Non-invasive deep vagus nerve stimulation method and system based on time domain interference

Abstract

This invention belongs to the field of brain function-related technology and proposes a non-invasive deep vagus nerve stimulation method and system based on temporal interferometry. The method includes: 3D modeling based on medical images of the patient's neck; simulating the electric field distribution under different electrode spatial layouts using the 3D geometric model of the neck to determine an optimal electrode pair combination library that maximizes the interference field strength in the deep vagus nerve target region; comparing features extracted from the patient's physiological signals with preset treatment goals; dynamically selecting different electrode pair combinations from the optimal electrode pair combination library based on the comparison results; and achieving precise in vivo spatial adjustment of the stimulation focus based on coordinate matching between the 3D geometric model of the neck and real space. This method integrates personalized medical image modeling, computer simulation, dynamic electrode gating, and high spatial registration, solving the pain point of disconnect between planning and execution in the field of neuromodulation.

Description

Technical Field

[0001] This invention belongs to the field of brain function-related technology, and in particular relates to a non-invasive deep vagus nerve stimulation method and system based on temporal interference. Background Technology

[0002] The statements in this section are merely background information related to the present invention and do not necessarily constitute prior art.

[0003] Percutaneous cervical vagus nerve stimulation (TCVNS) is a non-invasive neuromodulation technique that activates the vagus nerve by applying an electric current through electrodes placed on the skin of the neck. Compared to surgical implantable vagus nerve stimulation (VNS), it avoids surgical trauma and infection risks, and has fewer negative impacts on the patient's body, making it a research hotspot in recent years for epilepsy, depression, and inflammation control. However, conventional TCVNS technology, due to limitations in its physical principles, cannot achieve focused current on deep nerves. This is because the soft tissue of the neck is thick and exhibits significant individual differences, causing the applied current to diffuse and attenuate considerably during its conduction from the skin surface to the deep vagus nerve, resulting in insufficient stimulation depth and difficulty in effectively activating the deep vagal trunk; the stimulation effect is highly unstable and easily induces non-targeted side effects such as laryngeal and pharyngeal muscle groups, greatly limiting the maximum usable stimulation dose; TCVNS only controls the stimulation area by changing the electrode position; it lacks the ability of "spatial steering" of the electric field and cannot adjust the "focus" within the nerve; and it is significantly affected by neck anatomy and movement, resulting in large fluctuations in stimulation effects. The neck is a soft tissue area, and skin movement, muscle tension, and changes in head posture can alter the electrode fit; large fluctuations in contact impedance can lead to unstable output.

[0004] Existing technologies include Temporal Interference (TI) stimulation, which uses two or more high-frequency currents to interfere within tissues, "synthesizing" low-frequency modulated signals deep within the tissue, theoretically enabling non-invasive deep neuromodulation. However, current research and applications of this technology primarily focus on deep stimulation of the central nervous system (such as transcranial TI stimulation), aiming to address the depth and focusing issues of transcranial electrical stimulation. Applying TI technology to the peripheral nervous system, particularly vagus nerve stimulation in the neck, presents challenges drastically different from those for brain stimulation: the neck's complex anatomy, with closely adjacent nerves, blood vessels, and muscles, demands extremely high spatial selectivity for stimulation. Existing TI-based brain stimulation protocols, with electrode arrangements and parameter settings specifically designed for the brain, when directly applied to the neck, easily result in interference fields that detach from the target vagus nerve trunk or fail to form an optimal stimulation focus in the target area, leading to low stimulation efficiency or continued co-activation of non-target tissues. Therefore, there is an urgent need for a stimulation protocol specifically optimized for the neck's anatomy, capable of highly selective vagus nerve targeting. Summary of the Invention

