Peripheral nerve precise regulation method and system based on multi-channel flexible electrode

Abstract

The application provides a peripheral nerve precise regulation method and system based on a multi-channel flexible electrode, and is applied to the technical field of data processing. The physiological parameter information of a target user is processed based on a target peripheral nerve stimulation model to generate a muscle contraction characteristic comprehensive index, wherein the muscle contraction characteristic comprehensive index includes a muscle contraction amplitude, a contraction frequency, and the contraction frequency is clearly distinguished according to fast muscle, slow muscle and different movement modes. The training purpose information and the muscle contraction characteristic comprehensive index of the target user are processed based on the target peripheral nerve stimulation model to generate a preliminarily adjusted muscle contraction signal. The preliminarily adjusted muscle contraction signal is processed based on the physiological characteristic information of a target organ to generate a secondarily adjusted muscle contraction signal. The secondarily adjusted muscle contraction signal is processed based on the movement intensity interval information of the target organ to generate a target peripheral nerve stimulation signal.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to a method and system for precise peripheral nerve regulation based on multi-channel flexible electrodes. Background Art

[0002] Existing peripheral nerve stimulation methods have numerous shortcomings. Traditional methods struggle to comprehensively and accurately consider multiple aspects of information, including the target user's physiological parameters, training objectives, target organ physiological characteristics, and exercise intensity ranges, when generating muscle contraction signals. For example, they are unable to meticulously categorize muscle contraction amplitudes based on exercise phases, nor can they distinguish between fast-twitch and slow-twitch muscles, or between different exercise patterns. This results in the generated muscle contraction signals lacking specificity and accuracy.

[0003] When generating the initially adjusted muscle contraction signal, existing technologies are unable to efficiently generate a preset signal set, making it difficult to accurately extract muscle contraction signal features. They also have significant flaws in calculating the correlation matrix between muscle contraction signal features and constructing a priori knowledge graphs between muscle contraction signal features. This makes the initially adjusted signal difficult to meet actual needs, affecting the accuracy and effectiveness of peripheral nerve stimulation.

[0004] When performing secondary adjustments to the initially adjusted muscle contraction signal, existing technologies perform poorly in extracting the target organ's physiological characteristic parameters, comprehensively analyzing them, processing key data, and generating adjustment factors. This inability to fully consider the target organ's physiological characteristics results in deviations from the actual signal after secondary adjustments, making it impossible to achieve precise control of peripheral nerves.

[0005] It should be noted that the information disclosed in the above background technology section is only used to enhance the understanding of the background of the present disclosure, and therefore includes information that does not constitute prior art known to ordinary technicians in this field. Summary of the Invention

[0006] The purpose of the present application is to provide a method and system for precise control of peripheral nerves based on multi-channel flexible electrodes, which at least to a certain extent overcomes the problems existing in the prior art, and generates a comprehensive index of muscle contraction characteristics including contraction amplitude and frequency by processing physiological parameter information using a target peripheral nerve stimulation model, wherein the contraction amplitude is subdivided according to the movement stage, and the contraction frequency is distinguished by fast muscle, slow muscle and movement mode. Then, a preliminary adjusted muscle contraction signal is generated by combining the training purpose information and the comprehensive index. The preliminary signal is then processed based on the physiological characteristic information of the target organ to obtain a secondary adjusted signal. Finally, the secondary adjustment signal is processed based on the target organ movement intensity interval information to generate a target peripheral nerve stimulation signal. The entire method and system are intended to address the shortcomings of existing nerve stimulation technology, achieve precise control of peripheral nerves, provide more effective support for rehabilitation training, enhanced training, etc., and have important application value.

[0007] Other features and advantages of the present application will become apparent from the following detailed description, or may be learned in part by practice of the invention.

[0008] According to one aspect of the present application, a method for precise peripheral nerve regulation based on multi-channel flexible electrodes is provided, including: obtaining physiological parameter information of a target user, training purpose information of the target user, physiological characteristic information of a target organ, and exercise intensity interval information of the target organ; processing the physiological parameter information of the target user based on a target peripheral nerve stimulation model to generate a comprehensive muscle contraction characteristic index, wherein the comprehensive muscle contraction characteristic index includes muscle contraction amplitude and contraction frequency, the muscle contraction amplitude is subdivided into corresponding target amplitudes according to different exercise stages, and the contraction frequency is clearly distinguished according to fast muscles, slow muscles and different exercise patterns; processing the training purpose information and the comprehensive muscle contraction characteristic index of the target user based on the target peripheral nerve stimulation model to generate a preliminarily adjusted muscle contraction signal; processing the preliminarily adjusted muscle contraction signal based on the physiological characteristic information of the target organ to generate a secondary adjusted muscle contraction signal; processing the secondary adjusted muscle contraction signal based on the exercise intensity interval information of the target organ to generate a target peripheral nerve stimulation signal.

[0009] Another aspect of the present application is a peripheral nerve precision control device based on multi-channel flexible electrodes, characterized in that it includes: an acquisition module for acquiring physiological parameter information of a target user, training purpose information of the target user, physiological characteristic information of a target organ, and exercise intensity interval information of the target organ; a processing module for processing the physiological parameter information of the target user based on a target peripheral nerve stimulation model to generate a comprehensive index of muscle contraction characteristics, wherein the comprehensive index of muscle contraction characteristics includes muscle contraction amplitude and contraction frequency, and the muscle contraction amplitude is subdivided into corresponding target amplitudes according to different exercise stages, and the contraction frequency is clearly distinguished according to fast muscles, slow muscles and different exercise patterns; the training purpose information and the comprehensive index of muscle contraction characteristics of the target user are processed based on the target peripheral nerve stimulation model to generate a preliminarily adjusted muscle contraction signal; the preliminarily adjusted muscle contraction signal is processed based on the physiological characteristic information of the target organ to generate a secondary adjusted muscle contraction signal; the secondary adjusted muscle contraction signal is processed based on the exercise intensity interval information of the target organ to generate a target peripheral nerve stimulation signal.

[0010] According to another aspect of the present application, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a second processor, the method for precise peripheral nerve regulation based on multi-channel flexible electrodes is implemented.

[0011] This application provides a method and system for precise peripheral nerve regulation based on multi-channel flexible electrodes. The server uses a target peripheral nerve stimulation model to process physiological parameter information to generate a comprehensive muscle contraction characteristic index including contraction amplitude and frequency. The contraction amplitude is subdivided by movement stage, and the contraction frequency is differentiated by fast muscle, slow muscle, and movement mode. Then, a preliminary adjusted muscle contraction signal is generated by combining training purpose information and the comprehensive index. The preliminary signal is then processed based on the physiological characteristics of the target organ to obtain a secondary adjusted signal. Finally, the secondary adjusted signal is processed based on the target organ's movement intensity interval information to generate a target peripheral nerve stimulation signal. The generation of the preliminary adjustment signal involves operations such as generating a preset signal set, extracting features, calculating a correlation matrix, and constructing a priori knowledge graph. The secondary adjustment signal includes processes such as parameter extraction, comprehensive analysis, key data processing, and generation of adjustment factors. The entire method and system aims to address the shortcomings of existing nerve stimulation technology, achieve precise regulation of peripheral nerves, and provide more effective support for rehabilitation training, enhanced training, etc., with important application value.

[0012] It is to be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] Figure 1A flowchart showing a method for precise peripheral nerve regulation based on a multi-channel flexible electrode provided in one embodiment of the present application is shown;

[0014] Figure 2 A schematic structural diagram of a peripheral nerve precision control device based on multi-channel flexible electrodes provided in one embodiment of the present application is shown. DETAILED DESCRIPTION

[0015] The preferred embodiments of the present invention are described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0016] The following combination Figure 1 To describe the peripheral nerve precise control method based on multi-channel flexible electrodes according to an exemplary embodiment of the present application. Figure 1 As shown, the method is applied to the server and includes:

[0017] S101, obtaining physiological parameter information of a target user, training purpose information of the target user, physiological characteristic information of a target organ, and exercise intensity interval information of the target organ.

