Sleep breathing disorder monitoring method, system and electronic equipment

Abstract

The present invention discloses a sleep apnea monitoring method, comprising: acquiring EEG signals, EOG signals, and photoplethysmography signals; automatically staging sleep and calculating sleep duration based on the EEG and EOG signals; counting apnea and hypopnea events based on the photoplethysmography signals; and calculating the AHI index based on the calculated sleep duration (ST) and the number of apnea and hypopnea events (BSN). Furthermore, a corresponding head-mounted sleep apnea monitoring system and electronic device are also disclosed. The present invention enables real-time and accurate sleep staging and apnea diagnosis.

Description

Technical Field

[0001] The present invention belongs to the technical field of sleep breathing disorder monitoring, and in particular relates to a head-mounted sleep breathing disorder monitoring technology based on multimodal signals. Background Art

[0002] Sleep stages are a classification of sleep phases based on physiological signals during the night. These include rapid eye movement (REM) sleep and non-rapid eye movement (NREM) sleep. During REM sleep, muscle activity reaches its lowest level of the night, rapid eye movements can be observed, and this is when dreams are most likely to occur. NREM sleep can be further divided into three stages: N1, a light sleep stage from which the body is easily awakened; N2, a moderate sleep stage that accounts for the largest portion of total sleep time and marks the transition to a deeper sleep phase; and N3, a deep sleep stage, also known as slow-wave sleep, which is primarily responsible for restorative and repair activities. A typical sleep cycle for a healthy adult begins with wakefulness and enters stage N1 of NREM sleep, characterized by a gradual loss of attention and slow eye movements. A few minutes later, stage N2 enters, where muscle and eye movements decrease. After entering the deep sleep of stage N3, muscle tone further decreases, making awakening difficult. After stage N3, stages N2 and N1 pass through, transitioning to REM sleep. After REM sleep ends, you will enter the N1 stage again, and the cycle will repeat.

[0003] Obstructive sleep apnea (OSA) disorder is one of the common sleep disorders, characterized by repeated upper airway obstruction during sleep, which usually leads to decreased blood oxygen saturation. Under normal circumstances, it is difficult for patients to realize that they have sleep apnea, and they need someone beside them to discover and remind them, which poses a great hidden danger. At present, in clinical practice, the simplest and most direct way to improve sleep breathing is to wake the patient up and let him fall asleep again. The apnea will end when the patient is awakened, and the blood oxygen saturation will gradually return to the baseline level after normal breathing is restored. An apnea event refers to a complete cessation of oral and nasal airflow for more than 10 seconds during sleep; a hypopnea event refers to a decrease in the intensity (amplitude) of respiratory airflow during sleep by more than 50% compared to the baseline level, accompanied by a decrease in blood oxygen saturation by ≥4% compared to the baseline level. The apnea-hypopnea index (AHI) refers to the number of apneas and hypopneas per hour during sleep.

[0004] A polysomnogram (PSG) is a multi-channel signal graph collected clinically using a polysomnographic device (also known as a PSG device) to collect bioelectrical signals from different parts of the human body (such as the brain, eyes, and jaw), as well as physiological signals such as nasal breathing and snoring. After filtering and amplification, these signals are presented to the physician. Polysomnography is performed and recorded by professional physicians or technicians in specialized institutions such as hospitals or sleep centers. It is used to diagnose various sleep disorders and is the gold standard for determining sleep apnea-hypopnea syndrome (AHI). Therefore, current clinical diagnosis and calculation of the AHI typically requires a full night of sleep apnea monitoring using a PSG device. However, the use of PSG equipment has obvious disadvantages and shortcomings in terms of portability and intelligence, including: (1) PSG equipment is expensive and large in size, and needs to be used in hospitals, sleep research centers, and other places; (2) The use of PSG requires the assistance of professional technicians and cannot be completed independently; (3) It is usually necessary to collect data from at least 12 channels, and at least 22 data cables are connected between the subject and the collector, which will make the subject feel uncomfortable and affect the accuracy of the test; (4) After collecting the sleep data for the whole night, professional personnel are required to process and analyze it, which consumes a lot of energy and is not real-time. In addition, PSG is the gold standard for sleep breathing disorder monitoring and is used to diagnose a variety of sleep diseases, but for sleep breathing disorders, the types of signals required are also redundant.

[0005] OSA can occur at any stage of sleep, with a higher incidence in N1, N2, and REM stages compared to N3. When OSA occurs during REM sleep, apnea lasts longer and blood oxygen saturation decreases more rapidly. Therefore, diagnosing sleep-disordered breathing requires not only identifying apneas and hypopneas, but also accurately assessing sleep status. According to the Manual for the Interpretation of Sleep and Related Events, sleep staging relies primarily on electroencephalogram (EEG) and electrooculogram (EOG). Rapid eye movement (REM) sleep causes rapid eye movements, making REM sleep more accurately distinguishable through EOG. However, due to the complex placement and positioning of EOG electrodes, which are located on both sides of the face and easily compressible when sleeping on one's side, few head-mounted devices currently detect EOG. EEG signals are weak voltage signals at the μV level, and their amplitude and waveform are easily affected by the position of the acquisition and reference electrodes. However, existing monitoring equipment is non-standardized due to the different positions of the reference and acquisition electrodes. Most of the acquired signals do not meet the medical sleep staging standards in terms of amplitude and waveform, and cannot be interpreted by professional doctors. Therefore, the accuracy of sleep staging algorithms developed based on unknown signal quality is also uncertain. In addition, current sleep staging algorithms mostly analyze whole-night data, making it difficult to achieve real-time sleep staging.

[0006] For example, the invention patent application with publication number CN116035557A and patent name “A flexible wearable system for monitoring sleep apnea” discloses collecting information such as electrocardiogram, blood oxygen saturation, and body position changes on the chest and calculating AHI based on this information, but it only relies on body position changes to calculate sleep duration. Since calculating AHI requires sleep staging to obtain accurate sleep duration, and according to the “Handbook of Interpretation of Sleep and Related Events”, sleep staging mainly relies on EEG and EOG signals, relying solely on body movement to calculate sleep duration is prone to errors in AHI calculation. At the same time, the patent uses a machine learning method to calculate AHI, which is highly dependent on manually selected features, and the algorithm has poor stability and versatility. For example, the invention patent application with publication number CN119344671A and patent name "A portable sleep breathing disorder monitoring device based on EEG-blood oxygen saturation-ECG" discloses the use of surface-mount electrodes to collect frontal EEG and blood oxygen signals, and patch electrodes to collect chest ECG signals. However, the EEG electrodes are all attached to the forehead, there are no EOG electrodes, and the reference electrodes are not set according to international standards (10-20 international lead standard). The interference of eye movement cannot be eliminated, and it is difficult to collect accurate EEG signals with obvious amplitude.

