Systems and methods for insomnia screening and management
By verifying the coupling characteristics of multimodal physiological signals and intervening in biological time, the problems of data interference and individual biological rhythm adaptation in non-laboratory environments were solved, enabling precise screening and management of insomnia.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- SHANGHAI UNIV OF T C M
- Filing Date
- 2026-05-11
- Publication Date
- 2026-07-10
Smart Images
Figure CN122369765A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of digital healthcare and sleep health management technology, specifically to systems and methods for screening and managing insomnia. Background Technology
[0002] Insomnia, a prevalent sleep disorder, seriously affects people's physical and mental health and quality of life. With the development of digital healthcare technology, the focus of sleep health management is gradually shifting from hospital settings that rely on polysomnography monitoring to daily home settings. To achieve long-term tracking and refined intervention of insomnia progression, sleep screening and management systems based on wearable sensors and mobile computing terminals are gradually becoming important technological means in this field.
[0003] Existing sleep management solutions typically rely on smart wearable devices integrating accelerometers and photoelectric sensors to collect users' body movement signals and heart rate data, and estimate sleep stages and total sleep duration using preset algorithms. Meanwhile, various sleep health applications allow users to manually record subjective information such as sleep onset time and number of awakenings in log form. Based on the collected data, existing systems usually follow general clinical guidelines to push sleep hygiene advice, sleep-aiding music, or medication reminders to users at fixed physical times, attempting to improve users' sleep quality through standardized processes.
[0004] While existing technologies can provide basic sleep monitoring and guidance, several shortcomings remain. In open, non-laboratory environments, collected physiological signals are highly susceptible to non-physiological factors. For example, sudden changes in ambient temperature or improper fit can cause drastic fluctuations in skin temperature data. Current technologies lack the logic to cross-verify the inherent coupling mechanisms between different physiological signals, making it difficult to effectively eliminate environmental noise and leading to inaccurate analytical benchmarks. Secondly, existing solutions often treat objective monitoring data and subjective self-reported data in isolation, failing to establish quantitative indicators to assess the degree of deviation between the two. This results in an inability to accurately identify contradictory insomnia characterized by a separation between subjective experience and objective facts, easily leading to misdiagnosis or incorrect treatment guidance. Furthermore, existing intervention strategies are primarily based on sociophysical clock settings, neglecting the individual differences in circadian rhythm phases. For patients with shifted or advanced biological clocks, medication or light interventions based on standard physical times may fall into their physiologically insensitive or even antagonistic zones, thus weakening the actual effectiveness of the intervention. Summary of the Invention
[0005] To address the shortcomings of existing technologies, this invention provides a system and method for screening and managing insomnia, solving the technical problems of existing technologies such as the susceptibility of physiological data to environmental interference and distortion in non-laboratory environments, the lack of quantitative identification capabilities for paradoxical insomnia, and the difficulty in adapting fixed-time intervention strategies to individual circadian rhythm phases.
[0006] To achieve the above objectives, the present invention is implemented through the following technical solution: a system and method for screening and managing insomnia, including a data acquisition module configured to establish a communication connection with an external hardware device, synchronously acquiring subjective self-reported data streams from a mobile terminal, objective physiological data streams from a physiological signal sensor, and environmental monitoring data streams within a preset screening period, and constructing a multi-dimensional time-series data benchmark based on a timestamp alignment mechanism; The data verification module is configured to identify environmental interference based on the coupling characteristics between different physiological signals, and to perform logical consistency verification of the subjective self-reported data stream within a physical time window, outputting a valid data sequence with confidence weights. The feature calculation module is configured to receive the effective data sequence and calculate in parallel the sleep perception deviation index, which represents the degree of difference between subjective feelings and objective facts; the cumulative sleep pressure value, which represents the accumulation of homeostatic sleep driving force in the body; and the biological time, which represents the phase of the user's internal circadian rhythm. The strategy generation module is configured to determine the intervention path type that distinguishes between physiological insomnia and ambivalent insomnia based on the sleep perception deviation index, and to convert the standardized treatment time points into biological time windows that are adapted to individual users based on the biological time, thereby generating a dynamic management plan.
[0007] Preferably, the data acquisition module includes an interaction unit, a sensing unit, and a monitoring unit; The interactive unit is configured to acquire a subjective self-report data stream containing subjective sleep latency, subjective total sleep duration, and subjective number of awakenings; The sensing unit is configured to acquire an objective physiological data stream including body motion signal vectors, photoplethysmography pulse wave signals, and skin temperature signals. The monitoring unit is configured to acquire an environmental monitoring data stream containing information on ambient light intensity and exercise metabolic equivalent. The timestamp alignment mechanism executed by the data acquisition module includes periodically calibrating the external hardware device through the Network Time Protocol to ensure that the subjective self-report data stream, objective physiological data stream, and environmental monitoring data stream remain synchronized on the same time axis.
[0008] Preferably, the data verification module has built-in interference identification logic, which is configured to identify environmental interference and generate a temperature confidence coefficient by calculating the covariance characteristics of the multimodal signal. The specific process of the interference identification logic includes: calculating the absolute value of the derivative of the skin temperature signal with respect to time in real time; and calculating the covariance between the skin temperature change sequence and the instantaneous heart rate change sequence within a preset sliding time window. When the covariance value indicates negative correlation or no correlation, and the amplitude of the body motion signal is detected to exceed the static state threshold, it is determined that there is environmental interference, and the value of the temperature confidence coefficient is reduced. When the covariance value indicates positive correlation coupling, the temperature confidence coefficient is maintained at a high level; The valid data sequence output by the data verification module includes the temperature sequence weighted by the temperature confidence coefficient.