[0018] In one embodiment, the physiological parameter information of the target user includes basic information such as the target user's age, gender, and basic physical health status, as well as parameters directly related to muscle contraction, such as muscle fiber type, muscle strength, muscle endurance, nerve conduction velocity, etc. These parameters reflect the user's physical functions and physiological characteristics, and are crucial for the subsequent determination of appropriate neural stimulation strategies. For example, users of different ages and genders have different physiological characteristics of muscles and nerves, which will affect the effect of neural stimulation. The muscle fiber type determines the speed and strength characteristics of muscle contraction. Fast muscle fibers contract quickly and have great strength but are prone to fatigue, while slow muscle fibers are the opposite. Understanding these will help to develop targeted stimulation plans.

[0019] The training purpose information of the target user is mainly used to clarify the purpose of the user's training, including but not limited to rehabilitation training and enhancement training. Rehabilitation training aims to promote nerve regeneration and restore muscle function, and is suitable for patients whose neuromuscular function is impaired due to injury or disease, such as patients with limb paralysis after a stroke. Enhancement training is to improve muscle strength, endurance and other performance, and is generally used by athletes or people pursuing higher athletic ability. Different training purposes have different requirements for nerve stimulation. Rehabilitation training usually requires lower intensity and longer duration stimulation to promote nerve repair and muscle function recovery; enhancement training requires higher intensity and more targeted stimulation to stimulate the maximum potential of muscles.

[0020] Information on the physiological properties of target organs primarily involves the physiological characteristics of target organs (such as the muscles of the limbs and the nerves that control them), including muscle fiber type, nerve conduction velocity, muscle elasticity, and extensibility. As mentioned previously, muscle fiber type, as different types of muscle fibers respond differently to stimulation, is an important basis for developing stimulation strategies. Nerve conduction velocity affects the efficiency of nerve signal transmission. Slow conduction velocity may lead to delayed muscle contraction, and compensation for this factor needs to be considered when designing stimulation protocols. Muscle elasticity and extensibility affect the range of motion and flexibility of muscles. For training that requires extensive movement, understanding these muscle properties can help avoid injuries caused by excessive stretching or contraction.

[0021] The exercise intensity interval information of the target organ describes the range of exercise intensity required by the target organ in different exercise states, and is the key basis for determining the intensity and frequency of nerve stimulation. Exercise intensity intervals can be divided according to different exercise types and goals. For example, for simple daily activities, the exercise intensity is low; while for high-intensity sports and competitive activities, the exercise intensity is high. Understanding the exercise intensity interval of the target organ can enable nerve stimulation to more accurately match exercise needs and avoid excessive or insufficient stimulation. For example, when doing strength training, a higher stimulation intensity and frequency are required to meet the exercise needs of the muscles; while when doing endurance training, a relatively low stimulation intensity and a higher stimulation frequency are required to maintain sustained muscle contraction.

[0022] S102: Processing the target user's physiological parameter information based on the target peripheral nerve stimulation model to generate a comprehensive muscle contraction characteristic index.

[0023] In one embodiment, training data, clinical verification data and a preset initial model of peripheral nerve stimulation are obtained. A large amount of data related to the physiological parameters of the target user, the purpose of training, the physiological characteristics of the target organ and the exercise intensity range is collected. For example, muscle contraction data of users of different ages, genders and physical conditions during rehabilitation training and enhancement training are collected from rehabilitation centers, sports training institutions, etc., including information such as muscle contraction amplitude, frequency, stimulation intensity and duration, as well as physiological parameters and exercise intensity data of the corresponding target organs. Clinical verification data comes from the clinical practice of medical institutions, covering actual data of patients with neuromuscular dysfunction caused by different diseases or injuries during the process of receiving neurostimulation treatment. Such as muscle reaction data of stroke patients during rehabilitation training, muscle state change data of patients with muscular atrophy, etc., which are used to verify the effectiveness of the model in actual clinical scenarios.

[0024] The pre-set initial peripheral nerve stimulation model uses a deep learning neural network model, such as a multilayer perceptron (MLP), as its underlying structure. This model consists of an input layer, hidden layers, and an output layer. The input layer receives information such as the target user's physiological parameters, training purpose, target organ physiological characteristics, and exercise intensity ranges. The hidden layer extracts and transforms this input information using a nonlinear activation function. The output layer outputs preliminary muscle contraction signal predictions. Relevant model parameters, such as the weights and biases of the hidden layer neurons, are randomly initialized during the initial stage to prepare for subsequent training.

[0025] The training data is cleaned, features are extracted, and normalized to generate preprocessed feature data. The collected training data is cleaned to remove noise, outliers, and duplicate data. For example, in muscle contraction amplitude data, if there are outliers that deviate significantly from the normal range, such as sudden maximum or minimum values, statistical methods (such as those based on mean and standard deviation) can be used to identify and eliminate them. Key features are extracted from the training data, such as muscle contraction amplitude, contraction frequency, stimulation intensity, and duration. Taking muscle contraction amplitude as an example, target amplitudes are subdivided according to different stages of exercise, such as amplitude changes during simple daily activities and exercise training. Contraction frequency is extracted based on fast and slow twitch muscles and different movement patterns. The extracted features are normalized, mapping the data to a specific interval (e.g., [0, 1]) to eliminate dimensional differences between different features and improve model training effectiveness. For example, for muscle contraction amplitude data, the maximum and minimum values ​​are normalized to make amplitude data from different samples comparable.

[0026] The preprocessed feature data is processed to generate a stimulation model prediction parameter vector, which is used to represent prediction strategy information and parameter information for muscle contraction amplitude, frequency, stimulation intensity, and duration. The preprocessed feature data is processed, and through model training and learning, a stimulation model prediction parameter vector is generated. During model training, the backpropagation algorithm is used to continuously adjust the model's weights and biases to ensure that the model's prediction results better match the actual training data. The stimulation model prediction parameter vector is used to represent prediction strategy information and parameter information for muscle contraction amplitude, frequency, stimulation intensity, and duration. For example, the vector contains predicted values ​​for muscle contraction amplitude for different training purposes (rehabilitation training or plyometric training) and target organ physiological characteristics (organs with a high proportion of fast or slow muscle), as well as recommended parameters for the corresponding stimulation intensity and duration.

[0027] Based on the stimulation model prediction parameter vector, the preset initial model of peripheral nerve stimulation is optimized and trained to generate a trained peripheral nerve stimulation model. During the training process, the pre-processed feature data is input into the model, and the model adjusts its own parameters according to the prediction parameter vector so that the muscle contraction signal output by the model is closer to the muscle contraction in the actual training data. After multiple rounds of training, the model gradually learns the patterns and features in the data and generates a trained peripheral nerve stimulation model. When processing the input data, this model can more accurately predict the muscle contraction amplitude, frequency, stimulation intensity and duration, providing strong support for subsequent precise regulation.

[0028] Based on the clinical validation data, the trained peripheral nerve stimulation model is subjected to simulated peripheral nerve stimulation to generate validation results. Based on the validation results, the trained peripheral nerve stimulation model is evaluated and adjusted to generate a target peripheral nerve stimulation model. The trained peripheral nerve stimulation model is subjected to simulated peripheral nerve stimulation based on the clinical validation data. The clinical validation data is input into the trained model, which then outputs a predicted muscle contraction signal. The predicted signal output by the model is compared with the actual muscle contraction signal from the clinical validation data to generate a validation result. For example, the error between the predicted and actual signals, such as the mean squared error (MSE), is calculated to assess model accuracy. Based on the validation results, the trained peripheral nerve stimulation model is evaluated and adjusted. If the model error is large, the model structure is adjusted (e.g., increasing the number of hidden layer neurons, adjusting the activation function) or retraining the model (adjusting the learning rate, optimizing the algorithm) to continuously optimize model performance, ultimately generating a target peripheral nerve stimulation model.

