Tongue spatula system with acupoint nerve stimulation and real-time tongue muscle data acquisition functions and control method thereof for swallowing rehabilitation

Abstract

The invention relates to a tongue spatula system with acupoint nerve stimulation and real-time tongue muscle data acquisition functions and a control method for swallowing rehabilitation. The tongue spatula system comprises a tongue spatula body, a nerve stimulation module, a data acquisition module and a control unit electrically connected with the tongue spatula body and the nerve stimulation module. The core of the control method is that the control unit prestores or receives a rehabilitation target model containing target tongue muscle data parameters; in the training process, the control unit compares actual tongue muscle data acquired by the data acquisition module in real time with the model to determine deviation, and adaptively adjusts stimulation parameters of electrical stimulation applied by the nerve stimulation module through a preset control algorithm according to the deviation. A closed-loop and self-adaptive rehabilitation training process is formed by circularly executing the steps of stimulation application, data acquisition, deviation analysis and stimulation parameter self-adaptive adjustment.

Description

Technical Field

[0001] This invention relates to the field of medical devices and rehabilitation engineering technology, and in particular to a tongue depressor system with acupoint nerve stimulation and real-time tongue muscle data acquisition functions, and its control method for swallowing rehabilitation. Background Technology

[0002] Swallowing is a complex and delicate physiological activity involving the high degree of coordination of multiple nervous and muscular systems. However, in clinical practice, swallowing dysfunction caused by neurological diseases (such as stroke and Parkinson's disease), post-operative or radiotherapy / chemotherapy for head and neck tumors is a common sequela. It not only severely affects patients' nutritional intake and quality of life but can also lead to serious complications such as aspiration pneumonia. Therefore, effective and scientific swallowing rehabilitation training is crucial for patients' recovery.

[0003] In the field of existing rehabilitation technology, traditional rehabilitation training methods for dysphagia mainly rely on the professional knowledge and clinical experience of rehabilitation therapists. Therapists guide patients to perform a series of tongue and pharyngeal muscle training movements through verbal instructions and manual guidance. However, this type of method has obvious limitations: First, feedback during the training process mainly depends on the therapist's visual observation and the patient's subjective feelings, lacking objective and quantifiable data to accurately assess the quality of each movement, such as the strength of the tongue muscles and the coordination of the movements; second, the evaluation of rehabilitation effects is therefore highly subjective, making standardized tracking and comparison difficult; third, the adjustment cycle of the training program is long, and therapists usually need to complete one or more courses of treatment before they can adjust the program according to the overall situation, making it impossible to respond quickly to the patient's immediate performance during the training process.

[0004] To overcome these shortcomings, rehabilitation devices integrating sensing or electrical stimulation technologies have emerged. For example, some devices use pressure sensors to measure the pressure between the tongue and palate, providing quantitative data for assessing tongue muscle strength. Other devices employ neuromuscular electrical stimulation (NMES) technology, which stimulates target muscle groups by applying specific electrical pulses to the body surface to induce or enhance muscle contraction, assisting patients in completing training movements.

[0005] While these technologies have advanced the field of swallowing rehabilitation to some extent, they still face key technical bottlenecks in practical applications. On one hand, simple data acquisition devices are primarily used for assessment or as display tools for biofeedback training; they lack active intervention and guidance capabilities. After seeing feedback information, patients still need to rely on their own understanding and experimentation to adjust their tongue muscle movements, which has limited effectiveness for patients with poor proprioception or impaired cognitive abilities. On the other hand, most existing electrical stimulation devices employ an "open-loop" control mode. That is, therapists set a fixed set of stimulation parameters (such as intensity, frequency, and duration) for patients based on initial assessments and use these parameters throughout the training process. This open-loop mode cannot dynamically link with the patient's voluntary movement intentions and real-time performance. Regardless of whether the patient's current movement is performed well or poorly, the stimulation applied by the device remains unchanged. This results in stimulation that may be too strong, inhibiting the patient's active effort, or insufficient, failing to provide effective assistance, lacking specificity and adaptability. In short, existing technologies fail to effectively integrate real-time, objective data acquisition with dynamic, intelligent stimulation control, lacking a closed-loop feedback mechanism that can automatically optimize treatment intervention based on the patient's immediate functional performance. Summary of the Invention

[0006] To address the technical problems existing in the prior art, this invention provides a tongue depressor system with acupoint nerve stimulation and real-time tongue muscle data acquisition functions, and a control method for swallowing rehabilitation. It can quantify the patient's tongue muscle activity status in real time and intelligently and adaptively adjust nerve stimulation parameters based on the quantified data to achieve personalized and precise swallowing rehabilitation guidance training.

[0007] This invention discloses a tongue depressor system with acupoint nerve stimulation and real-time tongue muscle data acquisition functions, comprising:

[0008] Tongue pressure plate main body;

[0009] A nerve stimulation module, located on the tongue depressor body, is used to apply adjustable electrical stimulation to the user's tongue.

[0010] The data acquisition module is installed on the tongue depressor body and is used to collect the actual tongue muscle data of the user when performing swallowing-related actions in real time.

[0011] The control unit is electrically connected to the neurostimulation module and the data acquisition module;

[0012] The control unit is configured as follows:

[0013] Pre-store or receive a rehabilitation target model containing target tongue muscle data parameters;

[0014] The actual tongue muscle data collected in real time by the data acquisition module is compared with the rehabilitation target model to determine the deviation between the two; and, based on the deviation, the stimulation parameters of the electrical stimulation applied by the nerve stimulation module are adaptively adjusted to guide the user's tongue muscle movement to approach the rehabilitation target model.

[0015] Specifically, in the tongue depressor system technical solution of the present invention, firstly, the control unit, as the central hub of the system, pre-stores or is set by the therapist a rehabilitation target model as a training benchmark. This model includes tongue muscle data parameters for swallowing-related movements under ideal conditions, such as target pressure peaks and pressure distribution patterns. When the user places the tongue depressor body, which includes a nerve stimulation module and a data acquisition module, into their mouth and begins training, the data acquisition module (e.g., composed of a ring-shaped pressure sensor array) begins to work continuously, capturing the actual tongue muscle data generated by the user's tongue movements in real time and sending these data signals to the control unit. After receiving the actual tongue muscle data, the control unit's internal algorithm logic immediately compares and analyzes the data with the pre-stored rehabilitation target model, thereby accurately calculating the multi-dimensional deviations between the actual movements and the target movements, such as insufficient strength, poor coordination, or slow reaction. This step is a key link in achieving closed-loop control. Subsequently, based on the calculated deviations, the control unit generates new control commands according to a preset control scheme (such as PID control or expert system rules) and sends them to the nerve stimulation module. The neurostimulation module (e.g., composed of an electrode contact array and a stimulation signal generation circuit) adaptively adjusts various parameters of its output electrical stimulation, such as stimulation intensity, frequency, pulse width, or electrode activation mode, according to the instruction. The adjusted electrical stimulation is applied to a specific area of ​​the user's tongue, aiming to precisely guide, assist, or strengthen tongue muscle movements in the next iteration or cycle, thereby prompting the user's tongue muscle movements to continuously approach the rehabilitation target model. The entire process forms a continuous dynamic closed loop of "acquisition-comparison-judgment-adjustment-stimulation," enabling the system to respond to the user's performance in real time and provide personalized interventions.

