Throat electromyographic signal synchronous swallowing action monitoring and aspiration early warning device

Abstract

The invention relates to the field of medical electronics and rehabilitation engineering, and discloses a throat electromyographic signal synchronous swallowing action monitoring and aspiration early-warning device which comprises a shell, a display screen is fixedly connected to the top of the shell, a connecting end is arranged outside the shell, one end of a connecting wire is arranged in the middle of the connecting end, and the other end of the connecting wire is connected with the display screen. A signal collecting assembly is arranged outside the connecting line at the other end, the collecting assembly is electrically connected with a signal processing assembly and used for receiving collected throat electromyographic signals, analyzing the signals and outputting the signals, and the signal processing assembly comprises a main control circuit board. According to the invention, through the multi-source signal synchronization module, a mechanical acceleration peak value of laryngeal knot movement is captured by using an inertial sensor to serve as a hard time reference point, and a time window is constructed to segment and align throat electromyographic signals, so that the time resolution and the analysis accuracy of swallowing action feature extraction are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of medical electronics and rehabilitation engineering, in particular to a swallowing action monitoring and aspiration warning device synchronized with laryngeal myoelectric signals. BACKGROUND

[0002] Swallowing disorders are common in patients with nervous system or muscle damage, and if not timely intervention, it is easy to cause aspiration or suffocation. The existing non-invasive monitoring technology mostly relies on surface myoelectric signals or acoustic signals. Due to the weak amplitude and strong randomness of the surface myoelectric signals of the larynx, there is a lack of clear kinematics reference, and it is difficult to accurately define the starting and ending time of the action, resulting in blurred signal segmentation boundary. In the analysis of multiple source signals, the existing technology lacks a unified physical motion time reference, and there is a time deviation between different modal signals, making it difficult to achieve accurate registration and reducing the time resolution and accuracy of swallowing coordination analysis.

[0003] In terms of risk assessment and diagnosis, clinical swallowing radiography has radiation risk and large equipment volume, which is not suitable for daily real-time monitoring; bedside swallowing screening relies on the experience of medical staff and lacks objective quantitative indicators. The existing portable monitoring equipment usually only displays the original waveform, and does not have a normal swallowing feature model based on large sample data built-in, and cannot calculate the feature deviation degree of the monitoring data and the standard mode in real time. This technical limitation makes the aspiration risk warning mainly stay at the qualitative level, and cannot provide intuitive quantitative data support for clinical diagnosis.

[0004] In terms of hardware implementation, microvolt-level laryngeal myoelectric signals are easily disturbed by environmental noise. If the RF communication module integrated for wireless transmission lacks targeted shielding design, electromagnetic interference will be introduced and the signal SNR will be reduced. In addition, the existing devices generally lack a wearing state sensing function, which causes the device to continuously run in a non-wearing state or be triggered by environmental vibration, resulting in invalid data acquisition and false alarms, and at the same time increases the system power consumption and shortens the device battery life. SUMMARY

[0005] In view of the deficiencies of the prior art, the present application provides a swallowing action monitoring and aspiration warning device synchronized with laryngeal myoelectric signals, which solves the problems of difficulty in accurately defining the action time sequence by simply relying on myoelectric signals, large multi-source signal registration deviation, and lack of quantitative aspiration risk assessment based on standard models.

[0006] In order to achieve the above object, the present application is realized by the following technical scheme: the swallowing action monitoring and aspiration early warning device synchronized with the laryngeal electromyographic signal, comprising a shell, a display screen is fixedly connected to the top of the shell, a connecting end is arranged outside the shell, a connecting line is arranged at the middle of the connecting end, a signal acquisition assembly is arranged outside the other end of the connecting line, a signal processing assembly is electrically connected to the acquisition assembly, and the laryngeal electromyographic signal is received, analyzed and output; The signal processing assembly comprises a main control circuit board, the main control circuit board is arranged inside the shell, a working electrode is electrically connected to the outside of the interdigital electrode, a Bluetooth module and an inertial sensor are sequentially electrically connected to the top of the main control circuit board, an intelligent central control and a multi-dimensional state sensing system are arranged in the main control circuit board.

