Migraine risk indicator construction system based on evidence-based

Abstract

The invention relates to the related technical field of risk indicator construction management, in particular to an evidence-based migraine risk indicator construction system, which comprises a feature extraction module for acquiring physiological data and coupling abnormal features, an error correction module for mapping features and correcting errors, a dynamic adjustment module for adjusting a probability confidence interval, and an evidence-based migraine risk indicator management module for establishing an evidence-based migraine risk indicator. The technical problems that generation of the risk report is fixed to a preset migraine risk index, the risk report cannot adapt to individual differences, and the accuracy of risk assessment is limited are solved, physiological feature data are extracted, coupling abnormal features are analyzed, and the accuracy of risk assessment is improved. The method has the advantages that physiological changes before migraine attack are reflected more comprehensively, reliability of risk assessment is improved, historical intervention cases are introduced, confidence intervals of risk probability values are dynamically adjusted, dynamic adjustment is performed according to individual differences of patients and intervention response conditions, and adaptability of migraine risk reports is improved.

Description

Technical Field

[0001] The present invention relates to the technical field related to the construction and management of risk indicators, and specifically relates to a system for constructing migraine risk indicators based on evidence-based medicine. Background Art

[0002] The pathogenesis of migraine is complex and involves the interaction of multiple aspects such as neurovascular, neurotransmitters, genetic factors, and environmental factors. Due to the sudden and recurrent nature of migraine attacks, it is often difficult to predict the attacks in advance, resulting in a passive response state during the attacks and the inability to take effective preventive or relief measures in a timely manner. Most of the existing prediction methods rely on the subjective symptom reports of patients, simple physiological index monitoring, or risk assessment based on conventional statistical models. There are individual differences in the description and perception of symptoms by patients, which makes it difficult to ensure the accuracy and consistency of symptom reports and affects the reliability of risk prediction.

[0003] In summary, there is a technical problem in the prior art that the generation of risk reports is fixed to preset migraine risk indicators, unable to adapt to individual differences, and the accuracy of risk assessment is limited. Summary of the Invention

[0004] The present application provides a system for constructing migraine risk indicators based on evidence-based medicine, aiming to solve the technical problem in the prior art that the generation of risk reports is fixed to preset migraine risk indicators, unable to adapt to individual differences, and the accuracy of risk assessment is limited.

[0005] In view of the above problems, the technical solution of the present application is as follows: The present application provides a system for constructing evidence-based migraine risk indicators. Among them, the system includes: a feature extraction module for obtaining physiological feature data including cerebral oxygen saturation fluctuation curves and carotid artery blood flow spectra, and extracting coupled abnormal features, where the coupled abnormal features include oxygen oscillation phase desynchronization degree and vascular wall shear stress variation coefficient; an error correction module for mapping the coupled abnormal features to the migraine risk level space and formulating a risk probability surface, and the risk probability surface is corrected for periodic spectral offset with environmental infrasound exposure intensity and blue light irradiation dose parameters; a dynamic adjustment module for configuring the risk probability surface to be associated with a standardized risk feature vector and an initial risk probability value, and introducing historical intervention cases of the target user to dynamically adjust the confidence interval of the initial risk probability value; an update module for, at the same time, evaluating the intervention response of the standardized risk feature vector and the initial risk probability value according to the historical migraine association network, and when the intervention response evaluation value is lower than the intervention response threshold coordinated with the damage index, updating the risk probability surface with an iterative optimization instruction driven by Bayesian evidence theory; a synchronization module for constructing a migraine risk feature matrix on the updated risk probability surface, and the migraine risk feature matrix is used to formulate a migraine risk report and synchronize it to the user terminal.

[0006] In summary, one or more technical solutions provided in the present application achieve the extraction of physiological feature data for analyzing coupled abnormal features, more comprehensively reflecting the physiological changes before a migraine attack, improving the reliability of risk assessment. At the same time, introducing historical intervention cases to dynamically adjust the confidence interval of the risk probability value, so as to dynamically adjust according to the individual differences and intervention response situations of patients, and improve the technical effect of the adaptability of the migraine risk report. Description of the Drawings

[0007] Figure 1 is a schematic structural diagram of a system for constructing evidence-based migraine risk indicators provided by the present application; Figure 2 is a schematic diagram of the distributed iterative training process of the feature extraction module of the system for constructing evidence-based migraine risk indicators provided by the present application.

[0008] Description of the reference numerals: Feature extraction module M100, Error correction module M200, Dynamic adjustment module M300, Update module M400, Synchronization module M500. Detailed Embodiments Embodiment

[0009] The present application will be specifically described below with reference to the drawings. As Figure 1 shown, the present application provides a system for constructing evidence-based migraine risk indicators. Among them, the system includes: A feature extraction module M100 is used to obtain physiological feature data including a cerebral oxygen saturation fluctuation curve and a carotid artery blood flow spectrum, and extract coupled abnormal features, where the coupled abnormal features include a blood oxygen oscillation phase desynchronization degree and a coefficient of variation of vascular wall shear stress.

