Abdominal pressure dynamic change-based recessive gastroesophageal reflux event data analysis method and system

Abstract

The invention provides a recessive gastroesophageal reflux event data analysis system and method based on dynamic changes of abdominal pressure. The system and method are suitable for noninvasive detection of atypical reflux clinically. The system comprises an abdominal pressure dynamic detection unit, a throat vibration detection unit, a multi-mode signal processing module, a reflux event judgment module and a data output module. According to the system, by analyzing time sequence linkage between abdominal pressure mutation and throat vibration, a stomach content countercurrent event is recognized, silent type countercurrent characteristic sound is captured with the assistance of a bone conduction sound wave channel, and multi-path redundancy recognition is achieved. The method is suitable for postoperative patients, children, old people and other people who cannot complain, has the technical advantages of continuous wearing, behavior correlation analysis and clinical verifiability, and significantly improves the recognition rate and diagnosis accuracy of recessive gastroesophageal reflux events.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of medical devices and physiological signal detection, and specifically relates to a method and system for analyzing data of an implicit gastroesophageal reflux event based on dynamic changes in abdominal pressure. BACKGROUND

[0002] Gastroesophageal reflux disease (GERD) is a common digestive system disease, and long-term reflux can cause esophagitis, laryngitis, asthma, and even cancer. In actual clinical work, doctors are increasingly concerned about a group of patients who do not have typical symptoms but may have serious pathological reactions, i.e., the "implicit gastroesophageal reflux" population.

[0003] Such patients have the following significant clinical features and monitoring difficulties: 1. No subjective symptoms or vague complaints, making it extremely difficult to locate the problem based on medical history: Some patients, especially infants, the elderly, postoperative drowsy patients, and patients with nervous system disorders, cannot accurately express feelings such as "acid reflux" and "heartburn"; doctors often have to rely on indirect manifestations such as pharyngitis, coughing, and hoarseness to make inferences, which can easily lead to misdiagnosis or missed diagnosis.

[0004] 2. The reflux process is short and irregular in frequency, and non-acid reflux is often missed: Low-pressure reflux in a resting state may not have significant changes in intragastric pressure and is not accompanied by strong throat irritation. Traditional pH monitoring and impedance methods have very low sensitivity for detecting "low-amplitude, intermittent, and non-acid reflux".

[0005] 3. There is a lack of widely applicable dynamic monitoring equipment in clinical practice: Current mainstream methods such as 24-hour pH monitoring or multi-channel impedance detection require nasal intubation, which has poor patient tolerance, cannot be worn continuously for several days, and is not suitable for children and postoperative patients. In addition, the monitoring data are fragmented and costly, which is not conducive to long-term tracking and individualized assessment.

[0006] 4. "Silent reflux" is difficult to capture, becoming a diagnostic blind spot: Especially during sleep, in a supine position, or in a low muscle tone state after surgery, reflux may quietly ascend to the pharynx, only forming a subtle "wet rale" or "gurgling sound" at the glottis, without causing muscle vibration or significant changes in abdominal pressure. Doctors cannot detect it during routine examination or auscultation, and it is easily overlooked.

[0007] In summary, the identification of implicit reflux events, especially in atypical symptom populations, special body positions, children, and the elderly, has become a major difficulty and gap in the current management of gastroesophageal reflux disease.

[0008] Therefore, a new system is urgently needed, which is clinically wearable, easily accepted by patients, pathologically redundant, and has recognition logic closer to physiological mechanisms, and can realize real-time sensing of dynamic changes in abdominal pressure, capture throat muscle vibration response, identify non-muscle-derived "wet sound" signals caused by reflux, and establish a "cause-effect" event correlation mechanism, thereby improving the accuracy, specificity and clinical verifiability of the identification of silent reflux. SUMMARY

[0009] The present application aims to provide a silent gastroesophageal reflux event data analysis method and system based on dynamic changes in abdominal pressure, to realize non-invasive and full-pathway dynamic identification and statistics of reflux events, and to significantly improve the identification ability of "silent, atypical and transient" reflux, which has important clinical application value and promotion prospects.

[0010] The technical solutions adopted by the present application are as follows: The silent gastroesophageal reflux event data analysis system based on dynamic changes in abdominal pressure comprises: An abdominal pressure dynamic detection unit is used to collect low-frequency micro-pressure fluctuation signals related to changes in abdominal pressure in real time in the patient's body surface area and form an abdominal pressure change data sequence; A throat vibration detection unit is arranged at the patient's throat position and is used to synchronously collect vibration signals caused by muscle contraction or glottis closure in the throat; A multi-modal signal processing module is used to synchronously process the abdominal pressure signals and the throat vibration signals, including filtering, feature extraction, time alignment and linkage analysis; A reflux event determination module is used to determine a silent gastroesophageal reflux event according to a set timing logic within a certain time range after the abdominal pressure fluctuation signal appears, if the corresponding throat vibration signal is detected; A data storage and output module is used to save event records, count reflux frequency, generate behavior labels and periodic risk maps, and provide prompts or output results to terminal devices.

[0011] The abdominal pressure dynamic detection unit comprises a flexible micro-pressure sensor or a strain gauge structure arranged at the lower edge of the neck or above the clavicle, and is used to sense the tension or pressure wave changes caused by the transmission of intra-abdominal pressure changes along the body surface.

[0012] The throat vibration detection unit comprises a MEMS acceleration sensor or a piezoelectric vibration sensor sheet attached to the patient's Adam's apple position, and is used to collect micro-vibration signals generated by rapid glottis response.