[0029] In another embodiment, feature extraction is performed on the target user's physiological parameter information to generate muscle contraction amplitude features and contraction frequency features. The muscle contraction amplitude features include target amplitudes corresponding to different exercise stages, while the contraction frequency features are generated based on fast-twitch muscle, slow-twitch muscle, and different exercise patterns. The target user's physiological parameter information includes basic information such as age, gender, and general health status, as well as parameters closely related to muscle contraction, such as muscle fiber type, muscle strength, muscle endurance, and nerve conduction velocity. Users of different ages and genders have different physiological characteristics of muscles and nerves, which can affect muscle contraction performance. From this rich physiological parameter information, muscle contraction amplitude features and contraction frequency features are extracted. Regarding muscle contraction amplitude features, target amplitudes for muscle contraction vary during different exercise stages, such as daily activities, rehabilitation training, and competitive sports. For example, during daily walking, the contraction amplitude of leg muscles is small and relatively stable; whereas, during strength training, such as weightlifting, the contraction amplitude of upper limb muscles increases significantly. The contraction frequency features are generated based on fast-twitch muscle, slow-twitch muscle, and different exercise patterns. Fast muscle fibers contract quickly and are suitable for rapid explosive movements. For example, during sprinting, fast muscle fibers dominate contraction and have a higher contraction frequency. Slow muscle fibers contract slowly but have strong endurance. In endurance sports such as long-distance running, the contraction frequency of slow muscle fibers is relatively low.

[0030] The muscle contraction amplitude characteristics are quantitatively analyzed and processed to generate amplitude quantification factors. In the rehabilitation training scenario, multiple contraction amplitude measurements are performed on specific muscles at specific movement stages. Taking the hand grasping rehabilitation training of hemiplegic patients as an example, the biceps brachii of their hands are selected as the research object. During each grasping action, a high-precision displacement sensor or other measuring equipment is used to record the length change of the biceps brachii during contraction, and then the muscle contraction amplitude data is obtained. In one training cycle, 100 measurements were performed to form a data set containing 100 data. The mean is an important statistic that describes the trend in the data set. By calculating the mean of this set of data, the average level of contraction amplitude of the muscle in this movement stage can be obtained. Assume that these 100 measurement data are , mean The calculation formula is If the calculated mean is 5cm, this means that the average contraction amplitude of the muscle during this rehabilitation training phase is 5cm. The mean value can reflect the overall level of muscle contraction amplitude, but it cannot reflect the degree of dispersion of the data.

[0031] The standard deviation is used to measure the degree of dispersion of the data, that is, the fluctuation of the data. The larger the standard deviation, the greater the dispersion of the data, and the greater the difference in the muscle contraction amplitude at different measurement times; conversely, the smaller the standard deviation, the more concentrated the data. The calculation formula of standard deviation s is , if the calculated standard deviation is 1 cm, this indicates that during this rehabilitation training phase, the fluctuation range of the muscle contraction amplitude around the mean of 5 cm is relatively small, and most of the measured values ​​are between 4 cm and 6 cm (mean After obtaining the mean and standard deviation, these values ​​need to be further processed according to a specific quantization formula or algorithm to form an amplitude quantization factor. This specific quantization formula or algorithm will take into account factors such as the goal of rehabilitation training, the physiological characteristics of the muscle, and subsequent application requirements. For example, in a simple quantization algorithm, the amplitude quantization factor Q can be calculated using the following formula , where k is the adjustment coefficient set according to the actual situation. Assume that in this example Substituting the mean 5cm and standard deviation 1cm into the formula, we get This amplitude quantification factor 10 comprehensively reflects the size and stability of the muscle's contraction amplitude during this rehabilitation training phase. In subsequent analyses, the amplitude quantification factor can be used to compare differences in muscle contraction amplitude characteristics across different rehabilitation phases, patients, or training methods, providing a quantitative basis for developing personalized rehabilitation training programs and evaluating training effectiveness.

[0032] Contraction frequency characteristics are quantitatively analyzed and processed to generate frequency quantification factors. When studying muscle contraction frequency during different exercise modes, it is necessary to collect a large amount of data on the contraction frequencies of fast-twitch and slow-twitch fibers. For example, using two typical exercise modes, sprinting and long-distance running, real-time monitoring of the muscles involved in the exercise is performed using electromyography (EMG) or other advanced muscle activity monitoring equipment. In sprinting, the quadriceps femoris muscle is selected as the research subject. During multiple sprint training sessions, each lasting approximately 30 seconds, the contraction frequencies of fast-twitch and slow-twitch fibers are recorded during each contraction. Assuming 50 sprint training sessions are recorded, the contraction frequency of fast-twitch fibers fluctuates between 8 and 12 times per second, while the contraction frequency of slow-twitch fibers fluctuates between 3 and 6 times per second. Similarly, for long-distance running, the same muscle is selected and recorded during multiple long-distance training sessions (each lasting approximately 30 minutes), revealing a contraction frequency of 4 to 7 times per second for fast-twitch fibers and 4 to 6 times per second for slow-twitch fibers. This large amount of data provides rich material for subsequent analysis.

[0033] Exercise intensity is a key factor influencing muscle contraction frequency. In high-intensity sports like sprinting, fast-twitch fibers, due to their rapid contraction characteristics, are recruited in large numbers, resulting in a higher contraction frequency. Long-distance running, on the other hand, is a low-intensity, long-duration exercise. Slow-twitch fibers, due to their endurance advantages, are the primary fibers involved, and their contraction frequency is relatively low. Analysis of the collected data reveals that with increasing exercise intensity, fast-twitch fiber contraction frequency increases significantly, while slow-twitch fiber contraction frequency fluctuates within a certain range. For example, when sprinting speed increases by 20%, fast-twitch fiber contraction frequency increases from 10 to 12 times per second. In long-distance running, when speed changes are relatively small, slow-twitch fiber contraction frequency remains relatively stable at 5-6 times per second. Exercise duration is also closely related to muscle contraction frequency. During long-distance running, as exercise duration increases, fast-twitch fibers fatigue easily, causing their contraction frequency to gradually decrease. However, slow-twitch fibers, due to their excellent endurance, maintain a relatively stable or slightly increasing contraction frequency. For example, during a 60-minute long-distance run, the average contraction frequency of fast-twitch muscle fibers is 6 times per second in the first 30 minutes, dropping to 5 times per second in the last 30 minutes. Meanwhile, the contraction frequency of slow-twitch muscle fibers is 5 times per second in the first 30 minutes, rising to 5.5 times per second in the last 30 minutes. By analyzing muscle contraction frequency at different exercise durations, we can understand changes in muscle fatigue and adaptive adjustments during exercise.

[0034] After collecting a large amount of data and conducting comprehensive analysis, a specific algorithm is used to process this data to generate a frequency quantification factor. A weighted average algorithm is used to determine the weight based on the degree of influence of exercise intensity and exercise duration on the contraction frequency of fast muscle fibers and slow muscle fibers. Assuming that the weight of exercise intensity is , the weight of the motion duration is ,and For sprinting, it was determined through analysis Fast-twitch fiber contraction frequency , slow-twitch fiber contraction frequency , the calculation formula of the frequency quantization factor F is Substituting the values ​​into the formula yields: This frequency quantification factor of 8.5 comprehensively considers the contraction frequency of fast-twitch and slow-twitch muscles in sprinting, as well as the effects of exercise intensity and duration. This factor more comprehensively reflects the quantitative characteristics of muscle contraction frequency in sprinting. By comparing frequency quantification factors across different exercise modes, we can analyze the differences in the effects of different exercises on muscle contraction frequency, providing important quantitative evidence for the development of exercise training programs and the assessment of muscle function in rehabilitation treatments.

[0035] Based on the target peripheral nerve stimulation model, the amplitude and frequency quantization factors are processed to generate contraction feature assessment information and corresponding weight calculation results. Taking hand grasping and leg walking as examples, hand grasping emphasizes fine motor control and requires a high degree of precision in muscle contraction amplitude; leg walking, on the other hand, requires sustained and stable muscle contraction, placing certain demands on contraction frequency and endurance. Regarding target organs, for example, the ratio of fast-twitch to slow-twitch fibers in hand muscles differs from that in leg muscles. Hand muscles have a relatively low proportion of fast-twitch fibers and a relatively high proportion of slow-twitch fibers, which allows the hand to have greater endurance during prolonged, low-intensity grasping movements. Leg muscles, on the other hand, have varying fiber ratios in different areas, with thigh muscles having a higher proportion of fast-twitch fibers to meet the explosive power requirements of movements like walking and running.

