Closed-loop operation control method for mind-body integrated health management

Abstract

The application discloses a closed-loop operation control method for physical and mental integrated health management, and relates to the technical field of digital health management. The method constructs a metabolic syndrome patient portrait and comprehensively stratifies. Through a near-field communication component, health education is triggered in different home indoor life scenes, and a health education interaction behavior sequence is collected as a psychological behavior state observation signal. An artistic healing salon is organized as a controlled disturbance, and physiological responses before and after the disturbance are collected as disturbance response observation signals. The health education interaction behavior sequence, the disturbance response observation signals, physiological monitoring data and psychological scale data are asynchronously time-aligned, individual physical and mental coupling time lags and directions are identified, and the health education push timing and the salon scheduling are phase-aligned and adjusted according to the identification results and fed back to the patient portrait to form a closed loop. The intervention action is reversed into an observation probe, observation blanks are filled in without increasing wearable physiological devices, and the phase alignment of the intervention timing and the individual physical and mental rhythm is realized.

Description

Technical Field

[0001] This invention relates to the field of digital health management technology, and more specifically to a closed-loop operation control method for holistic health management of people with metabolic syndrome. Background Technology

[0002] Metabolic syndrome is a clinical syndrome comprised of multiple metabolic abnormalities, including hypertension, diabetes, lipid metabolism disorders, and central obesity. It affects a large population, and its onset and outcome are influenced by both physiological metabolic factors and psychological stress. Long-term chronic disease management of this population requires continuous tracking of physiological data such as blood pressure, blood glucose, and blood lipids, as well as assessment and intervention for psychological states such as anxiety, depression, and social isolation. However, under the real-world conditions of long-term home management, physiological monitoring data typically relies on patient self-testing, resulting in sparse and irregular sampling; psychological scale data are usually obtained through monthly or quarterly reviews, with even lower frequency. These two factors are misaligned on a timeline, forming two independent, sparse, and asynchronous sequences.

[0003] Chinese patent application CN112466423A discloses an intelligent chronic disease health management method and system. This method identifies chronic disease information of a target individual, generates management sub-plans for patients, chronic disease managers, and physicians based on this information, and pushes health education and intervention plans including psychological, nutritional, and exercise content to patients via mobile devices, while simultaneously monitoring and following up with the patient. However, this scheme positions the delivery of health education and intervention plans as a one-way output action, focusing only on delivering content to the patient and failing to utilize the interactive behaviors generated by the patient during the education process. This leaves a large gap of unobserved data between sparse physiological monitoring and low-frequency psychological assessments, making it difficult to perceive the drift in the patient's psychological and behavioral state between two formal assessments. Consequently, its closed-loop adjustment can only be based on group experience rules, failing to characterize the mind-body interaction relationship of a specific individual patient.

[0004] Chinese patent application CN113782208A discloses a home health assessment and intervention system based on intelligent rehabilitation equipment. This system collects individual health status and home environment data through a data acquisition module, uses deep learning to jointly identify physiological, psychological states, and behaviors through a simulation modeling and assessment module, and dynamically updates intervention strategies through a health intervention module. However, this approach uses synchronous fusion assessment of physiological and psychological data at the same time or within the same assessment period, assuming the two types of data are time-aligned. It cannot handle the irregular and misaligned timing of physiological and psychological observations in a home setting, and it fails to identify the individual differences in the temporal sequence between individual psychological states and physiological indicators. Therefore, it cannot determine whether, for a specific patient, changes in psychological state or deterioration of physiological indicators precedes changes in physiological state.

[0005] As can be seen from the existing technologies described above, the core bottleneck restricting the integrated management of the mind and body in people with metabolic syndrome lies in the fact that, under the conditions of sparse, low-frequency, and misaligned asynchronous sampling of both physiological monitoring and psychological assessment, existing solutions treat health education and psychological intervention as purely output actions, discarding the observational signals that could have been utilized during the intervention interaction. This results in a lack of intermediate-state observations between the two sparse physiological and psychological sequences, making it impossible to identify the individualized mind-body coupling relationship and its temporal sequence for a specific patient. Consequently, interventions can only be arranged using a uniform step size and group experience, leading to a mismatch between the timing of intervention and the individual's mind-body rhythm. How to obtain intermediate-state observations to fill the above-mentioned observational gaps without adding additional wearable physiological devices, identify the time lag and direction of individual mind-body coupling, and thereby align the timing of intervention with the phase of the individual's mind-body rhythm, is a technical problem that urgently needs to be solved in this field. Summary of the Invention

[0006] Addressing the core technical bottlenecks in existing home health management for metabolic syndrome—namely, the inability to identify individual mind-body coupling and achieve intervention phase alignment due to sparse asynchronous sampling of physiological monitoring and psychological assessment, the unidirectional output of intervention actions discarding their observational value, and the resulting failure to identify individual mind-body coupling—this invention provides a closed-loop operation control method for integrated mind-body health management. This method reverses two types of intervention actions—scene-based educational interaction and art therapy salons—into observation probes for psychological and behavioral states and controlled perturbation responses to psychological states. Furthermore, it performs time alignment on multiple asynchronous sparse sequences to identify the time lag and direction of individual mind-body coupling, thereby adjusting the intervention timing based on phase alignment. Without adding wearable physiological devices, this method achieves phase-aligned closed-loop control of the mind-body rhythms of metabolic syndrome patients from the principle that intervention actions are the source of observation.

[0007] The technical solution of this invention is as follows: A closed-loop operation and control method for holistic health management encompassing mind and body includes the following steps: Step S1: Construct a patient profile based on the basic health data, physiological monitoring data and psychological scale data of patients with metabolic syndrome. The patient profile includes a comprehensive stratification result determined by metabolic risk and psychological state. Step S2: Trigger scenario education content in different home indoor living scenarios through the patient's personal near-field communication component and push it to the patient's terminal. Collect the education interaction behavior sequence as a psychological behavior state observation signal. The education interaction behavior sequence is the interaction behavior data recorded by the patient's terminal in response to the scenario education content trigger. Step S3: Organize an art therapy salon as a controlled disturbance event of psychological state, and collect physiological response data before and after salon participation as disturbance response observation signals; Step S4: Construct a scene transfer topology based on the trigger transfer relationship of the near-field communication component among multiple scene tags, and use the scene transfer topology to correct the educational interaction behavior sequence; Step S5: Asynchronously time-align the educational interaction behavior sequence, the disturbance response observation signal, the physiological monitoring data, and the psychological scale data to identify the individualized mind-body coupling time lag and coupling direction of the patient. Step S6: Based on the mind-body coupling time lag and the coupling direction, the timing of pushing scene education content and the scheduling of art therapy salons are phase aligned and adjusted, and the phase alignment adjustment results are fed back to the patient profile to form a closed-loop operation control of monitoring, evaluation, intervention and feedback.

