A control system for pulsed electromagnetic fields for peripheral nerve injury repair

Abstract

This invention relates to the field of electromagnetic field control technology, specifically to a control system for pulsed electromagnetic fields used in the repair of peripheral nerve injuries. The system acquires the action potential voltage amplitude of each spatial node within the coil's coverage area, calculates the pre-stimulation conduction blockage coefficient based on the voltage distribution and time delay correlation of adjacent nodes, identifies blockage nodes, determines the principal normal direction of the blockage region through principal component analysis, and generates a theoretical output current value by combining the coil's projected area in that direction with the blockage coefficient. Subsequently, a low-frequency test pulse is applied, the post-stimulation conduction blockage coefficient is acquired, and a stimulation intensity correction coefficient is calculated. Finally, the corrected actual output intensity is used as the execution parameter for the next round of test pulses, forming a closed-loop adaptive control. This invention overcomes the problem that traditional uniform stimulation parameters cannot match the electrophysiological heterogeneity of nerve injury regions, achieving spatially precise and functionally adaptable electromagnetic control.

Description

Technical Field

[0001] This invention relates to the field of electromagnetic field control technology, and more specifically to a control system for pulsed electromagnetic fields used in the repair of peripheral nerve injuries. Background Technology

[0002] Following peripheral nerve injury, axonal regeneration and myelin formation depend on a complex electrophysiological microenvironment. The severed ends and adjacent regions exhibit complex heterogeneity in electrical activity. The core of the injury is characterized by disordered, high-amplitude abnormal discharges, while the distal ends show attenuation and delay in signal conduction. This spatial distribution difference in electrical activity determines that different regions have different response thresholds to electromagnetic stimulation and require different control modes.

[0003] Existing pulsed electromagnetic field control systems treat the damaged area as a uniformly electrically homogeneous whole in parameter settings, using a uniform low-frequency test pulse current for stimulation. However, the electrical activity states at different spatial locations within the actual damaged area vary significantly. The core area of ​​the damage exhibits disordered high-amplitude abnormal discharge, while the distal areas show signal attenuation and conduction delay. This results in the uniform low-frequency test pulse stimulation parameters failing to accurately match the actual needs of each area, and the energy distribution becoming disconnected from the spatial distribution of conduction obstacles, affecting the adaptability between stimulation parameters and neurophysiological states. Summary of the Invention

[0004] To address the technical problem that standardized low-frequency test pulse stimulation parameters in related technologies cannot accurately match the actual needs of different regions, this invention provides a control system for pulsed electromagnetic fields used in peripheral nerve injury repair. The specific technical solution adopted is as follows: This invention proposes a control system for pulsed electromagnetic fields used in the repair of peripheral nerve injuries, comprising a surface electromyography sensor array and an electromagnetic stimulation coil. The system includes: The acquisition module is used to acquire the voltage amplitude of the action potential through the surface electromyography sensor array; and to determine the pre-stimulation conduction block coefficient based on the numerical distribution and time delay correlation of the voltage amplitude of the upstream spatial node and its downstream adjacent spatial node in each surface electromyography sensor array. The theoretical strength analysis module is used to determine the blocking nodes based on the pre-stimulation conduction blockage coefficient; perform principal component analysis on all blocking node locations to determine the principal normal direction of the blocking region; and combine the projected area of ​​the coil in the principal normal direction and the pre-stimulation conduction blockage coefficient to determine the theoretical output current value. The stimulation correction module is used to apply low-frequency test pulse stimulation to the coil, obtain the post-stimulation conduction retardation coefficient of the three-dimensional region after pulse stimulation, and determine the stimulation intensity correction coefficient based on the pre-stimulation conduction retardation coefficient and the post-stimulation conduction retardation coefficient. The adjustment module is used to adjust the theoretical output current value according to the stimulation intensity correction coefficient to obtain the actual output intensity; the actual output intensity is used as the physical execution parameter for the next round of low-frequency test pulse stimulation of the electromagnetic stimulation coil.

[0005] Furthermore, based on the blind source separation algorithm and combined with the three-dimensional anatomical volume conductivity model as a priori constraint, the blind source separation is transformed into a semi-blind inverse problem. The multi-channel mixed signal collected by the surface electromyography sensor array is inversely mapped and reconstructed into action potentials of different spatial nodes in the three-dimensional space where the damaged nerve is located.

[0006] Furthermore, methods for determining the pre-stimulus conduction block coefficient include: The effective voltage value of the corresponding spatial node is determined based on the numerical distribution of the voltage amplitude of each spatial node at different sampling times. For two spatially adjacent spatial nodes, the amplitude variation coefficient is determined based on the difference in the effective voltage values ​​of the two spatial nodes; A time-series amplitude sequence is constructed based on the voltage amplitudes of the two spatial nodes at different sampling times; a time delay analysis is performed on the correlation of the time-series amplitude sequences of the two spatial nodes to determine the time delay coefficient; By combining the amplitude variation coefficient and the time delay coefficient, the pre-stimulus conduction blockage coefficient is obtained through hindrance analysis.

[0007] Furthermore, the methods for determining the time delay coefficient include: Based on cross-correlation analysis, correlation analysis is performed on the temporal amplitude sequences of two spatial nodes under different preset time offsets to determine the preset time offset that maximizes the correlation as the target offset time. The time delay coefficient is determined by comparing the target offset time with the standard time delay of a normal nerve.