[0005] To overcome the shortcomings of the prior art, this invention provides a non-invasive deep vagus nerve stimulation method and system based on temporal interference, which integrates personalized medical image modeling, computer simulation, dynamic electrode gating and high spatial registration to form a complete and precise stimulation process from virtual planning to physical operation, thus solving the pain point of disconnect between planning and execution in the field of neuromodulation.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: In a first aspect, the present invention provides a non-invasive deep vagus nerve stimulation method based on temporal interference, comprising: Based on the three-dimensional modeling of the patient's neck medical images, a three-dimensional geometric model of the neck is obtained. Based on the three-dimensional geometric model of the neck, the electric field distribution under different electrode spatial layouts is simulated to perform time-domain interference electric field simulation, and the optimal electrode pair combination scheme library that maximizes the interference field strength in the deep vagus nerve target area is determined. The system extracts features from the patient's physiological signals, compares the extracted features with the preset treatment goals, and dynamically selects different electrode pair combinations from the optimal electrode pair combination scheme library based on the comparison results. Based on the coordinate matching between the three-dimensional geometric model of the neck and the real space, it achieves in vivo precise spatial adjustment and closed-loop optimization of the stimulation focus.

[0007] Secondly, the present invention provides a non-invasive deep vagus nerve stimulation system based on temporal interferometry, comprising: The simulation module is configured to: perform three-dimensional modeling based on the patient's neck medical images to obtain a three-dimensional geometric model of the neck; and, based on the three-dimensional geometric model of the neck, perform time-domain interference electric field simulation by simulating the electric field distribution under different electrode spatial layouts to determine the optimal electrode pair combination scheme library that maximizes the interference field strength in the deep vagus nerve target area. The stimulation modulation module is configured to: extract features from the patient's physiological signals, compare the extracted features with the preset treatment target, dynamically select different electrode pair combinations from the optimal electrode pair combination scheme library based on the comparison results, and achieve in vivo precise spatial adjustment of the stimulation focus based on the coordinate matching between the three-dimensional geometric model of the neck and the real space.

[0008] Thirdly, the present invention provides an electronic device including a memory and a processor, and computer instructions stored in the memory and running on the processor, wherein the computer instructions, when executed by the processor, perform the method described in the first aspect.

[0009] Fourthly, the present invention provides a computer-readable storage medium for storing computer instructions, which, when executed by a processor, perform the method described in the first aspect.

[0010] Fifthly, the present invention provides a computer program product, including a computer program that, when executed by a processor, implements the method described in the first aspect.

[0011] The above one or more technical solutions have the following beneficial effects: This invention integrates personalized medical image modeling, computer simulation, dynamic electrode gating, and high spatial registration to form a complete and precise stimulation process from virtual planning to physical operation, thus solving the pain point of the disconnect between planning and execution in the field of neuromodulation.

[0012] In this invention, the physical characteristics of temporal interferometry are utilized to enable effective stimulation signals to penetrate to the deep vagus nerve trunk in the neck, achieving a stimulation depth close to that of implantable VNS. By combining individualized modeling simulation with real spatial location matching, the stimulation energy is precisely focused on the target nerve, solving the core defects of traditional non-invasive techniques such as current diffusion and poor targeting.

[0013] This invention combines the physical advantage of TI technology itself, which is that "high-frequency current does not over-activate superficial tissues," with the technical advantage of avoiding stimulation of non-target tissues through precise navigation. With this dual protection, side effects such as pain and muscle spasms can be eliminated to the greatest extent, laying the foundation for long-term and efficient treatment.

[0014] Advantages of additional aspects of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0015] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an improper limitation of the invention.

[0016] Figure 1 This is a functional relationship diagram in an embodiment of the present invention; Figure 2 This is an overall system architecture diagram in an embodiment of the present invention; Figure 3 This is a flowchart of non-invasive deep vagus nerve stimulation based on temporal interference in an embodiment of the present invention; Figure 4 These are two high-frequency sine wave diagrams in an embodiment of the present invention; Figure 5 This is a low-frequency envelope diagram generated by superimposing two high-frequency sine waves in an embodiment of the present invention. Figure 6 This is a simulation diagram of a grid-like electrode array in an embodiment of the present invention. Detailed Implementation

[0017] It should be noted that the following detailed descriptions are exemplary and intended to provide further illustration of the invention. Unless otherwise specified, 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.

[0018] It should be noted that the terminology used herein is for the purpose of describing particular implementations only and is not intended to limit the exemplary implementations of the present invention.

[0019] Where there is no conflict, the embodiments and features in the embodiments of the present invention can be combined with each other.

[0020] Terminology Explanation: Temporal interference stimulation: This refers to a technique that uses two or more high-frequency alternating currents of similar frequencies to propagate and superimpose within biological tissue, thereby "endogenously" generating a low-frequency envelope signal in a deep target region to achieve non-invasive deep neuromodulation. Its core principle is based on the characteristic that neurons are sensitive to low-frequency signals but insensitive to high-frequency signals.

