Music therapy intervention method and system for disease management of asthmatic children in OMO mode

Abstract

The application discloses a music therapy intervention method and system for asthma children disease management in an OMO mode, and belongs to the technical field of medical information processing and digital health intervention. The method comprises the following steps: generating an initial health portrait containing multi-dimensional health indicators through offline professional evaluation; collecting multi-modal physiological signals and symptom logs of asthma children through online continuous monitoring of a wearable device; in a cloud health management platform, determining the current course stage through an asthma state staging module, calculating a stress-relaxation state index through a physiological and psychological state analysis module, and generating a personalized music intervention prescription according to the stage and the index through a music acoustic parameter mapping module. The application solves the technical problems of poor targeting, insufficient individual adaptation and online and offline service fragmentation of existing music therapy schemes by constructing a pathology state driven acoustic parameter mapping model and an individualized online optimization loop.

Description

Technical Field

[0001] This invention relates to the field of medical information processing and digital health intervention technology, specifically to a music therapy intervention method and system for managing asthma in children under the OMO (Online-Merge-Offline) model. Background Technology

[0002] Asthma is the most common chronic inflammatory airway disease in childhood, with its core pathophysiology being variable airflow limitation and airway hyperresponsiveness. Effective disease management requires continuous and dynamic intervention for both "stable-period control" and "acute exacerbation warning." Traditional management models rely on periodic in-person outpatient visits and medication adjustments, which have inherent limitations such as large monitoring blind spots and difficulty in quantifying adherence.

[0003] With the development of digital health technology, the OMO (Online-Merge-Offline) model, which integrates online and offline services, provides a new paradigm for the long-term management of children with asthma through continuous monitoring and remote management via wearable devices. Meanwhile, music therapy, as a safe and easily accepted non-pharmacological intervention for children, is gradually being incorporated into asthma management as an adjunct approach due to its potential to regulate autonomic nerve function and influence respiratory rhythm through specific acoustic signals.

[0004] Current technological solutions that integrate music therapy into the OMO (Online-Merge-Offline) model have shortcomings in terms of precision, effectiveness, and safety of intervention.

[0005] Existing solutions mainly fall into two categories: one is fixed-schedule push based on standardized audio libraries. Its technical essence is a "time trigger," assuming that playing soothing music at fixed times has universal benefits, completely ignoring the dynamic fluctuations in asthma symptoms—a feedbackless open-loop system. The other is simple feedback regulation based on a single physiological signal (such as heart rate). Its essence is "univariate feedback control," but heart rate changes are non-specific and may originate from exercise, emotions, or asthma itself, which the system cannot distinguish. Neither type of solution establishes a causal mapping model between "specific pathological states of asthma" (such as bronchial tension, autonomic nervous system imbalance patterns driven by inflammation levels) and "precisely adjustable physical acoustic parameters" (beat, mode, harmonic complexity, etc.). The lack of targeted design for core asthma issues such as respiratory rhythm disorders and sympathetic over-excitation in music selection reduces intervention to a generalized relaxation attempt with random and uncertain effects.

[0006] Furthermore, existing technologies for adjusting acoustic parameters rely on preset general rules or limited combinations of experience. Their technical solutions lack an adaptive algorithm capable of learning online the differentiated physiological response functions of individual children to different acoustic parameters and performing closed-loop optimization accordingly. The system cannot automatically discover the most effective combination of acoustic parameters for a specific child through safe and orderly detection, nor can it dynamically adjust its strategy as the child grows, their condition evolves, or they adapt to their environment. This results in interventions remaining in a "one-size-fits-all" or "trial-and-error" state for a long time, failing to achieve truly personalized precision medicine.

[0007] Existing solutions often mechanically combine automated online recommendations with offline professional guidance. Online systems cannot deeply understand and utilize the precise staging and individual baseline determined by offline assessments; offline therapists also struggle to efficiently analyze the high-dimensional, multimodal physiological data continuously collected online. There is a lack of a unified, structured, and continuously updated "digital health profile" as the sole reliable source and technological hub for collaborative decision-making. This results in fragmented and even contradictory online and offline intervention strategies, failing to form a data-driven reinforcement loop of monitoring-decision-intervention-assessment, and weakening the synergistic effectiveness of the OMO (Online-Merge-Offline) model.

[0008] Therefore, how to overcome the above-mentioned technical bottlenecks and realize the targeted, adaptive and systematic application of music therapy in the management of children with asthma has become a key technical problem that urgently needs to be solved in this field. Summary of the Invention

[0009] The purpose of this invention is to provide a music therapy intervention method and system for the management of childhood asthma under the OMO model, which solves the technical problems of blind intervention, static nature and disconnected synergy caused by the lack of a causal mapping model between pathological characteristics and acoustic parameters in existing music therapy intervention programs for childhood asthma.

[0010] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows:

[0011] Music therapy interventions for managing childhood asthma under the OMO (Online-Merge-Offline) model include the following steps:

[0012] Offline professional assessment steps: Generate an initial health profile dataset for children with asthma and upload it to the cloud-based health management platform;

[0013] Online continuous monitoring steps: Collect respiratory waveform signals, heart rate variability characteristics, skin conductance level and body movement index of children with asthma through smart wearable devices, and receive symptom logs and environmental exposure information, and generate standardized physiological feature vectors for uploading;

[0014] Cloud-based intelligent decision-making steps: In the cloud-based health management platform, based on four features—respiratory rhythm variability coefficient, high-frequency heart rate variability power, skin conductance fluctuation amplitude, and body motion entropy value—the asthma state staging module determines the current disease stage as stable, warning, or pre-acute exacerbation phase according to preset rules and machine learning models; the physiological and psychological state analysis module calculates the stress-relaxation state index (SRI) using a weighted linear combination formula; and the music acoustic parameter mapping module, through its embedded acoustic-physiological coupling response rule set, generates a music intervention prescription containing five dimensions—beat rate, mode type, pitch range, harmonic complexity, and dynamic envelope slope—based on the current disease stage and the SRI value range.

[0015] Intervention implementation and feedback steps: The music intervention prescription is sent to the parent's mobile terminal for playback, and the data is sent back to the cloud health management platform to update the health profile;

[0016] Offline collaborative intervention steps: A structured treatment report is generated based on online data and pushed to the therapist's workbench. The therapist adjusts the offline intervention plan and updates the health profile based on the report.

[0017] Furthermore, the structured clinical assessment was conducted in accordance with the Global Initiative for Asthma (GINA) guidelines and standardized psychological scales.

[0018] Furthermore, the standardized physiological feature vector is obtained by processing respiratory waveforms, photoplethysmography (PPG) signals, skin conductance response (GSR) signals, and triaxial acceleration data, and includes respiratory rhythm variation coefficient, normalized power value of high-frequency heart rate variability, skin conductance fluctuation amplitude, and body motion entropy value.

[0019] Furthermore, the asthma state staging module employs rule-based reasoning and a machine learning classifier in collaboration to determine the stage of the disease.

[0020] Furthermore, the Stress-Relaxation State Index (SRI) is calculated through the following steps: the four characteristic values ​​of respiratory rhythm variability coefficient, high-frequency heart rate variability power, skin conductance fluctuation amplitude, and body motion entropy value are respectively subjected to minimum-maximum normalization based on the preset characteristic value range boundary of the corresponding age group; the four normalized characteristic values ​​are then used to calculate the SRI through a fixed weighted linear combination formula.