[0007] Preferably, the signal acquisition assembly comprises a silica gel pad, a reference electrode is arranged at the middle of the silica gel pad, the working electrode is arranged outside the silica gel pad, and the skin is contacted and the signal is acquired.

[0008] Preferably, the intelligent central control and the multi-dimensional state sensing system realize accurate monitoring and early warning of the swallowing process through the following functional modules: Firstly, the system is provided with a wearing state monitoring module. The working principle of the module is based on edge electric field induction technology, and the capacitance value change feedback by the interdigital electrode is monitored in real time. When the human skin contacts or approaches the device, the dielectric constant changes cause the capacitance value to change suddenly, and the module determines whether the device is in an effective wearing state according to this, and controls the wake-up start or sleep of the system accordingly, so as to reduce the invalid power consumption.

[0009] Secondly, the system is provided with a multi-source signal synchronization module for solving the time axis alignment problem of electrophysiological signals and mechanical motion signals. The running logic of the module is: taking the peak value time of the laryngeal prominence motion acceleration detected by the inertial sensor as the reference time point, and constructing a time window with a preset length centering on the reference point. The system intercepts the laryngeal electromyographic signal segment collected by the working electrode and the reference electrode in the time window, so as to realize the synchronous segmentation and accurate registration of the motion signal and the electromyographic signal in the time dimension.

[0010] Thirdly, the system is provided with a swallowing feature extraction module for quantifying feature analysis of the multi-source signals after synchronous segmentation. For the laryngeal electromyographic signal, the module calculates the root mean square value to represent the effective value of the signal, and calculates the integral electromyographic value to represent the total amount of muscle activity; for the laryngeal prominence motion data, the module extracts the maximum displacement amplitude and motion duration. The above parameters jointly construct a multi-dimensional swallowing feature vector.

[0011] Finally, the system is provided with an aspiration risk assessment and interaction module for performing quantitative risk determination. A normal swallowing feature threshold model trained based on healthy population data is pre-stored in the system. During operation, the module inputs the real-time multi-dimensional swallowing feature vector constructed into the model for comparison, and calculates the feature deviation between the real-time feature vector and the normal model. When the calculated feature deviation value exceeds the safety threshold pre-set by the system, it is determined that the current swallowing action has aspiration risk. At this time, the system immediately generates a red warning signal, drives the display terminal to pop up an alarm window, and synchronously plots the aligned laryngeal muscle electromyographic signal change curve and the laryngeal prominence movement trajectory curve on the display screen for intuitive observation by medical staff.

[0012] Preferably, the main control circuit board is externally electrically connected with an electromagnetic shielding cover for covering the analog signal processing area on the main control circuit board to shield external electromagnetic interference and radio frequency interference generated by the Bluetooth module.

[0013] Preferably, the shell is externally fixedly connected with an LED display lamp, and the LED display lamp is electrically connected with the main control circuit board.

[0014] Preferably, the silicone pad is of self-adhesive silicone rubber material.

[0015] Preferably, the middle part of the silicone pad is provided with a MEMS microphone for collecting swallowing sound.

[0016] Preferably, the bottom of the shell is provided with a fixing frame, and the main control circuit board is externally fixedly connected with a battery.

[0017] The present application provides a laryngeal electromyographic signal synchronous swallowing action monitoring and aspiration early warning device. It has the following advantages: 1. The present application uses an inertial sensor to capture the mechanical acceleration peak of the laryngeal prominence movement as a hard time reference point to construct a time window for dividing and aligning the laryngeal electromyographic signal through a multi-source signal synchronous module. This design effectively solves the technical problem of being difficult to define the starting point and ending point of the swallowing action in the prior art by relying solely on the electromyographic signal, eliminates the time lag between different signal sources, and improves the time resolution and analysis accuracy of the swallowing action feature extraction.

[0018] 2. The present application discards the traditional subjective observation method and establishes a multi-dimensional feature vector model based on time domain (such as movement duration) and frequency domain (such as integral electromyographic value) through an aspiration risk assessment and interaction module. By calculating the deviation of the real-time feature vector from the normal swallowing feature threshold model, the device can output a quantitative aspiration probability and synchronously plot the aligned electromyographic and movement curves on the display screen, providing intuitive and objective data support for medical staff and realizing the leap from qualitative observation to quantitative early warning.