[0021] External cervical vagus nerve: refers to the segment of the vagus nerve that runs within the carotid sheath in the neck, and is the specific anatomical target for non-invasive stimulation in this invention.

[0022] Microarray electrode plate: refers to a flexible circuit board that integrates multiple (usually arranged in a matrix) miniaturized, independent electrode contacts, which can be attached to the skin surface and can dynamically select different combinations of electrode pairs as stimulation ports through a switching matrix.

[0023] Multiplexing switch: refers to a controlled electronic switch array connected between the signal generator and the microarray electrode plate, used to dynamically route the stimulation signal to one or more specified electrode pairs on the microarray according to control instructions.

[0024] Stimulation focus: In temporal interference stimulation, the spatial region where the intensity of the effective low-frequency envelope signal generated by the superposition of two or more high-frequency currents within the tissue is sufficient to activate neurons. Example 1 like Figures 1-6 As shown, this embodiment discloses a non-invasive deep vagus nerve stimulation method based on temporal interferometry, including: Based on the three-dimensional modeling of the patient's neck medical images, a three-dimensional geometric model of the neck is obtained. Based on the three-dimensional geometric model of the neck, the electric field distribution under different electrode spatial layouts is simulated to perform time-domain interference electric field simulation, and the optimal electrode pair combination scheme library that maximizes the interference field strength in the deep vagus nerve target area is determined. The system extracts features from the patient's physiological signals, compares the extracted features with the preset treatment goals, and dynamically selects different electrode pair combinations from the optimal electrode pair combination scheme library based on the comparison results. Based on the coordinate matching between the three-dimensional geometric model of the neck and the real space, it achieves in vivo precise spatial adjustment of the stimulation focus.

[0025] This embodiment implements a non-invasive deep vagus nerve stimulation method based on temporal interferometry, which is achieved through three subsystems: an individualized planning system, a stimulation execution subsystem, and an image-guided navigation subsystem. The individualized planning system generates a three-dimensional anatomical model based on the patient's medical images and outputs an individualized stimulation plan based on simulation calculations. The stimulation execution subsystem activates the corresponding microarray electrode pairs according to the stimulation plan, thereby applying multi-frequency stimulation current to the patient's neck. The image-guided navigation subsystem tracks the electrode plate position in real time and feeds back the spatial coordinates to the main controller to achieve calibration and closed-loop optimization of the stimulation focus.

[0026] In this embodiment, the individualized planning system includes an individualized positioning and modeling module, which is used to acquire and process the anatomical structure information of an individual's neck in order to determine the optimal placement position of the electrode array.

[0027] Specifically, it includes: (1) Medical imaging unit: used to acquire medical imaging data of an individual’s neck, preferably magnetic resonance imaging data.

[0028] (2) Three-dimensional modeling and simulation calculation unit: Based on medical image data, reconstruct the three-dimensional geometric model and electrical model of the neck, including skin, fat, muscle, blood vessels and vagus nerve; and perform computer simulation of time-domain interference electric field based on this three-dimensional geometric model and electrical model of the neck. By simulating the electric field distribution under different electrode spatial layouts, calculate and determine the optimal electrode pair combination scheme library that maximizes the interference field strength in the deep vagus nerve target area.

[0029] The construction of the three-dimensional geometric model and electrical model of the neck is specifically as follows: Acquire the patient's T1 MRI images; Manually label the tissues to be segmented in a portion of the MRI image to form an initial labeled image A; The pixel intensities of different tissues segmented in the initial labeled image A are extracted and substituted into the Gaussian mixture model to calculate the model parameters; The remaining MRI images are segmented using a trained model.

[0030] In this embodiment, a Gaussian mixture model is used for tissue segmentation of medical images in personalized 3D modeling. The Gaussian mixture model considers the grayscale distribution of an image as a mixture of multiple Gaussian distribution components, with each component corresponding to a tissue type.

[0031] Assuming the image has 1 pixel, each pixel The strength is , tag as ,in For the number of categories (e.g., organization type).