[0010] An error correction module M200 is used to map the coupled abnormal features to a migraine risk level space, and formulate a risk probability surface, where the risk probability surface is corrected for periodic spectral shift with environmental infrasound exposure intensity and blue light irradiation dose parameters.

[0011] Specifically, the cerebral oxygen saturation fluctuation curve refers to the change trajectory of oxygen saturation in brain tissue continuously monitored by means such as near-infrared spectroscopy technology, which is used to reflect the dynamic state of the balance between cerebral blood oxygen supply and demand; the carotid artery blood flow spectrum is a distribution curve of parameters such as carotid artery blood flow velocity and flow rate over time or cardiac cycle obtained by using Doppler ultrasound technology, which reflects hemodynamic characteristics.

[0012] The blood oxygen oscillation phase desynchronization degree refers to the degree of phase difference of blood oxygen oscillation in different regions or different time periods in the cerebral oxygen saturation fluctuation curve. Phase desynchronization implies abnormalities in local cerebral blood circulation or oxygen metabolism regulation; the coefficient of variation of vascular wall shear stress is a relative index measuring the degree of shear stress fluctuation generated by blood flow on the vascular wall, which reflects the stability of the vascular wall under the impact of blood flow. The higher the coefficient of variation, the greater the shear stress fluctuation, and the more easily the vascular endothelial function may be damaged.

[0013] The migraine risk level space is a mathematical model space constructed based on multi-dimensional physiological features and risk factors, which is used to quantify various characteristic parameters into migraine attack risk levels; the periodic spectral shift error correction refers to using signal processing technologies such as Fourier transform to analyze the periodic change rules of environmental infrasound exposure intensity and blue light irradiation dose parameters in the frequency domain, and adjust the risk probability surface to eliminate errors caused by periodic interference.

[0014] Obtain the cerebral oxygen saturation fluctuation curve and carotid artery blood flow spectrum data. Specifically, receive the changes in cerebral oxygen saturation in regions such as the prefrontal lobe and parietal lobe of the brain from a near-infrared spectrometer. After analysis, it is found that within 24 hours before the onset of migraine patients, the fluctuation amplitude of cerebral oxygen saturation increases on average compared with the baseline level. At the same time, analyze the carotid artery blood flow spectrum, and combine the peak systolic velocity and end-diastolic velocity of the carotid artery in migraine patients to determine abnormal blood flow characteristic data.

[0015] According to the abnormal blood flow characteristic data, the oxygen oscillation phase desynchronization degree is extracted. By performing wavelet transform decomposition on the cerebral oxygen saturation fluctuation curve, the phase information in different frequency bands is determined, and the coefficient of variation of vascular wall shear stress is determined. According to the hemodynamic formula τ = 4μQ / (πr³) (where τ is the shear stress, μ is the blood viscosity, Q is the blood flow, and r is the blood vessel radius), combined with the carotid blood flow spectrum data, the proportion of the coefficient of variation of shear stress in migraine patients can be obtained. These coupled abnormal characteristics can more comprehensively reflect the physiological changes before the onset of migraine.

[0016] When mapping the coupled abnormal characteristics to the migraine risk level space, periodic spectral shift error correction is performed with the parameters of environmental infrasound exposure intensity (unit: dB) and blue light irradiation dose (unit: J / cm²). Specifically, collect the environmental infrasound (frequency range 0.1Hz - 20Hz) exposure intensity data and monitor the blue light irradiation dose. By calculating the periodic spectral components of the environmental infrasound exposure intensity and blue light irradiation dose parameters, least squares method is used for error correction to adjust the risk probability surface, ensuring that the characteristic data input into the model has high correlation and representativeness.

[0017] The dynamic adjustment module M300 is used to configure the standardized risk feature vector and the initial risk probability value associated with the risk probability surface, and introduce the historical intervention cases of the target user to dynamically adjust the confidence interval of the initial risk probability value.

[0018] The update module M400 is used to, at the same time, evaluate the intervention response of the standardized risk feature vector and the initial risk probability value according to the historical migraine association network. When the intervention response evaluation value is lower than the intervention response threshold coordinated with the damage index, an iterative optimization instruction driven by Bayesian evidence theory is used to update the risk probability surface.

[0019] The synchronization module M500 is used to construct a migraine risk feature matrix on the updated risk probability surface, and the migraine risk feature matrix is used to draw up a migraine risk report and synchronize it to the user terminal.

[0020] Specifically, the standardized risk feature vector refers to a vector formed by normalizing a variety of coupled abnormal characteristics (such as oxygen oscillation phase desynchronization degree, coefficient of variation of vascular wall shear stress, etc.) extracted from physiological characteristic data. The normalization process usually uses linear transformation or Z - score standardization to make its values fall within a unified range, facilitating comparison and comprehensive analysis between different characteristics; the initial risk probability value is the basic probability value of the migraine attack risk set based on the big data of the general population or prior knowledge, reflecting the possible probability of having migraine in the absence of specific individual information; the confidence interval is a quantitative interval for estimating the accuracy of the initial risk probability value according to the statistical significance level, used to measure the reliability of the estimated value.