[0019] 3、The application sets up an electromagnetic shield on the hardware, effectively covers the analog signal processing area on the main control circuit board, physically isolates the external environment electromagnetic interference and the radio frequency noise generated by the on-board Bluetooth module, ensures the signal-to-noise ratio of the micro-volt level electromyographic signal, combines the wearing state monitoring function based on the interdigital electrode capacitance induction, realizes the intelligent wake-up and sleep control, prolongs the equipment endurance and reduces the false alarm rate. BRIEF DESCRIPTION OF DRAWINGS

[0020] Figure 1 It is a perspective view of the application; Figure 2 It is a shell structure schematic view of the application; Figure 3 It is a main control circuit board structure schematic view of the application; Figure 4 It is a silica gel pad structure schematic view of the application; Figure 5 It is a functional module block diagram of the intelligent central control and multi-dimensional state sensing system of the application.

[0021] Among them, 1, shell; 2, display screen; 3, LED display lamp; 4, connecting line; 5, fixing frame; 6, connecting end; 7, electromagnetic shield; 8, battery; 9, main control circuit board; 10, interdigital electrode; 11, Bluetooth module; 12, inertial sensor; 13, MEMS microphone; 14, working electrode; 15, reference electrode; 16, silica gel pad. DETAILED DESCRIPTION

[0022] The technical solutions in the embodiments of the application will be clearly and completely described below with reference to the drawings in the specification of the application. Obviously, the described embodiments are only a part of the embodiments of the application, rather than all the embodiments. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative labor fall within the protection scope of the application.

[0023] Please refer to the drawings in the specification of the application Figure 1 - the drawings in the specification of the application Figure 5 The throat electromyographic signal synchronous swallowing action monitoring and aspiration early warning device provided by the embodiments of the application comprises a shell 1, the top of the shell 1 is fixedly connected with a display screen 2, the outside of the shell 1 is provided with a connecting end 6, the middle of the connecting end 6 is provided with one end of a connecting line 4, the other end of the connecting line 4 is provided with a signal acquisition assembly outside, the signal acquisition assembly is electrically connected with a signal processing assembly, and the signal processing assembly is used for accepting the acquired throat electromyographic signal and analyzing and outputting the structure. The signal processing assembly comprises a main control circuit board 9 arranged inside the shell 1, the top of the main control circuit board 9 is sequentially electrically connected with a interdigital electrode 10, a Bluetooth module 11 and an inertial sensor 12, the outside of the interdigital electrode 10 is electrically connected with a working electrode 14, and the inside of the main control circuit board 9 is provided with an intelligent central control and multi-dimensional state sensing system. The signal acquisition assembly comprises a silica gel pad 16, the middle of the silica gel pad 16 is provided with a reference electrode 15 for contacting the skin and acquiring signals, and the outside of the silica gel pad 16 is provided with the working electrode 14. The outside of the main control circuit board 9 is electrically connected with an electromagnetic shield 7 for covering the analog signal processing area on the main control circuit board 9, shielding external electromagnetic interference and radio frequency interference generated by the Bluetooth module 11. The shell 1 is externally fixedly connected with an LED display lamp 3, and the LED display lamp 3 is electrically connected with the main control circuit board 9. The silica gel pad 16 is made of self-adhesive silicone rubber material. The middle of the silica gel pad 16 is provided with a MEMS microphone 13 for collecting swallowing sound. The bottom of the shell 1 is provided with a fixing frame 5, and the outside of the main control circuit board 9 is fixedly connected with a battery 8.

[0024] After the device is started, the main control circuit board 9 continuously monitors the capacitance change of the contact interface by using the interdigital electrode 10. When the silica gel pad 16 made of self-adhesive silicone rubber material is attached to the skin of the human neck, the change of the dielectric constant of the contact surface causes the capacitance value fed back by the interdigital electrode 10 to suddenly change, and the intelligent central control system identifies the wearing state accordingly, and the control system automatically switches between the working mode and the low-power sleep mode, effectively reducing invalid power consumption and preventing false triggering.