[0032] Prior distribution:

[0033] Likelihood function (Gaussian distribution):

[0034] in, The mean, Let Variance be the variance.

[0035] Full probability distribution:

[0036] Based on the previous step, we can obtain... The segmented result can be obtained using the following formula.

[0037] Calculate the posterior probability:

[0038] in, Represents pixels Category The posterior probability.

[0039] Update model parameters:

[0040]

[0041]

[0042] Repeat until convergence.

[0043] After convergence, the final category of each pixel is:

[0044] The determination of the optimal electrode pair combination scheme is as follows: The modified NMR image was used to generate a high-quality tetrahedral mesh via cgal; An array of electrodes (n) is arranged in a coordinate system with the cricoid cartilage skin landmark on the ventral side of the neck as the origin. Using one electrode as the reference electrode and the other electrodes as moving electrodes (n-1), obtain n-1 simulation results to form the leadfield matrix; The optimization aims to maximize the electric field strength of the vagus nerve, and the simplex algorithm is used to obtain the optimal electrode parameters and locations.

[0045] This embodiment uses a coordinate system established with the cricoid cartilage skin landmark on the ventral side of the neck as the origin to define, describe, or guide the location of cervical vagus nerve stimulation electrodes. This includes, but is not limited to: integrating this coordinate system into the operating software interface to assist in positioning; using this coordinate system as a reference for modeling and result analysis in the simulation model; and specifying electrode attachment specifications according to this coordinate system. Any implementation that places the electrode center point within the aforementioned preferred coordinate range is also acceptable.

[0046] The electrode array consists of multiple independent microelectrode contacts, each of which is designed to be much smaller than a conventional transdermal electrode (e.g., in diameter or side length), preferably ranging from 0.5 mm to 5 mm. As a specific, and not limiting, example, the electrode contact could be 2 mm in size.

[0047] The center-to-center spacing (pitch) between the electrode contacts is designed to provide sufficiently high spatial resolution while ensuring stimulation coverage; preferably, the spacing ranges from 3 mm to 7 mm. As a specific, and not limiting, example, the electrode spacing could be 5 mm.

[0048] In this embodiment, the stimulation execution subsystem includes a physiological signal acquisition module, an intelligent control algorithm module, and a signal generation and control module.

[0049] The physiological signal acquisition module is used to monitor the physiological state of the subject in real time. Specifically, the physiological signal acquisition module includes: an electrocardiogram (ECG) signal acquisition unit, an electroencephalogram (EEG) signal acquisition unit, and an electromyography (EMG) signal acquisition unit.

[0050] Among them, the electrocardiogram (ECG) signal acquisition unit is used to acquire ECGs, focusing on monitoring indicators that reflect vagal tone, such as heart rate and heart rate variability.

[0051] EEG signal acquisition unit: used to acquire electroencephalograms and monitor changes in brain rhythms related to disease states or stimulus responses.

[0052] Electromyography (EMG) signal acquisition unit: used to acquire surface EMG signals of the neck, monitor and quantify the degree of coordinated activation of neck muscles such as the sternocleidomastoid muscle due to current diffusion.

[0053] Signal processing and feature extraction module: Connected to the physiological signal acquisition module, it is used to filter, reduce noise, and amplify the acquired raw physiological signals, and extract key feature values ​​related to stimulation effects and side effects.

[0054] The intelligent control algorithm module, acting as the "brain" of the system, receives signals from the signal processing and feature extraction module, compares them with preset treatment targets, and then generates control commands to adjust the output parameters of the signal generation and control module in real time.

[0055] The intelligent control algorithm module receives the feature vector x(t) output by the signal processing and feature extraction module, which includes at least the heart rate and HRV indicators extracted based on electrocardiogram, the muscle coactivation intensity indicators extracted based on neck electromyography, and safety indicators such as contact impedance and registration error related to electrode attachment.

[0056] Intelligent control algorithm module: enter: Physiological signals: ECG (electrocardiogram), EMG (electromyography of the neck), EEG (electroencephalogram) optional; Equipment status: skin-electrode contact resistance, actual output current, temperature rise / overcurrent indicator; Spatial information: current gated electrode pair number, array attitude / registration error, predicted focal coordinates (from navigation module); Output: TI signal parameters: Two / multiple high-frequency carrier frequencies f1, f2, difference frequency Δf = |f1 f2 |; Amplitude A1, A2; Phase difference Δφ; Duty cycle / stimulus-rest ratio.