[0007] Dreem2 is a wireless sleep apnea monitoring headband developed by Dreem. It monitors EEG signals, heart rate, and body movement data to help users understand and improve sleep quality. It is one of the most accurate consumer sleep recorders on the market and has been approved by the US FDA as a Class II medical device. Dreem2 features six EEG electrodes: four located on the forehead (in the frontal lobe) and two located at the back of the head (in the occipital lobe). When the device is operating, one of the electrodes serves as a reference electrode and does not participate in potential measurement. The forehead portion of the headband also features a pulse oximeter for measuring heart rate. However, the Dreem2's EEG electrodes are positioned relatively close together. The four forehead electrodes are all located in the frontal lobe, resulting in minimal potential differences in the collected signals. The signals collected by multiple EEG electrodes are similar, resulting in a high level of redundant information. Eye movement generates variations in the electrical signal, which spreads across the entire head, with amplitudes reaching up to 100mV. EEG signals, on the other hand, are weak, typically with amplitudes around 50μV. Therefore, EEG signals are affected by eye movements, requiring the use of electrooculogram (EOG) signals to eliminate interference. Furthermore, the Dreem2 headband's reference electrodes are located on the forehead or occipital region. However, the forehead reference electrode is located closer to the EEG electrode, resulting in higher signal similarity and a smaller potential difference, making it difficult to calculate a valid EEG signal. Furthermore, during sleep, the occipital electrode is prone to movement with head movement, leading to inaccurate and unstable reference potentials.

[0008] In summary, PSG devices have problems such as complex installation, non-wearability, and poor comfort; existing commercial wearable sleep apnea monitoring devices have not yet achieved simultaneous sleep state judgment and sleep apnea monitoring and accurate calculation of AHI; and the existing head sleep apnea monitoring device Dreem2 also has problems such as only using EEG signals and being easily interfered by eye movements, and unreasonable reference electrode settings. Summary of the Invention

[0009] To address the aforementioned technical issues, the present invention provides a sleep apnea monitoring method, applicable to a head-mounted sleep apnea monitoring system. While meeting sleep diagnosis requirements, this method optimizes electrode selection and placement, addressing issues such as EEG signal interference from eye movements and excessive common-mode interference caused by close electrode spacing. Furthermore, the method simultaneously performs multi-channel data analysis on EEG, EOG, and blood oxygenation signals, enabling real-time and accurate sleep staging and apnea diagnosis. Furthermore, the present invention also discloses a corresponding head-mounted sleep apnea monitoring system and electronic device.

[0010] The technical solutions of the present invention are as follows:

[0011] A first aspect of the present invention discloses a sleep apnea monitoring method, which can be applied to a head-mounted sleep apnea monitoring system, comprising:

[0012] Acquire EEG signals, EOG signals and photoplethysmography signals;

[0013] Automatically classify sleep stages and calculate sleep duration based on EEG and EOG signals;

[0014] Apnea and hypopnea event statistics are calculated based on the blood oxygen saturation calculated from the photoplethysmography signal;

[0015] The AHI index was calculated based on the statistical sleep duration (ST) and the number of apnea and hypopnea events (BSN).

[0016] As an optional solution, the sleep breathing disorder monitoring method further includes: preprocessing the photoplethysmography signal; and calculating the blood oxygen saturation based on the preprocessed photoplethysmography signal.

[0017] As an optional solution, the sleep breathing disorder monitoring method further includes: outputting the severity level of obstructive sleep apnea (OSA) according to preset rules based on the AHI index and blood oxygen saturation.

[0018] As an optional solution, the sleep breathing disorder monitoring method further includes: when the AHI index is greater than a preset value, issuing a wake-up signal or a wake-up instruction.

[0019] As an optional solution, automatic sleep staging and sleep duration statistics are available, including:

[0020] Pre-processing the EEG and EOG signals to form EEG and EOG signals of preset sizes;

[0021] Based on a preset sleep cycle, the EEG and EOG signals of preset sizes are input into the sleep staging network to obtain sleep staging prediction results. The sleep staging network can extract signal features and their time series relationships. The sleep staging prediction results include W stage, N1 stage, N2 stage, N3 stage and REM stage.

[0022] The sleep duration ST is calculated based on the sleep stage prediction results. The sleep duration ST = preset sleep cycle * the total number of N1, N2, N3 and REM stages in the sleep stage prediction results;

[0023] Apnea and hypopnea event statistics, including:

[0024] The difference between two adjacent blood oxygen saturations is calculated, and the number of times the difference value exceeds the preset threshold is counted as the number of apnea and hypopnea events (BSN).

[0025] As an optional solution, the sleep staging network includes: a convolutional function layer, which is configured to: input the EEG and EOG signals into one branch respectively, and after three convolution operations, randomly set a part of the features to 0 through the Dropout layer to obtain the output features of the two branches; transform the output features of the two branches to the same size through the Reshape layer, and splice them through the Concat layer to obtain the convolution features; sort the convolution features of different time periods in chronological order to form a convolution feature sequence; the convolution operation includes convolution calculation, batch normalization, activation function and maximum pooling operation; the time series function layer, which is configured to: select the convolution features of different times in the convolution feature sequence, learn the temporal relationship of the convolution features, and obtain the corrected mixed features after transformation and dimensionality reduction; the sleep classification function layer, which is configured to: output the corrected mixed features through the fully connected layer and the Softmax layer to form a probability array with a value range of 0-1 for representing the prediction probability of different sleep stages, where the sleep stage corresponding to the maximum probability is the final sleep staging prediction result.