[0008] The beneficial effects of this invention are as follows: Firstly, this invention uses the sequence of interactive educational scenarios as a signal for observing psychological and behavioral states, and the art therapy salon as a controlled perturbation of psychological states, collecting physiological responses before and after. This allows for the acquisition of previously missing intermediate state observations between two formal assessments. The mechanism lies in the fact that the patient's active approach to the near-field communication component triggering the educational activity and reading serves as a sample of behavior voluntarily exposed by the patient. Salon participation, on the other hand, represents a controlled perturbation applied to the psychological dimension. These two activities constitute observable input-output pairs in the psychological-behavioral pathway and the psychological-to-physiological pathway, respectively. In contrast, simply treating educational activities as a one-way push and discarding interactive behavior fails to obtain the aforementioned intermediate state observations. This invention thus increases observation density without adding any wearable physiological devices.

[0009] Secondly, this invention asynchronously aligns educational interaction sequences, disturbance response observation signals, physiological monitoring data, and psychological scale data, identifying the time delay and direction of mind-body coupling. This solves the problem of being unable to characterize individual mind-body causal relationships when multiple sparse sequences are misaligned. The mechanism lies in using kernel-weighted irregular sampling cross-correlation to directly estimate the time shift that maximizes the correlation between the psychological and physiological observation branches under irregular sampling conditions, thereby obtaining individualized mind-body coupling time delays and directions. In contrast, synchronous fusion assessment and default data time alignment cannot identify the temporal sequence between individual psychological and physiological processes.

[0010] Third, this invention adjusts the timing of intervention based on the identified mind-body coupling time lag, allowing for early intervention within a window where psychological states have already shown signs of drift but physiological indicators have not yet exceeded limits. This upgrades the intervention from a group experience adjustment with a uniform step size to a causal time lag constraint adjustment aligned with the individual's mind-body rhythm. The mechanism lies in formalizing the time-delayed directed mind-body coupling relationship as a constraint imposed on closed-loop control, causing the intervention timing of the control center to shift forward with the individual's time lag, thus intervening before physiological deterioration manifests.

[0011] Fourth, the above three aspects are coupled to form a synergistic effect. The mechanism is that the intervention action is reversed into the intermediate state observation obtained by the observation probe, which is exactly the data source for identifying the time lag of the individual's mind-body coupling; and the identified time lag is exactly the source of the lead time for phase-aligned intervention; the intervention action both generates observation and is calibrated by the phase alignment and is re-driven, so that the four links of monitoring, assessment, intervention and feedback are connected end to end, forming a phase alignment closed loop driven by the individual's mind-body coupling that cannot be achieved by any single link alone. Its overall effect is significantly better than the existing scheme that simply superimposes the three of observation, assessment and intervention. Attached Figure Description

[0012] Figure 1 This is the overall flowchart of the closed-loop operation control method for the holistic health management of mind and body of the present invention; Figure 2 This is a flowchart of step S2 of the present invention, which involves collecting the sequence of educational interaction behaviors as a signal for observing psychological and behavioral states. Figure 3 This is a flowchart of asynchronous time alignment and mind-body coupling time delay identification in step S5 of the present invention; Figure 4 This is a flowchart of the phase alignment adjustment and closed-loop operation control in step S6 of the present invention. Detailed Implementation

[0013] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only for explaining the present invention and are not intended to limit the present invention. This embodiment takes a 180-day home-based holistic management program for a group of patients with metabolic syndrome conducted by a chronic disease management center of a general hospital as an example, and illustrates the specific implementation of each step.

[0014] like Figure 1 As shown, the closed-loop operation control method for integrated mind-body health management of the present invention includes steps S1 to S6. These are elaborated below. Step S1: Constructing a patient profile and performing comprehensive stratification. This step acquires basic health data, physiological monitoring data, and psychological scale data of patients with metabolic syndrome to construct a patient profile. The basic health data includes the patient's age, gender, height, weight, waist circumference, and previous chronic disease diagnoses; the physiological monitoring data includes systolic blood pressure, diastolic blood pressure, fasting blood glucose, and blood lipid data measured at home and uploaded via the patient's terminal, wherein systolic and diastolic blood pressure are in mmHg, fasting blood glucose is in mmol / L, and blood lipids are in mmol / L; the psychological scale data includes anxiety and depression scale scores completed by the patient according to a preset period.

[0015] After obtaining the above data, this step comprehensively stratifies patients according to two dimensions: metabolic risk and psychological state. The metabolic risk dimension is based on whether systolic blood pressure, fasting blood glucose, waist circumference, and blood lipids exceed the corresponding clinical thresholds. The number of abnormal indicators is counted, and patients are divided into three metabolic risk levels: low, medium, and high. Specifically, patients with 0 to 1 abnormal indicator are classified into the low metabolic risk level, indicating mild metabolic disorders and a low risk of long-term complications. Patients with 2 abnormal indicators are classified into the medium metabolic risk level, indicating the coexistence of two abnormal metabolic indicators and belonging to the borderline high-risk group, requiring enhanced lifestyle intervention. Patients with 3 or more abnormal indicators are classified into the high metabolic risk level, indicating that the number of abnormal indicators meets the diagnostic criteria for metabolic syndrome, and their risk of cardiovascular disease and diabetes is significantly increased. The psychological state dimension is based on anxiety and depression scale scores, classifying patients into categories such as emotional stability, anxiety tendency, depression tendency, and social isolation tendency. The comprehensive stratification results, composed of metabolic risk level and psychological state category, are written into the patient profile. This patient profile is continuously updated in subsequent step S6 based on feedback from phase alignment adjustment results, thus maintaining a depiction of the patient's latest physical and mental state throughout the management cycle.

[0016] It should be noted that the low, medium, and high metabolic risk levels used in this step are a custom simplified stratification model for digital home-based chronic disease management. The stratification indicator cut-off points are strictly determined according to the metabolic abnormality judgment cut-off points recommended in the "Guidelines for the Prevention and Treatment of Metabolic Syndrome in Chinese Adults (2022 Edition, 2024 Edition)". The stratification and grading rules are designed independently based on the medical logic that the more metabolic abnormality indicators listed in the guidelines, the higher the risk of cardiovascular and cerebrovascular complications. It is not a clinical diagnosis and treatment standard stratification, but is only used for the automatic calculation of patient profiles and the matching of intervention intensity in this method.

[0017] It should also be noted that the metabolic syndrome risk stratification provided by clinical guidelines requires physician interpretation of information such as target organ damage and cardiovascular history for clinical diagnosis and treatment decisions. The stratification in this step, however, is a simplified stratification automatically calculated by the patient's terminal in a home setting. It is only used internally for patient profiling and intervention intensity matching, and its results are not considered as a disease diagnosis. The two methods are based on different information sources, have different applicable scenarios, and are used differently; therefore, there is no medical conflict between them.