[0021] The historical migraine association network is a knowledge graph constructed based on a large amount of data of migraine patients, including physiological characteristics related to migraine, environmental factors, intervention measures, and the association relationships between various physiological characteristics, which are mined through technologies such as graph neural networks; the intervention response evaluation value is a quantitative index to measure the degree of response of users to previous intervention measures (such as drugs, lifestyle adjustments, etc.), and is calculated by analyzing data such as changes in physiological characteristics and headache attack frequency before and after the intervention.

[0022] The damage index is a comprehensive index to evaluate the degree of damage of migraine attacks to the patient's physical function and quality of life, and is calculated according to specific weights based on factors such as headache intensity, duration, and accompanying symptoms. The iterative optimization instruction driven by Bayesian evidence theory refers to an algorithm instruction that gradually optimizes and adjusts the risk probability surface by continuously updating the prior knowledge combined with new evidence (such as new physiological characteristic data, intervention response results, etc.) using Bayes' theorem; the migraine risk characteristic matrix is a two-dimensional or high-dimensional matrix integrating the standardized risk characteristic vectors, dynamically adjusted initial risk probability values, and relevant association information, which is used to comprehensively and systematically describe the migraine risk status.

[0023] When configuring the risk probability surface to associate the standardized risk characteristic vectors and the initial risk probability values, by collecting the physiological characteristic data of the target user for a continuous period of time (for example, one month), the standardized oxygen oscillation phase desynchronization degree, vascular wall shear stress variation coefficient, etc. are obtained, and these characteristic vectors are associated with the initial risk probability values to construct the risk probability surface; after introducing the historical intervention cases of the target user, the confidence interval of the initial risk probability value is dynamically adjusted accordingly, narrowing its width, improving the accuracy of risk assessment, and making the risk probability value more in line with the actual situation of the user individual.

[0024] When evaluating the intervention response of the standardized risk characteristic vectors and the initial risk probability values according to the historical migraine association network, an intervention response threshold coordinated with the damage index is set. When the intervention response evaluation value is lower than the intervention response threshold coordinated with the damage index, the iterative optimization instruction driven by Bayesian evidence theory is triggered. Using the newly collected physiological characteristic data as evidence and applying Bayesian evidence theory, the risk probability surface can more accurately reflect the real-time risk status of the user; when constructing the migraine risk characteristic matrix, information such as the updated risk probability surface data and the trend of physiological characteristic changes is integrated. The migraine risk report formulated based on this matrix can list in detail key contents such as the migraine attack risk level of the user and the contribution degree of the main risk factors, and is synchronously sent to the user terminal in real time.

[0025] Furthermore, the migraine risk characteristic matrix is used to formulate a migraine risk report and synchronize it to the user terminal, and the synchronization module M500 is also used to execute the following method: Integrate an ultraviolet intensity sensor and a barometric pressure change detection unit to monitor the light intensity and altitude change during outdoor activities in real time; based on the light intensity and altitude change, evaluate the environmental induction index, and make a determination in combination with the individual tolerance threshold; if the determination result is yes, then combine the user location data to perform automatic route planning and navigation.

[0026] Specifically, an ultraviolet intensity sensor is a device that can detect the environmental ultraviolet radiation intensity in real time. Its measurement unit is usually microwatts per square centimeter (μW / cm²), which can be used to evaluate the impact of outdoor light on the human body; the barometric pressure change detection unit monitors the change in atmospheric pressure through a barometric pressure sensor, usually in hectopascals (hPa), and can be used to infer the altitude change, because the atmospheric pressure decreases as the altitude increases; the environmental induction index is a quantitative index that evaluates the possibility of these factors inducing migraines after comprehensively considering various environmental factors such as light intensity and altitude change. This index is usually established based on a large amount of statistical data and the environmental exposure responses of migraine patients.

[0027] The individual tolerance threshold refers to the maximum intensity or change range of environmental induction factors that each user can tolerate according to their own physiological characteristics and historical migraine attack situations, and it is an important parameter for individualized risk assessment; automatic route planning and navigation refers to generating a navigation route that avoids high migraine risk areas for the user based on the user's location data and migraine risk assessment results, using map services and route planning.

[0028] The integrated ultraviolet intensity sensor and barometric pressure change detection unit monitor the light intensity and altitude change during outdoor activities in real time. The sensor measures the ultraviolet intensity once every fixed period. When the ultraviolet intensity exceeds the ultraviolet intensity threshold (for example, 300 μW / cm²), it is recorded as high-intensity light; at the same time, when the barometric pressure sensor detects that the barometric pressure is dropping beyond the lower limit of the barometric pressure, it is inferred that the altitude may be rising; based on these data, the environmental induction index is evaluated. Through statistical analysis, it is found that when the environmental induction index exceeds the individual tolerance threshold, the risk of migraine attack increases.