[0026] The second aspect of the present invention discloses a head-mounted sleep breathing disorder monitoring system, comprising a head-mounted sleep breathing disorder monitoring device and a computer program; the head-mounted sleep breathing disorder monitoring device comprises a head-mounted wearable mechanism, and electrooculogram electrodes, electroencephalogram electrodes, reference electrodes, a PPG sensor and a hardware circuit arranged on the head-mounted wearable mechanism; the electrooculogram, electroencephalogram and reference electrodes each have a pair, and when the head-mounted wearable mechanism is worn, the pair of electrooculogram electrodes are respectively located below the outer canthus of the left eye and above the outer canthus of the right eye of the human body for collecting electrooculogram signals; the pair of electroencephalogram electrodes are respectively located on the left and right sides of the human frontal lobe On the right side, a pair of reference electrodes are respectively in contact with the mastoid processes at the junction of the left and right external auricles and cheeks of the human body. The EEG electrodes are used in conjunction with the reference electrodes to collect EEG signals; the PPG sensor is used to obtain the photoelectric volumetric pulse wave signal of the human body; the hardware circuit is used to pre-process the electrooculogram, EEG signals and photoelectric volumetric pulse wave signals; the computer program has one or more, which are stored in the hardware circuit and / or an electronic device that is communicatively connected to the head-mounted sleep breathing disorder monitoring device; when the computer program is run, it is used to execute the sleep breathing disorder monitoring method described in the first aspect of the present invention and any optional scheme.

[0027] As an optional solution, the EOG and EEG electrodes are detachably connected to the wearable head-mounted device through a flexible Velcro; the Velcro includes a velvet surface and a hook surface, the velvet surface is fixed to the inner side of the wearable head-mounted device and has an area larger than the hook surface, and the EOG and EEG electrodes are bonded to the velvet area through the hook surface.

[0028] As an optional solution, when the head-mounted wearable mechanism is worn, the EEG electrodes are arranged 4-6 cm to the left and right of the inner midline of the eye mask and 0.5-2.5 cm from the upper edge; the EOG electrodes are arranged 6-8 cm to the left and right and 0.5-2.5 cm from the lower edge and 4-6 cm from the upper edge.

[0029] As an optional solution, the hardware circuit includes a microcontroller, a signal acquisition module, a first communication module and a power supply module electrically connected to the microcontroller; the signal acquisition module is electrically connected to the electrooculogram electrodes, the electroencephalogram electrodes and the PPG sensor, and is used to pre-process the electrooculogram, electroencephalogram and human photoelectric volume pulse wave signals; the first communication module is used to realize the communication connection between the microcontroller and the electronic device.

[0030] As an optional solution, the wearable mechanism is a flexible structure; the hardware circuit is a flexible circuit board; the electrooculogram, electroencephalogram and reference electrodes are flexible electrodes, among which the electrooculogram and electroencephalogram electrodes are also constructed as a serpentine pattern with stretchable lines; the wires used to connect various electrodes, PPG sensors, intervention components and internal circuits are all constructed as serpentine traces with stretchable lines.

[0031] As an optional solution, the wearable mechanism is an eye mask; the eye mask has an eye mask body consisting of a flexible cover and an elastic strap; the flexible cover includes an inner cover and an outer cover that are connected to each other; the electrooculogram electrodes, the electroencephalogram electrodes, and the PPG sensor are arranged on the inner cover; the reference electrode is arranged on the elastic strap; and the hardware circuit is arranged between the inner cover and the outer cover.

[0032] The third aspect of the present invention discloses an electronic device, including a processor, a memory and a second communication module; the electronic device is communicatively connected to a head-mounted sleep breathing disorder monitoring device through the second communication module; the processor is used to call a computer program stored in the memory to implement the sleep breathing disorder monitoring method of the first aspect of the present invention and any optional scheme.

[0033] The present invention has the following beneficial effects:

[0034] (1) Compared with the traditional method of calculating the AHI index by taking the collected sleep duration of the whole night as the sleep duration, the error caused by waking up during the night is ignored, which easily leads to an underestimation of the AHI index. The present invention uses a head-mounted integrated device to detect and count apnea and hypopnea events while monitoring the sleep state, further improving the accuracy of AHI index calculation.

[0035] (2) The present invention uses the EEG signals and EOG signals collected by the head-mounted sleep breathing disorder monitoring device as the basis for sleep staging. Among them, EEG can provide the characteristic waves of the brain for accurate distinction (N1, N2 and N3), and EOG can reflect the activity of the eyeballs, which helps to identify the rapid eye movement (REM) period. In addition, the combination of the two signals can eliminate the signal interference of EOG artifacts. Compared with using a single EEG or EOG signal, the information contained is more comprehensive, which can achieve more accurate sleep staging.

[0036] (3) The head-mounted sleep breathing disorder monitoring device provided by the present invention is an integrated wearable device for the head. By optimizing the electrode type selection and electrode position design, it solves the problem of EEG signals being interfered with by eye movements and the problem of excessive common-mode interference caused by the electrodes being too close together. It can simultaneously collect EEG, EOG and blood oxygen saturation signals, and simultaneously perform precise sleep staging and sleep breathing disorder monitoring, thereby calculating the AHI index in real time and accurately.

[0037] In particular, the position of the reference electrode has been optimized, placed on the inside of the eye mask strap, at the intersection of the auricle and cheek (near the mastoid process) when worn, to provide a reference potential signal. The potential difference between the potential signals collected by the two EEG electrodes on the left and right forehead and the reference signals of the left and right reference electrodes forms two EEG signals. Furthermore, the reference electrodes are located at a location with thin skin and thick bones, and are far from the cerebral cortex, making them less susceptible to interference from factors such as sweat, body movement, and brain activity, thus providing a stable reference signal.

[0038] Furthermore, the positions of EEG and EOG electrodes on the head and face were optimized, and the electrode layout was optimized according to the forehead length, width, eye distance and eye width of Asian adults, so that the electrode positions can meet the international PSG wearing standards and ensure the high consistency and interpretability of signal characteristics and clinical equipment.

[0039] (4) The present invention uses flexible Velcro to achieve the coordinated use of the velvet surface and the hook surface, allowing for flexible adjustment and accurate positioning of the electrode position, accommodating deviations caused by differences in human facial structure. When first wearing the device, the user can precisely adjust the electrode position based on their facial structure, and no adjustment is required for subsequent wears.

[0040] (5) The present invention improves the convolutional neural network in the sleep staging algorithm. The network combination of CNN and LSTM can effectively extract signal features and time series relationships. In CNN, multiple different convolutional layers are used to extract local features of EEG and EOG signals from different dimensions, which can effectively capture the rhythmic changes of EEG and eye movement signals. In LSTM, the feature sequence is used instead of the original signal, which not only reduces the computational complexity and achieves lightweight model, but also helps LSTM quickly and sensitively perceive the changes in sleep stages over time.