[0016] According to the technical solution of the present invention, the tongue depressor body is integrally molded from medical-grade silicone, and its tongue-facing surface is divided into a central nerve stimulation area and a peripheral data acquisition area. The nerve stimulation module is an array of electrode contacts disposed in the central nerve stimulation area, and the data acquisition module is a ring-shaped pressure sensor array disposed in the peripheral data acquisition area and surrounding the electrode contact array. Specifically, in this solution, the tongue depressor body is made of medical-grade silicone with good biocompatibility through an integral molding process. This ensures the safety of the device when in direct contact with the user's mouth, avoids adverse physiological reactions, and the integral molding structure also gives it good waterproof performance, facilitating cleaning and disinfection, and meeting the hygiene requirements of medical devices. More importantly, this solution clearly divides the central nerve stimulation area and the peripheral data acquisition area on the tongue-facing surface of the tongue depressor body. The nerve stimulation module, specifically an electrode contact array, is disposed in the central area; while the data acquisition module, specifically a ring-shaped pressure sensor array, is disposed in the peripheral area, forming a surrounding layout of the central electrode contact array. This spatial partitioning design effectively isolates the application of stimulation signals from the sensing of electromyographic signals. The electrical stimulation signal is primarily concentrated in the central region, acting on acupoints or neuromuscular nodes on the tongue, while pressure data is comprehensively captured by the peripheral sensor array. This layout significantly reduces electromagnetic interference or crosstalk caused by the electrical stimulation signal to the weak physiological signals collected by the pressure sensors, thus ensuring the signal-to-noise ratio and accuracy of the collected tongue muscle pressure data. High-precision raw data input is fundamental for the effective operation of subsequent closed-loop control algorithms; therefore, this structural design provides a physical guarantee for the accurate and reliable operation of the entire system. Furthermore, the ring-shaped pressure sensor array can simultaneously collect pressure information generated during tongue movement from multiple directions, not only measuring the total pressure magnitude but also analyzing the pressure distribution pattern and the migration trajectory of the pressure center point, providing a rich data foundation for subsequent multi-dimensional motion coordination analysis.

[0017] According to the technical solution of the present invention, the control unit determines the deviation in the following ways:

[0018] The force dimension deviation is calculated by comparing the difference between the actual peak pressure and the target value; and...

[0019] The coordination dimension deviation is determined by calculating the cosine similarity between the actual pressure distribution map and the target map, or by calculating the average Euclidean distance between the actual pressure center point trajectory and the target trajectory.

[0020] Specifically, this approach breaks down complex tongue muscle movement into two key dimensions for quantitative assessment: strength and coordination. For strength, the approach directly compares the user's actual peak pressure during movement with the target peak pressure preset in the rehabilitation model, calculating the difference to quantify insufficient or excessive strength. This is an intuitive and effective assessment method that directly reflects whether the user's maximum tongue muscle contraction capacity meets the standard. For coordination, the approach provides two more refined calculation paths: First, by calculating the cosine similarity between the real-time collected pressure distribution map and the target map, the complex spatial pressure distribution pattern is transformed into a single numerical value, which quantifies the degree of conformity between the actual movement pattern and the standard pattern. Second, by calculating the average Euclidean distance between the trajectory of the actual center of pressure (CoP) and the target trajectory, the accuracy of the direction, range, and stability of tongue movement is assessed. Both methods transform motor coordination issues, which previously relied on the therapist's subjective observation, into objective, repeatable, and quantitative indicators. By specifying deviations into strength and coordination dimensions, the system can perform more precise diagnostic analysis of users' functional impairments, thus providing clear direction for subsequent adaptive parameter adjustments. For example, the system can clearly distinguish whether a user simply lacks strength or has sufficient strength but inaccurate motor control, thereby implementing more targeted stimulation programs, avoiding general and inefficient adjustment methods, and improving the targeting of rehabilitation training.

[0021] According to the technical solution of the present invention, the adjustable stimulation parameters include a stimulation intensity adjustable from 0.1mA to 20mA, a stimulation frequency adjustable from 1Hz to 150Hz, and an electrode activation mode that allows independent control of the on / off state of each electrode in the array. Specifically, in this solution, the adjustable stimulation parameters include a stimulation intensity adjustable from 0.1mA to 20mA, a stimulation frequency adjustable from 1Hz to 150Hz, and an electrode activation mode that allows independent control of the on / off state of each electrode in the array. First, the wide and finely adjustable range of stimulation intensity (0.1mA to 20mA) and frequency (1Hz to 150Hz) gives the system better flexibility, enabling it to adapt to the differentiated needs of different users and different stages of rehabilitation. For example, lower frequencies can be used for muscle relaxation or activation, higher frequencies can be used for muscle strengthening training, and fine adjustment of stimulation intensity ensures that while providing effective assistance, the user's comfort and safety are maximized. Second, and most importantly, the "electrode activation mode" control dimension is introduced. This indicates that the control unit can not only adjust the overall intensity and speed of stimulation, but also control the spatial distribution of the stimulation. By independently controlling the on / off state and even the output weight of each contact point in the electrode array, the system can generate stimulation areas of arbitrary shape and location. This function, combined with the coordination dimension deviation calculated in the aforementioned scheme, enables highly targeted reinforcement training. When the system detects insufficient muscle activation in a specific area of ​​the user's tongue (manifested as a low-pressure area in the pressure distribution map), it can precisely enhance the output of the corresponding electrode in that area, thereby achieving targeted and intensive stimulation of weak muscle groups. This precise spatial dimension control capability is key to guiding and correcting complex movement patterns, better enhancing the system's training effect in improving tongue muscle coordination.

[0022] According to the technical solution of the present invention, the control unit further includes human-computer interaction software, which has: a therapist-side interface for therapists to set the rehabilitation goal model and safety thresholds for stimulation parameters, and a patient-side interface providing real-time feedback and training guidance to the user. Specifically, this solution aims to add a specific human-computer interaction software and divide it into a clearly defined therapist-side interface and a patient-side interface. The therapist-side interface empowers professional rehabilitation therapists with the ability to perform top-level design and safety monitoring of the entire rehabilitation process. The therapist can tailor a rehabilitation goal model for the user based on their professional assessment and set safety thresholds for stimulation parameters. This design ensures that the application of the technology is always under the guidance and supervision of professionals, guaranteeing the scientific nature and safety of the rehabilitation plan, making the system a professional tool to enhance therapists' capabilities, rather than a completely automated replacement. On the other hand, the patient-side interface focuses on improving the user's experience and participation during training. By transforming abstract tongue muscle data into intuitive visual or auditory feedback and providing clear training guidance, it enables users to more easily understand training tasks and perceive their own motor performance in real time. This instant feedback mechanism, similar to biofeedback therapy, helps users develop a clearer sense of proprioception and actively control and adjust their tongue muscle movements. Simultaneously, presenting feedback information through gamification significantly enhances the enjoyment of training, thereby improving user compliance and initiative. This is a crucial factor in ensuring the final rehabilitation outcome, especially for rehabilitation processes that require long-term commitment. Therefore, this software design, through its rational functional division, effectively combines professionalism, safety, and user compliance.