[0029] Combined with the individual tolerance threshold, the system makes a determination. If the determination result is yes (that is, the environmental induction factor exceeds the individual tolerance threshold), then combine the user location data (through longitude and latitude coordinates), and use the map service API to automatically plan a navigation route that avoids high-intensity light areas and sections with sharp altitude changes; by monitoring environmental factors in real time and evaluating their induction risks, potential migraine-inducing environmental conditions can be detected in a timely manner, providing corresponding avoidance suggestions for users, enhancing the practicality and user experience of the system, expanding the migraine risk assessment from simple physiological monitoring to a comprehensive assessment of environmental and physiological factors, and providing more comprehensive migraine prevention support for users.

[0030] Further, the synchronization module M500 is further configured to execute the following method: Integrate a sound wave spectrum analysis unit in the noise-canceling headphones to continuously monitor the energy proportion of the noise in the wave frequency band in the ambient noise, and determine the acoustic induction index; when the acoustic induction index exceeds the individualized tolerance threshold, activate the active noise-canceling system for dynamic attenuation, and synchronously adjust the noise-canceling intensity gradient in combination with the user's heart rate variability data.

[0031] Specifically, the sound wave spectrum analysis unit is an electronic component integrated in the noise-canceling headphones, and its function is to decompose the spectrum of the ambient noise, that is, to split the complex sound signal according to different frequency components, and calculate the energy proportion of each frequency band; the acoustic induction index is based on the sound wave spectrum analysis result, combined with the sensitivity data of migraine patients to the noise in the wave frequency band, and is a quantitative index used to evaluate the risk degree of the current ambient noise inducing migraine.

[0032] The individualized tolerance threshold refers to the maximum acceptable value of the acoustic induction index set for each user according to the user's personal auditory sensitivity, migraine history, and previous responses to noise; dynamic attenuation is the process of automatically adjusting the noise-canceling depth according to the intensity and frequency changes of the noise monitored in real time by generating sound waves with opposite phases to the external noise to cancel the noise.

[0033] The heart rate variability data refers to the change of the user's heart rate within a certain period of time, usually obtained through a wearable heart rate monitoring device, which reflects the activity state of the autonomic nervous system and is closely related to stress and pain perception; the noise-canceling intensity gradient refers to the rate and amplitude of the active noise-canceling system adjusting the noise-canceling intensity under different noise conditions, and synchronous adjustment means changing the noise-canceling intensity in real time according to the heart rate variability data to achieve the best comfort and noise-canceling effect.

[0034] Preferably, the noise in the wave frequency band stimulates the auditory nerve, which in turn affects the balance of brain neurotransmitters. By integrating the sound wave spectrum analysis unit in the noise-canceling headphones to continuously monitor the ambient noise, the sound wave spectrum analysis unit analyzes the spectrum of the ambient noise, divides the sound frequency range into multiple frequency bands, and determines the energy proportion of each frequency band; migraine patients are particularly sensitive to the noise in a fixed frequency band. Evaluate the acoustic induction index. When the acoustic induction index exceeds the individualized tolerance threshold, the active noise-canceling system is immediately activated. At the same time, the user's heart rate variability data is continuously obtained, and according to the change rate of the heart rate variability, the gradient of the noise-canceling intensity is synchronously adjusted to dynamically increase, reducing the negative impact of the noise in the wave frequency band on migraine patients and improving the effectiveness of the system in practical applications.

[0035] Further, physiological characteristic data including a cerebral oxygen saturation fluctuation curve and a carotid artery blood flow spectrum are obtained, and coupling abnormal characteristics are extracted. The feature extraction module M100 is further configured to execute the following method: Based on the physiological characteristic data, analyze the time-varying causal relationship between cerebral oxygen saturation and carotid artery blood flow velocity, and determine the attenuation rate of prefrontal-brainstem functional connectivity; based on the attenuation rate of prefrontal-brainstem functional connectivity, construct a coupled differential equation of vascular wall shear stress and intracranial pressure conduction, and determine the compensation critical point, where the compensation critical point is used to quantify the physiological function compensation state before a migraine attack.

[0036] Specifically, the time-varying causal relationship means that there is a dynamic causal association between cerebral oxygen saturation and carotid artery blood flow velocity at different time points, that is, a change in cerebral oxygen saturation at a certain moment may be caused by fluctuations in carotid artery blood flow velocity in a previous period of time, and this causal association changes over time; the attenuation rate of prefrontal-brainstem functional connectivity is an index to measure the rate of decline in the strength of the functional connection between the prefrontal lobe and the brainstem of the brain over time, usually based on functional magnetic resonance imaging (fMRI) or near-infrared spectroscopy imaging (fNIRS) data, and is determined by analyzing the synchronous change in blood oxygen level fluctuations in the two brain regions. The higher the attenuation rate, the faster the functional synergy between the two decreases.

[0037] The coupled differential equation of vascular wall shear stress and intracranial pressure conduction is a mathematical model used to describe the dynamic conduction relationship between the shear stress on the vascular wall and the change in intracranial pressure. The equation includes hemodynamic parameters (such as blood flow velocity, blood vessel diameter, etc.) and cerebrospinal fluid dynamics parameters (such as intracranial pressure, cerebrospinal fluid flow resistance, etc.). By solving this equation, the law of their interaction can be revealed. The compensation critical point refers to the critical state when the body's physiological function compensation mechanism reaches its limit before a migraine attack. At this time, a small change in physiological indicators triggers a headache attack. The compensation critical point is determined by analyzing the change trend of the solution of the coupled differential equation under specific conditions, and is used to quantify the limit state of the body's physiological function compensation ability before a migraine attack.