[0057] Spatial parameters: electrode pair combination (channel selection strategy), and, if necessary, amplitude ratio / phase difference for micro-focusing.

[0058] The control algorithm module in this embodiment is a conventional machine learning algorithm.

[0059] This embodiment performs windowed processing on ECG, EMG, EEG, and equipment operation data, and extracts feature quantities for closed-loop control within each processing window to form feature vectors that are input into the intelligent control algorithm module. To reduce the impact of time-domain interference stimuli on the recorded signal, band-stop / notch filtering can be applied to the recorded signal to suppress the carrier frequency and its harmonics and intermodulation components.

[0060] ECG features: Construct RR interval sequences based on ECG signals and extract features reflecting heart rate and heart rate variability from them. These features may include, but are not limited to, heart rate indicators, time-domain HRV indicators, frequency-domain HRV indicators, and combinations thereof.

[0061] EMG features: These are features extracted from surface electromyography signals to characterize muscle activation intensity, muscle burst patterns, and spectral distribution. These features may include, but are not limited to, amplitude features, burst frequency / duration features, and spectral statistical features. They are used to identify non-target muscle co-activation, abnormal contraction, or motion artifacts, and serve as inputs for closed-loop risk control.

[0062] EEG features: Extract features representing brain state and task response from EEG signals. These features may include, but are not limited to, frequency band power features, frequency band ratios, time-frequency representation features, event-related features, and combinations thereof. When task event markers exist, event-related indicators can be further extracted as optional features.

[0063] Equipment and safety features: Extract features from equipment operation data for safety control and quality assessment. These features may include, but are not limited to, contact impedance, output consistency / closed-loop error, calibration / registration error, lead detachment or saturation markings, etc.

[0064] Feature vectors and closed-loop control inputs: The aforementioned feature quantities are arranged in a preset order to form a feature vector, which is then used to generate TI stimulus parameters and spatial parameters according to a preset closed-loop control strategy. The closed-loop control strategy is used to enable spatial gating and focus fine-tuning of stimulus output permission, stimulus intensity and difference frequency parameter updates, and electrode pair / phase amplitude ratio, while meeting safety constraints. The closed-loop control strategy can be implemented using rule-based control, threshold gating, PID / adaptive control, optimization control (e.g., constraint optimization / model predictive control), state estimation filtering, or one or a combination of machine learning / deep learning models; the specific implementation method is not limited.

[0065] The preset treatment target is defined as a target interval or target vector corresponding to the above characteristics, and can be set by the clinician or automatically generated by the individualized baseline calibration process. The intelligent control algorithm module generates control commands to adjust the output parameters of the signal generation and control module in real time by calculating the deviation e(t) or cost function J(t) between the characteristics and the target. The control commands include at least the amplitude, difference frequency Δf, phase difference, and duty cycle of two or more high-frequency signals, and can further control a multi-channel gating switch to switch electrode pair combinations, realizing dynamic adjustment of the stimulation focus in the body space. When the efficacy index does not reach the target interval, the control algorithm increases the stimulation intensity or adjusts Δf by a preset step size; when the side effect index (e.g., neck electromyography RMS) exceeds the safety threshold, the control algorithm reduces the intensity or automatically switches to an electrode pair combination with lower cost to non-target tissues in the protocol library, thereby inhibiting non-target co-activation while ensuring efficacy.

[0066] Therapeutic efficacy optimization based on ECG / EEG: If the heart rate decrease is insufficient or the target EEG rhythm change does not meet expectations, the control algorithm can increase the stimulation current intensity or adjust the stimulation frequency Δf by a preset step size to enhance the therapeutic effect. Suppression of side effects based on electromyography: If the amplitude of the neck electromyography signal exceeds the safety threshold, the control algorithm can immediately reduce the current intensity by step size, or automatically switch to another set of electrode pair combinations in the optimal electrode pair combination scheme library that can reduce muscle co-activation. When all physiological indicators have entered the preset "optimal treatment window" and remain stable, the system maintains the current parameters for continuous stimulation.