[0008] Furthermore, the method for determining the blocking node includes: Spatial nodes whose pre-stimulation conduction blockage coefficient values ​​are greater than a preset coefficient threshold are designated as blockage nodes.

[0009] Furthermore, the method for determining the principal normal direction of the blocking region includes: Construct a covariance matrix based on the three-dimensional coordinates of each discrete point in the set of coordinates of the blocked region; The covariance matrix is ​​decomposed into eigenvalues ​​to obtain three mutually orthogonal eigenvectors. The eigenvector with the smallest corresponding eigenvalue is selected as the principal normal direction of the blocking region.

[0010] Furthermore, the method for determining the theoretical output current value includes: Obtain the preset safe current that a single coil is allowed to pass through; The analytical coefficients are determined by combining the pre-stimulus conduction block coefficients of all spatial nodes; Determine the projected area of ​​the coil in the principal normal direction and normalize it to obtain the spatial modulation coefficient of the coil. The theoretical output current value is obtained by adjusting the preset safety current based on the analysis coefficient and spatial modulation coefficient.

[0011] Furthermore, the preset safety current is adjusted based on the analysis coefficients and spatial modulation coefficients to obtain the theoretical output current value, including: The product of the analysis coefficient and the spatial modulation coefficient is calculated and normalized as an adjustment index; The product of the adjustment index and the preset safe current is used as the theoretical output current value.

[0012] Furthermore, the method for determining the stimulus intensity correction factor includes: By comparing the values ​​of the conduction block coefficient after stimulation with those before stimulation, the stimulation intensity correction coefficient is determined.

[0013] Furthermore, the method for determining the actual output intensity includes: The product of the stimulus intensity correction factor and the theoretical output current value is calculated as the actual output intensity.

[0014] The present invention has the following beneficial effects: This invention constructs a conduction block quantification system based on spatial nodes as basic units. Combined with a surface electromyography (EMG) sensor array, it analyzes the action potential voltage amplitude of each node within a three-dimensional region and the time delay correlation between adjacent nodes, thus describing the local electrophysiological heterogeneity of damaged nerves. Based on this, principal component analysis is used to extract the spatial principal normal direction of the blocked region, and the projected area of ​​the coil in this direction is fused with the pre-stimulation conduction block coefficient to generate a theoretical output current value with spatial directivity and functional weight, thereby breaking the limitation of traditional systems that treat the damaged region as a uniform whole in electrical characteristics. Furthermore, by applying low-frequency test pulses and comparing the changes in conduction block coefficients before and after stimulation, a stimulation intensity correction coefficient is dynamically obtained, and the actual output intensity of the next round of test pulses is adjusted accordingly, forming a closed-loop feedback mechanism based on measured electrophysiological responses. This effectively solves the problem of mismatch between uniform stimulation parameters and the spatial heterogeneity of electrical activity in the nerve injury region in existing technologies, significantly improving the spatial consistency and functional adaptability between the pulsed electromagnetic field energy distribution and the local nerve functional state. Attached Figure Description

[0015] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 This is a structural diagram of a pulsed electromagnetic field control system for peripheral nerve injury repair provided in one embodiment of the present invention. Detailed Implementation

[0017] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of a pulsed electromagnetic field control system for peripheral nerve injury repair proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0018] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0019] Furthermore, in all division and logarithmic operations involved in this invention, a protection mechanism is employed to prevent computational crashes or invalid values ​​due to a zero denominator or zero input. This protection mechanism can be implemented reasonably according to the actual situation. For example, when the denominator of a division operation or the argument of a logarithmic function is zero, a protection parameter with the same dimension as or without dimension can be added. The value of this protection parameter can be a very small value greater than zero, thereby ensuring the robustness and feasibility of the algorithm under extreme conditions. In addition, the normalization functions mentioned in this invention, unless otherwise specified, all employ maximum-minimum normalization to normalize the normalization result to the [0, 1] interval or other continuous intervals. The maximum and minimum values ​​used in the maximum-minimum normalization can be obtained according to the actual situation. For example, when multiple values ​​can be obtained in the implementation process and it is necessary to compare the size relationship between different values, multiple values ​​can be counted to obtain the maximum and minimum values. However, when only a single value can be obtained in the implementation process or it is not necessary to compare the size relationship between different values, the maximum and minimum values ​​can be obtained by counting based on a large amount of historical experimental data or prior data obtained in the early stage. When performing maximum-minimum analysis based on a large amount of historical experimental data, the maximum and minimum values ​​are used as the cutoff values ​​for analysis. When the data is greater than the historical maximum value or less than the historical minimum value, the value is truncated and determined as the corresponding cutoff value for maximum-minimum normalization calculation.

[0020] The following description, in conjunction with the accompanying drawings, details a specific scheme for a pulsed electromagnetic field control system for peripheral nerve injury repair provided by the present invention.

[0021] Please see Figure 1 The diagram illustrates a control system structure for a pulsed electromagnetic field used in the repair of peripheral nerve injuries, according to an embodiment of the present invention. The system includes: The acquisition module 101 is used to acquire the voltage amplitude of the action potential through the surface electromyography sensor array; and to determine the pre-stimulation conduction block coefficient based on the numerical distribution and time delay correlation of the voltage amplitude of the upstream spatial node and its downstream adjacent spatial node in each surface electromyography sensor array.