[0013] The multi-modal signal processing module specifically comprises: A filtering processing unit is used to remove breathing and heartbeat interference and retain the abdominal pressure mutation segment and the pharyngeal vibration characteristic segment; a peak change, a rate of change of the abdominal pressure signal, and an amplitude, a frequency, or an envelope feature of the vibration signal; a time alignment unit for determining whether the laryngopharyngeal vibration signal appears within a set time window after the abdominal pressure change.

[0014] The judgment logic of the reflux event determination module is: When the abdominal pressure detection unit collects a significant pressure change, and detects a laryngopharyngeal vibration signal within a preset time range after the change, and the intensity and mode of both exceed the threshold set by the system, it is identified as an implicit reflux event.

[0015] The data storage and output module includes: an event statistics unit for recording the time, frequency, duration, and intensity of the reflux event; a behavior label generation unit for identifying the triggering behavior associated with the reflux event in combination with the patient's eating, body position, and / or coughing behavior data; a periodic modeling unit for analyzing the patient's circadian reflux rhythm based on event data from multiple consecutive days to generate a time distribution heat map and a risk trend graph.

[0016] The system is integrated into a wearable neckband, and the abdominal pressure dynamic detection unit and the laryngopharyngeal vibration detection unit are embedded in the lower front edge and the front area of the neckband, respectively. The multi-modal signal processing module, the reflux event determination module, the data storage and output module, and the power module are set on one side of the neckband, and the overall structure has flexibility, adhesion, and daily wear adaptability.

[0017] A data analysis method for implicit gastroesophageal reflux events based on dynamic changes in abdominal pressure, comprising the following steps: a. Real-time collection of patient abdominal pressure fluctuation data by the abdominal pressure dynamic detection unit; b. Simultaneous collection of laryngopharyngeal muscle vibration signals by the laryngopharyngeal vibration detection unit; c. Band-pass filtering, feature extraction, and time window synchronization analysis in the multi-modal signal processing module; d. Determine whether the abdominal pressure signal meets the rise rate and amplitude threshold, and whether the laryngopharyngeal vibration signal is detected within a certain time window after the signal; e. If the preset determination condition is met, record it as an implicit reflux event by the reflux event determination module, and generate statistical results or alarm output by the data storage module.

[0018] The method is suitable for special groups at risk of implicit reflux, including GERD patients without typical symptoms, children, the elderly, postoperative patients, and sleep reflux patients.

[0019] The reflux event determination module adopts a gated window-based pattern recognition algorithm, and combines threshold filtering, noise suppression and feature extraction strategies to improve the accuracy and specificity of reflux event recognition.

[0020] The present application effectively breaks through the limitations of existing technology in the recognition of latent gastroesophageal reflux events by constructing a dual-channel linkage recognition mechanism for abdominal pressure dynamic signals and laryngeal vibration signals, and has the following beneficial effects: 1. Improve the accuracy and specificity of reflux recognition: Traditional reflux monitoring methods rely on esophageal pH or impedance probe insertion detection, which is highly invasive and cannot identify atypical or upper pharyngeal reflux events. The present application synchronously collects abdominal pressure changes and laryngeal vibration signals and performs linkage matching on the time axis to determine the occurrence of reflux based on the logic of "abdominal pressure mutation first, laryngeal response later", and constructs a recognition rule that conforms to the pathophysiological process, significantly reducing false positives and false negatives, and significantly improving recognition accuracy and specificity.

[0021] 2. Covering silent reflux and other atypical events: For the clinically common "silent reflux with no obvious symptoms and no significant abdominal pressure mutation", the present application innovatively introduces a bone conduction micro-sound perception channel in the laryngeal detection path, which can identify low-amplitude, short-duration but frequency-characteristic obvious "air-liquid interface disturbance" sound waves (such as gurgling sound, water bubble sound), even if the abdominal pressure and pharyngeal vibration paths are not triggered, it can also identify "low pressure reflux events", making up for the recognition blind area of traditional devices for slight reflux in the upper pharynx, and is particularly suitable for high-risk groups such as postoperative patients, infants and children, or unconscious patients.

[0022] 3. Improve the behavior discrimination ability and anti-interference of the system: The system integrates a behavior perception module and introduces a joint analysis mechanism for pharyngeal vibration signals and bone conduction signals, which can effectively exclude interference from non-reflux physiological behaviors such as swallowing, coughing, and language. For example: During eating, even if there is an abdominal pressure mutation and a pharyngeal vibration response, if there is no bone conduction signal or it is language characteristic, it can be determined as a non-reflux behavior; on the contrary, in the sleep state, a low-frequency abdominal pressure rise + weak bone conduction wet sound signal combination can more reliably identify silent reflux at night. This mechanism significantly enhances the adaptive recognition ability of the system in a natural living environment.

[0023] 4. Realize non-invasive, wearable, long-term dynamic monitoring: The present system uses a high-elasticity flexible medical neckband to integrate abdominal pressure detection and pharyngeal vibration detection units, which is light and comfortable, adjustable in tightness, does not need to be inserted or manually operated, and is suitable for children, the elderly, postoperative patients and other sensitive groups; At the same time, it is equipped with a low-power MCU and a Bluetooth communication module, which can support continuous use and data recording for more than 24 hours, making it convenient for doctors to track reflux event trends across multiple time periods and across day and night, and assist in diagnosis and efficacy evaluation.