[0018] In its implementation, the patient profile is organized as a multi-dimensional state record, containing both static attribute fields corresponding to basic health data and dynamic state fields corresponding to physiological monitoring data, psychological scale data, and comprehensive stratification results. Static attribute fields are written once upon patient enrollment, while dynamic state fields are updated throughout the management cycle with newly collected physiological monitoring and psychological scale data. The comprehensive stratification results are stored as a combination of metabolic risk level and psychological state category, such as a combination of high metabolic risk and anxiety, or a combination of moderate metabolic risk and depression, facilitating rapid retrieval of corresponding scenario-based educational content and art therapy salon formats by level and category in subsequent steps. It should be noted that the patient profile constructed in this step is not fixed once established, but rather serves as a feedback carrier for closed-loop operation control. It is repeatedly updated after receiving the phase alignment adjustment results in step S6, ensuring that the patient profile consistently reflects the patient's latest metabolic risk and psychological state throughout the entire management cycle.

[0019] The stratification used in this step is a routine assessment based on comparisons of multiple clinical thresholds and scale scoring. The thresholds for each threshold are the aforementioned stratification indicator cutoff points, strictly adhering to the metabolic abnormality assessment cutoff points recommended in the *Guidelines for the Prevention and Treatment of Metabolic Syndrome in Chinese Adults (2022 and 2024 Editions)*. For example, systolic blood pressure exceeding 130 mmHg, fasting blood glucose exceeding 6.1 mmol / L, and waist circumference exceeding 90 cm for men or 85 cm for women are used as the criteria for determining abnormal indicators. These will not be elaborated further here. The output of this step is a patient profile carrying the comprehensive stratification results, serving as the initial basis for selecting subsequent scenario-based educational content and matching it with the art therapy salon.

[0020] Step S2 involves triggering evangelism in a near-field communication scenario and collecting a sequence of evangelistic interaction behaviors as a signal for observing psychological and behavioral states. For example... Figure 2 As shown, this step involves distributing a portable near-field communication (NFC) device to the patient. This NFC device can be implemented as a smart card or wristband with a built-in NFC chip, pre-stored with the patient's identifier. Scene tags are then fixedly deployed in different indoor living scenarios within the patient's home. These scene tags are NFC tags that can be sensed by the NFC device, and each scene tag is associated with a scene category. The indoor living scenarios refer to fixed functional areas within the home. The tag deployment location, device ownership, and daily maintenance for these scenarios can all be determined and completed within the patient's home, eliminating the need for unified deployment of NFC tags in public areas. When a patient actively approaches a scene tag while carrying a near-field communication device, or when the patient's terminal's indoor positioning module (such as a Bluetooth beacon or UWB) detects that the patient has entered the scene area and stayed for more than a preset time (such as 120 seconds), the push of the scene education content should also be triggered and recorded as a 'passive trigger' interaction. The near-field communication device and the scene tag will communicate near-field to trigger the scene education content corresponding to the scene category, and the scene education content will be pushed to the patient's terminal for display.

[0021] Unlike existing timed mass messaging or passive reception-based education methods, this step not only pushes educational content to the patient's terminal but also collects a sequence of educational interaction behaviors as a signal for observing psychological and behavioral states. Specifically, each time the patient's terminal responds to a scenario-based educational content trigger, it records the active trigger time, scenario category, educational content reading completion rate, and consecutive trigger interval for that interaction. These interaction behavior data are then arranged chronologically to form the educational interaction behavior sequence. The active trigger time is the moment when near-field communication sensing occurs; the scenario category is the scenario category associated with the triggered scenario tag; the educational content reading completion rate is the percentage of educational content read by the patient's terminal, calculated by combining the scroll position reached and the dwell time on the patient's terminal page, with a value ranging from 0% to 100%; and the consecutive trigger interval is the difference between two adjacent active trigger times, expressed in hours.

[0022] Furthermore, this step analyzes the sequence of educational interaction behaviors, marking consecutive incomplete reading patterns or consecutive extended trigger intervals for the same scenario category as candidate psychological avoidance behavior markers. These candidate psychological avoidance behavior markers are then confirmed by the scenario transition topology correction in step S4 before being used as the degenerate components of the psychological behavior state observation signal. This is based on the fact that a patient with good behavioral compliance intentions typically actively triggers and fully reads the educational content in the corresponding scenario. However, when a patient's psychological avoidance motivation increases, statistically, it is more likely to manifest as consecutive rapid skimming, consecutive incomplete reading, or a gradually lengthening interval between active triggers of the same educational content. These behavioral correlations are empirical patterns already documented in health behavior compliance research and psychological intervention practice, and can serve as auxiliary observational evidence for psychological behavior states, rather than a sufficient and necessary causal determination. Therefore, if the completion rate of reading the educational content is consistently below the preset completion rate threshold (e.g., below 30% for three consecutive times), or if the interval between consecutive triggers gradually increases and exceeds the preset interval threshold (e.g., gradually increasing from the usual 12 hours to more than 72 hours), it is marked as a candidate psychological avoidance behavior marker to indicate a possible trend of regression in the patient's psychological and behavioral state.

[0023] It should be noted that the changes in the aforementioned interactive behaviors are not solely caused by psychological avoidance. Similar changes in interactive behavior can occur due to factors such as the patient being busy with daily tasks and not reading, the educational content not matching their current needs, the patient not having been to the scene that day, or equipment malfunctions in the near-field communication component or the patient's terminal. Therefore, this method does not directly determine the patient's psychological state based on a single or single-path interactive behavior. The candidate psychological avoidance behavior markers are merely auxiliary observation signals for psychological behavior states: the confusing situation of the patient not having been to the scene is distinguished and eliminated by the scene transfer topology in step S4; equipment malfunctions are excluded by the online status of the near-field communication component and heart rate records recorded by the patient's terminal; and occasional situations such as incompatible content and temporary busyness are suppressed by marking consecutive patterns of the same scene category rather than single events. The psychological avoidance behavior markers confirmed after the above corrections and eliminations still need to be merged into the psychological observation branch with the psychological scale data in step S5 to jointly participate in the identification of the mind-body coupling time lag, rather than being used as a separate conclusion for determining the patient's psychological state.

[0024] Therefore, this step reverses what existing technologies would consider negative interactions—such as failed or ineffective push notifications—into high-density auxiliary observation signals that reflect trends in the patient's psychological and behavioral state. This allows for the acquisition of auxiliary observations of the patient's psychological and behavioral state at near-daily frequency between two formal psychological scale assessments. It should be noted that these auxiliary observations do not replace psychological scale assessments, nor are they used alone to determine the patient's emotional state. Their role is to provide intermediate-state observations between the two formal assessments for the asynchronous time alignment in step S5. The output of this step is a sequence of educational interaction behaviors carrying candidate markers of psychological avoidance behavior, serving as one of the sources of psychological observation branches for subsequent asynchronous time alignment.