[0022] Following peripheral nerve injury, axonal regeneration and myelin formation depend on a complex electrophysiological microenvironment. The severed ends and adjacent regions exhibit complex heterogeneity in electrical activity. The core of the injury is characterized by disordered, high-amplitude abnormal discharges, while the distal ends show attenuation and delay in signal conduction. This spatial distribution difference in electrical activity determines that different regions have different response thresholds to electromagnetic stimulation and require different control modes.

[0023] Existing pulsed electromagnetic field control systems treat the damaged area as a uniformly electrically homogeneous whole in parameter settings, using a uniform low-frequency test pulse current for stimulation. However, the electrical activity states at different spatial locations within the actual damaged area vary significantly. The core area of ​​the damage exhibits disordered high-amplitude abnormal discharge, while the distal areas show signal attenuation and conduction delay. This results in the uniform low-frequency test pulse stimulation parameters failing to accurately match the actual needs of each area, and the energy distribution becoming disconnected from the spatial distribution of conduction obstacles, affecting the adaptability between stimulation parameters and neurophysiological states.

[0024] Based on the blind source separation algorithm, and combined with the three-dimensional anatomical volumetric conductivity model as a prior constraint, the blind source separation is transformed into a semi-blind inverse problem. The multi-channel mixed signal collected by the surface electromyography sensor array is inversely mapped and reconstructed into action potentials of different spatial nodes in the three-dimensional space where the injured nerve is located. In this process, a spatial forward transfer matrix is ​​constructed by combining the three-dimensional anatomical volumetric conductivity model of the measured limb, and the blind source separation algorithm is used to extract independent neural electrical activity sources. By solving the electromagnetic inverse problem, these sources are mapped and reconstructed into action potentials of specific coordinate nodes in the three-dimensional space.

[0025] It should be noted that by combining the three-dimensional anatomical volumetric conductivity model as a priori constraint, blind source separation is transformed into a semi-blind inverse problem that can be solved by those skilled in the art based on known electromagnetic source imaging techniques.

[0026] It should be noted that the system comprises multiple independently controlled arrays of stimulation coils distributed across a three-dimensional region, as well as an array of electromyographic sensors covering the same region. In subsequent steps, the current to each coil is independently allocated. Specifically: A 64-channel high-density electromyography (EMG) sensor patch (electrode spacing 5mm) was attached around the skin surface of the patient's limb, covering the proximal end, the projection of the injury core area, and the distal end of the nerve. Multi-channel mixed potential signals were continuously recorded for more than 5 seconds at a sampling rate of 1000Hz.

[0027] Based on MRI or CT images of the patient's limbs, medical image processing software is used to segment tissues such as skin, fat, muscle, nerve, and bone, generating a three-dimensional tetrahedral finite element mesh. Each tissue is assigned a known conductivity value (e.g., 0.35 S / m for muscle), and a spatial node to be reconstructed is defined every 1-2 mm along the nerve's course.

[0028] Using finite element simulation software (such as COMSOL), a unit current source is applied sequentially at each spatial node, and the potential value generated by the current source at each electrode position on the body surface is calculated, thus obtaining an M-row × N-column transfer matrix (M is the number of electrodes, N is the number of nodes). Each element in the matrix represents the transfer coefficient of a node to a certain electrode.

[0029] The measured multi-channel electromyography signal matrix is ​​substituted into the regularized minimum norm estimation algorithm (regularization parameter approximately 0.01) to solve the underdetermined equations, thereby obtaining the action potential estimates of each spatial node at each sampling time, forming the initial node time-series signal matrix.

[0030] Blind source separation (e.g., using independent component analysis, ICA) is performed on the initially reconstructed node time-series signal to separate multiple statistically independent time components from the mixed signal; the correlation coefficient between each component and the column vector of each node in the forward transfer matrix is ​​calculated, and the component is assigned to the spatial node with the highest correlation. The unmatched node signals are set to zero, and finally the pure action potential time-series waveform of each node is obtained.

[0031] In this embodiment of the invention, a specially designed coil can be placed on the skin surface corresponding to the injury site. For deep sciatic nerves or brachial plexus nerves, a coil with stronger focusing may be required. For nerves in the limbs, a loop coil is often placed around the limb so that the nerve passes through the center of the coil to obtain a uniform field strength distribution. Specifically, a non-invasive high-density surface electromyography (HD-sEMG) sensor array patch is used, which is attached to the limb surface in a two-dimensional flexible grid array at the proximal end, the projection position of the injury core area, and the distal end of the injured nerve, thereby obtaining different spatial nodes.

[0032] In this embodiment of the invention, the sampling frequency is between 100Hz and 1000Hz. This embodiment of the invention uses 1000Hz for sampling to obtain the action potential at different sampling times, and to determine the voltage amplitude of the action potential of different spatial nodes at different sampling times.

[0033] The consistency of voltage amplitude changes between adjacent spatial nodes can be used to analyze whether neural conduction abnormalities occur between them. In this embodiment of the invention, the pre-stimulation conduction blockade coefficient is used to characterize the blocking effect of electrical signal conduction at different spatial nodes before low-frequency test pulse stimulation. The larger the value of the pre-stimulation conduction blockade coefficient, the more significant the blocking effect on neural conduction between adjacent spatial node locations.