[0024] 5. Enhanced adaptability to individual differences and parameter adjustment capability: System parameters can be individually set or model trained according to different patient physiological characteristics, disease stages or historical data, and the recognition sensitivity and specificity can be dynamically adjusted to adapt to the performance differences under different BMI, age and disease state, and to improve the overall adaptability.

[0025] 6. Clear visual clinical auxiliary value: The time stamp, intensity level and behavior label of all reflux events are automatically recorded and summarized into trend charts, heat maps and risk rhythm charts. Doctors can view daily / hourly reflux frequency distribution, most likely time period (such as after meals or at night), and induced behavior distribution through mobile App or backend platform, providing quantitative visual basis for clinical diagnosis and individualized treatment of gastroesophageal reflux disease (GERD) or atypical throat reflux.

[0026] In summary, the present application has made breakthrough progress in structure construction, signal processing, decision logic and application scenarios, not only improving the sensitivity and specificity of reflux event recognition, but also solving the problem of difficult identification of silent, atypical and night reflux in traditional technology. It realizes a practical technical solution with clinical verifiability, strong applicability and wide popularization prospect while ensuring comfort, and has obvious clinical beneficial effects. BRIEF DESCRIPTION OF DRAWINGS

[0027] Figure 1 is a schematic view of the neckband and the integrated device thereon of the present application; Figure 2 is a schematic view of a patient wearing the neckband of the present application; Figure 3 is a schematic view of the structure and function of the analysis system of the present application; Figure 4 is a schematic view of the flow of the analysis method of the present application.

[0028] In the figure, 1 is an abdominal pressure dynamic detection unit; 2 is a throat vibration detection unit; 3 is a multi-modal signal processing module; 4 is a reflux event determination module; 5 is a data storage and output module; and 6 is a neckband. DETAILED DESCRIPTION

[0029] In order to make the purpose, technical scheme and advantages of the present application clearer and more understandable, the present application will be further described in detail below in combination with the drawings and examples. It should be understood that the specific examples described herein are only used to explain the present application and do not limit the present application.

[0030] Referring to Figures 1 to 3 , the data analysis system for the latent gastroesophageal reflux events based on the dynamic changes of abdominal pressure comprises: The abdominal pressure dynamic detection unit 1 is used for collecting low-frequency micro-pressure fluctuation signals related to abdominal pressure changes of a patient's body surface area in real time and forming an abdominal pressure change data sequence. The throat vibration detection unit 2 is arranged at a throat part of the patient and is used for synchronously collecting vibration signals caused by muscle contraction or glottis closure of the throat part; The multi-modal signal processing module 3 is used for synchronously processing the abdominal pressure signals and the throat vibration signals, including filtering, feature extraction, time alignment and linkage analysis; The reflux event determination module 4 is used for determining a silent gastroesophageal reflux event according to a set timing logic, if the corresponding throat vibration signal is detected within a time range after the abdominal pressure fluctuation signal appears. The data storage and output module 5 is used for saving event records, counting reflux frequency, generating behavior labels and periodic risk maps, and providing prompts or output results to a terminal device.

[0031] Specifically, the abdominal pressure dynamic detection unit 1 is arranged at a position close to the clavicle at the lower edge of the neckband, adopts a flexible attachment type structure, and internally contains a plurality of array type micro-pressure sensors or strain gauges (such as bendable sensors based on MEMS technology) for real-time sensing of low-frequency pressure fluctuations or skin tension changes transmitted along the body surface due to changes in abdominal cavity pressure. To improve the fit and signal stability, the sensor is covered with a medical silicone layer and is fixed in contact with the skin through a low-sensitization adhesive film. The abdominal pressure signal has a typical slow-rising slow-falling or rapid-rising shape, and the system collects and forms an abdominal pressure change time sequence in real time.

[0032] The throat vibration detection unit 2 is arranged at the corresponding laryngeal prominence part in front of the neckband and closely adheres to the throat skin. The unit selects a MEMS micro-accelerometer or a piezoelectric vibration sensor, which has high sensitivity and can accurately capture the instantaneous vibration response of the throat muscles caused by reflux impact, such as glottis sudden closure, laryngeal reflex contraction or slight coughing action. Such vibration signals have characteristics such as high frequency, short duration and sharp shape, and can be distinguished from general swallowing, language or coughing actions.

[0033] The multi-modal signal processing module 3 is integrated in the side control unit of the neckband and includes a low-power MCU, an analog-to-digital conversion chip and a feature processing chip. The module processes the raw data of the two sensors in parallel, and the specific process includes: Band-pass filtering: the abdominal pressure signal is retained in the 0.1~~1Hz interval (respiration and heartbeat components are excluded), and the throat vibration signal is retained in the 20~~50Hz interval (language or cough interference is excluded); Feature extraction: extracting the mutation amplitude, rising rate and duration of the abdominal pressure signal, and extracting the main frequency band, amplitude envelope and signal intensity integral of the throat vibration signal; Time window alignment analysis: After determining whether there is a sudden change in abdominal pressure signal, a matching laryngeal vibration signal is found within a set time window, and the synchronization of the characteristics of the two is confirmed.

[0034] The reflux event determination module 4 performs logical recognition based on the signal time linkage relationship. The specific rules are as follows: When the abdominal pressure signal appears a rapid rising segment exceeding a preset threshold; and a laryngeal vibration event with vibration intensity exceeding the standard is detected within a set time window thereafter; the system determines that it is a silent gastroesophageal reflux event; and further determines its level (such as mild, moderate, severe), which is classified according to signal intensity and characteristic pattern.

