Neural electrical stimulation parameter optimization method based on modified intensity-time model and neural stimulation system

Abstract

The application discloses a nerve electric stimulation parameter optimization method and a nerve stimulation system based on a modified strength-time model, and the method comprises the following steps: performing prior tests on a target nerve, including: applying electric stimulation with a plurality of preset pulse widths, and collecting minimum domain stimulation intensities of each pulse width t i I i ; fitting a modified strength-time model by using a plurality of pairs of data t i , I i , and solving individualized parameters I rh , Q th 、 c :, I rh is a base strength, which is defined as the minimum stimulation current intensity required to cause nerve excitation when the pulse width tends to infinity; Q th is a threshold charge constant related to the membrane capacitance characteristic, c is a modified time constant; the base strength between 1.85 and 2.15 times is taken as the threshold stimulation current intensity optimization value, and the pulse width of a single electric stimulation required is calculated as the pulse width optimization value.​

Description

Technical Field

[0001] This invention relates to the field of biomedicine, and more particularly to a method and system for optimizing neural electrical stimulation parameters based on a modified intensity-time model, applicable to the optimization of operating parameters of neurally modulated medical electronic devices. Background Technology

[0002] Electrical nerve stimulation (including spinal cord stimulation, deep brain stimulation, vagus nerve stimulation, etc.) has been proven to be an effective treatment for chronic pain, movement disorders (such as Parkinson's disease), epilepsy, and other refractory diseases. The selection of stimulation parameters, especially the proper matching of current intensity and pulse width, directly affects the therapeutic effect, energy consumption, and patient tolerance and comfort.

[0003] Currently, the classic model describing the relationship between the intensity of a nerve threshold stimulus and the duration (t) is mainly the exponential model: in, I The intensity of nerve electrical stimulation, Let t be the base intensity, t be the pulse duration, and τ be the membrane time constant. This exponential model is based on the RC circuit theory of neuronal membrane capacitance charging, and its physical meaning is clear.

[0004] While the exponential model has high fitting accuracy under short-pulse stimulation, its parameter fitting requires nonlinear regression, which is computationally complex. Furthermore, its parameter τ does not directly correspond to the concepts of "base intensity" and "time value" familiar to clinicians. Engineering implementation requires additional exponential calculations, which places a significant burden on the processor and battery of low-power devices.

[0005] Most existing neurostimulators use fixed parameters or simple linear mapping, lacking the ability to adaptively optimize based on individualized threshold curves for each patient. This results in the inability to automatically adjust parameters after changes in the electrode-tissue interface, micro-displacement, or tissue response. Long-term use can easily lead to understimulation (resulting in decreased efficacy) or overstimulation (resulting in discomfort or side effects).

[0006] The disclosure of the above background technical content is only for the purpose of assisting in understanding the concept and technical solution of this application, and does not necessarily provide technical instruction. Summary of the Invention

[0007] The purpose of this invention is to provide a method based on the Weiss model by introducing a modified time constant. c A modified intensity-time model for optimizing neural electrical stimulation parameters, especially for extremely short pulse widths of less than 100 μs, predicts current parameters that are closer to the true threshold.

[0008] To achieve the above objectives, the technical solution adopted by the present invention is as follows: A method for optimizing neural electrical stimulation parameters based on a modified intensity-time model includes the following steps: The target nerve was subjected to the following prior tests: applying electrical stimulation with multiple preset pulse widths and acquiring the minimum domain stimulation intensity at each pulse width. Using multiple pairs ( t i , I i The data was fitted to the following modified intensity-time model to obtain individualized parameters. I rh , Q th , c : in, I The threshold stimulation current intensity required to induce excitation, t The pulse width of a single electrical stimulation. I rh The base intensity is defined as the minimum stimulation current intensity required to induce nerve excitation when the pulse width approaches infinity. Q th The threshold charge constant is related to the membrane capacitance characteristics. c To correct the time constant; t i This indicates the first action applied to the target nerve during the prior test. i The pulse width of the secondary pulse stimulation. I i For the collected and t i The corresponding minimum domain stimulus intensity; by k The basalt intensity is used as the optimized threshold stimulation current intensity, and the corresponding required pulse width for a single electrical stimulation is calculated as the optimized pulse width, where 1.85 ≤ k ≤2.15.

[0009] Furthermore, following any one or a combination of the aforementioned technical solutions, the corresponding conduction velocity is also measured in the prior test; Based on the aforementioned conduction velocity, select k The numerical value: the faster the conduction speed, the... k The larger the value, the slower the conduction speed. k The smaller the value, the better.

[0010] Furthermore, based on any one or a combination of the aforementioned technical solutions, the fiber type of the target nerve is determined to be Aβ, Aδ, or C type according to the conduction velocity; If the fiber type of the target nerve is Aβ, then k The value ranges from 2 to 2.15; If the fiber type of the target nerve is Aδ, then k The value ranges from 1.9 to 2; If the fiber type of the target nerve is type C, then k The values ​​range from 1.85 to 1.92.