[0034] Furthermore, in some embodiments of the present invention, the method for determining the pre-stimulus conduction blockage coefficient includes: determining the effective voltage value of the corresponding spatial node based on the numerical distribution of the voltage amplitude of each spatial node at different sampling times; for two spatially adjacent spatial nodes, determining the amplitude variation coefficient based on the difference in the effective voltage values ​​of the two spatial nodes; constructing a time-series amplitude sequence based on the voltage amplitudes of the two spatial nodes at different sampling times; performing a time delay analysis on the correlation of the time-series amplitude sequences of the two spatial nodes to determine the time delay coefficient; and performing a blockage analysis by combining the amplitude variation coefficient and the time delay coefficient to obtain the pre-stimulus conduction blockage coefficient.

[0035] Traditional neurophysiological assessments rely heavily on macroscopic indicators (such as compound muscle action potentials, CMAP), which are insufficient to reflect the local conduction blockage of nerve fibers at the millimeter scale within the lesion area. This limits the individualized and dynamic control of pulsed electromagnetic field stimulation parameters.

[0036] In this embodiment of the invention, the effective voltage value of the corresponding spatial node is determined by the numerical distribution of the voltage amplitude of each spatial node at different sampling times. Specifically, the root mean square of the voltage amplitude of each spatial node can be calculated to realize the calculation and analysis of the effective voltage value.

[0037] Then, the difference in effective voltage values ​​(root mean square) of the action potentials of adjacent spatial nodes is calculated to construct an amplitude variation coefficient, which is used to quantify the energy attenuation of the signal during spatial propagation. The absolute value of the difference between the effective voltage values ​​of two spatial nodes can be directly calculated to achieve difference analysis, and then normalized as the amplitude variation coefficient. Specifically, the normalization process can be achieved by dividing the absolute value of the difference by a preset maximum voltage value. This preset maximum voltage value can be set to 5mV based on historical prior experience, without any restrictions.

[0038] In this embodiment of the invention, adjacent spatial nodes refer to other spatial nodes within a preset spatial range of a certain spatial node, such as other spatial nodes within a preset 5 mm range (a spherical spatial range with a radius of 5 mm).

[0039] Simultaneously, by performing cross-correlation and other time-delay analyses on the temporal amplitude sequences of the two nodes, a time-delay coefficient is extracted to characterize the abnormality of nerve conduction velocity. These two physical quantities jointly characterize the functional integrity of the local neural pathway from the two dimensions of amplitude fidelity and temporal synchronization, respectively. Optionally, in some embodiments of the present invention, the method for determining the time-delay coefficient includes: based on cross-correlation analysis, performing correlation analysis on the temporal amplitude sequences of the two spatial nodes under different preset time offsets, determining the preset time offset that maximizes the correlation as the target offset time; and determining the time-delay coefficient by comparing the target offset time with the standard time delay of normal nerves.

[0040] It is understandable that electrical signal transmission has a corresponding time delay effect through different nerve points. Under normal circumstances, the nerve transmission speed is very fast and the time delay is very small. Under abnormal conditions, the time delay will increase. Therefore, time delay analysis is used to determine the time delay coefficient to indicate the abnormality of nerve conduction speed.

[0041] Different preset time offsets represent analyses performed under different time delays. They can be set to use 1 microsecond as a preset time offset for analysis, that is, the time delay is determined to be 1 microsecond, 2 microseconds, 3 microseconds, until 100 microseconds is determined as the end point of a time delay, thus obtaining different time delay offsets. Then, the temporal amplitude sequence of the leading spatial node of neural transmission is time-delayed, and its correlation with the temporal amplitude sequence of another spatial node is calculated. When the correlation is the largest, the corresponding preset time offset is the time taken for neural transmission, which is used as the target offset time.

[0042] In this embodiment of the invention, the correlation can be calculated using the Pearson correlation coefficient, or other different correlation calculation methods such as the Spearman correlation coefficient can be used, and there is no limitation on this.

[0043] The standard time delay of a normal nerve can be set according to the actual spatial node distance. It is calculated based on the distance between adjacent spatial nodes and the standard nerve conduction velocity. For example, it can be set to 100 microseconds when the distance is 5 millimeters. The time delay coefficient is obtained by calculating the absolute value of the difference between the target offset time and the standard time delay of the normal nerve, and the ratio of the absolute value of the difference to the standard time delay. In other words, the larger the value of the time delay coefficient, the longer the conduction delay, thus realizing time delay analysis.

[0044] A larger amplitude variation coefficient indicates a higher proportion of potential energy loss, meaning a greater energy attenuation. The amplitude variation coefficient and the time delay coefficient can be combined for hindrance analysis to obtain the pre-stimulation conduction hindrance coefficient. Specifically, since a larger time delay coefficient indicates a longer conduction delay, and a larger amplitude variation coefficient indicates a higher proportion of potential energy loss, both indicate a certain degree of conduction hindrance. Therefore, the product of the amplitude variation coefficient and the time delay coefficient can be directly calculated as the pre-stimulation conduction hindrance coefficient.

[0045] The theoretical strength analysis module 102 is used to determine the blocking nodes based on the pre-stimulation conduction blockage coefficient; perform principal component analysis on all blocking node locations to determine the principal normal direction of the blocking region; and combine the projected area of ​​the coil in the principal normal direction and the pre-stimulation conduction blockage coefficient to determine the theoretical output current value.