[0025] Based on specific scenarios, the different home living scenarios include a family kitchen scenario, a family dining room scenario, a family living room scenario, and a family bedroom scenario. Correspondingly, the scenario-based educational content includes home dietary education content corresponding to the family kitchen and family dining room scenarios, home light exercise education content corresponding to the family living room scenario, and home routine management education content corresponding to the family bedroom scenario. For example, when a patient actively triggers the scenario in the family kitchen, home dietary education content such as low-salt and low-fat diet combinations is pushed to the patient's terminal; when a patient actively triggers the scenario in the family dining room, home dietary education content such as controlling total food intake and adjusting the order of eating is pushed; when a patient actively triggers the scenario in the family living room, home light exercise education content suitable for patients with metabolic syndrome to perform indoors, such as stepping in place and stretching, is pushed; when a patient actively triggers the scenario in the family bedroom, home routine management education content such as regular routines and pre-sleep relaxation is pushed. All of the above scenario tags are fixedly deployed in the corresponding functional areas within the patient's home, and do not involve tag deployment, equipment ownership, or maintenance in public places. By matching scenarios with content, health education can reach patients in the scenarios where they are most likely to engage in the corresponding health behaviors, thereby improving the scenario relevance of health education.

[0026] Step S3 involves organizing an art therapy salon as a controlled perturbation of the psychological state and collecting perturbation response observation signals. This step involves organizing patients with metabolic syndrome to participate in art therapy salons. These art therapy salons are small-group, non-pharmacological psychological intervention activities, including four types: painting therapy salons, music therapy salons, dance therapy salons, and drama therapy salons. This step treats each art therapy salon as a controlled perturbation event applied to the patient's psychological dimension and collects physiological response data before and after the salon participation as perturbation response observation signals. Specifically, within a preset time window before and after the patient participates in the art therapy salon, physiological response data is collected, such as systolic blood pressure, diastolic blood pressure, and heart rate within 30 minutes before and 30 minutes after the salon. The physiological response data before and after the salon are paired to constitute the perturbation response observation signals.

[0027] The significance of treating art therapy salons as controlled disturbance events lies in the fact that the timing of the salons is known and controllable. Therefore, the physiological changes before and after the salons constitute a natural, controlled experiment of the transmission of psychological states to physiological indicators. By observing the disturbance response signals accumulated through participation in multiple salons, it is possible to reflect the direction and magnitude of the physiological responses of specific patients after their psychological states have been intervened. This provides observational data from physiological pathways for subsequently identifying the direction of individual mind-body coupling.

[0028] Specifically, for the same patient attending multiple art therapy salons within the management period, the time of each salon and the physiological response data of the paired individuals before and after each salon were recorded. These paired individuals were then arranged according to the salon's opening time to form the patient's perturbation response observation signal sequence. Because the salon schedules differ among patients and the intervals between salons for the same patient are irregular, this perturbation response observation signal sequence also exhibits an irregular, sparse sampling characteristic on the time axis. This characteristic will be processed along with other signals in the asynchronous time alignment step S5. The collection time window for the physiological response data before and after the salon is set within 30 minutes before the salon begins and 30 minutes after it ends, to ensure that the collected physiological responses closely approximate the immediate effect of the controlled perturbation of the salon, reducing interference from other irrelevant factors. It should be noted that, unlike home physiological monitoring data that relies on patient self-testing and has random sampling times, the perturbation response observation signal is an actively collected paired observation near a known perturbation time, thus more clearly reflecting the causal response of psychological state to physiological indicators.

[0029] Furthermore, this step matches the patient with an appropriate art therapy salon format based on the psychological state described in the patient profile. Specifically, music therapy salons are matched for anxiety states, painting therapy salons for depression states, and dance therapy salons and drama therapy salons for social isolation states (i.e., interpersonal withdrawal). Music therapy salons alleviate the tension of anxious patients through soothing music listening and ensemble playing; painting therapy salons provide a way for depressed patients to externalize their emotions through free painting and color expression; dance therapy salons create low-pressure social scenarios through small-group improvisational physical interaction, releasing suppressed emotions and improving the loneliness of socially isolated patients; drama therapy salons reduce the fear of social expression and increase the willingness to interact with others in socially isolated patients through small-group role-playing and situational expression. Dance therapy and drama therapy are both standard forms of expressive arts therapy, and their effectiveness in improving loneliness and interpersonal withdrawal has been documented in existing psychological intervention research and clinical practice. Both require only an indoor space and a group facilitator for regular implementation, without the need for specialized planting sites, materials, or maintenance conditions, making them convenient for weekly scheduling in chronic disease management centers. By matching psychological state with the salon format, controlled perturbations better align with the patient's psychological needs, enhancing the effectiveness of perturbation response observation signals. The output of this step is the paired perturbation response observation signals before and after salon participation, serving as one source for subsequent asynchronous time-aligned physiological observation branches.

[0030] Step S4: Construct a scene transfer topology and correct the sequence of educational interaction behaviors. This step constructs a scene transfer topology based on the trigger transfer relationships between multiple scene labels of the near-field communication component. Specifically, the scene category associated with each scene label is used as a node, and on this basis, 'virtual scene' nodes inferred from patient terminal sensor data are introduced, such as 'bedridden silent scene' (based on accelerometer and sleep monitoring) and 'screen immersion scene' (based on screen usage time and application type). The transfer relationship from the previous scene category to the next scene category in two adjacent active triggers is used as a directed edge, and the frequency of occurrence of each directed edge within the management cycle is counted to form the scene transfer topology. The nodes of the scene transfer topology only include scene categories corresponding to each functional area inside the home, such as the family kitchen scene, family dining room scene, family living room scene, and family bedroom scene, and do not involve outdoor public spaces.

[0031] The purpose of constructing the scene transition topology is to correct the candidate psychological avoidance behavior markers marked in step S2. When marking solely based on the sequence of educational interaction behaviors, a confusing situation arises: a patient may not actively trigger a scene, either because they passed through the scene but did not actively approach the scene tag due to psychological avoidance, or because they never actually passed through the scene. The former constitutes psychological avoidance behavior, while the latter should not be counted. This step uses the scene transition topology to distinguish between these: if, in the patient's scene transition topology, there is a high-frequency transition path between the preceding and subsequent scenes adjacent to an untriggered scene that passes through that untriggered scene, then it is determined that the patient passed through the scene but did not actively trigger it, and the candidate psychological avoidance behavior marker corresponding to this non-triggering is thus confirmed as a psychological avoidance behavior marker; conversely, if there is no transition path passing through the scene, it is determined that the patient did not pass through the scene, and the candidate psychological avoidance behavior marker corresponding to this non-triggering is removed and not counted as a psychological avoidance behavior marker. Furthermore, if the patient's terminal record shows that the near-field communication component was offline or malfunctioning during that period, the absence of a trigger will not be counted as a marker of psychological avoidance behavior, thus eliminating confusion caused by equipment malfunction. This avoids misjudging a patient's lack of experience with the scenario or equipment malfunction as psychological avoidance, transforming the labeling of psychological avoidance behavior markers from a single behavioral inference to a determination after scene transition topology correction, thereby improving the accuracy of psychological behavior state observation signals.