[0011] Furthermore, in accordance with any or a combination of the aforementioned technical solutions, the method for optimizing neural electrical stimulation parameters provided by the present invention further includes: The target nerve is stimulated with pulsed current intensity and pulse width optimization values, and nerve response signals are collected simultaneously. Based on the amplitude of the neural response signal, the stimulation current intensity and pulse width of the output pulse stimulation signal are adjusted.

[0012] Furthermore, based on any one or a combination of the aforementioned technical solutions, if the amplitude of the acquired neural response signal is lower than a preset first amplitude threshold, the stimulation current intensity and / or the pulse width are increased based on the current output pulse stimulation signal. If the amplitude of the acquired neural response signal is higher than the preset second amplitude threshold, and the second amplitude threshold is greater than or equal to the first amplitude threshold, then the stimulation current intensity and / or pulse width are reduced based on the current output pulse stimulation signal.

[0013] Furthermore, following any one or a combination of the aforementioned technical solutions, the neural electrical activity induced by stimulation is collected in real time as a neural response signal, wherein the neural electrical activity includes one or more of the following: compound action potential, local field potential, and patient perception threshold marker from external input; When it is determined that the output needs to be adjusted based on the neural response signal, the step size for adjustment is set to be between the base strength obtained from the fitting. I rh The step size is set to be between 1% and 6% of the base strength obtained from the fitting. I rh 5% to 10%.

[0014] Furthermore, following any one or a combination of the aforementioned technical solutions, the prior test includes at least one electrical stimulation with a pulse width less than or equal to 100 μs and one electrical stimulation with a pulse width greater than or equal to 1000 μs, and the minimum domain stimulation intensity corresponding to each pulse width is measured using an electrophysiological recording system.

[0015] Furthermore, following any one or a combination of the aforementioned technical solutions, the modified intensity-time model is fitted using the nonlinear least squares method to obtain the base intensity. I rh Threshold charge constant Q th and correction time constant c ; The following evaluation function is used to check the goodness of fit: in, R 2 For goodness of fit, I pred_i To minimize the stimulus intensity I i Corresponding pulse width t i Substitute the threshold stimulus current intensity calculated from the fitted modified intensity-time model, For all minimum domain stimulus in the prior test I i The average value; If the goodness of fit R 2 If the preset evaluation index is met, the fitting is considered successful; otherwise, the modified intensity-time model is refitted.

[0016] Furthermore, following any one or a combination of the aforementioned technical solutions, the error function between the predicted value of the modified intensity-time model and the measured value of the minimum domain stimulus intensity is defined as follows: 2 ,in, n The number of pulse stimuli applied in the prior test; Minimization Error To obtain the optimal parameters I rh , Q th , c .

[0017] Furthermore, in accordance with any or a combination of the aforementioned technical solutions, the method for optimizing neural electrical stimulation parameters provided by the present invention further includes: If the baseline strength obtained by fitting I rh If the safety range is exceeded, a warning message will be issued.

[0018] Furthermore, in accordance with any or a combination of the aforementioned technical solutions, the method for optimizing neural electrical stimulation parameters provided by the present invention further includes: Using multiple pairs ( t i , I i The data was fitted to the following classic intensity-time model to obtain individualized parameters. I’ rh , Q’ th : in, I’ rh The basic intensity in the classical intensity-time model, Q’ th This represents the threshold charge constant in the classical intensity-time model. If the pulse width calculated using the modified intensity-time model exceeds a preset pulse width threshold, then the classical intensity-time model is used instead of the modified intensity-time model to determine the pulse width. k × I’ rh As the optimized value of the threshold stimulation current intensity, or, with k × I’ rh and k × I rh The weighted summation result is used as the optimized value of the threshold stimulation current intensity, wherein the pulse width threshold is the definition value of the ultrashort pulse, and the pulse width threshold is between 0 and 200 μs.

[0019] According to another aspect of the present invention, a nerve stimulation system is provided, comprising: An electrical stimulation pulse generation module, which is configured to output programmable current pulses; A control module configured to control the output pulse width, amplitude, and frequency of the electrical stimulation pulse generation module; The neural stimulation system uses the optimization method described above to determine the neural electrical stimulation parameters, including: Under the control of the control module, the electrical stimulation pulse generation module performs a priori testing on the target nerve; The control module fits the modified intensity-time model and calculates... k The base intensity corresponds to the required pulse width.

[0020] Furthermore, in accordance with any or a combination of the aforementioned technical solutions, the system further includes a neural response signal acquisition module, which is configured to detect the response amplitude of neural electrical activity induced by stimulation and send the detection result to the control module; The control module is configured to adjust the output pulse width and amplitude of the electrical stimulation pulse generation module according to the response amplitude, so that the real-time response amplitude tends to the target amplitude range.

[0021] Furthermore, following any one or a combination of the aforementioned technical solutions, the system further includes a monitoring and alarm module, which is configured to issue a warning message in response to the base intensity obtained by fitting the modified intensity-time model exceeding a preset safety range.

[0022] Furthermore, following any one or more of the aforementioned technical solutions, the system is integrated into one or more of the following devices: an implantable pulse generator, a transcutaneous electrical nerve stimulation device, a deep brain stimulator, and a spinal cord stimulator.