[0041] (6) The present invention can calculate blood oxygen saturation in real time based on the PPG signal collected by the head-mounted sleep apnea monitoring device, and monitor changes in oxygen saturation in real time. By identifying nocturnal hypoxic events and periodic downward trends, it can accurately capture obstructive sleep apnea (OSA) and hypopnea events. Compared with traditional respiratory airflow monitoring, the blood oxygen saturation monitoring sensor is smaller, more comfortable to wear, and has stronger system integration.

[0042] (7) The head-mounted sleep apnea monitoring device provided by the present invention utilizes flexible fabric and wireless data transmission, making it highly comfortable to wear and eliminating the need for data cables. Furthermore, the device utilizes flexible and stretchable electrodes and wires that adapt to the deformation and expansion of the eye mask, establishing a long-term, stable skin-electrode contact interface and enabling long-term, accurate data collection at night. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 Schematic diagram of the structure of the sleep apnea monitoring eye mask Figure 1 ;

[0044] Figure 2 Schematic diagram of the structure of the sleep apnea monitoring eye mask Figure 2 ;

[0045] Figure 3 A front perspective view of the sleep monitoring eye mask in a wearing state;

[0046] Figure 4 A left perspective view of the sleep monitoring eye mask in the wearing state;

[0047] Figure 5 This is a schematic diagram of the electrode structure and connecting wire structure of the sleep monitoring eye mask;

[0048] Figure 6 Schematic diagram of bonding electrodes to Velcro;

[0049] Figure 7 This is a circuit connection block diagram of a sleep apnea monitoring system;

[0050] Figure 8 This is a schematic diagram of the AHI calculation process;

[0051] Figure 9 This is a schematic diagram of the network structure of the sleep staging module;

[0052] Figure 10 The structure diagram of the convolutional neural network for feature extraction;

[0053] Figure 11 This is a flow chart of blood oxygen saturation calculation.

[0054] Figure markings: 1-eye mask body, 2-inner cover, 3-outer cover, 4-ear-hook elastic strap, 5-left EEG electrode, 6-right EEG electrode, 7-left EO electrode, 8-right EO electrode, 9-reference left electrode, 10-reference right electrode, 11-PPG sensor, 12-hardware circuit, 13-power supply battery, 14-switch, 15-LED light, 161-Velcro hook surface, 162-Velcro suede surface. DETAILED DESCRIPTION

[0055] The technical solutions of the present invention will be clearly and completely described below with reference to specific embodiments and accompanying drawings. Obviously, the embodiments described are only some embodiments of the present invention, not all embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0056] In the description of the present invention, if there are terms indicating directions or positional relationships such as "upper", "lower", "front", "back", "inside", "outside", "left", "right", etc., they are based on the directions or positional relationships shown in the drawings and are only for the convenience of describing the present invention and simplifying the description. They do not indicate or imply that the devices or elements referred to must have a specific direction, be constructed and operate in a specific direction. Therefore, they should not be understood as limiting the present invention. In addition, if there are terms such as "first" and "second", they are used to distinguish similar objects and are not necessarily used to describe a specific order or relative importance. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood in combination with specific circumstances. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, but may include other units that are not clearly listed or inherent to these products or devices.

[0057] like Figure 1 As shown, Embodiment 1 of the present invention discloses a sleep apnea monitoring eye mask, which primarily comprises an eye mask body 1, a pair of EEG electrodes (electrodes 5 and 6), a pair of EOG electrodes (electrodes 7 and 8), a pair of reference electrodes (electrodes 9 and 10), a PPG sensor 11, a hardware circuit 12, and a power supply battery 13. The eye mask body 1 comprises a mask surface and an ear-hook elastic strap 4. The mask surface is divided into an inner mask surface 2 and an outer mask surface 3. Both the inner mask surface 2 and the outer mask surface 3 can be made of cotton fabric with good flexibility, comfort, and breathability, or can be made of materials such as silk, leather, and spandex. The inner mask surface 2 and the outer mask surface 3 can be fixedly connected at the edges to form a single unit. The ear-hook elastic strap 4 comprises two pieces, fixedly connected to either side of the eye mask body 1, for securing the eye mask body 1 to the head. Materials such as spandex, polyester, and spandex can be used. In other embodiments, the elastic strap can also be a headband, and this is not a limitation of the present invention. It can be understood that the eye mask body composed of the inner cover 2, the outer cover 3 and the ear-hanging elastic strap 4 is a head-mounted wearable mechanism for carrying the above-mentioned electrodes and devices.

[0058] On the inner side of the eye mask body 1 (i.e. the side that contacts the skin), a pair of EEG electrodes, namely electrodes 5 and 6, are symmetrically fixed on the left and right sides above the inner cover 2, and a pair of EOG electrodes, namely electrodes 7 and 8, are fixed on the lower left and upper right sides of the inner cover 2. Figure 3 As shown, when the eye mask is worn, the EEG electrodes are located exactly on the left and right sides of the human frontal lobe (Fp1 and Fp2 in the standard EEG position) to collect EEG potential signals; the EOG electrodes are located approximately 1 cm below the outer canthus of the left eye and approximately 1 cm above the outer canthus of the right eye. The EOG signal is the potential difference between the two EOG electrodes. A pair of reference electrodes, namely the left reference electrode 9 and the right reference electrode 10, are respectively fixed to the inner sides of the left and right ear-hook elastic straps 4. Figure 4 As shown, when the eye mask is worn, a pair of reference electrodes contact the mastoid processes at the junction of the left and right auricles and cheeks, respectively, providing a stable, low-amplitude reference potential. The potential difference between the left and right EEG electrodes and the left and right reference electrodes generates two EEG signals. These electrodes can be attached to the inner mask 2 or the ear-hook elastic strap 4 by gluing them together using a soft adhesive such as silicone.