[0021] Furthermore, the music acoustic parameter mapping module has an embedded acoustic-physiological coupling response rule set, and the acoustic parameter configurations included in the music intervention prescription involve beat rate, mode type, pitch range, harmonic complexity, and dynamic envelope slope.

[0022] Furthermore, when the current stage of the disease is the warning period, the beat rate in the acoustic parameter configuration is no more than sixty beats per minute, and the mode type is limited to natural minor or pentatonic scale.

[0023] Furthermore, when the current disease stage is a stable period, the beat rate in the acoustic parameter configuration is seventy to eighty beats per minute, and the key type is major.

[0024] Furthermore, it also includes local emergency response procedures: when the respiratory rate is detected to be continuously exceeding the threshold and accompanied by a sudden drop in the high-frequency component of heart rate variability, it automatically switches to a preset breathing synchronization guidance track.

[0025] Furthermore, the audio specified in the music intervention prescription is derived from a clinically validated audio library.

[0026] Furthermore, the cloud-based health management platform employs a role-based access control mechanism and end-to-end encrypted transmission.

[0027] Furthermore, the structured treatment report includes online intervention execution rate, stress-relaxation state index (SRI) trend, respiratory rhythm variation coefficient box plot, symptom log heatmap, and abnormal event marking.

[0028] Furthermore, the workflow of the music acoustic parameter mapping module includes an individualized online acoustic parameter optimization loop.

[0029] Furthermore, the online acoustic parameter optimization loop includes the following steps: evaluating the average instantaneous gradient of the Stress-Relaxation State Index (SRI) after the initial music intervention prescription is executed; if its value is lower than a preset first gradient threshold, the optimization process is triggered.

[0030] Furthermore, the optimization process performs individual response detection, which sequentially changes a single dimension in the acoustic parameter vector; and based on the detection results, it updates online a sensitivity vector (s) characterizing the sensitivity of the acoustic parameters to the physiological response gradient; the k-th component of the sensitivity vector (s) The update formula is: ,in Let be the change in the k-th dimension parameter. .

[0031] Furthermore, based on the updated sensitivity vector (s), an optimized acoustic parameter adjustment direction is generated by solving a constrained linear optimization problem and then executed; the optimization problem is: finding the parameter adjustment amount. To maximize the objective function And it satisfies parameter boundary constraints and single adjustment range constraints.

[0032] In addition, this invention also discloses a music therapy intervention system for the management of childhood asthma under the OMO model, used to perform the music therapy intervention method for the management of childhood asthma under the OMO model as described above, including:

[0033] The offline assessment module is used to perform structured clinical assessments on children with asthma, generate an initial health profile dataset containing multi-dimensional health indicators, and upload it to the cloud-based health management platform.

[0034] The smart wearable device is used for continuous online monitoring, collecting respiratory waveform signals, heart rate variability characteristics, skin conductance levels and body movement index of children with asthma, and receiving symptom logs and environmental exposure information. After processing, it generates standardized physiological feature vectors and uploads them to the cloud-based health management platform.

[0035] The cloud-based health management platform includes:

[0036] The asthma state staging module is used to determine the current disease stage as stable, warning, or pre-acute exacerbation phase based on four features: respiratory rhythm variation coefficient, high-frequency heart rate variability power, skin conductance fluctuation amplitude, and body motion entropy value, according to preset rules and machine learning models.

[0037] The physiological and psychological state analysis module is used to calculate the stress-relaxation state index using a weighted linear combination formula.

[0038] The music acoustic parameter mapping module has an embedded acoustic-physiological coupling response rule set, which is used to generate a music intervention prescription with five dimensions, including beat rate, mode type, pitch range, harmonic complexity and dynamic envelope slope, based on the current disease stage and the numerical range of the stress-relaxation state index.

[0039] The parent's mobile terminal is used to receive and execute the music intervention prescription, play the corresponding audio, and send the execution data back to the cloud-based health management platform to update the health profile;

[0040] The therapist's workbench is used to receive structured treatment reports generated based on online data, allowing therapists to adjust offline intervention plans and update health profiles based on the reports.

[0041] Compared with the prior art, the present invention has the following beneficial effects:

[0042] Existing music selection techniques often rely on therapists' subjective experience or general soothing music libraries, resulting in weak correlation with the real-time pathological state of children with asthma. This invention establishes a decision-making logic based on objectively determined disease stages and quantified stress-relaxation indices, by constructing and coordinating an asthma state staging module, a physiological and psychological state analysis module, and a music acoustic parameter mapping module. Based on clinical guidelines and multimodal physiological signals, it accurately distinguishes different stages such as stable and warning periods. Through an embedded acoustic-physiological coupling response rule set, specific pathophysiological states are mapped to precisely defined combinations of acoustic parameters, including beat rate, mode type, and harmonic complexity. This transforms music intervention from an art dependent on personal experience into a targeted regulatory technology driven by the child's real-time physiological and pathological data, with precisely reproducible and adjustable parameters, thus improving the objectivity, repeatability, and clinical efficacy of intervention measures.

[0043] Existing solutions, with their fixed rules or preset content, cannot adapt to individual differences. This invention introduces a personalized online acoustic parameter optimization loop. When the initial intervention response is insufficient, it can automatically and safely initiate parameter probing and learn the individual child's acoustic-physiological response sensitivity model (Jacobi matrix) online based on real-time feedback data. This allows for the dynamic calculation of optimized parameter combinations that better enhance physiological benefits. This represents a leap from general group rules to real-time individual adaptation, ensuring the long-term effectiveness and accuracy of the intervention plan.

[0044] This invention eliminates the service fragmentation common in OMO (Online-Merge-Offline) models, constructing a data-driven collaborative closed loop centered on health profiles. In traditional models, online automated push notifications and offline professional guidance often operate independently. This invention seamlessly connects online intelligent decision-making and offline professional intervention through an initialized and continuously updated unified health profile. Structured treatment reports generated in the cloud provide offline therapists with comprehensive data-driven decision support, while new offline assessment conclusions are instantly fed back to the health profile to optimize the online model. This bidirectional, real-time feedback based on the same data source truly achieves logical consistency and synergistic effects across intervention strategies, forming a management synergy and significantly improving the coherence and effectiveness of overall disease management.

[0045] This invention deploys an emergency response module at the edge, independent of cloud communication. When a specific physiological signal pattern indicating a risk of acute deterioration is detected, it can immediately and autonomously activate a pre-defined respiratory synchronization guidance track without network support, providing immediate basic physiological regulation support for the child. This mitigates the risk of response delays or interruptions in extreme situations associated with cloud systems that rely on network connections, adding a crucial safety layer to the entire intervention plan.

[0046] This invention introduces a stress-relaxation state index and its changing trend analysis, enabling the effect of music intervention to be quantitatively evaluated in a continuous numerical form. Combined with an online acoustic parameter optimization mechanism, it achieves dynamic adaptive adjustment of the intervention strategy. Attached Figure Description

[0047] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained from these drawings without creative effort.

[0048] Figure 1 This is a block diagram of the overall system architecture of the present invention.

[0049] Figure 2 This is a block diagram of the internal modules of the cloud-based health management platform of the present invention.

[0050] Figure 3 This is an overall flowchart of the method described in this invention.

[0051] Figure 4 This is a diagram of the system login interface described in this invention.

[0052] Figure 5 This is a graph showing the effect of the Stress-Relaxation State Index (SRI) intervention on children with asthma, as described in this invention.

