A sleep monitoring system based on brain-computer interface and PPG module

Abstract

This invention relates to a sleep monitoring system based on a brain-computer interface and a PPG module, belonging to the field of sleep monitoring technology. The system includes: a brain-computer interface module for acquiring EEG signals and extracting EEG features; a PPG module for acquiring pulse wave signals and extracting heart rate and heart rate variability features; an EEG triggering module for determining trigger anchor points; a hysteresis window module for generating a hysteresis tracking window based on the trigger anchor points and the individual hysteresis baseline; a pulse wave delay matching module for obtaining pulse wave delay response features within the window; a feedforward acquisition control module for enhancing PPG acquisition; and a fusion judgment module for determining the sleep monitoring results. By generating a hysteresis tracking window located after the trigger anchor point through the EEG trigger point and matching the PPG delay response features within this window, the system avoids missed or false positives caused by hard alignment of EEG and PPG signals at the same time, thus improving the accuracy of micro-awakening recognition.

Description

Technical Field

[0001] This invention belongs to the field of sleep monitoring technology and relates to a sleep monitoring system based on brain-computer interface and PPG module. Background Technology

[0002] A sleep monitoring system is a device or system used to collect physiological signals during a user's sleep and identify sleep state, sleep events, and sleep continuity based on changes in these signals. A sleep monitoring system based on a brain-computer interface and a PPG module refers to a system that simultaneously collects electroencephalogram (EEG) signals using the BCG module and photoplethysmography (PPG) signals using the PPG module during sleep monitoring. This allows for the acquisition of information such as changes in brain electrical activity, heart rate, and heart rate variability, in order to analyze micro-arousals, sleep fragmentation, and autonomic nervous system response states. Currently, EEG and PPG characteristics at the same time point or within the same time window are typically fused synchronously to determine whether a micro-arousal event has occurred in the user.

[0003] However, in current technology, EEG signals reflect the instantaneous activity of the central nervous system, while PPG signals reflect the results of peripheral blood flow and autonomic nervous system regulation. There is a time lag in the physiological response between the two. If EEG characteristics and pulse wave characteristics are directly aligned at the same time or within the same time window, it is easy to find situations where micro-awakening-related mutations have occurred in the EEG but the delayed response of the pulse wave has not yet arrived. This makes it difficult for the system to accurately associate EEG mutations with PPG response changes, thereby reducing the accuracy of micro-awakening event recognition. Summary of the Invention

[0004] In view of this, the purpose of the present invention is to provide a sleep monitoring system based on a brain-computer interface and a PPG module, which solves the problem that the electroencephalogram (EEG) signal and the PPG signal are difficult to align accurately in existing sleep monitoring due to the sluggish physiological response, resulting in a decrease in the accuracy of micro-awakening event recognition.

[0005] To achieve the above objectives, the present invention provides the following technical solution: A sleep monitoring system based on a brain-computer interface and a PPG module includes: The brain-computer interface module collects brain signals during the user's sleep process and extracts the characteristics of brain frequency band energy changes and brain signal abrupt change slope characteristics; The PPG module collects pulse wave signals during the user's sleep process and extracts heart rate change characteristics and heart rate variability change characteristics. The dual-track caching module caches the characteristics of brainwave frequency band energy changes, brainwave mutation slopes, heart rate changes, and heart rate variability changes, respectively. The EEG triggering module determines the trigger anchor point based on the characteristics of EEG frequency band energy changes and the characteristics of EEG mutation slope. Hysteresis window module, which generates a hysteresis tracking window located after the trigger anchor point based on the trigger anchor point and the individual hysteresis baseline; The pulse wave delay matching module performs coherent matching on heart rate change features and heart rate variability change features within the hysteresis tracking window to obtain pulse wave delay response features. The feedforward acquisition control module controls the PPG module to improve the sampling rate and LED drive current within the hysteresis tracking window; The fusion judgment module determines the sleep monitoring results based on the trigger anchor point, pulse wave delay response characteristics, and individual hysteresis baseline.

[0006] Furthermore, the dual-track caching module caches the characteristics of brainwave frequency band energy changes, brainwave mutation slopes, heart rate changes, and heart rate variability changes, including: Based on the sampling time sequence of EEG signals, the characteristics of EEG frequency band energy change and the characteristics of EEG mutation slope are written into the short-term buffer window of EEG. Based on the sampling time sequence of the pulse wave signal, the heart rate change characteristics and heart rate variability change characteristics are written into the pulse wavelength buffer window; Based on the start and end times of the short-term buffer window of EEG and the start and end times of the buffer window of pulse wavelength, a time index relationship between EEG features and pulse wave features is established. Among them, the buffer duration of the pulse wavelength time buffer window is greater than that of the EEG short-time buffer window, so that the pulse wavelength time buffer window retains the pulse wave characteristics after the trigger anchor point.

[0007] Furthermore, the EEG triggering module obtains trigger anchor points based on the characteristics of EEG frequency band energy changes and the characteristics of EEG mutation slopes, including: Read the brainwave frequency band energy change characteristics and brainwave mutation slope characteristics corresponding to the current brainwave analysis window from the short-term buffer window; The difference between the characteristics of brainwave frequency band energy change and the sleep reference frequency band energy is calculated to obtain the brainwave frequency band mutation amount; Based on the variation amplitude of the EEG slope characteristics between adjacent EEG analysis windows, the EEG slope mutation amount is obtained. When the frequency band mutation amount of EEG reaches the frequency band mutation threshold, the slope mutation amount of EEG reaches the slope mutation threshold, and the direction of frequency band energy enhancement corresponding to the frequency band mutation amount is consistent with the direction of micro-awakening EEG response, the time point corresponding to the current EEG analysis window is marked as the trigger anchor point.

[0008] Furthermore, the individual hysteresis baseline in the hysteresis window module is obtained in the following way: Baseline EEG and baseline pulse wave signals were collected during the user's waking and calm phase before sleep monitoring began. The starting point of the EEG response is marked based on the starting position of the frequency band energy change in the baseline EEG signal; Mark the starting point of the pulse wave response based on the starting position of heart rate change or heart rate variability change in the baseline pulse wave signal; The time interval between the EEG response start point and the pulse wave response start point was calculated, and the time interval was used as the individual lag baseline.

[0009] Furthermore, the hysteresis window module generates a hysteresis tracking window located after the trigger anchor point based on the trigger anchor point and the individual hysteresis baseline, including: Read the brainwave frequency band energy change characteristics and brainwave mutation slope characteristics corresponding to the trigger anchor point, and generate brainwave trigger intensity level; The delay compensation coefficient is called based on the EEG trigger intensity level, and the individual delay baseline is corrected using the delay compensation coefficient to obtain the current delay duration. The window starts at the trigger anchor point and ends at the window according to the current hysteresis duration, forming a hysteresis tracking window. The hysteresis tracking window is sent to the pulse wave delay matching module, which then retrieves the pulse wave delay response features within the hysteresis tracking window.

[0010] Furthermore, coherent matching is performed on the heart rate variability characteristics and heart rate variability characteristics to obtain pulse wave delay response characteristics, including: Heart rate change features and heart rate variability change features that fall within the hysteresis tracking window are extracted from the pulse wavelength buffer window, and multiple pulse wave candidate response segments are formed according to a preset step size. Based on the EEG frequency band energy change characteristics and EEG mutation slope characteristics corresponding to the trigger anchor point, an EEG mutation reference sequence is generated, and the normalized cross-correlation value between the EEG mutation reference sequence and each pulse wave candidate response segment is calculated respectively. Based on the normalized cross-correlation value and the delay deviation of each pulse wave candidate response segment relative to the individual hysteresis baseline, a coherent matching score is generated, and the pulse wave candidate response segments whose coherent matching scores meet the matching threshold are used as pulse wave delay response features.