[0067] The signal generation and control module is used to generate two high-frequency AC signals with frequencies of f1 and f2, respectively. The frequency difference |f1 - f2| = Δf is the effective stimulation frequency of the target nerve (1-100 Hz). The signal generation and control module has the ability to independently adjust the signal intensity, phase and timing of each pathway.

[0068] The configurable microarray electrode plate integrates multiple microelectrodes arranged in a high-density array on a flexible substrate, which can be attached to the skin surface of the neck. The electrodes in the microarray are connected to the signal generation and control module through a multiplexer switch, enabling the system to dynamically select different electrode pair combinations based on the optimization results of the individualized positioning and modeling module, so as to output two or more high-frequency currents, thereby achieving precise in vivo spatial adjustment of the stimulation focus.

[0069] The system dynamically selects from a library of optimal electrode pair combinations generated by the individualized positioning and modeling module. This library contains pre-stored predicted interference field strength indices for the target vagus nerve region and cost indices for non-target tissues (such as the sternocleidomastoid muscle, laryngeal muscles, and / or adjacent blood vessels) for each electrode pair combination. It can also further pre-store sensitivity parameters for electrode attachment offset and posture changes. During stimulation, the intelligent control algorithm module acquires real-time efficacy-related features of ECG and / or EEG, as well as side effect-related features of cervical electromyography, and safety features such as electrode contact impedance and virtual-real registration error. Under the premise of meeting safety thresholds and the lower limit of interference field strength in the target region, the system performs multi-target scoring and ranking of candidate electrode pair combinations, prioritizing those with high efficacy in the target region, low non-target cost, and greater robustness to posture deviations. The system uses a multi-channel gating switch to switch the output port to achieve in vivo spatial adjustment of the stimulation focus. After selecting an electrode pair combination, the system can also achieve minute displacement of the focus by adjusting the amplitude ratio and / or phase difference of the two signals to inhibit muscle co-activation while ensuring therapeutic efficacy.

[0070] Specifically: 1. Pre-treatment stage: Establish an "electrode pair performance database".

[0071] After individualized modeling and simulation, the system not only obtains a set of "optimal" electrode pairs, but also performs simulation evaluation on all possible combinations of electrode pairs on the microarray that meet the safe spacing requirements and can serve as stimulation ports, generating a structured database. Each record contains: Electrode pair ID: Identifies which two electrodes are involved.

[0072] Predicted electric field indices: the maximum electric field strength, average electric field strength, and neural volume exceeding the activation threshold generated by this combination on the target vagus nerve.

[0073] Side effect prediction indicators: the maximum electric field strength or activation risk score predicted by the combination in key non-target tissues (such as the sternocleidomastoid muscle and the vicinity of the recurrent laryngeal nerve).

[0074] Spatial attribute: The three-dimensional coordinates of the predicted stimulus focus generated by this combination in the cricoid cartilage coordinate system.

[0075] 2. Dynamic selection logic and strategy.

[0076] During actual stimulation, the master control system dynamically selects and switches electrode pairs from the aforementioned database based on the following strategy: a. Initial startup strategy: the optimal choice based on comprehensive analysis; Upon initial system startup, a highly efficient stimulation starting point must be established while ensuring safety. At this time, the control algorithm queries the pre-stored "electrode pair-performance database" and weights-scores all candidate combinations based on predicted efficacy indicators (such as the average electric field strength on the vagus nerve) and predicted safety indicators (such as the highest electric field in the muscle region). The algorithm automatically selects the electrode pair combination with the highest overall score as the initial stimulation protocol, thus achieving the optimal balance between efficacy and safety from the outset of treatment.

[0077] b. Strategies for optimizing therapeutic effects: progressive adjustment of intensity and space; When real-time physiological feedback (such as insufficient improvement in heart rate variability) indicates that the therapeutic effect is not as expected, the system initiates an optimization process. First, the algorithm attempts to gradually increase the stimulation current intensity of the currently used electrode pair within a safe range. If the current intensity has reached the preset safe upper limit and the therapeutic effect is still insufficient, the system will initiate spatial adjustment: based on the spatial attributes in the database, it will search for other electrode pair combinations whose predicted stimulation focus position is close to the current focus, but whose predicted electric field intensity is significantly higher, and automatically switch to that combination. This can enhance the energy accumulation in the target area by changing the current path and interference mode without exceeding the safe current upper limit.

