric interference through a Butterworth low-pass filter according to the time-series ECG data, and obtains the ECG preprocessing data by removing baseline drift through a median filter according to the ECG filtered data. The frequency expression formula of the Butterworth low-pass filter is:
[0046]
[0047] Among them, a is the normalized adjustment parameter, w is the frequency of the Butterworth low-pass filter, n is the filter order, and w p is the passband edge frequency.
[0048] Specifically, the EEG preprocessing model segments the time-series EEG data through a preset time window to obtain EEG segments, decomposes the EEG segments through complete empirical mode decomposition to obtain EEG decomposition data, selects the intrinsic mode components of the electrooculogram signals of the EEG decomposition data, identifies the electrooculogram peak position through a constant false alarm algorithm based on the intrinsic mode components of the electrooculogram signals, sets a small segment of the signal centered on the electrooculogram peak position in the intrinsic mode components of the electrooculogram signals to zero to obtain a zeroed signal, and obtains the EEG preprocessing data through inverse CEEMAN reconstruction based on the zeroed signal.
[0049] Specifically, the thoraco-abdominal motion preprocessing model is used to eliminate the large fluctuation data in the time-series thoraco-abdominal motion data caused by interference such as coughing and talking. First, the outliers are removed from the time-series thoraco-abdominal motion data through the box plot method to obtain the cleaned data, and then the thoraco-abdominal motion preprocessing data is obtained by smoothing and denoising the cleaned data through the SG filter.
[0050] Specifically, the ECG feature extraction model is used to extract feature data of each ECG segmented data in the ECG segmented sequence. The ECG feature extraction model performs R wave positioning according to the ECG segmented data through the PT algorithm to obtain an R wave signal, extracts the interval signal of the R wave signal according to the R wave time interval to obtain an RR interval signal, extracts the peak signal of the R wave signal, performs cubic spline interpolation processing on the peak signal to obtain an EDR signal, calculates the time domain feature according to the RR interval signal through the HRV linear analysis method, calculates the frequency domain feature according to the EDR signal through the HRV linear analysis method, and splices the time domain feature and the frequency domain feature to obtain the ECG feature data.
[0051] In this embodiment, the time domain features include: the RR interval average value, the root mean square of the difference between adjacent RR intervals, the standard deviation of the RR interval, the number of adjacent RR interval differences greater than 50ms (NN50), the ratio of NN50 to the number of RR intervals, and heart rate; the frequency domain features include: ultra-low frequency components, low frequency components, high frequency components, the ratio of high and low frequencies, normalized low frequency components, and normalized high frequency components.
[0052] Specifically, the EEG feature extraction model is used to extract feature data of each EEG segmented data in the EEG segmented sequence, and obtain EEG time domain features and EEG frequency domain features through wavelet transformation calculation according to the EEG segmented data, and obtain EEG entropy features through time series nonlinear detection method according to the EEG segmented data, and the time domain features, the frequency domain features, and the EEG entropy features are combined to obtain a feature set, and the EEG feature data is obtained by feature selection through the RelifF algorithm according to the feature set. The calculation formula of the RelifF algorithm is:
[0053]
[0054] Where w(A) is the weight of feature A, E is a randomly selected sample, Si is the i-th closest sample of the same type to sample E, m is the number of iterations, k is the number of nearest neighbor samples, and D i (REM) is the i-th sample in the REM category that is closest to sample E, P(REM) is the proportion of the REM category in the data set, and diff(A,E,S i ) is the feature A in sample R and sample S i The difference measure between .
[0055] In this embodiment, the EEG entropy features include approximate entropy, power spectrum entropy, Renyi entropy, and permutation entropy.