[0025] When entering the monitoring state, the signal acquisition assembly and the signal processing assembly work cooperatively to acquire multi-dimensional physiological data. The reference electrode 15 located in the middle of the silica gel pad 16 provides a human body reference potential, and cooperates with the external working electrode 14 to form a differential signal acquisition loop, and real-time captures the weak myoelectric signal generated by the muscle contraction of the laryngeal part; at the same time, the MEMS microphone 13 collects the acoustic signal in the swallowing process. The inertial sensor 12 detects the spatial attitude change of the device in real time, and acquires the acceleration and angular velocity data generated by the up-and-down movement of the laryngeal prominence, and provides a time reference of mechanical movement for subsequent signal processing. In this process, the electromagnetic shield 7 covering the analog signal processing area of the main control circuit board 9 shields by grounding, physically isolates the external environmental electromagnetic noise and the radio frequency interference when the on-board Bluetooth module 11 works, and ensures that the high-gain analog circuit outputs the original signal with high signal-to-noise ratio.

[0026] The intelligent central control and multi-dimensional state perception system inside the main control circuit board 9 gathers and calculates the above-mentioned multi-source data. The system uses the motion peak time captured by the inertial sensor 12 to align the timing and window segment of the electromyographic signal, extracts the time domain and frequency domain feature vectors, and inputs them into the preset model for comparison to quantify the aspiration risk. The final analysis result is visualized on the display screen 2 at the top of the shell 1, which draws a waveform curve in real time and displays the risk level, and the LED display lamp 3 provides optical feedback of the power state or warning information according to the system logic, and the battery 8 provides stable DC power support for the whole machine.

[0027] In a specific implementation, the intelligent central control and multi-dimensional state perception system is configured to perform parallel collection and synchronous processing of multi-source signals. The input variables defined by the system include: the laryngeal prominence motion acceleration vector collected by the inertial sensor 12 , the raw electromyographic signal amplified by the difference between the working electrode 14 and the reference electrode 15 , and the real-time capacitance value fed back by the interdigital electrode 10 . The display screen 2 is the output terminal of the system, which is controlled by the system and is used to visualize the processed signal features and aspiration risk assessment results in time and frequency domain dimensions.

[0028] Specifically, based on the above hardware architecture, the intelligent central control and multi-dimensional state perception system is logically divided into four cooperatively working execution modules, namely: wearing state monitoring module, multi-source signal synchronization module, swallowing feature extraction module, and aspiration risk assessment and interaction module. The four modules are stored in the memory of the main control circuit board 9 in the form of embedded software code and are called and executed by the processing unit. After the system is powered on and initialized, the wearing state monitoring module is first run, and the module is in a resident running state, while the remaining three modules are controlled by the output instructions of the wearing state monitoring module and are in a power-off or low-power sleep mode in a non-wearing state to prolong the battery life of the battery 8.

[0029] The wearing state monitoring module works based on the principle of capacitive sensing, and its input end is connected to the interdigital electrode 10. The interdigital electrode 10 constructs a planar capacitor structure at the main control circuit board 9 or the flexible connection, which radiates an edge electric field to the external space. When the device is not worn, the medium of the edge electric field is air, and its relative permittivity is close to 1; when the device is attached to the human neck skin, the medium of the edge electric field changes to the skin tissue and subcutaneous moisture, and its relative permittivity is much larger than that of air. This sudden change of medium will cause the capacitance value at both ends of the interdigital electrode 10 to change.

[0030] The wearing state monitoring module reads the real-time capacitance value of the interdigital electrode 10 through the built-in capacitance digital converter (CDC) at a preset sampling frequency To eliminate the influence of environmental temperature and humidity drift and circuit parasitic capacitance, the system is provided with a dynamic reference capacitance value , which is automatically updated each time the system is reset or in a long-term stationary state. The wearing state monitoring module calculates the real-time capacitance change amount according to the following formula : ; Wherein, is the original capacitance data collected at the moment; is the current non-wearing state reference capacitance value. The system is provided with a predetermined threshold value

[0031] , which is calibrated by experiment and corresponds to the minimum capacitance increment when the silica gel pad 16 is fully attached to the skin. The wearing state monitoring module determines the current wearing state by comparing the logical function : ; Wherein, represents the wearing state flag bit, with a value of 1 representing a valid wearing state and a value of 0 representing a non-wearing or falling off state; represents the length of time during which the capacitance change amount continuously exceeds the threshold value; is a preset anti-shake time threshold value for filtering out false trigger signals caused by temporary hand contact.

