Opening device-based swallowing muscle rehabilitation training system and method

Abstract

The application provides an open mouth piece-based swallowing muscle rehabilitation training system and method, and relates to the field of medical data processing.The method comprises the following steps: obtaining swallowing influence data of a patient; determining a plurality of key surface electrode detection positions of the patient according to the swallowing influence data of the patient; fixing the open mouth state of the patient through an open mouth piece, and setting a plurality of surface electrodes according to the plurality of key surface electrode detection positions of the patient; collecting swallowing detection data through the plurality of surface electrodes; determining swallowing disorder characteristics of the patient according to the swallowing detection data; and determining a swallowing muscle rehabilitation training scheme of the patient according to the swallowing disorder characteristics of the patient, wherein the swallowing muscle rehabilitation training scheme at least comprises a muscle electric stimulation scheme, and has the advantages of improving the effect of swallowing muscle rehabilitation.

Description

Technical Field

[0001] This invention relates to the field of medical data processing, and in particular to a swallowing muscle rehabilitation training system and method based on an oral mouth opener. Background Technology

[0002] Oropharyngeal dysphagia is a geriatric syndrome that threatens the physical and mental health of the elderly. Elderly individuals with oropharyngeal dysphagia may experience choking when drinking water, aspiration, or swallowing of food, posing a risk of serious consequences such as aspiration pneumonia and suffocation. Swallowing is a complex, coordinated movement involving many small muscles. The decline in muscle mass with age is inevitable, leading to decreased strength in the muscles related to swallowing and resulting in oropharyngeal dysphagia. Therefore, strengthening the swallowing muscles is crucial for the elderly. Rehabilitation training is an essential measure for dysphagia. Existing treatment options include: oral and tongue exercises (using cotton swabs and tongue depressors to strengthen voluntary movements and muscle strength of the lips and tongue), hot and cold stimulation to induce the swallowing reflex (using cotton swabs moistened with cold water to stimulate the soft palate and pharynx to induce swallowing), and swallowing method training (neck relaxation training, articulation training, breathing training, repetitive swallowing, alternating swallowing, swallowing on the healthy side, and nodding swallowing), etc.

[0003] Current pharyngeal muscle rehabilitation programs often adopt a "one-size-fits-all" approach, using the same training content and intensity for all patients. This standardized approach ignores individual differences among patients, such as age, gender, underlying diseases, and the severity of swallowing disorders. For some patients, the standardized program may be too simple and fail to achieve the desired rehabilitation effect. For others, the program may be too complex or too intense, leading to discomfort or injury during training.

[0004] Therefore, there is a need to provide swallowing muscle rehabilitation training systems and methods based on mouth opening devices to improve the effectiveness of swallowing muscle rehabilitation. Summary of the Invention

[0005] This invention provides a swallowing muscle rehabilitation training method based on an oral mouth opener, comprising: acquiring swallowing impact data of a patient; determining multiple key surface electrode detection positions of the patient based on the swallowing impact data; fixing the patient's mouth-opening state with an oral mouth opener, and setting multiple surface electrodes according to the multiple key surface electrode detection positions of the patient; collecting swallowing detection data through the multiple surface electrodes; determining the swallowing disorder characteristics of the patient based on the swallowing detection data; and determining a swallowing muscle rehabilitation training program for the patient based on the swallowing disorder characteristics, wherein the swallowing muscle rehabilitation training program includes at least a muscle electrical stimulation program.

[0006] Furthermore, acquiring swallowing impact data for the patient includes: identifying the influencing diseases and medications of the swallowing disorder; acquiring the patient's medical history; extracting key medical record data from the patient's medical history based on the influencing diseases and medications of the swallowing disorder; and acquiring the patient's swallowing performance data, wherein the swallowing impact data for the patient includes at least key medical record data and swallowing performance data.

[0007] Furthermore, based on the patient's swallowing impact data, multiple surface electrode detection locations are determined, including: acquiring key medical record data, swallowing performance data, and swallowing muscle damage information from multiple historical patients; identifying similar historical patients based on the key medical record data from multiple historical patients and the patient's key medical record data; identifying the first key swallowing muscle based on the swallowing muscle damage information from similar historical patients; determining the correlation information between swallowing performance and swallowing muscle damage based on the swallowing performance data and swallowing muscle damage information from multiple historical patients; identifying the second key swallowing muscle based on the patient's swallowing performance data and the correlation information between swallowing performance and swallowing muscle damage; deduplicating the first and second key swallowing muscles to determine the target swallowing muscle; acquiring the correlation information between the swallowing muscle and the surface electrode detection locations; and determining multiple surface electrode detection locations for the patient based on the target swallowing muscle and the correlation information between the swallowing muscle and the surface electrode detection locations.

[0008] Furthermore, determining the association information between swallowing muscles and surface electrode detection locations includes: identifying multiple surface electrode detection locations; fixing the mouth-opening state of multiple historical patients using an oral gag; setting multiple surface electrodes based on the multiple surface electrode detection locations; collecting swallowing detection data from multiple historical patients through the multiple surface electrodes; determining the historical patient group corresponding to each swallowing muscle based on the swallowing muscle damage information of multiple historical patients; and for each swallowing muscle, determining the associated surface electrode detection location based on the swallowing detection data of historical patients included in the historical patient group corresponding to the swallowing muscle.

[0009] Furthermore, based on the swallowing test data of historical patients included in the historical patient group corresponding to the swallowing muscles, the surface electrode detection locations associated with the swallowing muscles are determined, including: calculating the correlation coefficient between each surface electrode detection location and the degree of damage to the swallowing muscles based on the swallowing test data of historical patients included in the historical patient group corresponding to the swallowing muscles, and screening candidate surface electrode detection locations for the swallowing muscles from multiple surface electrode detection locations; calculating the correlation coefficient between any two candidate surface electrode detection locations for the swallowing muscles based on the swallowing test data of historical patients included in the historical patient group corresponding to the swallowing muscles; and screening surface electrode detection locations associated with the swallowing muscles from the candidate surface electrode detection locations for the swallowing muscles based on the correlation coefficient of any two candidate surface electrode detection locations screened.

[0010] Furthermore, based on the swallowing test data, the swallowing disorder characteristics of the patient are determined, including: for each target swallowing muscle, based on the detection location of the key surface electrodes associated with the target swallowing muscle, the local swallowing test data corresponding to the target swallowing muscle is determined from the swallowing test data; the local swallowing test data corresponding to the target swallowing muscle is decomposed through multivariate empirical mode decomposition; the functional characteristics of the target swallowing muscle are extracted from the decomposition results; and the degree of damage to the target swallowing muscle is determined based on the functional characteristics of the target swallowing muscle. The swallowing disorder characteristics of the patient include the degree of damage to each target swallowing muscle.