[0056] Specifically, the blood oxygen feature extraction model extracts feature data of each segment of blood oxygen saturation segmented data in the blood oxygen saturation segmented sequence based on statistical methods and nonlinear methods. The blood oxygen feature extraction model calculates basic statistical features according to the blood oxygen saturation segmented data by statistical methods, and calculates nonlinear features according to the blood oxygen saturation segmented data by nonlinear methods. The basic statistical features and the nonlinear features are combined to obtain the blood oxygen feature data. The basic statistical features include minimum value, mean value, and variance, and the nonlinear features include CTM, LZC, and approximate entropy.
[0057] Specifically, the motion feature extraction model is used to extract feature data of each segment of thoracoabdominal motion segment data in the thoracoabdominal motion segment sequence. The motion feature extraction model determines the signal trough position according to the thoracoabdominal motion segment data by a moving average method, calibrates the thoracoabdominal motion segment data by a qualitative diagnosis calibration method to obtain the tidal volume, and performs parameter quantitative analysis according to the trough position and the tidal volume to obtain the thoracoabdominal motion feature data. The expression of the qualitative diagnosis calibration method is:
[0058] W T =M*(K*V RC +V AB ),
[0059] Among them, W T is the calibrated tidal volume, K is the relative scaling factor between the uncalibrated thoracic and abdominal signals, M is the scaling factor for scaling the relative tidal volume in arbitrary units to the tidal volume in absolute volume units, V RC is the chest tidal volume, V AB is the abdominal tidal volume.
[0060] In this embodiment, the chest-abdomen movement characteristic data include respiratory rate, inspiration / expiration time, inspiration time fraction, chest-respiration contribution ratio, chest-abdomen phase difference, and chest-abdomen inspiration airflow.
[0061] A system for sleep breathing diagnosis based on simple brain electrodes using the sleep breathing diagnosis method based on simple brain electrodes, characterized by comprising: a data acquisition module, a preprocessing module, a segmentation module, a feature extraction module, and a data analysis module;
[0062] The data acquisition module is used to acquire EEG data through simple brain electrodes, acquire sleep breathing data through PSG equipment, add time stamps to the sleep breathing data and the EEG data to obtain time series sleep breathing data and time series EEG data, wherein the time series sleep breathing data includes: time series ECG data, time series blood oxygen saturation data, and time series chest and abdominal movement data;
[0063] The preprocessing module is used to calculate the ECG preprocessing data through the ECG preprocessing model according to the time-series ECG data, calculate the EEG preprocessing data through the EEG preprocessing model according to the time-series EEG data, preset a blood oxygen saturation threshold, remove the time-series blood oxygen saturation data according to the blood oxygen saturation threshold to obtain the blood oxygen saturation preprocessing data, and calculate the chest and abdominal motion preprocessing data through the chest and abdominal motion preprocessing model according to the time-series chest and abdominal motion data;
[0064] The segmentation module is used to preset a unit time length, and segment the ECG preprocessing data, the EEG preprocessing data, the blood oxygen saturation preprocessing data, and the chest and abdominal movement preprocessing data according to the unit time length to obtain an ECG segmentation sequence, an EEG segmentation sequence, a blood oxygen saturation segmentation sequence, and a chest and abdominal movement segmentation sequence;
[0065] The feature extraction module is used to calculate the ECG feature data through the ECG feature extraction model according to the ECG segmented sequence, calculate the EEG feature data through the EEG feature extraction model according to the EEG segmented sequence, calculate the blood oxygen feature data through the blood oxygen feature extraction model according to the blood oxygen saturation segmented sequence, and calculate the chest and abdominal motion feature data through the motion feature extraction model according to the chest and abdominal motion segmented sequence;
[0066] The data analysis module is used to normalize the ECG characteristic data, the EEG characteristic data, the blood oxygen characteristic data, and the chest and abdominal movement characteristic data to obtain ECG normalized data, EEG normalized data, blood oxygen normalized data, and chest and abdominal movement normalized data, and to perform parallel fusion of the ECG normalized data, the EEG normalized data, the blood oxygen normalized data, and the chest and abdominal movement normalized data to obtain a splicing vector, and to obtain sleep breathing diagnostic data through classifier calculation based on the splicing vector.