[0009] Preferably, the data verification module has built-in logic gating logic and is configured to execute the following verification rules: Physical contradiction verification rule: Calculate the total time spent in bed between the user's recorded bedtime and wake-up time. If the total time spent in bed is less than the user's recorded subjective total sleep time, then a physical contradiction is determined to exist. Extreme value difference verification rule: Calculate the absolute difference between the subjective sleep latency and the objective sleep latency measured based on objective physiological data stream. If the difference exceeds the preset tolerance threshold, the data is judged to have abnormal deviation. When any of the above rules are triggered, the data verification module will mark the corresponding subjective self-report data stream as invalid, and the data marked as invalid will not participate in the calculation of the sleep perception deviation index.
[0010] Preferably, the feature calculation module is configured to calculate the sleep perception deviation index by calculating the relative error rate and weighted summation; The specific calculation logic is as follows: Calculate the ratio of the absolute value of the difference between subjective sleep latency and objective sleep latency to the objective sleep latency, and multiply it by a preset sleep latency bias weighting coefficient. Calculate the ratio of the difference between objective total sleep duration and subjective total sleep duration to objective total sleep duration, and multiply it by a preset maintenance bias weighting coefficient; Calculate the ratio of subjective awakening times to preset reference awakening times, and multiply it by a preset awakening perception weighting coefficient; The sleep perception deviation index is obtained by adding the above three calculation results together.
[0011] Preferably, the feature calculation module is configured to calculate the cumulative sleep pressure value based on the physiological mechanism of "daytime load accumulation - nighttime compensatory release"; The specific calculation logic is as follows: taking the moment of awakening on the day as the starting point of integration, the real-time metabolic equivalent obtained by acceleration signal conversion and the ambient light intensity after logarithmic transformation are integrated and accumulated, and the total amount of pressure released during the daytime nap is subtracted. The total pressure is calculated based on the product of the duration of the nap and a sleep depth weighting coefficient determined by the magnitude of the heart rate drop.
[0012] Preferably, the feature calculation module is configured to calculate the biological time using multimodal fusion localization technology; The specific calculation logic is as follows: The objective physiological data stream during the nighttime period was processed to extract the time of lowest body temperature and the time of lowest resting heart rate; Using the temperature confidence coefficient generated by the data verification module as a weight, the lowest body temperature time and the lowest resting heart rate time after phase delay compensation are weighted and fused to determine the biological clock phase reference point. Based on the difference between the biological clock phase reference point and the reference phase time of a standard healthy population, a mapping relationship between physical time and biological time is established.
[0013] Preferably, the strategy generation module is configured to perform path decision-making based on the numerical range of the sleep perception deviation index: When the sleep perception deviation index is greater than the preset high deviation threshold, it is determined to be ambivalent insomnia, the cognitive reconstruction priority path is activated, an instruction sequence containing sleep restriction therapy and cognitive reconstruction training is generated, and the priority of drug-assisted sleep recommendations is reduced. When the sleep perception deviation index is less than or equal to the preset low deviation threshold, it is determined to be physiological insomnia, the physical therapy priority path is activated, a drug treatment plan or physical therapy plan is generated, and the cognitive behavioral therapy module is locked.
[0014] Preferably, the strategy generation module is further configured to perform feedforward control and time-domain remapping functions; The time-domain remapping function includes: receiving the biological clock phase reference point corresponding to the biological time, and shifting the suggested time in the standard treatment plan into the actual physical execution time based on the difference between the reference point and the standard reference phase; The feedforward control function includes: during the daytime period of the screening cycle, comparing the current cumulative sleep pressure value with a preset minimum pressure threshold; if the current cumulative sleep pressure value is less than the minimum pressure threshold, calculating the additional exercise load required for compensation, and converting it into feedback instructions for adjusting the exercise load based on exercise metabolic data.
[0015] A second aspect of the present invention provides a method for screening and managing insomnia, using the system for screening and managing insomnia as described in any of the preceding claims, the method comprising: The data acquisition module simultaneously collects subjective self-report data streams, objective physiological data streams, and environmental monitoring data streams within a preset screening cycle, and constructs a multi-dimensional time-series data benchmark. The data verification module identifies environmental interference based on the coupling characteristics between physiological signals and performs logical consistency verification on subjective input data to generate a valid data sequence with confidence weights. The feature calculation module calculates the sleep perception deviation index, cumulative sleep pressure value, and individual biological time in parallel based on the effective data sequence. The strategy generation module determines the intervention path type based on the sleep perception deviation index and performs time-domain remapping of the intervention measures based on the biological time to generate a dynamic management plan that includes daytime load regulation and nighttime sleep management.
[0016] This invention provides a system and method for screening and managing insomnia. It offers the following advantages: 1. This invention improves the confidence level of monitoring data in non-laboratory environments by constructing a data verification mechanism based on the coupling characteristics of multimodal physiological signals. The system utilizes the physiological correlation between skin temperature and heart rate changes, identifies non-physiological signal fluctuations caused by sudden changes in ambient temperature or abnormal wear by calculating covariance features, and introduces a temperature confidence coefficient to weight the raw data. This mechanism effectively solves the technical problem of data distortion caused by external noise interference in open environments when using a single sensor, providing a high-fidelity data foundation for subsequent feature decomposition.
[0017] 2. This invention achieves quantitative differentiation and precise triage of insomnia subtypes by calculating the Sleep Perception Bias Index. This index comprehensively calculates the difference rate between subjective perception and objective monitoring of sleep latency, sleep duration, and number of awakenings. It can accurately identify ambivalent insomnia patients with a significant tendency for subjective exaggeration and physiological insomnia patients with objective sleep impairment. Based on this index, the system automatically matches intervention paths prioritizing cognitive restructuring or physical therapy, avoiding misdiagnosis caused by neglecting subjective bias in traditional screening, thereby reducing the risk of ineffective treatment or drug abuse.