[0023] The beneficial effects of the technical solution provided by this invention are as follows: a. A novel threshold model is proposed that balances fitting accuracy and computational simplicity, can be integrated with existing clinical parameter systems, and supports closed-loop adaptive regulation to optimize stimulation parameters; b. This invention provides the introduction of a corrected time constant. c The corrected model, especially when the pulse width is extremely short (e.g., <100μs), yields predicted current parameters that are closer to the true threshold, ensuring the effectiveness of electrical stimulation and avoiding energy waste or overstimulation risks caused by model errors. c. Retain clinically familiar parameters (pulse width and threshold stimulation current intensity), be compatible with electrodiagnostic indicators familiar to clinicians, facilitate direct integration with existing clinical workflows, and reduce the threshold for technology conversion; d. By monitoring neural responses in real time and dynamically adjusting stimulation intensity, the stimulation effect of the closed-loop system of this invention can be significantly improved compared with traditional open-loop fixed-parameter stimulation. It can also automatically compensate for threshold drift caused by electrode micro-displacement, tissue reaction or physiological state changes, avoiding long-term understimulation (decreased efficacy) or overstimulation (discomfort, side effects). e. The model contains only three parameters, and the fitting uses a nonlinear least squares method, which can be completed in seconds on a low-power microcontroller; compared with the exponential model, this model does not involve exponential function operations, significantly reducing memory usage and computational overhead, which is beneficial to extending the battery life of implantable devices; f. A dual-model adaptive mechanism combining the modified intensity-time model and the classical intensity-time model is introduced, which retains the advantages of the modified model in short pulses while ensuring the rationality of the electrical stimulation parameters under long pulses. Attached Figure Description

[0024] To more clearly illustrate the technical solutions in the embodiments of this application 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 recorded in this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0025] Figure 1 A flowchart illustrating a method for optimizing neural electrical stimulation parameters based on a modified intensity-time model, provided as an exemplary embodiment of the present invention; Figure 2 A block diagram of a nerve stimulation system provided as an exemplary embodiment of the present invention; Figure 3 A flowchart illustrating a dual-model adaptive mechanism provided as an exemplary embodiment of the present invention; Figure 4 Intensity-time plots of the modified model and the Weiss model provided for the first numerical embodiment of the present invention; Figure 5 The intensity-time curves of the modified model and the Weiss model provided for the second numerical embodiment of the present invention. Detailed Implementation

[0026] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0027] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, apparatus, product, or device that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or devices.

[0028] Classical strength-time models, such as the Weiss formula: ,in, IThe threshold stimulation current intensity required to induce excitation, t The pulse width is the duration of a single electrical stimulation. I’ rh The basic intensity in the classical intensity-time model, Q’ th This represents the threshold charge constant in the classical intensity-time model. In the classical Weiss formula, when the pulse width... t When the threshold stimulation current intensity approaches zero, it tends to be infinitely large. However, in real physical environments, the current is actually limited by factors such as the capacitance effect of the electrode-tissue interface, the geometric decay of the axial current distribution, or the kinetic limit of the ion channel, and will not tend to be infinite.

[0029] This invention aims to provide a modified intensity-time model that eliminates the singularity of the classical Weiss formula and solves the numerical instability problems that arise during curve fitting or computer simulation. Furthermore, for certain high-frequency electrical stimulations (such as the pulse leader phase of deep brain stimulation (DBS) or transcranial magnetic stimulation (TMS), the measured intensity-time curves often exhibit a "plateauing" trend in the extremely short pulse region. The traditional classical Weiss formula significantly overestimates the threshold current in this region, while the modified intensity-time model intends to better fit this asymptotic behavior.

[0030] In one embodiment of the present invention, a method for optimizing neural electrical stimulation parameters based on a modified intensity-time model is provided, such as... Figure 1 The method for optimizing neural electrical stimulation parameters includes the following steps: S100: Perform the following prior tests on the target nerve, including: applying multiple preset pulse widths. t i Electrical stimulation was applied, and the minimum domain stimulation intensity was acquired for each pulse width. I i ; The prior test includes at least one electrical stimulation with a pulse width of less than or equal to 100 μs and one electrical stimulation with a pulse width of greater than or equal to 1000 μs, and the minimum domain stimulation intensity corresponding to each pulse width is measured using an electrophysiological recording system.

[0031] S200: Utilizing multiple pairs ( t i , I i The data was fitted to the following modified intensity-time model to obtain individualized parameters. I rh , Q th , c : in, I The threshold stimulation current intensity required to induce excitation, t The pulse width of a single electrical stimulation. I rh The base intensity is defined as the minimum stimulation current intensity required to induce nerve excitation when the pulse width approaches infinity. Q th The threshold charge constant is related to the membrane capacitance characteristics. c To correct the time constant; t i This indicates the first action applied to the target nerve during the prior test. i The pulse width of the secondary pulse stimulation. I i For the collected and t i The corresponding minimum domain stimulus intensity; I rh , Q th , c It is called an individualized parameter because even if the stimulated neural sites are the same, the parameters will differ for different individuals. t i , I i The data may vary, which will affect the parameters obtained from the fit. I rh , Q th , c The value will also be different.