[0059] It is worth noting that the potential signal collected by the two EEG electrodes on the left and right sides of the forehead in the present invention forms two EEG signals with the potential difference from the reference signal of the left and right reference electrodes. Compared with the occipital region, the area near the mastoid process is less affected by head movement and is not subject to complex external signal interference (such as electrocardiogram signals). Compared with the forehead and the tip of the nose, the skin near the mastoid process is thinner and sweats less, which ensures the long-term stability of the electrode-skin interface. At the same time, the mastoid process behind the ear is a raised bone, and the thicker bone structure is conducive to isolating the brain's activity signals, and the mastoid process is located in the middle and rear of the entire skull, far away from the main electrical activity area of the cerebral cortex, and can provide a reference potential that is not affected by brain activity as a reference signal.

[0060] Furthermore, the present invention designs specific electrode placements based on Asian anatomical characteristics. Since the forehead length (hairline to upper eye edge) is approximately 6-8 cm, the forehead width (from the hairline to the inner side of the eye) is approximately 10-12 cm, the interocular distance (the horizontal distance between the inner corners of the eyes) is approximately 2.5-3.5 cm, and the eye length (from canthus to canthus) is approximately 2.2-3.0 cm, the EEG electrodes (electrodes 5 and 6) can be placed 4-6 cm to the left and right of the midline and 0.5-2.5 cm from the upper edge, respectively. The LEOG electrode (electrode 7) can be placed 6-8 cm to the left and right and 0.5-2.5 cm from the lower edge. The REOG electrode (electrode 8) can be placed 4-6 cm from the upper edge.

[0061] Furthermore, the present invention can also be used in conjunction with flexible Velcro to ensure positioning. The flexible Velcro can be used to move the monitoring electrodes (i.e., EEG electrodes and EOG electrodes), ensuring that they can be reused after being set once, and that the position is always within the ideal range. Velcro generally includes a velvet surface and a hook surface, and the velvet surface and the hook surface can be used together to achieve the adhesion of the two. Figure 6 As shown, for example, the monitoring electrode size is 1×1 cm 2 The monitoring electrode is attached to the Velcro hook surface 161 of the same shape and size through the Velcro adhesive, and the Velcro velvet surface 162 is 2×2 cm in size. 2The Velcro velvet surface 162 is attached to the inner cover 2 by the Velcro adhesive. In this way, when wearing it for the first time, the user can adjust the specific position of the monitoring electrode on the Velcro velvet surface 162 by himself according to the instructions based on the characteristics of his facial structure, and no adjustment is required for subsequent wearing.

[0062] In order to ensure good contact between the electrodes and the skin, all types of electrodes (EEG electrodes, EOG electrodes, and reference electrodes) need to be flexible and able to deform and adapt to different wearing conditions like the fabric of the eye mask. In addition, the ear-hook elastic straps of the eye mask need to be stretched repeatedly when worn. Therefore, the reference electrode and connecting wires must also have sufficient stretchability to avoid breakage during the stretching process, thereby ensuring stable signal acquisition and transmission. Figure 5 As shown in (a) in the figure, these electrodes can be made of ultra-thin metal foil with a serpentine pattern and cut by high-precision laser cutting. Compared with hard electrodes such as silver / silver chloride and gold cup electrodes, flexible electrodes have higher skin conformality and comfort, which can effectively avoid the problem of poor contact due to uneven skin. Figure 5 As shown in (b), the wires between the electrodes and the hardware circuit can also be routed in a serpentine pattern, effectively preventing circuit breakage caused by eye mask stretching and deformation. Furthermore, ultra-thin, stretchable polyimide film can be used to insulate and encapsulate the serpentine circuits through methods such as heat-seal and hot-pressing, ensuring stable and reliable signal transmission.

[0063] The PPG sensor 11 is mainly composed of a light source (eg, LED) and a photodetector. The PPG sensor 11 is fixed to the upper middle portion of the inner cover 2 by digging a hole in the surface of the inner cover 2. Figure 3 As shown, when the eye mask is worn, the PPG sensor 11 is located at the center of the forehead and monitors the forehead's photoplethysmographic (PPG) signal (hereinafter referred to as the "PPG signal") via reflected PPG (photoplethysmography) signals. PPG is a non-invasive technology that measures changes in blood flow within blood vessels by detecting optical changes in the skin. When light emitted by a light source is absorbed by the skin, blood vessels, and blood, some of the light is reflected and received by the detector. However, as hemoglobin binds to oxygen, it absorbs light of different wavelengths, affecting the light intensity received by the sensor. In this embodiment, the PPG sensor 11 can specifically be a MAX30102 sensor.

[0064] Combine Figure 2 and Figure 3As shown, the hardware circuit 12 and power supply battery 13 can be fixed between the inner cover 2 and the outer cover 3 by adhesive bonding. When the eye mask is worn, it is positioned precisely on the center of the forehead to balance the overall weight. The multiple pin interfaces of the hardware circuit are connected to the EEG electrodes, EOG electrodes, reference electrodes, and PPG sensors via wires, synchronously collecting EEG, EOG, and photoplethysmography signals. A switch 14, electrically connected to the hardware circuit 12, is also located at the top of the outer cover 3. Switch 14 controls the on / off state of the sleep apnea monitoring eye mask.

[0065] like Figure 7 As shown, the hardware circuit 12 integrates a microcontroller, primarily responsible for data operations and module function control. An STM32 controller can be used. The hardware circuit 12 also integrates a signal acquisition module with multiple signal acquisition units. The signal acquisition units integrate analog amplifier circuits and analog-to-digital conversion circuits, respectively used for amplifying and converting EEG, EOG, and photoplethysmography signals. For example, the analog amplifier circuit in the EEG processing unit first amplifies the small-amplitude EEG signal; the analog-to-digital conversion circuit then converts the amplified analog signal into a digital signal. The hardware circuit 12 also integrates a Bluetooth communication module and a power management module. The microcontroller transmits the converted analog-to-digital signal to a mobile terminal via the Bluetooth communication module. The power management module is connected to a power supply battery 13 and primarily powers the microcontroller and various functional modules in the hardware circuit 12. Furthermore, the hardware circuit 12 is equipped with an LED indicator module to indicate the system's on and off status. The hardware circuit 12 can utilize a flexible printed circuit (FPC) made from materials such as polyimide or polyester film, which offers high reliability, light weight, thinness, and good flexibility.