[0046] The pre-stimulation conduction blockade coefficient characterizes the degree of significant blockage in nerve conduction between adjacent spatial nodes. Based on this, all spatial nodes can be divided to identify the blocking nodes that produce the nerve conduction blockade effect.

[0047] In this embodiment of the invention, spatial nodes whose pre-stimulation conduction blockage coefficient is greater than a preset coefficient threshold are designated as blockage nodes.

[0048] The preset coefficient threshold is the threshold value of the conduction block coefficient before stimulation. Optionally, the preset coefficient threshold can be set according to the actual conduction scenario requirements, such as using the Otsu threshold method for threshold analysis. Optionally, its value can be, for example, 0.5.

[0049] The set of adjacent blocking nodes is not random, but tends to form a blocking region with a specific spatial orientation along the nerve axon. The blocking region has a corresponding principal component direction, which is used to characterize the optimal penetration or action direction of the nerve conduction blocking effect in space.

[0050] The method for determining the principal normal direction of the hindrance region includes: constructing a covariance matrix based on the three-dimensional coordinates of each discrete point in the coordinate set of the hindrance region; performing eigenvalue decomposition on the covariance matrix to obtain three mutually orthogonal eigenvectors; and selecting the eigenvector with the smallest corresponding eigenvalue as the principal normal direction of the hindrance region.

[0051] The process of determining the principal normal direction of the blocking region involves extracting the best-fit plane from a set of discrete spatial points using Principal Component Analysis (PCA), and then using the normal vector of this plane as the principal normal direction. The specific steps are as follows: Collect the spatial coordinates of all nodes identified as blocking nodes to form a three-dimensional point set, denoted as the blocking region coordinate set. Assume this set contains N points, and the coordinates of each point are represented as follows: Where i = 1, 2, 3, ..., N; calculate the geometric center (centroid) of the point set, whose coordinates are the arithmetic mean of the coordinates of each dimension. Center the coordinates of all points relative to the centroid to obtain the mean-free coordinates. Use these centered coordinates to construct a 3×3 covariance matrix. Perform eigenvalue decomposition on the covariance matrix to obtain three real eigenvalues ​​and three mutually orthogonal unit eigenvectors. These three eigenvectors together define the principal axis direction of the data point distribution. Select the eigenvector corresponding to the smallest eigenvalue and calculate the dot product of this eigenvector and the neural anatomical direction reference vector. If the dot product is non-negative, it is directly used as the principal normal direction. If the dot product is negative, multiply the eigenvector by the negative one to determine the principal normal direction of the blockade region. The principal normal direction represents the vertical direction of the optimal fitting plane of the blockade region in three-dimensional space, providing a key spatial geometric basis for the subsequent optimization of the electromagnetic field stimulation direction.

[0052] The neuroanatomical orientation reference vector is preset based on the long axis direction of the measured limb or anatomical prior data from medical images.

[0053] The method for determining the theoretical output current value includes: obtaining the preset safe current that a single coil can pass through; determining the analysis coefficient by combining the pre-stimulation conduction retardation coefficients of all spatial nodes; determining the projected area of ​​the coil in the principal normal direction and normalizing it to obtain the spatial modulation coefficient of the coil; and adjusting the preset safe current based on the analysis coefficient and the spatial modulation coefficient to obtain the theoretical output current value.

[0054] The theoretical output current value is used to distribute the current to different coils, thereby avoiding the influence of uniform parameter stimulation on the adaptability between neurophysiological states.

[0055] The preset safe current refers to the maximum current value that a single coil can pass through for a long period of time or intermittently, under the premise of ensuring tissue safety and avoiding thermal damage or electrochemical side reactions. It is preset according to the patient's physical signs and clinical experiments in actual scenarios. Specifically, it can be, for example, 500mA.

[0056] In this embodiment of the invention, the mean of the pre-stimulus conduction blockage coefficients of all spatial nodes can be calculated, and the mean can be normalized to obtain the analytical coefficients. Specifically, the mean can be normalized to its minimum and maximum values ​​based on a preset global minimum blockage coefficient of 0 and a global maximum blockage limit value. It should be noted that the global maximum blockage limit value is a limit value obtained based on the analysis of a large amount of historical experimental data. For example, the global maximum blockage limit value can be 0.8, depending on the actual situation.

[0057] The process involves determining the projected area of ​​the coil along the principal normal direction and normalizing it to obtain the spatial modulation coefficient of the coil. The intensity of the pulsed electromagnetic field's effect on nerve tissue depends not only on the magnitude of the current flowing through the coil but, more importantly, on the effective component of the induced electric field along the normal direction of the blocked region. According to Faraday's law of electromagnetic induction, the direction of the induced electric field generated by the coil is perpendicular to its rate of change of magnetic field, and its penetration or driving capability in a specific direction is closely related to the orientation of the coil plane relative to that direction. The closer the principal axis of the coil plane is to being parallel to the principal normal direction of the blocked region, the larger the normal component of the induced electric field generated by the coil at that interface, thus more effectively depolarizing axons in a conduction-blocked state across the membrane.

[0058] Specifically, the process of normalizing the projected area to obtain the spatial modulation coefficient can be, for example, by obtaining the maximum projected area among all coils, and then recording the ratio of the projected area of ​​each coil to the maximum projected area as the spatial modulation coefficient of that coil. It should be noted that when the maximum projected area is 0, it indicates a significant problem with the coil configuration. In this case, an error needs to be reported, and the coils need to be reconfigured.