[0023] Based on the above system scheme, the present invention also provides a control method for swallowing rehabilitation, which is applied to the tongue depressor system described in the present invention, and includes the following steps:

[0024] S1. Set a rehabilitation target model that includes target tongue muscle data parameters;

[0025] S2. Control the nerve stimulation module to apply electrical stimulation to the user's tongue, and at the same time, control the data acquisition module to collect the user's actual tongue muscle data in real time when attempting to complete the target action;

[0026] S3. Analyze the real-time collected actual tongue muscle data and the rehabilitation target model to calculate the deviation between the two;

[0027] S4. Based on the aforementioned deviation, automatically adjust the stimulation parameters for the electrical stimulation used in the next training cycle; and,

[0028] S5. Repeat steps S2 to S4 to form a closed-loop, adaptive rehabilitation training process.

[0029] Specifically, in the control method of this invention, step S1 is the initialization stage, which sets a clear "target point" for the entire closed-loop control process, namely a rehabilitation target model containing quantifiable indicators. Step S2 is the closed-loop initiation and execution stage, where the two core functions of the system—neural stimulation and data acquisition—are activated simultaneously. The neural stimulation module applies electrical stimulation to guide or assist the user, while the data acquisition module objectively records the user's actual performance. Step S3 is the feedback information generation stage, where the actual tongue muscle data collected in real time is compared with the set target model, and the deviation between the two is calculated, thereby quantifying the user's functional performance into an error signal that the control system can judge. Step S4 is the core innovation of this method, which embodies the intelligent adaptive characteristics of the system. That is, the control system does not execute a static program, but automatically adjusts the electrical stimulation parameters for the next training cycle based on the deviation calculated in step S3, achieving dynamic optimization of the treatment intervention. Finally, step S5, by cyclically executing steps S2 to S4, strings together the single adjustment behavior into a continuous and iterative rehabilitation training process. This cyclical mechanism ensures that each training movement is optimized and corrected based on the previous performance, thus forming a complete, adaptive closed-loop rehabilitation process.

[0030] According to the technical solution of the present invention, in step S1, the rehabilitation target model is set as a curve function that includes the change of target pressure over time. The expected trajectory of the target pressure center point changing over time The time-series data template. Specifically, in this scheme, the rehabilitation target model is no longer a single or set of static numerical values, but is defined as a curve function that contains the target stress changing over time. The expected trajectory of the target pressure center point changing over time The time-series data template. This definition method has significant technical advantages. First, it expands the rehabilitation goal from an endpoint state (such as reaching a certain pressure value) to a complete dynamic process. A functional swallowing-related action itself has complex temporal characteristics, including the establishment, maintenance, and release of force, as well as the precise movement of the tongue in space. The curve can accurately depict the ideal muscle force output pattern throughout the entire movement cycle, and The trajectory depicts the standard path of tongue movement in space. Secondly, this multi-dimensional, dynamic template provides a more comprehensive and refined benchmark for subsequent deviation calculations. The system can compare not only values ​​at a single instant but also waveform similarity and trajectory consistency over the entire time series. This allows the system to evaluate the user's performance at various stages of the movement (such as force exertion speed, plateau stability, and relaxation control), thereby guiding and correcting the entire movement process, rather than simply optimizing a single parameter. Therefore, defining the target model as a time-series data template is a technical prerequisite for achieving high-fidelity guidance and evaluation of complex tongue muscle movements, enhancing the scientific rigor and precision of the control method.

[0031] According to the technical solution of the present invention, in step S4, the process of automatically adjusting the stimulation parameters based on the deviation includes:

[0032] When the calculated force dimension deviation exceeds the first preset threshold, the electrical stimulation intensity of the next cycle is automatically increased.

[0033] When the calculated coordination dimension deviation exceeds the second preset threshold, the electrode activation mode is automatically adjusted to enhance the electrode output corresponding to the pressure-weak area.

[0034] When the calculated deviation in response speed dimension exceeds the third preset threshold, the electrical stimulation frequency of the next cycle is automatically increased.

[0035] Specifically, this scheme elaborates on the logic of adaptive parameter adjustment based on deviation in step S4, constructing a rule-based control scheme based on multi-dimensional deviation input. This scheme establishes a clear causal relationship between the calculated deviations in different dimensions and the corresponding stimulation parameter adjustment actions. When the strength dimension deviation exceeds the first preset threshold, it indicates that the user's tongue muscle strength is significantly insufficient, and the system directly increases the electrical stimulation intensity in the next cycle—the most direct and effective means of strength compensation and strengthening. When the coordination dimension deviation exceeds the second preset threshold, it indicates that the user's main problem lies in imprecise motor control. In this case, the system does not indiscriminately increase overall stimulation but intelligently adjusts the electrode activation mode, specifically strengthening the electrode output corresponding to areas of weak pressure, achieving targeted activation of specific muscle groups, thereby correcting unbalanced force patterns. When the response speed dimension deviation exceeds the third preset threshold, it indicates low neuromuscular excitability, and the system increases the electrical stimulation frequency to improve nerve recruitment efficiency and muscle reaction speed. The superiority of this control logic lies in its high degree of specificity. It simulates the clinical reasoning process of experienced rehabilitation therapists, enabling the selection of the most appropriate intervention method based on the nature of the problem (whether it is a strength, coordination, or reaction speed issue). This targeted approach to regulation avoids ineffective or inappropriate stimulation, ensuring that each parameter adjustment has a clear therapeutic purpose, thereby significantly improving the efficiency of closed-loop control and the effectiveness of rehabilitation training.

[0036] According to the technical solution of the present invention, in step S4, the adjustment amount ΔI of the stimulation intensity is calculated by a PID control algorithm, wherein the proportional term (P) responds to the current deviation, the integral term (I) responds to the historical accumulated deviation, and the derivative term (D) responds to the rate of change of the deviation, so as to achieve smooth and stable control of tongue muscle strength. In other words, this solution provides a more precise and stable algorithm implementation method for the stimulation intensity adjustment in step S4—the PID (proportional-integral-derivative) control algorithm. PID control provides a continuous and smooth control mechanism. Specifically, the proportional term (P) makes the adjustment amount of the stimulation intensity proportional to the current pressure deviation, ensuring the system's rapid response to deviations. The integral term (I) accumulates historical deviations; if the user continues to be unable to reach the target pressure, even if the single deviation is small, the cumulative effect of the integral term will cause the stimulation intensity to steadily increase until the steady-state error is eliminated. This effectively avoids the problem of the user being in a "bottleneck period" for a long time. The differential term (D) responds to the rate of change of the deviation. When the deviation is rapidly decreasing (i.e., the user's performance is improving), the differential term exerts an inhibitory effect, preventing overshoot caused by excessively rapid adjustment of stimulus intensity, thus avoiding overstimulation of the user and making the entire adjustment process more stable. In summary, the introduction of the PID control algorithm applies classical engineering control theory to the rehabilitation process, making the system's guidance and assistance of tongue muscle strength no longer a simple switch between "strong" and "weak," but a dynamic, smooth, and predictive fine-tuning. This not only improves the user's comfort during training but also makes the growth process of tongue muscle strength closer to a natural physiological learning pattern, thereby achieving a more stable and solid rehabilitation effect.