[0032] Furthermore, this step combines the scene transfer topology with virtual scene data, communication behavior data, and motion data to construct a behavioral diversity index, serving as a quantitative indicator of social isolation. This index integrates multi-dimensional data using methods such as principal component analysis; a lower value indicates a more singular behavioral pattern and a higher risk of social isolation. Specifically, it statistically analyzes the number of home indoor scene categories actually accessed by the patient in the scene transfer topology and the coverage of their transfer paths. When the number of accessed home indoor scene categories decreases over time, and the transfer paths shrink to a few scenes such as the family bedroom, it indicates a contraction in the patient's daily activity range at home, thus outputting a social isolation spatial behavior index. This social isolation spatial behavior index is an additional observation naturally generated during the correction of candidate psychological avoidance behavior markers. It can be used in step S3 to match dance therapy salons and drama therapy salons for social isolation status, and as a record of social isolation dimensions in subsequent health record archiving. The output of this step is a corrected sequence of educational interaction behaviors and a scene transfer topology.

[0033] Step S5 involves asynchronous time alignment and mind-body coupling delay and direction identification. For example... Figure 3 As shown, this step asynchronously aligns the corrected sequence of educational interaction behaviors, disturbance response observation signals, physiological monitoring data, and psychological scale data to identify the individualized mind-body coupling time lag and coupling direction of the patient.

[0034] Because the aforementioned signals originate from different acquisition methods, their sampling times are irregular and misaligned: the educational interaction behavior sequence is generated by the patient's initiative, with random sampling times; the disturbance response observation signal is generated according to the salon schedule; the physiological monitoring data is generated by the patient's self-test; and the psychological scale data is generated according to the review cycle. Before performing asynchronous time alignment, the physiological observation branch signals need to be 'corrected for drug effects'. The system should record all drug adjustment events (time, type, dosage) and correct the physiological monitoring data within a certain time window after drug adjustment according to a preset pharmacological model (such as drug half-life and expected effect curve), removing the systematic bias caused by the drug and making the remaining variation more purely reflect the mind-body coupling effect. This step first merges the aforementioned signals into two observation branches: the educational interaction behavior sequence carrying psychological avoidance behavior markers and the psychological scale data are merged into the psychological observation branch, and the disturbance response observation signal and the physiological monitoring data are merged into the physiological observation branch. Each branch is a discrete observation sequence with irregular sampling times. The observations within each branch are standardized to zero mean and unit variance to obtain standardized psychological and physiological observation branch signals, thus eliminating dimensional differences.

[0035] To address the irregular sampling times and the misalignment of the two branches, this step establishes a kernel-weighted irregular sampling cross-correlation, using time shift as the independent variable, to estimate the correlation between the psychological observation branch signal and the physiological observation branch signal. The kernel-weighted irregular sampling cross-correlation is calculated using the following formula: ,in: The irregular sampling cross-correlation weighted by the kernel function is denoted as , where is a scalar and its value range is . Dimensionless, derived from this formula with a given time shift. The calculations show that the signal in the psychological observation branch shifts relative to the physiological observation branch signal. The degree of correlation between the latter two branches; Let be a time shift quantity, and be a scalar quantity, with a value range of . The unit is d, which is obtained by discretely taking values ​​within the search interval and represents the amount of translation of the psychological observation branch signal relative to the physiological observation branch signal on the time axis. The upper limit of the time shift search is denoted by , which is a scalar with a value of 21 and a unit of days (d). It is determined based on the longest reasonable period for the transmission of psychological states in metabolic syndrome to physiological indicators, and characterizes the maximum range of the investigated mind-body coupling time delay. If the value is too small, it will not be able to cover the long coupling delay; if it is too large, it will introduce irrelevant spurious correlations. Here, is the sequence number of the psychological observation branch, and is a scalar with a value range of . to An integer, dimensionless, determined by the order of observations in the psychological observation branches, representing the psychological observation branches referenced. The total number of observations in the psychological observation branch is a scalar with a range of positive integers and no dimension. It is determined by the number of observations accumulated by the psychological observation branch within the management cycle and represents the number of psychological observations involved in the calculation. Here, is the sequence number of the physiological observation branch, and is a scalar with a value range of . to An integer, dimensionless, determined by the order of observations in the physiological observation branches, representing the physiological observation branches referenced. The total number of observations in the physiological observation branch is a scalar with a range of positive integers and is dimensionless. It is determined by the number of observations accumulated by the physiological observation branch within the management cycle and represents the number of physiological observations involved in the calculation. For the first The sampling time of each psychological observation branch is a scalar, with a value range of times within the management period, and the unit is d. It is determined by the occurrence time of the psychological observation and represents the position of the psychological observation on the time axis. For the first The sampling time of each physiological observation branch is a scalar, with a value range of times within the management cycle, and the unit is d. It is determined by the occurrence time of the physiological observation and represents the position of the physiological observation on the time axis. For the standardized psychological observation branch signal at time... The value of is a scalar, with a range of real numbers and no dimension. It is obtained by standardizing the observations of the psychological observation branch according to zero mean and unit variance, and represents the deviation of the patient's psychological and behavioral state from its mean at that moment. For the standardized physiological observation branch signal at time The value of is a scalar, with a range of real numbers and no dimension. It is obtained by standardizing the physiological observations by zero mean and unit variance, and represents the deviation of the patient's physiological indicators from their mean at that moment. Let be the Gaussian kernel function, defined as Its value range is Dimensionless, used to assign weights to irregularly sampled observation pairs based on their temporal proximity, so that the closer the time intervals are... The greater the contribution of the observation; Let be the independent variable of the Gaussian kernel function, and take in this formula. , is a scalar with units of d, representing the time residual between physiological observation and translated psychological observation; denoted as kernel bandwidth, and denoted as a scalar with a value of 1.5 in d, determined based on the average sampling interval of home physiological and psychological observations, characterizing the time smoothing scale of the kernel function. If the value is too small, the estimation variance will increase due to too few effective observation pairs; if the value is too large, the coupling delay will become fuzzy due to excessive smoothing. The function is an exponential function with a base of the natural constant and is dimensionless. After obtaining the irregular sample cross-correlation weighted by the kernel function, this step searches for a time shift that maximizes the time shift within the time shift search interval. This time shift is identified as the patient's individualized mind-body coupling time delay, and the coupling direction is determined by the sign of the cross-correlation at the corresponding time shift. The mind-body coupling time delay and coupling direction are determined by the following formula: , in: The time delay is the time delay between mind and body, and is a scalar with a range of values ​​of . The unit is d, which is obtained by taking the maximum point of irregular sampling cross-correlation weighted by the kernel function within the search interval. It represents the optimal alignment time shift between the patient's psychological observation branch signal and physiological observation branch signal, and its absolute value represents the time span of the interaction between psychology and physiology. This is the maximum point operator, which means returning the value of the independent variable that maximizes the objective function within a specified interval; The definition is the same as the aforementioned formula; , The definition is the same as the aforementioned formula; Let be the coupling direction, and be a scalar with a range of values. Dimensionless, determined by the sign of the time lag in mind-body coupling, it characterizes the causal sequence between the psychological and physiological processes. Psychological states precede physiological indicators; that is, psychological changes precede physiological changes. Physiological indicators precede psychological states; that is, physiological changes precede psychological states. The two characteristics do not show a significant temporal order. It is a sign function, taking values ​​for positive, zero, and negative values ​​respectively. , , , dimensionless. As can be seen from the above, under the condition that the sampling time is irregular and the two branches are not aligned, this step directly estimates the optimal alignment time shift between the psychological observation branch and the physiological observation branch, without the need to resample each signal at equal intervals, thus avoiding the problem of time delay information being destroyed by interpolation under sparse and irregular sampling.