[0032] When the wearing state flag bit jumps from 0 to 1, the wearing state monitoring module outputs a system wake-up instruction. This instruction triggers the power management unit of the main control circuit board 9, which in turn powers the inertial sensor 12, the Bluetooth module 11, and the front-end analog amplification circuit of the signal acquisition component, causing the system to enter a full-function working mode. At the same time, the module controls the LED display lamp 3 to light up, providing feedback to the user that the device is ready.

[0033] Conversely, when the wearing state flag bit jumps from 1 to 0 and maintains for a certain period of time, the wearing state monitoring module outputs a sleep instruction. The system immediately cuts off the power supply to other peripherals except the capacitance detection circuit, stops data acquisition and Bluetooth transmission, and controls the LED display lamp 3 to turn off, causing the device to automatically enter a micro-ampere-level low-power standby mode. This logic ensures that the device only performs swallowing monitoring and aspiration warning when it is effectively attached to the human skin, avoiding the collection of invalid data and false alarms.

[0034] ​Further, when the system is in the wake-up working state, the multi-source signal synchronization module and the swallowing feature extraction module are cooperatively operated to perform time sequence registration and feature quantization on the collected original physiological signals. The multi-source signal synchronization module is mainly used to solve the technical problem that a single electromyography signal is difficult to accurately define the starting point and the ending point of the swallowing action. The module first receives the laryngeal prominence motion acceleration signal collected by the inertial sensor 12 in real time , and performs low-pass filtering processing on the laryngeal prominence motion acceleration signal to remove high-frequency jitter noise and extract an acceleration waveform reflecting the vertical motion component of the laryngeal prominence.

[0035] The operation logic of the multi-source signal synchronization module is set as follows: signal segmentation is performed with the peak time of the laryngeal prominence motion acceleration as the reference point. During the operation process, the module continuously monitors the acceleration waveform, and when it is detected that the acceleration amplitude exceeds a preset motion trigger threshold , a peak search algorithm is started. The algorithm locks the time when the acceleration reaches the maximum value in the subsequent preset search interval, denoted as . This time represents the physical time when the laryngeal prominence is lifted to the highest point or the most intense motion, and has high time stability, which satisfies the following conditions: ; ; wherein is the acceleration amplitude at the time , and is the neighborhood time radius for local extreme value determination.

[0036] Based on the determined reference point , the multi-source signal synchronization module constructs a preset length time window including the whole swallowing process. The definition formula of the time window is as follows: ; wherein refers to the time point of the most intense motion (maximum acceleration) of the laryngeal prominence in the identified swallowing action, which is used as the synchronization reference anchor point of the whole system; refers to the neighborhood radius of the peak search algorithm, which is used to determine the search range of the local maximum value; refers to the forward interception time length (suggested value: 0.5s-1.0s), which is used to cover the electromyography signal before the oral preparation period of swallowing; refers to the backward interception time length (suggested value: 1.0s-1.5s), which is used to cover the electromyography signal in the reset period of swallowing.

[0037] The swallowing feature extraction module receives the signal segment after synchronization segmentation, and performs feature calculation in the time domain and the frequency domain to construct a multi-dimensional feature vector. For the intercepted laryngopharyngeal electromyography signal segment The module first calculates its root mean square (RMS) value to characterize the effective contraction strength of the relevant muscle groups during swallowing. The calculation formula is as follows: ; in, Refers to the synchronization time window The total number of sampling points contained herein satisfies ,in This refers to the system sampling frequency; Refers to the first time window The amplitude of the pharyngeal electromyography signal voltage at each discrete sampling point (unit: mV or This data is the data after removing the DC bias.