[0032] Specifically, the modified intensity-time model is fitted using the nonlinear least squares method to obtain the base intensity. I rh Threshold charge constant Q th and correction time constant c In one embodiment of the present invention, the error function between the predicted value of the modified intensity-time model and the measured value of the minimum domain stimulus intensity is defined as: 2 ,in, n The number of pulse stimuli applied in the prior test; Minimization Error To obtain the optimal parameters I rh , Q th , c .

[0033] The following evaluation function is used to check the goodness of fit: in, R 2 For goodness of fit, I pred_i To minimize the stimulus intensity I i Corresponding pulse width t i Substitute the threshold stimulus current intensity calculated from the fitted modified intensity-time model, For all minimum domain stimulus in the prior test I i The average value; If the goodness of fit R 2 If the preset evaluation index is met, the fitting is considered successful; otherwise, the modified intensity-time model is refitted.

[0034] S300: k The basalt intensity is used as the optimized threshold stimulation current intensity, and the corresponding required pulse width for a single electrical stimulation is calculated as the optimized pulse width, where 1.85 ≤ k ≤2.15.

[0035] In this embodiment, the constant term introduced by the intensity-time model is corrected. c This provides a non-zero lower limit plateau for the current response, especially with extremely short pulse widths (e.g., <100 μs). Because the lipid bilayer structure of the cell membrane acts as a capacitor, the charge injection rate is limited by the capacitor charging time. At this point, even with a significant increase in current intensity, it is difficult to further shorten the excitation time. (Adding a constant term...) c This is equivalent to introducing an effective pulse width offset.

[0036] In one embodiment of the present invention, k This is a value randomly selected within the interval [1.85, 2.15]; in different embodiments, the selected value is... k The value is affected by the nerve diameter; the membrane time constant τ of a thicker nerve is larger than that of a thinner nerve, requiring a larger [value]. k Only values ​​that can measure shorter characteristic times are obtained. Based on this, in this embodiment, the corresponding conduction velocity is also measured in the prior test; according to the conduction velocity, a selection is made. k The numerical value: the faster the conduction speed, the... k The larger the value, the slower the conduction speed. k The smaller the value, the better.

[0037] In one specific embodiment, the fiber type of the target nerve is determined to be Aβ, Aδ, or C based on the conduction velocity. For example, the conduction velocity of Aβ type nerves is typically 30 to 70 m / s, the conduction velocity of Aδ type nerves is typically 5 to 30 m / s, and the conduction velocity of C type nerves is typically 0.5 to 2 m / s. If the fiber type of the target nerve is Aβ, then k The value ranges from 2 to 2.15; If the fiber type of the target nerve is Aδ, then k The value ranges from 1.9 to 2; If the fiber type of the target nerve is type C, then k The values ​​range from 1.85 to 1.92.

[0038] In one embodiment of the present invention, before step S300, the fitted basis strength is first... I rh Safety assessment: If the value exceeds the preset safety range, a warning message is issued, such as when the stimulator's operating state deviates from expectations (e.g., the required intensity abnormally increases beyond the safety limit), reminding clinicians to check the electrode position or system status. If the condition is deemed safe, after step S300, a pulse stimulation signal is output to the target nerve using the optimized threshold stimulation current intensity and pulse width values, and adaptive adjustments can be made based on the output pulse stimulation signal. S400: Acquires neural response signals while outputting pulse stimulation signals; Specifically, the neural response signal acquisition module is used to acquire the neural electrical activity induced by stimulation in real time as a neural response signal. The neural electrical activity includes one or more of the following: compound action potential, local field potential, and patient perception threshold marker from external input. S500: Adjust the stimulation current intensity and pulse width of the output pulse stimulation signal according to the amplitude of the neural response signal.

[0039] In one specific embodiment, if the amplitude of the acquired neural response signal is lower than a preset first amplitude threshold, the stimulation current intensity and / or pulse width are increased based on the current output pulse stimulation signal; the possible upward adjustment step size is between the base intensity obtained in step S200. I rh 1% to 6%.

[0040] If the amplitude of the acquired neural response signal is higher than a preset second amplitude threshold, and the second amplitude threshold is greater than or equal to the first amplitude threshold, then the stimulation current intensity and / or pulse width are reduced based on the current output pulse stimulation signal. The possible downsizing step size is between the base intensity obtained in step S200. I rh 5% to 10%.

[0041] In an advanced embodiment of the present invention, a dual-model adaptive mechanism is provided, such as... Figure 3 As shown, the modified intensity-time model and the classical intensity-time model are used as alternative dual models, and an empirical pulse width threshold is used to achieve the splitting of the modified model and the classical model: Using multiple pairs ( t i , I i The data was fitted to the following classic intensity-time model to obtain individualized parameters. I’ rh , Q’ th : in, I’ rh The basic intensity in the classical intensity-time model, Q’ th This represents the threshold charge constant in the classical intensity-time model. As a defining value for ultrashort pulses, the empirical pulse width threshold ranges from 0 to 200 μs. In this embodiment, the pulse width threshold is set to 100 μs. In an optional embodiment, if the pulse width calculated using the modified intensity-time model is ≥100μs, then the classical intensity-time model is used instead of the modified intensity-time model to determine the pulse width. k × I’ rh As the optimized value for the threshold stimulation current intensity; In another optional embodiment, if the pulse width calculated using the modified intensity-time model is ≥100μs, then... k × I’ rh and k × I rh The weighted summation result is used as the optimized value of the threshold stimulus current intensity, for example, α× k × I rh +β× k × I’ rh, where α and β are weighting coefficients, and the sum of the two is 1.