[0018] 3. This invention establishes a dynamic intervention timing mapping mechanism based on individual biological time, improving the physiological response efficiency of intervention measures. By integrating body temperature and heart rate characteristics to locate the biological clock phase reference point, the system converts the general standard physical therapy time into a biological time window adapted to the user's internal circadian rhythm, ensuring that light exposure, drug administration, or exercise intervention falls within the body's sensitive period. Combined with feedforward control logic based on cumulative sleep pressure values, the system can guide daytime activity load in reverse according to nighttime sleep needs, realizing a closed-loop management throughout the entire cycle from daytime behavior regulation to nighttime sleep improvement. Attached Figure Description
[0019] Figure 1 This is a schematic diagram of the overall system architecture of the present invention; Figure 2This is a schematic diagram of the system workflow of the present invention; Figure 3 This is a schematic diagram illustrating the specific structure and data interaction of the data acquisition module of the present invention; Figure 4 This is a schematic diagram of the logical processing flow of the data verification module of the present invention; Figure 5 This is a diagram showing the internal logic and data flow of the feature calculation module of the present invention.
[0020] Among them, 10 is the data acquisition module; 20 is the data verification module; 30 is the feature calculation module; and 40 is the strategy generation module. Detailed Implementation
[0021] The present invention will now be described in detail with reference to the accompanying drawings and specific embodiments. It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other.
[0022] See attached document Figure 1 , Figure 1 This is a schematic diagram of the overall architecture of a system for insomnia screening and management according to an embodiment of the present invention. The system includes: a data acquisition module 10, a data verification module 20, a feature calculation module 30, and a strategy generation module 40. The data acquisition module 10, the data verification module 20, the feature calculation module 30, and the strategy generation module 40 interact with each other through an internal data bus or a standard application programming interface, forming a closed-loop control logic from data perception, quality control, feature quantification to intervention decision-making.
[0023] The data acquisition module 10, serving as the system's front-end input interface, is configured to establish a communication connection with external hardware devices and construct a multi-dimensional time-series data benchmark within a preset screening period. In this embodiment, the screening period is selected as 24-72 hours. The external hardware devices include physiological signal sensors for wearing on the user's wrist or other parts of the body, and mobile terminal devices for providing a human-computer interaction interface.
[0024] The data acquisition module 10 performs the task of synchronously acquiring multi-source heterogeneous data. The acquired data stream covers three dimensions: subjective self-reported data stream from mobile terminal, objective physiological data stream from physiological signal sensor, and environmental monitoring data stream.
[0025] The subjective self-report data stream includes the time points and status evaluation information input by the user; The objective physiological data stream includes continuously sampled body movement signals, pulse wave signals, and body temperature signals; The environmental monitoring data stream includes information on ambient light intensity and exercise metabolic equivalent.
[0026] The data acquisition module 10 uses a timestamp alignment mechanism to ensure that the subjective self-report data stream, objective physiological data stream, and environmental monitoring data stream are strictly synchronized on the same time axis, providing a temporal basis for subsequent correlation analysis.
[0027] The data verification module 20 is connected to the output of the data acquisition module 10 and is used to clean and assess the confidence level of the raw data before feature calculation. To address the vulnerability of data acquired in non-laboratory environments to interference, the data verification module 20 incorporates environmental interference identification logic and subjective data gating logic.
[0028] The data verification module 20 is configured to identify non-physiological signal fluctuations caused by abnormal wearing conditions, sudden changes in ambient temperature, or device detachment, based on the coupling characteristics between different physiological signals, such as the physiological correlation between changes in body temperature and heart rate. Simultaneously, the data verification module 20 performs logical consistency verification on subjective input data within a physical time window, eliminating invalid inputs that violate physical principles or physiological limits.
[0029] After processing, the data verification module 20 outputs a valid data sequence with confidence weights and invalid data removal markers to ensure that the subsequent calculation module only performs calculations based on high-confidence data, thereby improving the robustness of the system evaluation.
[0030] The feature calculation module 30 is connected to the data verification module 20 and receives the verified valid data sequence. The feature calculation module 30 is the core computing unit of the system, responsible for converting discrete signal data into feature indicators with clinical screening and management significance.
[0031] The feature calculation module 30 integrates multiple physiological model algorithms to perform parallel calculations on three key dimensions describing the characteristics of insomnia: the sleep perception bias index, which represents the degree of difference between subjective feelings and objective facts; the cumulative sleep pressure value, which represents the accumulation of homeostatic sleep driving forces in the body; and the biological time, which represents the phase of the user's internal circadian rhythm.
[0032] The feature calculation module 30 locates the biological clock phase point by fusing body temperature and heart rate features, and establishes a mapping relationship between physical clock time and individual biological time, thereby solving the technical problem that general physical time cannot accurately correspond to the individual physiological rhythm window.
[0033] The strategy generation module 40 is connected to the feature calculation module 30, and generates structured control instructions and management plans based on the calculated feature indicators. The strategy generation module 40 determines the type of intervention path based on the sleep perception deviation index, distinguishes between physiological insomnia and ambivalent insomnia, and adjusts the weights of cognitive behavioral therapy and physical therapy according to the type of intervention path.
[0034] Meanwhile, the strategy generation module 40 performs time-domain remapping of all drug administration, physical therapy, and behavioral recommendations based on biological time, converting standardized treatment time points into biological time windows adapted to individual users to ensure that intervention measures fall within the sensitive period of physiological response.