[0035] This module uses a gated window matching logic and supports simple machine learning optimization rules to improve the accuracy and specificity of detection and reduce false positives.

[0036] The data storage and output module 5 is used for recording, analyzing and feeding back reflux event data. The specific functions include: Event storage and statistics: record the timestamp, signal intensity, duration and level of each reflux event; Behavior label generation: if the system integrates body position, cough, swallowing and other behavior sensors, it can associate reflux events with corresponding triggering behaviors, such as "occurring within 10 minutes of eating", "occurring in a supine position", etc. Periodic risk modeling: analyze continuous multi-day data to draw daily event density graphs, time trend lines and high-risk time heat maps to assist doctors in determining the regularity and severity of reflux occurrence; This module synchronizes with the mobile terminal App through Bluetooth, and doctors or patients can remotely view statistical charts and warning records.

[0037] The entire system structure is integrated into a flexible neck strap 6 that can be worn around the neck. The neck strap 6 is made of high-elastic breathable medical fabric, has good skin-friendliness and fit, and supports wearing during sleep. The abdominal pressure detection unit 1 and the laryngeal vibration detection unit 2 are respectively arranged at the lower edge and the front of the neck strap 6, and are adjusted in tightness by magic tape or fastener to ensure fit and non-slip. The remaining modules and battery module are uniformly arranged on one side of the neck strap 6, and are balancedly distributed to avoid discomfort when worn. The overall weight of the system is not more than 80g, and can be used continuously for more than 24 hours.

[0038] The system of the present application is suitable for a plurality of special groups of people at risk of "silent gastroesophageal reflux", including but not limited to: GERD patients without typical chest pain and burning sensation, sleep reflux patients, patients who cannot complain or have language barriers (such as children and the elderly), postoperative recovery patients, and people with chronic pharyngitis, hoarseness, asthma and other atypical manifestations.

[0039] In addition, see Figure 4A data analysis method for latent gastroesophageal reflux events based on dynamic changes in abdominal pressure includes the following steps: a. Real-time acquisition of patient abdominal pressure fluctuation data via abdominal pressure dynamic detection unit 1; b. Simultaneously acquire pharyngeal muscle vibration signals through the pharyngeal vibration detection unit 2; c. Perform bandpass filtering, feature extraction, and time window synchronization analysis in multimodal signal processing module 3; d. Determine whether the abdominal pressure signal meets the rise rate and amplitude thresholds, and whether a pharyngeal vibration signal is detected within a certain time window after the signal; e. If the preset judgment conditions are met, the backflow event judgment module 4 records it as a hidden backflow event, and the data storage module 5 generates statistical results or alarm output.

[0040] More specifically, the steps are as follows: Step a: Acquisition of dynamic abdominal pressure signals A flexible abdominal pressure dynamic detection unit 1, positioned at the lower anterior edge of the neck band 6 or above the clavicle, collects real-time signals of changes in body surface tension using miniature piezoelectric or strain gauge sensors. This signal indirectly reflects intra-abdominal pressure waves. A sampling frequency of 10–30 Hz is recommended to capture subtle dynamic changes such as breath-holding, straining, and pressure buildup in the stomach. The raw signal is a continuous time-series waveform, which, after initial smoothing and noise reduction, is sent to the next processing module.

[0041] Step b: Acquisition of pharyngeal vibration signals Simultaneously, the pharyngeal vibration detection unit 2, attached directly in front of the Adam's apple, collects subtle vibration signals from the muscle tissue beneath the patient's pharyngeal skin, with a sampling frequency of 100-200Hz, focusing on capturing: The rapid closure of the glottis is caused by the backflow of esophageal contents impacting the larynx; Reflexive coughing, tension or contraction of the throat; Other micro-tremors, such as throat-clearing movements.

[0042] This signal is short-duration and high-frequency, usually appearing as a momentary spike, and is a direct representation of the backflow "effect".

[0043] III. Step c: Multimodal signal processing In the multimodal signal processing module 3, the signals from the two channels are acquired synchronously and processed in parallel, specifically including: 1. Filtering: The abdominal pressure signal is filtered using a 0.1–1 Hz bandpass filter to remove heartbeat, respiration and low-frequency drift; the pharyngeal vibration signal is filtered using a 20–50 Hz bandpass filter to remove speech / cough / noise interference; if the system integrates a behavior perception module, it can simultaneously recognize actions such as coughing and swallowing to assist in the judgment.

[0044] 2. Feature extraction, comprising: Abdominal pressure feature: detecting the peak, starting point, maximum slope of the "rapid rising waveband"; Pharyngeal shock feature: extracting the instantaneous energy, main frequency interval, duration, signal peak and envelope feature; The feature standard is based on pre-clinical statistics, and the threshold parameters are set.

[0045] 3. Time alignment and linkage analysis: taking the abdominal pressure rising waveband as the starting reference; searching for whether a significant pharyngeal shock event occurs within the 0.3-1.5 second determination window thereafter; if the abdominal pressure mutation and the pharyngeal shock feature satisfy the "cause-effect" time sequence relationship in the time domain, entering the next step of determination.