[0059] It should be noted that this application normalizes the projected area of ​​the coil in the principal normal direction. The purpose of this is to eliminate the absolute area difference between different coils caused by differences in physical size, thereby obtaining a dimensionless relative coefficient (i.e., spatial modulation coefficient) to quantify the relative level of geometric coupling efficiency of each coil to the blocking region.

[0060] The normalization methods used in this application are mostly for standardizing data with different dimensions and value ranges to facilitate subsequent data integration and analysis. Without normalization, data with different standards will affect the final output results.

[0061] Specifically, different coils may have different shapes, number of turns, and diameters, resulting in significant differences in their absolute projected areas. Normalization (e.g., dividing the projected area of ​​each coil by the largest projected area among all coils) can map the contribution potential of each coil to the same comparable interval (e.g., [0,1]). This spatial modulation coefficient is directly used for subsequent current allocation weight calculations, enabling dynamic allocation of current resources based on contribution potential: coils with high projected weights receive higher analysis coefficients (i.e., larger adjustment indices), thus undertaking more stimulation tasks; low-weight coils are moderately suppressed.

[0062] Therefore, by using the projected area of ​​the coil in the principal normal direction of the stagnation region as a weight, the geometric coupling efficiency of the coil to the stagnation region is quantified. The output current distribution of the coil should be related to its projected weight in the principal normal direction of the stagnation region. Correlating the analysis coefficients with the projected weights allows the system to dynamically allocate current resources according to contribution potential: coils with high projected weights receive higher analysis coefficients, thus undertaking more stimulation tasks within safety limits; low-weight coils are moderately suppressed.

[0063] Among them, the analysis coefficient reflects the degree of blockage obtained by combining the pre-stimulus conduction blockage coefficients of all spatial nodes. The more severe the blockage, the higher the required control intensity should theoretically be. The spatial modulation coefficient represents the geometric relationship between the coil and the blockage region. Combining the data from both dimensions, it is possible to prioritize and strengthen the control of the high-blockage region, ensuring that the allocated current can be efficiently converted into an effective induced electric field in the target direction.

[0064] Specifically, in this embodiment of the invention, the product of the analysis coefficient and the spatial modulation coefficient can be directly calculated and normalized as an adjustment index; the product of the adjustment index and the preset safety current is used as the theoretical output current value.

[0065] The adjustment index can be comprehensively analyzed from two dimensions: the overall resistance situation and the spatial state of each coil, thereby determining the required current intensity. In this embodiment of the invention, the product value of the parameters in the two dimensions is directly calculated, and the normalized value of the product value is used as the adjustment index. It should be noted that the normalization process can be, for example, maximum and minimum value normalization, where the maximum and minimum values ​​can be obtained by combining historical prior experience, and there is no limitation thereto.

[0066] The analysis coefficient characterizes the global degree of obstruction and reflects the average degree of obstruction of all spatial nodes. When a certain region is severely obstructed but the projected area of ​​the coil at that location is small, the overall analysis coefficient may still be high, thereby increasing the baseline current of all coils. For coils with small projected areas, even if the global baseline is increased, the absolute current increase is still limited. Therefore, coils with high geometric coupling efficiency are preferred for stimulation because injecting a large current into a coil that is almost perpendicular to the principal normal direction of the obstructed region results in a very low effective component of the induced electric field in that direction. Most of the energy is wasted and may even stimulate non-target regions.

[0067] It should be noted that if strong stimulation is required in clinical practice for areas that can only be covered by small projection area coils, the system can compensate through a closed-loop feedback mechanism (i.e., stimulation intensity correction coefficient): if the conduction block coefficient of the area does not decrease significantly after the first round of stimulation, the system will increase the stimulation intensity correction coefficient (e.g., from 1.0 to 1.2), thereby increasing the theoretical output current value as a whole, and indirectly increasing the absolute current of the small projection area coil.

[0068] Therefore, this application aims to achieve adaptive control through a dual mechanism of prioritizing the use of high-efficiency coils and compensating for inefficient coils with global strength enhancement. This design ensures effectiveness while also considering safety and energy efficiency.

[0069] The theoretical output current value represents the optimal excitation current that should be applied to a single coil to achieve effective and precise electromagnetic nerve modulation, while taking into account safety, the degree of nerve function damage, and the spatial geometry of the coil.

[0070] The stimulation correction module 103 is used to apply low-frequency test pulse stimulation to the coil, obtain the post-stimulation conduction retardation coefficient of the three-dimensional region after pulse stimulation, and determine the stimulation intensity correction coefficient based on the pre-stimulation conduction retardation coefficient and the post-stimulation conduction retardation coefficient.

[0071] In electromagnetic neuromodulation based on multi-coil arrays, it is necessary to dynamically evaluate and precisely correct the stimulation intensity to match individualized neural functional states. The actual response of neural tissue to electromagnetic fields is significantly uncertain due to the influence of local microenvironment, damage heterogeneity, and individual physiological differences. Therefore, relying solely on the pre-stimulation conduction block coefficient is insufficient to guarantee the stability and reliability of the stimulation effect. Intensity correction requires combining tissue analysis after low-frequency test pulse stimulation.