[0036] To facilitate understanding of the mechanism of irregular sampling cross-correlation weighted by kernel function, the following explanation is provided. Under conventional equal-interval sampling conditions, the cross-correlation of two signals can be directly obtained by summing them point-by-point with a fixed step size. However, both the psychological and physiological observation branches in this invention are obtained from irregularly timed acquisitions. The observation times of the two branches are almost never exactly different by a fixed time shift. If they are forcibly resampled to an equal-interval grid, the interpolation process will generate data points out of thin air in the originally unobserved time periods, causing the estimated time delay to be biased. This step introduces a Gaussian kernel function to weight the observation pairs according to their temporal proximity. For a given time shift, the closer the physiological observation time is to the shifted psychological observation time, the greater its weight in the summation. In this way, the correlation between the two branches can be estimated on continuous time shifts without resampling, and the curve of the correlation changing with the time shift shows a maximum value at the actual mental-physical coupling time delay. The kernel bandwidth determines the aforementioned weighted time smoothing scale. A value that is too small will result in too few effective observation pairs participating in the summation for each time shift, increasing the estimation variance. A value that is too large will obscure the location of the maximum value of the correlation curve due to over-smoothing. Therefore, in this embodiment, the kernel bandwidth is set to 1.5 days, matching the average sampling interval of home physiological and psychological observations. The output of this step is the patient's individualized psychosomatic coupling lag and coupling direction, accompanied by a 'drug interference confidence level' label. This confidence level is calculated based on the frequency and amplitude of concurrent drug adjustment events. When the confidence level is lower than a preset threshold, the system should activate a conservative adjustment mode in step S6 as the basis for phase alignment adjustment in step S6.

[0037] Step S6: Phase alignment intervention timing adjustment and closed-loop operation control. (For example...) Figure 4 As shown, this step adjusts the timing of the push of scenario-based educational content and the scheduling of art therapy salons based on the mind-body coupling time lag and the coupling direction, and feeds back the phase alignment adjustment results to the patient profile, forming a closed-loop operation control of monitoring, evaluation, intervention and feedback.

[0038] To ensure closed-loop security, the system should have a built-in abnormal state detection and circuit breaker mechanism. When the system detects accelerated deterioration of physiological indicators, sudden changes in psychological state, or high-frequency oscillations in closed-loop adjustment, it should automatically trigger the circuit breaker, suspend automated adjustment, switch to manual takeover mode, and revert to the conservative default strategy until the administrator confirms the termination. This step first constructs a time-delayed directed mind-body coupling constraint, using the aforementioned mind-body coupling lag as a lead time. All adjustments to intervention timing must adhere to hard constraints on the magnitude of each adjustment; for example, the timing of scenario-based educational outreach should not be adjusted by more than ±3 days per adjustment, and the salon scheduling should not be adjusted by more than one cycle per adjustment. The time-delayed directed mind-body coupling constraint predicts future physiological observation branch signals based on current psychological observation branch signals, expressed as follows: ,in: For future moments The predicted value of the physiological observation branch signal is a scalar with a range of real numbers and no dimension. It is calculated by this formula based on the current psychological observation branch signal and represents the predicted deviation of the patient's physiological indicators from their mean after the time lag of mind-body coupling. The current time is a scalar, and its value range is the time within the management cycle. The unit is d, which is determined by the current running time of the system and represents the starting time of the prediction. The definition is the same as the formula mentioned above, serving as the lead time for prediction; The value of the standardized physiological observation branch signal at the current moment is a scalar with a range of real numbers and is dimensionless. It is obtained by standardizing the current physiological observation with zero mean and unit variance, and represents the deviation of the current physiological index from its mean. The definition is the same as the previous formula, and it serves as the direction coefficient of the coupling effect; Let be the mind-body coupling gain, and be a scalar with a range of values. Dimensionless, its value is the maximum value of the irregular sampling cross-correlation weighted by the kernel function at the aforementioned mind-body coupling time delay, i.e. The intensity of the transmission from the deviation of the psychological state to the deviation of the physiological indicator is identified by step S5. The larger the value, the stronger the conduction. The value of the standardized psychological observation branch signal at the current moment is a scalar with a range of real numbers and no dimension. It is obtained by standardizing the current psychological observation with zero mean and unit variance, and represents the deviation of the current psychological behavior state from its mean. The baseline value of the standardized psychological observation branch signal is a scalar, taking a value of 0, dimensionless, and determined by the mean of the standardized psychological observation branch signal, representing the neutral baseline of the psychological and behavioral state. Based on the time-delayed directed mind-body coupling constraint, this step further provides a phase-aligned intervention triggering decision. When the psychological and behavioral state observation signal shows a psychological state drift and the physiological monitoring data has not yet exceeded the limit, the push of scenario-based educational content or the scheduling of an art therapy salon is triggered in advance within the advance window determined by the mind-body coupling time delay. The intervention triggering decision is given by the following formula: , in: The intervention triggers the decision; is a scalar, and its value range is... Dimensionless, calculated by this formula based on three conditions: psychological state drift, physiological margin, and advance window, characterizing whether intervention is triggered at the current moment. Characterization triggers intervention, The representation does not trigger; The definition is the same as the aforementioned formula; is the psychological state drift quantity, which is a scalar with a range of non-negative real numbers and is dimensionless. It is determined by the absolute value of the deviation of the current standardized psychological observation branch signal from its recent baseline, and represents the drift amplitude of the psychological behavioral state. The psychological state drift threshold is a scalar with a value of 1.0. It is dimensionless and is determined based on the standard signal deviating from one standard deviation as a criterion for significant drift. It represents the minimum psychological drift amplitude required to trigger an intervention. Too small a value will lead to frequent triggering, while too large a value will lead to missed triggering. The definition is the same as the formula mentioned above; The physiological threshold is a scalar value of 2.0, dimensionless, determined by the deviation of the standardized physiological observation branch signal from two standard deviations as the criterion for exceeding the physiological limit. It represents the upper bound of the physiological index before it has deteriorated, and is only considered when... Intervention should only be initiated within the early intervention window before physiological indicators exceed the limit; The detection time of mental state drift is a scalar, with a value range of moments within the management period, and the unit is d. It is determined by the moment when the mental state drift first reaches the mental state drift threshold, and represents the starting point of the advance window. Let be the advance ratio coefficient, be a scalar, and take a value of 0.5, with a range of . Dimensionless, determined based on the strategy of intervening early by taking the midpoint between the psychological drift detection time and the predicted physiological limit exceedance time, characterizing the degree of advance of intervention relative to the predicted physiological limit exceedance time. The smaller the value, the greater the lead time. The definition is the same as the aforementioned formula; The advance window length, expressed in seconds (d), is the time window from the drift detection time. As can be seen from the formula, this step triggers intervention only when the patient's psychological and behavioral state has significantly drifted, but the physiological indicators have not yet exceeded their limits, and the current time falls within the advance window calculated from the drift detection time. Through the advance window determined by the mind-body coupling time lag, intervention can proceed before the physiological indicators show signs of deterioration, thus aligning the intervention timing with the individual's mind-body rhythm. In contrast, existing intervention schemes using a uniform fixed step length often mismatch the intervention timing with the individual's mind-body time lag, either failing to intervene in time when the psychological state has drifted but the physiological state has not yet deteriorated, or intervening only after the physiological state has deteriorated.