[0042] Short pulse stimulation of less than 100 μs has a significant impact on the nonlinear properties of the nerve membrane (such as the delay in activation gating), including the correction time constant. c It can better describe this behavior that deviates from the ideal Ohm's law; under short pulse stimulation of 100 μs or more, the nerve membrane tends to be in a steady state. The classical model (Weiss's law) is accurate enough and simpler to calculate. Therefore, it can be used to replace it, or its calculation results can be referenced for weighted summation. This can avoid overfitting and avoid the situation where numerical instability will occur in the long pulse region due to the forced use of the modified model, thus ensuring the universality of the output parameters.

[0043] In one embodiment of the present invention, a neural stimulation system is provided, including, as shown in the figure below. Figure 2 The module shown: An electrical stimulation pulse generation module, which is configured to output programmable current pulses; A control module (i.e., a programmer) is configured to control the output pulse width, amplitude, and frequency of the electrical stimulation pulse generation module; The neural stimulation system uses the optimization method described above to determine the neural electrical stimulation parameters, including: Under the control of the control module, the electrical stimulation pulse generation module performs a priori testing on the target nerve; The control module fits the modified intensity-time model and calculates... k The base intensity corresponds to the required pulse width.

[0044] See also Figure 2 The system further includes a neural response signal acquisition module and a monitoring and alarm module. The neural response signal acquisition module is configured to detect the response amplitude of neural electrical activity induced by stimulation and send the detection result to the control module. The control module is configured to adjust the output pulse width and amplitude of the electrical stimulation pulse generation module according to the response amplitude, so that the real-time response amplitude tends to the target amplitude range. The monitoring and alarm module is configured to issue a warning message in response to the base intensity obtained by fitting the modified intensity-time model exceeding the preset safety range. For example, it will issue a warning when the stimulator's working state deviates from the expected state (such as when the required intensity abnormally increases beyond the safety limit), reminding clinical personnel to check the electrode position or system status.

[0045] The system is integrated into one or more of the following devices: implantable pulse generator (IPG), transcutaneous electrical nerve stimulation (TENS) device, deep brain stimulator (DBS), and spinal cord stimulator (SCS).

[0046] The neural stimulation system provided in this embodiment and the neural electrical stimulation parameter optimization method based on the modified intensity-time model provided in the above embodiment belong to the same inventive concept. Here, by referencing the entire text, all contents of the embodiment of the neural electrical stimulation parameter optimization method based on the modified intensity-time model are incorporated into this neural stimulation system embodiment.

[0047] The programmer can set electrical stimulation parameters in real time and send them to the stimulation pulse generator, as well as receive feedback of neural electrical activity from the recording electrodes (in this embodiment, muscle contraction state information detected by an accelerometer / six-axis gyroscope). The electrical stimulation pulse generator receives the electrical stimulation parameters from the programmer and outputs an electrical stimulation pulse sequence with set frequency, current intensity, and pulse width. The stimulation electrodes are attached to the limb area requiring electrical stimulation. The recording electrodes, equipped with accelerometers / six-axis gyroscopes, are also attached to the limb area requiring electrical stimulation and can send the muscle movement state to the programmer in real time. The programmer receives the effect of the current electrical stimulation parameters for model fitting, stimulation protocol delivery, and closed-loop adjustment. The parameter fitting process is as follows: Figure 1 As shown, specific numerical examples are as follows: First numerical embodiment: In step S100, a pulse width scanning sequence is set, covering the range from ultrashort pulses (<100μs) to long pulses (>1000μs). Short pulses help correct the model's improvement in the short pulse region, while long pulses help accurately estimate the base intensity. The pulse sequence selected in this specific numerical embodiment is as follows: [t1=30μs,t2=60μs,t3=100μs,t4=300μs,t5=600μs,t6=1000μs]; The target of stimulation is the radial flexor carpi radialis muscle. The programmer sequentially starts with the t1 pulse, gradually increasing the current intensity in 0.5mA steps until the accelerometer / six-axis gyroscope in the recording electrode module detects regular muscle movement. The current intensity is then recorded as the threshold current I. i The values ​​obtained are I1=8.87mA, I2=6.91mA, I3=5.02mA, I4=3.96mA, I5=3.51mA, and I6=3.32mA, respectively.

[0048] In step S200, all the data pairs obtained in step S100 are used to fit the modified intensity-time model through the nonlinear fitting module of the programmer, that is, the data pairs (30μs, 8.87mA), (60μs, 6.91mA), (100μs, 5.02mA), (300μs, 3.96mA), (600μs, 3.51mA), and (1000μs, 3.32mA) are substituted into the modified model: Nonlinear least squares methods, such as the Levenberg-Marquardt algorithm, are used to solve for the basis strength. I rh Threshold charge constant Q th and correction time constant c This makes the threshold predicted by the model and the measured value I... i The overall error is minimized. Define the error function: 2 ,in, n The number of pulse stimuli applied in the prior test; Minimization Error To obtain the optimal parameters I rh =3.04, Q th =269.7, c =15.82.