[0036] The model will perform weighted calculations on the amplitude and frequency quantization factors based on the proportion of muscle fiber types in the target organ. Assume that in a certain target organ, the proportion of fast muscle fibers is p and the proportion of slow muscle fibers is 1-p. For the amplitude quantization factor A and the frequency quantization factor F, the weights related to fast muscle fibers are set as , the weight associated with slow-twitch fibers is ,and When the target organ has a high proportion of fast muscle fibers (p=0.6), in order to highlight the importance of the quantitative factors related to fast muscle fiber contraction, we set In this way, when calculating the comprehensive amplitude quantization value and the integrated frequency quantization value When This weighted calculation method can more accurately reflect the effects of different fiber types on muscle contraction, thereby more reasonably evaluating the state of muscle contraction.

[0037] After obtaining the quantified comprehensive amplitude and frequency values, the model determines the rationality of muscle contraction amplitude and frequency at different stages of movement. For example, during a certain stage of fine grasping, previous research suggests that the reasonable muscle contraction amplitude ranges from 2-5 cm, and the contraction frequency ranges from 3-6 times per second. If the calculated comprehensive amplitude and frequency values ​​fall within these ranges, the model deems the current muscle contraction state reasonable; if they fall outside these ranges, an abnormality is present. The model also considers the characteristics of the movement pattern in its judgment. During rapid grasping movements, muscle contraction amplitude and frequency may increase momentarily; as long as they return to normal within a reasonable timeframe, this is considered reasonable.

[0038] By assessing the rationality of muscle contraction amplitude and frequency, the model generates contraction characteristic assessment information. This assessment includes whether the muscle contraction intensity is sufficient, the frequency is stable, and whether the contraction state meets the requirements of the current exercise pattern and target organ. For example, if the assessment results indicate that muscle contraction intensity is weak and the frequency fluctuates significantly during a particular exercise phase, adjustments to the neurostimulation parameters may be necessary to improve muscle contraction effectiveness. During this calculation and assessment process, the model also determines the weights of each quantitative factor in the overall assessment, generating weighted calculation information. These weights not only reflect the importance of quantitative factors associated with different fiber types but also provide a basis for further optimizing neurostimulation strategies. For example, if the weight of the amplitude quantitative factor associated with fast-twitch fibers is found to have a greater impact on the assessment results during a particular exercise pattern, the neurostimulation design can prioritize adjustments to stimulation parameters related to fast-twitch fiber contraction, such as intensity and frequency, to achieve more precise muscle contraction control.

[0039] The contraction feature evaluation information and the corresponding weight calculation result information are processed to generate a comprehensive muscle contraction feature index, wherein the muscle contraction feature comprehensive index is used to characterize the physiological state characteristics of the target user's muscles during the contraction process, such as amplitude and frequency. The muscle contraction feature comprehensive index is a comprehensive characterization of the physiological state characteristics of the target user's muscles during the contraction process, such as amplitude and frequency. For example, in the rehabilitation training effect evaluation scenario, by comprehensively considering the contraction feature evaluation information and the weight calculation result information, the muscle contraction feature comprehensive index can intuitively reflect the recovery of the patient's muscle function. If, after a period of rehabilitation training, the quantitative factor corresponding to the muscle contraction amplitude feature performs better and better in the comprehensive index, and the contraction frequency feature gradually approaches the normal range, it means that the patient's muscle function is gradually recovering, providing an important basis for doctors to adjust the nerve stimulation strategy.

[0040] S103: Processing the training purpose information and muscle contraction characteristic comprehensive index of the target user based on the target peripheral nerve stimulation model to generate a preliminarily adjusted muscle contraction signal.

[0041] In one embodiment, a target peripheral nerve stimulation model processes a dataset consisting of a target user's training purpose information and comprehensive muscle contraction characteristic indicators to generate a preset muscle contraction signal set. The target peripheral nerve stimulation model receives the target user's training purpose information (such as rehabilitation training, strengthening training, etc.) and a dataset consisting of comprehensive muscle contraction characteristic indicators (including muscle contraction amplitude, contraction frequency, etc.). Taking a patient undergoing rehabilitation training after a stroke as an example, their comprehensive muscle contraction characteristic indicators show a small muscle contraction amplitude and unstable contraction frequency. During the simulation process, the model systematically varies parameters such as stimulation intensity, frequency, and duration. Taking stimulation intensity as an example, the model may start at a low intensity and gradually increase the intensity to observe the theoretical response of muscle contraction at different intensities. For stimulation frequency, multiple combinations from slower to faster frequencies will be tried. The stimulation duration will also vary from short to long. In the case of this stroke rehabilitation patient, the model may first simulate a stimulation pattern with a lower stimulation intensity (such as 0.5mA), a lower frequency (such as 20Hz), and a shorter duration (such as 500ms), and calculate the predicted muscle contraction under this pattern; then, it adjusts the parameters to simulate a stimulation pattern with a higher stimulation intensity (such as 1.2mA), a higher frequency (such as 50Hz), and a longer duration (such as 800ms), and so on, generating neural stimulation patterns with multiple different parameter combinations.

[0042] By simulating different neural stimulation modes, the model obtains a series of possible muscle contraction signals, which together constitute a set of preset muscle contraction signals. Each signal is a prediction of the patient's muscle contraction response under a specific combination of stimulation parameters. In this set, different signals have different characteristics. For example, some signals may correspond to larger predicted values ​​of muscle contraction amplitude, which may be generated under a combination of higher stimulation intensity and appropriate frequency; while some signals may have more stable contraction frequency predictions, which may be achieved by adjusting the relationship between stimulation duration and frequency. This set of preset muscle contraction signals contains a wealth of possibilities, providing a broad selection basis for the subsequent screening of muscle contraction signals that best suit the patient's current condition. In subsequent steps, the model will further analyze and screen these signals to find the signal that best meets the patient's rehabilitation needs, so as to achieve precise neural stimulation regulation and promote the recovery of the patient's muscle function.

[0043] The preset muscle contraction signal set is processed to generate contraction intensity characteristics and frequency change characteristics, as well as their frequency of occurrence under different training purposes. Within the preset muscle contraction signal set, the changes in muscle contraction intensity under different signals are observed to determine the contraction intensity characteristics. At the same time, the changes in muscle contraction frequency over time or at different training stages are analyzed to obtain frequency change characteristics. The frequency of occurrence of these characteristics under different training purposes (such as different stages of rehabilitation training and different types of plyometric training) is statistically analyzed. For example, in the early stages of rehabilitation training, signals with lower contraction intensities and slower frequency changes appear more frequently. As the rehabilitation process progresses, the frequency of signals with higher contraction intensities and more regular frequency changes gradually increases.

[0044] The contraction intensity feature, frequency change feature and corresponding occurrence frequency information are processed to generate feature correlation matrix information. The calculation formula for generating feature correlation matrix information is: ;in, The values ​​of 1 and 2 represent the contraction intensity feature and frequency change feature respectively; n is the number of samples; is the kth sample value in the contraction strength feature vector X, which represents the observed value of contraction strength under a certain training purpose; is the frequency change eigenvector The k-th sample value in represents the corresponding frequency change observation value under the same training purpose; is the weight vector The kth element in reflects the frequency of occurrence of the kth sample under different training purposes. The higher the frequency, The larger it is, the more important it is used to adjust the sample's importance when calculating correlation. is the contraction strength feature vector The mean of , used to measure the average level of contraction strength characteristics; Frequency change eigenvector The mean of , used to measure the average level of frequency change characteristics; is the element in the feature correlation matrix R, is 1 (correlation between itself and itself), It represents the correlation coefficient between the contraction intensity feature and the frequency change feature. Its value range is between -1 and 1. The closer the absolute value is to 1, the stronger the correlation between the two features; the closer the absolute value is to 0, the weaker the correlation.