[0011] Furthermore, based on the patient's swallowing dysphagia characteristics, a rehabilitation training program for the patient's swallowing muscles is determined, including: obtaining the correlation information between swallowing muscles and rehabilitation training exercises; determining the patient's rehabilitation training exercises based on the patient's swallowing dysphagia characteristics and the correlation information between swallowing muscles and rehabilitation training programs; determining the exercise parameters of the patient's rehabilitation training exercises based on the patient's swallowing dysphagia characteristics; and determining the patient's muscle electrical stimulation program based on the patient's swallowing dysphagia characteristics.

[0012] Furthermore, based on the patient's swallowing dysphagia characteristics, the exercise parameters for the patient's rehabilitation training are determined, including: determining the probability distribution of damage to each swallowing muscle and the correlation coefficient between any two swallowing muscles based on swallowing muscle damage information from multiple historical patients; generating multiple sample patients based on the probability distribution of damage to each swallowing muscle and the correlation coefficient between any two swallowing muscles; determining the optimal swallowing muscle rehabilitation training program for each sample patient; identifying a first similar sample patient from the multiple sample patients based on the patient's swallowing dysphagia characteristics and rehabilitation training program; and determining the exercise parameters for the patient's rehabilitation training based on the optimal swallowing muscle rehabilitation training program for the first similar sample patient.

[0013] Furthermore, based on the patient's swallowing dysphagia characteristics, a muscle electrical stimulation protocol is determined, including: determining the association information between swallowing muscles and electrical stimulation locations; determining key electrical stimulation locations based on the patient's swallowing dysphagia characteristics and the association information between swallowing muscles and electrical stimulation locations; identifying a second similar sample patient from multiple sample patients based on the patient's swallowing dysphagia characteristics and key electrical stimulation locations; and determining the patient's electrical stimulation parameters based on the optimal swallowing muscle rehabilitation training protocol of the first similar sample patient, wherein the patient's muscle electrical stimulation protocol includes key electrical stimulation locations and electrical stimulation parameters.

[0014] This invention provides a swallowing muscle rehabilitation training system based on an oral mouth opener, comprising: a data acquisition module for acquiring swallowing impact data of a patient; a location determination module for determining multiple key surface electrode detection locations of the patient based on the swallowing impact data; a swallowing detection module for fixing the patient's mouth-opening state with an oral mouth opener, setting multiple surface electrodes based on the multiple key surface electrode detection locations of the patient, and collecting swallowing detection data through the multiple surface electrodes; a feature determination module for determining the swallowing disorder characteristics of the patient based on the swallowing detection data; and a training planning module for determining a swallowing muscle rehabilitation training program for the patient based on the swallowing disorder characteristics, wherein the swallowing muscle rehabilitation training program includes at least a muscle electrical stimulation program.

[0015] Compared with existing technologies, the swallowing muscle rehabilitation training system and method based on an oral mouth opener provided by this invention has at least the following beneficial effects:

[0016] By acquiring swallowing-related data such as swallowing time, force, and coordination, a comprehensive prediction of the patient's swallowing function can be made. This allows for the determination of key surface electrode detection locations and further precise acquisition of electrophysiological signals during swallowing, providing accurate data for subsequent analysis. Based on the collected swallowing data, the characteristics of the patient's swallowing disorder can be accurately analyzed. A personalized swallowing muscle rehabilitation training program, including muscle electrical stimulation, can be tailored to the patient's specific condition. Compared to traditional rehabilitation methods, precise assessment and personalized treatment can more effectively improve the patient's swallowing function. Patients can see rehabilitation effects in a shorter time, thus shortening the rehabilitation cycle and improving rehabilitation efficiency. Attached Figure Description

[0017] This specification will be further described by way of exemplary embodiments, which will be described in detail with reference to the accompanying drawings. These embodiments are not limiting; in these embodiments, the same reference numerals denote the same structures, wherein:

[0018] Figure 1 This is a flowchart illustrating a swallowing muscle rehabilitation training method based on an mouth opener, according to some embodiments of this specification.

[0019] Figure 2 This is a flowchart illustrating the process of determining multiple key surface electrode detection locations of a patient according to some embodiments of this specification;

[0020] Figure 3 This is a schematic flowchart illustrating the screening of surface electrode detection locations associated with swallowing muscles according to some embodiments of this specification;

[0021] Figure 4This is a schematic diagram illustrating the association between swallowing muscles and rehabilitation training exercises according to some embodiments of this specification;

[0022] Figure 5 This is a schematic diagram of a module of a swallowing muscle rehabilitation training system based on an mouth opener, according to some embodiments of this specification. Detailed Implementation

[0023] To more clearly illustrate the technical solutions of the embodiments in this specification, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are merely some examples or embodiments of this specification. For those skilled in the art, these drawings can be applied to other similar scenarios without creative effort. Unless obvious from the context or otherwise specified, the same reference numerals in the drawings represent the same structures or operations.

[0024] Figure 1 This is a flowchart illustrating a swallowing muscle rehabilitation training method based on an mouth opener, according to some embodiments of this specification. Figure 1 As shown, the swallowing muscle rehabilitation training method based on an oral mouth opener may include the following steps.

[0025] S110. Obtain data on the patient's swallowing effects.

[0026] In some embodiments, S110 specifically includes:

[0027] Determine the diseases and medications that affect swallowing disorders;

[0028] Obtain the patient's medical history;

[0029] Based on the diseases and medications that affect swallowing disorders, key medical record data are extracted from the patient's historical medical records.

[0030] Obtain patient swallowing performance data, including at least key medical record data and swallowing performance data.

[0031] Among these, diseases affecting swallowing refer to various conditions that can directly or indirectly impair swallowing function, leading to clinical symptoms such as difficulty swallowing, choking, and aspiration. These diseases may interfere with the normal physiological mechanisms of the swallowing process by damaging the nervous system, muscular system, structural organs, or metabolic function. For example, neurological diseases (such as cerebrovascular accidents, Parkinson's disease, and multiple sclerosis), muscle diseases (such as muscular dystrophy and polymyositis), and head and neck tumors and inflammations can all lead to impaired swallowing muscle function.

[0032] Drugs that affect swallowing disorders are those that interfere with the physiological process of swallowing through direct effects (such as inhibiting nerve conduction or relaxing muscles) or indirect effects (such as causing dry mouth or impaired consciousness). For example, antipsychotic drugs may trigger extrapyramidal reactions, leading to swallowing coordination disorders; chemotherapy drugs may cause oral mucositis, indirectly affecting the function of swallowing muscles.