[0053] Figure 6 This is a graph showing the change in the coefficient of variation of respiratory rhythm under music intervention in the system of this invention.

[0054] Figure 7 This is a chart showing the stress-relaxation state index (SRI) characteristics of the system of the present invention at different stages of disease progression.

[0055] Figure 8 This is a comparative analysis of the SRI change trends before and after the optimization of the acoustic parameters of the system of the present invention. Detailed Implementation

[0056] In the following description, only certain exemplary embodiments are briefly described. As those skilled in the art will recognize, the described embodiments can be modified in various ways without departing from the spirit or scope of the embodiments of the invention. Therefore, the drawings and description are considered to be exemplary in nature and not restrictive.

[0057] The following is combined with Figures 1-8 The embodiments of the present invention will be described in detail below.

[0058] Example 1: See Figures 4-8This embodiment discloses a music therapy intervention system for the management of childhood asthma under the OMO (Online-Merge-Offline) model, including:

[0059] The offline assessment module is used to perform structured clinical assessments on children with asthma, generate an initial health profile dataset containing multi-dimensional health indicators, and upload it to the cloud-based health management platform.

[0060] The smart wearable device is used for continuous online monitoring, collecting respiratory waveform signals, heart rate variability characteristics, skin conductance levels and body movement index of children with asthma, and receiving symptom logs and environmental exposure information. After processing, it generates standardized physiological feature vectors and uploads them to the cloud-based health management platform.

[0061] The cloud-based health management platform includes:

[0062] The asthma state staging module is used to determine the current disease stage as stable, warning, or pre-acute exacerbation phase based on four features: respiratory rhythm variation coefficient, high-frequency heart rate variability power, skin conductance fluctuation amplitude, and body motion entropy value, according to preset rules and machine learning models.

[0063] The physiological and psychological state analysis module is used to calculate the stress-relaxation state index using a weighted linear combination formula.

[0064] The music acoustic parameter mapping module has an embedded acoustic-physiological coupling response rule set, which is used to generate a music intervention prescription with five dimensions, including beat rate, mode type, pitch range, harmonic complexity and dynamic envelope slope, based on the current disease stage and the numerical range of the stress-relaxation state index.

[0065] The parent's mobile terminal is used to receive and execute the music intervention prescription, play the corresponding audio, and send the execution data back to the cloud-based health management platform to update the health profile;

[0066] The therapist's workbench is used to receive structured treatment reports generated based on online data, allowing therapists to adjust offline intervention plans and update health profiles based on the reports.

[0067] See Figures 1-3 The music therapy intervention method for managing childhood asthma under the OMO model includes the following steps:

[0068] Offline professional assessment steps: Pediatricians and music therapists conduct structured clinical assessments of children with asthma, collecting basic lung function indicators, frequency of acute attacks, history of allergen exposure, medication adherence records, and psychological scale scores to generate an initial health profile dataset, which is then encrypted and uploaded to the cloud-based health management platform.

[0069] Online continuous monitoring steps: A multi-channel physiological sensor array in a smart wearable device is used to collect real-time respiratory waveform signals, heart rate variability characteristics, skin conductance levels, and body movement index of children with asthma. Symptom logs and environmental exposure information recorded by parents are also received. The multi-channel physiological sensor array includes a piezoelectric breathing band, a photoplethysmography (PPG) sensor, a skin conductance response (GSR) electrode, and a triaxial accelerometer. The collected signals are preprocessed to generate standardized physiological feature vectors, which are then fused with the symptom logs and environmental exposure information and uploaded to the cloud-based health management platform.

[0070] Cloud-based intelligent decision-making steps: In the cloud-based health management platform, based on the health profile dataset, the standardized physiological feature vector, the symptom log, and the environmental exposure information, the current disease stage is determined through the asthma state staging module, the stress-relaxation state index is calculated through the physiological and psychological state analysis module, and a unique music intervention prescription is generated through the music acoustic parameter mapping module based on the current disease stage and the stress-relaxation state index. The music intervention prescription specifies the audio material library index, playback duration, start time window, and acoustic parameter configuration. The acoustic parameters include tempo, mode type, pitch range, harmonic complexity, and dynamic envelope slope.

[0071] Intervention execution and feedback steps: The music intervention prescription is sent to the parent's mobile terminal and local audio playback is triggered; at the same time, the prescription execution record and real-time physiological feedback data are sent back to the cloud health management platform to update the health profile and optimize subsequent intervention strategies;

[0072] Offline collaborative intervention steps: A structured treatment report is generated based on the online intervention data within the period and pushed to the therapist's workbench; the therapist adjusts the offline intervention plan according to the report and writes the new assessment conclusions into the health profile, completing the data loop.

[0073] Furthermore, the structured clinical assessment specifically includes: a comprehensive evaluation based on the diagnostic and staging criteria for childhood asthma in the Global Initiative for Asthma (GINA) guidelines, combined with the Childhood Anxiety and Depression Scale and the Childhood Behavioral Scale; the collected assessment data includes forced expiratory volume in one second (FEV1), forced vital capacity (FVC), the ratio of FEV1 to FVC, the number of acute exacerbations in the past twelve months, known allergen exposure history, daily medication use records and adherence scores, tobacco exposure in the home environment, and self-reported sleep quality; the initial health profile dataset is encapsulated in JSON format and includes the child's identity identifier, clinical stage label, physiological baseline parameters, psychological state score vector, and environmental risk factor weight matrix.

[0074] In the online continuous monitoring step, the preprocessing unit built into the smart wearable device performs real-time preprocessing on the collected signals, specifically including:

[0075] 1) Respiratory signal processing: Real-time filtering based on moving average (window width 200ms) and zero-crossing detection are used to extract the respiratory cycle. Coefficient of variation of respiratory rhythm The coefficient of variation was calculated using a formula based on 20 consecutive respiratory cycles. ,in The periodic standard deviation, The periodic average;

[0076] 2) PPG signal processing: The first-order differential thresholding method is used to detect the R-wave and calculate the interval between adjacent R-waves. Heart rate variability high-frequency power normalized value By calculating the past 5 minutes The power spectral density percentage of the sequence in the 0.15-0.4 Hz frequency band was obtained;

[0077] 3) GSR ​​signal processing: Calculate the mean skin conductance level every 10 seconds. Skin conductance fluctuation amplitude Defined as Standard deviation over the past 2 minutes;

[0078] 4) Acceleration processing: Calculate the vector magnitude of triaxial acceleration. Entropy of body movement Through calculation The approximate entropy of the sequence within a 20-second window is obtained, with an embedding dimension of m=2 and a tolerance of r=0.2×standard deviation.

[0079] All feature values ​​are uploaded to the cloud every 30 seconds.

[0080] Furthermore, the workflow of the asthma state staging module includes: calling the rule reasoning submodule to perform preliminary screening based on the criteria for determining symptom control levels in the Global Initiative for Asthma (GINA) guidelines; the criteria include daytime symptom frequency, nighttime symptoms, activity limitation, and frequency of emergency medication use; if all stable phase conditions are not met, the machine learning classifier submodule is activated; the machine learning classifier adopts a gradient boosting decision tree architecture, and the input features include the average respiratory rate over the past 72 hours, the proportion of low-frequency power in heart rate variability, the weighted score of symptom logs, and the cumulative value of environmental risk factors, and the output is one of three mutually exclusive labels: stable phase, warning phase, or pre-acute exacerbation phase.