[0046] Four, step d: logical determination of reflux event The following logic is performed by the reflux event determination module 4: 1. Determination condition The abdominal pressure signal satisfies: maximum change amplitude > A1, change rate > A2; the pharyngeal shock signal satisfies: amplitude > B1, duration > B2; and the pharyngeal shock signal appears within the effective time window ΔT (such as within 1 second) after the abdominal pressure mutation; then it is determined as a latent pharyngeal response caused by a gastric content reflux stimulation.

[0047] A1 is the maximum change amplitude of abdominal pressure, that is, the maximum amplitude change of the abdominal pressure signal within a preset window, which is used to identify high-amplitude pressure accumulation behavior with potential clinical significance; A2 is the abdominal pressure change rate, which represents the change speed of the abdominal pressure signal per unit time, and is used to eliminate slow fluctuations and physiological noise. Only when the abdominal pressure signal satisfies the threshold requirements of A1 and A2 at the same time, the system determines that the signal has "suspected inducibility".

[0048] B1 is the vibration amplitude, which is used to ensure that the vibration event is significant and to exclude slight muscle contraction or swallowing behavior; B2 is the vibration duration, which is used to exclude transient high-frequency interference or random vibration. Only when the amplitude of the pharyngeal shock signal exceeds B1 and the duration exceeds B2, and its occurrence time falls within the ΔT time window after the abdominal pressure mutation, the system determines that this event is a "latent gastroesophageal reflux event".

[0049] This method comprehensively considers the time sequence causal relationship between the sudden increase of abdominal pressure in the stomach content upward process and the muscle response of the pharynx and throat, takes the "abdominal pressure mutation first and pharyngeal response later" mode as the physiological basis, adopts clear signal strength and time domain determination rules, and improves the identification ability of atypical reflux events. Parameters A1, A2, B1, B2 and ΔT can be adjusted individually according to different populations, clinical states or data training models, to realize more flexible identification strategy and higher specificity and sensitivity.

[0050] 2. Auxiliary criterion If the system accesses the body position / food intake / cough sensing unit, false triggering can be excluded according to the behavioral context; dynamic threshold correction based on statistical rules can be introduced to improve robustness.

[0051] 3. Event output and labeling Each identified event is automatically labeled, including: time point, abdominal pressure peak, pharyngeal shock intensity, judgment level; level classification is based on amplitude, frequency, and duration to determine mild, moderate, and severe reflux.

[0052] Five, step e: data storage and analysis output After the judgment confirms the reflux event, the following processing is automatically performed by the data storage module 5: 1. Event record: including time stamp, signal intensity, analysis characteristics, level; associated with behavioral events (eating, supine, sleep) storage.

[0053] 2. Statistical analysis: statistics of daily and hourly reflux frequency; calculate the most likely time period (such as 1h after meal, 0-3am at night); generate heat map, trend chart and rhythm model.

[0054] 3. Prompt and export: set threshold, if the reflux event exceeds the set upper limit within 24h, push the prompt; data is synchronized to App or doctor's background through Bluetooth or Wi-Fi for follow-up and intervention.

[0055] In addition, in clinical cases, some patients have special reflux performance - that is, without significant abdominal pressure rise or muscle vibration, only showing a slight "gurgling sound" or "wet whistling" in the throat. This kind of signal is difficult to detect with the naked eye, patients feel vague, and conventional equipment cannot identify it. It belongs to non-pressure-driven reflux, and its typical clinical feature is that reflux liquid silently slides into the upper pharynx or larynx area, only forming low-intensity abnormal sound wave disturbance at the gas-liquid interface.

[0056] To solve the problem of missing detection of such "silent reflux", the present application introduces a "reverse flow sound wave sensing redundant channel" in the pharyngeal and laryngeal detection path. Its core is to add a bone conduction type miniature microphone element based on the structure of the existing pharyngeal and laryngeal vibration detection unit. This element relies on the neck bone as a sound wave transmission medium, and can accurately pick up non-muscle-derived small sound waves caused by pharyngeal and laryngeal gas-liquid disturbance, especially suitable for capturing "bubble breaking sound", "wet whistling" or "short gurgling" when reflux moves up.

[0057] The system processing logic is also optimized. In the signal processing module, a new independent audio recognition channel is added, which is specifically used to analyze the signals collected by the bone conduction channel, and the core algorithm includes: Frequency band pass filter: limit the signal in the range of 300Hz~800Hz, exclude language, cough, swallowing interference; Non-periodic burst detection: identify waveform mutation events, filter out continuous noise; Short-time energy envelope extraction: focus on analyzing "sound burst" segments with duration <1 second, low amplitude but obvious peak protrusion.

[0058] If the signal collected in the channel meets the above-mentioned sound wave reflux characteristics, and its appearance is not accompanied by abdominal pressure rise or obvious throat vibration, the system will mark it as "low pressure reflux event" or "silent reflux event", and record it in the data statistics module with different colors or labels.

[0059] This redundant channel not only enhances the system's ability to identify atypical reflux patterns, but also allows for high-sensitivity monitoring without interference or subjective cooperation in scenarios such as nighttime sleep, surgical anesthesia, use by the elderly or infants, and solves the blind spot that abdominal pressure detection paths cannot cover.

[0060] The implementation mechanism of the "silent gastroesophageal reflux event recognition system based on the dynamic change of abdominal pressure and the dual-path recognition mechanism of throat vibration" is based on the physiological and pathological mechanisms of gastroesophageal reflux. It constructs a closed-loop recognition logic composed of five layers of linkage: cause perception, physiological response capture, timing determination, event recognition, and data feedback, ensuring that the system has high sensitivity and high specificity under non-invasive, real-time, and continuous monitoring conditions, and adapts to the complex reality of "silent reflux" that is difficult to capture and easy to miss in clinical diagnosis.