[0033] Key medical record data extracted from a patient's medical history may include the severity of any swallowing disorders the patient has had or is currently experiencing, the dosage and duration of any medications the patient has taken or is currently taking that affect swallowing. Key medical record data may also include basic patient information such as age and gender.

[0034] Data on a patient's swallowing performance can be obtained through the following tests: Kubota water swallowing test, repeated saliva swallowing test, and video fluoroscopic swallowing imaging.

[0035] Kubota drinking water experiment:

[0036] Environment and position: The patient sits upright in a quiet environment free from disturbances, keeping the head in a neutral position.

[0037] Testing tools: Use a warm water cup with a fixed graduation (30ml) to avoid changes in liquid flow rate due to differences in cup shape.

[0038] Operating steps:

[0039] Instruct the patient to drink 30ml of warm water all at once, and not in divided doses.

[0040] Observe and record the swallowing time (from the moment the cup comes into contact with the water until swallowing is completed).

[0041] Record any instances of coughing or choking, categorized into levels 1-5:

[0042] Level 1: Swallowed smoothly within 5 seconds without choking.

[0043] Level 2: Swallow within 5 seconds without choking.

[0044] Level 3: Swallowed in one go, but with choking.

[0045] Level 4: Swallowed in two or more gulps, with choking or coughing.

[0046] Level 5: Frequent choking and coughing, unable to swallow completely.

[0047] Repeated saliva swallowing test:

[0048] Patient preparation: The patient should sit upright with their head and neck relaxed and maintain natural breathing.

[0049] Operating steps:

[0050] Instruct the patient to swallow, and the assessor places their fingers below the patient's Adam's apple to feel the muscle contractions caused by swallowing.

[0051] Record the number of swallows within 30 seconds (from the first swallow to the end of the 30th second). To avoid reducing the number of swallows due to the patient's tension or external interference, repeat 2-3 times and take the average.

[0052] Video-assisted fluoroscopy (FLFS) swallowing imaging:

[0053] Contrast agent selection: Use barium agents of different viscosities (such as thin liquid, paste, solid) to simulate daily diet.

[0054] Position and angle: The patient sits upright or lies on their side (depending on the condition), and the camera records the swallowing process in lateral and anterior-posterior positions.

[0055] Operating steps:

[0056] Instruct the patient to swallow barium solutions of different viscosities in sequence (5-10 ml each time).

[0057] Dynamic observation of the pharyngeal phase (residual epiglottis and pyriform fossa) and the esophageal phase (esophageal peristalsis and opening of the cricopharyngeal muscle).

[0058] Identify retention (barium residue >10%), aspiration (barium entering the airway), and leakage (barium entering the subglottic region).

[0059] Quantitative indicators:

[0060] Pharyngeal transit time: the time from when barium enters the pharynx to when it completely passes through the cricopharyngeal muscle (normal <1 second).

[0061] Residue classification: Grade 0 (no residue), Grade 1 (small residue, <10%), Grade 2 (significant residue, ≥10%).

[0062] Aspiration grading: The Penetration-Aspiration Scale (PAS) is used, with 1-8 levels to quantify the severity of aspiration.

[0063] Patient swallowing data may include: choking level, number of swallows in 30 seconds, pharyngeal transit time, residual volume level, and aspiration level.

[0064] S120. Based on the patient's swallowing impact data, determine the locations of multiple key surface electrode detection points for the patient.

[0065] Figure 2 This is a flowchart illustrating the process of determining multiple key surface electrode detection locations for a patient, as shown in some embodiments of this specification. Figure 2 As shown, in some embodiments, S120 specifically includes:

[0066] Acquire key medical record data, swallowing performance data, and swallowing muscle damage information from multiple historical patients;

[0067] Based on key medical record data from multiple historical patients and key medical record data from the patient, similar historical patients were identified;

[0068] Based on information on swallowing muscle damage in patients with similar histories, the first key swallowing muscle was identified;

[0069] Based on swallowing performance data and swallowing muscle damage information from multiple historical patients, the correlation between swallowing performance and swallowing muscle damage was determined.

[0070] Based on the patient's swallowing performance data and the correlation between swallowing performance and swallowing muscle damage, the second key swallowing muscle was identified;

[0071] The first and second key swallowing muscles were deduplicated to identify the target swallowing muscle.

[0072] Obtain the correlation information between the swallowing muscles and the detection location of the surface electrodes;

[0073] Based on the target swallowing muscle and the association information between the swallowing muscle and the surface electrode detection location, multiple surface electrode detection locations of the patient are determined.

[0074] The methods for obtaining key medical record data and swallowing performance data of historical patients are similar to those for obtaining key medical record data and swallowing performance data of patients, and will not be repeated here.

[0075] Swallowing muscles can include the tongue muscles, masticatory muscles, cheek muscles, and pharyngeal muscles.

[0076] Information on swallowing muscle damage can be determined manually (e.g., by doctors, experts, etc.) based on clinical grading scales, as well as electromyography, imaging (e.g., fiberoptic laryngoscopy swallowing function test, video swallowing angiography, etc.) and functional tests (Kubota water swallowing test, repeated saliva swallowing test, high-resolution manometry, etc.). The information on swallowing muscle damage can include the degree of damage to each swallowing muscle.

[0077] For example, a clinical grading scale may include:

[0078] Tongue muscle assessment:

[0079] Grade 1: Can fit snugly against the upper palate and the left and right gums.

[0080] Grade 2: Can press firmly against the upper palate but cannot press against the left and right gums.

[0081] Level 3: Can be lifted but cannot reach the palate.

[0082] Level 4: Cannot be lifted.

[0083] Assessment of masticatory and buccal muscles:

[0084] Level 1: Can fully turn the corners of the mouth to the left and right, puff out air and tap the cheeks without leaking air, and bite down with strong teeth.

[0085] Grade 2: When puffing out air and tapping the cheek, air leaks out; the upper and lower teeth bite together strongly on one side and weakly on the other.

[0086] Level 3: The air can't be puffed up tightly; there is a biting action, but the force is weak.

[0087] Level 4: Complete inability to puff out air, inability to bite.

[0088] Pharyngeal muscle assessment:

[0089] Grade 1: Both soft palates are raised forcefully.

[0090] Grade 2: One side of the soft palate is raised forcefully.

[0091] Grade 3: Inability to elevate the soft palate.

[0092] Grade 4: The soft palate cannot be raised.