[0081] Based on continuous monitoring data analysis, the Stress-Relaxation State Index (SRI) exhibits different typical value ranges at different disease stages. The SRI during the stable period is generally in the higher range, the SRI during the warning period is in the middle range, while the SRI in the pre-acute exacerbation stage is in the lower range and fluctuates significantly. These range characteristics are used to assist in the determination of disease stages and the selection of music intervention strategies in the cloud-based health management platform.

[0082] The steps for calculating the Stress-Relaxation Index (SRI) by the physiological and psychological state analysis module include:

[0083] Input feature preparation: Obtaining the coefficient of variation of respiratory rhythm High-frequency heart rate variability power normalization value Skin conductance fluctuation amplitude Body entropy value The current values ​​of the four features.

[0084] Feature normalization: The min-max normalization method is used to normalize each feature value. Map to the [0,1] interval:

[0085] ;

[0086] in, and This refers to the reasonable range boundaries of this characteristic determined based on statistical data of children with asthma in the target age group (6-12 years old), for example: The range is [0.05, 0.5]. The range is [0.1, 0.8]. The range is [0.02, 0.5]. , The range is [0.1, 2.0].

[0087] SRI calculation: This involves normalizing the eigenvectors. Enter the following formula:

[0088] .

[0089] in, The preset fixed weight coefficients, and satisfy the following conditions: As a preferred implementation, take .

[0090] The SRI value ranges from [0, 1]. A higher value indicates that parasympathetic nerve activity is more dominant and the relaxation state is better.

[0091] Furthermore, the music acoustic parameter mapping module embeds an acoustic-physiological coupling response rule set based on clinical controlled trial data. The rule set defines the specific music acoustic parameter combinations to be activated under different disease stages and different stress-relaxation state index ranges. Among the acoustic parameters, the beat rate is divided into four levels: no more than 60 beats per minute, 61 to 70 beats per minute, 71 to 80 beats per minute, and more than 80 beats per minute. The mode type is divided into three categories: natural minor, pentatonic scale, and major. The pitch range is divided into three levels based on middle C: one octave above and below, one and a half octaves above and below, and two octaves above and below. The harmonic complexity is divided into three levels: low, medium, and high according to the proportion of non-fundamental frequency component energy. The dynamic envelope slope is divided into three categories: slow rise and fall, medium, and rapid.

[0092] When the asthma stage classification module determines that the current disease stage is the warning period, the music acoustic parameter mapping module forcibly enables the following acoustic parameter configurations: the beat rate is no more than 60 beats per minute, the mode type is limited to natural minor or pentatonic scale, the pitch range is compressed to within two octaves above and below middle C, the harmonic complexity is set to low level, and the dynamic envelope slope adopts a slow rise and slow fall mode.

[0093] Furthermore, when the asthma state staging module determines that the current disease stage is stable, the music acoustic parameter mapping module enables the following acoustic parameter configurations: beat rate of 70 to 80 beats per minute, mode type expanded to major key system, pitch range widened to 1.5 octaves above and below middle C, and harmonic complexity set to medium level.

[0094] Furthermore, it also includes local emergency response procedures:

[0095] In the edge computing unit within the smart wearable device or parent's mobile terminal, respiratory rate and high-frequency components of heart rate variability are monitored in real time. When the respiratory rate exceeds the age-corrected threshold for five consecutive minutes and the high-frequency components of heart rate variability decrease by more than 50% within ten minutes, an acute deterioration risk is identified. Subsequently, without waiting for cloud instructions, the local audio playback module is directly triggered to switch to a preset respiratory synchronization guidance track. The respiratory synchronization guidance track consists of alternating inhalation and exhalation prompts with a fixed beat. The interval between the beats is preset to four or six seconds according to the child's age, corresponding to an inhalation-to-exhalation ratio of 1:2.

[0096] The audio materials in the music intervention prescription are pre-recorded by a team of professional music therapists and have undergone double-blind clinical validation. Each audio segment corresponds to a fixed set of acoustic parameters and has a stable physiological regulatory effect under specific pathological conditions. The audio material library supports dynamic expansion. After newly recorded audio is clinically validated, its acoustic parameter feature vector and corresponding physiological response curve are used to update the acoustic-physiological coupling response model.

[0097] Furthermore, the cloud-based health management platform adopts a role-based access control model, defining four roles: guardian of the child, registered music therapist, pediatrician, and public health administrator. Health profile data is only accessible to guardians of the child, registered music therapists, and pediatricians, and differentiated data read and write permissions are assigned according to their roles. All data transmission is end-to-end encrypted, and the audio prescription generation process is recorded in a complete audit log.

[0098] The structured treatment report generated during the offline collaborative intervention steps includes the online intervention execution rate within the cycle, the daily average trend line of the stress-relaxation state index, the box plot of the respiratory rhythm variation coefficient, the symptom log heatmap, and abnormal event markers. The abnormal event markers include a single stress-relaxation state index drop of more than 0.3 units or a respiratory rate that is consistently higher than the threshold for more than 30 minutes. The report is automatically pushed to the therapist's workbench through the interface of the hospital information system.

[0099] To facilitate a better understanding of the present invention by those skilled in the art, the following detailed description is provided in conjunction with specific embodiments.

[0100] A music therapy intervention method for managing childhood asthma under the MO model relies on the organic synergy of four functional modules: offline professional assessment, online continuous monitoring, cloud-based intelligent decision-making, and local emergency response, forming a data-driven closed-loop system with a health profile of children with asthma as the central hub.

[0101] Starting with the overall system architecture, the following sections detail the specific composition of each component, signal processing flow, core algorithm logic, and intervention execution mechanism to ensure that those skilled in the art can fully reproduce the technical solution of this invention based on this specification.

[0102] During the offline professional assessment phase, a structured clinical assessment of the target children with asthma is conducted jointly by licensed physicians qualified to diagnose and treat respiratory diseases and certified registered music therapists. The assessment process strictly follows the diagnostic and staging criteria for childhood asthma in the Global Initiative for Asthma (GINA) guidelines and incorporates standardized psychological assessment tools such as the SCARED (Childhood Anxiety and Depression Scale) and the CBCL (Childhood Behavioral Scale) for comprehensive evaluation.

[0103] The assessment includes, but is not limited to: basic lung function indicators (such as forced expiratory volume in one second (FEV1), forced vital capacity (FVC) and its ratio FEV1 / FVC), number of acute exacerbations in the past twelve months, known allergen exposure history, daily medication use records and adherence scores, household tobacco exposure, and self-reported sleep quality. All raw assessment data are structured and entered into an initial health profile dataset, which includes the child's identification identifier, clinical stage label, baseline physiological parameters, psychological status scores, and environmental risk factor data. This dataset is uploaded to the cloud-based health management platform via an encrypted channel that meets the national information security level protection requirements (Level 3) and serves as the benchmark for generating subsequent online intervention strategies.

[0104] During the online continuous monitoring phase, children with asthma wear smart wearable devices that integrate multi-channel physiological sensor arrays. These smart wearable devices are deployed in the form of flexible wristbands or chest patches to ensure comfort during long-term wear and stability of signal acquisition.