[0035] In addition, the strategy generation module 40 also has a feedforward control function, which can generate feedback instructions for adjusting exercise load during the daytime period of the screening cycle based on the comparison results of the accumulated sleep pressure value and the preset threshold. These instructions are then pushed to the user via the mobile terminal, forming a complete closed loop from nighttime monitoring feedback to guide daytime behavior, thereby improving the sleep quality of the next night.
[0036] See attached document Figure 2 , Figure 2 This is a schematic diagram of the system workflow according to an embodiment of the present invention.
[0037] The system's operation begins with the activation of the data acquisition module 10, which continuously records multidimensional data during the screening period. The acquired data is sent to the data verification module 20 in real time, which calculates the covariance characteristics and logical boundaries of the signal, removes low-quality data, and marks the confidence level.
[0038] Subsequently, based on valid data, the feature calculation module 30 updates the user's deviation index and rhythm phase parameters daily, and calculates sleep pressure in real time by integrating the data.
[0039] At the end of the screening cycle or when a preset examination time is triggered, the strategy generation module 40 calls the deviation index, rhythm phase parameters, and sleep pressure calculation results to generate classification judgments for insomnia subtypes, schedule adjustment instructions for biological clock phases, and compensation suggestions for daytime workload, which are finally output to the user terminal or medical management platform. This enables continuous tracking and dynamic intervention of the user's sleep health status.
[0040] See attached document Figure 3 , Figure 3 The diagram illustrates the specific structure and data interaction of a data acquisition module according to an embodiment of the present invention. As the sensing front end of the system, the data acquisition module is configured to construct a multi-dimensional time-series data benchmark. The physical implementation of the data acquisition module relies on a mobile terminal application and sensor hardware worn by the user. Logically, the data acquisition module includes an interaction unit, a sensing unit, and a monitoring unit, which operate in parallel within a preset screening period.
[0041] The interaction unit, implemented using the display and input interfaces of a mobile terminal, is used to acquire sleep assessment parameters that reflect the patient's subjective perception. The interaction unit generates a structured data entry interface within a preset daily time window, collecting subjective parameters including subjective sleep latency, subjective total sleep duration, and subjective awakening count. Here, subjective sleep latency is defined as the time difference between the user's recorded bedtime and the user's subjectively perceived sleep time; subjective total sleep duration is defined as the user's subjectively perceived actual sleep duration; and subjective awakening count is defined as the number of times the user remembers regaining consciousness after falling asleep until finally waking up. The interaction unit is configured with input logic validation rules. When a negative duration value or a value exceeding a preset physiological limit threshold is detected, a prompt is triggered, requiring re-entry to ensure data validity. In this embodiment, the physiological limit threshold is selected as 0 to 24 hours.
[0042] The sensing unit is based on a microelectromechanical system (MEMS) sensor integrated into the wearable device, used to continuously acquire objective physiological signals at a fixed sampling rate. The data stream acquired by the sensing unit includes body motion signal vectors, photoplethysmography (PPG) pulse wave signals, and skin temperature signals.
[0043] Specifically, the body motion signal vector is acquired by a three-axis or six-axis accelerometer. This vector contains acceleration components of the device in three orthogonal directions within a Cartesian coordinate system, used for subsequent analysis of changes in body position and activity intensity. The acquisition of accelerometer data and the decomposition of coordinate axes are well-known techniques to those skilled in the art, and can be achieved by reading values from the sensor registers.
[0044] The photoplethysmography (PPG) signal is obtained by a photoelectric sensor by emitting a light beam of a specific wavelength and receiving the intensity of light reflected from the skin. Changes in the PPG signal waveform reflect changes in blood volume in the microvascular bed. The sensing unit is equipped with analog front-end circuitry to perform analog-to-digital conversion and preliminary bandpass filtering on the raw optical signal, outputting a digitized timing signal. This signal contains cardiac cycle information, used for subsequent extraction of instantaneous heart rate and heart rate variability features.
[0045] Skin temperature signals are acquired by a contact temperature sensor, with the sensor probe in close contact with the user's wrist skin to record the Celsius temperature values over time. Because wrist temperature is significantly affected by the environment, the raw temperature sequence output by the sensing unit needs to be combined with other modal data for confidence verification during subsequent processing.
[0046] The monitoring unit is used to quantify the load of external environmental factors and daytime activity. The monitoring unit includes an ambient light sensor and a motion metabolism algorithm component. The ambient light sensor collects ambient light intensity sequences to assess the impact of light on the timing of the biological clock. The motion metabolism algorithm component calculates real-time metabolic equivalents based on acceleration data. This component calculates the signal vector magnitude of the acceleration signal, i.e., the square root of the sum of the squares of the acceleration components along each axis, and substitutes this vector magnitude into a preset regression equation to convert it into a metabolic equivalent value, thereby quantifying the intensity of daytime physical activity. The specific hardware selection and basic driving methods of the light sensor and acceleration sensor are well-known to those skilled in the art and will not be elaborated upon here.
[0047] The data acquisition module encapsulates the data acquired by the interaction unit, sensing unit, and monitoring unit into data packets with absolute timestamps. All time-series data undergoes time synchronization before transmission, and is periodically calibrated with the standard network time protocol of the mobile terminal or cloud server to ensure that the data acquired by different sensors are strictly aligned on the timeline, providing a unified time-series reference for subsequent multimodal data verification.
[0048] See attached document Figure 4 , Figure 4 A schematic diagram of the logical processing flow of a data verification module according to an embodiment of the present invention is shown. As a pre-processing step for feature decomposition, the data verification module is configured to receive the raw data stream from the data acquisition module and perform cleaning and logical verification for environmental interference and subjective input anomalies. The data verification module internally includes an interference identification unit and a logic gating unit, which process objective physiological data and subjective self-reported data, respectively.