[0038] Based on the physiological characteristic data, use time series analysis such as causal test to analyze the time-varying causal relationship between cerebral oxygen saturation and carotid artery blood flow velocity. Taking near-infrared spectroscopy imaging data as an example, analyze the time series of monitored cerebral oxygen saturation and carotid artery blood flow velocity. For example, it is found that when the carotid artery blood flow velocity increases by 10% at time t, the cerebral oxygen saturation rises by an average of 5% at time t + 2 minutes, and this causal relationship significantly indicates the existence of a causal association; by calculating the change in causal strength within multiple time windows, determine the attenuation rate of prefrontal-brainstem functional connectivity, that is, the functional connectivity between the two per minute, to provide support for the subsequent construction of the coupled differential equation.

[0039] Based on the determined attenuation rate of prefrontal-brainstem functional connectivity, a coupled differential equation of vascular wall shear stress and intracranial pressure conduction was constructed; the compensatory critical point was determined by simulation analysis using the coupled differential equations, and the risk of migraine attacks increased at the compensatory critical point; by deeply exploring the physiological mechanisms behind the physiological characteristic data, quantifying the changes in neural functional connections and constructing a coupling model, a key basis was provided for more accurate assessment of the risk of migraine attacks.

[0040] Furthermore, the compensation critical point is used to quantify the physiological function compensation state before the onset of migraine, and the feature extraction module M100 is also used to execute the following method: Monitor the pulsation propagation velocity of the middle meningeal artery and establish a nonlinear regression function of the pulsation propagation velocity and cerebrospinal fluid pressure fluctuation; if the standard deviation of the arterial pulsation propagation velocity exceeds the baseline value, combine the time window to deeply explore the associated physiological characteristics that change synergistically with the compensatory state of the physiological function, and configure a time series association mechanism.

[0041] Specifically, the middle meningeal artery pulsation propagation velocity refers to the rate at which the blood flow pulsation of the middle meningeal artery (the blood flow pulse generated by the heartbeat) propagates in the blood vessels, usually measured in meters per second (m / s), which can reflect the stability of cerebral hemodynamics; spinal fluid pressure fluctuation refers to the change in cerebrospinal fluid pressure in the ventricular system and subarachnoid space over time, usually measured in millimeters of mercury (mmHg), and its fluctuation is closely related to cerebral blood flow regulation and intracranial pressure balance.

[0042] The nonlinear regression function is a mathematical model used to describe the nonlinear relationship between two variables, which refers to the complex correlation between the pulsation propagation velocity and the cerebrospinal fluid pressure fluctuation. This relationship is expressed by fitting historical data to find the optimal function form. The standard deviation of the arterial pulsation propagation velocity is an indicator of the degree of dispersion of the pulsation propagation velocity around its mean value. The larger the standard deviation, the more drastic the fluctuation of the pulsation propagation velocity.

[0043] The baseline value refers to the average level of the standard deviation of the arterial pulsation propagation velocity in a normal physiological state or without the risk of migraine attack, which is used as a reference value to judge whether it is abnormal; the time window refers to the length of the time period used to analyze the data, which is used to extract characteristic information within a specific time period from continuous data. The associated physiological characteristics of the coordinated changes in the physiological function compensation state refer to other related physiological indicators that change simultaneously with the physiological function compensation state, such as heart rate variability, blood oxygen saturation, vascular wall shear stress, etc. These characteristics may show specific change patterns before the onset of migraine; the temporal association mechanism is used to explore the correlation between different physiological characteristics in the time series, determine the order, mutual influence and common change rules between different physiological characteristics, so as to more comprehensively understand the physiological changes before the onset of migraine.

[0044] Monitor the pulsatile propagation velocity of the middle meningeal artery. If the average value of the monitored pulsatile propagation velocity is 0.5 m / s and the standard deviation is 0.12 m / s, while the standard deviation of the baseline value is only 0.05 m / s, it indicates that the standard deviation significantly exceeds the baseline value. At this time, the system combines the set time window to deeply mine the associated physiological characteristics that co-vary with the physiological function compensation state. Specifically, within the time window with abnormal standard deviation of the pulsatile propagation velocity, the standard deviation of heart rate variability decreases, the fluctuation amplitude of blood oxygen saturation increases, and the coefficient of variation of vascular wall shear stress increases; based on these characteristics, a time-series association mechanism is established. Through Granger causality test, it is found that the change in pulsatile propagation velocity precedes the change in heart rate variability, the fluctuation of blood oxygen saturation occurs almost simultaneously with the change in pulsatile propagation velocity, and the change in vascular wall shear stress lags behind the change in pulsatile propagation velocity.