[0011] Furthermore, the feedforward acquisition control module controls the PPG module to increase the sampling rate and LED drive current within the hysteresis tracking window. Based on the hysteresis tracking window generated by the hysteresis window module, it immediately generates a sampling enhancement control command, carrying the start and end times of the hysteresis tracking window, enabling the PPG module to enter the enhanced acquisition state only within this critical time range, including: After the trigger anchor point is marked, a sampling enhancement control instruction is generated based on the hysteresis tracking window. The sampling enhancement control instruction includes a sampling rate increase instruction and an LED drive current increase instruction. The sampling enhancement control command is sent to the PPG module, which enables the PPG module to acquire pulse wave signals in enhanced acquisition state within the hysteresis tracking window. Enhanced acquisition state includes increasing the sampling rate and increasing the LED drive current within a preset safe drive range. After the hysteresis tracking window ends, a data acquisition recovery command is sent to the PPG module to restore the PPG module to normal data acquisition state.

[0012] Furthermore, the fusion judgment module determines sleep monitoring results based on trigger anchor points, pulse wave delay response characteristics, and individual hysteresis baselines, including: Based on the characteristics of brainwave frequency band energy change and brainwave mutation slope corresponding to the trigger anchor point, brainwave trigger weights are generated. Based on the coherent matching score of the pulse wave delayed response features and the delay deviation of the pulse wave delayed response features relative to the individual hysteresis baseline, pulse wave response weights are generated. Based on the EEG trigger weight and pulse wave response weight, the system outputs the micro-awakening judgment result, the EEG artifact pending confirmation result, or the single EEG abnormality result.

[0013] Furthermore, based on the EEG trigger weight and pulse wave response weight, the system outputs micro-awakening determination results, EEG artifact confirmation results, or single EEG abnormality results, including: When the EEG trigger weight reaches the EEG trigger threshold and the pulse wave response weight reaches the pulse wave response threshold, the micro-awakening judgment result is output. When the EEG trigger weight reaches the EEG trigger threshold, the pulse wave response weight does not reach the pulse wave response threshold, and the pulse wave signal quality within the hysteresis tracking window does not meet the effective acquisition conditions, the EEG artifact pending confirmation result is output. When the EEG trigger weight reaches the EEG trigger threshold, the pulse wave response weight does not reach the pulse wave response threshold, and the pulse wave signal quality within the hysteresis tracking window meets the effective acquisition conditions, a single EEG abnormality result is output.

[0014] Furthermore, the fusion judgment module also generates a sleep monitoring report based on the sleep monitoring results, including: Based on the trigger anchor point corresponding to the micro-awakening judgment result, mark the occurrence time of the micro-awakening event; The degree of sleep fragmentation is generated based on the number of micro-awakening events per unit time and the time interval between adjacent micro-awakening events; The output includes a sleep monitoring report showing the timing of micro-awakening events, the degree of sleep fragmentation, and the delayed response status of pulse waves.

[0015] The beneficial effects of this invention are as follows: 1. This invention generates a hysteresis tracking window located after the EEG trigger anchor point, and matches PPG delayed response features within the window to avoid missed or false judgments caused by hard alignment of EEG and PPG at the same time, thereby improving the accuracy of micro-awakening recognition.

[0016] 2. This invention obtains the individual latency baseline through the waking and calming phase, and modifies the latency tracking window based on the EEG trigger intensity, enabling the system to adapt to the differences in central nervous system response and peripheral pulse wave response delay among different users, thereby improving the universality of sleep monitoring.

[0017] 3. This invention improves the PPG module's sampling rate within the hysteresis tracking window and increases the LED driving current within a preset safe driving range by using feedforward acquisition control after EEG triggering, thereby enhancing the PPG signal acquisition quality within the critical window.

[0018] 4. This invention uses EEG trigger weights, pulse wave response weights, and PPG signal quality to determine branches, distinguishing between micro-awakening, EEG artifacts to be confirmed, and single EEG abnormalities, avoiding misjudgment of a single signal, and improving the reliability and interpretability of sleep monitoring results.

[0019] Other advantages, objectives, and features of the invention will be set forth in part in the description which follows, and in part will be apparent to those skilled in the art from the following examination, or may be learned from practice of the invention. The objectives and other advantages of the invention can be realized and obtained through the following description. Attached Figure Description

[0020] To make the objectives, technical solutions, and advantages of the present invention clearer, the preferred embodiments of the present invention will be described in detail below with reference to the accompanying drawings, wherein: Figure 1 This is an architecture diagram of a sleep monitoring system based on a brain-computer interface and a PPG module, according to an embodiment of the present invention. Detailed Implementation

[0021] The following specific examples illustrate the implementation of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and various details in this specification can be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of the present invention. Unless otherwise specified, the following embodiments and features can be combined with each other.

[0022] The accompanying drawings are for illustrative purposes only and are schematic diagrams, not actual pictures. They should not be construed as limiting the invention. To better illustrate the embodiments of the invention, some parts in the drawings may be omitted, enlarged, or reduced, and do not represent the actual product dimensions. It is understandable to those skilled in the art that some well-known structures and their descriptions may be omitted in the drawings.

[0023] In the accompanying drawings of the embodiments of the present invention, the same or similar reference numerals correspond to the same or similar components. In the description of the present invention, it should be understood that if terms such as "upper," "lower," "left," "right," "front," and "rear" indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, they are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, the terms used to describe positional relationships in the drawings are only for illustrative purposes and should not be construed as limiting the present invention. For those skilled in the art, the specific meaning of the above terms can be understood according to the specific circumstances.

[0024] Please see Figure 1 This is a sleep monitoring system based on a brain-computer interface and a PPG module.

[0025] Example Figure 1 This embodiment illustrates a specific implementation of a sleep monitoring system based on a brain-computer interface and a PPG module, including: The brain-computer interface module collects brain signals during the user's sleep process and extracts the characteristics of brain frequency band energy changes and brain signal abrupt change slope characteristics; Specifically, the brain-computer interface module can be placed on the user's head using a headband, forehead patch, or cap-type EEG acquisition structure to continuously acquire EEG signals during the user's sleep. The brain-computer interface module can perform preprocessing on the acquired EEG signals, such as filtering, removing power frequency interference, and removing abnormal spikes, to reduce the impact of electrode contact fluctuations, environmental noise, and body movement interference on EEG analysis. The brain-computer interface module performs windowing processing on the preprocessed EEG signals according to the preset EEG analysis windows, and extracts the EEG frequency band energy change features and EEG abrupt change slope features in each EEG analysis window. The EEG frequency band energy change features are used to characterize the energy change of the target EEG frequency band relative to the sleep baseline state in the current EEG analysis window, and the EEG abrupt change slope features are used to characterize the rate of change of EEG frequency band energy between adjacent EEG analysis windows. The target EEG frequency band can be selected according to the sleep monitoring needs and the frequency band related to the wakefulness response.

[0026] The PPG module collects pulse wave signals during the user's sleep process and extracts heart rate change characteristics and heart rate variability change characteristics. Specifically, the PPG module can adopt a wristband, finger clip, ear clip, or patch-type photoplethysmography (PPG) acquisition structure to continuously acquire pulse wave signals during the user's sleep. The PPG module includes a light-emitting diode (LED) and a photodetector. The LED emits detection light into the skin tissue, and the photodetector receives the light signal after reflection or transmission through the tissue and converts the light intensity change into a pulse wave signal. During the acquisition process, the PPG module can perform DC component removal, bandpass filtering, and abnormal peak removal on the pulse wave signal to reduce the impact of ambient light, contact pressure changes, and body movement interference on subsequent analysis. After obtaining the preprocessed pulse wave signal, the PPG module identifies continuous peaks, troughs, or effective pulse cycles in the pulse wave and extracts heart rate variability features based on the time interval between adjacent pulse cycles. These heart rate variability features characterize the rise, fall, or short-term fluctuations in the user's heart rate over time during sleep. The PPG module can also extract heart rate variability features based on the interval changes between multiple consecutive pulse cycles, which characterizes the heart rate interval fluctuations caused by autonomic nervous system regulation.