[0061] First, from the physiological mechanism, gastroesophageal reflux events are usually caused by the upward movement of gastric contents due to increased intra-abdominal pressure or relaxation of the sphincter. Its path passes through the esophagus, throat, and can cause pharyngeal mucosa discomfort, or even lead to glottal stimulation, coughing, and other reactions. However, in clinical practice, a large number of reflux events do not accompany typical heartburn and other subjective symptoms, especially in sleep, postoperative sedation, children, or aphasia patients, which are manifested as "silent reflux" or "non-acid reflux", and are often misdiagnosed as pharyngitis, asthma, hoarseness, etc. Such events may be accompanied by a short-term increase in abdominal pressure before they occur, and can cause muscle vibration or sound wave disturbance after entering the throat. Therefore, if the short-term linkage relationship between "abdominal pressure mutation" and "throat reaction" can be accurately identified, the objective detection of silent reflux can be achieved.

[0062] To this end, the system first monitors the body surface micro-pressure fluctuation signal continuously through the abdominal pressure dynamic detection unit arranged in the clavicle region of the neck. The signal is derived from the upward transmission of abdominal cavity pressure through the diaphragm and subcutaneous tissue, and especially under the conditions of holding breath, force, stomach distension, etc., it will show sudden, aperiodic low-frequency fluctuations. Through the flexible attached micro-pressure sensing array, the unit can capture such changes and form an abdominal pressure waveform data stream for determining whether there is an "abdominal pressure mutation event" with inducing potential.

[0063] Secondly, the laryngeal vibration detection unit is attached to the Adam's apple position and is specially used to collect muscle responses caused by stomach contents impacting the throat, such as glottis closure, pharyngeal muscle vibration, short and brief coughing, etc. These signals show vibration bands with short duration, high frequency and peak out, which are quite different from normal language and swallowing vibration characteristics, and have good separation. Such signals are the "effect type" evidence that reflux events really act on the upper respiratory tract.

[0064] Thirdly, the multi-modal signal processing module in the system synchronously collects, filters, extracts features and time window aligns the above two signals, ensuring that the data is accurately synchronized in the time scale. By setting a typical time window, it is determined whether there is a causal time sequence relationship of "abdominal pressure mutation first and pharyngeal vibration following", and accordingly enters the reflux event determination logic module. According to the signal amplitude, rising rate, duration, vibration intensity and other indicators, it is determined whether a set of physiological meaning clear determination threshold group (A1, A2, B1, B2, ΔT) is met, so as to realize structured event recognition.

[0065] It is worth mentioning that, in order to solve the clinical blind area that some "silent reflux" events may occur without abdominal pressure fluctuation and muscle response, a bone conduction redundant channel is further introduced in the design of the system. A bone conduction type micro microphone element is embedded in the laryngeal detection unit structure, which can pick up non-muscle-derived sound signals caused by liquid disturbance, such as "gurgling sound", "wet whistling sound" or "fine explosion sound". These signals are often produced when liquid slides into the throat and the gas-liquid interface is disturbed, and traditional vibration or pressure sensors cannot identify them. The bone conduction method stably transmits such low-intensity sound through the neck bones, so that the system has the ability to identify "non-pressure-driven" and "non-muscle response type" reflux.

[0066] In the recognition level, the system integrates the above three data to form a determination logic closed loop: if the abdominal pressure fluctuation + pharyngeal vibration both meet the threshold, it can be determined as "dynamic reflux"; if there is no abdominal pressure and no pharyngeal vibration but bone conduction is activated, it is determined as "silent reflux"; if there is no matching signal or the time sequence is not consistent, it is excluded. The system has built-in signal feature discrimination logic, which can also adjust the threshold according to different groups of people, disease types or training models, to improve the recognition sensitivity and adaptability.

[0067] Finally, at the level of data feedback and clinical application, all events are recorded with characteristic parameters, time points, signal levels, and so on, and are marked with behavior perception units (such as eating, body position, and sleep recording) to generate trend charts and cycle risk modeling. The data can be uploaded to the doctor's terminal through Bluetooth or Wi-Fi to realize remote analysis, disease intervention, and treatment response evaluation, and truly form a complete data closed loop from event recognition to clinical feedback.

[0068] In summary, the system combines the abdominal pressure pathway, the pharyngeal vibration pathway, and the bone conduction sound wave pathway to construct a highly coupled implicit gastroesophageal reflux event recognition mechanism. The implementation mechanism not only conforms to the actual physiological process of gastroesophageal reflux events, but also conforms to the operable path of clinical verification, and has high feasibility and application prospect.

[0069] The following are several typical clinical application cases based on the present application, which prove the practical application value, technical advantages, and key role in precise treatment of the present scheme: Case One: Static abdominal pressure reflux event of a postoperative patient with sensitive pharyngeal reflex Patient information: 47-year-old female, 3 days after sleeve gastrectomy.

[0070] Background characteristics: Less food intake, no typical chest burning, and report of "occasional tight throat".

[0071] Monitoring performance: Abdominal pressure short-time surge (peak +24%) in night supine state; followed by rapid vibration of the throat (non-swallowing, no language) within 0.6 seconds; no typical air-liquid sound wave detected in the bone conduction audio channel.