[0093] Key medical records from multiple unstructured historical patients and key medical records from individual patients can be transformed into computable feature vectors. For example, age is normalized to the [0,1] interval, and gender is represented using binary encoding. Using ICD-10 encoding or a custom disease classification system, the names of diseases affecting dysphagia are converted into numerical vectors, disease severity is converted into numerical values ​​according to clinical standards (such as NIHSS, mRS, HY classification), the names of medications affecting dysphagia are converted into numerical vectors, daily medication dosage is normalized according to patient weight, and treatment duration is normalized to [0,1] as a proportion of the total disease course. Basic information, disease information, and medication information are combined into a feature vector. Using similarity calculation methods (e.g., Euclidean distance, cosine similarity, etc.), the similarity between the feature vectors of historical patients and the feature vectors of individual patients is calculated, and historical patients whose feature vector similarity is greater than a feature vector similarity threshold are considered similar historical patients.

[0094] Based on the information on swallowing muscle damage in patients with similar histories, the damaged swallowing muscles in these patients are identified. If there are multiple patients with similar histories, the damaged swallowing muscles in these patients are deduplicated, and the damaged swallowing muscles identified after deduplication are designated as the first critical swallowing muscles.

[0095] Chi-square test or rank correlation analysis (such as Spearman correlation coefficient) was performed on each swallowing performance and the degree of damage to each muscle to determine the association between swallowing performance and swallowing muscle damage. For example, if the chi-square test p-value between the choking level and the degree of damage to the pharyngeal constrictor muscle is <0.05 and the correlation coefficient is >0.6, it indicates that the two are significantly correlated, and there is an association between choking and the pharyngeal constrictor muscle.

[0096] In some embodiments, determining the association information between the swallowing muscle and the detection location of the surface electrode includes:

[0097] Determine the detection locations of multiple surface electrodes;

[0098] The mouth-opening state of multiple historical patients is fixed by an mouth opener. Multiple surface electrodes are set according to the detection positions of multiple surface electrodes, and swallowing detection data of multiple historical patients are collected through multiple surface electrodes.

[0099] Based on the swallowing muscle damage information of multiple historical patients, the historical patient group corresponding to each swallowing muscle was determined;

[0100] For each swallowing muscle, the location of the surface electrode associated with the swallowing muscle is determined based on the swallowing test data of the historical patients in the corresponding historical patient group.

[0101] Specifically, multiple surface electrode detection sites need to cover all swallowing muscles. These sites can be determined manually (e.g., by doctors, specialists, etc.).

[0102] Surface electrodes are non-invasive bioelectrical signal acquisition devices that detect muscle electrical activity by being attached to the skin surface. Their core principle is based on measuring surface potential differences, allowing the acquisition of physiological signals without penetrating the skin.

[0103] The mouth opener can adjust its opening width according to the patient's mouth opening degree, avoiding excessive stretching that could lead to muscle fatigue. The mouth opener is placed between the patient's upper and lower molars and secured to the head with an elastic band to ensure a stable mouth opening. After stabilizing the patient's mouth opening, the patient can be instructed to perform empty swallowing, and swallowing data can be collected during this process. This swallowing data can include the continuous voltage signal output by the surface electrodes at each detection position during the patient's empty swallowing.

[0104] For each swallowing muscle, based on the swallowing muscle damage information of historical patients, it is determined whether the damaged swallowing muscle of the historical patient includes that swallowing muscle. If it does, the historical patient is classified into the historical patient group corresponding to that swallowing muscle. It is understood that the historical patients included in the historical patient groups corresponding to any two swallowing muscles may have at least partial overlap.

[0105] In some embodiments, determining the surface electrode detection location associated with the swallowing muscle based on swallowing test data of historical patients included in the historical patient group corresponding to the swallowing muscle includes:

[0106] Based on the swallowing test data of historical patients in the historical patient group corresponding to the swallowing muscles, the correlation coefficient between each surface electrode detection location and the degree of damage to the swallowing muscles is calculated, and candidate surface electrode detection locations for the swallowing muscles are screened from multiple surface electrode detection locations.

[0107] Based on the swallowing test data of historical patients in the historical patient group corresponding to the swallowing muscle, calculate the correlation coefficient between any two candidate surface electrode detection positions of the swallowing muscle.

[0108] Based on the correlation coefficient between any two candidate surface electrode detection locations of the swallowing muscle, surface electrode detection locations associated with the swallowing muscle are selected from the candidate surface electrode detection locations of the swallowing muscle.

[0109] Specifically, for each swallowing muscle and each surface electrode detection location, the degree of damage to that swallowing muscle in the historical patients included in the historical patient group can be converted into a numerical value. The average electromyography (EMG) value is calculated based on the continuous voltage signal output by the surface electrode at the corresponding detection location during the historical patient's empty swallowing process. The converted numerical value of the degree of damage to the swallowing muscle in each historical patient in the corresponding historical patient group and the average EMG value are then substituted into the formula for calculating the nonlinear correlation coefficient (e.g., Spearman's rank correlation coefficient, mutual information, etc.) to obtain the nonlinear correlation coefficient between the surface electrode detection location and the swallowing muscle. Surface electrode detection locations with a nonlinear correlation coefficient greater than a threshold (e.g., 0.7) are selected as candidate surface electrode detection locations for swallowing muscles.

[0110] For any two candidate surface electrode detection locations of the swallowing muscle, the average electromyographic values ​​of the two candidate surface electrode detection locations of each historical patient in the corresponding historical patient group of the swallowing muscle can be substituted into the nonlinear correlation coefficient (e.g., Spearman rank correlation coefficient, mutual information, etc.) calculation formula to calculate the correlation coefficient between the two candidate surface electrode detection locations.

[0111] Figure 3 This is a schematic flowchart illustrating the screening of surface electrode detection locations associated with swallowing muscles according to some embodiments of this specification, such as... Figure 3 As shown, the following procedure can be used to filter the surface electrode detection locations associated with the swallowing muscle from the candidate surface electrode detection locations of the swallowing muscle based on the correlation coefficient between any two candidate surface electrode detection locations:

[0112] S121. Calculate the variance of the correlation coefficient based on the correlation coefficient between any two candidate surface electrode detection positions of the swallowing muscle.

[0113] S122. Determine whether the variance of the correlation coefficient is less than the threshold of the variance of the correlation coefficient. If yes, complete the screening and use the remaining candidate surface electrode detection positions as the surface electrode detection positions associated with the swallowing muscle. If no, proceed to S123. The threshold of the variance of the correlation coefficient can be selected manually based on experience or obtained through simulation experiments.

[0114] S123. Determine whether the number of remaining candidate surface electrode detection positions is greater than the number threshold (e.g., 5). If not, use the remaining candidate surface electrode detection positions as the surface electrode detection positions associated with the swallowing muscle. If yes, execute S124.