[0045] Through the above analysis, the understanding of the physiological function compensation state before migraine attack is deepened. Also, by mining the time-series association of multiple physiological characteristics, it provides strong support for accurately identifying the early warning signals of migraine, effectively enriching the early warning ability of the entire risk assessment system, enabling the system to capture the complex physiological changes before migraine attack more timely and comprehensively.

[0046] Furthermore, the feature extraction module M100 is also used to execute the following method: The physiological function compensation state includes the eye movement pattern; based on the time-series association mechanism, activate the micro pressure sensor array embedded in the contact lens to receive the eyelid microtremor frequency and intraocular pressure fluctuation data; through the eyelid microtremor frequency and intraocular pressure fluctuation data, set the association mapping index item between the eye movement pattern and the migraine prodromal period.

[0047] Specifically, the micro pressure sensor array refers to a group of tiny pressure sensors embedded in the contact lens, which can monitor the eyelid microtremor frequency and intraocular pressure fluctuation in real time; the eyelid microtremor frequency refers to the frequency when the eyelid trembles involuntarily, with the unit of Hertz (Hz), reflecting the tension degree of the eyelid muscle and the neural regulation state; the intraocular pressure fluctuation data refers to the change of intraocular pressure over time, usually in millimeters of mercury (mmHg), and abnormal intraocular pressure fluctuation may be related to the change of cerebrospinal fluid pressure and migraine attack.

[0048] The eye movement pattern refers to a series of characteristics of eye movement, including eyelid microtremor frequency, intraocular pressure fluctuation, eye movement trajectory, etc. These characteristics may present specific change patterns during the migraine prodromal period; the association mapping index item refers to the index that establishes an association relationship between the eye movement pattern and the migraine prodromal period. By analyzing the change characteristics of the eye movement pattern, key parameters are extracted as the index for quickly identifying the arrival of the migraine prodromal period.

[0049] Based on the temporal correlation mechanism, start the micro pressure sensor array embedded in the contact lens. Specifically, start the micro pressure sensor array embedded in the contact lens to synchronously receive the eyelid microtremor frequency and intraocular pressure fluctuation data; by collecting these data and combining with the temporal correlation mechanism, set the correlation mapping index items between the eye movement pattern and the prodromal period of migraine; by introducing this new physiological feature dimension of the eye movement pattern, the data source for risk assessment is further enriched. Furthermore, the change in the eye movement pattern can reflect the early changes in the autonomic nervous system and cerebral hemodynamics, providing a more comprehensive basis for early warning of migraine attacks.

[0050] Furthermore, the feature extraction module M100 is also used to execute the following method: Introduce the exosome concentration. When the correlation coefficient between the exosome concentration and the compensation critical point meets the correlation coefficient threshold, include the exosome concentration in the physiological function compensation state evaluation system; in the physiological function compensation state evaluation system, use the quantum key distribution protocol to generate a biofeedback training key, and the biofeedback training key matches the historical intervention cases of the target user.

[0051] Specifically, the exosome concentration refers to the number of exosomes contained in a unit volume of body fluid (such as blood, cerebrospinal fluid, etc.). Exosomes are small vesicles secreted by cells, carrying various biomolecules (such as proteins, nucleic acids, etc.), and their concentration changes can reflect the physiological state and pathological processes of cells. The correlation coefficient threshold is a set numerical standard used to judge whether the correlation between the exosome concentration and the compensation critical point is strong enough. Usually, the value ranges from 0.3 to 0.5. When the correlation coefficient exceeds this threshold, it is considered that there is a significant correlation between the two.

[0052] The physiological function compensation state evaluation system is a systematic model for comprehensively evaluating the compensation capabilities of each system of the body when facing physiological dysfunction. It contains multiple physiological indicators and evaluation parameters for quantifying the compensation state of the body; the quantum key distribution protocol is an encryption communication protocol based on the principles of quantum mechanics. It establishes a secure key that cannot be stolen between communication parties through the transmission of quantum states to protect the transmission security of the biofeedback training key; the biofeedback training key is a set of encoded parameters generated based on the user's physiological data and intervention history, used to guide the user to perform targeted biofeedback training to help regulate physiological functions and relieve migraine symptoms.

[0053] Introduce the exosome concentration as a new evaluation index, determine the correlation coefficient between the exosome concentration and the compensation critical point. Assuming the correlation coefficient threshold is set to 0.4, through statistical analysis, it is found that the correlation coefficient between the two is 0.45, exceeding the threshold, indicating that the exosome concentration can reflect the change in the body's compensation state. Therefore, it is included in the physiological function compensation state evaluation system, increasing the index dimension of the evaluation system and enriching the evaluation data source.

[0054] In the evaluation system, a quantum key distribution protocol is used to generate a biofeedback training key. Correspondingly, the user has a good response to the biofeedback training, and the frequency of their headache attacks decreases. It is encrypted and transmitted to the user's terminal device to guide the user to perform effective biofeedback training. In the above steps, introducing exosome concentration expands the breadth and depth of the evaluation system, enabling it to more accurately capture the changes in the compensatory state of the body and enhancing the comprehensiveness and accuracy of risk assessment. Using the quantum key distribution protocol ensures the security of the biofeedback training key, ensuring that the user can receive personalized, secure, and reliable training guidance and improving the effectiveness of the intervention measures.