[0037] According to the technical solution of the present invention, the method further includes the steps of: recording and storing a rehabilitation log, wherein the rehabilitation log contains the original tongue muscle data, multi-dimensional deviation values, and adjusted stimulation parameters for each training cycle; and automatically generating a visualized rehabilitation report containing curves showing changes in strength, coordination, and reaction speed based on the rehabilitation log. Specifically, this solution adds the functions of data recording, storage, and report generation to the closed-loop training process, thereby forming a complete closed loop of rehabilitation management. The method requires the system to record and store a rehabilitation log containing the original tongue muscle data, multi-dimensional deviation values, and adjusted stimulation parameters for each training cycle. This detailed data recording itself has significant value, providing a completely transparent and traceable objective record for each training process. On this basis, the method further requires the system to automatically generate a visualized rehabilitation report based on the rehabilitation log, which contains curves showing changes in key indicators such as strength, coordination, and reaction speed. The technical effects brought about by this function are multifaceted: First, it provides an objective and quantitative basis for the evaluation of rehabilitation effects. Therapists can break away from the traditional assessment methods that rely on subjective feelings and qualitative descriptions, and intuitively grasp the user's rehabilitation progress, plateaus, or regression by analyzing data curves. Secondly, these quantitative reports provide strong data support for therapists to adjust macro-level rehabilitation plans, making their judgments more scientific and evidence-based. Finally, visualized progress reports can also serve as positive feedback to users, allowing them to clearly see the results of their efforts, thereby enhancing their confidence in rehabilitation and their motivation for continued participation. Therefore, this step, by digitizing and visualizing the training process, completes a macro-level closed loop from "training-assessment-adjustment," helping to improve the scientific management level and overall efficiency of the entire rehabilitation process.

[0038] The technical effects of this invention, a tongue depressor system with acupoint nerve stimulation and real-time tongue muscle data acquisition functions, and its control method for swallowing rehabilitation, include: First, because the system includes a data acquisition module, it can acquire the user's actual tongue muscle data in real time, making the originally subjective and difficult-to-quantify tongue muscle movement process objective and quantifiable. Second, the control unit in the system is configured to compare the acquired actual data with a preset rehabilitation target model and calculate the deviation. This setting transforms rehabilitation training from a vague goal-guided process into a control problem with a clear error signal, providing a prerequisite for precise control. Third, the core of this technical solution lies in the fact that the control unit can adaptively adjust the stimulation parameters of the nerve stimulation module according to the calculated deviation. This technical feature establishes a real-time feedback pathway from "motor performance" to "stimulation intervention," breaking the traditional open-loop stimulation device model where stimulation parameters are fixed once set. When insufficient tongue muscle strength is detected, the system can automatically increase the stimulation intensity; when coordination problems are detected, the electrode activation mode can be adjusted to strengthen weak muscle groups. This adaptive adjustment based on real-time deviation makes each stimulation highly targeted. Ultimately, through the cyclical execution of this closed-loop process, the entire rehabilitation training evolves into a continuous iterative and self-optimizing process. This process not only more effectively guides and strengthens the user's tongue muscle function, causing its movement pattern to continuously approach the ideal model, but also records all data throughout the training process (including tongue muscle performance data and adjusted stimulation parameters), providing therapists with detailed and objective quantitative evidence for phased assessments and macro-level program adjustments. In summary, this technical solution, by integrating real-time data acquisition, deviation analysis, and adaptive stimulation parameter adjustment functions, constructs a closed-loop feedback control system, thereby achieving personalized and adaptive swallowing rehabilitation training and improving the accuracy and efficiency of training. Attached Figure Description

[0039] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0040] Figure 1 This is an overall functional framework diagram of the tongue depressor system of the present invention;

[0041] Figure 2 This is a schematic diagram of the tongue depressor device of the present invention;

[0042] Figure 3 This is a hardware circuit connection block diagram (schematic diagram) of the tongue depressor system of the present invention.

[0043] Figure 4 This is a flowchart of the control method of the present invention;

[0044] Figure 5 This is a schematic diagram of the rehabilitation target model of the present invention (typical variation diagram of the target pressure curve);

[0045] Figure 6 This is a schematic diagram of the rehabilitation target model of the present invention (a schematic diagram of the CoP trajectory formation process);

[0046] Figure 7 This is a logic diagram of the present invention based on adjusting stimulus parameters according to deviation;

[0047] Figure 8 This is a schematic diagram of the human-computer interaction interface of the present invention. Detailed Implementation

[0048] To more clearly illustrate the technical solution of the present invention, the present invention will be described in detail below with reference to the accompanying drawings and specific embodiments. The embodiments described in this section are intended to illustrate the present invention by way of example, and are not intended to limit the present invention. Those skilled in the art, under the guidance of the present invention, can make various modifications or additions to the following embodiments, and such modifications or additions should all fall within the protection scope of the present invention.

[0049] Example 1

[0050] like Figure 1 As shown, this embodiment provides a tongue depressor system with acupoint nerve stimulation and real-time tongue muscle data acquisition functions, referring to... Figure 1 The system's overall functional block diagram shown is mainly composed of three parts: a tongue depressor device 1, a data processing and control host 2, and human-computer interaction software 3 running on the host. Each part forms an organic functional whole through control signal flow and data signal flow.

[0051] like Figure 2 As shown, the tongue depressor device 1 is the core hardware unit that directly contacts the patient's mouth and performs nerve stimulation and data acquisition functions. The device mainly includes a tongue depressor body 10, a nerve stimulation module 11 integrated thereon, and a data acquisition module 12.

[0052] The tongue depressor body 10 is made of medical-grade liquid silicone material conforming to ISO 10993 biocompatibility standards, manufactured through a precision injection molding process. The body is flat, with a slight curvature conforming to the anatomical curve of the human palate. It is slightly wider at the front (approximately 15-20 mm) and narrows to 10-12 mm at the rear, with a total length of approximately 35-45 mm and a thickness controlled within the range of 3-5 mm. This size design ensures sufficient functional area while minimizing the patient's foreign body sensation and nausea reflex. The silicone material has a Shore hardness of 20-40, balancing sufficient structural strength with good flexibility, allowing the tongue depressor to adapt to individual differences in the oral cavity morphology of different patients to a certain extent.

[0053] The tongue depressor body 10 has its tongue-facing surface divided into two functional areas: a central nerve stimulation area 101 and a surrounding data acquisition area 102. This partitioned design achieves spatial independence between the stimulation and sensing functions, effectively avoiding mutual interference of electrical signals while ensuring the coordinated operation of both functions. Inside the tongue depressor body 10, a flexible printed circuit board made of polyimide substrate is embedded. This circuit board carries the connection lines for all electronic components and is integrated with the silicone body through a packaging process.