[0115] S124. For each remaining candidate surface electrode detection position, calculate the average correlation coefficient between the candidate surface electrode detection position and other candidate surface electrode detection positions, and calculate the average correlation coefficient of the candidate surface electrode detection positions.

[0116] SS125. The candidate surface electrode detection position with the smallest mean correlation coefficient is selected as the candidate surface electrode detection position to be eliminated in the current iteration, and the remaining candidate surface electrode detection positions are updated.

[0117] S126. Calculate the variance of the correlation coefficient based on the correlation coefficient of any two currently remaining candidate surface electrode detection positions, and then execute S122.

[0118] Based on the association information between swallowing muscles and surface electrode detection locations, the associated surface electrode detection locations of the target swallowing muscles can be found. All associated surface electrode detection locations of the target swallowing muscles are deduplicated, and the remaining surface electrode detection locations after deduplication are used as multiple surface electrode detection locations for the patient.

[0119] Understandably, similarity analysis of key medical record data is used to filter out highly matched cases from historical databases. This matching mechanism avoids the limitations of relying on a single data dimension and ensures the representativeness of the reference sample. Damaged muscle information from similar historical patients is directly extracted to quickly identify the first key muscle. Combined with the current patient's swallowing performance data, individualized damage characteristics are mined to determine the second key muscle. Data is collected only at the electrode locations corresponding to the target muscles, avoiding the inefficiency of full-coverage detection.

[0120] S130. Fix the patient's open mouth state with an opening device, and set multiple surface electrodes according to the detection positions of multiple key surface electrodes of the patient.

[0121] S140. Swallowing detection data is collected through multiple surface electrodes.

[0122] The swallowing test data can include the continuous voltage signal output by the surface electrode at each surface electrode detection position during the patient's empty swallowing process.

[0123] S150. Based on the swallowing test data, determine the characteristics of the patient's swallowing disorder.

[0124] In some embodiments, S150 specifically includes:

[0125] For each target swallowing muscle, based on the detection location of the key surface electrodes associated with the target swallowing muscle, the corresponding local swallowing detection data is determined from the swallowing detection data. Through multivariate empirical mode decomposition, the local swallowing detection data corresponding to the target swallowing muscle is decomposed, and the functional characteristics of the target swallowing muscle are extracted from the decomposition results. Based on the functional characteristics of the target swallowing muscle, the degree of damage to the target swallowing muscle is determined. Among them, the swallowing disorder characteristics of the patient include the degree of damage to each target swallowing muscle.

[0126] The local swallowing detection data corresponding to the target swallowing muscle can include continuous voltage signals output by multiple key surface electrode detection locations associated with the target swallowing muscle during the patient's empty swallowing process.

[0127] Multivariate empirical mode decomposition is used to decompose the local swallowing detection data corresponding to the target swallowing muscle, including:

[0128] Objective: To decompose multi-channel local swallowing detection data, extract intrinsic mode functions related to muscle function, and eliminate noise and redundant information.

[0129] step:

[0130] Input: Continuous voltage signals output by multiple key surface electrodes associated with the target swallowing muscles during the patient's empty swallowing process.

[0131] Projection and extremum detection: Calculate the projection direction of the signal in multidimensional space and detect local extrema.

[0132] Envelope fitting: Fit the upper and lower envelopes using interpolation and calculate the mean curve.

[0133] Intrinsic mode function extraction: Iteratively extract components that satisfy the intrinsic mode function conditions (such as zero mean, the number of extreme points and the number of zero-crossing points differing by no more than 1).

[0134] Termination condition: Decomposition stops when the remaining signal is a monotonic function or a constant.

[0135] Output: Intrinsic mode function and residual signal for each key surface electrode detection location.

[0136] The functional characteristics of the target swallowing muscles can include the time-domain characteristics (e.g., root mean square, integrated electromyography, mean absolute value, etc.) and frequency-domain characteristics (e.g., median frequency, mean power frequency, etc.) of the intrinsic mode functions at each key surface electrode detection location. The root mean square reflects the intensity of muscle activity. The integrated electromyography reflects the total energy of muscle contraction. The mean absolute value reflects the average level of signal amplitude. The median frequency reflects the proportion of different muscle fiber types. The mean power frequency reflects the shift to lower frequencies during muscle fatigue.

[0137] The degree of damage to the target swallowing muscles can be determined using a damage assessment model based on the functional characteristics of the target swallowing muscles. This damage assessment model may include:

[0138] 1. Feature Input Layer

[0139] Input: Time-domain and frequency-domain characteristics of the intrinsic mode functions of each key surface electrode detection location of the target swallowing muscle.

[0140] 2. Feature Processing Layer

[0141] Normalization: Scaling features to the [0,1] interval to eliminate the influence of dimensions.

[0142] Dimensionality reduction: Extract key features through principal component analysis or linear discriminant analysis to reduce redundant information.

[0143] 3. Determine the core layer

[0144] Classification model: Features are mapped to damage levels (e.g., 0 = normal, 1 = mild damage, 2 = moderate damage, 3 = severe damage) through support vector machines.

[0145] 4. Output layer

[0146] Damage level: Classification result (e.g., pharyngeal constrictor muscle: grade 2).

[0147] The training process for the damage assessment model may include:

[0148] 1. Data partitioning

[0149] Training set: 70% of the samples, used for model parameter optimization.

[0150] Validation set: 15% of the samples, used for hyperparameter tuning (such as kernel function parameters of SVM).

[0151] Test set: 15% of the samples, used for final performance evaluation.

[0152] 2. Model Training

[0153] Optimize the kernel function (such as the RBF kernel) and the penalty coefficient C, and find the optimal parameters through grid search or Bayesian optimization.

[0154] 3. Performance Evaluation

[0155] The damaged model is trained using metrics such as accuracy, recall, F1 score, and ROC curve.

[0156] Understandably, data from a single key surface electrode detection location may be incomplete due to muscle anatomical variations, electrode attachment deviations, or signal interference. Multi-channel signals can cover different regions of the target muscle, capturing a more comprehensive range of activity patterns. Cross-validation of multi-channel signals can identify and eliminate anomalous signals (such as artifacts or noise), improving the robustness of feature extraction. The temporal or frequency domain features of a single key surface electrode detection location may be limited by signal quality or muscle activity heterogeneity. Joint analysis of multi-channel signals can extract richer feature combinations, enabling more accurate quantification of muscle activity intensity and fatigue levels.

[0157] S160. Determine the patient's swallowing muscle rehabilitation training plan based on the characteristics of the patient's swallowing disorder.

[0158] Among them, the swallowing muscle rehabilitation training program should include at least a muscle electrical stimulation program.