[0055] Furthermore, as Figure 2 shown, the feature extraction module M100 is also used to execute the following method: In the physiological function compensatory state evaluation system, a data drift unit is configured. The data drift unit is used to analyze the change in the KL divergence of the distribution corresponding to the historical intervention cases of the target user. Perform population subgroup difference analysis on the target user. If it exceeds the difference threshold, activate the federated learning framework of the central server to perform distributed iterative training on each population subgroup.

[0056] Specifically, the data drift unit is a module for monitoring and analyzing data distribution changes. It can detect whether the distribution characteristics of data shift under different time periods or different conditions, which is crucial for evaluating the stability and reliability of the model. The KL divergence of the distribution is an asymmetric measure for measuring the difference between two probability distributions, used to quantify the difference degree between the data distribution of the target user's historical intervention cases and the current data distribution. The larger its value, the greater the distribution difference.

[0057] Population subgroup difference analysis refers to dividing the population into different subgroups according to specific characteristics (such as age, gender, medical history, etc.) and analyzing the differences in physiological function compensatory state, intervention response, etc. between each subgroup to identify groups with different risk characteristics and intervention needs. The difference threshold is a preset standard value used to judge whether the difference between population subgroups is significant. When the KL divergence exceeds this threshold, it is considered that the difference is large enough and further model optimization measures need to be taken. The federated learning framework is a distributed machine learning method that allows multiple devices or institutions to jointly train a model without sharing the original data, protects data privacy through encryption technology, and at the same time realizes knowledge sharing and model optimization, improving the generalization ability of the model for different population subgroups.

[0058] Configure a data drift unit in the physiological function compensation state evaluation system to analyze the change in the distribution KL divergence corresponding to the historical intervention cases of the target user. Exemplarily, the intervention case data of user A in the past 6 months are collected, including physiological indicators, intervention measures, and effects before and after the intervention, and are divided into a training data set (data in the first 3 months) and a test data set (data in the last 3 months); determine the distribution KL divergence of the two data sets. If the difference threshold is set to 0.15 and the calculated KL divergence is 0.20, exceeding the difference threshold indicates that the data distribution has changed significantly, which may be due to changes in the user's physiological state, intervention response, or external environment, etc.

[0059] At this time, trigger the analysis of subgroup differences in the population for the target user. Divide the population into different subgroups according to characteristics such as age (e.g., divided into less than 30 years old, 30 - 50 years old, and more than 50 years old), gender, medical history (e.g., whether there is a family history of migraine), etc., indicating that there are significant differences between different subgroups. At the same time, activate the federated learning framework of the central server, encrypt the data of each population subgroup and distribute it to different computing nodes for distributed iterative training, while protecting the privacy data of users; in the above steps, by monitoring data drift and analyzing subgroup differences in the population, the dynamic adaptability and accuracy of the evaluation system are ensured, and the applicability changes of the model in different populations and different periods are timely discovered, and targeted optimization is carried out through federated learning, enabling the system to continuously provide reliable risk assessment and intervention suggestions, effectively improving the long-term effectiveness and wide applicability of the solution.

[0060] In summary, the beneficial effects of the embodiments of the present application are: Due to the adoption of a feature extraction module for obtaining physiological feature data including cerebral oxygen saturation fluctuation curves and carotid artery blood flow spectra and extracting coupled abnormal features, an error correction module for mapping the coupled abnormal features to the migraine risk level space and formulating a risk probability surface, where the risk probability surface is corrected for periodic spectral shift errors with environmental infrasound exposure intensity and blue light irradiation dose parameters, a dynamic adjustment module for configuring the risk probability surface to be associated with a standardized risk feature vector and an initial risk probability value and introducing historical intervention cases of the target user to dynamically adjust the confidence interval of the initial risk probability value, an update module for, at the same time, evaluating the intervention response of the standardized risk feature vector and the initial risk probability value according to the historical migraine association network, and when the intervention response evaluation value is lower than the intervention response threshold coordinated with the damage index, updating the risk probability surface with an iterative optimization instruction driven by Bayesian evidence theory, and a synchronization module for constructing a migraine risk feature matrix on the updated risk probability surface, where the migraine risk feature matrix is used to formulate a migraine risk report and synchronize it to the user terminal. By providing a migraine risk index construction system based on evidence-based practice, the present application realizes the extraction of physiological feature data to analyze coupled abnormal features, more comprehensively reflects the physiological changes before a migraine attack, improves the reliability of risk assessment, and at the same time, introduces historical intervention cases to dynamically adjust the confidence interval of the risk probability value, so as to make dynamic adjustments according to the individual differences and intervention response situations of patients, achieving the technical effect of improving the adaptability of the migraine risk report.

[0061] In summary, any step can be stored as computer instructions or programs in an unrestricted computer memory and can be called and recognized by an unrestricted computer processor, and no redundant restrictions are imposed here.