[0038] Simultaneously, the swallowing feature extraction module calculates the integrated electromyography (iEMG) value of the pharyngeal electromyography signal to characterize the cumulative discharge of muscles during the swallowing cycle. Integrated electromyography value The calculation formula is as follows: ; in, Refers to the absolute value of the electromyographic signal amplitude (full-wave rectification). Refers to the sampling period, which is the reciprocal of the sampling frequency. ), introduced here In order to approximate continuous integrals in a mathematical and physical sense This gives the result physical dimensions (Vs).

[0039] For the Adam's apple movement data, the swallowing feature extraction module performs a double integral operation on the acceleration signal within the time window to obtain the vertical displacement curve of the Adam's apple, and extracts the maximum displacement amplitude from it. This feature reflects the degree of Adam's apple elevation and is an important indicator for assessing the effectiveness of airway protection mechanisms. The calculation logic is as follows: ; ; ; in, Refers to the dynamic acceleration signal after removing the gravitational component; The time-varying curve of the vertical displacement of the Adam's apple relative to its initial position; It refers to the maximum physical displacement of the Adam's apple from its resting position to its highest point during a single swallow.

[0040] In addition, the module also calculates whether the acceleration signal exceeds the motion trigger threshold. The duration of the movement was determined by the duration of the Adam's apple movement. .

[0041] Finally, the swallowing feature extraction module normalizes the four independent feature parameters calculated above and combines them in a pre-set order to construct a multi-dimensional swallowing feature vector that describes the swallowing action : ; Among them, refers to the duration of laryngeal prominence movement; is the root mean square value; is the integral electromyography value; is the maximum displacement amplitude.

[0042] The feature vector is then transmitted to the aspiration risk assessment and interaction module as the data basis for subsequent pattern recognition and risk determination.

[0043] Further, the aspiration risk assessment and interaction module, as the decision output unit of the system, is configured to make logical judgments and visual feedback based on the quantitative data output by the swallowing feature extraction module. The module has pre-stored a normal swallowing feature threshold model, which is a reference system generated based on a large number of data samples of standard swallowing actions of healthy adults through statistical analysis or machine learning training. In specific embodiments, the normal swallowing feature threshold model is parameterized as a four-dimensional standard feature mean vector and a corresponding feature weight vector The standard feature mean vector contains the statistical mean values of the four key feature dimensions in the normal swallowing state, defined as: ; Among them, is the standard root mean square value, is the standard integral electromyography value, is the standard maximum displacement amplitude, is the standard movement duration. The feature weight vector assigns different weight coefficients to each feature parameter according to their sensitivity differences to aspiration risk, and satisfies the normalization condition.

[0044] The core logic of the aspiration risk assessment and interaction module when running is to input the multi-dimensional swallowing feature vector constructed in real time into the normal swallowing feature threshold model for comparison operation. The module uses the weighted Euclidean distance algorithm to quantify the difference between the current swallowing action and the standard swallowing action, and calculates the feature deviation degree . Its calculation formula is as follows: ; wherein, is the characteristic deviation value; is the weight coefficient of the th characteristic dimension; is the real-time calculated value of the th characteristic component; is the corresponding mean value of the th standard characteristic in the model; is the standard deviation of the th characteristic in the normal swallowing characteristic model, used to eliminate the scale difference between different dimension characteristics.

[0045] A clinically verified safety threshold is set in the system. The aspiration risk assessment and interaction module compares the calculated characteristic deviation with the safety threshold numerically. When the judgment result meets , the module determines that the current swallowing action is abnormal and identifies that there is an aspiration risk; when the judgment result meets , the module determines that the current swallowing action is within the safe range.

[0046] Once it is determined that there is an aspiration risk, the aspiration risk assessment and interaction module immediately triggers the alarm response mechanism. The system generates a red warning control signal and transmits it to the display screen 2. The display screen 2 responds to the signal, pops up a prominent alarm window at the top layer of the interface, and prompts the high aspiration risk word in the form of text or icon, while driving the LED display lamp 3 to flash red light at high frequency for physical warning, so as to realize the sound and light synchronous instant feedback.