[0159] In some embodiments, S160 specifically includes:

[0160] Obtain the correlation information between swallowing muscles and rehabilitation training exercises. This correlation information can include training exercises associated with each swallowing muscle that aid in its rehabilitation, for example... Figure 4 This is a schematic diagram illustrating the association between swallowing muscles and rehabilitation training exercises, based on some embodiments of this specification, such as... Figure 4 As shown, supraglottic swallowing, which involves holding one's breath and closing the glottis before swallowing, followed by rapid swallowing and coughing to clear any remaining food, can enhance the closing ability of the pharyngeal constrictor muscles. Therefore, there is a correlation between supraglottic swallowing and the pharyngeal constrictor muscles. Another example is the Shaker exercise, where the patient lies flat and only looks up at their toes for one minute, repeating this multiple times to strengthen the hypohyoid muscles. Information on the correlation between swallowing muscles and rehabilitation exercises can be obtained manually or through big data analysis.

[0161] Based on the patient's swallowing disorder characteristics and the correlation information between swallowing muscles and rehabilitation training programs, the patient's rehabilitation training exercises are determined;

[0162] Based on the characteristics of the patient's swallowing disorder, determine the exercise parameters for the patient's rehabilitation training exercises;

[0163] Based on the characteristics of the patient's swallowing disorder, determine the patient's muscle electrical stimulation protocol.

[0164] This process involves identifying the target swallowing muscle damaged by the patient based on the characteristics of the patient's swallowing disorder, finding rehabilitation exercises associated with the damaged target swallowing muscle based on the correlation information between the swallowing muscle and rehabilitation training programs, deduplicating the rehabilitation exercises associated with each damaged target swallowing muscle, and using the remaining rehabilitation exercises after deduplication as the patient's rehabilitation training exercises.

[0165] In some embodiments, the exercise parameters for the patient's rehabilitation training exercises are determined based on the patient's swallowing dysphagia characteristics, including:

[0166] Based on the swallowing muscle damage information of multiple historical patients, the probability distribution of damage to each swallowing muscle and the correlation coefficient of damage between any two swallowing muscles were determined.

[0167] Based on the probability distribution of damage to each swallowing muscle and the correlation coefficient between any two swallowing muscles, multiple sample patients are generated.

[0168] Determine the optimal swallowing muscle rehabilitation training program for each patient sample;

[0169] Based on the patient’s swallowing disorder characteristics and rehabilitation training program, the first similar sample patient was identified from multiple sample patients;

[0170] Based on the optimal swallowing muscle rehabilitation training program for patients in the first similar sample, the motion parameters of the patients' rehabilitation training exercises were determined.

[0171] This allows for the organization and analysis of swallowing muscle damage information from multiple historical patients, and the statistical analysis of the damage probability of each swallowing muscle, i.e., the frequency of damage to that swallowing muscle in all historical patients.

[0172] For any two swallowing muscles, the degree of damage to these two swallowing muscles in each historical patient can be converted into numerical values, which can then be substituted into the formula for calculating nonlinear correlation coefficients (e.g., Spearman rank correlation coefficient, mutual information, etc.) to obtain the correlation coefficient of damage to the two swallowing muscles.

[0173] Multiple sample patients can be generated based on the probability distribution of damage to each swallowing muscle and the correlation coefficient between any two swallowing muscles damaged, using the following steps:

[0174] S161. Initialization: Create an empty swallowing muscle damage record table for each sample patient.

[0175] S162. Muscle-by-muscle simulation: For each swallowing muscle, random sampling is performed based on its damage probability to determine whether the muscle is damaged and the degree of damage in the sample patients. Specifically, for each swallowing muscle, the mean of the damage correlation coefficient between that muscle and other swallowing muscles can be calculated to obtain the mean damage correlation coefficient for that muscle. The swallowing muscles are then sorted according to the mean damage correlation coefficient from smallest to largest, and muscle-by-muscle simulation is performed based on the sorting results.

[0176] S163. Considering Correlation: During the simulation, for muscles whose damage level has been determined, the probability or degree of damage to other related muscles is adjusted based on their correlation coefficients with other muscles to reflect the correlation between them. For example, if the correlation coefficient between two muscles is high, then the probability that the other muscle will also be damaged when one muscle is damaged will increase.

[0177] S164. Repeat the process: Repeat S162 and S163 until the extent of damage to all swallowing muscles has been determined.

[0178] S165. Generate a sample: Save the swallowing muscle damage record of each sample patient as a generated sample patient.

[0179] For example, suppose there are three swallowing muscles: A, B, and C, with probabilities of damage of 0.3, 0.4, and 0.2, respectively. The correlation coefficient between A and B is 0.6, between A and C is 0.3, and between B and C is 0.4. Now, we need to generate 100 sample patients.

[0180] initialization:

[0181] Create 100 empty swallowing muscle damage record tables.

[0182] Muscle-by-muscle simulation:

[0183] For muscle A, in each sample patient, whether it is damaged is determined with a probability of 0.3. That is, a random number between 0 and 1 is generated. If the random number is less than or equal to 0.3, muscle A of the sample patient is considered to be damaged. If the random number is greater than 0.3, muscle A of the sample patient is considered not damaged.

[0184] For muscle B, its probability of injury is adjusted after considering its correlation with muscle A. For example, if muscle A is injured in a sample of patients, the probability of muscle B being injured might increase to 0.5. Then, the adjusted probability is used to determine whether muscle B is injured.

[0185] Similarly, for muscle C, after considering its correlation with muscles A and B, its probability of damage is adjusted, and a decision is made as to whether it is damaged.

[0186] Generate samples:

[0187] The swallowing muscle damage record of each sample patient was saved, resulting in 100 generated sample patients.

[0188] Understandably, the probability of damage to different muscles during actual swallowing is not independent. For example, the cricopharyngeal muscle and the pharyngeal constrictor muscle need to contract collaboratively during the initiation of swallowing; if the cricopharyngeal muscle is damaged, the probability of damage to the pharyngeal constrictor muscle may increase significantly. By dynamically adjusting the probability through correlation coefficients, this physiological correlation is simulated, making the generated sample patients closer to the actual situation. If samples were generated solely based on independent probabilities, combinations that do not conform to clinical reality might be produced (e.g., highly correlated muscles are not damaged).

[0189] The optimal swallowing muscle rehabilitation training program for each patient sample can be determined manually (e.g., by doctors, experts, etc.) or through big data analysis.