[0105] The multi-channel physiological sensor array specifically includes: a piezoelectric breathing band for continuously acquiring respiratory waveform signals generated by combined chest and abdominal movements; a photoplethysmography (PPG) sensor for acquiring raw heart rate and heart rate variability (HRV) signals based on transmission or reflection optical principles; impedance-type skin conductance response (GSR) electrodes for measuring changes in skin conductance caused by sympathetic nerve activity; and a triaxial microelectromechanical system (MEMS) accelerometer for recording body movement amplitude, frequency, and direction information. The analog signals output by each sensor are sampled by an independent channel analog-to-digital converter (ADC), with sampling rates set at 20 Hz for respiratory signals, 100 Hz for PPG signals, 50 Hz for GSR signals, and 50 Hz for acceleration signals, with a quantization bit depth of no less than 12 bits. The digitized raw signals are transmitted to a local edge computing unit via Bluetooth Low Energy 5.0, which can be built into the wearable device itself or a mobile terminal carried by the parent.

[0106] The edge computing unit performs primary signal preprocessing operations, and the processing flow includes the following steps:

[0107] Wavelet transform denoising and baseline drift correction are performed on the respiratory waveform signal to extract the respiratory cycle, inspiratory-to-expiratory ratio and respiratory rhythm variation coefficient;

[0108] Adaptive filtering is applied to PPG signals to eliminate motion artifacts. The heart rate interval sequence is extracted from the preprocessed signal, and then the time-domain indices (such as SDNN, RMSSD) and frequency-domain indices (such as low-frequency power LF, high-frequency power HF and their ratio LF / HF) of heart rate variability are calculated.

[0109] The GSR signal was smoothed using a sliding window, and the mean skin conductance level (SCL) and the frequency of skin conductance response (SCR) events per unit time were calculated.

[0110] Vector synthesis is performed on the triaxial acceleration data, the movement index (MI) is calculated, and its complexity is further evaluated using the sample entropy algorithm, denoted as the movement entropy (ME).

[0111] The above processing results are integrated into a standardized physiological feature vector. ,in This is the coefficient of variation of respiratory rhythm. This represents the normalized value of high-frequency HRV power as a percentage of total power. The amplitude of skin conductance fluctuation is represented by , and ME is the body entropy value. This vector is fused with symptom logs (including cough frequency, wheezing perception intensity, and number of nighttime awakenings) and environmental exposure information (such as PM2.5 concentration, pollen index, and indoor humidity) synchronously recorded on the parent's mobile terminal to form a multi-source input data packet, which is then encrypted and uploaded to the cloud-based health management platform via HTTPS protocol.

[0112] Within the cloud-based health management platform, three interconnected core processing modules are deployed:

[0113] The system comprises three modules: an asthma state staging module, a physiological and psychological state analysis module, and a music acoustic parameter mapping module. The asthma state staging module receives the child's historical health profile and current multi-source input data. It first calls the rule-based reasoning submodule to perform a preliminary screening based on the GINA guidelines for symptom control levels (e.g., daytime symptom frequency ≤ 2 times / week, no nighttime symptoms, no activity limitation, and emergency medication use ≤ 2 times / week). If all stable phase conditions are met, it is directly labeled as "stable phase"; otherwise, the machine learning classifier submodule is activated. This classifier uses a gradient boosting decision tree (GBDT) architecture. Input features include the average respiratory rate over the past 72 hours, the proportion of low-frequency HRV, the weighted score of symptom logs, and the cumulative value of environmental risk factors. The output is one of three mutually exclusive labels: stable phase, warning phase, or pre-acute exacerbation phase. During the training phase, the classifier uses a labeled dataset from a multi-center clinical cohort, with a sample size of no less than 3000 children with asthma. Each case includes at least 30 days of continuous monitoring records and physician-confirmed disease stage labels, with a cross-validation accuracy of no less than 92%.

[0114] The physiological and psychological state analysis module receives data from the respiratory rhythm variation coefficient. High-frequency heart rate variability power normalization value Skin conductance fluctuation amplitude Body entropy value Standardized physiological feature vectors .

[0115] in:

[0116] Defined as: the denoised respiratory cycle sequence Calculate its coefficient of variation, i.e. ,in The standard deviation of the periodic series, This is the mean.

[0117] Defined as: the power in the high-frequency band (0.15-0.4 Hz) of the heart rate variability power spectrum. With total power The ratio (usually defined as ≤0.4 Hz) is, i.e. .

[0118] Defined as: skin conductance level within a time window The standard deviation.

[0119] Defined as: the composite vector amplitude sequence of a triaxial accelerometer The entropy value is calculated using the sample entropy algorithm, with the parameter set to: embedding dimension. tolerance ( (where is the standard deviation of the sequence).

[0120] The module constructs a Stress-Relaxation Index (SRI) based on the theory of autonomic nervous system function regulation. It employs a group-wide boundary-based approach. , Mini-maximum normalization and weighted linear combination formula Perform calculations, where The function is determined based on the direction of the feature's contribution to the relaxed state (e.g., for...). and Since the contribution direction is negative, we take... )

[0121] The SRI value ranges from 0 to 1. A higher value indicates a greater dominance of parasympathetic activity and a deeper state of relaxation in the child; conversely, a lower value suggests overactivation of the sympathetic nervous system and a potential stress response. This index is used as a continuous variable input to the next stage to drive fine-tuning of acoustic parameters.

[0122] Furthermore, in practical use, the music acoustic parameter mapping module embeds a pre-trained acoustic-physiological coupling response model. This model is based on a prospective randomized controlled trial that included 400 asthmatic children aged 6-12 years, who were randomly assigned to different acoustic parameter combination intervention groups. Each group received standardized music intervention for four weeks, five times a week, for twenty minutes each time. During the intervention, the SRI change trajectory was continuously recorded, and a mixed-effects model was used to analyze the marginal effect of each acoustic parameter on SRI. Finally, five key acoustic parameter dimensions were identified: beats per minute (BPM), mode type, pitch range, harmonic complexity, and dynamic envelope slope. Each dimension is divided into several discrete levels: BPM is divided into four levels: ≤60, 61–70, 71–80, and >80; mode type is divided into three categories: natural minor, pentatonic scale, and major; pitch range is based on middle C (C4) and is divided into three levels: ±1 octave, ±1.5 octave, and ±2 octave; harmonic complexity is divided into three levels according to the proportion of non-fundamental frequency components in the Fourier spectrum: low (<15%), medium (15%–30%), and high (>30%); dynamic envelope slope refers to the rate of amplitude change between the start and end segments of the audio signal and is divided into three categories: gradual rise / fall (rise / fall time ≥ 3 s), medium (1–3 s), and rapid (<1 s).

[0123] The acoustic-physiological coupled response model operates using the aforementioned decision function. In specific implementation, for the "warning period," the decision function adopts the following rules:

[0124] like Then the preset "deep relaxation" parameter combination will be output. BPM ≤ 60, mode type is natural minor, pitch range is one octave above and below middle C, harmonic complexity is low, dynamic envelope slope is gradually rising and falling;

[0125] like The output will be the "moderate guidance" parameter combination. BPM ≤ 60, mode type is pentatonic, pitch range is two octaves above and below middle C, harmonic complexity is low, and dynamic envelope slope is gradually rising and falling. For the "stable period", the decision function adopts another set of rules: if The output will show the "maintain stability" parameter combination. The BPM is 70-80, the key type is major, the pitch range is 1.5 octaves above and below middle C, the harmonic complexity is medium, and the dynamic envelope slope is medium.