[0066] Embodiment 2 of the present invention provides a mobile terminal that, together with the sleep apnea monitoring eye mask of Embodiment 1, constitutes a sleep apnea monitoring system. The mobile terminal includes a processor, a memory electrically connected to the processor, a communication module, and an alarm warning module. The mobile terminal can be an electronic device such as a smartphone, tablet computer, smartwatch, smart bracelet, or dedicated supporting terminal. The alarm warning module can be implemented using a built-in speaker in the mobile terminal.

[0067] like Figure 8As shown, the functions of the mobile terminal mainly include automatic sleep staging and apnea and hypopnea event statistics, and based on the sleep staging and apnea and hypopnea event statistics results, the AHI value is calculated and updated in real time. Accordingly, the processor of the mobile terminal is equipped with a sleep analysis module, an apnea and hypopnea event statistics module, and an AHI calculation module. Among them, the sleep analysis module mainly includes a sleep preprocessing unit, a sleep staging unit, and a sleep duration statistics unit, which are respectively used to preprocess EEG and EOG signals, obtain sleep stage prediction results, and perform sleep duration statistics. The apnea and hypopnea event statistics module mainly includes a blood oxygen preprocessing unit, a blood oxygen saturation calculation unit, and an event statistics unit, which are respectively used to preprocess PPG signals, calculate blood oxygen saturation, and apnea and hypopnea event statistics. The AHI calculation module calculates the AHI value in real time based on the sleep staging results of the sleep analysis module and the event statistics results of the apnea and hypopnea event statistics module.

[0068] In a specific embodiment, the acquisition frequency of EEG and EOG in the hardware circuit is set to 100 Hz, and 30 seconds is used as the cycle for sleep staging. The EEG and EOG signals are low-pass filtered in the range of 0-30 Hz by the sleep preprocessing unit in the sleep analysis unit. In addition, in order to avoid the marginal effect caused by filtering, the length of the input signal in the filtering needs to be greater than the length of one cycle (30 seconds), and it is re-cut to the length of one cycle after filtering. In this example, a time lag method is adopted, that is, the filtering start time is 10 seconds after the data acquisition is completed, and a signal with a length of 50 seconds is input for filtering. The filtered signals are longitudinally spliced to form an EEG signal (array 1) of size (2, 3000) and an EOG signal (array 2) of size (1, 3000) as the result of sleep preprocessing and input into the sleep staging module.

[0069] like Figure 9 As shown in Figure 2, the sleep staging unit mainly includes three functional layers connected in sequence, namely the convolutional functional layer, the time series functional layer and the sleep classification functional layer. Figure 10As shown, arrays 1 and 2 pass through three consecutive convolutional layers. In each convolutional layer, the arrays undergo a one-dimensional convolution calculation (Conv1D) to obtain high-dimensional features. Batch normalization (BN) is then used to transform the feature values to a standard normal distribution with a mean of 0 and a variance of 1. The standard feature array undergoes nonlinear mapping using the ReLU activation function, and the maximum pooling layer (MaxPool) selects the local maximum of the array to reduce the array size, resulting in a single-layer convolution feature array. The single-layer convolution feature array is then fed into the next convolutional layer, repeating the convolution calculation, batch normalization, activation function, and maximum pooling operations. After three convolutional layers, the resulting three-layer convolution feature array passes through a dropout layer, which randomly sets some features to 0 to reduce the risk of network overfitting, resulting in the output features of the three-layer convolution (i.e., the output features of one branch). Because the convolutional layer parameters of the two branches, array 1 and array 2, differ, and the extracted features are of different sizes, the output features of both branches are transformed to the same size through the Reshape layer and concatenated through the Concat layer to obtain the final convolutional features (CNN features). CNN features from different time periods are sorted chronologically to form a convolutional feature sequence. In the time series function layer (LSTM), CNN features from different time periods in the convolutional feature sequence are selected to learn the temporal relationship between the CNN features. The CNN features are then transformed and reduced in dimension to obtain a modified hybrid feature. Finally, in the sleep classification function layer, the modified hybrid feature is mapped from high-dimensional features to an output array of size (1,5) through a fully connected layer (FC). The Softmax layer converts the output array into a probability array based on the size of the output array. Each element in the probability array has a value between 0 and 1, and the sum of all elements is 1. The values in the probability array represent the predicted probabilities of five different sleep stages: W (wakefulness), N1, N2, N3, and REM. The sleep stage corresponding to the highest probability is the final prediction.

[0070] The sleep duration statistics unit can calculate the wearer's sleep duration (ST) based on the sleep stage prediction results, that is, the cycle length (30s) × the total number of sleep stage prediction results (N1+N2+N3+REM).

[0071] Combine Figure 11As shown, the red light (Red) and infrared light (IR) emitted by the light source in the PPG sensor 11 are reflected by the skin and received by the detector, forming a photoplethysmography signal. The blood oxygen saturation is calculated using the beat-to-beat blood oxygen calculation method as an example. First, the red light and infrared light in the photoplethysmography signal are each low-pass filtered at 5Hz by the blood oxygen preprocessing unit. After filtering, the peaks and troughs of the two signal waveforms are respectively found. The absolute value of the difference between the peak and trough of a particular beat is used as the AC value of the beat, and the average of all low-pass filtered signals between one peak and the next is used as the DC value DC of the beat.

[0072] Then, the blood oxygen saturation calculation unit can obtain the blood oxygen saturation of this beat using the following formula (1).

[0073] (1)

[0074] Where, Indicates blood oxygen saturation, represents the red light AC component, represents the DC component of red light, Indicates the infrared light AC component, Indicates the DC component of infrared light.

[0075] In the above method, the calculation frequency of the blood oxygen saturation value is related to the heart beat frequency (heart rate). In order to stabilize the sampling rate of the blood oxygen saturation signal, the time window method can also be used to calculate the blood oxygen saturation value every second to ensure that the sampling rate of the blood oxygen saturation is 1Hz. First, set a time window of a specific length. (Take 5 seconds as an example). Every 1 second in this time window, the newly collected photoplethysmogram is stored and the photoplethysmogram of the first 1 second is released. Within the data of the time window, the main peak and trough values of the photoplethysmogram are found. The absolute value of the difference between the peak and trough of each beat is averaged as the AC value within this time window. The average of all values is used as the DC value DC of the beat within this time window. Then, the blood oxygen saturation is calculated according to formula (2).