[0039] In addition, closed-loop operation control should also include compliance monitoring and mandatory wake-up sub-processes: if no active or passive scenario-based education triggering behavior is detected for N consecutive days, the system should automatically start the mandatory wake-up procedure, including but not limited to telephone reminders, timed and fixed-point push of core education content, and mark it as 'compliance crisis' to trigger manual intervention.

[0040] This step feeds back the phase alignment adjustment results, including the adjusted timing of scenario-based educational content delivery and the scheduling of art therapy salons, to the patient profile, updating the comprehensive stratification results and intervention arrangements in the patient profile. This creates a closed-loop control system from steps S1 to S6, connecting monitoring, assessment, intervention, and feedback. In this closed loop, the intervention acts both as an observation probe generating intermediate-state observations and as an intervention method driven by phase alignment adjustment, enabling the entire management process to continuously and adaptively adjust to changes in the patient's physical and mental state.

[0041] Specifically, within the management cycle, the aforementioned closed loop is repeatedly executed according to a preset operating cycle: In each operating cycle, steps S2 and S3 continuously collect interactive behavior data and disturbance response observation signals generated by scene-based educational interactions and participation in art therapy salons; step S4 corrects these data; step S5 updates the patient's individualized mind-body coupling lag and coupling direction accordingly; step S6 recalculates the advance window and adjusts the intervention timing based on the updated mind-body coupling lag, and then feeds the adjustment results back to the patient profile; this process is repeated continuously, so that the identification of mind-body coupling lag is continuously refined with the accumulation of observations, and the phase alignment of the intervention timing continuously approaches the patient's true mind-body rhythm. Thus, the closed loop formed by this invention is not a one-time open-loop adjustment, but an adaptive closed loop that continuously converges over time. The observations generated by the intervention action calibrate the mind-body coupling lag, and the mind-body coupling lag calibrates the timing of the intervention action. The two calibrate and promote each other, forming a synergistic closed loop that cannot be achieved by either link alone.

[0042] Furthermore, as a group-level enhancement to phase alignment adjustment, this embodiment statistically analyzes the distribution of mind-body coupling lags among patients within the group. When the distribution of these mind-body coupling lags exhibits a bimodal structure with zero lag as the boundary, patients whose coupling direction prioritizes psychological state over physiological indicators are classified as psychologically driven subtypes, and those whose coupling direction prioritizes physiological indicators over psychological state are classified as physiologically driven subtypes. Art therapy salons are prioritized for the psychologically driven subtypes, ensuring that interventions primarily target their leading psychological pathways; scene-based educational content is prioritized for the physiologically driven subtypes, ensuring that interventions primarily target their leading physiological behavioral pathways. This bimodal classification is a group structure that naturally emerges during the statistical analysis of individual mind-body coupling lags. Based on this, different intervention priorities are applied to the two subtypes, further enhancing the targeted nature of the intervention.

[0043] Furthermore, this embodiment archives the operational data of each stage of the closed-loop operation control of monitoring, assessment, intervention, and feedback according to individual and group dimensions to form an integrated physical and mental health record for the metabolic syndrome population. The operational data of each stage includes patient profiles and their comprehensive stratification results, sequences of educational interaction behaviors and markers of psychological avoidance behaviors, disturbance response observation signals, physical-mental coupling lag and direction, phase alignment adjustment results, and social isolation spatial behavior indicators. Based on the integrated physical and mental health record of the metabolic syndrome population, chronic disease management quality assessment indicators are output, such as the rate of improvement in physiological indicators within the group, the rate of improvement in psychological state, and the timing of intervention, providing data support for chronic disease management quality assessment and intervention program optimization.

[0044] Furthermore, this embodiment includes an outpatient follow-up reminder in the closed-loop operation control. When the improvement of the physiological monitoring data is below a first threshold and the trend of the psychological scale data shows continuous deterioration, an outpatient follow-up reminder is pushed to the patient's terminal. The improvement of the physiological monitoring data is determined by the magnitude of change in the current physiological monitoring data relative to the initial physiological monitoring data of the management cycle; the trend of change in the psychological scale data is determined by the trend of multiple consecutive psychological scale data. The outpatient follow-up reminder is only pushed when both insufficient physiological improvement and continuous psychological deterioration are simultaneously met, avoiding excessive disturbance caused by frequent reminders based on a single dimension.

[0045] To illustrate the effectiveness of this invention, statistical analysis was conducted on the operational data of a group of metabolic syndrome patients involved in this embodiment over a 180-day management period. The results showed that, compared to the control management method using unidirectional educational outreach and synchronous integrated assessment with fixed-step intervention, this invention increased the observation density of psychological and behavioral states between two formal psychological scale assessments from once a month in the control method to nearly once a day; after asynchronous alignment using kernel-weighted irregular sampling cross-correlation, the identifiable rate of mind-body coupling lag within the group reached 86.5%; after adjusting the intervention timing using phase alignment, the proportion of interventions before physiological indicators exceeded limits, i.e., the intervention timing hit rate, increased from approximately 41.2% in the control method to 78.6%; at the end of the management period, the improvement rate of key physiological indicators such as systolic blood pressure and fasting blood glucose within the group increased from approximately 52.4% in the control method to 71.3%, and the improvement rate of psychological scale scores increased from approximately 38.7% in the control method to 63.5%. The above data shows that by reversing the intervention action into an observation probe, identifying the individual's mind-body coupling time delay and aligning the intervention accordingly, the present invention significantly improves the level of precision in the integrated management of mind and body in people with metabolic syndrome without increasing the number of wearable physiological devices.