[0047] In addition, the display terminal is configured to perform waveform synchronous drawing function. Regardless of the risk judgment result, the aspiration risk assessment and interaction module drives the display screen 2 to enter the real-time monitoring interface. The interface takes time axis as the horizontal coordinate, and draws or superimposes two groups of curves in the same window: one group is the change curve of the laryngeal muscle electrical signal collected and processed through the working electrode 14 and the reference electrode 15, and the other group is the laryngeal prominence movement trajectory curve collected and integrated through the inertial sensor 12. The two groups of curves follow the time reference established by the multi-source signal synchronous module, ensuring that the wave peaks and troughs of the waveforms correspond in time sequence, so that medical staff can intuitively observe the coordination of swallowing action and the timing relationship of muscle excitation, and assist in clinical diagnosis and rehabilitation assessment.

[0048] Working principle: After the device is powered on, the main control circuit board 9 first monitors the change of environmental capacitance through the interdigital electrode 10. When the self-adhesive silicone rubber material silica gel pad 16 is attached to the skin of the human neck, the capacitance value sensed by the interdigital electrode 10 changes abruptly due to the change of medium. The intelligent central control and multi-dimensional state sensing system identifies that the device is in an effective wearing state, and then controls the system to wake up from the sleep mode and enter the working mode. If no wearing signal is detected, the system maintains low-power sleep to achieve power management.

[0049] After entering the working mode, the signal acquisition component and the signal processing component cooperate to perform multi-source data acquisition. The reference electrode 15 on the silica gel pad 16 cooperates with the working electrode 14 to collect the myoelectric signal of the throat part through a differential input method; the MEMS microphone 13 synchronously picks up the acoustic signal during the swallowing process; the inertial sensor 12 real-time collects the acceleration and angular velocity data generated by the laryngeal prominence movement. In this process, the electromagnetic shield 7 covering the analog signal processing area of the main control circuit board 9 is processed by physical isolation and grounding to shield external environmental electromagnetic noise and radio frequency interference generated by the on-board Bluetooth module 11 during operation, ensuring that the analog front-end circuit outputs high signal-to-noise ratio original signals.

[0050] The intelligent central control and multi-dimensional state sensing system processes the collected data in real time. The multi-source signal synchronization module takes the time when the laryngeal prominence movement acceleration peak value detected by the inertial sensor 12 as a hard time reference, constructs a time window of a preset length, and cuts and time-aligns the continuous myoelectric signal according to the window, realizing the synchronous segmentation of mechanical movement signal and electrophysiological signal. The swallowing feature extraction module then calculates the synchronized signal segment to extract the root mean square value of the myoelectric signal, the integral myoelectric value, and the maximum displacement amplitude and movement duration of the laryngeal prominence movement, and constructs a multi-dimensional swallowing feature vector reflecting the current swallowing state.

[0051] The aspiration risk assessment and interaction module inputs the real-time constructed feature vector into the system preset normal swallowing feature threshold model for comparison operation to calculate the feature deviation. If the deviation value exceeds the preset safety threshold, the system determines that the current action has aspiration risk, immediately controls the display screen 2 to pop up a red warning window, and synchronously draws the myoelectric signal change curve and the laryngeal prominence movement trajectory curve; the LED display lamp 3 emits light signals of corresponding colors according to the system instructions, providing intuitive monitoring feedback in the form of sound and light and images.

Claims

1. A swallowing action monitoring and aspiration early warning device synchronized with pharyngeal electromyographic signals, characterized in that, include: The housing (1) has a display screen (2) fixedly connected to its top. A connection end (6) is provided on the outside of the housing (1). A connection line (4) is provided in the middle of the connection end (6). A signal acquisition component is provided on the outside of the other end of the connection line (4). The acquisition component is electrically connected to a signal processing component for receiving the acquired pharyngeal electromyography signals and analyzing and outputting the structure. The signal processing component includes a main control circuit board (9), which is disposed inside the housing (1). The top of the main control circuit board (9) is electrically connected to an interdigital electrode (10), a Bluetooth module (11), and an inertial sensor (12). The outside of the interdigital electrode (10) is electrically connected to a working electrode (14). The main control circuit board (9) is equipped with an intelligent central control and multi-dimensional state perception system.