[0027] The dual-track caching module caches the characteristics of brainwave frequency band energy changes, brainwave mutation slopes, heart rate changes, and heart rate variability changes, respectively. Specifically, the dual-track caching module caches EEG frequency band energy variation characteristics, EEG mutation slope characteristics, heart rate variation characteristics, and heart rate variability variation characteristics, including: Based on the sampling time sequence of EEG signals, the characteristics of EEG frequency band energy change and the characteristics of EEG mutation slope are written into the short-term buffer window of EEG. Based on the sampling time sequence of the pulse wave signal, the heart rate change characteristics and heart rate variability change characteristics are written into the pulse wavelength buffer window; Based on the start and end times of the short-term buffer window of EEG and the start and end times of the buffer window of pulse wavelength, a time index relationship between EEG features and pulse wave features is established. Among them, the buffer duration of the pulse wavelength time buffer window is greater than that of the EEG short-time buffer window, so that the pulse wavelength time buffer window retains the pulse wave characteristics after the trigger anchor point.

[0028] Specifically, the dual-track buffer module receives the EEG frequency band energy change characteristics, EEG mutation slope characteristics, and corresponding EEG timestamps output by the brain-computer interface module, and continuously writes the above EEG characteristics into the short-term EEG buffer window according to the sampling time sequence of the EEG signal. The short-term EEG buffer window can adopt a circular buffer structure. When the characteristics of a new EEG analysis window are written, the historical EEG characteristics that have exceeded the buffer time are removed in sequence, so that the system always retains the transient change information of EEG in the most recent period of time, which makes it easier for the EEG triggering module to identify the mutation trend of the EEG side in a timely manner, without having to retain a large amount of historical EEG data for a long time. In another data path, the dual-track buffer module receives the heart rate change characteristics, heart rate variability change characteristics, and corresponding pulse wave timestamps output by the PPG module, and writes the above pulse wave characteristics into the pulse wavelength time buffer window according to the sampling time sequence of the pulse wave signal. Since PPG reflects the results of peripheral blood flow and autonomic nervous regulation, its response is usually later than EEG mutation. Therefore, the buffer duration of the pulse wavelength time buffer window is set to be greater than the buffer duration of the EEG short-time buffer window, so that after the EEG trigger anchor point is generated, the pulse wave change characteristics for a period of time after the trigger anchor point can still be retained, providing a data basis for the delayed response matching in the subsequent hysteresis tracking window. To ensure that EEG features and pulse wave features can be accurately correlated in subsequent processing, the dual-track caching module also establishes a time index relationship between EEG features and pulse wave features based on the start and end times of the EEG short-term caching window and the start and end times of the pulse wavelength time caching window. Specifically, the start and end times, center time points, and corresponding EEG features of each EEG analysis window can be timestamped with the start and end times, center time points, and corresponding pulse wave features of each pulse wave analysis window. This allows the hysteresis window module to extract pulse wave feature segments located after the trigger anchor point from the pulse wavelength time caching window according to the time index after obtaining the trigger anchor point. For example, the short-term EEG buffer window can retain only the EEG frequency band energy change features and EEG mutation slope features within the most recent few EEG analysis windows to quickly determine whether an EEG mutation has occurred. The pulse wavelength buffer window retains heart rate change features and heart rate variability change features over a longer period of time. When the EEG triggering module marks a trigger anchor point in a certain EEG analysis window, the system does not immediately use the PPG features at that anchor point as the fusion basis. Instead, it uses the time index relationship already established in the dual-track buffer module to retrieve the pulse wave features after the trigger anchor point from the pulse wavelength buffer window for the pulse wave delay matching module to retrieve the delay response features. Through the aforementioned dual-track caching method, the system forms an asymmetric caching mechanism in its data structure, which captures instantaneous mutations via a short window for EEG and retains delayed responses via a pulse wavelength window. This not only improves the real-time performance of EEG mutation triggering but also avoids the loss of effective pulse wave responses during fusion at the same time point due to PPG response lag. This provides a continuous and indexable data foundation for subsequent delayed tracking and phase compensation based on trigger anchor points.

[0029] The EEG triggering module determines the trigger anchor point based on the characteristics of EEG frequency band energy changes and the characteristics of EEG mutation slope. Specifically, the EEG triggering module obtains trigger anchor points based on the characteristics of EEG frequency band energy changes and the characteristics of EEG mutation slopes, including: Read the brainwave frequency band energy change characteristics and brainwave mutation slope characteristics corresponding to the current brainwave analysis window from the short-term buffer window; The difference between the characteristics of brainwave frequency band energy change and the sleep reference frequency band energy is calculated to obtain the brainwave frequency band mutation amount; Based on the variation amplitude of the EEG slope characteristics between adjacent EEG analysis windows, the EEG slope mutation amount is obtained. When the frequency band mutation amount of EEG reaches the frequency band mutation threshold, the slope mutation amount of EEG reaches the slope mutation threshold, and the direction of frequency band energy enhancement corresponding to the frequency band mutation amount is consistent with the direction of micro-awakening EEG response, the time point corresponding to the current EEG analysis window is marked as the trigger anchor point.

[0030] Specifically, the electrical triggering module reads the EEG frequency band energy change characteristics and EEG mutation slope characteristics corresponding to the current EEG analysis window from the EEG short-term buffer window, and simultaneously reads the timestamp of the EEG analysis window. Since the EEG short-term buffer window continuously writes EEG features according to the sampling time order, the EEG triggering module can determine whether there are sudden EEG changes in each new analysis window after it enters the buffer. The current EEG analysis window can be understood as the latest analysis window that has entered or is being detected in the EEG short-term buffer window, and its corresponding time point can be the center time, start time or preset representative time point of the analysis window. After reading the brainwave frequency band energy change characteristics of the current EEG analysis window, the EEG triggering module calculates the difference between it and the sleep reference frequency band energy to obtain the EEG frequency band mutation variable. The sleep reference frequency band energy can come from historical stable segments after the user enters stable sleep, or from the average or median value of several stable EEG analysis windows within the current sleep stage. Preferably, to reduce the impact of differences in EEG amplitude among different users on the judgment result, a normalized difference method can be used to calculate the EEG frequency band mutation variable.

[0031] in, Indicates the first Each EEG analysis window corresponds to a mutation variable in the EEG frequency band. Indicates the first The characteristics of brainwave frequency band energy changes corresponding to each brainwave analysis window. Indicates the baseline frequency band energy during sleep. This represents a stable term used to avoid a denominator of zero. A positive value indicates that the target frequency band energy of the current EEG analysis window is enhanced relative to the sleep baseline frequency band energy. If the value is large, it indicates that there is a significant trend of sudden increase in frequency band energy within the analysis window. While obtaining the EEG band mutation variable, the EEG triggering module also obtains the EEG slope mutation variable based on the change amplitude of the EEG mutation slope characteristics between adjacent EEG analysis windows. This processing is used to determine whether the EEG changes have the characteristic of rapid enhancement in a short period of time, rather than being caused solely by slow drift or natural changes during sleep. Preferably, the EEG slope mutation variable can be expressed as:

[0032] in, Indicates the first The EEG slope mutation variable corresponding to each EEG analysis window Indicates the first The characteristics of the slope of brain electrical aberrations corresponding to each brain electrical analysis window. Indicates the first The EEG analysis window corresponds to the slope characteristics of EEG mutations, so the EEG triggering module not only focuses on whether the current EEG energy is rising, but also on whether the rate of change of EEG changes has changed abruptly, thereby reducing the risk of false triggering caused by a single energy threshold judgment. After completing the above calculations, the EEG triggering module compares the EEG band mutation variable with the band mutation threshold and the EEG slope mutation variable with the slope mutation threshold. The band mutation threshold and slope mutation threshold can be configured according to the user's sleep baseline state, historical detection data, or preset experience range. Only when the EEG band mutation variable reaches the band mutation threshold, the EEG slope mutation variable reaches the slope mutation threshold, and the direction of the band energy enhancement corresponding to the EEG band mutation variable is consistent with the direction of the micro-awakening EEG response, will the EEG triggering module mark the time point corresponding to the current EEG analysis window as the trigger anchor point. Here, the direction of the micro-awakening EEG response refers to the direction of the target band energy enhancement related to micro-awakening, such as a short-term enhancement of the target band energy relative to the sleep baseline state, rather than a decrease or irregular fluctuation. For example, when a user is in a stable sleep state, the target EEG frequency band energy is in a low and stable range. If the target frequency band energy in a current EEG analysis window is significantly higher than the sleep baseline frequency band energy, and the slope change between adjacent analysis windows is also significantly increased, and this change is manifested as a frequency band energy enhancement direction related to micro-awakening, then the EEG triggering module marks the center time of the analysis window as the trigger anchor point. Conversely, if there is only an increase in frequency band energy but no significant slope change, or a large slope change but the frequency band energy enhancement direction is inconsistent with the micro-awakening EEG response direction, then no trigger anchor point is marked to reduce false triggering caused by slow EEG fluctuations, short-term noise, or changes in non-target directions. The EEG triggering module further transforms the basic EEG features output by the brain-computer interface module into EEG frequency band mutations and EEG slope mutations that can be used to trigger events. It uses a combination of threshold and directional consistency to jointly determine whether a trigger anchor point is generated, enabling the system to prioritize capturing instantaneous changes on the central nervous system side. This trigger anchor point is then used as the time reference for the subsequent hysteresis window module to generate a hysteresis tracking window, thereby avoiding the direct mechanical alignment of EEG features and pulse wave features at the same moment.

[0033] Hysteresis window module, which generates a hysteresis tracking window located after the trigger anchor point based on the trigger anchor point and the individual hysteresis baseline; Specifically, the individual hysteresis baseline in the hysteresis window module is obtained in the following way: Baseline EEG and baseline pulse wave signals were collected during the user's waking and calm phase before sleep monitoring began. The starting point of the EEG response is marked based on the starting position of the frequency band energy change in the baseline EEG signal; Mark the starting point of the pulse wave response based on the starting position of heart rate change or heart rate variability change in the baseline pulse wave signal; The time interval between the EEG response start point and the pulse wave response start point was calculated, and the time interval was used as the individual lag baseline.

[0034] Specifically, the lag window module acquires an individual lag baseline before formal sleep monitoring. Specifically, after the user wears the brain-computer interface module and PPG module, baseline EEG signals and baseline pulse wave signals are collected simultaneously during the awake and calm phase, and the corresponding timestamps are recorded. Through this pre-calibration process, the subsequent lag tracking window no longer uses a fixed experience duration, but has a user-individualized basis. For baseline EEG signals, the hysteresis window module can extract target frequency band energy changes according to a preset analysis window. When the target frequency band energy in the baseline EEG signal shows a identifiable change relative to the initial state of wakefulness and calmness, the corresponding time point is marked as the EEG response start point. For baseline pulse wave signals, the hysteresis window module identifies the pulse wave response start position based on heart rate change characteristics or heart rate variability change characteristics, and marks the corresponding time point as the pulse wave response start point. Subsequently, the hysteresis window module calculates the time interval between the EEG response start point and the pulse wave response start point, and uses this time interval as the individual hysteresis baseline. Preferably, the individual hysteresis baseline can be expressed as:

[0035] in, Indicates the individual's baseline lag; This indicates the time corresponding to the start point of the EEG response; This indicates the time corresponding to the start point of the pulse wave response; For example, during the waking and calm phase, the system first identifies the starting position of the target frequency band energy change in the baseline EEG signal, and then identifies the starting position of the heart rate change in the baseline pulse wave signal. The time difference between the two is used as the individual lag baseline for this sleep monitoring of the user. In this way, the lag window module can provide an individualized time reference for the generation of the subsequent lag tracking window based on the delay relationship between the user's own central nervous system response and peripheral pulse wave response.

[0036] Specifically, the hysteresis window module generates a hysteresis tracking window located after the trigger anchor point based on the trigger anchor point and the individual hysteresis baseline, including: Read the brainwave frequency band energy change characteristics and brainwave mutation slope characteristics corresponding to the trigger anchor point, and generate brainwave trigger intensity level; The delay compensation coefficient is called based on the EEG trigger intensity level, and the individual delay baseline is corrected using the delay compensation coefficient to obtain the current delay duration. The window starts at the trigger anchor point and ends at the window according to the current hysteresis duration, forming a hysteresis tracking window. The hysteresis tracking window is sent to the pulse wave delay matching module, which then retrieves the pulse wave delay response features within the hysteresis tracking window.

[0037] Specifically, after the EEG triggering module records the triggering defect, the hysteresis window module reads the corresponding EEG frequency band energy change characteristics and EEG mutation slope characteristics, and generates an EEG trigger intensity level based on the magnitude of their changes. Specifically, the EEG frequency band energy change characteristics and EEG mutation slope characteristics can be compared with their respective low, medium, and high grading thresholds. When both are in the low-level range, a low trigger intensity level is generated; when at least one of them is in the medium-level range and has not reached the high-level range, a medium trigger intensity level is generated; when at least one of them reaches the high-level range and the other is not lower than the medium-level range, a high trigger intensity level is generated. Through this grading method, the hysteresis window module can convert the magnitude and rate of change of EEG mutations into the grading parameters required for subsequent window correction. After obtaining the EEG trigger intensity level, the hysteresis window module calls the hysteresis compensation coefficient corresponding to that level. The hysteresis compensation coefficient can be pre-stored in the system parameter table; for example, a low trigger intensity level corresponds to the first hysteresis compensation coefficient, a medium trigger intensity level corresponds to the second hysteresis compensation coefficient, and a high trigger intensity level corresponds to the third hysteresis compensation coefficient. Each hysteresis compensation coefficient is used to adjust the window length corresponding to the individual's hysteresis baseline. That is, the current hysteresis duration can be expressed as:

[0038] in, Indicates the first The current lag duration corresponding to each triggering defect Indicates the individual's baseline lag. Indicates the relationship with the first The system can also set a minimum and maximum allowed duration for the current hysteresis duration to avoid the window being too short or too long, so that the corrected current hysteresis duration is limited to the preset duration range. Subsequently, the hysteresis window module uses the time point corresponding to the triggering defect as the starting point of the window and configures the window ending point according to the current hysteresis duration, forming a hysteresis tracking window located after the triggering defect. The hysteresis tracking window can be represented as:

[0039] in, Indicates the first Each trigger defect corresponds to a hysteresis tracking window. Indicates the first The time point corresponding to each triggering defect Indicates the first The current hysteresis duration corresponding to each trigger defect is tracked by a hysteresis tracking window that is always located after the trigger defect. This window is used to cover the delayed response range that may occur in peripheral pulse waves relative to EEG mutations, thus avoiding the forced alignment of EEG signals and PPG signals at the same time. For example, the parameter table can be set as follows: the hysteresis compensation coefficient corresponding to the low trigger intensity level is less than or equal to 1, the hysteresis compensation coefficient corresponding to the medium trigger intensity level is close to 1, and the hysteresis compensation coefficient corresponding to the high trigger intensity level is greater than or equal to 1. If the EEG frequency band energy change and EEG mutation slope characteristics corresponding to a certain trigger defect are obvious, the hysteresis window module generates a high trigger intensity level and calls the corresponding hysteresis compensation coefficient to correct the individual hysteresis baseline, thereby forming a relatively long or sufficient hysteresis tracking window to cover possible delayed heart rate changes or heart rate variability changes. After the hysteresis tracking window is formed, the hysteresis window module sends the start and end times of the window to the pulse wave delay matching module. The module then extracts the corresponding heart rate change features and heart rate variability change features from the pulse wave time buffer window based on the start and end times of the window, and retrieves the pulse wave delay response features. Through the above processing, the system can dynamically configure the PPG tracking interval according to the individual hysteresis baseline and the current EEG trigger intensity, so that subsequent pulse wave matching does not depend on a fixed time window, thereby improving the reliability of the phase correspondence between the EEG instantaneous response and the PPG hysteresis response.