[0072] Doctor review: No aspiration phenomenon in bedside electronic laryngoscope, and the patient reported "significant nocturnal acid reflux" the next morning.

[0073] The event occurred in the high-risk period of delayed gastric emptying after surgery, reflecting the anastomosis logic; as there is no sound wave involved, it is only triggered by the double physical channels, which conforms to the "static pressure-throat reaction" type of reflux performance.

[0074] Case Two: Non-sound wave reflux event of a child with mild LPR (laryngopharyngeal reflux) Patient information: 9-year-old boy, suffered from allergic rhinitis, long-term throat clearing, and night cough.

[0075] Background characteristics: No acid reflux complaint, parents only observed that he "swallowed water and cleared his throat sometimes at night".

[0076] Monitoring performance: In the night supine position, abdominal pressure slowly climbed and broke through the threshold several times; immediately after, a throat vibration signal was detected, with a duration of about 0.35 seconds and a low but stable amplitude; no abnormal "gurgling" or "whistling" signal was detected in the bone conduction sound channel.

[0077] Doctor's review: Part of the reflux goes up without crossing the glottis, but there is indeed gas stimulation found by 24-hour pharyngeal pH monitoring.

[0078] Such cases are in the early stage of LPR, typical reflux is not dominated by liquid, gas phase stimulation is dominant, no gas-liquid sound wave is produced, but pharyngeal reflex is indeed induced, and it is accurately captured under the coordination of the double channels.

[0079] Case three, gas type diurnal reflux in elderly patients Patient information: 72-year-old male, often taking antacids due to decreased digestive function.

[0080] Background characteristics: Coughing several times a day, but no obvious heartburn.

[0081] Monitoring performance: Rapid fluctuation of abdominal pressure (typical gas release mode) was detected within 30 minutes after meals; short pharyngeal shock (0.2 seconds) occurred within 0.8 seconds, with obvious amplitude, different from speech or swallowing; only background low-frequency noise was recorded in the bone conduction sound channel, without wave peak events.

[0082] Doctor's review: Combined with video materials and patient complaints, it is judged that stomach gas reflux stimulates the glottis, causing mild coughing.

[0083] Typical "gas phase reflux", it is difficult to form obvious gas-liquid sound, but the pharyngeal response is clear; it can be captured by relying on abdominal pressure + vibration.

[0084] Case four, supine reflux during non-eating period Patient information: 55-year-old female, day shift nurse, night shift sleep fragmentation.

[0085] Background characteristics: Repeatedly awakened at night, with the habit of clearing the throat.

[0086] Monitoring performance: After 20 minutes of supine position, a sudden change in low-amplitude abdominal pressure wave was recorded; about 0.4 seconds later, a slight vibration in the throat was detected (low intensity but structure matched); no bone conduction sound events were detected.

[0087] Doctor's review: Follow-up gastroscopy and esophageal impedance examination suggest mild resting reflux.

[0088] Atypical quiet period reflux, which belongs to a high-risk ignored event, is successfully detected, indicating that the system is sensitive to "non-acoustic, non-subjective" events.

[0089] Case five, swallowing reflux error exclusion verification Patient information: 33-year-old male, professional teacher, feels morning acid reflux.

[0090] Background characteristics: The system detected abdominal pressure and pharyngeal shock matching several times during eating, but some were active swallowing.

[0091] Monitoring performance: In the explicit feeding segment (labeled synchronous behavior module), several sudden changes in abdominal pressure → pharyngeal shock response; bone conduction has no signal, and the system identifies it as a non-reflux event; the true reflux event occurs at night independent of swallowing behavior.

[0092] The system can accurately distinguish subjective swallowing behavior from silent reflux, with good specificity, even if the bone conduction channel has no signal, the judgment is still accurate.

[0093] Case six, bone conduction trigger verification of silent reflux at night Patient information: 49-year-old female, complains of hoarseness in the morning, no typical burning sensation.

[0094] Background characteristics: prone to pharyngeal irritation during sleep, past night pH detection is not typical, preliminary suspicion of silent nocturnal reflux.

[0095] Monitoring performance: Wear the device for continuous monitoring for 8 hours. No significant reflux events were detected during the waking period; bone conduction signal burst activation occurred twice at 0:52 and 2:37 at night, with a frequency of 450-600 Hz and a duration of about 0.7s. Abdominal pressure and pharyngeal shock channel signals are stable and do not reach the threshold.

[0096] The system identifies the event through the bone conduction sound wave sensing channel and records it as "low pressure silent reflux"; endoscopic examination found signs of chronic irritation of the posterior pharyngeal wall, verifying the complementary value of bone conduction path to nocturnal reflux.

[0097] Case seven, bone conduction independent recognition under intermittent clearing symptoms in children Patient information: 7-year-old boy, low weight, presents with night cough and daytime frequent clearing, no heartburn or chest pain.

[0098] Background characteristics: Parents report that night cough is related to sleeping position, and have suspected post-nasal drip or airway allergy.

[0099] Monitoring performance: Wear the system for two consecutive nights, no events on the first night; at 22:14 on the second night, the system detects a short "gurgle" sound wave (main frequency 510 Hz, duration 0.6s) through the bone conduction channel, and there is no significant change in abdominal pressure and pharyngeal shock. Clearing behavior occurs within 30 seconds.

[0100] The system is labeled as "silent pharyngeal reflux" and is included in the "bone conduction independent trigger category" in the next day's data statistics. After laryngoscopy, mild pharyngeal reflux inflammation was found. This result rules out allergic factors and confirms the important clinical recognition value of bone conduction for atypical reflux in children.