[0072] That is, by applying low-frequency test pulses within a safe range to the coil, the conduction blockage coefficient after stimulation is determined based on the conduction blockage state of each node in the three-dimensional region after stimulation; combined with the known conduction blockage coefficient before stimulation, the difference between the two is used to quantify the actual improvement of nerve function by the current stimulation, and then the stimulation intensity correction coefficient is calculated.

[0073] Understandably, the specific method for obtaining the post-stimulation conduction block coefficient is similar to that for the pre-stimulation conduction block coefficient, and will not be elaborated further. The post-stimulation conduction block coefficient represents the degree of conduction block in neural tissue after stimulation.

[0074] In this embodiment of the invention, the stimulation intensity correction coefficient is determined by comparing the values ​​of the conduction block coefficient after stimulation with the conduction block coefficient before stimulation.

[0075] In some embodiments of the present invention, the difference between the post-stimulation conduction block coefficient and the pre-stimulation conduction block coefficient can also be calculated, and the ratio of this difference to the pre-stimulation conduction block coefficient can be calculated to achieve normalization processing, representing the stimulus intensity correction coefficient. The larger the value, the more obvious the post-stimulation block, and the more intense the stimulus needs to be. When the pre-stimulation conduction block coefficient is 0, it means that no block is generated, and the stimulus intensity correction coefficient is directly set to the minimum value of 0. According to the actual scenario requirements, the stimulus intensity correction coefficient is determined by combining the post-stimulation conduction block coefficient and the pre-stimulation conduction block coefficient.

[0076] The adjustment module 104 is used to adjust the theoretical output current value according to the stimulation intensity correction coefficient to obtain the actual output intensity; the actual output intensity is used as the physical execution parameter for the next round of low-frequency test pulse stimulation of the electromagnetic stimulation coil.

[0077] In electromagnetic neuromodulation, although the theoretical output current value has taken into account safety boundaries, functional requirements, and spatial geometric factors, it is still essentially a predicted value and cannot fully reflect the actual dynamic response of neural tissue to stimulation. If it is directly used for continuous control, insufficient or excessive stimulation may occur due to individual physiological differences or changes in the local microenvironment, thus weakening the accuracy of control.

[0078] In this embodiment of the invention, the theoretical output current value is numerically corrected based on the stimulation intensity correction coefficient obtained in the previous step to generate the actual output intensity, which is then used as the physical execution parameter for each coil in the next round of low-frequency test pulse stimulation.

[0079] Preferably, in this embodiment of the invention, the product of the stimulus intensity correction coefficient and the theoretical output current value is calculated as the actual output intensity. That is, the actual output intensity is obtained by adjusting the ratio based on the product. Of course, in other embodiments of the invention, a correction function, such as a two-dimensional linear function, can also be set, with the theoretical output current value as the independent variable and the slope as the stimulus intensity correction coefficient. The function is fitted using a large amount of historical data to determine the intercept, and the actual output intensity is output. The actual output intensity is the current value after numerical correction.

[0080] On the other hand, to ensure the reliability of the overall analysis, the stimulus intensity correction coefficient can be linearly mapped to a preset correction limit range (e.g., [0.8, 1.2], set according to actual adjustment needs). For example, the maximum and minimum values ​​of the stimulus intensity correction coefficient can be determined, the maximum value mapped to 1.2, and the minimum value mapped to 0.8. Calculations and analysis are then performed to determine the linear mapping function, thus realizing the linear mapping function. No restrictions are placed on this. The product of the mapped correction coefficient and the theoretical output current value is taken as the actual output intensity.

[0081] Of course, in other embodiments of the present invention, a big data model (such as a convolutional neural network) can also be trained through historical prior experience to obtain a corrected model, and the actual output intensity can be determined based on the corrected model. The specific model training and result output are well known in the art and will not be described in detail here.

[0082] In this embodiment of the invention, the actual output intensity is used as the current value of each coil in the next round of low-frequency test pulse stimulation to achieve effective low-frequency test pulse stimulation. It should be noted that if it is the first round of low-frequency test pulse stimulation, a preset initial stimulation current value, such as 300 mA, can be used based on prior experience, and there is no limitation on this.

[0083] This invention constructs a conduction block quantification system based on spatial nodes as basic units. Combined with a surface electromyography (EMG) sensor array, it analyzes the action potential voltage amplitude of each node within a three-dimensional region and the time delay correlation between adjacent nodes, thus describing the local electrophysiological heterogeneity of damaged nerves. Based on this, principal component analysis is used to extract the spatial principal normal direction of the blocked region, and the projected area of ​​the coil in this direction is fused with the pre-stimulation conduction block coefficient to generate a theoretical output current value with spatial directivity and functional weight, thereby breaking the limitation of traditional systems that treat the damaged region as a uniform whole in electrical characteristics. Furthermore, by applying low-frequency test pulses and comparing the changes in conduction block coefficients before and after stimulation, a stimulation intensity correction coefficient is dynamically obtained, and the actual output intensity of the next round of test pulses is adjusted accordingly, forming a closed-loop feedback mechanism based on measured electrophysiological responses. This effectively solves the problem of mismatch between uniform stimulation parameters and the spatial heterogeneity of electrical activity in the nerve injury region in existing technologies, significantly improving the spatial consistency and functional adaptability between the pulsed electromagnetic field energy distribution and the local nerve functional state.