[0067] An electronic device comprises a memory, a processor and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the sleep breathing diagnosis method based on simple brain electrodes as described above is implemented.
[0068] A storage medium containing computer executable instructions, characterized in that the computer executable instructions are used to execute the sleep breathing diagnosis method based on simple brain electrodes as described above when executed by a computer processor.
[0069] The computer storage medium of the embodiment of the present invention may adopt any combination of one or more computer-readable media. The computer-readable medium may be a computer-readable signal medium or a computer-readable storage medium. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples (non-exhaustive list) of computer-readable storage media include: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In this document, a computer-readable storage medium may be any tangible medium containing or storing a program that may be used by or in combination with an instruction execution system, device or device.
[0070] Computer-readable signal media may include data signals propagated in baseband or as part of a carrier wave, which carry computer-readable program code. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. Computer-readable signal media may also be any computer-readable medium other than a computer-readable storage medium, which may send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, apparatus, or device.
[0071] The program code included on the computer readable medium can be transmitted with any appropriate medium, including but not limited to wireless, electric wire, optical cable, RF, etc., or any suitable combination of the above. The computer program code for performing the operation of the present invention can be written in one or more programming languages or their combinations, and the programming language includes object-oriented programming languages-such as Java, Smalltalk, C++, and also includes conventional procedural programming languages-such as "C" language or similar programming languages. The program code can be executed completely on the user's computer, partially on the user's computer, as an independent software package, partially on the user's computer and partially on the remote computer, or completely on the remote computer or server. In the case of a remote computer, the remote computer can be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computer (for example, using an Internet service provider to connect through the Internet).
[0072] The above description is only a preferred embodiment of the present invention and does not limit the present invention in any form. Although the present invention has been disclosed as a preferred embodiment as above, it is not used to limit the present invention. Any technical personnel in this field can make some changes or modify the technical contents disclosed above into equivalent embodiments without departing from the scope of the technical solution of the present invention. However, any simple modification, equivalent change and modification made to the above embodiments according to the technical essence of the present invention without departing from the content of the technical solution of the present invention still fall within the scope of the technical solution of the present invention.
Claims
1. A sleep breathing diagnosis method based on simple brain electrodes, characterized in that: include: S1: obtaining EEG data through simple brain electrodes, obtaining sleep breathing data through PSG equipment, adding timestamps to the sleep breathing data and the EEG data to obtain time-series sleep breathing data and time-series EEG data, wherein the time-series sleep breathing data includes: time-series ECG data, time-series blood oxygen saturation data, and time-series chest and abdominal movement data; S2: Calculate the ECG preprocessing data through the ECG preprocessing model according to the time-series ECG data, calculate the EEG preprocessing data through the EEG preprocessing model according to the time-series EEG data, preset a blood oxygen saturation threshold, remove the time-series blood oxygen saturation data according to the blood oxygen saturation threshold to obtain blood oxygen saturation preprocessing data, and calculate the chest and abdominal motion preprocessing data through the chest and abdominal motion preprocessing model according to the time-series chest and abdominal motion data; S3: Preset a unit time length, and perform segmentation processing on the ECG preprocessing data, the EEG preprocessing data, the blood oxygen saturation preprocessing data, and the chest and abdominal movement preprocessing data according to the unit time length to obtain an ECG segmentation sequence, an EEG segmentation sequence, a blood oxygen saturation segmentation sequence, and a chest and abdominal movement segmentation sequence; S4: obtaining ECG feature data by using an ECG feature extraction model according to the ECG segmented sequence, obtaining EEG feature data by using an EEG feature extraction model according to the EEG segmented sequence, obtaining blood oxygen feature data by using a blood oxygen feature extraction model according to the blood oxygen saturation segmented sequence, and obtaining chest and abdominal motion feature data by using a motion feature extraction model according to the chest and abdominal motion segmented sequence; S5: Normalize the ECG characteristic data, the EEG characteristic data, the blood oxygen characteristic data, and the chest and abdominal movement characteristic data to obtain ECG normalized data, EEG normalized data, blood oxygen normalized data, and chest and abdominal movement normalized data; perform parallel fusion on the ECG normalized data, the EEG normalized data, the blood oxygen normalized data, and the chest and abdominal movement normalized data to obtain a splicing vector; and calculate the sleep breathing diagnostic data through a classifier based on the splicing vector.