[0076] According to the "Sleep and Related Events Manual" published by the American Academy of Sleep Medicine, when the blood oxygen saturation drops by ≥4% compared to the baseline value, it can be used to judge a hypopnea event. When the wearer experiences apnea, the respiratory airflow almost stops and the blood oxygen saturation drops faster. The difference in blood oxygen saturation can be calculated according to formula (2): It is understandable that for the time window method, The event statistics unit counts the number of times the difference in blood oxygen levels between two consecutive seconds exceeds a specified threshold, which is the number of apnea and hypopnea events (BSN). In this embodiment, the threshold is 4%, but it can be adjusted based on the severity of the sleep apnea and individual differences.

[0077] (2)

[0078] The AHI calculation module calculates the AHI index based on the sleep duration ST, the number of apnea and hypopnea events BSN according to the following formula (3).

[0079] (3)

[0080] Furthermore, the present invention can also set a wake-up threshold. When the AHI exceeds the wake-up threshold, the mobile terminal will wake the wearer via an alarm. In this embodiment, the wake-up threshold is set to 15, which means that the wearer with moderate sleep apnea will be awakened. In other embodiments, bone conduction headphones, ringing modules, etc. can be added to the ear-hook elastic strap 4 of the sleep apnea monitoring eye mask. After receiving relevant instructions from the mobile terminal, the corresponding operation is executed to wake the wearer.

[0081] It's worth noting that, because the present invention allows real-time analysis of ST and BSN from every 30-second sleep cycle signal, AHI calculation can be performed in real time, eliminating the need for a full night of monitoring. In fact, AHI changes in real time with sleep onset, allowing real-time AHI calculation to reveal the wearer's AHI changes. Based on the dynamic, real-time AHI, the severity of a wearer's apnea throughout the night can be accurately analyzed, and a wake-up threshold can be set as needed, enabling timely sleep intervention in severe apnea cases to prevent unforeseen events.

[0082] Furthermore, the present invention can also determine the severity of obstructive sleep apnea (OSA) based on the AHI index and blood oxygen saturation. As shown in Table 1, according to the "Guidelines for the Diagnosis and Treatment of Obstructive Sleep Apnea in Adults," the severity of obstructive sleep apnea (OSA) in adults can be categorized as mild, moderate, and severe, based on the AHI and blood oxygen saturation. Therefore, the wearer's OSA severity level can be output based on this standard.

[0083] Table 1 Obstructive sleep apnea (OSA) severity in adults

[0084]

[0085] It is understandable that in other embodiments, the functions implemented by the mobile terminal can also be integrated into the hardware circuit of the sleep breathing disorder monitoring eye mask, and the relevant data calculation and analysis can be performed by the microcontroller.

[0086] In summary, the present invention designs an integrated wearable head device based on EEG, EOG and blood oxygen multimodal signals, and adopts flexible electronic fabrics to ensure high wearing comfort; under the premise of meeting the needs of sleep diagnosis, the selection of electrodes is optimized to solve the problems of EEG signals being interfered with by eye movements and excessive common-mode interference caused by the close distance between electrodes; the reference electrode is set near the mastoid of the external auricle to ensure the stability of the reference potential; it can simultaneously perform wireless transmission and multi-channel data analysis of EEG, EOG and blood oxygen signals to perform sleep staging and sleep apnea diagnosis.

[0087] When the sleep breathing disorder monitoring system disclosed in the embodiment of the present invention is used, the power switch 13 is first turned on to start the sleep breathing disorder monitoring eye mask. The LED light 14 flashes to indicate that the system is turned on, and wireless communication between devices is established by the mobile terminal. When the user wears the smart eye mask, the user can slightly adjust the position of the eye mask by hand to ensure that the EEG electrodes (5, 6), the EOG electrodes (7, 8), the reference electrodes (9, 10) and the PPG sensor 11 are in the correct position and in good contact. After the mobile terminal establishes a connection with the sleep breathing disorder monitoring eye mask, the LED light remains on and sleep breathing disorder monitoring begins. The hardware circuit in the sleep breathing disorder monitoring eye mask collects EEG, EOG and blood oxygen saturation signals in real time, amplifies and converts them into analog-to-digital signals respectively, and then transmits them wirelessly to the mobile terminal. The mobile terminal will analyze the EEG and EOG data in real time based on the various signals received (EEG signals, EOG signals, etc.) to automatically stage sleep and calculate the blood oxygen saturation signal to perform statistics on sleep apnea and hypopnea events, and calculate the AHI in real time. If apnea triggers an alarm, the wearer will be awakened by the alarm of the sleep breathing disorder monitoring eye mask. After the sleep is over, the power switch 13 is turned off, and the mobile terminal stops receiving data and no longer displays the analysis results of sleep stages and apnea.

[0088] Finally, it should be noted that although the embodiments of the present invention have been described above in conjunction with the accompanying drawings, the present invention is not limited to the above-mentioned specific embodiments and application fields. The above-mentioned specific embodiments are merely illustrative and instructive, and not restrictive. Under the guidance of this specification, those skilled in the art can also make many forms without departing from the scope of protection of the claims of the present invention, and all of these forms are protected by the present invention.

Claims

1. A sleep apnea monitoring method, characterized in that: Applied to head-mounted sleep apnea monitoring systems, including: Acquire EEG signals, EOG signals and photoplethysmography signals; Automatically classify sleep and calculate sleep duration based on the EEG and EOG signals; Apnea and hypopnea event statistics are calculated based on the blood oxygen saturation calculated from the photoplethysmography signal; The AHI index was calculated based on the statistical sleep duration (ST) and the number of apnea and hypopnea events (BSN); The head-mounted sleep breathing disorder monitoring system includes a head-mounted sleep breathing disorder monitoring device and a computer program. The head-mounted sleep breathing disorder monitoring device includes a head-mounted wearable mechanism, and electrooculogram (EOG) electrodes, electroencephalogram (EEG) electrodes, a reference electrode, a PPG sensor, and a hardware circuit arranged on the head-mounted wearable mechanism. The EOG, EEG, and reference electrodes each have a pair. When the head-mounted wearable mechanism is worn, the pair of EOG electrodes are respectively located below the outer canthus of the left eye and above the outer canthus of the right eye. The pair of EEG electrodes are respectively located on the left and right sides of the frontal lobe of the human body. The pair of reference electrodes are respectively in contact with the mastoid processes at the junction of the left and right external auricles and cheeks of the human body. The EEG signal is obtained through an EEG electrode and a reference electrode; the EOG signal is obtained through an EOG electrode; and the photoplethysmography signal is obtained through a PPG sensor. The automatic sleep staging is implemented through a sleep staging network, which includes: The convolution function layer is configured to: input the EEG and EOG signals into one branch respectively, and after three convolution operations, randomly set some features to 0 through the Dropout layer to obtain the output features of the two branches; transform the output features of the two branches to the same size through the Reshape layer, and splice them through the Concat layer to obtain the convolution features; sort the convolution features of different time periods in chronological order to form a convolution feature sequence; the convolution operation includes convolution calculation, batch normalization, activation function and maximum pooling operation; The time series function layer is configured to select convolution features at different times in the convolution feature sequence, learn the temporal relationship between the convolution features, and obtain the corrected mixed features after transformation and dimensionality reduction. The sleep classification function layer is configured to pass the modified mixed features through the fully connected layer and the softmax layer to output a probability array with a value range of 0-1, which is used to represent the predicted probability of different sleep stages. The sleep stage corresponding to the maximum probability is the final sleep stage prediction result.