[0045] Assume that during the rehabilitation training, data of 10 training phases n=10 are collected, and the contraction intensity feature vector , frequency change feature vector , weight vector .

[0046] First calculate .

[0047] Then calculate (Correlation coefficient between contraction intensity and frequency change characteristics):

[0048]

[0049] but , indicating that there is a strong positive correlation between the contraction intensity feature and the frequency change feature. is 1 (used to indicate the correlation between itself and itself).

[0050] Based on the feature correlation matrix information, a priori knowledge graph is generated to characterize the feature association relationship of muscle contraction signals under different training purposes. The priori knowledge graph graphically displays the connection between contraction intensity features, frequency change features, and training purposes. In the graph, nodes represent different features (such as contraction intensity and frequency change) and training purposes, and the weights of edges are determined by the feature correlation matrix information. The larger the weight, the stronger the association between the two nodes. For example, there may be a strong edge connection between the rehabilitation training purpose node and the high contraction intensity and stable frequency change feature nodes, indicating that in rehabilitation training, these two features often appear simultaneously and are closely related to each other.

[0051] Based on the prior knowledge graph, the target user's training purpose information and comprehensive muscle contraction characteristic indicators are processed to generate a preliminarily adjusted muscle contraction signal. Using the prior knowledge graph, combined with the target user's training purpose information and comprehensive muscle contraction characteristic indicators, the model optimizes and adjusts the current muscle contraction signal to generate a preliminarily adjusted muscle contraction signal. For the above-mentioned stroke rehabilitation patients, the model adjusts the signals in the preset muscle contraction signal set based on the correlation between the rehabilitation training stage and the specific contraction intensity and frequency change characteristics in the prior knowledge graph, as well as the patient's current comprehensive muscle contraction characteristic indicators, to determine a preliminarily adjusted muscle contraction signal that better meets the patient's current rehabilitation needs. For example, if the prior knowledge graph shows that in the middle stage of rehabilitation training, signals with higher contraction intensity and stable frequency changes are more conducive to muscle function recovery, the model will screen and adjust the signals in the preset set so that the preliminarily adjusted muscle contraction signal has these characteristics, providing a basis for further precise regulation in the future.

[0052] S104: Processing the initially adjusted muscle contraction signal based on the physiological characteristic information of the target organ to generate a secondary adjusted muscle contraction signal.

[0053] In one embodiment, the muscle contraction signal after the preliminary adjustment is subjected to target parameter extraction and data classification processing to generate stimulation intensity information, stimulation frequency information, and stimulation duration information. The muscle contraction signal after the preliminary adjustment contains a variety of potential information that can be used for precise regulation. Target parameter extraction and data classification processing are performed on it to separate the stimulation intensity information, stimulation frequency information, and stimulation duration information. Taking a patient undergoing rehabilitation training as an example, the muscle contraction signal after the preliminary adjustment may appear as a series of electrical signal changes at a certain moment. Through a specific signal processing algorithm, the current stimulation intensity, such as 0.8mA; the stimulation frequency, assuming it is 30Hz; and the stimulation duration, such as each stimulation lasting 600ms, are extracted from these signals. This information is the basis for subsequent analysis and adjustment, and they directly affect the muscle contraction effect and the effectiveness of nerve stimulation.

[0054] Based on the target organ's physiological characteristics and the corresponding characteristic standard version number, a comprehensive analysis of stimulation intensity, frequency, and duration is performed to generate parameter importance assessment information and physiological suitability assessment information. The target organ's physiological characteristics include its muscle fiber type and nerve conduction velocity. Different muscle fiber types, such as fast-twitch and slow-twitch fibers, respond differently to different stimulation parameters. Fast-twitch fibers contract quickly and are more suited to higher-frequency and higher-intensity stimulation; slow-twitch fibers respond better to lower-frequency and lower-intensity stimulation. Nerve conduction velocity also influences stimulation effectiveness; slower-conducting nerves require longer stimulation durations to elicit effective muscle contraction. The characteristic standard version number ensures that the analysis uses standards consistent with current medical and scientific understanding. For example, for a target muscle with a high proportion of fast-twitch fibers and normal nerve conduction velocity, if the current stimulation intensity and frequency are both low, a comprehensive analysis may conclude that this stimulation parameter combination is ineffective in activating fast-twitch fibers. This comprehensive analysis generates parameter importance assessment information and physiological suitability assessment information. The parameter importance evaluation information is used to determine the importance of each stimulation parameter under the current physiological state, and the physiological adaptability evaluation information measures the degree of match between these parameters and the physiological characteristics of the target organ.

[0055] Based on the parameter importance evaluation information and physiological adaptability evaluation information, the key data in the preliminarily adjusted muscle contraction signal are screened and correlated to generate a key muscle contraction data set. In the above example where the proportion of fast muscle fibers is relatively high, if the evaluation finds that the stimulation intensity has the greatest impact on the muscle contraction effect, then the data related to the higher stimulation intensity will be screened out. At the same time, considering the correlation between the stimulation frequency and duration and the intensity, the frequency and duration data related to them within a reasonable range will also be screened out. These screened data are correlated with each other to form a key muscle contraction data set. The data in this set is determined based on the physiological characteristics of the target organ and the current stimulation effect evaluation, and is crucial for the subsequent further optimization of the muscle contraction signal.

[0056] Key muscle contraction data sets are integrated and processed, combining the physiological characteristics of the target organ with weight information from a pre-set multi-knowledge base to generate a muscle contraction signal adjustment factor. The pre-set multi-knowledge base is a database containing extensive expertise, encompassing a wide range of information on muscle contraction under different physiological conditions. A key component of this is the optimal stimulation parameter range for different muscle fiber types under different exercise modes. For example, in a fast, explosive exercise like sprinting, fast-twitch fibers predominate. Their optimal stimulation parameters may be higher intensities (e.g., 1-2 mA) and frequencies (80-100 Hz). These fibers respond quickly to high-frequency, high-intensity stimulation, generating powerful contraction forces. In endurance exercises like long-distance running, slow-twitch fibers predominate. They are more suited to lower intensities (0.5-1 mA) and frequencies (30-50 Hz) to maintain prolonged contractions without fatigue. The relationship between nerve conduction velocity and stimulation effectiveness is also crucial to the knowledge base. A slower nerve conduction velocity means that nerve signals take longer to reach the muscle. Therefore, a longer stimulation duration may be necessary to induce effective muscle contraction. For example, under normal nerve conduction velocity, the stimulation duration may be 500ms, while when the nerve conduction velocity is slower, it needs to be extended to 800ms to achieve the same muscle contraction effect. This knowledge provides a scientific basis for subsequent analysis and calculations.

[0057] The weight information is set according to the importance of different knowledge in practical applications. In the research and practice of muscle contraction, certain knowledge is more critical for specific physiological states and movement needs. For example, in rehabilitation training, for the goal of muscle strength recovery, the knowledge weight of the effect of stimulation intensity on muscle contraction strength may be higher. Because in the early stages of rehabilitation, the appropriate stimulation intensity can directly promote the activation and growth of muscle fibers, and play a key role in the recovery of muscle strength. In some training scenarios with higher requirements for movement accuracy, the knowledge weight of the relationship between stimulation frequency and muscle contraction accuracy will be greater, because precise control of stimulation frequency helps to achieve more refined muscle movements.

[0058] During the integration operation, the data from the key data set, physiological characteristic information, and knowledge from the knowledge base are combined. Suppose the key data set indicates that the current stimulation intensity is slightly low, for example, 0.6 mA. However, multiple knowledge bases indicate that, under the physiological conditions of the target organ (e.g., upper limb muscles with a high proportion of fast-twitch fibers), appropriately increasing the stimulation intensity (e.g., to 0.8-1 mA) can significantly improve muscle contraction. This knowledge has a high weight (e.g., 0.8). In this case, when calculating the adjustment factor, this weight will be used to increase the stimulation intensity. Assuming other factors (such as the knowledge related to stimulation frequency and duration, which have weights of 0.1 and 0.1, respectively) are also included in the calculation, the adjustment factor for stimulation intensity is calculated as: (target intensity - current intensity) × stimulation intensity knowledge weight, or (0.9 - 0.6) × 0.8 = 0.24. This means that when generating the muscle contraction signal adjustment factor, the stimulation intensity will be adjusted toward an increase of 0.24 mA. Ultimately, by comprehensively considering key data, physiological characteristics, and knowledge base knowledge, and combining their respective weights, a muscle contraction signal adjustment factor is calculated. This adjustment factor represents the direction (such as increasing or decreasing stimulation intensity, frequency, duration, etc.) and degree (specific adjustment value) of optimizing the current muscle contraction signal, providing guidance for generating more accurate and effective secondary adjusted muscle contraction signals.