[0190] For each sample patient, the similarity between their swallowing disorder characteristics and those of the current patient is calculated. This can be done using distance metrics (such as Euclidean distance or Manhattan distance) or similarity coefficients (such as cosine similarity or Pearson correlation coefficient). The overlap between the sample patient's rehabilitation training programs and the current patient's rehabilitation training programs is also calculated. If a patient's rehabilitation training programs include four items, and three of these items are also present in the sample patient's rehabilitation training programs, the overlap is 0.75. A weighted sum of the similarity between the sample patient's and the current patient's swallowing disorder characteristics, as well as the overlap between the sample patient's and the current patient's rehabilitation training programs, is used to obtain the comprehensive similarity between the sample patient and the current patient. Sample patients with a comprehensive similarity greater than a threshold (e.g., 0.7) are designated as the first similar sample patients. For the patient's rehabilitation exercises, the exercise parameters of the first similar sample patient in that exercise can be used to determine the patient's exercise parameters for that exercise. For example, the mean of the exercise parameters of the first similar sample patient in that exercise can be used as the patient's exercise parameters for that exercise.

[0191] Understandably, the similarity of swallowing disorder characteristics (such as Euclidean distance and cosine similarity) is used to assess the degree of physiological similarity between patients, while the overlap of rehabilitation training programs (e.g., 0.75) is used to assess the suitability of training programs. For example, if two patients are highly similar in swallowing muscle activation patterns and training programs, their motor parameters (such as repetition count and resistance intensity) may be valuable. By referencing parameters from similar samples, ineffective training or patient discomfort caused by inappropriate parameters can be reduced, thus minimizing the waste of rehabilitation resources. For instance, it avoids patients dropping out of training due to excessively high parameters or poor therapeutic effects due to excessively low parameters.

[0192] In some embodiments, a muscle electrical stimulation protocol is determined based on the patient's swallowing dysphagia characteristics, including:

[0193] Determine the association between swallowing muscles and the location of electrical stimulation;

[0194] Based on the patient's swallowing dysphagia characteristics and the correlation information between swallowing muscles and electrical stimulation locations, the key electrical stimulation locations for the patient are determined.

[0195] Based on the patient’s swallowing dysphagia characteristics and key electrical stimulation locations, a second similar sample patient was identified from multiple sample patients;

[0196] Based on the optimal swallowing muscle rehabilitation training program for patients in the first similar sample, the electrical stimulation parameters of the patients were determined. The patient's muscle electrical stimulation program included key electrical stimulation locations and electrical stimulation parameters.

[0197] The association information between swallowing muscles and electrical stimulation sites can include the electrical stimulation site associated with each swallowing muscle. These sites are those that, when treated with electrical stimulation, have a positive impact on the recovery of the swallowing muscle. This association information can be determined manually (e.g., by doctors, experts, etc.), through big data analysis, or through simulation experiments.

[0198] Specifically, based on the characteristics of the patient's swallowing disorder, the target swallowing muscle that is damaged in the patient can be identified. Based on the association information between the swallowing muscle and the electrical stimulation location, the electrical stimulation location associated with the patient's damaged target swallowing muscle can be found. The electrical stimulation locations associated with each damaged target swallowing muscle in the patient can be deduplicated, and the remaining electrical stimulation locations after deduplication can be used as the patient's key electrical stimulation locations.

[0199] The method for identifying the second similar sample patient is similar to the method for identifying the first similar sample patient, and will not be repeated here.

[0200] The patient's electrical stimulation parameters can be determined based on the electrical stimulation parameters of the first similar sample of patients. For example, the patient's electrical stimulation parameters can be obtained by averaging the electrical stimulation parameters of the first similar sample of patients.

[0201] Understandably, personalized muscle electrical stimulation (EMS) plans are developed for patients by analyzing the association between swallowing muscles and EMS locations, screening key EMS locations, matching with second-similar patient samples, and optimizing EMS parameters. By pre-determining the association between swallowing muscles and EMS locations (e.g., the cricopharyngeal muscle corresponding to a specific electrode location in the neck), it is ensured that EMS acts directly on the target muscle, avoiding ineffective stimulation. For example, if the cricopharyngeal muscle is damaged, directly stimulating its associated EMS location can more effectively promote its recovery. Duplicate EMS locations associated with damaged target swallowing muscles are eliminated to avoid multiple stimulations of the same location, reducing patient discomfort. For example, if two damaged muscles are associated with the same EMS location, after deduplication, only one stimulation is required, improving treatment efficiency. Patient parameters are determined using the optimal swallowing muscle rehabilitation training plan (including EMS parameters) from first-similar patient samples, avoiding directly trying high-risk parameters. For example, if similar patient samples do not experience adverse reactions at a certain EMS intensity, the current patient can refer to this intensity to reduce muscle fatigue or damage caused by excessive stimulation. Calculating the mean value of electrical stimulation parameters for patients in the first similar sample can smooth out individual differences and avoid adverse consequences caused by abnormal parameters in a particular sample (such as being too high or too low).

[0202] Figure 5 This is a schematic diagram of a module of a swallowing muscle rehabilitation training system based on an mouth opener, as shown in some embodiments of this specification. Figure 5 As shown, the swallowing muscle rehabilitation training system based on an oral mouth opener may include a data acquisition module, a location determination module, a swallowing detection module, a feature determination module, and a training planning module.

[0203] The data acquisition module is used to acquire data on the patient's swallowing effects;

[0204] The location determination module is used to determine the detection locations of multiple key surface electrodes for the patient based on the patient's swallowing impact data;

[0205] The swallowing detection module is used to fix the patient's mouth opening state through an mouth opener. Multiple surface electrodes are set according to the patient's multiple key surface electrode detection positions, and swallowing detection data is collected through multiple surface electrodes.

[0206] The feature determination module is used to determine the characteristics of a patient's swallowing disorder based on swallowing test data;

[0207] The training planning module is used to determine the patient's swallowing muscle rehabilitation training program based on the characteristics of the patient's swallowing disorder. The swallowing muscle rehabilitation training program includes at least a muscle electrical stimulation program.

[0208] The swallowing muscle rehabilitation training system based on an mouth opener can be used to perform the swallowing muscle rehabilitation training methods based on the mouth opener described above, which will not be repeated here.

[0209] Finally, it should be understood that the embodiments described in this specification are merely illustrative of the principles of the embodiments described herein. Other variations may also fall within the scope of this specification. Therefore, alternative configurations of the embodiments described herein are intended to be consistent with the teachings of this specification, rather than as examples or limitations. Accordingly, the embodiments described herein are not limited to those explicitly introduced and described herein.