2. The sleep apnea monitoring method according to claim 1, wherein: Also includes: Preprocessing the photoplethysmography signal; The blood oxygen saturation is calculated based on the preprocessed photoplethysmography signal.

3. The sleep apnea monitoring method according to claim 1, wherein: Also includes: Based on the AHI index and the blood oxygen saturation, the severity of the obstructive sleep apnea (OSA) is output according to a preset rule.

4. The sleep apnea monitoring method according to any one of claims 1 to 3, wherein: The automatic sleep staging and sleep duration statistics include: Pre-processing the EEG and EOG signals to form EEG and EOG signals of preset sizes; Based on a preset sleep cycle, the EEG and EOG signals of preset sizes are input into a sleep staging network to obtain a sleep staging prediction result; the sleep staging network can extract signal features and their time series relationship; the sleep staging prediction results include stage W, stage N1, stage N2, stage N3 and REM stage; Count the sleep duration ST according to the sleep stage prediction result, wherein the sleep duration ST = the preset sleep cycle * the total number of N1 stage, N2 stage, N3 stage and REM stage in the sleep stage prediction result; The apnea and hypopnea event statistics include: The difference between two adjacent blood oxygen saturations is calculated, and the number of times the difference value exceeds the preset threshold is counted as the number of apnea and hypopnea events (BSN).

5. A head-mounted sleep apnea monitoring system, characterized in that: Includes a head-mounted sleep apnea monitoring device and a computer program; The head-mounted sleep breathing disorder monitoring device includes a head-mounted wearable mechanism, and electrooculogram electrodes, electroencephalogram electrodes, reference electrodes, a PPG sensor, and a hardware circuit arranged on the head-mounted wearable mechanism; The electrooculogram (EOG), electroencephalogram (EEG) and reference electrodes each have a pair. When the head-mounted wearable device is worn, the pair of EOG electrodes are respectively located below the outer canthus of the left eye and above the outer canthus of the right eye of the human body to collect EOG signals. A pair of EEG electrodes are respectively located on the left and right sides of the human frontal lobe, and a pair of reference electrodes are respectively in contact with the mastoid processes at the junction of the left and right external auricles and cheeks of the human body. The EEG electrodes and the reference electrodes are used in conjunction with each other to collect EEG signals. The PPG sensor is used to obtain a human body photoplethysmogram signal; The hardware circuit is used for preprocessing electrooculogram (EOG), electroencephalogram (EEG) and photoplethysmography (PEP) signals; The computer program has one or more components stored in a hardware circuit and / or an electronic device in communication with a head-mounted sleep apnea monitoring device; When the computer program is executed, it is used to perform the sleep breathing disorder monitoring method according to any one of claims 1 to 4.

6. The head-mounted sleep apnea monitoring system according to claim 5, wherein: The EOG and EEG electrodes are detachably connected to the wearable head-mounted device via a flexible Velcro strip; the Velcro strip includes a velvet surface and a hook surface, the velvet surface is fixed to the inner side of the wearable head-mounted device and has an area larger than the hook surface, and the EOG and EEG electrodes are adhered to the velvet surface area via the hook surface.

7. The head-mounted sleep apnea monitoring system according to claim 5, wherein: When the head-mounted wearable mechanism is worn, the EEG electrodes are respectively arranged 4-6 cm to the left and right of the inner midline of the eye mask and 0.5-2.5 cm from the upper edge; the electrooculogram electrodes are respectively arranged 6-8 cm to the left and right and 0.5-2.5 cm from the lower edge and 4-6 cm from the upper edge.

8. The head-mounted sleep apnea monitoring system according to any one of claims 5 to 7, wherein: The hardware circuit includes a microcontroller, and a signal acquisition module, a first communication module and a power supply module electrically connected to the microcontroller; The signal acquisition module is electrically connected to the electrooculogram electrodes, the electroencephalogram electrodes and the PPG sensor, and is used to pre-process the electrooculogram, electroencephalogram and human photoelectric pulse wave signals; The first communication module is used to realize communication connection between the microcontroller and the electronic device.

9. The head-mounted sleep apnea monitoring system according to any one of claims 5 to 7, wherein: The wearable mechanism is a flexible structure; the hardware circuit is a flexible circuit board; the electrooculogram, electroencephalogram and reference electrodes are flexible electrodes, wherein the electrooculogram and electroencephalogram electrodes are also constructed as a serpentine pattern with stretchable lines; the wires used to connect various electrodes, PPG sensors, intervention components and internal circuits are all constructed as serpentine traces with stretchable lines.

10. The head-mounted sleep apnea monitoring system according to claim 9, wherein: The wearable mechanism is an eye mask; the eye mask has an eye mask body consisting of a flexible cover and an elastic strap; the flexible cover includes an inner cover and an outer cover that are connected to each other; the electrooculogram electrodes, electroencephalogram electrodes, and PPG sensors are arranged on the inner cover; the reference electrode is arranged on the elastic strap; and the hardware circuit is arranged between the inner cover and the outer cover.

11. An electronic device, characterized in that: It comprises a processor, a memory and a second communication module; the electronic device is communicatively connected to a head-mounted sleep breathing disorder monitoring device via the second communication module; the processor is used to call a computer program stored in the memory to implement the sleep breathing disorder monitoring method according to any one of claims 1 to 4.

Images