[0054] The nerve stimulation module 11 is located within the central nerve stimulation area 101 and mainly consists of an electrode contact array 111 and a stimulation signal generation circuit 112. The electrode contact array 111 adopts a 4×4 or 5×5 matrix layout, with a total of 16 or 25 independent micro-electrode contacts. Each electrode contact has a diameter of approximately 1.5-2 mm and is made of medical-grade SUS316L stainless steel with a mirror-polished surface to reduce mechanical stimulation of the tongue mucosa. The electrode contacts protrude slightly from the silicone surface by about 0.1-0.2 mm; this micro-protrusion design ensures good electrical contact with the moist tongue mucosa while avoiding excessive pressure. The spatial layout of the array is precisely calculated to cover the surface projection areas of key acupoints such as the Lianquan, Jinjin, and Yuye acupoints under the tongue, as well as the nerve innervation areas of major swallowing-related muscle groups such as the genioglossus and styloglossus muscles.

[0055] The stimulation signal generation circuit 112 is integrated into the handle portion 113 of the tongue depressor device. This handle is made of the same medical-grade silicone material as the main body, forming a sealed circuit cavity. The core components of the circuit include a 32-bit ARM Cortex-M4 microcontroller, a 16-channel 12-bit digital-to-analog converter (DAC), and 16 independent programmable constant current source drive circuits. The microcontroller receives digital control commands from the data processing and control host 2 and parses these commands in real time to generate corresponding PWM control signals. The DAC converts the digital signals into analog voltage signals, which then drive the constant current source circuits to output precisely controlled current pulses to the selected electrode contacts. The constant current source architecture is designed to overcome individual differences in the impedance of the tongue mucosa among different patients, ensuring that the actual electrical stimulation dose applied to the tissue remains constant, thereby achieving a consistent physiological response.

[0056] The stimulation signal generation circuit 112 supports high-precision dynamic adjustment of multiple stimulation parameters. Stimulation intensity is achieved by adjusting the output current of the constant current source, with an adjustment range of 0.1 mA to 20 mA and a step accuracy of 0.1 mA, meeting the full range of needs from weak sensory threshold stimulation to strong functional electrical stimulation. Stimulation frequency is achieved by controlling the pulse generation time interval, with an adjustment range of 1 Hz to 150 Hz. 1-10 Hz is suitable for muscle relaxation and blood circulation promotion, 20-50 Hz for muscle contraction training, and 80-150 Hz for enhancing neural excitability. Pulse width can be adjusted from 50 microseconds to 1000 microseconds; optimizing the pulse width can minimize patient discomfort while ensuring stimulation effectiveness. The stimulation waveform supports various forms, including symmetrical biphasic square waves, asymmetrical biphasic square waves, and sine waves. The biphasic square wave is preferred because it effectively prevents charge accumulation in tissues. Electrode activation mode allows for independent control of the on / off state of each electrode in the array, as well as the weighted distribution of its output current, thereby enabling precise positioning, dynamic movement, or range adjustment of the stimulation area.

[0057] The data acquisition module 12, located within the peripheral data acquisition area 102, mainly consists of a ring-shaped pressure sensor array 121 and a data acquisition and processing circuit 122. The ring-shaped pressure sensor array 121 employs a distributed layout, with 16 independent pressure sensing units evenly distributed along an elliptical loop around the periphery of the central nerve stimulation area. Each sensing unit uses a flexible piezoresistive thin-film sensor with a thickness of only 0.2 mm, made of conductive polymer composite material, whose resistance changes inversely with the applied pressure. The sensor's pressure measurement range is 0-50 Newtons, with a resolution of 0.1 Newtons and a response time of less than 5 milliseconds, accurately capturing transient pressure changes generated by tongue muscle movement. The ring-shaped layout design allows the system to monitor the tongue's movement in all directions from all angles. By analyzing the differences and change patterns in the readings of each sensing unit, the position coordinates and movement trajectory of the pressure center point can be calculated in real time, thereby quantitatively assessing the coordination, directional accuracy, and range of motion of the tongue muscles.

[0058] The data acquisition and processing circuit 122 is also integrated into the handle portion 113, and is jointly controlled by the microcontroller along with the stimulation signal generation circuit 112. The main components of this circuit include a 16-channel analog multiplexer, a pre-amplifier signal conditioning circuit, and a 16-bit high-precision analog-to-digital converter (ADC). The microcontroller controls the multiplexer to sequentially connect each pressure sensing unit, converting its weak resistance change signal into a voltage signal. The pre-amplifier signal conditioning circuit performs low-pass filtering on these voltage signals to eliminate high-frequency noise interference, and amplifies the signal to the optimal input range of the ADC using a precision operational amplifier. The ADC cyclically scans all 16 sensing units at a sampling frequency of 100 Hz to ensure the capture of the complete dynamic process of rapid tongue movements. The acquired digital data is processed in real time by a built-in algorithm to extract key quantitative indicators such as peak pressure, pressure integral, pressure distribution map, pressure center point trajectory, and neuromuscular reaction time.

[0059] Furthermore, the data processing and control host 2 is the core controller of the entire system, undertaking key functions such as data processing, algorithm calculation, parameter adjustment, and user interaction. For example... Figure 3 The hardware circuit connection diagram shown illustrates that the host computer adopts an embedded industrial panel PC hardware form factor, equipped with an ARM Cortex-A53 quad-core processor with a clock speed of 1.8 GHz, 4GB of LPDDR4 RAM, and 64GB of eMMC storage, which can meet the performance requirements of real-time data processing and complex algorithm calculations. The host computer integrates a 10.1-inch capacitive multi-touch display with a resolution of 1920×1200 pixels and supports 256 color depth, providing a good hardware foundation for data visualization and user interaction.

[0060] The communication between the main unit and the tongue depressor device 1 is achieved via Bluetooth Low Energy (BLE) 5.0 protocol, with a communication distance of up to 10 meters and a data transmission rate of 2 Mbps, fully meeting the bandwidth requirements for real-time control and data feedback. The Bluetooth module integrated within the tongue depressor device uses a Nordic nRF52840 chip solution, supporting 128-bit AES encryption to ensure the security and privacy protection of medical data transmission. The wireless connection provides patients with greater freedom of movement during training, avoiding the potential restrictions and safety hazards associated with wired connections.

[0061] The host computer's software system comprises a real-time data processing engine, a closed-loop control algorithm engine, a data storage management system, and a communication protocol stack. The real-time data processing engine is responsible for real-time analysis, filtering, and feature extraction of raw sensor data from the tongue depressor device, with a processing latency controlled within 10 milliseconds to ensure timely response to changes in the patient's tongue muscle movements. The closed-loop control algorithm engine implements various adaptive adjustment schemes, including fuzzy logic control based on expert rules, PID proportional-integral-derivative control, and predictive control algorithms based on machine learning, allowing therapists to select the most suitable control scheme based on the patient's specific condition.

[0062] In addition, the human-computer interaction software 3 runs on the data processing and control host 2, and mainly includes two relatively independent subsystems: the therapist-side operation interface 31 and the patient-side training interface 32. The therapist-side operation interface 31 provides functions such as patient management, treatment plan formulation, parameter setting, real-time monitoring, and report generation. The patient management module supports the creation and maintenance of patient files, recording the patient's basic information, medical history, assessment results, and historical training data. The treatment plan formulation module has a built-in library of rehabilitation target models covering various standard swallowing movements, such as tongue tip elevation model, tongue base retraction model, circumflex tongue movement model, and lateral sweeping model. Each model defines a corresponding target pressure time series function. and the trajectory function of the target pressure center point Therapists can choose a suitable standard model from a model library, or create a personalized target model by adjusting key parameters or guiding the patient to perform baseline movements.