[0040] The pulse wave delay matching module performs coherent matching on heart rate change features and heart rate variability change features within the hysteresis tracking window to obtain pulse wave delay response features. In this embodiment, coherent matching is performed on heart rate variability characteristics and heart rate variability characteristics to obtain pulse wave delay response characteristics, including: Heart rate change features and heart rate variability change features that fall within the hysteresis tracking window are extracted from the pulse wavelength buffer window, and multiple pulse wave candidate response segments are formed according to a preset step size. Based on the EEG frequency band energy change characteristics and EEG mutation slope characteristics corresponding to the trigger anchor point, an EEG mutation reference sequence is generated, and the normalized cross-correlation value between the EEG mutation reference sequence and each pulse wave candidate response segment is calculated respectively. Based on the normalized cross-correlation value and the delay deviation of each pulse wave candidate response segment relative to the individual hysteresis baseline, a coherent matching score is generated, and the pulse wave candidate response segments whose coherent matching scores meet the matching threshold are used as pulse wave delay response features.

[0041] Specifically, after receiving the hysteresis tracking window sent by the hysteresis window module, the pulse wave delay matching module extracts the heart rate change features and heart rate variability change features falling within the time range from the pulse wave time buffer window according to the start and end time of the hysteresis tracking window. Subsequently, the pulse wave delay matching module performs sliding segmentation on the extracted pulse wave features according to a preset step size to form multiple pulse wave candidate response segments. Each candidate response segment corresponds to a delay time relative to the trigger anchor point. Through this processing, the system can retain multiple possible PPG delay response positions after the trigger anchor point, instead of only taking the PPG features at a fixed time for judgment. After forming pulse wave candidate response segments, the pulse wave delay matching module generates an EEG mutation reference sequence based on the EEG frequency band energy change characteristics and EEG mutation slope characteristics corresponding to the trigger anchor point. This EEG mutation reference sequence is used to characterize the change pattern of EEG mutations near the trigger anchor point, such as the direction of frequency band energy enhancement, the enhancement amplitude, and the steepness of the change. After normalizing each pulse wave candidate response segment, the pulse wave delay matching module calculates the normalized cross-correlation value between it and the EEG mutation reference sequence to obtain the degree of coherence between the two in terms of change trend. The higher the normalized cross-correlation value, the more consistent the candidate response segment is with the EEG mutation reference sequence in terms of change trend. Furthermore, the pulse wave delay matching module calculates the delay deviation of each pulse wave candidate response segment relative to the individual hysteresis baseline. Specifically, the start time, center time, or maximum response time of the candidate response segment can be used as its response time, and the time difference between this response time and the trigger anchor point is calculated to obtain the actual delay time. Then, the actual delay time is compared with the individual hysteresis baseline to obtain the delay deviation. Thus, the system not only determines whether the PPG segment and the EEG mutation have a consistent trend, but also determines whether the time of occurrence of the PPG segment conforms to the user's individual hysteresis pattern.

[0042] In a preferred embodiment, the coherence matching score can be expressed as:

[0043] in, Indicates the first Coherence matching score of candidate pulse wave response segments Indicates the first Normalized cross-correlation values ​​between candidate pulse wave response fragments and EEG mutation reference sequences Indicates the first The actual delay time of each pulse wave candidate response segment relative to the trigger anchor point. Indicates the individual's baseline lag. The formula represents the delay deviation penalty coefficient. In this formula, the normalized cross-correlation value is used to evaluate morphological coherence, and the delay deviation is used to evaluate physiological lag consistency. The higher the coherence matching score, the more consistent the candidate response fragment is with the EEG mutation trend and the closer its actual delay time is to the individual lag baseline. For example, after a certain trigger anchor point, there are multiple pulse wave candidate response segments within the hysteresis tracking window. One of the candidate response segments shows a short-term increase in heart rate accompanied by changes in heart rate variability, and its occurrence time is close to the user's individual hysteresis baseline. In this case, the candidate response segment obtains a high coherence matching score. Conversely, if a candidate response segment has heart rate fluctuations, but its occurrence time deviates significantly from the individual hysteresis baseline, or its trend is inconsistent with the EEG mutation reference sequence, its coherence matching score decreases.

[0044] Finally, the pulse wave delay matching module uses pulse wave candidate response segments whose coherence matching scores meet the matching threshold as pulse wave delay response features. If there are multiple candidate response segments that meet the matching threshold, the candidate response segment with the highest coherence matching score is selected as the pulse wave delay response feature. In this way, the system can filter out PPG delay responses within the hysteresis tracking window that are both coherent with the trend of EEG mutation and conform to the individual hysteresis time pattern, thereby improving the reliability of the correspondence between EEG response and autonomic nervous system response in micro-awakening determination.

[0045] The feedforward acquisition control module controls the PPG module to improve the sampling rate and LED drive current within the hysteresis tracking window; Specifically, the feedforward acquisition control module controls the PPG module to increase the sampling rate and LED drive current within the hysteresis tracking window, including: After the trigger anchor point is marked, a sampling enhancement control instruction is generated based on the hysteresis tracking window. The sampling enhancement control instruction includes a sampling rate increase instruction and an LED drive current increase instruction. The sampling enhancement control command is sent to the PPG module, which enables the PPG module to acquire pulse wave signals in enhanced acquisition state within the hysteresis tracking window. Enhanced acquisition state includes increasing the sampling rate and increasing the LED drive current within a preset safe drive range. After the hysteresis tracking window ends, a data acquisition recovery command is sent to the PPG module to restore the PPG module to normal data acquisition state.