[0049] Calculate goodness of fit R 2 The score reached 0.99, which meets the evaluation metric of greater than 0.9, and the fit was successful.

[0050] Fit the classical intensity-time model under the same experimental conditions. The fundamental intensity in the classical intensity-time model is obtained. I’ rh =3.34, Q’ th =174.7; The corrected model in this embodiment is: The Weiss model is For the intensity-time curves of both, please refer to [link / reference]. Figure 4 The intensity-time curves of the modified model and the Weiss model are compared, as shown in Table 1: Table 1. Comparison of Threshold Stimulus Current Intensity Prediction Results between the Modified Model and the Classical Model Table 1 shows that compared to the classic Weiss model, the modified model has a significantly lower overall error level and higher overall fitting accuracy. Except for the 100μs pulse width, the errors at other points are all controlled within ±5%. Especially in the short pulse width segments of 30μs and 60μs, the errors of the modified model decreased from 3.27% and -9.55% of the classic Weiss model to 0.68% and -4.49%, respectively, significantly improving the prediction bias of the traditional classic Weiss model under short pulse widths.

[0051] Although the error corresponding to a 100μs pulse width is greater than that of the classic Weiss model, the predicted value of 5.37mA differs from the measured value of 5.02mA by only 0.35mA, which is still within the acceptable range. Compared to the extreme deviation at the short pulse width of 60μs observed in the classic Weiss model, the error distribution of the modified model is more stable, indicating that the bias term introduced by the modified model... c It effectively corrects the inherent biases of the traditional classic Weiss model.

[0052] Energy consumption comparison: based on the energy consumption formula Q = I 2 × t Table 2 shows the energy consumption parameters corresponding to Table 1: Table 2. Comparison of energy consumption parameters between the modified model and the classic model. A negative energy consumption comparison indicates a decrease in the energy consumption of the modified model, while a positive value indicates an increase. As shown in Table 2, compared to the classic Weiss model, the modified model has significantly lower energy consumption. This means that under these conditions, the stimulus scheme corresponding to the modified model has lower energy consumption, less device heat generation, and better battery life.

[0053] At the 60μs point, the energy consumption of the modified model is slightly higher. This is because the current predictions of the modified model at these points are higher than those of the Weiss model and are closer to the measured values, so the energy consumption is also closer to the actual situation.

[0054] Using a time-based strategy, at twice the base intensity, the corrected model has a pulse width of 72.9 μs and an output current of 6.08 mA; the Weiss model has a pulse width of 52.31 μs and an output current of 6.68 mA. Due to the improved accuracy of the prediction models, the corresponding muscle and nerve regions can be activated with more precise pulse width and current parameters, resulting in a better therapeutic experience for patients.

[0055] In addition, during the electrical stimulation therapy, the recording electrodes will still record the muscle movement status in real time. If the current intensity of the recommended parameters does not meet expectations or is too strong, the programmer will increase or decrease the current intensity accordingly, forming a closed-loop control of the treatment process to ensure the best treatment effect.

[0056] Second numerical embodiment: The selected pulse sequence is as follows: [t1=30μs,t2=60μs,t3=100μs,t4=300μs,t5=600μs,t6=1000μs]; The target muscle for stimulation is the vastus lateralis of the thigh. The programmer sequentially starts with the t1 pulse, gradually increasing the current intensity in 0.5mA steps until the accelerometer / six-axis gyroscope in the recording electrode module detects regular muscle movement. The current intensity is then recorded as the threshold current I. i .

[0057] The values ​​obtained are I1=44.5mA, I2=28.5mA, I3=20.5mA, I4=13.0mA, I5=9.5mA, and I6=9.0mA.

[0058] Substitute the data pairs (30μs, 8.87mA), (60μs, 6.91mA), (100μs, 5.02mA), (300μs, 3.96mA), (600μs, 3.51mA), and (1000μs, 3.32mA) into the correction model: Obtain the optimal parameters I rh =6.89, Q th =1462, c =9.9, and the goodness of fit is 9.9. R 2 pass.

[0059] Fit the classical intensity-time model under the same experimental conditions. The fundamental intensity in the classical intensity-time model is obtained. I’ rh =8.41, Q’ th =1114; The corrected model in this embodiment is: The Weiss model is For the intensity-time curves of both, please refer to [link / reference]. Figure 5 The intensity-time curves of the modified model and the Weiss model are compared, as shown in Table 3: Table 3. Comparison of Threshold Stimulus Current Intensity Prediction Results between the Modified Model and the Classical Model Table 1 shows that, except for pulse widths of 300 μs and 1000 μs, the errors at other points were all controlled within ±2.5%. Especially in the 60 μs short pulse width segment, the error of the corrected model decreased from -5.33% of the classic Weiss model to -2.42%, significantly improving the prediction bias of the traditional classic Weiss model under short pulse widths. This indicates that the bias term introduced by the corrected model... c It effectively corrects the inherent biases of the traditional classic Weiss model.

[0060] Although the error corresponding to a 300μs pulse width is greater than that of the classic Weiss model, the predicted current intensity in the 300μs pulse width range is 1.39mA lower than the measured value, which is within the allowable range.