[0049] The interference identification unit primarily addresses the problem of sensor data distortion caused by changes in the wearing environment, particularly considering that wrist skin temperature signals are easily affected by non-physiological factors. During normal sleep physiological regulation, a decrease in core body temperature is typically accompanied by a steady decrease in heart rate and reduced body movement, showing a positive correlation. When limbs are exposed to cold environments or heat sources, skin temperature fluctuates drastically and inconsistently with heart rate changes, showing a negative or no correlation. The interference identification unit utilizes this coupling characteristic between multidimensional physiological signals, calculating the covariance characteristics of multimodal signals to identify environmental interference and generate corresponding confidence weights.
[0050] The specific confidence level calculation logic is as follows: The interference identification unit calculates the temperature confidence coefficient in real time, which reflects the reliability of the skin temperature signal at the current moment. The calculation formula is as follows: ; In the formula, For a moment Temperature confidence coefficient; It represents the absolute value of the derivative of the skin temperature signal with respect to time, which is obtained in a discrete sampling system by calculating the absolute value of the first-order difference between adjacent sampling points; This is a preset sensitivity adjustment coefficient used to control the steepness of the Sigmoid function. In this embodiment, a value of 0.8-1.2 is selected; The rate of change is a preset threshold, and the rate of change threshold In this embodiment, a temperature range of 0.01-0.03℃ / s is selected. Decision factor, decision factor The value of depends on the covariance of the skin temperature signal and the instantaneous heart rate signal, as well as the amplitude of the body motion signal vector.
[0051] For the determination factor The specific determination logic is as follows: the system calculates the covariance value of the temperature change sequence and the heart rate change sequence within a preset sliding time window, and the length of the sliding time window is selected as 200-400 seconds in this embodiment. When the covariance value is less than zero, it indicates a negative correlation, that is, decoupling of physiological signals; or when its absolute value is less than a preset lower threshold of correlation, it indicates no correlation. Furthermore, if the amplitude of the body movement signal exceeds the resting state threshold, the determination factor is... When the value is 1, the calculated result of the exponent term in the formula increases, leading to an increase in the denominator and a significant decrease in the temperature confidence coefficient; otherwise, When the value is 0, the temperature confidence coefficient remains high, while the correlation lower limit threshold is selected as 0.2-0.4 in this embodiment, and the static state threshold is selected as 0.01-0.03g in this embodiment.
[0052] The confidence calculation logic executed by the interference identification unit achieves the following signal processing effect: when a significant and drastic temperature change is detected, i.e., whether it is an increase or decrease in temperature, and this change does not conform to the physiological coupling law of heart rate increase due to temperature decrease, the calculated temperature confidence coefficient approaches zero, indicating that the current temperature data is mainly dominated by environmental factors; conversely, if the temperature change is gradual or synchronized with the heart rate change, the coefficient remains high. The interference identification unit finally outputs a confidence-weighted effective temperature sequence, which is the product of the original temperature value and the confidence coefficient, and will be used for subsequent phase calculation.
[0053] The logic gating unit is used to perform boundary checks on the physical consistency and physiological rationality of subjective data input by users, in order to eliminate invalid data caused by misoperation or arbitrary filling. The logic gating unit has a pre-built verification rule base to verify subjective sleep latency, subjective total sleep duration, and related time nodes.
[0054] The verification rules executed by the logic gating unit include: Physical contradiction verification rule: Calculate the total time spent in bed between the time the user enters to go to bed and the time the user enters to wake up. If the time spent in bed is less than the total subjective sleep time entered by the user, it is determined that there is a physical logical contradiction.
[0055] Extreme value difference verification rule: Calculate the absolute difference between the subjective sleep latency and the objective sleep latency measured by the sensing unit. If the difference exceeds a preset tolerance threshold, the data is determined to have an abnormal deviation, and the tolerance threshold is selected as 20-40 minutes in this embodiment.
[0056] When either the physical contradiction verification rule or the extreme value difference verification rule is triggered, the logic gating unit marks the subjective data for the corresponding date as invalid. Subjective data marked as invalid will not participate in the subsequent calculation of the average deviation index, thus preventing extreme outliers from interfering with the overall evaluation results.
[0057] See attached document Figure 5 , Figure 5 The diagram illustrates the internal logic and data flow of a feature calculation module according to an embodiment of the present invention. As the core computing unit of the system, the feature calculation module is configured to receive valid data sequences from the data verification module and calculate key physiological indicators used to guide clinical screening and intervention decisions. The feature calculation module internally operates three sub-units in parallel: a deviation calculation unit, a stress prediction unit, and a phase mapping unit.
[0058] The deviation calculation unit is used to quantify the degree of separation between the patient's subjective sleep perception and objective monitoring facts, thereby providing a quantitative basis for distinguishing between ambivalent insomnia and physiological insomnia. This unit first calls a preset sleep staging algorithm to process the objective physiological signals and extract the objective sleep latency and objective total sleep duration. For the specific implementation of the sleep staging algorithm, those skilled in the art can use a body movement determination algorithm based on the Cole-Kripke model or a staging algorithm based on heart rate variability. Sleep staging algorithms are well-known technologies in this field and will not be elaborated upon here.