[0046] Specifically, after the EEG triggering module marks the trigger anchor point, the feedforward acquisition control module does not wait for the pulse wave signal to change naturally. Instead, based on the hysteresis tracking window generated by the hysteresis window module, it immediately generates a sampling enhancement control command. This sampling enhancement control command includes a sampling rate increase command and a light-emitting diode drive current increase command, and carries the start and end times of the hysteresis tracking window. This allows the PPG module to enter the enhanced acquisition state only within this critical time range. In this way, the system utilizes the early response characteristics of the EEG signal relative to the pulse wave signal to perform feedforward control on the PPG module, instead of passively compensating after the pulse wave signal has been interfered with. In enhanced acquisition mode, the PPG module increases the sampling frequency of the pulse wave signal according to the sampling rate increase instruction, so that the heart rate change characteristics and heart rate variability change characteristics within the hysteresis tracking window have higher time resolution. At the same time, the PPG module increases the LED drive current within the preset safe drive range according to the LED drive current increase instruction, so as to enhance the light signal penetration and echo signal strength. The preset safe drive range can be pre-limited by the hardware specifications of the PPG module, the allowable operating current of the LED, and the user's wearing safety requirements, so that the enhanced acquisition will not exceed the safe operating range of the device. For example, when the EEG triggering module marks the trigger anchor point at a certain moment, the hysteresis window module generates a hysteresis tracking window located after the trigger anchor point. Based on this, the feedforward acquisition control module controls the PPG module to increase the sampling rate within the hysteresis tracking window and appropriately increase the LED driving current. If the user experiences slight turning over or changes in contact pressure after micro-awakening, the enhanced acquisition state can improve the recognizability of the pulse wave signal within that time period, making it easier for the subsequent pulse wave delay matching module to retrieve the delayed response segment corresponding to the heart rate change or heart rate variability change. After the hysteresis tracking window ends, the feedforward acquisition control module sends an acquisition recovery command to the PPG module, so that the PPG module returns to the normal sampling rate and normal LED drive current. This recovery process can be automatically triggered according to the end time of the hysteresis tracking window, or it can be triggered after the pulse wave delay matching module completes the response retrieval within the window. Through the windowed enhanced acquisition method, the system only increases the PPG hardware resources in the critical hysteresis response stage after the EEG is triggered, so as to avoid the PPG module being in a high power consumption state for a long time. The feedforward acquisition control module converts EEG trigger anchors into hardware-enhanced acquisition control signals for the PPG module, improving signal acquisition quality within the critical window while waiting for pulse wave delay responses. Its technical advantages are twofold: firstly, it reduces the risk of PPG feature loss due to body movement, changes in contact pressure, or low perfusion within the hysteresis tracking window; secondly, it allows for resumption of normal acquisition after the window ends, balancing sleep monitoring accuracy with device power consumption control.

[0047] The fusion judgment module determines the sleep monitoring results based on the trigger anchor point, pulse wave delay response characteristics, and individual hysteresis baseline.

[0048] Specifically, the fusion judgment module determines sleep monitoring results based on trigger anchor points, pulse wave delay response characteristics, and individual hysteresis baselines, including: Based on the characteristics of brainwave frequency band energy change and brainwave mutation slope corresponding to the trigger anchor point, brainwave trigger weights are generated. Based on the coherent matching score of the pulse wave delayed response features and the delay deviation of the pulse wave delayed response features relative to the individual hysteresis baseline, pulse wave response weights are generated. Based on the EEG trigger weight and pulse wave response weight, the system outputs the micro-awakening judgment result, the EEG artifact pending confirmation result, or the single EEG abnormality result.

[0049] Specifically, after receiving the trigger anchor point and pulse wave delay response features, the fusion determination module first generates EEG trigger electrical weights based on the EEG frequency band energy change features and EEG abrupt change slope features corresponding to the trigger anchor point. Specifically, the EEG frequency band energy change features reflect the enhancement amplitude of the target EEG frequency band at the trigger anchor point, and the EEG abrupt change slope features reflect the steepness of this enhancement process. The fusion determination module can normalize these two features and then weight and fuse them to obtain the EEG trigger electrical weights. Preferably, the EEG trigger electrical weights can be expressed as:

[0050] in, Indicates the trigger weight of brainwaves. This represents the normalized value of the energy change characteristics of the EEG frequency band corresponding to the trigger anchor point. This represents the normalized value of the EEG mutation slope feature corresponding to the trigger anchor point. The fusion coefficient represents the characteristics of changes in brainwave frequency band energy. This brainwave trigger weight is used to characterize the reliability of brainwave-side mutations at the current trigger anchor point. The higher the value, the more likely there is a transient response related to micro-awakening on the brainwave side. On the pulse wave side, the fusion determination module generates pulse wave response weights based on the coherence matching score of the pulse wave delayed response feature and the delay deviation of the pulse wave delayed response feature relative to the individual hysteresis baseline. The coherence matching score is used to reflect the consistency of the change trend between the pulse wave candidate response segment and the EEG mutation reference sequence, and the delay deviation is used to reflect whether the time of occurrence of the pulse wave delayed response feature is close to the user's own individual hysteresis baseline. In one embodiment, the pulse wave response weights can be expressed as:

[0051] in, Indicates the pulse wave response weight. The coherent matching score represents the characteristics of the pulse wave delayed response. This indicates the amount of delay deviation of the pulse wave delayed response characteristics relative to the individual's hysteresis baseline. This represents the penalty coefficient corresponding to the amount of delay deviation. With this setting, the pulse wave response weight depends not only on whether the PPG change is significant, but also on whether the PPG change occurs within a time range that conforms to the individual's physiological lag pattern. Subsequently, the fusion judgment module outputs sleep monitoring results based on the weight of the EEG trigger current and the weight of the pulse wave response. Specifically, when the weight of the EEG trigger current is high and the weight of the pulse wave response also meets the corresponding requirements, it indicates that the EEG mutation at the trigger anchor point has obtained PPG delayed response support within the hysteresis tracking window, and the fusion judgment module can output the micro-arousal judgment result. When the weight of the EEG trigger current is high, but no effective pulse wave response is obtained within the hysteresis tracking window, the fusion judgment module combines the subsequent pulse wave signal quality judgment and outputs the EEG artifact pending confirmation result or the single EEG abnormality result. Thus, the system no longer directly gives the micro-arousal judgment based solely on the EEG mutation, but introduces the PPG delayed response relationship for cross-confirmation. For example, if the EEG frequency band energy change is significant and the EEG mutation slope is large at a certain trigger anchor point, the EEG trigger weight is high. In the hysteresis tracking window after the trigger anchor point, if the pulse wave delay matching module further retrieves a short-term increase in heart rate or a change in heart rate variability, and the coherence matching score of the response segment is high and the delay deviation is small, the pulse wave response weight is high. At this time, the fusion judgment module has a high reliability in judging the event as a micro-awakening event. Conversely, if a mutation occurs on the EEG side but the PPG side lacks response features that conform to the individual hysteresis pattern, the system will not simply output the micro-awakening result, but will enter the judgment path of EEG artifact to be confirmed or single EEG abnormality. By integrating the instantaneous EEG response and the delayed pulse wave response into a corresponding verification relationship, the misjudgment caused by electrode disturbance and short-term noise when relying solely on EEG can be reduced, and the trigger source can be missed due to response lag when relying solely on PPG, thereby improving the reliability and interpretability of sleep micro-awake monitoring results.

[0052] Specifically, based on the EEG trigger weight and pulse wave response weight, the system outputs micro-awakening determination results, EEG artifact confirmation results, or single EEG abnormality results, including: When the EEG trigger weight reaches the EEG trigger threshold and the pulse wave response weight reaches the pulse wave response threshold, the micro-awakening judgment result is output. When the EEG trigger weight reaches the EEG trigger threshold, the pulse wave response weight does not reach the pulse wave response threshold, and the pulse wave signal quality within the hysteresis tracking window does not meet the effective acquisition conditions, the EEG artifact pending confirmation result is output. When the EEG trigger weight reaches the EEG trigger threshold, the pulse wave response weight does not reach the pulse wave response threshold, and the pulse wave signal quality within the hysteresis tracking window meets the effective acquisition conditions, a single EEG abnormality result is output.