2. The sleep breathing diagnosis method based on simple brain electrodes according to claim 1, characterized in that: The ECG preprocessing model obtains ECG filtered data by filtering power frequency interference and myoelectric interference through a Butterworth low-pass filter according to the time-series ECG data, and obtains the ECG preprocessing data by removing baseline drift through a median filter according to the ECG filtered data. The frequency expression formula of the Butterworth low-pass filter is: Among them, a is the normalized adjustment parameter, w is the frequency of the Butterworth low-pass filter, n is the filter order, and w p is the passband edge frequency.
3. The sleep breathing diagnosis method based on simple brain electrodes according to claim 1, characterized in that: The EEG preprocessing model segments the time-series EEG data through a preset time window to obtain EEG segments, decomposes the EEG segments through complete empirical mode decomposition to obtain EEG decomposition data, selects the intrinsic mode components of the electrooculogram signals of the EEG decomposition data, identifies the electrooculogram peak position through a constant false alarm algorithm based on the intrinsic mode components of the electrooculogram signals, sets a small segment signal centered on the electrooculogram peak position in the intrinsic mode components of the electrooculogram signals to zero to obtain a zero-set signal, and obtains the EEG preprocessing data through inverse CEEMAN reconstruction based on the zero-set signal.
4. The sleep breathing diagnosis method based on simple brain electrodes according to claim 1, characterized in that: The thoraco-abdominal motion preprocessing model is used to eliminate the large fluctuation data in the time-series thoraco-abdominal motion data caused by interference such as coughing and talking. First, the outliers are removed by the box plot method according to the time-series thoraco-abdominal motion data to obtain the cleaned data, and then the thoraco-abdominal motion preprocessing data is obtained by smoothing and denoising the cleaned data through the SG filter.
5. The sleep breathing diagnosis method based on simple brain electrodes according to claim 1, characterized in that: The ECG feature extraction model is used to extract feature data of each ECG segmented data in the ECG segmented sequence. The ECG feature extraction model performs R wave positioning through the PT algorithm according to the ECG segmented data to obtain an R wave signal, extracts the interval signal of the R wave signal according to the R wave time interval to obtain an RR interval signal, extracts the peak signal of the R wave signal, performs cubic spline interpolation processing on the peak signal to obtain an EDR signal, calculates the time domain feature through the HRV linear analysis method according to the RR interval signal, calculates the frequency domain feature through the HRV linear analysis method according to the EDR signal, and splices the time domain feature and the frequency domain feature to obtain the ECG feature data.
6. The sleep breathing diagnosis method based on simple brain electrodes according to claim 1, characterized in that: The EEG feature extraction model is used to extract feature data of each EEG segmented data in the EEG segmented sequence, calculate the EEG time domain features and EEG frequency domain features according to the EEG segmented data through wavelet transformation, calculate the EEG entropy features according to the EEG segmented data through a time series nonlinear detection method, collect the time domain features, the frequency domain features, and the EEG entropy features to obtain a feature set, perform feature selection according to the feature set through the RelifF algorithm to obtain the EEG feature data, and the RelifF algorithm calculation formula is: Where w(A) is the weight of feature A, E is a randomly selected sample, Si is the i-th closest sample of the same type to sample E, m is the number of iterations, k is the number of nearest neighbor samples, and D i (REM) is the i-th sample in the REM category that is closest to sample E, P(REM) is the proportion of the REM category in the data set, and diff(A,E,S i ) is the feature A in sample R and sample S i The difference measure between .