[0084] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0085] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

Claims

1. A control system for pulsed electromagnetic fields for peripheral nerve injury repair, characterized by, The system comprises a surface electromyography sensor array and an electromagnetic stimulation coil, and includes: The acquisition module is used to acquire the voltage amplitude of the action potential through the surface electromyography sensor array; and to determine the pre-stimulation conduction block coefficient based on the numerical distribution and time delay correlation of the voltage amplitude of the upstream spatial node and its downstream adjacent spatial node in each surface electromyography sensor array. The theoretical strength analysis module is used to determine the blocking nodes based on the pre-stimulation conduction blockage coefficient; perform principal component analysis on all blocking node locations to determine the principal normal direction of the blocking region; and combine the projected area of ​​the coil in the principal normal direction and the pre-stimulation conduction blockage coefficient to determine the theoretical output current value. The stimulation correction module is used to apply low-frequency test pulse stimulation to the coil, obtain the post-stimulation conduction retardation coefficient of the three-dimensional region after pulse stimulation, and determine the stimulation intensity correction coefficient based on the pre-stimulation conduction retardation coefficient and the post-stimulation conduction retardation coefficient. The adjustment module is used to adjust the theoretical output current value according to the stimulation intensity correction coefficient to obtain the actual output intensity; the actual output intensity is used as the physical execution parameter for the next round of low-frequency test pulse stimulation of the electromagnetic stimulation coil.

2. A control system for pulsed electromagnetic fields for peripheral nerve injury repair as claimed in claim 1, wherein, Based on the blind source separation algorithm, combined with the three-dimensional anatomical volume conductivity model as a prior constraint, the blind source separation is transformed into a semi-blind inverse problem. The multi-channel mixed signal collected by the surface electromyography sensor array is inversely mapped and reconstructed into the action potentials of different spatial nodes in the three-dimensional space where the damaged nerve is located.

3. A control system for pulsed electromagnetic fields for peripheral nerve injury repair as claimed in claim 1, wherein, Methods for determining the pre-stimulus conduction block coefficient include: The effective voltage value of the corresponding spatial node is determined based on the numerical distribution of the voltage amplitude of each spatial node at different sampling times. For two spatially adjacent spatial nodes, the amplitude variation coefficient is determined based on the difference in the effective voltage values ​​of the two spatial nodes; A time-series amplitude sequence is constructed based on the voltage amplitudes of the two spatial nodes at different sampling times; a time delay analysis is performed on the correlation of the time-series amplitude sequences of the two spatial nodes to determine the time delay coefficient; By combining the amplitude variation coefficient and the time delay coefficient, the pre-stimulus conduction blockage coefficient is obtained through hindrance analysis.

4. A control system for pulsed electromagnetic fields for peripheral nerve injury repair as claimed in claim 3, wherein, Methods for determining the time delay factor include: Based on cross-correlation analysis, correlation analysis is performed on the temporal amplitude sequences of two spatial nodes under different preset time offsets to determine the preset time offset that maximizes the correlation as the target offset time. The time delay coefficient is determined by comparing the target offset time with the standard time delay of a normal nerve.

5. A control system for pulsed electromagnetic fields for peripheral nerve injury repair as claimed in claim 1, wherein, Methods for determining blocking nodes include: Spatial nodes whose pre-stimulation conduction blockage coefficient values ​​are greater than a preset coefficient threshold are designated as blockage nodes.

6. A control system for pulsed electromagnetic fields for peripheral nerve injury repair as claimed in claim 1, wherein, Methods for determining the principal normal direction of the blocking region include: Construct a covariance matrix based on the three-dimensional coordinates of each discrete point in the set of coordinates of the blocked region; The covariance matrix is ​​decomposed into eigenvalues ​​to obtain three mutually orthogonal eigenvectors. The eigenvector with the smallest corresponding eigenvalue is selected as the principal normal direction of the blocking region.

7. A control system for pulsed electromagnetic fields for peripheral nerve injury repair as claimed in claim 1, wherein, Methods for determining the theoretical output current value include: Obtain the preset safe current that a single coil is allowed to pass through; The analytical coefficients are determined by combining the pre-stimulus conduction block coefficients of all spatial nodes; Determine the projected area of ​​the coil in the principal normal direction and normalize it to obtain the spatial modulation coefficient of the coil. The theoretical output current value is obtained by adjusting the preset safety current based on the analysis coefficient and spatial modulation coefficient.

8. A control system for pulsed electromagnetic fields for peripheral nerve injury repair as claimed in claim 7, wherein, The preset safety current is adjusted based on the analysis coefficients and spatial modulation coefficients to obtain the theoretical output current value, including: The product of the analysis coefficient and the spatial modulation coefficient is calculated and normalized as an adjustment index; The product of the adjustment index and the preset safe current is used as the theoretical output current value.

9. A control system for pulsed electromagnetic fields for peripheral nerve injury repair as defined in claim 1, wherein, Methods for determining the stimulus intensity correction factor include: By comparing the values ​​of the conduction block coefficient after stimulation with those before stimulation, the stimulation intensity correction coefficient is determined.

10. A control system for a pulsed electromagnetic field for peripheral nerve injury repair as described in claim 1, characterized in that, Methods for determining the actual output intensity include: The product of the stimulus intensity correction factor and the theoretical output current value is calculated as the actual output intensity.

Images