[0063] The parameter setting module allows therapists to set personalized stimulation parameter safety thresholds and initial parameter combinations for each patient. The system employs multiple safety check mechanisms to ensure that the set parameters are always within a safe range. The real-time monitoring module displays the patient's real-time tongue muscle data in a high-refresh-rate graphical format, including dynamic pressure curves, real-time pressure distribution heatmaps, pressure center point trajectory diagrams, and a side-by-side comparison with the target model, enabling therapists to intuitively observe the patient's training status and the system's adjustment process. The report generation module automatically summarizes and analyzes the data after each training session, generating a detailed rehabilitation report that includes quantitative assessment indicators, progress trend analysis, and suggested adjustment plans.

[0064] The patient-side training interface 32 can adopt a design similar to an animated interface, using simple and intuitive animated guidance, real-time visual feedback, and incentive mechanisms to improve patient participation and compliance. The training guidance function transforms complex tongue muscle movement requirements into easily understandable visual instructions, such as demonstrating the key points of the target movement through a dynamic 3D tongue model. The real-time feedback function transforms abstract pressure data into intuitive visual elements, such as a virtual cursor that moves in real-time based on the position of the tongue's pressure center point. Patients need to control their tongue muscle movements to guide the cursor to the target area on the screen; this interactive method helps enhance the fun and sense of accomplishment during training.

[0065] The distributed monitoring structure of the annular pressure sensor array 121 can not only measure the overall strength performance of the tongue muscles, but also capture the spatial distribution characteristics and temporal evolution of tongue muscle movement, which helps to provide an objective data basis for quantitatively assessing tongue muscle coordination. By calculating the position and trajectory of the pressure center point in real time, the system can identify subtle defects in tongue muscle movement such as deviation, tremor, and uneven force.

[0066] This embodiment constructs a fully functional intelligent swallowing rehabilitation training platform using the aforementioned architecture. The system not only achieves integrated stimulation and monitoring at the hardware level, but also establishes an adaptive adjustment scheme based on real-time physiological feedback at the control level, providing a technical foundation and hardware support for the subsequent implementation of closed-loop rehabilitation control methods.

[0067] Example 2

[0068] This embodiment, based on the tongue depressor system hardware platform described in Embodiment 1, elaborates on the specific implementation of a closed-loop control method for swallowing rehabilitation. This control method is an adaptive adjustment process that uses the patient's real-time tongue muscle function as feedback signals, nerve stimulation parameters as control variables, and approximating a preset rehabilitation goal as its control objective. For example... Figure 4 The flowchart of the closed-loop control method shown is a complete automated rehabilitation training system constructed through five progressive steps, specifically steps S1 to S5.

[0069] Step S1: Technical implementation of the rehabilitation goal model setting stage (setting a rehabilitation goal model that includes target tongue muscle data parameters) is as follows:

[0070] The rehabilitation goal model setting stage is the foundation of the entire closed-loop control method, and its technical implementation relies on a dedicated algorithm module running on the data processing and control host 2. This stage is completed by qualified rehabilitation therapists through the therapist-side interface 31 of the human-computer interaction software 3. The system's built-in standardized rehabilitation goal model library adopts a hierarchical and categorized data organization structure, establishing a set of standard models covering various typical training movements based on the pathological type, severity, and rehabilitation stage of swallowing disorders. These models include, but are not limited to, the tongue tip resisting the palate model, the tongue body lateral movement model, the tongue root retraction model, the circular sweeping model, and the tongue-pointing model. Each model is based on statistical analysis of a large amount of clinical data, possessing high scientific validity and universality.

[0071] like Figure 5 The diagram shown illustrates the rehabilitation goal model. Each rehabilitation goal model is mathematically defined as a multi-dimensional time series data template, mainly containing three core parameter functions: the duration of the standard movement (T), the curve function of the target pressure changing over time, and so on. And the expected trajectory function of the target pressure center point changing over time. Taking the tongue tip resisting the palate movement model as a specific example, its standard duration T is set to 3 seconds, and the target pressure curve is... It exhibits a typical trapezoidal distribution, comprising a 0.5-second pressure rise phase, a 2-second pressure maintenance phase, and a 0.5-second pressure fall phase. The pressure peak is set within the range of 5 to 25 Newtons based on the patient's age, gender, and functional assessment results. Target pressure center point trajectory. It is defined as a relatively stable coordinate region located at about 1 / 3 of the front end of the tongue depressor, with an allowable offset range of ±3 mm. This setting ensures both the accuracy of the movement and takes into account the reasonable variation range of normal physiological movements.

[0072] Therapists can select the most suitable standard model from the model library based on the patient's specific assessment results, or personalize existing models using the system's customization functions. These customization functions support fine-tuning of key model parameters, including adjusting the target peak pressure, modifying movement duration, redefining pressure distribution patterns, and setting personalized center point trajectories. Furthermore, the system provides a baseline movement recording function, allowing therapists to guide patients through a relatively standard movement. The system records the pressure data characteristics of this movement in real time and uses this as a basis to generate a patient-specific initial target model. This personalized modeling approach better adapts to the functional foundation and rehabilitation potential of different patients. After the model is set up, the relevant data is encoded and stored in the memory of the data processing and control host, providing a benchmark reference for subsequent real-time comparative analysis.

[0073] Step S2: Execution mechanism for stimulation initiation and real-time data acquisition (controlling the neurostimulation module to apply electrical stimulation to the user's tongue, and simultaneously controlling the data acquisition module to acquire real-time data on the user's actual tongue muscles as they attempt to complete the target action). Specifically as follows:

[0074] The stimulation initiation and real-time data acquisition phase is the starting point for the closed-loop control method. This phase achieves complete synchronization between nerve stimulation application and tongue muscle data acquisition through timing control and multi-threaded parallel processing mechanisms. After the patient correctly places the tongue depressor device 1 in their mouth and activates the training program, the data processing and control host 2 sends a control command data packet containing initial stimulation parameters to the tongue depressor device via a Bluetooth communication interface.

[0075] During step S2, upon receiving a control command, the microcontroller within the tongue depressor device immediately activates the parallel operation mode of the nerve stimulation module 11 and the data acquisition module 12. Based on the received parameter configuration, the nerve stimulation module generates a corresponding basic waveform signal via a digital-to-analog converter. This signal, after being amplified by a constant current source drive circuit, outputs precisely controlled current pulses to the activated specific electrodes in the electrode contact array 111. The initial stimulation parameters are typically set to a low guiding level, with typical values ​​of 2-5 mA intensity, 20-30 Hz frequency, and 200-300 microsecond pulse width. This parameter configuration produces a noticeable sensory effect without causing discomfort to the patient, primarily activating the patient's voluntary movement intention and providing guidance for the direction of movement. Simultaneously, the data acquisition module begins continuous operation at a fixed sampling frequency of 100 Hz. The microcontroller sequentially connects the 16 sensing units in the annular pressure sensor array 121 via a multiplexer, completing real-time acquisition of actual tongue muscle data as the user attempts to perform the target movement.