[0059] The initially adjusted muscle contraction signal is processed based on the muscle contraction information adjustment factor to generate a secondary adjusted muscle contraction signal. Assuming that the adjustment factor indicates that the stimulation intensity needs to be increased by 0.2 mA, the stimulation frequency needs to be increased by 5 Hz, and the stimulation duration needs to be extended by 100 ms, then the initially adjusted muscle contraction signal is modified according to this adjustment scheme. After such processing, the secondary adjusted muscle contraction signal is more in line with the physiological needs of the target organ and the current training or treatment purpose. In rehabilitation training scenarios, the secondary adjusted muscle contraction signal can more effectively promote muscle function recovery and provide more reliable support for achieving precise peripheral nerve regulation.

[0060] S105 , processing the secondary adjusted muscle contraction signal based on the exercise intensity interval information of the target organ to generate a target peripheral nerve stimulation signal.

[0061] In one embodiment, the target organ's exercise intensity interval information and the secondary adjusted muscle contraction signal are quantitatively analyzed and processed to generate a lower limit exercise intensity quantification factor, an upper limit exercise intensity quantification factor, a contraction intensity quantification factor, a contraction frequency quantification factor, and a stimulation duration quantification factor. Taking an athlete doing strength training as an example, the exercise intensity interval of their target organ (such as the biceps) may vary within a certain range, assuming the lower limit is 30% of the maximum contraction intensity and the upper limit is 80% of the maximum contraction intensity. By analyzing the secondary adjusted muscle contraction signal, the quantitative value of the current contraction intensity within this interval is determined, and then the contraction intensity quantification factor is generated; based on the frequency and duration information in the signal, the contraction frequency quantification factor and the stimulation duration quantification factor are generated respectively. These quantification factors are the basis for subsequent analysis, and they convert the characteristics of the exercise intensity and muscle contraction signals into numerical values ​​that can be used for calculation and comparison.

[0062] The real-time physiological data of the target organ during muscle contraction is processed for feature extraction to generate muscle tension features, muscle fatigue features, nerve conduction velocity features, and metabolite concentration features. In the aforementioned strength training scenario, sensors are used to monitor the muscle tension of the biceps brachii in real time. As training progresses, muscle tension changes, and these changes are recorded to generate muscle tension features. Muscle fatigue is assessed by analyzing the duration and intensity of muscle contraction to generate muscle fatigue features. Nerve conduction velocity is obtained using relevant detection technologies to generate nerve conduction velocity features. Metabolite concentrations in blood or muscle tissue are detected to generate metabolite concentration features. These features reflect the real-time physiological state of the muscle during contraction and provide a basis for in-depth understanding of the working conditions of the muscle.

[0063] Based on the lower limit quantification factor for exercise intensity, the upper limit quantification factor for exercise intensity, and the contraction strength quantification factor, muscle tension characteristics and nerve conduction velocity characteristics are fused to generate quantitative features related to exercise intensity and contraction strength. These factors quantitatively represent the range of exercise intensity for the target organ and the current muscle contraction strength. They reflect the boundaries of exercise intensity and the relative position of the current contraction strength within this range. Muscle tension characteristics directly reflect the amount of force generated during muscle contraction, while nerve conduction velocity characteristics influence the efficiency of nerve signal transmission to the muscle, thereby affecting the response speed and intensity of muscle contraction. These factors are crucial for accurately assessing muscle contraction status. For example, in a weightlifting athlete, the exercise intensity range of the target muscle (e.g., quadriceps femoris) during a training session was quantitatively analyzed. The lower limit quantification factor for exercise intensity was set to 30% (representing the lowest effective exercise intensity level that the muscle can withstand during the session), and the upper limit quantification factor was set to 80% (representing the highest exercise intensity level that is safe and effective for training results). During training, when athletes approach the upper limit of exercise intensity (e.g., 75%), muscle status is monitored. Low muscle tension at this point indicates that the force generated by the muscles is not commensurate with the current exercise intensity. Normal nerve conduction velocity, however, indicates that nerve signals are being transmitted to the muscles, but the muscles are not fully responding. This suggests that muscle contraction strength may not be optimal and cannot fully meet the demands of the current exercise intensity. In this case, further analysis and adjustments are necessary to optimize muscle contraction.

[0064] To more accurately measure the correlation between exercise intensity and contraction strength, a weighted average method is used to calculate the associated quantitative feature values. First, the weights for each quantitative factor and feature must be determined. Weights are typically determined based on extensive experimental data, professional knowledge, and practical experience. For example, in strength training exercises like weightlifting, previous research and practice have found that the upper limit of exercise intensity has a significant impact on determining whether muscle contraction strength is appropriate. Therefore, it is given a higher weight, assuming 0.4. The contraction strength factor directly reflects the actual state of muscle contraction and is weighted 0.3. The muscle tension feature intuitively reflects the strength of muscle contraction and is weighted 0.2. The nerve conduction velocity feature is weighted 0.1. A weighted average is then calculated based on these weights. Assuming the upper limit of exercise intensity is 0.8 (corresponding to 80% of the upper limit intensity), the contraction strength factor is 0.6 (indicating that the current contraction intensity is 60% of the upper limit intensity), the quantified value of the muscle tension feature is 0.5 (assuming a range of 0-1 represents relative tension), and the quantified value of the nerve conduction velocity feature is 0.8 (also using a range of 0-1 to represent the ratio relative to normal velocity). According to the weighted average formula: associated quantitative characteristic value = exercise intensity upper limit quantitative factor × its weight + contraction intensity quantitative factor × its weight + muscle tension characteristic × its weight + nerve conduction velocity characteristic × its weight, that is: 0.8×0.4+0.6×0.3+0.5×0.2+0.8×0.1=0.68.

[0065] The calculated correlation quantization eigenvalue is 0.68, which can help determine the degree of match between the current muscle contraction strength and exercise intensity. Generally speaking, an ideal correlation quantization eigenvalue range is set (for example, 0.8-1 indicates a good match between muscle contraction strength and exercise intensity). If 0.68 is below this range, it means that the current muscle contraction strength is not well adapted to the exercise intensity requirements. Based on this, trainers or systems can adjust the neural stimulation strategy accordingly, such as appropriately increasing the stimulation intensity and optimizing the stimulation frequency, to improve muscle contraction strength, so that it can better adapt to the exercise intensity and achieve better training results. It also helps prevent sports injuries caused by insufficient muscle contraction or excessive fatigue.

[0066] Based on the contraction frequency quantification factor and the stimulation duration quantification factor, muscle fatigue characteristics and metabolite concentration characteristics are fused to generate quantitative features that correlate contraction characteristics with physiological state. The contraction frequency quantification factor is a quantified value derived from muscle contraction frequency, reflecting the frequency of muscle contractions per unit time. In strength training, contraction frequency directly affects muscle working patterns and energy consumption. For example, a high contraction frequency may be suitable for explosive power training but can easily lead to rapid muscle fatigue, while a lower contraction frequency is more suitable for endurance training. The stimulation duration quantification factor indicates the duration of each stimulation, which determines the total duration of muscle stimulation. Different stimulation durations have different effects on muscles. Longer stimulation durations may produce a stronger contraction response, but also increase the risk of muscle fatigue. The muscle fatigue characteristic quantitatively describes the degree of muscle fatigue. During strength training, as training continues, the muscle's contraction capacity gradually decreases, which is a manifestation of muscle fatigue. Muscle fatigue characteristics can be measured in various ways, such as a decrease in muscle force and a decrease in muscle contraction speed. The metabolite concentration profile focuses on changes in the concentration of metabolites (such as lactate) produced during muscle contraction. High-intensity or prolonged muscle contractions produce a large amount of metabolites, the accumulation of which can affect normal muscle function and lead to muscle fatigue. Higher metabolite concentrations generally indicate greater muscle fatigue.