[0059] The deviation calculation unit calculates the sleep perception deviation index based on extracted objective indicators and subjective indicators obtained by the interaction unit. This index is a composite quantitative value that comprehensively reflects the perceptual deviation between the sleep onset stage and the sleep maintenance stage. The calculation formula is as follows: ; In the formula, The sleep perception bias index; and These represent subjective sleep latency and objective sleep latency, respectively, with the numerator taken as the absolute value to capture bidirectional bias. and These represent the subjective total sleep duration and the objective total sleep duration, respectively. The number of times of subjective awakening; The preset reference number of awakenings is constant, which is selected as 1-2 times in this embodiment; The weighting coefficient for sleep onset deviation is selected as 0.4-0.6 in this embodiment; To maintain the bias weighting coefficient, a value of 0.3-0.5 is selected in this embodiment; The arousal perception weighting coefficient is selected as 0.1-0.2 in this embodiment; To prevent extremely small positive numbers with a denominator of zero, a range of 0.0005-0.005 is selected in this embodiment. The sleep perception deviation index calculation formula, by calculating the relative error rate and weighted summation, can eliminate the dimensional influence caused by differences in baseline sleep duration among different patients and output a standardized deviation score.
[0060] The stress prediction unit is used to construct a sleep stress model based on the conservation of bioenergy, and to estimate the accumulation of homeostatic sleep drive forces in real time. This model is based on the physiological mechanism of "daytime load accumulation - nighttime compensatory release", which converts external environmental load into internal physiological stress values.
[0061] The stress prediction unit updates the accumulated sleep stress value in real time through integral calculation. The calculation formula is as follows: ; In the formula, For a moment The cumulative sleep stress value; The system records the awakening time for the current day, with the score range from the awakening time to the current time. ; For a moment The real-time metabolic equivalent is used to characterize the rate of adenosine accumulation produced by physical activity; For a moment The ambient light intensity was simulated using a logarithmic function to demonstrate the nonlinear stimuli effect of light on the arousal system. and These are the motion load conversion coefficient and the light load conversion coefficient, respectively. In this embodiment... Select 0.4-0.6, In this embodiment, a value of 1.5-2.5 is selected; The pressure relief efficiency coefficient is selected as 3.0-4.0 in this embodiment. For all of the day The sum of the total pressure released during each rest period, and the pressure release during a single rest period. It is equal to the product of the duration of the nap and the corresponding sleep depth weighting coefficient, which is determined by the average heart rate decrease during the nap.
[0062] The phase mapping unit addresses the issue that physical clock time cannot accurately reflect the phase of an individual's biological clock, establishing a personalized biological time reference. This unit employs multimodal fusion positioning technology, utilizing the phase correlation between body temperature rhythm and heart rate rhythm to estimate the biological clock reference point.
[0063] The phase mapping unit first performs cosine curve fitting or low-pass filtering on the nighttime data stream to eliminate high-frequency noise interference, and then extracts the times of lowest body temperature and lowest resting heart rate. Subsequently, using the temperature confidence coefficient output by the aforementioned data verification module, the two feature times are weighted and fused to calculate the biological clock phase reference point, as shown in the following formula: ; In the formula, This serves as a reference point for the biological clock phase, corresponding to the phase of the lowest body temperature. The moment of lowest body temperature; This is the moment of lowest resting heart rate; The average value of the temperature confidence coefficient within the nighttime window is used to dynamically adjust the confidence weight: when the temperature data quality is high, the algorithm tends to rely on the body temperature phase; when the temperature is severely disturbed, the algorithm automatically tends to rely on the phase-compensated heart rate phase. The physiological fixed delay constant for the heart rate phase relative to the body temperature phase is selected as 60-120 minutes in this embodiment.
[0064] After determining the reference point, the phase mapping unit establishes a mapping relationship between physical time and biological time: ; In the formula, Biological time; Physical time; In this embodiment, 04:00-05:00 is selected as the reference phase time for a standard healthy population. Through this mapping, if the user's phase... Later than the reference phase When a user's time is 06:00, their biological time will be 2 hours ahead of their physical time, thus accurately reflecting their physiological time difference.
[0065] The strategy generation module 40 is connected to the feature calculation module 30 and is responsible for transforming abstract physiological features into specific clinical intervention instructions. The strategy generation module is internally configured with a path decision subunit, a time-domain mapping subunit, and a feedforward feedback subunit, which correspond to treatment path selection, intervention timing calibration, and daytime load adjustment, respectively.
[0066] The path decision-making subunit dynamically selects an appropriate insomnia management path based on the numerical range of the sleep perception deviation index. This subunit pre-stores high-level deviation thresholds and low-level deviation thresholds, and classifies intervention strategies into three categories based on these thresholds: cognitive restructuring priority, physical therapy priority, and mixed intervention.
[0067] The specific execution logic of path decision is as follows: When the sleep perception bias index calculated by the system exceeds the high-level bias threshold, the patient is deemed to have a significant tendency to exaggerate subjectively, i.e., exhibiting obvious characteristics of ambivalent insomnia. In this embodiment, the high-level bias threshold is selected as 0.4-0.6. At this time, the system activates the cognitive restructuring priority path. Under this path, the system calls a pre-set cognitive correction task library to generate a sequence of instructions containing sleep restriction therapy and cognitive restructuring training. The system also automatically masks or reduces the display priority of medication-assisted sleep suggestions in the user interface to prevent the patient from developing a psychological dependence on medication.
[0068] When the sleep perception deviation index is less than or equal to the low deviation threshold, the patient's subjective feelings are considered to be basically consistent with objective physiological facts, which is typical physiological insomnia. In this embodiment, the low deviation threshold is selected as 0.1-0.3. At this time, the system activates the physical therapy priority path. Under this path, the system prioritizes recommending drug treatment or physical therapy plans and locks the cognitive behavioral therapy module to avoid unnecessary cognitive intervention that may increase patient anxiety.