[0053] Specifically, the fusion judgment module compares the EEG trigger weight with the EEG trigger threshold and the pulse wave response weight with the pulse wave response threshold. The EEG trigger threshold is used to determine whether the EEG mutation at the trigger anchor point reaches the intensity of the micro-awakening related response, and the pulse wave response threshold is used to determine whether there is a pulse wave delayed response corresponding to the EEG mutation within the hysteresis tracking window. When the EEG trigger weight reaches the EEG trigger threshold and the pulse wave response weight reaches the pulse wave response threshold, it indicates that there is a transient change on the EEG side and a corresponding delayed response on the PPG side within the hysteresis tracking window. The fusion judgment module outputs the micro-awakening judgment result. This method can avoid directly judging micro-awakening based on a single EEG mutation and improve the reliability of the judgment. When the EEG trigger weight reaches the EEG trigger threshold, the pulse wave response weight does not reach the pulse wave response threshold, and the pulse wave signal quality within the hysteresis tracking window does not meet the effective acquisition conditions, the fusion judgment module outputs the EEG artifact pending confirmation result. The effective acquisition conditions can be judged based on the continuity of the pulse wave cycle, the effective pulse amplitude, the integrity of the peak and trough, the short-term amplitude mutation, and the contact stability. At this time, no effective response is formed on the PPG side, which may be caused by motion artifacts, low perfusion, or loose contact. Therefore, the EEG trigger result is not directly rejected. When the EEG trigger weight reaches the EEG trigger threshold, the pulse wave response weight does not reach the pulse wave response threshold, and the pulse wave signal quality within the hysteresis tracking window meets the effective acquisition conditions, the fusion judgment module outputs a single EEG abnormality result. This situation indicates that there is a mutation on the EEG side, but the PPG side does not show a delayed response that conforms to the individual hysteresis pattern under the effective acquisition state. Therefore, this event is not directly output as a micro-awakening event. For example, if the EEG trigger weight is high at the trigger anchor point, and a short-term increase in heart rate or heart rate variability occurs within the hysteresis tracking window, a micro-arousal determination result is output. If EEG triggers are also present, but the PPG waveform within the hysteresis tracking window is disrupted due to turning over, an EEG artifact pending confirmation result is output. If the PPG waveform quality is stable but no delayed response is observed, a single EEG abnormality result is output. Through these branches, the system can distinguish between genuine micro-arousals, PPG acquisition failures, and single EEG abnormalities, improving the accuracy and interpretability of sleep monitoring results.

[0054] Specifically, the fusion judgment module also generates a sleep monitoring report based on the sleep monitoring results, including: Based on the trigger anchor point corresponding to the micro-awakening judgment result, mark the occurrence time of the micro-awakening event; The degree of sleep fragmentation is generated based on the number of micro-awakening events per unit time and the time interval between adjacent micro-awakening events; The output includes a sleep monitoring report showing the timing of micro-awakening events, the degree of sleep fragmentation, and the delayed response status of pulse waves.

[0055] Specifically, after obtaining the sleep monitoring results, the fusion judgment module associates the events identified as micro-arousals with their corresponding trigger anchors and uses the time point corresponding to the trigger anchor as the occurrence time of the micro-arousal event. Since the trigger anchor comes from the transient mutation on the EEG side, marking the micro-arousal event with this time point can more accurately reflect the location where the central nervous system response first appears, rather than mistakenly taking the PPG response time that appears later as the starting point of the event. After completing the micro-awakening event labeling, the fusion judgment module counts the micro-awakening events according to a preset statistical period. Specifically, it can count the number of micro-awakening events per unit time and calculate the time interval between adjacent micro-awakening events. When the number of micro-awakening events per unit time is large and the time interval between adjacent events is short, it indicates that the user's sleep process is frequently interrupted, and the corresponding sleep fragmentation degree is high. Preferably, the sleep fragmentation degree can be expressed as:

[0056] in, Indicates the degree of sleep fragmentation. This indicates the number of micro-awakening events within the statistical period. Indicates the duration of the statistical period. This represents the average time interval between adjacent micro-arousal events. This represents the weighting coefficient corresponding to the number of micro-awakening events. This represents a stable term used to avoid a denominator of zero; For example, if the system marks multiple micro-awake events within the same statistical period and the intervals between these micro-awake events are short, the degree of sleep fragmentation is high. If the number of micro-awake events is small and the intervals between adjacent events are long, the degree of sleep fragmentation is low. In this way, the sleep monitoring report can not only record individual micro-awake events, but also reflect the continuity of the user's entire sleep process. Finally, the fusion judgment module outputs a sleep monitoring report, which includes at least the occurrence time of micro-arousals, the degree of sleep fragmentation, and the pulse wave delayed response status. The pulse wave delayed response status can include whether pulse wave delayed response features were retrieved, the corresponding matching score, and the delay deviation relative to the individual's delayed response baseline. Through this report, it is possible to understand EEG triggering events, PPG delayed response, and the degree of sleep fragmentation simultaneously, thereby improving the interpretability of sleep monitoring results.

[0057] Example 2 This embodiment provides a detailed description using a home-based wearable sleep monitoring example.

[0058] Before sleep, the user wears a head-mounted brain-computer interface module and a wristband-mounted PPG module. The system first simultaneously acquires baseline EEG and baseline pulse wave signals during the user's waking and calm phase, obtaining an individual lag baseline based on the time interval between the EEG response start point and the pulse wave response start point. Upon entering sleep monitoring, the brain-computer interface module continuously acquires EEG signals and extracts EEG frequency band energy change characteristics and EEG mutation slope characteristics, while the PPG module continuously acquires pulse wave signals and extracts heart rate change characteristics and heart rate variability change characteristics. When the EEG triggering module identifies EEG mutations related to micro-arousal and marks the trigger anchor point, the lag window module generates a lag tracking window located after the trigger anchor point based on the trigger anchor point and the individual lag baseline. The pulse wave delay matching module retrieves pulse wave delay response features corresponding to EEG mutations within this window. The fusion judgment module then outputs the micro-arousal judgment result, sleep fragmentation degree, and pulse wave delay response status. This embodiment is suitable for home-based overnight sleep monitoring and can improve the reliability of micro-arousal event recognition by utilizing PPG hysteresis response relationship without requiring EEG signals and PPG signals to be forcibly aligned at the same time.

[0059] Example 3, This embodiment provides a detailed description using an example of respiratory-related microarousal monitoring.

[0060] If a user experiences respiratory instability, a short-term hypoxia trend, or a recovery response after apnea during sleep, the EEG may initially show increased target band energy and a higher slope of EEG spikes, while heart rate changes or heart rate variability changes on the PPG side usually appear with a lag. After the EEG trigger module marks the trigger anchor point, the hysteresis window module generates a hysteresis tracking window based on the individual hysteresis baseline. The feedforward acquisition control module controls the PPG module to increase the sampling rate within this window and increase the LED drive current within a preset safe driving range to enhance the acquisition quality of the PPG delayed response after respiratory-related events. If the pulse wave delay matching module finds a heart rate change or heart rate variability change within the hysteresis tracking window that is coherent with the EEG spike and has a small delay deviation, the fusion judgment module outputs a micro-awakening judgment result. If the PPG signal within the hysteresis tracking window does not meet the effective acquisition conditions due to turning over, low perfusion, or poor contact, the EEG artifact pending confirmation result is output. This embodiment demonstrates the application effect of the present invention in phase compensation of the central nervous system transient response and the autonomic nervous system delayed response in respiratory-related sleep events.

[0061] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A sleep monitoring system based on a brain-computer interface and a PPG module, characterized in that: include: The brain-computer interface module collects brain signals during the user's sleep process and extracts the characteristics of brain frequency band energy changes and brain signal abrupt change slope characteristics; The PPG module collects pulse wave signals during the user's sleep process and extracts heart rate change characteristics and heart rate variability change characteristics. A dual-track caching module, which respectively caches the brainwave frequency band energy change characteristics and the brainwave mutation slope characteristics, as well as the heart rate change characteristics and the heart rate variability change characteristics; The EEG triggering module determines the trigger anchor point based on the EEG frequency band energy change characteristics and the EEG mutation slope characteristics; The hysteresis window module generates a hysteresis tracking window located after the trigger anchor point based on the trigger anchor point and the individual hysteresis baseline; The pulse wave delay matching module performs coherent matching on the heart rate change features and the heart rate variability change features within the hysteresis tracking window to obtain pulse wave delay response features. The feedforward acquisition control module controls the PPG module to increase the sampling rate and the LED drive current within the hysteresis tracking window; The fusion determination module determines the sleep monitoring results based on the trigger anchor point, the pulse wave delay response characteristics, and the individual hysteresis baseline.