[0126] like The output will be the "actively maintain" parameter combination. The BPM is 70-80, the key is major, the pitch range is two octaves above and below middle C, the harmonic complexity is high, and the dynamic range slope is medium. This rule set and its parameter combinations... to It has been proven effective in pre-clinical trials, for example, in children in the early warning stage. Fifteen minutes after the combined intervention, the average SRI increased. The number of units (mean ± standard deviation) (95% confidence interval [0.24, 0.32]) was measured, and the respiratory depth increased by 12.7%. When the condition was determined to be in the pre-acute phase, the system immediately stopped all active music playback and switched to passive white noise masking mode. This mode outputs pink noise at 40–60 dB SPL with a center frequency of 1 kHz to shield against sudden ambient noise and reduce the risk of bronchospasm induced by the startle reflex. At the same time, the system automatically pushes emergency treatment guidelines to the parents' devices, including instructions for use of inhaled short-acting β2 receptor agonists, suggestions for body position adjustment, and navigation links to the nearest secondary or higher-level medical institutions.

[0127] Based on the collaborative calculation results of the above three modules, a unique music intervention prescription is automatically generated. This prescription exists in the form of a structured data object, containing fields such as: audio material library index (UUID format), playback duration (unit: seconds), start time window (UTC timestamp), and acoustic parameter configuration vector (including specific values ​​such as BPM, mode, and pitch).

[0128] The audio library, pre-recorded by a team of professional music therapists according to evidence-based medicine principles, contains 240 original audio tracks, each strictly controlled to be between 18 and 22 minutes in length, with a sampling rate of 48 kHz and a bit depth of 24 bits, stored in a cloud-based object storage service. Each audio track underwent double-blind clinical validation before being added to the library, confirming a statistically significant positive moderating effect on SRI under specific pathological conditions (p < 0.01). The generated prescription is sent to the parent's mobile terminal application via a two-way authenticated TLS 1.3 secure channel, triggering the local audio playback module to execute playback. During playback, the system continuously transmits the current physiological feature vector and playback progress markers in real time, used to update the short-term response curve in the health profile online.

[0129] In the offline intervention phase, the system automatically generates structured treatment reports at three time granularities: weekly, monthly, and quarterly. These reports are in PDF / A-3 format, embedding interactive charts and covering: the online intervention execution rate within the cycle (actual playback counts / planned playback counts), the daily average SRI trend line, a box plot of the respiratory rhythm coefficient of variation, a symptom log heatmap, and abnormal event markers (such as a single SRI drop exceeding 0.3 units or a respiratory rate consistently above the threshold for more than 30 minutes). This report is automatically pushed to the therapists' workstations at contracted medical institutions via the hospital information system's HL7 FHIR interface. During face-to-face interventions, therapists adjust their breathing training programs (such as diaphragmatic breathing rhythm and pursed-lip breathing duration), emotional guidance techniques (such as situational dialogues designed based on recent SRI low periods), and family interaction suggestions (such as recommending specific parent-child music games to enhance the intervention effect) based on this report, and write the new assessment conclusions into the health profile, completing the data loop.

[0130] Furthermore, the edge computing unit incorporates a local emergency response module that operates independently of the cloud communication link. When the respiratory rate exceeds the age-corrected threshold (6–8 years: >30 breaths / min; 9–12 years: >25 breaths / min) for five consecutive minutes and is accompanied by a decrease of more than 50% in the high-frequency component of HRV (HF power) within 10 minutes, the system determines that there is a risk of acute deterioration. Without waiting for cloud instructions, it directly triggers the local audio to switch to a preset respiratory synchronization guidance track. This track consists of alternating inspiratory cues (440 Hz pure tone, lasting 1.5 seconds) and expiratory cues (330 Hz pure tone, lasting 3.0 seconds) with a fixed beat interval preset to four seconds (6–8 years) or six seconds (9–12 years), corresponding to an inspiratory-to-expiratory ratio of 1:2. The acoustic parameters of this track are fixed in the read-only storage area of ​​the device firmware and cannot be modified, deleted, or overwritten by the user, ensuring that basic physiological regulation support can still be provided in the event of communication interruption, network latency, or cloud service unavailability.

[0131] The cloud-based health management platform adopts a role-based access control (RBAC) model, defining four main roles: child guardian, registered music therapist, pediatrician, and public health administrator.

[0132] Health profile data is only accessible to the first three roles, following the principle of least privilege: guardians can only view their own children's data and prescription execution records; therapists can access the full data of the children under their care but cannot modify clinical staging labels; physicians have full read and write permissions. All data transmission uses AES-256-GCM end-to-end encryption, with the key negotiated and generated using the national cryptographic algorithm SM2. The audio prescription generation process records a complete audit log, including operation time, operator identity, input data hash value, and output prescription fingerprint. The log is retained for no less than 10 years, complying with the relevant requirements of the Personal Information Protection Law and the Guiding Principles for Medical Device Software Registration Review. The platform connects with regional public health information systems through standardized APIs, automatically reporting asthma management quality indicators, including annual acute exacerbation rate, medication adherence rate, and number of unplanned visits, supporting disease control departments in conducting group-level intervention effectiveness assessments.

[0133] In a preferred implementation, the music acoustic parameter mapping module incorporates an individualized calibration mechanism when generating prescriptions. Within 72 hours of the initial intervention, the child's SRI response slope to the initial acoustic configuration is analyzed. If the response slope is lower than the population mean by one standard deviation, the harmonic complexity or dynamic envelope slope of the next cycle is automatically fine-tuned until the optimal individual response point is found. This mechanism is implemented using a Bayesian optimization algorithm, with the objective function being a weighted sum of maximizing the SRI increment and minimizing the playback interruption rate, ensuring that the intervention is both effective and sustainable.

[0134] Furthermore, the audio library supports dynamic expansion. Once newly recorded audio passes double-blind clinical validation, its acoustic parameter feature vectors and corresponding SRI response curves are injected into the training set of the acoustic-physiological coupled response model, triggering online incremental learning of the model. This process employs a federated learning framework, where each medical institution updates its locally trained model, only uploading encrypted gradients to the central server for aggregation, thus avoiding the flow of original patient data across institutions and balancing model evolution with privacy protection.

[0135] This invention constructs a dynamic intervention mechanism based on pathological stage identification, multimodal physiological and psychological state quantification, and an acoustic parameter-physiological response mapping model. Under an OMO (Online-Merge-Offline) architecture, it achieves deep collaboration between precise online delivery and professional offline services, ensuring that each music intervention is strictly matched to the actual physiological needs of asthmatic children. The entire technical solution is characterized by high modularity, clear data flow, traceable algorithms, and a complete emergency response mechanism, providing a complete technical paradigm for the standardized and intelligent application of non-pharmacological interventions in the management of chronic respiratory diseases.

[0136] Example 2: This example is a further optimization based on Example 1. In this example, the workflow of the music acoustic parameter mapping module includes an individualized online acoustic parameter optimization loop, which specifically includes the following steps:

[0137] S131 Initial Response Assessment: The initial period following the execution of the initial music intervention prescription generated based on the acoustic-physiological coupling response rule set. Internally, the stress-relaxation state index is calculated in real time. and calculate exist Average instantaneous gradient within ;

[0138] S132 Optimization Trigger Judgment: The average instantaneous gradient With the preset positive gradient threshold Compare; if Then determine the current combination of acoustic parameters. It did not elicit an effective physiological regulatory response, thus failing to trigger the individualized parameter optimization process;

[0139] S133 Individual Response Detection and Learning: Execute the individual response detection strategy, sequentially combining at least one alternative acoustic parameter. Applied to a short detection period and calculate each Corresponding physiological response gradient Simultaneously, based on the detection data, a sensitivity vector representing the current individual acoustic-physiological response sensitivity of the child is updated online. ;

[0140] S134 Optimal Parameter Decision and Execution: Based on the Updated Sensitivity Vector Calculate the acoustic parameter adjustment direction that maximizes the desired physiological response gradient. ;in accordance with For the current parameter combination Adjustments are made to generate an optimized combination of acoustic parameters. And apply it to the subsequent main intervention period.