[0067] In strength training scenarios, contraction frequency, stimulation duration, muscle fatigue, and metabolite concentrations are closely linked. When contraction frequency is high and stimulation duration is long, muscles need to contract continuously and rapidly, which consumes a large amount of energy and leads to a continuous accumulation of metabolites. For example, during a high-intensity squat training session, athletes perform squats rapidly (high contraction frequency) and each set lasts for a long time (long stimulation duration). As training progresses, muscle fatigue gradually increases, manifesting as decreased muscle strength and slower movements. Simultaneously, due to the continuous muscle contraction, metabolites such as lactic acid accumulate in muscle tissue and blood, increasing their concentrations. In this state, the muscles are in a state of high fatigue, and their contraction characteristics are specifically correlated with their current physiological state. High contraction frequency and long stimulation duration lead to increased muscle fatigue and increased metabolite concentrations. These physiological changes, in turn, affect muscle contraction characteristics, such as contraction force and velocity.

[0068] In order to accurately evaluate the sustainability of muscle contraction and the impact of the current physiological state on contraction, it is necessary to integrate the contraction frequency quantification factor, stimulation duration quantification factor, muscle fatigue characteristics, and metabolite concentration characteristics through weighted summation. Assume that the weight of the contraction frequency quantification factor is , the weight of the stimulus duration quantization factor is , the weight of muscle fatigue feature is , the weight of the metabolite concentration feature is ,and For example, in a specific strength training study, setting Assume that in a strength training monitoring, the contraction frequency quantification factor is (The value range is assumed to be 0-1, 1 indicates a very high contraction frequency), the value of the stimulation duration quantization factor is , the quantified value of muscle fatigue characteristic is (1 means extreme fatigue), the quantified value of the metabolite concentration characteristic is (1 indicates extremely high concentration). The calculation formula for the quantitative feature Z associated with the contraction characteristics and physiological state is: For example, if (higher contraction frequency), (longer stimulation duration), (moderate fatigue), (medium metabolite concentration), then: .

[0069] The calculated quantitative eigenvalue associated with the contraction characteristics and physiological state (such as 0.66 in the above example) can intuitively reflect the sustainability of muscle contraction and the degree to which the current physiological state affects contraction. A standard range is typically set to assess muscle condition. For example, a standard range of 0-0.4 indicates that the muscle is in good condition; 0.4-0.7 indicates that the muscle is beginning to show signs of fatigue but still maintains some contraction capacity; and 0.7-1 indicates that the muscle is severely fatigued and its contraction capacity is significantly affected. By comparing the calculated quantitative eigenvalue with the standard range, the current state of the muscle can be determined. In the above example, 0.66 indicates that the muscle is beginning to show signs of fatigue. Although it can still contract, its condition requires attention and adjustments to the training plan or neural stimulation strategy to avoid excessive fatigue and potential injury, ensuring the safety and effectiveness of training.

[0070] Based on the target organ's movement pattern characteristics, movement phase characteristics, recovery cycle characteristics, and potential injury risk characteristics, quantitative features related to exercise intensity and contraction intensity, as well as quantitative features related to contraction characteristics and physiological state, are analyzed and processed to generate muscle contraction-nerve stimulation correlation information. This muscle contraction-nerve stimulation correlation information characterizes the relationship between muscle contraction and peripheral nerve stimulation, as well as the degree of correlation with the movement characteristics of the target organ. For strength-trained athletes, movement patterns, recovery cycles, and potential injury risks vary across different exercise phases (e.g., warm-up, training, and cool-down). During the training phase, exercise intensity is high. If the quantitative features related to exercise intensity and contraction intensity indicate insufficient contraction intensity, and the quantitative features related to contraction characteristics and physiological state indicate muscle fatigue, combined with potential injury risk (e.g., muscle strain risk due to overtraining), a comprehensive analysis is performed to generate muscle contraction-nerve stimulation correlation information. This information characterizes the relationship between muscle contraction and peripheral nerve stimulation, as well as the degree of correlation with the movement characteristics of the target organ, providing important insights for adjusting neurostimulation strategies.

[0071] Based on the correlation information between muscle contraction and nerve stimulation, the secondary adjusted muscle contraction signal is processed to generate a target peripheral nerve stimulation signal. If the correlation information indicates that the current muscle contraction intensity is insufficient and the muscle begins to fatigue, the nerve stimulation signal needs to be adjusted to improve the exercise effect and avoid injury. For example, appropriately increase the stimulation intensity, adjust the stimulation frequency, or shorten the stimulation duration to generate a target peripheral nerve stimulation signal that is more in line with the current muscle state and exercise requirements. This signal is transmitted to the peripheral nerves of the target organ through multi-channel flexible electrodes to achieve precise control of muscle contraction, enabling muscles to respond more effectively to exercise needs while ensuring the health of muscles and nerves.

[0072] The server first obtains the target user's physiological parameters, training purpose, target organ physiological characteristics, and exercise intensity interval information. Next, the server processes this physiological parameter information using the target peripheral nerve stimulation model to generate a comprehensive muscle contraction characteristic index, including contraction amplitude and frequency. Contraction amplitude is subdivided by exercise phase, and contraction frequency is differentiated by fast twitch, slow twitch, and exercise mode. Then, combining the training purpose information and the comprehensive index, a preliminary adjusted muscle contraction signal is generated. This preliminary signal is then processed based on the target organ physiological characteristics to produce a secondary adjusted signal. Finally, the secondary adjusted signal is processed based on the target organ exercise intensity interval information to generate the target peripheral nerve stimulation signal.

[0073] In this process, the acquisition of the target peripheral nerve stimulation model is crucial, and it requires multiple steps such as training data processing, model optimization training, clinical verification, and evaluation and adjustment. When generating the initial adjustment signal, it involves operations such as generating a preset signal set, extracting features, calculating the correlation matrix, and constructing a priori knowledge graph. The secondary adjustment signal includes processes such as parameter extraction, comprehensive analysis, key data processing, and generating adjustment factors. The entire method and system aims to address the shortcomings of existing neurostimulation technology, achieve precise regulation of peripheral nerves, and provide more effective support for rehabilitation training, enhanced training, etc., and has important application value.

[0074] In one embodiment, Figure 2 As shown, the present application also provides a peripheral nerve precision control device based on a multi-channel flexible electrode, comprising:

[0075] An acquisition module 201 is used to acquire physiological parameter information of a target user, training purpose information of the target user, physiological characteristic information of a target organ, and exercise intensity interval information of a target organ;

[0076] The processing module 202 is used to process the physiological parameter information of the target user based on the target peripheral nerve stimulation model to generate a comprehensive index of muscle contraction characteristics, wherein the comprehensive index of muscle contraction characteristics includes muscle contraction amplitude and contraction frequency, the muscle contraction amplitude is subdivided into corresponding target amplitudes according to different exercise stages, and the contraction frequency is clearly distinguished according to fast muscles, slow muscles and different exercise patterns; based on the target peripheral nerve stimulation model, the training purpose information and the comprehensive index of muscle contraction characteristics of the target user are processed to generate a preliminarily adjusted muscle contraction signal; based on the physiological characteristic information of the target organ, the preliminarily adjusted muscle contraction signal is processed to generate a secondary adjusted muscle contraction signal; based on the exercise intensity interval information of the target organ, the secondary adjusted muscle contraction signal is processed to generate a target peripheral nerve stimulation signal.

[0077] Each embodiment in this application is described in a related manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for evaluating the embodiment of the method for precise control of peripheral nerves based on multi-channel flexible electrodes, electronic devices, electronic devices, and readable storage media, since they are basically similar to the embodiment of the method for precise control of peripheral nerves based on multi-channel flexible electrodes described above, the description is relatively simple, and the relevant parts can be referred to the partial description of the embodiment of the method for precise control of peripheral nerves based on multi-channel flexible electrodes described above.