2. The swallowing action monitoring and aspiration early warning device with synchronized pharyngeal electromyographic signals according to claim 1, characterized in that, The signal acquisition component includes a silicone pad (16), a reference electrode (15) is provided in the middle of the silicone pad (16) for contacting the skin and acquiring signals, and a working electrode (14) is provided on the outside of the silicone pad (16).

3. The swallowing action monitoring and aspiration early warning device with synchronized pharyngeal electromyographic signals according to claim 2, characterized in that, The intelligent central control and multi-dimensional state perception system includes: Wearing status monitoring module is used to determine whether the device is in an effective wearing state based on the change in capacitance value fed back by the interdigital electrode (10), and to control the system's wake-up and sleep modes; A multi-source signal synchronization module is used to align the laryngeal movement data collected by the inertial sensor (12) with the pharyngeal electromyography signals collected by the working electrode (14) and the reference electrode (15) on the time axis. The swallowing feature extraction module is used to extract swallowing time-domain and frequency-domain features based on aligned data; The aspiration risk assessment and interaction module is used to calculate the aspiration probability based on the extracted features and drive the display screen (2) to provide visual feedback.

4. The swallowing action monitoring and aspiration early warning device with synchronized pharyngeal electromyographic signals according to claim 3, characterized in that, The specific operating logic of the multi-source signal synchronization module and the swallowing feature extraction module is as follows: The multi-source signal synchronization module uses the peak moment of the laryngeal movement acceleration detected by the inertial sensor (12) as the reference point to construct a time window of a preset length, and extracts the pharyngeal electromyographic signal segments collected by the working electrode (14) and the reference electrode (15) within the time window, thereby completing the synchronous segmentation of motion signals and electromyographic signals. The swallowing feature extraction module processes the synchronously segmented signal segments to extract the root mean square value and integral electromyographic value of the pharyngeal electromyographic signal, as well as the maximum displacement amplitude and duration of the laryngeal movement data, and constructs a multidimensional swallowing feature vector.

5. The swallowing action monitoring and aspiration early warning device with synchronized pharyngeal electromyographic signals according to claim 1, characterized in that, The specific operating logic of the aspiration risk assessment and interaction module is as follows: The system has a pre-stored normal swallowing feature threshold model. The aspiration risk assessment and interaction module inputs the multi-dimensional swallowing feature vector, which is constructed in real time, into the normal swallowing feature threshold model for comparison. When the calculated feature deviation exceeds the preset safety threshold, it is determined that there is a risk of accidental inhalation. The system immediately generates a red warning signal and pops up an alarm window on the display terminal. The display terminal is configured to simultaneously plot the change curve of pharyngeal electromyography signal and the trajectory curve of laryngeal movement on the display screen (2) for medical staff to observe intuitively.

6. The swallowing action monitoring and aspiration early warning device with synchronized pharyngeal electromyographic signals according to claim 1, characterized in that, The main control circuit board (9) is electrically connected to an electromagnetic shield (7) to cover the analog signal processing area on the main control circuit board (9) and shield external electromagnetic interference and radio frequency interference generated by the Bluetooth module (11).

7. The swallowing action monitoring and aspiration early warning device with synchronized pharyngeal electromyographic signals according to claim 1, characterized in that, An LED display lamp (3) is fixedly connected to the outside of the housing (1), and the LED display lamp (3) is electrically connected to the main control circuit board (9).

8. The swallowing action monitoring and aspiration early warning device with synchronized pharyngeal electromyographic signals according to claim 2, characterized in that, The silicone pad (16) is made of self-adhesive silicone rubber.

9. The swallowing action monitoring and aspiration early warning device with synchronized pharyngeal electromyographic signals according to claim 2, characterized in that, A MEMS microphone (13) is provided in the middle of the silicone pad (16) for collecting swallowing sounds.

10. The swallowing action monitoring and aspiration early warning device with synchronized pharyngeal electromyographic signals according to claim 1, characterized in that, A mounting bracket (5) is provided at the bottom of the housing (1), and a battery (8) is fixedly connected to the outside of the main control circuit board (9).

Images