[0141] In step S131, the average instantaneous gradient The calculation formula is:

[0142] ;

[0143] in, At the start of the intervention, To calculate the difference time interval of the instantaneous gradient, for The number of difference calculations within the time period satisfies The first gradient threshold The range of values ​​is to .

[0144] In step S133, the individual response sensitivity model is implemented using a sensitivity vector. To approximate this, the vector estimates the exponential gradient of the stress-relaxation state near the current operating point. For acoustic parameter vectors An approximately linear response to unit changes in each dimension. The acoustic parameter vector... ,in For the beat rate, The numerical code corresponding to the mode type. This represents the numerical value corresponding to the pitch range (unit: octave). This represents the percentage value corresponding to harmonic complexity. This represents the numerical value corresponding to the dynamic envelope slope.

[0145] The sensitivity vector The online update rules are as follows:

[0146] 1) Updates are performed only when a single continuous numerical parameter (BPM, pitch range R, harmonic complexity H) is changed during detection;

[0147] 2) For the two categorical variables, tune type M and dynamic envelope slope E, sensitivity learning is not performed; selection is made directly based on the preset rule set.

[0148] 3) The update formula for continuous parameters is:

[0149] ;

[0150] in:

[0151] Application test parameter combination Afterwards, during the detection period The average gradient of the stress-relaxation state index observed internally;

[0152] It is a prediction of gradient change based on the sensitivity before the update;

[0153] For the first The change in the dimension parameter during this detection (e.g., BPM changes from 75 to 65, then...) );

[0154] 4) Parameter boundary: learning rate The initial values ​​of each component of the sensitivity vector are 0.

[0155] If multiple parameters are changed simultaneously during the detection process, the detection data will not be used to update the sensitivity vector. This is used only for comparing response gradients. The sensitivity vector... The initial value can be set to a vector of all zeros.

[0156] In specific implementation, the direction of acoustic parameter adjustment is calculated in step S134. The process is a constrained optimization problem:

[0157]

[0158] The constraints are:

[0159] ;

[0160] ;

[0161] in, and These are vectors consisting of the physical lower and upper limits of each acoustic parameter dimension, respectively. This is the step scaling factor, and its value range is... to ; This represents the maximum allowable parameter change in a single optimization, used to ensure the safety and stability of the optimization process. Let be the infinite norm of the vector.

[0162] In practice, the specific implementation is as follows:

[0163] 1. System initialization and parameter settings:

[0164] The system presets key parameters for online loop optimization: initial evaluation period. Seconds (5 minutes); instantaneous gradient difference interval seconds; first gradient threshold This threshold, determined based on a statistical analysis of baseline physiological variability in 200 children with asthma, represents the lower limit of the clinically significant rate of change in SRI; short detection period Seconds (3 minutes); learning rate Step scaling factor Parameter variation constraints The settings are as follows: BPM variation not exceeding ±10, tuning coding. Pitch range Variation not exceeding ±1 level, harmonic complexity percentage Variation not exceeding ±10%, dynamic envelope slope encoding The change shall not exceed ±1 level.

[0165] 2. Optimize the loop operation process:

[0166] Taking an intervention session of an 8-year-old child with stable asthma (patient ID: PA_001) as an example. The initial health profile shows a baseline SRI of approximately 0.50. The cloud-based asthma staging module determines that the child is currently in a "stable phase," and the physiological and psychological state analysis module calculates the current SRI to be 0.52. The music acoustic parameter mapping module generates an initial parameter combination based on the original rule set (see the original instruction manual's example). : , (The code corresponds to "major key") (The encoding corresponds to "1.5 octaves above and below Middle C") (This corresponds to 20%, i.e., "intermediate" harmonic complexity). (The encoding corresponds to a "medium" envelope slope). That is... .

[0167] Step S131 (Initial Response Assessment): Parent's mobile device starts playing. The corresponding music. The system is in The moment begins Within seconds, every The instantaneous gradient of SRI is calculated once per second. The measured SRI data sequence is as follows: The average instantaneous gradient is calculated according to the formula, where... :

[0168] Step S132 (Optimize trigger judgment): The system determines the initial parameters. Failure to elicit a valid physiological response (SRI increases too slowly) immediately triggers a personalized parameter optimization process. Current parameters The current sensitivity vector is initialized to a zero vector. .

[0169] Step S133 (Individual Response Detection and Learning: First Detection): The system performs the first detection. To explore a better BPM in the current stable state, an alternative parameter combination slightly below the lower limit of the rule recommendation range is generated. ), other parameters remain unchanged: that is , .application to Seconds. During the detection period, the SRI varied from 0.531 to 0.538. The observed average response gradient. .

[0170] According to the definition in the sensitivity vector update formula Since this detection only changed the first parameter (BPM, i.e.) ),and Therefore, the gradient is predicted. .

[0171] According to the sensitivity vector update formula, since this probe only changed the first parameter (BPM), only the sensitivity vector needs to be updated. The first component The updated formula is: .

[0172] Substitute the data from this probe into the calculation:

[0173]

[0174] After the update, the sensitivity vector After this update, The small negative value indicates that, under the current conditions, reducing BPM may not have resulted in the expected rapid increase in SRI.

[0175] Step S133 (Individual Response Detection and Learning: Second Detection): Based on the current sensitivity vector The system decided to conduct a second probe, shifting its focus to adjusting another important parameter: harmonic complexity. .

[0176] Generate alternative parameter combinations :Will Increased from 0.20 to 0.30 ( The BPM is restored to 75, while other parameters remain unchanged. , .

[0177] application Up to 180 seconds. During the detection period, the SRI changed significantly from 0.538 to 0.550. Observation gradient .

[0178] Since this detection only changed the fourth parameter ( Therefore, only the sensitivity vector is updated. The fourth component Predicting gradients .

[0179] According to the updated formula:

[0180]

[0181] After the update, the sensitivity vector .

[0182] At this point, the sensitivity vector The system shows that, for the current patient PA_001, increasing harmonic complexity ( It has a strong positive sensitivity to increasing the SRI gradient.

[0183] To ensure the robustness and security of the online learning process, multiple security constraints are built into the optimized loop:

[0184] Valid update criterion: Only when the observed absolute value of the gradient is... Greater than the physiological noise threshold (For example, Only when this happens will the sensitivity vector be processed. Updates are made to avoid mislearning caused by instantaneous fluctuations in the signal.

[0185] Confidence accumulation: The system is a sensitivity vector Maintain a confidence counter for each component When a probe is determined to be a valid update, the confidence level of the corresponding component is... Increase; conversely, decrease. The final sensitivity value used for decision-making is... ,in This is a preset confidence threshold (e.g., 3). This mechanism ensures that the learning results are based on multiple valid observations.