[0061] Energy consumption comparison: based on the energy consumption formula Q = I 2 × t Table 4 shows the energy consumption parameters corresponding to Table 3: Table 4. Comparison of energy consumption parameters between the modified model and the classical model. A negative energy consumption comparison indicates a decrease in the energy consumption of the modified model, while a positive value indicates an increase. As shown in Table 2, compared to the classic Weiss model, the modified model has significantly lower energy consumption. This means that under these conditions, the stimulus scheme corresponding to the modified model has lower energy consumption, less device heat generation, and better battery life.

[0062] At 60μs and 100μs, the energy consumption of the modified model is slightly higher. This is because the current predictions of the modified model at these points are higher than those of the Weiss model and are closer to the measured values, so the energy consumption is also closer to the actual situation.

[0063] Using a time-based strategy, for the modified model, at twice the base intensity (2×6.89), the pulse width was 202.29 μs and the output current was 13.78 mA; for the Weiss model, at twice the base intensity (2×8.41), the pulse width was 132.46 μs and the output current was 16.82 mA. Due to the improved accuracy of the prediction model, the corresponding muscle and nerve regions can be activated with more precise pulse width and current parameters, resulting in a better therapeutic experience for patients.

[0064] In addition, during the electrical stimulation therapy, the recording electrodes will still record the muscle movement status in real time. If the current intensity of the recommended parameters does not meet expectations or is too strong, the programmer will increase or decrease the current intensity accordingly, forming a closed-loop control of the treatment process to ensure the best treatment effect.

[0065] Third numerical embodiment: Based on the data from the second numerical embodiment, the optimized pulse width was calculated using a modified model to be 202.29 μs (≥100 μs), which satisfies... Figure 3 The dual-model adaptive mechanism shown calculates the optimized threshold stimulation current intensity with weight coefficients α=0.5 and β=0.5: 0.5×2×6.89+0.5×2×8.41=15.3mA; Will I =15.3mA Substituted into the corrected model and Weiss model The pulse widths were obtained as 163.94μs and 161.68μs respectively. The optimized pulse width value was then calculated with weighting coefficients α=0.5 and β=0.5: 0.5×163.94μs+0.5×161.68μs=162.81μs.

[0066] Thus, the optimized stimulation parameters obtained under the dual-model adaptive mechanism include an optimized threshold stimulation current intensity of 15.3 mA and an optimized pulse width of 162.81 μs.

[0067] Compared to the measured data of 13mA and 300μs (see...), Figure 3 In the second embodiment, due to the error caused by numerical instability resulting from overfitting of the correction model, the optimized stimulation parameters obtained were 13.78 mA and 202.29 μs. Compared with the optimized stimulation parameters under the dual-model adaptive mechanism, these parameters are closer to the measured data in Table 3. This indicates that the dual-model adaptive mechanism brings the advantage of data universality in optimizing neural electrical stimulation parameters in pulse widths above 100 μs.

[0068] This invention does not limit the values ​​of weighting coefficients α and β to 0.5. For a part of the nerve, the values ​​of weighting coefficients α and β can be assigned through a finite number of trials. If the trials show that the prediction accuracy of the typical Weiss model is more accurate in pulse width segments above 100μs, then the value of weighting coefficient β is increased; otherwise, the value of weighting coefficient α is increased.

[0069] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0070] The above description is only a specific embodiment of this application. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of this application, and these improvements and modifications should also be considered within the scope of protection of this application.

Claims

1. A method for optimizing neural electrical stimulation parameters based on a modified intensity-time model, characterized in that, Includes the following steps: The target nerve was subjected to the following prior tests: applying electrical stimulation with multiple preset pulse widths and acquiring the minimum domain stimulation intensity at each pulse width. Using multiple pairs ( t i , I i The data was fitted to the following modified intensity-time model to obtain individualized parameters. I rh , Q th , c : ； in, I The threshold stimulation current intensity required to induce excitation, t The pulse width of a single electrical stimulation. I rh The base intensity is defined as the minimum stimulation current intensity required to induce nerve excitation when the pulse width approaches infinity. Q th The threshold charge constant is related to the membrane capacitance characteristics. c To correct the time constant; t i This indicates the first action applied to the target nerve during the prior test. i The pulse width of the secondary pulse stimulation. I i For the collected and t i The corresponding minimum domain stimulus intensity; by k The basalt intensity is used as the optimized threshold stimulation current intensity, and the corresponding required pulse width for a single electrical stimulation is calculated as the optimized pulse width, where 1.85 ≤ k ≤2.

15.

2. The method for optimizing neural electrical stimulation parameters according to claim 1, characterized in that, The corresponding conduction velocity was also measured in the prior test; Based on the aforementioned conduction velocity, select k The numerical value: the faster the conduction speed, the... k The larger the value, the slower the conduction speed. k The smaller the value, the better.

3. The method for optimizing neural electrical stimulation parameters according to claim 2, characterized in that, Based on the conduction velocity, the fiber type of the target nerve is determined to be type Aβ, type Aδ, or type C. If the fiber type of the target nerve is Aβ, then k The value ranges from 2 to 2.15; If the fiber type of the target nerve is Aδ, then k The value ranges from 1.9 to 2; If the fiber type of the target nerve is type C, then k The values ​​range from 1.85 to 1.