Claims

1. A peripheral nerve precision control device based on multi-channel flexible electrodes, characterized in that: include: An acquisition module is used to acquire physiological parameter information of a target user, training purpose information of the target user, physiological characteristic information of a target organ, and exercise intensity interval information of a target organ; a processing module, configured to process the target user's physiological parameter information based on the target peripheral nerve stimulation model to generate a comprehensive muscle contraction characteristic index, wherein the comprehensive muscle contraction characteristic index includes muscle contraction amplitude and contraction frequency. The muscle contraction amplitude is subdivided into corresponding target amplitudes according to different exercise stages, and the contraction frequency is clearly distinguished according to fast twitch, slow twitch, and different exercise modes; Based on the target peripheral nerve stimulation model, the training purpose information and comprehensive muscle contraction characteristic indicators of the target user are processed to generate a preliminary adjusted muscle contraction signal; processing the initially adjusted muscle contraction signal based on physiological characteristic information of the target organ to generate a secondary adjusted muscle contraction signal; The muscle contraction signal after secondary adjustment is processed based on the exercise intensity interval information of the target organ to generate a target peripheral nerve stimulation signal.

2. The device according to claim 1, wherein The acquisition module is configured to acquire a target peripheral nerve stimulation model, including: Obtain training data, clinical validation data, and a preset initial peripheral nerve stimulation model; Perform data cleaning, feature extraction and normalization on the training data to generate preprocessed feature data; Processing the preprocessed feature data to generate a stimulation model prediction parameter vector, wherein the stimulation model prediction parameter vector is used to represent prediction strategy information and parameter information for muscle contraction amplitude, frequency, stimulation intensity, and duration; performing an optimization training process on a preset peripheral nerve stimulation initial model based on the stimulation model prediction parameter vector to generate a trained peripheral nerve stimulation model; Performing simulated peripheral nerve stimulation on the trained peripheral nerve stimulation model based on clinical validation data to generate validation results; The trained peripheral nerve stimulation model is evaluated and adjusted based on the verification results to generate a target peripheral nerve stimulation model.

3. The device according to claim 1, wherein The processing module is configured to process the physiological parameter information of the target user based on the target peripheral nerve stimulation model to generate a comprehensive index of muscle contraction characteristics, including: Perform feature extraction on the target user's physiological parameter information to generate muscle contraction amplitude features and contraction frequency features. The muscle contraction amplitude features include the target amplitudes corresponding to different exercise stages, and the contraction frequency features are generated based on fast muscle, slow muscle, and different exercise patterns. Quantitatively analyze and process the muscle contraction amplitude characteristics to generate amplitude quantification factors; Perform quantitative analysis on the contraction frequency characteristics to generate frequency quantization factors; processing the amplitude quantization factor and the frequency quantization factor based on the target peripheral nerve stimulation model to generate contraction feature evaluation information and corresponding weight calculation result information; The contraction feature evaluation information and the corresponding weight calculation result information are processed to generate a comprehensive muscle contraction feature index, wherein the comprehensive muscle contraction feature index is used to characterize the amplitude and frequency of the target user's muscles during the contraction process.

4. The device according to claim 1, wherein The processing module is configured to process the training purpose information and the comprehensive index of muscle contraction characteristics of the target user based on the target peripheral nerve stimulation model to generate a preliminary adjusted muscle contraction signal, including: Processing a data set consisting of training purpose information and comprehensive muscle contraction characteristic indicators of a target user based on a target peripheral nerve stimulation model to generate a preset muscle contraction signal set; Processing a preset muscle contraction signal set to generate contraction intensity characteristics and frequency change characteristics as well as their occurrence frequency information under different training purposes; Processing the contraction intensity features, frequency change features and corresponding occurrence frequency information to generate feature correlation matrix information; Generate a priori knowledge graph based on feature correlation matrix information to characterize the feature correlation relationship of muscle contraction signals under different training purposes; Based on the prior knowledge graph, the training purpose information and comprehensive indicators of muscle contraction characteristics of the target user are processed to generate a preliminary adjusted muscle contraction signal.

5. The device according to claim 4, characterized in that The processing module is configured to process the training purpose information and the comprehensive index of muscle contraction characteristics of the target user based on the target peripheral nerve stimulation model to generate a preliminary adjusted muscle contraction signal, and further includes: The calculation formula for generating feature correlation matrix information is: ； in, , The values ​​of 1 and 2 represent the contraction intensity feature and frequency change feature respectively; n is the number of samples; is the kth sample value in the contraction strength feature vector x, which represents the observed value of contraction strength under a certain training purpose; is the frequency change eigenvector The k-th sample value in represents the corresponding frequency change observation value under the same training purpose; is the weight vector The kth element in reflects the frequency of occurrence of the kth sample under different training purposes. The higher the frequency, The larger it is, the more important it is used to adjust the sample's importance in calculating correlation. is the contraction strength feature vector The mean of is the frequency change eigenvector The mean of is an element in the feature correlation matrix R.

6. The device according to claim 4, characterized in that The processing module is configured to process the initially adjusted muscle contraction signal based on the physiological characteristic information of the target organ to generate a secondary adjusted muscle contraction signal, including: The target parameters of the initially adjusted muscle contraction signal are extracted and data is classified to generate stimulation intensity information, stimulation frequency information, and stimulation duration information; Based on the physiological characteristic information of the target organ and the corresponding characteristic standard version number information, the stimulation intensity information, stimulation frequency information, and stimulation duration information are comprehensively analyzed to generate parameter importance assessment information and physiological adaptability assessment information. The physiological characteristic information of the target organ includes the muscle fiber type and nerve conduction velocity of the target organ; Based on the parameter importance evaluation information and the physiological adaptability evaluation information, the key data in the initially adjusted muscle contraction signal are screened and correlated to generate a key muscle contraction data set; Perform integration operations on key muscle contraction data sets, combine the physiological characteristics of target organs with the weight information in the preset multi-knowledge base, and generate muscle contraction signal adjustment factors; The initially adjusted muscle contraction signal is processed based on the muscle contraction information adjustment factor to generate a secondary adjusted muscle contraction signal.

7. The device according to claim 1, wherein The processing module is configured to process the secondary adjusted muscle contraction signal based on the exercise intensity interval information of the target organ to generate a target peripheral nerve stimulation signal, including: Quantitative analysis is performed on the target organ's exercise intensity interval information and the secondary adjusted muscle contraction signal to generate an exercise intensity lower limit quantification factor, an exercise intensity upper limit quantification factor, a contraction intensity quantification factor, a contraction frequency quantification factor, and a stimulation duration quantification factor; Perform feature extraction and processing on the real-time physiological data of the target organ during muscle contraction to generate muscle tension features, muscle fatigue features, nerve conduction velocity features, and metabolite concentration features; Based on the lower limit quantification factor of exercise intensity, the upper limit quantification factor of exercise intensity, and the contraction intensity quantification factor, the muscle tension characteristics and nerve conduction velocity characteristics are fused to generate the quantitative characteristics associated with exercise intensity and contraction intensity; Based on the contraction frequency quantification factor and the stimulation duration quantification factor, the muscle fatigue characteristics and metabolite concentration characteristics are fused to generate the quantitative characteristics associated with the contraction characteristics and physiological state; Based on the motion pattern characteristics, motion phase characteristics, recovery cycle characteristics, and potential injury risk characteristics of the target organ, quantitative characteristics of the association between exercise intensity and contraction intensity, and quantitative characteristics of the association between contraction characteristics and physiological state are analyzed and processed to generate muscle contraction and nerve stimulation correlation information. The muscle contraction and nerve stimulation correlation information is used to characterize the relationship between muscle contraction and peripheral nerve stimulation, as well as the degree of correlation with the motion characteristics of the target organ; The secondary adjusted muscle contraction signal is processed based on the muscle contraction and nerve stimulation correlation information to generate a target peripheral nerve stimulation signal.

Images