[0186] Rollback mechanism: If consecutive Next (for example) None of the probes triggered a valid update (i.e., If the system determines that the current individualized optimization cannot obtain stable benefits, it will automatically revert to the preset parameter combination generated by the acoustic-physiological coupling response rule set and generate a "requires offline review" flag to be pushed to the therapist's workbench.

[0187] Step S134 (Optimal Parameter Decision and Execution: Generating an Optimal Prescription): After two probes, the system compares the observed response gradients: , The latter is significantly higher than the former, and System decision-making adopts As the optimized parameter combination .

[0188] The cloud-based music acoustic parameter mapping module generates new music intervention prescriptions, specifying the use of... The matched audio material will continue to play until the end of this intervention session. Meanwhile, and the corresponding sensitivity vector The health profile of the child is stored as initial parameters and sensitivity prior knowledge for the next intervention session.

[0189] The Stress-Relaxation State Index (SRI), based on features such as high-frequency heart rate variability (HF-HRV) and respiratory rhythm variability coefficient, is an internationally recognized quantitative indicator for assessing autonomic nervous system function. One of the core pathophysiological mechanisms of asthma is autonomic nervous system imbalance (sympathetic dominance / parasympathetic inhibition). Therefore, improving SRI through intervention using musical acoustic parameters essentially involves direct regulation of underlying pathological aspects of asthma. Specifically, beat rate guides respiratory rhythm towards stability through a "rhythmic synchronization" effect; specific modalities and harmonic complexity influence emotions and stress responses through the auditory cortex-limbic system pathway. The "pathological state → acoustic parameter" mapping model and online optimization loop established in this invention integrates the aforementioned dispersed physiological principles into a measurable, feedback-enabled, and adjustable precision intervention system for the first time, providing a reproducible technical path for the effectiveness of digital therapy in chronic disease management.

[0190] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including both the preferred embodiments and all changes and modifications falling within the scope of the invention.

[0191] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. It should be noted that 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 music therapy intervention method for managing childhood asthma under the OMO (Online-Merge-Offline) model, characterized in that... Includes the following steps: Offline professional assessment steps: Generate an initial health profile dataset for children with asthma and upload it to the cloud-based health management platform; Online continuous monitoring steps: Collect respiratory waveform signals, heart rate variability characteristics, skin conductance level and body movement index of children with asthma through smart wearable devices, and receive symptom logs and environmental exposure information, and generate standardized physiological feature vectors for uploading; Cloud-based intelligent decision-making steps: In the cloud-based health management platform, based on four features—respiratory rhythm variation coefficient, high-frequency heart rate variability power, skin conductance fluctuation amplitude, and body motion entropy value—the asthma state staging module determines the current disease stage as stable, warning, or pre-acute exacerbation phase according to preset rules and machine learning models; the physiological and psychological state analysis module calculates the stress-relaxation state index (SRI) using a weighted linear combination formula; and the music acoustic parameter mapping module, through its embedded acoustic-physiological coupling response rule set, generates a music intervention prescription containing five dimensions—beat rate, mode type, pitch range, harmonic complexity, and dynamic envelope slope—based on the current disease stage and the SRI value range. Intervention implementation and feedback steps: The music intervention prescription is sent to the parent's mobile terminal for playback, and the data is sent back to the cloud health management platform to update the health profile; Offline collaborative intervention steps: A structured treatment report is generated based on online data and pushed to the therapist's workbench. The therapist adjusts the offline intervention plan and updates the health profile based on the report.

2. The music therapy intervention method for managing childhood asthma under the OMO model according to claim 1, characterized in that, The structured clinical assessment was conducted in accordance with the Global Initiative for Asthma (GINA) guidelines and standardized psychological scales.

3. The music therapy intervention method for managing childhood asthma under the OMO model according to claim 1, characterized in that, The standardized physiological feature vector is obtained by processing respiratory waveforms, photoplethysmography (PPG) signals, skin conductance response (GSR) signals, and triaxial acceleration data, and includes respiratory rhythm variation coefficient, normalized power value of high-frequency heart rate variability, skin conductance fluctuation amplitude, and body motion entropy value.

4. The music therapy intervention method for managing childhood asthma under the OMO model according to claim 1, characterized in that, The asthma state staging module uses rule-based reasoning and machine learning classifiers to work together to determine the stage of the disease.

5. The music therapy intervention method for managing childhood asthma under the OMO model according to claim 1, characterized in that, The Stress-Relaxation State Index (SRI) is calculated through the following steps: the four characteristic values ​​of respiratory rhythm variability coefficient, high-frequency heart rate variability power, skin conductance fluctuation amplitude, and body entropy value are respectively subjected to minimum-maximum normalization based on the preset characteristic value range boundaries of the corresponding age group. The SRI is obtained by calculating the four normalized eigenvalues ​​using a fixed weighted linear combination formula.

6. The music therapy intervention method for managing childhood asthma under the OMO model according to claim 3, characterized in that, The music acoustic parameter mapping module contains an embedded set of acoustic-physiological coupling response rules. The acoustic parameter configurations included in the music intervention prescription involve beat rate, mode type, pitch range, harmonic complexity, and dynamic envelope slope.

7. The music therapy intervention method for managing childhood asthma under the OMO model according to claim 6, characterized in that, When the current disease stage is the warning period, the beat rate in the acoustic parameter configuration is no more than sixty beats per minute, and the mode type is limited to natural minor or pentatonic scale.

8. The music therapy intervention method for managing childhood asthma under the OMO model according to claim 6, characterized in that, When the current disease stage is stable, the beat rate in the acoustic parameter configuration is seventy to eighty beats per minute, and the key type is major.

9. The music therapy intervention method for managing childhood asthma under the OMO model according to claim 1, characterized in that, It also includes local emergency response steps: when the respiratory rate is detected to be continuously exceeding the threshold and accompanied by a sudden drop in the high-frequency component of heart rate variability, it automatically switches to a preset breathing synchronization guidance track.

10. A music therapy intervention system for managing childhood asthma under the OMO model, used to perform the music therapy intervention method for managing childhood asthma under the OMO model as described in any one of claims 1-9, characterized in that, include: The offline assessment module is used to perform structured clinical assessments on children with asthma, generate an initial health profile dataset containing multi-dimensional health indicators, and upload it to the cloud-based health management platform. The smart wearable device is used for continuous online monitoring, collecting respiratory waveform signals, heart rate variability characteristics, skin conductance levels and body movement index of children with asthma, and receiving symptom logs and environmental exposure information. After processing, it generates standardized physiological feature vectors and uploads them to the cloud-based health management platform. The cloud-based health management platform includes: The asthma state staging module is used to determine the current disease stage as stable, warning, or pre-acute exacerbation stage based on four features: respiratory rhythm variation coefficient, high-frequency heart rate variability power, skin conductance fluctuation amplitude, and body motion entropy value, according to preset rules and machine learning models. The physiological and psychological state analysis module is used to calculate the stress-relaxation state index using a weighted linear combination formula. The music acoustic parameter mapping module has an embedded acoustic-physiological coupling response rule set, which is used to generate a music intervention prescription with five dimensions, including beat rate, mode type, pitch range, harmonic complexity and dynamic envelope slope, based on the current disease stage and the numerical range of the stress-relaxation state index. The parent's mobile terminal is used to receive and execute the music intervention prescription, play the corresponding audio, and send the execution data back to the cloud-based health management platform to update the health profile; The therapist's workbench is used to receive structured treatment reports generated based on online data, allowing therapists to adjust offline intervention plans and update health profiles based on the reports.

Images