[0101] Case eight, bone conduction response to gas-liquid disturbance reflux in supine position Patient information: 65-year-old male, has a history of GERD, recent symptoms are well controlled, occasionally feels foreign body in the throat.

[0102] Background feature: Wearing monitoring system during postprandial recumbency lunch break, observing "non-stimulating" reflux risk.

[0103] Monitoring performance: The system recorded a bone conduction signal activation (frequency range 530~670 Hz, duration about 1 second) at 14:47, without accompanying abdominal pressure mutation or pharyngeal shock characteristics, and the behavior module simultaneously prompted "recumbent" state.

[0104] The system determines it as "low pressure gas-liquid interface disturbance type reflux", and marks it with the "redundant path trigger" label. No repeated events occurred within the next 24 hours. Follow-up endoscopy suggests cardia relaxation and mild esophageal mucosa irritation, verifying the system detection results.

[0105] The above only describes the preferred embodiments of the present application and is not intended to limit the present application. Any modifications, equivalent replacements, and improvements made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A system for data analysis of silent gastroesophageal reflux events based on dynamic changes in abdominal pressure, characterized in that, Comprising: an abdominal pressure dynamic detection unit (1) for real-time acquisition of low-frequency micro-pressure fluctuation signals related to abdominal pressure changes in the patient's body surface area and forming an abdominal pressure change data sequence; a laryngeal vibration detection unit (2) arranged at the patient's laryngeal site for synchronous acquisition of vibration signals caused by laryngeal muscle contraction or glottis closure; a multi-modal signal processing module (3) for synchronous processing of abdominal pressure signals and laryngeal vibration signals, including filtering, feature extraction, time alignment and linkage analysis; a reflux event determination module (4) for determining a silent gastroesophageal reflux event according to the set timing logic within a certain time range after the abdominal pressure fluctuation signal appears, if the corresponding laryngeal vibration signal is detected; a data storage and output module (5) for saving event records, counting reflux frequency, generating behavior labels and periodic risk maps, and providing prompts or output results to terminal devices.

2. The system of claim 1, wherein, The abdominal pressure dynamic detection unit (1) includes a flexible micro-pressure sensor or strain gauge structure arranged below the neck or above the collarbone for sensing the tension or pressure wave changes of the abdominal cavity pressure changes along the body surface.

3. The system of claim 1, wherein, The laryngeal vibration detection unit (2) includes a MEMS acceleration sensor or a piezoelectric vibration sensor attached to the patient's Adam's apple position for collecting micro-vibration signals generated by the rapid response of the glottis.

4. The system of claim 1, wherein, The multi-modal signal processing module (3) specifically includes: a filtering processing unit for removing breathing and heartbeat interference and retaining abdominal pressure mutation segments and pharyngeal vibration characteristic segments; a feature extraction unit for extracting peak changes, change rates of abdominal pressure signals, and amplitude, frequency or envelope features of vibration signals; a time alignment unit for determining whether the laryngeal vibration signal appears within the set time window after the abdominal pressure change.

5. The system of claim 1, wherein, The judgment logic of the reflux event determination module (4) is: When the abdominal pressure detection unit collects obvious pressure changes and detects laryngeal vibration signals within the preset time range after the changes, and the intensity and mode of both exceed the system's threshold standard, it is identified as a silent reflux event.

6. The system of claim 1, wherein, The data storage and output module (5) includes: an event statistics unit for recording the time, frequency, duration and intensity of reflux events; a behavior label generation unit for identifying the trigger behaviors that may be associated with reflux events in combination with the patient's eating, body position and / or coughing behavior data; a periodic modeling unit for analyzing the patient's circadian reflux rhythm based on event data of multiple consecutive days to generate a time distribution heat map and a risk trend graph.

7. The system of claim 1, wherein, The system is integrated into a wearable neckband (6), the abdominal pressure dynamic detection unit (1) and the laryngeal vibration detection unit (2) are respectively embedded in the lower front edge and the front area of the neckband (6), the multi-modal signal processing module (3), the reflux event determination module (4), the data storage and output module (5) and the power module are arranged on one side of the neckband (6), and the overall structure has flexibility, adhesion and daily wear adaptability.

8. A method for analyzing data of an implicit gastroesophageal reflux event based on dynamic changes in abdominal pressure, characterized in that, The method comprises the following steps: a. Real-time acquisition of abdominal pressure fluctuation data of the patient by the abdominal pressure dynamic detection unit (1); b. Collecting throat muscle vibration signals by the throat vibration detection unit (2) synchronously; c. Performing band-pass filtering, feature extraction and time window synchronous analysis in the multi-modal signal processing module (3); d. Determining whether the abdominal pressure signal meets the rising rate and amplitude threshold, and whether the throat vibration signal is detected within a certain time window after the signal; e. If the preset determination condition is met, recording a silent reflux event by the reflux event determination module (4), and generating statistical results or alarm output by the data storage module (5).

9. The method of claim 8, wherein, The method is suitable for special groups at risk of silent reflux, including GERD patients without typical symptoms, children, the elderly, postoperative patients and sleep reflux patients.

10. The method according to claim 8 or 9, characterized in that, The reflux event determination module (4) adopts a pattern recognition algorithm based on a gated window, and combines threshold filtering, noise suppression and feature extraction strategies to improve the accuracy and specificity of reflux event recognition.

Images