[0076] Step S3: Algorithm logic for real-time deviation analysis and calculation (analyzing the actual tongue muscle data collected in real time and the rehabilitation target model to calculate the deviation between the two), as follows:

[0077] The core algorithm module of step S3 runs on the CPU of the data processing and control host, employing a multi-threaded parallel computing architecture to ensure real-time performance. This step is responsible for analyzing the actual tongue muscle data collected in real time in step S2 and the rehabilitation target model set in step S1 to calculate the deviation between the two. During operation, after receiving the real-time pressure data stream from the tongue depressor device, the algorithm module first performs data preprocessing, including outlier detection and removal, data smoothing filtering, and coordinate system normalization. Then, it performs a multi-dimensional quantitative comparative analysis of the processed actual tongue muscle data and the preset rehabilitation target model data.

[0078] Deviation calculations are performed in parallel across three key dimensions, forming a comprehensive deviation assessment vector. The force dimension deviation calculation employs two evaluation metrics: peak pressure difference and pressure integral difference. The peak pressure difference is defined as:

[0079]

[0080] in This represents the peak pressure in the target model. This indicates the maximum pressure value actually measured during the current action cycle.

[0081] The integral difference of pressure is defined as:

[0082]

[0083] This indicator reflects the difference in the total work done by the tongue muscles throughout the entire movement. The weighted combination of these two indicators constitutes the final force dimension deviation.

[0084]

[0085] The weighting coefficients α and β can be adjusted according to the training focus.

[0086] The calculation of the coordination dimension deviation is based on two methods: pressure distribution similarity analysis and pressure center point trajectory deviation measurement. Pressure distribution similarity is implemented using a cosine similarity algorithm, which calculates the actual pressure distribution vector at a given time t. With the target distribution vector The vector angle is calculated, and the similarity value range is [0,1]. The closer the value is to 1, the more similar the distributions are. The consistency deviation is defined as:

[0087]

[0088] The deviation of the pressure center point trajectory is calculated using a dynamic time warping algorithm to determine the average Euclidean distance between the actual CoP trajectory and the target trajectory.

[0089]

[0090] The final consistency dimension deviation is achieved through

[0091]

[0092] The calculation yielded the result.

[0093] The response speed dimension deviation is calculated through neuromuscular reaction time analysis. The system records the time interval from the moment the action command is issued to the moment when the tongue muscle pressure first exceeds the preset response threshold (usually set to 1 Newton). and compared with the standard reaction time defined in the rehabilitation goal model. The response speed deviation is defined as a comparison.

[0094]

[0095] When the actual reaction time exceeds the target time, the deviation value is positive, indicating a slow neuromuscular response; when the actual reaction time is shorter than the target time, the deviation value is negative, indicating a normal or relatively quick response.

[0096] Through the above multi-dimensional deviation calculation, the system can obtain a deviation vector that comprehensively describes the patient's functional performance after each action cycle:

[0097]

[0098] This vector provides a quantitative basis for subsequent adaptive parameter adjustments.

[0099] Step S4: Adaptive adjustment mechanism for stimulation parameters based on bias (automatically adjusting the stimulation parameters for electrical stimulation in the next training cycle based on the bias), as follows:

[0100] Step S4 uses an intelligent judgment algorithm running on the data processing and control host to convert the deviation analysis results calculated in step S3 into optimized stimulus parameter configurations. For example... Figure 7 The diagram shown illustrates the logic of adjusting stimulus parameters based on deviation. Step S4 provides two optional control schemes: an expert system control scheme based on multi-dimensional rules and a continuous adjustment control scheme based on PID algorithm. The therapist can choose the most suitable scheme according to the patient's specific situation and training stage.

[0101] In the specific implementation of step S4, the expert system control scheme based on multi-dimensional rules simulates the clinical judgment logic of rehabilitation therapists, establishing an "IF-THEN" rule base containing 24 core rules. This rule base directly reflects the specific process of automatically adjusting stimulus parameters based on deviations. When the calculated force dimension deviation... When the pressure exceeds the first preset threshold α (usually set to 20% of the target pressure), the system automatically increases the electrical stimulation intensity for the next cycle. The adjustment amount is calculated using the following formula:

[0102]

[0103] Where k1 is the force compensation coefficient, typically 0.3-0.5 mA per Newton deviation. When the calculated coordination dimension deviation... When the pressure exceeds the second preset threshold β, the system automatically adjusts the electrode activation mode to enhance the electrode output corresponding to the pressure-weak area. Specifically, this involves increasing the output weight of the electrodes in the weak area to 1.2-1.5 times, while reducing the electrode weight in other areas to 0.8-0.9 times. When the calculated response speed dimension deviation... When the frequency exceeds the third preset threshold γ (usually set to 200 milliseconds), the system automatically increases the electrical stimulation frequency for the next cycle, with each adjustment ranging from 5 to 10 Hz.

[0104] Step S4 also provides a continuous adjustment control scheme based on a PID control algorithm, which calculates the technical characteristics of the stimulus intensity adjustment amount through the PID control algorithm. The mathematical model of the PID controller is as follows:

[0105]

[0106] Among them, the proportion term In response to the current deviation, the integral term Response to historical cumulative deviation, differential term The rate of change of response deviation is used to achieve smooth and stable control of tongue muscle strength. Proportional coefficient. The typical value range is 0.2-0.8 mA per Newton, and the integral coefficient is... Typical values ​​range from 0.05 to 0.2 mA per Newton-second, with differential coefficients... Typical values ​​range from 0.01 to 0.05 mA / s per Newton. The parameter adjustment process employs a limiting protection mechanism to ensure that the adjusted stimulus parameters remain within a safe and effective range.

[0107] Step S5: Implementation of iterative training and process recording (repeatedly executing steps S2 to S4 to form a closed-loop, adaptive rehabilitation training process). Details are as follows:

[0108] Step S5 is the stage where steps S2 to S4 are executed cyclically to form a closed-loop adaptive rehabilitation training process. This step constitutes the continuous operation mode of the closed-loop control method. Step S5 forms a continuous and adaptive rehabilitation training process by cyclically executing the core steps of step S2 (stimulus initiation and real-time data acquisition), step S3 (real-time deviation analysis and calculation), and step S4 (adaptive adjustment of stimulus parameters based on deviation). The system uses a preset training cycle as the basic time unit, and the duration of each cycle can be set to 3-10 seconds according to the characteristics of the rehabilitation target model set in step S1. Within each cycle, the system re-executes the neural stimulation of step S2 according to the stimulation parameters optimized in step S4, while continuously collecting the patient's tongue muscle response data. At the end of the cycle, deviation analysis is performed in step S3, and parameter configuration for the next cycle is generated. This cycle is repeated until the preset training duration is completed or the training goal is achieved.