7. The sleep breathing diagnosis method based on simple brain electrodes according to claim 1, characterized in that: The motion feature extraction model is used to extract feature data of each segment of thoracoabdominal motion segment data in the thoracoabdominal motion segment sequence. The motion feature extraction model determines the signal trough position according to the thoracoabdominal motion segment data by a moving average method, calibrates the thoracoabdominal motion segment data by a qualitative diagnosis calibration method to obtain the tidal volume, and performs parameter quantitative analysis according to the trough position and the tidal volume to obtain the thoracoabdominal motion feature data. The expression of the qualitative diagnosis calibration method is: W T =M*(K*V RC +V AB ), Among them, W T is the calibrated tidal volume, K is the relative proportionality factor between the uncalibrated thoracic and abdominal signals, M is the proportionality factor for scaling the relative tidal volume in arbitrary units to the tidal volume in absolute volume units, and V RC is the chest tidal volume, V AB is the abdominal tidal volume.
8. A system for sleep breathing diagnosis based on simple EEG electrodes using the sleep breathing diagnosis method based on simple EEG electrodes according to claim 1, characterized in that: include: Data acquisition module, preprocessing module, segmentation module, feature extraction module, data analysis module; The data acquisition module is used to acquire EEG data through simple brain electrodes, acquire sleep breathing data through PSG equipment, add time stamps to the sleep breathing data and the EEG data to obtain time series sleep breathing data and time series EEG data, wherein the time series sleep breathing data includes: time series ECG data, time series blood oxygen saturation data, and time series chest and abdominal movement data; The preprocessing module is used to calculate the ECG preprocessing data through the ECG preprocessing model according to the time-series ECG data, calculate the EEG preprocessing data through the EEG preprocessing model according to the time-series EEG data, preset a blood oxygen saturation threshold, remove the time-series blood oxygen saturation data according to the blood oxygen saturation threshold to obtain the blood oxygen saturation preprocessing data, and calculate the chest and abdominal motion preprocessing data through the chest and abdominal motion preprocessing model according to the time-series chest and abdominal motion data; The segmentation module is used to preset a unit time length, and segment the ECG preprocessing data, the EEG preprocessing data, the blood oxygen saturation preprocessing data, and the chest and abdominal movement preprocessing data according to the unit time length to obtain an ECG segmentation sequence, an EEG segmentation sequence, a blood oxygen saturation segmentation sequence, and a chest and abdominal movement segmentation sequence; The feature extraction module is used to calculate the ECG feature data through the ECG feature extraction model according to the ECG segmented sequence, calculate the EEG feature data through the EEG feature extraction model according to the EEG segmented sequence, calculate the blood oxygen feature data through the blood oxygen feature extraction model according to the blood oxygen saturation segmented sequence, and calculate the chest and abdominal motion feature data through the motion feature extraction model according to the chest and abdominal motion segmented sequence; The data analysis module is used to normalize the ECG characteristic data, the EEG characteristic data, the blood oxygen characteristic data, and the chest and abdominal movement characteristic data to obtain ECG normalized data, EEG normalized data, blood oxygen normalized data, and chest and abdominal movement normalized data, and to perform parallel fusion of the ECG normalized data, the EEG normalized data, the blood oxygen normalized data, and the chest and abdominal movement normalized data to obtain a splicing vector, and to obtain sleep breathing diagnostic data through classifier calculation based on the splicing vector.
9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the sleep breathing diagnosis method based on simple brain electrodes as described in any one of claims 1-7 is implemented.
10. A storage medium containing computer executable instructions, characterized in that: The computer executable instructions are used to execute the sleep breathing diagnosis method based on simple brain electrodes as described in any one of claims 1-7 when executed by a computer processor.