[0069] For other numerical ranges, the system activates a hybrid intervention pathway, providing cognitive and physical therapy recommendations in parallel.
[0070] The time-domain mapping subunit is used to address the mismatch between fixed-time treatment plans and individual patient circadian rhythms. This subunit receives a circadian rhythm phase reference point from the phase mapping unit and converts the recommended time in the standard treatment plan, based on the standard circadian rhythm (in this embodiment, the standard population should receive light exposure at 07:00), into a personalized physical execution time adapted to the user.
[0071] The calculation logic for time-domain mapping is as follows: The system has a pre-installed standard treatment protocol library, which contains standard recommended timing for each intervention procedure. When the standard morning light time is set to two hours after the reference phase, if the reference phase is 04:00, then The time is 06:00. The time-domain mapping subunit calculates the actual physical execution time based on the current biological clock phase reference point, using the following formula: ; In the formula, This refers to the actual physical execution time. This is the recommended timing in the standard treatment plan; This is the reference point for the biological clock phase obtained from the current calculation; This is the reference phase time for a standard healthy population. Using this formula, if the user's phase lags (... ), then the execution time This will be postponed accordingly, thereby ensuring that light therapy, exercise intervention, or drug administration falls within the sensitive window of the patient's biological rhythm, avoiding treatment resistance caused by phase delay or advance.
[0072] The feedforward feedback subunit constructs a closed-loop control system between daytime activities and nighttime sleep. This subunit operates periodically during the daytime hours of the screening cycle, performing feedforward prediction and compensation control based on cumulative sleep pressure values.
[0073] The specific execution logic of feedforward feedback is as follows: The system reads the current cumulative sleep pressure value at preset daytime checkpoints and compares it with the minimum pressure threshold required to ensure deep sleep at night.
[0074] If the current cumulative sleep pressure value is less than the minimum pressure threshold, it indicates that the current daytime workload is insufficient to generate enough nighttime sleep drive. In this embodiment, the minimum pressure threshold is selected as 40-60 units. At this time, the feedforward feedback subunit calculates the required additional exercise load compensation using the following formula: ; In the formula, This is for additional exercise load; The minimum pressure threshold; This represents the cumulative sleep stress value at the current moment. The exercise load conversion factor is selected as 1.5-2.5 in this embodiment. After calculating the additional exercise load, the subunit retrieves the exercise metabolism database and converts the load into specific exercise instructions. The conversion is based on the following: the instruction duration equals the additional exercise load divided by the standard metabolic equivalent of the corresponding exercise type. The feedforward feedback control mechanism realizes the leap from simple monitoring to proactive health management, actively intervening in nighttime sleep quality by adjusting daytime behavior.
[0075] The methods for screening and managing insomnia described below correspond to the systems for screening and managing insomnia described above.
[0076] This invention also provides a method for screening and managing insomnia, the method comprising: The data acquisition module simultaneously collects subjective self-report data streams, objective physiological data streams, and environmental monitoring data streams within a preset screening cycle, and constructs a multi-dimensional time-series data benchmark. The data verification module identifies environmental interference based on the coupling characteristics between physiological signals and performs logical consistency verification on subjective input data to generate a valid data sequence with confidence weights. The feature calculation module calculates the sleep perception deviation index, cumulative sleep pressure value, and individual biological time in parallel based on the effective data sequence. The strategy generation module determines the intervention path type based on the sleep perception deviation index and performs time-domain remapping of the intervention execution points based on the biological time to generate a dynamic management plan that includes daytime load regulation and nighttime sleep management.
[0077] The method in this embodiment can be used to execute the above system embodiment, and its principle and technical effect are similar, so it will not be described again here.
[0078] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.
Claims
1. A system for screening and managing insomnia, characterized in that, include: The data acquisition module is configured to establish a communication connection with external hardware devices and simultaneously collect subjective self-reported data streams from mobile terminals, objective physiological data streams from physiological signal sensors, and environmental monitoring data streams within a preset screening cycle. It also constructs a multi-dimensional time-series data benchmark based on a timestamp alignment mechanism. The data verification module is configured to identify environmental interference based on the coupling characteristics between different physiological signals, and to perform logical consistency verification of the subjective self-reported data stream within a physical time window, outputting a valid data sequence with confidence weights. The feature calculation module is configured to receive the effective data sequence and calculate in parallel the sleep perception deviation index, which represents the degree of difference between subjective feelings and objective facts; the cumulative sleep pressure value, which represents the accumulation of homeostatic sleep driving force in the body; and the biological time, which represents the phase of the user's internal circadian rhythm. The strategy generation module is configured to determine the type of intervention path that is adapted to different data feature patterns based on the sleep perception deviation index, and to convert the standardized intervention reference time point into a biological execution time window adapted to the individual user based on the biological time, thereby generating a dynamic management plan.
2. The system for insomnia screening and management according to claim 1, characterized in that, The data acquisition module includes an interaction unit, a sensing unit, and a monitoring unit; The interactive unit is configured to acquire a subjective self-report data stream containing subjective sleep latency, subjective total sleep duration, and subjective number of awakenings; The sensing unit is configured to acquire an objective physiological data stream including body motion signal vectors, photoplethysmography pulse wave signals, and skin temperature signals. The monitoring unit is configured to acquire an environmental monitoring data stream containing information on ambient light intensity and exercise metabolic equivalent. The timestamp alignment mechanism executed by the data acquisition module includes periodically calibrating the external hardware device through the Network Time Protocol to ensure that the subjective self-report data stream, objective physiological data stream, and environmental monitoring data stream remain synchronized on the same time axis.