Claims

1. A swallowing muscle rehabilitation training system based on an oral mouth opener, characterized in that, include: The data acquisition module is used to acquire data on the patient's swallowing effects; The location determination module is used to determine the detection locations of multiple key surface electrodes for the patient based on the patient's swallowing impact data; The swallowing detection module is used to fix the patient's mouth opening state through an oral gag. Multiple surface electrodes are set according to the patient's multiple key surface electrode detection positions, and swallowing detection data is collected through multiple surface electrodes. The feature determination module is used to determine the characteristics of a patient's swallowing disorder based on swallowing test data; The training planning module is used to determine the patient's swallowing muscle rehabilitation training program based on the characteristics of the patient's swallowing disorder, wherein the swallowing muscle rehabilitation training program includes at least a muscle electrical stimulation program. The data acquisition module acquires data on the patient's swallowing impact, including: Determine the diseases and medications that affect swallowing disorders; Obtain the patient's medical history; Based on the diseases and medications that affect swallowing disorders, key medical record data are extracted from the patient's historical medical records. Acquire patient swallowing performance data, wherein the patient swallowing impact data includes at least key medical record data and swallowing performance data; The swallowing detection module determines multiple surface electrode detection locations for the patient based on the patient's swallowing impact data, including: Acquire key medical record data, swallowing performance data, and swallowing muscle damage information from multiple historical patients; Based on key medical record data from multiple historical patients and key medical record data from the patient, similar historical patients were identified; Based on information on swallowing muscle damage in patients with similar histories, the first key swallowing muscle was identified; Based on swallowing performance data and swallowing muscle damage information from multiple historical patients, the correlation between swallowing performance and swallowing muscle damage was determined. Based on the patient's swallowing performance data and the correlation between swallowing performance and swallowing muscle damage, the second key swallowing muscle was identified; The first and second key swallowing muscles were deduplicated to identify the target swallowing muscle. Obtain the correlation information between the swallowing muscles and the detection location of the surface electrodes; Based on the target swallowing muscle and the association information between the swallowing muscle and the surface electrode detection location, multiple surface electrode detection locations of the patient are determined.

2. The swallowing muscle rehabilitation training system based on an oral mouth opener according to claim 1, characterized in that, The swallowing detection module determines the association information between the swallowing muscles and the detection location of the surface electrodes, including: Determine the detection locations of multiple surface electrodes; The mouth-opening state of multiple historical patients is fixed by an mouth opener. Multiple surface electrodes are set according to the detection positions of multiple surface electrodes, and swallowing detection data of multiple historical patients are collected through multiple surface electrodes. Based on the swallowing muscle damage information of multiple historical patients, the historical patient group corresponding to each swallowing muscle was determined; For each swallowing muscle, the location of the surface electrode associated with the swallowing muscle is determined based on the swallowing test data of the historical patients in the corresponding historical patient group.

3. The swallowing muscle rehabilitation training system based on an oral mouth opener according to claim 2, characterized in that, The swallowing detection module determines the detection location of the surface electrode associated with the swallowing muscle based on the swallowing detection data of historical patients in the historical patient group corresponding to the swallowing muscle, including: Based on the swallowing test data of historical patients in the historical patient group corresponding to the swallowing muscles, the correlation coefficient between each surface electrode detection location and the degree of damage to the swallowing muscles is calculated, and candidate surface electrode detection locations for the swallowing muscles are screened from multiple surface electrode detection locations. Based on the swallowing test data of historical patients in the historical patient group corresponding to the swallowing muscle, calculate the correlation coefficient between any two candidate surface electrode detection positions of the swallowing muscle. Based on the correlation coefficient between any two candidate surface electrode detection locations of the swallowing muscle, surface electrode detection locations associated with the swallowing muscle are selected from the candidate surface electrode detection locations of the swallowing muscle.

4. The swallowing muscle rehabilitation training system based on an oral mouth opener according to claim 1 or 2, characterized in that, The feature determination module determines the patient's swallowing disorder characteristics based on swallowing test data, including: For each target swallowing muscle, based on the detection location of the key surface electrodes associated with the target swallowing muscle, the corresponding local swallowing detection data is determined from the swallowing detection data. Through multivariate empirical mode decomposition, the local swallowing detection data corresponding to the target swallowing muscle is decomposed, and the functional characteristics of the target swallowing muscle are extracted from the decomposition results. Based on the functional characteristics of the target swallowing muscle, the degree of damage to the target swallowing muscle is determined. Among them, the swallowing disorder characteristics of the patient include the degree of damage to each target swallowing muscle.

5. The swallowing muscle rehabilitation training system based on an oral mouth opener according to claim 1 or 2, characterized in that, The training planning module determines a swallowing muscle rehabilitation training plan for the patient based on the characteristics of the patient's swallowing disorder, including: Obtain information on the correlation between swallowing muscles and rehabilitation exercises; Based on the patient's swallowing disorder characteristics and the correlation information between swallowing muscles and rehabilitation training programs, the patient's rehabilitation training exercises are determined; Based on the characteristics of the patient's swallowing disorder, determine the exercise parameters for the patient's rehabilitation training exercises; Based on the characteristics of the patient's swallowing disorder, determine the patient's muscle electrical stimulation protocol.

6. The swallowing muscle rehabilitation training system based on an oral mouth opener according to claim 5, characterized in that, The training planning module determines the exercise parameters for the patient's rehabilitation training exercises based on the patient's swallowing dysphagia characteristics, including: Based on the swallowing muscle damage information of multiple historical patients, the probability distribution of damage to each swallowing muscle and the correlation coefficient of damage between any two swallowing muscles were determined. Based on the probability distribution of damage to each swallowing muscle and the correlation coefficient between any two swallowing muscles, multiple sample patients are generated. Determine the optimal swallowing muscle rehabilitation training program for each patient sample; Based on the patient’s swallowing disorder characteristics and rehabilitation training program, the first similar sample patient was identified from multiple sample patients; Based on the optimal swallowing muscle rehabilitation training program for patients in the first similar sample, the motion parameters of the patients' rehabilitation training exercises were determined.

7. The swallowing muscle rehabilitation training system based on an oral mouth opener according to claim 6, characterized in that, The training planning module determines the patient's muscle electrical stimulation protocol based on the patient's swallowing dysphagia characteristics, including: Determine the association between swallowing muscles and the location of electrical stimulation; Based on the patient's swallowing dysphagia characteristics and the correlation information between swallowing muscles and electrical stimulation locations, the key electrical stimulation locations for the patient are determined. Based on the patient’s swallowing dysphagia characteristics and key electrical stimulation locations, a second similar sample patient was identified from multiple sample patients; Based on the optimal swallowing muscle rehabilitation training program for patients in the first similar sample, the electrical stimulation parameters of the patients are determined, wherein the patient's muscle electrical stimulation program includes key electrical stimulation locations and electrical stimulation parameters.

Images