[0046] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A closed-loop operation control method for holistic mind-body health management, characterized in that: include: Patient profiles are constructed based on basic health data, physiological monitoring data, and psychological scale data of patients with metabolic syndrome. These patient profiles include comprehensive stratification results determined by metabolic risk and psychological status. The patient's personal near-field communication device triggers educational content in different home indoor living scenarios and pushes it to the patient's terminal. The triggering includes active triggering and passive triggering. Active triggering is triggered when the patient actively approaches the scene tag while carrying the near-field communication device. Passive triggering is triggered when the indoor positioning module of the patient's terminal detects that the patient has entered the scene area and stayed for more than a preset time. The educational interaction behavior sequence is collected as a psychological and behavioral state observation signal. The educational interaction behavior sequence is the interaction behavior data recorded by the patient's terminal in response to the triggering of the scene educational content. Organizing art therapy salons as controlled disturbances to psychological states, and collecting physiological response data before and after salon participation as disturbance response observation signals; The educational interaction behavior sequence, the disturbance response observation signal, the physiological monitoring data and the psychological scale data are asynchronously time-aligned to identify the individualized mind-body coupling time lag and coupling direction of the patient. Based on the aforementioned mind-body coupling time lag and coupling direction, the timing of pushing out scenario-based educational content and the scheduling of art therapy salons are phase-aligned and adjusted, and the phase alignment adjustment results are fed back to the patient profile, forming a closed-loop operation control of monitoring, evaluation, intervention and feedback.

2. The closed-loop operation control method for integrated physical and mental health management according to claim 1, characterized in that, The collection of educational interaction behavior sequences includes: recording the active triggering time, scene category, educational content reading completion rate, and continuous triggering interval of the near-field communication component in various home indoor living scenarios; marking the pattern of continuous incomplete reading or continuous triggering interval extension for the same scene category in the educational interaction behavior sequence as candidate psychological avoidance behavior markers, wherein the candidate psychological avoidance behavior markers are the degenerate components to be corrected in the psychological behavior state observation signal.

3. The closed-loop operation control method for integrated physical and mental health management according to claim 2, characterized in that, Also includes: A scene transfer topology is constructed based on the trigger transfer relationship of the near-field communication component among multiple scene tags; The scenario transition topology is used to distinguish between two scenarios: patients passing through a scenario without actively triggering it and patients not passing through a scenario. Candidate psychological avoidance behavior markers belonging to the scenario where patients passed through a scenario without actively triggering it are identified as psychological avoidance behavior markers, while candidate psychological avoidance behavior markers belonging to the scenario where patients did not pass through a scenario are eliminated to correct the degradation component of the psychological behavior state observation signal. The behavioral diversity index constructed by fusing the scenario transition topology with virtual scenario data, communication behavior data, and motion data is used as a quantitative indicator of social isolation. The virtual scenario data includes bedridden silent scenarios and screen immersion scenarios inferred from patient terminal sensor data.

4. The closed-loop operation control method for integrated physical and mental health management according to claim 3, characterized in that, The asynchronous time alignment includes: establishing a kernel-weighted irregular sampling cross-correlation for the educational interaction behavior sequence, the scene transition topology, the disturbance response observation signal, the physiological monitoring data, and the psychological scale data, which have irregular and misaligned sampling times; searching for a time shift that maximizes the kernel-weighted irregular sampling cross-correlation between the psychological observation branch and the physiological observation branch; identifying the time shift as the mind-body coupling time delay; and determining the coupling direction by the sign of the corresponding kernel-weighted irregular sampling cross-correlation.

5. The closed-loop operation control method for integrated physical and mental health management according to claim 4, characterized in that, Also includes: The distribution of the psychosomatic coupling time lags of each patient in the group was statistically analyzed. When the distribution of the mind-body coupling lag presents a bimodal structure with zero lag as the boundary, patients whose coupling direction is psychological state leading physiological indicators are classified as psychologically driven subtypes, and patients whose coupling direction is physiological indicators leading psychological state are classified as physiologically driven subtypes. Art therapy salons are prioritized for the psychologically driven subtype, and educational content for the scene is prioritized for the physiologically driven subtype.

6. The closed-loop operation control method for integrated physical and mental health management according to claim 5, characterized in that, The phase alignment adjustment includes: constructing a time-delayed directed mind-body coupling constraint with the mind-body coupling delay as the lead time, and applying the time-delayed directed mind-body coupling constraint to the closed-loop operation control; when the psychological behavior state observation signal shows psychological state drift and the physiological monitoring data has not yet exceeded the limit, the push of scene education content or the scheduling of art therapy salon is triggered in advance within the advance window determined by the mind-body coupling delay, so that the closed-loop operation control is transformed from a data-driven model to a model with mind-body causal time-delay constraints.

7. The closed-loop operation control method for integrated physical and mental health management according to claim 1, characterized in that, The art therapy salon includes painting therapy salon, music therapy salon, dance therapy salon and drama therapy salon; based on the psychological state in the patient profile, music therapy salon is matched for anxiety state, painting therapy salon is matched for depression state, and dance therapy salon and drama therapy salon are matched for social isolation state.

8. The closed-loop operation control method for integrated physical and mental health management according to claim 1, characterized in that, The different home indoor living scenarios include the family kitchen scenario, the family dining room scenario, the family living room scenario, and the family bedroom scenario; the scenario education content includes home diet education content corresponding to the family kitchen scenario and the family dining room scenario, home light exercise education content corresponding to the family living room scenario, and home routine management education content corresponding to the family bedroom scenario.

9. The closed-loop operation control method for integrated physical and mental health management according to claim 1, characterized in that, Also includes: The operational data of each link in the closed-loop operation control of monitoring, evaluation, intervention and feedback are archived according to individual and group dimensions to form a physical and mental integrated health record for people with metabolic syndrome. Based on the physical and mental integrated health record of people with metabolic syndrome, chronic disease management quality assessment indicators are output.

10. The closed-loop operation control method for integrated physical and mental health management according to claim 1, characterized in that, The closed-loop operation control also includes: when the improvement of the physiological monitoring data is lower than the first threshold and the trend of the psychological scale data is continuously deteriorating, pushing an outpatient follow-up visit reminder to the patient terminal.

Images