3. The system for insomnia screening and management according to claim 1, characterized in that, The data verification module has built-in interference identification logic, which is configured to identify environmental interference and generate temperature confidence coefficients by calculating the covariance characteristics of multimodal signals. The specific process of the interference identification logic includes: calculating the absolute value of the derivative of the skin temperature signal with respect to time in real time; and calculating the covariance between the skin temperature change sequence and the instantaneous heart rate change sequence within a preset sliding time window. When the covariance value indicates negative correlation or no correlation, and the amplitude of the body motion signal is detected to exceed the static state threshold, it is determined that there is environmental interference, and the value of the temperature confidence coefficient is reduced. When the covariance value indicates positive correlation coupling, the temperature confidence coefficient is maintained at a high level; The valid data sequence output by the data verification module includes the temperature sequence weighted by the temperature confidence coefficient.
4. The system for insomnia screening and management according to claim 1, characterized in that, The data verification module has built-in logic gating logic and is configured to execute the following verification rules: Physical contradiction verification rule: Calculate the total time spent in bed between the user's recorded bedtime and wake-up time. If the total time spent in bed is less than the user's recorded subjective total sleep time, then a physical contradiction is determined to exist. Extreme value difference verification rule: Calculate the absolute difference between the subjective sleep latency and the objective sleep latency measured based on objective physiological data stream. If the difference exceeds the preset tolerance threshold, the data is judged to have abnormal deviation. When any of the above rules are triggered, the data verification module will mark the corresponding subjective self-report data stream as invalid, and the data marked as invalid will not participate in the calculation of the sleep perception deviation index.
5. The system for insomnia screening and management according to claim 1, characterized in that, The feature calculation module is configured to calculate the sleep perception deviation index by calculating the relative error rate and weighted summation. The specific calculation logic is as follows: Calculate the ratio of the absolute value of the difference between subjective sleep latency and objective sleep latency to the objective sleep latency, and multiply it by a preset sleep latency bias weighting coefficient. Calculate the ratio of the difference between objective total sleep duration and subjective total sleep duration to objective total sleep duration, and multiply it by a preset maintenance bias weighting coefficient; Calculate the ratio of subjective awakening times to preset reference awakening times, and multiply it by a preset awakening perception weighting coefficient; The sleep perception deviation index is obtained by adding the above three calculation results together.
6. The system for insomnia screening and management according to claim 1, characterized in that, The feature calculation module is configured to calculate the cumulative sleep pressure value based on the physiological mechanism of "daytime load accumulation - nighttime compensatory release"; The specific calculation logic is as follows: taking the moment of awakening on the day as the starting point of integration, the real-time metabolic equivalent obtained by acceleration signal conversion and the ambient light intensity after logarithmic transformation are integrated and accumulated, and the total amount of pressure released during the daytime nap is subtracted. The total pressure is calculated based on the product of the duration of the nap and a sleep depth weighting coefficient determined by the magnitude of the heart rate drop.
7. The system for insomnia screening and management according to claim 3, characterized in that, The feature calculation module is configured to calculate the biological time using multimodal fusion localization technology; The specific calculation logic is as follows: The objective physiological data stream during the nighttime period was processed to extract the time of lowest body temperature and the time of lowest resting heart rate; Using the temperature confidence coefficient generated by the data verification module as a weight, the lowest body temperature time and the lowest resting heart rate time after phase delay compensation are weighted and fused to determine the biological clock phase reference point. Based on the difference between the biological clock phase reference point and the reference phase time of a standard healthy population, a mapping relationship between physical time and biological time is established.
8. The system for insomnia screening and management according to claim 1, characterized in that, The strategy generation module is configured to perform path decisions based on the numerical range of the sleep perception deviation index: When the sleep perception deviation index is greater than the preset high deviation threshold, it is identified as a perception deviation-dominated state, the cognitive reconstruction priority path is activated, an instruction sequence containing sleep restriction execution logic and cognitive reconstruction training content is generated, and the priority of external auxiliary intervention suggestions is reduced. When the sleep perception deviation index is less than or equal to the preset low deviation threshold, it is identified as a physiologically driven dominant state, the physical regulation priority path is activated, a non-cognitive auxiliary regulation scheme or physical relaxation scheme is generated, and the cognitive behavior regulation module is locked.
9. The system for insomnia screening and management according to claim 1, characterized in that, The strategy generation module is also configured to perform feedforward control and time-domain remapping functions; The time-domain remapping function includes: receiving the biological clock phase reference point corresponding to the biological time, and shifting the suggested time in the standard treatment plan into the actual physical execution time based on the difference between the reference point and the standard reference phase; The feedforward control function includes: during the daytime period of the screening cycle, comparing the current cumulative sleep pressure value with a preset minimum pressure threshold; if the current cumulative sleep pressure value is less than the minimum pressure threshold, calculating the additional exercise load required for compensation, and converting it into feedback instructions for adjusting the exercise load based on exercise metabolic data.
10. A method for screening and managing insomnia, characterized in that, Using the system for insomnia screening and management according to any one of claims 1-9, the method comprises: The data acquisition module simultaneously collects subjective self-report data streams, objective physiological data streams, and environmental monitoring data streams within a preset screening cycle, and constructs a multi-dimensional time-series data benchmark. The data verification module identifies environmental interference based on the coupling characteristics between physiological signals and performs logical consistency verification on subjective input data to generate a valid data sequence with confidence weights. The feature calculation module calculates the sleep perception deviation index, cumulative sleep pressure value, and individual biological time in parallel based on the effective data sequence. The strategy generation module determines the intervention path type based on the sleep perception deviation index and performs time-domain remapping of the intervention execution points based on the biological time to generate a dynamic management plan that includes daytime load regulation and nighttime sleep management.