[0062] Furthermore, the above technical solutions only represent the preferred technical solutions of the technical solutions of the embodiments of the present application. Some changes that may be made by those skilled in the art to some parts thereof all reflect the principles of the novel embodiments of the present application. Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the scope of the present application.

Claims

1. Construct a system based on evidence-based migraine risk indicators, characterized by: include: A feature extraction module is used to obtain physiological feature data including cerebral oxygen saturation fluctuation curve and carotid artery blood flow spectrum, and extract coupling abnormality features, wherein the coupling abnormality features include blood oxygen oscillation phase desynchronization and vascular wall shear stress variation coefficient; An error correction module is used to map the coupling abnormality characteristics to the migraine risk level space and formulate a risk probability surface, wherein the risk probability surface is used to perform periodic spectrum offset error correction based on the ambient infrasound exposure intensity and the blue light irradiation dose parameters; A dynamic adjustment module, used to configure the risk probability surface to associate with the standardized risk feature vector and the initial risk probability value, and introduce the historical intervention cases of the target user to dynamically adjust the confidence interval of the initial risk probability value; An updating module, for simultaneously, performing intervention response evaluation on the standardized risk feature vector and the initial risk probability value according to the historical migraine association network, and updating the risk probability surface with iterative optimization instructions driven by Bayesian evidence theory when a number of pre-response evaluation values ​​are lower than an intervention response threshold coordinated with the injury index; The synchronization module is used to construct a migraine risk feature matrix based on the updated risk probability surface, and the migraine risk feature matrix is ​​used to prepare a migraine risk report and synchronize it to the user terminal.

2. The system for constructing an evidence-based migraine risk index according to claim 1, wherein: The migraine risk feature matrix is ​​used to prepare a migraine risk report and synchronize it to a user terminal, and also includes: Integrated UV intensity sensor and air pressure change detection unit to monitor light intensity and altitude changes during outdoor activities in real time; Based on the light intensity and altitude changes, the environmental induced index is evaluated and determined in combination with the individual tolerance threshold; If the judgment result is yes, automatic planning and navigation are performed in combination with the user's location data.

3. The system for constructing an evidence-based migraine risk index according to claim 2, wherein: Also includes: Integrate a sound wave spectrum analysis unit in the noise reduction headphones to monitor the ambient noise in real time. The energy proportion of the noise in the wave frequency band is used to determine the acoustic induced index; When the acoustic induced index exceeds the individualized tolerance threshold, the active noise reduction system is activated for dynamic attenuation, and the noise reduction intensity gradient is synchronously adjusted in combination with the user's heart rate variability data.

4. The evidence-based migraine risk index construction system according to claim 1, characterized in that: Acquire physiological characteristic data including brain oxygen saturation fluctuation curve and carotid artery blood flow spectrum, extract coupling abnormality features, and also include: According to the physiological characteristic data, analyzing the time-varying causal relationship between cerebral oxygen saturation and carotid artery blood flow velocity, and determining the attenuation rate of prefrontal lobe-brainstem functional connectivity; Based on the attenuation rate of the prefrontal lobe-brainstem functional connectivity, a coupled differential equation of vascular wall shear stress and intracranial pressure conduction is constructed to determine the compensation critical point, which is used to quantify the physiological function compensation state before the onset of migraine.

5. The system for constructing an evidence-based migraine risk index according to claim 4, wherein: The compensation critical point is used to quantify the physiological function compensation state before the migraine attack, and also includes: Monitor the pulsation propagation velocity of the middle meningeal artery and establish a nonlinear regression function between the pulsation propagation velocity and the cerebrospinal fluid pressure fluctuation; If the standard deviation of the arterial pulsation propagation velocity exceeds the baseline value, combined with the time window, the associated physiological characteristics that change synergistically with the compensatory state of the physiological function are deeply mined, and a time series association mechanism is configured.

6. The system for constructing an evidence-based migraine risk index according to claim 5, wherein: The physiological function compensation state includes eye movement patterns; Based on the timing association mechanism, a micro pressure sensor array embedded in the contact lens is activated to receive eyelid micro-tremor frequency and intraocular pressure fluctuation data; The association mapping index item between the eye movement pattern and the prodromal stage of migraine is set by using the eyelid microtremor frequency and intraocular pressure fluctuation data.

7. The system for constructing an evidence-based migraine risk index according to claim 6, wherein: Introducing exosome concentration, when the correlation coefficient between the exosome concentration and the compensation critical point meets the correlation coefficient threshold, the exosome concentration is included in the physiological function compensation state evaluation system; In the physiological function compensation state assessment system, a quantum key distribution protocol is used to generate a biofeedback training key, and the biofeedback training key matches the historical intervention case of the target user.

8. The system for constructing an evidence-based migraine risk index according to claim 7, wherein: include: In the physiological function compensation state assessment system, a data drift unit is configured, and the data drift unit is used to analyze the distribution KL divergence changes corresponding to the historical intervention cases of the target user; Perform population subgroup difference analysis on the target users. If the difference threshold is exceeded, activate the federated learning framework of the central server to perform distributed iterative training on each population subgroup.

Images