2. The sleep monitoring system based on brain-computer interface and PPG module according to claim 1, characterized in that: The dual-track caching module caches the EEG frequency band energy variation characteristics, the EEG abrupt change slope characteristics, the heart rate variation characteristics, and the heart rate variability variation characteristics, respectively, including: Based on the sampling time sequence of the EEG signals, the EEG frequency band energy change characteristics and the EEG abrupt change slope characteristics are written into the EEG short-time buffer window; Based on the sampling time sequence of the pulse wave signal, the heart rate change characteristics and the heart rate variability change characteristics are written into the pulse wavelength time buffer window; Based on the start and end times of the short-term EEG buffer window and the start and end times of the pulse wavelength buffer window, a time index relationship is established between EEG features and pulse wave features. The buffer duration of the pulse wavelength time buffer window is greater than the buffer duration of the EEG short-term buffer window, so that the pulse wavelength time buffer window retains the pulse wave characteristics after the trigger anchor point.

3. A sleep monitoring system based on a brain-computer interface and a PPG module according to claim 2, characterized in that: The EEG triggering module obtains trigger anchor points based on the EEG frequency band energy change characteristics and the EEG abrupt change slope characteristics, including: Read the brainwave frequency band energy change characteristics and brainwave mutation slope characteristics corresponding to the current brainwave analysis window from the brainwave short-term buffer window; The difference between the brainwave frequency band energy change characteristics and the sleep reference frequency band energy is calculated to obtain the brainwave frequency band mutation amount; Based on the variation amplitude of the EEG abrupt change slope characteristics between adjacent EEG analysis windows, the EEG slope abrupt change amount is obtained. When the EEG frequency band mutation amount reaches the frequency band mutation threshold, the EEG slope mutation amount reaches the slope mutation threshold, and the frequency band energy enhancement direction corresponding to the EEG frequency band mutation amount is consistent with the micro-awakening EEG response direction, the time point corresponding to the current EEG analysis window is marked as the trigger anchor point.

4. A sleep monitoring system based on a brain-computer interface and a PPG module according to claim 1, characterized in that: The individual hysteresis baseline in the hysteresis window module is obtained in the following way: Baseline EEG and baseline pulse wave signals were collected during the user's waking and calm phase before sleep monitoring began. The starting point of the EEG response is marked based on the starting position of the frequency band energy change in the baseline EEG signal; Based on the starting position of heart rate change or the starting position of heart rate variability change in the baseline pulse wave signal, mark the starting point of pulse wave response; The time interval between the EEG response start point and the pulse wave response start point is calculated, and the time interval is used as the individual's lag baseline.

5. A sleep monitoring system based on a brain-computer interface and a PPG module according to claim 4, characterized in that: The hysteresis window module generates a hysteresis tracking window located after the trigger anchor point based on the trigger anchor point and the individual hysteresis baseline, including: Read the brainwave frequency band energy change characteristics and brainwave mutation slope characteristics corresponding to the trigger anchor point to generate brainwave trigger intensity level; Based on the EEG trigger intensity level, the corresponding hysteresis compensation coefficient is invoked, and the individual hysteresis baseline is corrected using the hysteresis compensation coefficient to obtain the current hysteresis duration; The hysteresis tracking window is formed by using the trigger anchor point as the starting point of the window and configuring the window ending point according to the current hysteresis duration. The hysteresis tracking window is sent to the pulse wave delay matching module, so that the pulse wave delay matching module can retrieve pulse wave delay response features within the hysteresis tracking window.

6. A sleep monitoring system based on a brain-computer interface and a PPG module according to claim 5, characterized in that: Coherent matching is performed on the heart rate variability features and the heart rate variability features to obtain pulse wave delay response features, including: Heart rate change features and heart rate variability change features that fall within the hysteresis tracking window are extracted from the pulse wavelength buffer window, and multiple pulse wave candidate response segments are formed according to a preset step size. Based on the EEG frequency band energy change characteristics and EEG mutation slope characteristics corresponding to the trigger anchor point, an EEG mutation reference sequence is generated, and the normalized cross-correlation value between the EEG mutation reference sequence and each of the pulse wave candidate response segments is calculated respectively. Based on the normalized cross-correlation value and the delay deviation of each pulse wave candidate response segment relative to the individual hysteresis baseline, a coherent matching score is generated, and the pulse wave candidate response segments whose coherent matching scores meet the matching threshold are used as the pulse wave delay response features.

7. A sleep monitoring system based on a brain-computer interface and a PPG module according to claim 6, characterized in that: The feedforward acquisition control module controls the PPG module to increase the sampling rate and LED drive current within the hysteresis tracking window. Based on the hysteresis tracking window generated by the hysteresis window module, it immediately generates a sampling enhancement control command, carrying the start and end times of the hysteresis tracking window, enabling the PPG module to enter the enhanced acquisition state only within this critical time range, including: After the trigger anchor point is marked, a sampling enhancement control instruction is generated based on the hysteresis tracking window. The sampling enhancement control instruction includes a sampling rate increase instruction and a light-emitting diode drive current increase instruction. The sampling enhancement control command is sent to the PPG module, causing the PPG module to acquire pulse wave signals in an enhanced acquisition state within the hysteresis tracking window. The enhanced acquisition state includes increasing the sampling rate and increasing the LED driving current within a preset safe driving range. After the hysteresis tracking window ends, a data acquisition recovery command is sent to the PPG module to restore the PPG module to normal data acquisition state.

8. A sleep monitoring system based on a brain-computer interface and a PPG module according to claim 6, characterized in that: The fusion determination module determines the sleep monitoring results based on the trigger anchor point, the pulse wave delay response characteristics, and the individual hysteresis baseline, including: Based on the brainwave frequency band energy change characteristics and brainwave mutation slope characteristics corresponding to the trigger anchor points, brainwave trigger weights are generated. Based on the coherent matching score of the pulse wave delayed response features and the delay deviation of the pulse wave delayed response features relative to the individual hysteresis baseline, a pulse wave response weight is generated. Based on the EEG trigger weight and the pulse wave response weight, the micro-awakening determination result, the EEG artifact pending confirmation result, or the single EEG abnormality result are output.

9. A sleep monitoring system based on a brain-computer interface and a PPG module according to claim 8, characterized in that: The process of outputting a micro-awakening determination result, an EEG artifact pending confirmation result, or a single EEG abnormality result based on the EEG trigger weight and the pulse wave response weight includes: When the EEG trigger weight reaches the EEG trigger threshold and the pulse wave response weight reaches the pulse wave response threshold, the micro-awakening determination result is output. When the EEG trigger weight reaches the EEG trigger threshold, the pulse wave response weight does not reach the pulse wave response threshold, and the pulse wave signal quality in the hysteresis tracking window does not meet the effective acquisition conditions, the EEG artifact pending confirmation result is output. When the EEG trigger weight reaches the EEG trigger threshold, the pulse wave response weight does not reach the pulse wave response threshold, and the pulse wave signal quality within the hysteresis tracking window meets the effective acquisition conditions, the single EEG abnormality result is output.

10. A sleep monitoring system based on a brain-computer interface and a PPG module according to claim 9, characterized in that: The fusion determination module also generates a sleep monitoring report based on the sleep monitoring results, including: Based on the trigger anchor point corresponding to the micro-awakening determination result, mark the occurrence time of the micro-awakening event; The degree of sleep fragmentation is generated based on the number of micro-awakening events per unit time and the time interval between adjacent micro-awakening events; The output includes a sleep monitoring report that includes the occurrence time of the micro-awakening events, the degree of sleep fragmentation, and the pulse wave delayed response status.

Images