92.

4. The method for optimizing neural electrical stimulation parameters according to claim 1, characterized in that, Also includes: The target nerve is stimulated with pulsed current intensity and pulse width optimization values, and nerve response signals are collected simultaneously. Based on the amplitude of the neural response signal, the stimulation current intensity and pulse width of the output pulse stimulation signal are adjusted.

5. The method for optimizing neural electrical stimulation parameters according to claim 4, characterized in that, If the amplitude of the acquired neural response signal is lower than the preset first amplitude threshold, the stimulation current intensity and / or pulse width will be increased based on the current output pulse stimulation signal. If the amplitude of the acquired neural response signal is higher than the preset second amplitude threshold, and the second amplitude threshold is greater than or equal to the first amplitude threshold, then the stimulation current intensity and / or pulse width are reduced based on the current output pulse stimulation signal.

6. The method for optimizing neural electrical stimulation parameters according to claim 4, characterized in that, Real-time acquisition of stimulus-induced neural electrical activity as neural response signals, wherein the neural electrical activity includes one or more of compound action potentials, local field potentials, and patient perception threshold markers from external input; When it is determined that the output needs to be adjusted based on the neural response signal, the step size for adjustment is set to be between the base strength obtained from the fitting. I rh The step size is set to be between 1% and 6% of the base strength obtained from the fitting. I rh 5% to 10%.

7. The method for optimizing neural electrical stimulation parameters according to claim 1, characterized in that, The prior test includes at least one electrical stimulation with a pulse width less than or equal to 100 μs and one electrical stimulation with a pulse width greater than or equal to 1000 μs, and the minimum domain stimulation intensity corresponding to each pulse width is measured using an electrophysiological recording system.

8. The method for optimizing neural electrical stimulation parameters according to claim 1, characterized in that, The modified intensity-time model is fitted using the nonlinear least squares method to obtain the base intensity. I rh Threshold charge constant Q th and correction time constant c ; The following evaluation function is used to check the goodness of fit: ； in, R 2 For goodness of fit, I pred_i To minimize the stimulus intensity I i Corresponding pulse width t i Substitute the threshold stimulus current intensity calculated from the fitted modified intensity-time model, For all minimum domain stimulus in the prior test I i The average value; If the goodness of fit R 2 If the preset evaluation index is met, the fitting is considered successful; otherwise, the modified intensity-time model is refitted.

9. The method for optimizing neural electrical stimulation parameters according to claim 1, characterized in that, Define the error function between the predicted values ​​of the modified intensity-time model and the measured values ​​of the minimum domain stimulus intensity as follows: 2 ,in, n The number of pulse stimuli applied in the prior test; Minimization Error To obtain the optimal parameters I rh , Q th , c .

10. The method for optimizing neural electrical stimulation parameters according to claim 1, characterized in that, Also includes: If the baseline strength obtained by fitting I rh If the safety range is exceeded, a warning message will be issued.

11. The method for optimizing neural electrical stimulation parameters according to any one of claims 1 to 10, characterized in that, Also includes: Using multiple pairs ( t i , I i The data was fitted to the following classic intensity-time model to obtain individualized parameters. I ’ rh , Q’ th : ； in, I’ rh The basic intensity in the classical intensity-time model, Q’ th This represents the threshold charge constant in the classical intensity-time model. If the pulse width calculated using the modified intensity-time model exceeds a preset pulse width threshold, then the classical intensity-time model is used instead of the modified intensity-time model to determine the pulse width. k × I’ rh As the optimized value of the threshold stimulation current intensity, or, with k × I’ rh and k × I rh The weighted summation result is used as the optimized value of the threshold stimulation current intensity, wherein the pulse width threshold is the definition value of the ultrashort pulse, and the pulse width threshold is between 0 and 200 μs.

12. A neural stimulation system, characterized in that, include: An electrical stimulation pulse generation module, which is configured to output programmable current pulses; A control module configured to control the output pulse width, amplitude, and frequency of the electrical stimulation pulse generation module; The neural stimulation system uses the optimization method described in any one of claims 1 to 11 to determine the neural electrical stimulation parameters, including: Under the control of the control module, the electrical stimulation pulse generation module performs a priori testing on the target nerve; The control module fits the modified intensity-time model and calculates... k The base intensity corresponds to the required pulse width.

13. The neural stimulation system according to claim 12, characterized in that, The system also includes a neural response signal acquisition module, which is configured to detect the response amplitude of neural electrical activity induced by stimulation and send the detection result to the control module; The control module is configured to adjust the output pulse width and amplitude of the electrical stimulation pulse generation module according to the response amplitude, so that the real-time response amplitude tends to the target amplitude range.

14. The neural stimulation system according to claim 12, characterized in that, The system also includes a monitoring and alarm module, which is configured to issue a warning message in response to the base intensity obtained by fitting the modified intensity-time model exceeding a preset safety range.

15. The neural stimulation system according to claim 12, characterized in that, The system is integrated into one or more of the following devices: an implantable pulse generator, a transcutaneous electrical nerve stimulation device, a deep brain stimulator, and a spinal cord stimulator.

Images