[0109] Step S5 also includes rehabilitation log recording and visualization report generation functions. Data recording throughout the iteration process employs a multi-layered storage architecture. The system records and stores the raw tongue muscle data, multi-dimensional deviation values, and adjusted stimulation parameters for each training cycle, forming a detailed rehabilitation log. The raw data layer records the values ​​of 16 pressure sensors at each sampling moment, the number of the currently activated electrode and its output current value, timestamp, and other basic information. The data storage format uses an efficient binary structure, with the amount of raw data generated in a single training session being approximately 2-5 megabytes. The feature data layer records the key quantitative indicators extracted for each training cycle, including parameters such as peak pressure, pressure integral, pressure center point coordinates, and reaction time, as well as the corresponding target values ​​and deviation values. The parameter adjustment layer records the specific details of each parameter optimization, including the stimulation intensity, frequency, pulse width, and electrode activation state before and after adjustment, as well as the deviation conditions triggering the adjustment and the control scheme used.

[0110] After training, the system automatically generates a visualized rehabilitation report based on the rehabilitation log, including graphs showing changes in strength, coordination, and reaction speed. The system's data analysis engine comprehensively processes multi-level recorded data to generate a detailed visualized rehabilitation report. This report includes data displays in various chart formats, such as... Figure 8The rehabilitation report example shown in the human-computer interaction software interface diagram primarily includes a strength change trend curve, a coordination improvement radar chart, a reaction speed optimization trajectory chart, and a dynamic adjustment process chart of stimulation parameters during training. The strength change trend curve displays the evolution of both the target and actual pressure curves on the horizontal axis and the vertical axis, visually reflecting the patient's strength improvement through the reduction in the distance between the curves. The coordination improvement radar chart uses polygons to represent the patient's pressure control ability in different spatial directions; the closer the polygon is to a circle, the better the coordination. The reaction speed optimization trajectory chart records the trend of neuromuscular reaction time throughout the training process; a downward trend indicates improved nerve conduction efficiency.

[0111] In summary, the closed-loop control method described in this embodiment combines passive electrical stimulation therapy with the patient's active training exercises, achieving a technological leap from the traditional open-loop rehabilitation model to an intelligent closed-loop rehabilitation model. Traditional electrical stimulation rehabilitation equipment typically uses fixed stimulation parameter configurations, which cannot be dynamically adjusted according to the patient's real-time responses, resulting in problems such as poor stimulation effects or overstimulation. The closed-loop control method of this invention, by monitoring the patient's tongue muscle function in real time, can promptly identify problems during the training process and automatically optimize parameters, ensuring that each nerve stimulation is precisely matched to the patient's current functional state, significantly improving the pertinence and effectiveness of rehabilitation training.

[0112] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A tongue pressure plate system with acupoint nerve stimulation and real-time tongue muscle data acquisition functions, characterized in that, include: Tongue pressure plate main body; A nerve stimulation module, located on the tongue depressor body, is used to apply adjustable electrical stimulation to the user's tongue. The data acquisition module is installed on the tongue depressor body and is used to collect the actual tongue muscle data of the user when performing swallowing-related actions in real time. The control unit is electrically connected to the neurostimulation module and the data acquisition module; The control unit is configured as follows: Pre-store or receive a rehabilitation target model containing target tongue muscle data parameters; The actual tongue muscle data collected in real time by the data acquisition module is compared with the rehabilitation target model to determine the deviation between the two; and, based on the deviation, the stimulation parameters of the electrical stimulation applied by the nerve stimulation module are adaptively adjusted to guide the user's tongue muscle movement to approach the rehabilitation target model.

2. The tongue pressure plate system with acupoint nerve stimulation and real-time tongue muscle data acquisition function according to claim 1, characterized in that, The tongue pressure plate body is integrally molded from medical-grade silicone, and its tongue-facing side surface is divided into a central nerve stimulation area and a peripheral data acquisition area; the nerve stimulation module is an array of electrode contacts located in the central nerve stimulation area, and the data acquisition module is a ring-shaped pressure sensor array located in the peripheral data acquisition area and surrounding the electrode contact array.

3. The tongue pressure plate system with acupoint nerve stimulation and real-time tongue muscle data acquisition function according to claim 1, characterized in that, The control unit determines the deviation in the following ways: The force dimension deviation is calculated by comparing the difference between the actual peak pressure and the target value; and... The coordination dimension deviation is determined by calculating the cosine similarity between the actual pressure distribution map and the target map, or by calculating the average Euclidean distance between the actual pressure center point trajectory and the target trajectory.

4. The tongue pressure plate system with acupoint nerve stimulation and real-time tongue muscle data acquisition function according to claim 1, characterized in that, The adjustable stimulation parameters include stimulation intensity adjustable from 0.1mA to 20mA, stimulation frequency adjustable from 1Hz to 150Hz, and electrode activation mode that allows independent control of the on / off state of each electrode in the array.

5. The tongue pressure plate system with acupoint nerve stimulation and real-time tongue muscle data acquisition function according to claim 1, characterized in that, The control unit also includes human-computer interaction software, which has: a therapist-side interface for therapists to set the rehabilitation goal model and the safety threshold of the stimulation parameters, and a patient-side interface for providing real-time feedback and training guidance to the user.

6. A control method for swallowing rehabilitation, applied to the tongue depressor system as described in claim 1, characterized in that, Includes the following steps: S1. Set a rehabilitation target model that includes target tongue muscle data parameters; S2. Control the nerve stimulation module to apply electrical stimulation to the user's tongue, and at the same time, control the data acquisition module to collect the user's actual tongue muscle data in real time when attempting to complete the target action; S3. Analyze the real-time collected actual tongue muscle data and the rehabilitation target model to calculate the deviation between the two; S4. Based on the aforementioned deviation, automatically adjust the stimulation parameters for the electrical stimulation used in the next training cycle; and, S5. Repeat steps S2 to S4 to form a closed-loop, adaptive rehabilitation training process.

7. The control method for swallowing rehabilitation according to claim 6, characterized in that, In step S1, the rehabilitation target model is set as a curve function that includes the change of target pressure over time. The expected trajectory of the target pressure center point changing over time Time series data template.

8. The control method for swallowing rehabilitation according to claim 5 or 6, characterized in that, In step S4, the process of automatically adjusting the stimulation parameters based on the deviation includes: When the calculated force dimension deviation exceeds the first preset threshold, the electrical stimulation intensity of the next cycle is automatically increased. When the calculated coordination dimension deviation exceeds the second preset threshold, the electrode activation mode is automatically adjusted to enhance the electrode output corresponding to the pressure-weak area. When the calculated deviation in response speed dimension exceeds the third preset threshold, the electrical stimulation frequency of the next cycle is automatically increased.

9. The control method for swallowing rehabilitation according to claim 6, characterized in that, In step S4, the adjustment amount ΔI of the stimulation intensity is calculated by a PID control algorithm, wherein the proportional term (P) responds to the current deviation, the integral term (I) responds to the historical cumulative deviation, and the derivative term (D) responds to the rate of change of the deviation, so as to achieve smooth and stable control of tongue muscle strength.

10. The control method for swallowing rehabilitation according to claim 6, characterized in that, The method also includes the steps of: recording and storing a rehabilitation log containing raw tongue muscle data, multi-dimensional deviation values, and adjusted stimulation parameters for each training cycle; and automatically generating a visual rehabilitation report based on the rehabilitation log, including curves showing changes in strength, coordination, and reaction speed.

Images