[0078] c. Side effect avoidance strategies: rapid response and proactive switching; When side effect signals are detected in real time (such as excessive neck electromyography amplitude), the system prioritizes safety and comfort. The first step is to immediately reduce the stimulation current intensity of the current electrode pair. If the side effects disappear after reducing the current, this state is maintained; if reducing the current leads to a significant decrease in efficacy, or if the side effects persist, the second step is executed: the system quickly searches the database to find alternative electrode pair combinations that predict the lowest side effect indicators while minimizing the decline in predicted efficacy indicators, and then switches to them. The core of this strategy is that when local current distribution causes problems, actively changing the stimulation entry point fundamentally avoids activation of sensitive non-target tissues.

[0079] d. Focus fine-tuning strategy; In the image-guided navigation interface, when the operator needs to make fine adjustments based on the real-time display of the predicted focus and its relationship to the neuroanatomical location, the system supports manual fine-tuning. The operator simply clicks on the desired new stimulation target on the 3D model, and the algorithm calculates the spatial coordinates of the target in real time, retrieves the best-matching electrode pair combination from the database based on the predicted focus coordinates, and then automatically performs the switch.

[0080] e. Stimulus scanning strategy; To cover longer nerve segments or find the optimal response point for an individual, the system can execute a preset scanning stimulation mode. In this mode, the control algorithm sequentially calls a series of electrode pairs arranged in an orderly manner along the course of the vagus nerve, based on the predicted focal coordinates in the database, and automatically switches to the next set after stimulating one set of electrodes for a specific duration.

[0081] In this embodiment, the image-guided navigation module is used to achieve precise spatial reproduction of the stimulation protocol. It tracks the positions of the electrode pads and the patient in the real world using a spatial positioning device and accurately registers them with a virtual, individualized 3D digital model. This allows for real-time visualization of the stimulation focus position on the screen, providing the operator with spatial guidance. It includes: (1) Spatial positioning unit: preferably an ultrasound positioning system, used to track the three-dimensional spatial coordinates of positioning markers attached to the microarray electrode plate and the patient's body surface in real time.

[0082] (2) Spatial registration unit: The real spatial coordinates obtained through the spatial positioning unit are precisely matched with the virtual spatial coordinates in the individualized 3D model to realize the coordinate system unity between the virtual world and the real world.

[0083] (3) Visual interaction unit: The registered three-dimensional model of the neck is displayed in real time on the display device, and the predicted stimulation focus calculated based on the current activated electrode pair is dynamically rendered, providing the operator with a "what you see is what you get" navigation interface.

[0084] The control and processing unit is electrically connected to the signal generation and control module and the multiplexing switch. It is used to control the output parameters of each signal and execute pre-stored or received optimized electrode pair configuration instructions from the simulation unit.

[0085] Optionally, distributed discrete electrodes can replace integrated microarray plates. While integrated microarray electrode plates are the preferred solution, this embodiment also covers the use of multiple independent, discrete surface electrodes. Based on simulation results, the operator can manually and precisely attach individual electrodes to the optimal points determined by the neck simulation and connect them to a selector switch via wires. This method reduces the manufacturing cost of the electrode plate while still achieving basic stimulation point switching functionality.

[0086] Optionally, for special applications requiring higher stimulation efficiency and signal-to-noise ratio (such as in animal experiments), the electrode array in this embodiment can be replaced with a "microneedle electrode" or a "retractable needle electrode array". When the electrode is attached, the microneedles can penetrate into the stratum corneum of the skin, but the depth is far from reaching the nerve trunk, thereby reducing skin contact resistance and improving stimulation efficiency, while still maintaining the "minimally invasive" characteristics and avoiding traditional surgical implantation.

[0087] Optionally, by attaching optical positioning markers to the microarray electrode plates and the patient's neck, the system can track their three-dimensional spatial coordinates in real time. By spatially registering the acquired real-world spatial coordinates with the individualized three-dimensional model, the positions of the electrode plates and predicted stimulation focus in the virtual model can be displayed in real time and accurately on a display device.

[0088] Optionally, the focus can be moved by adjusting the amplitude of the two signals and the spatial position of the electrode pair. Alternatively, the electrode pair and current amplitude can be fixed, allowing for precise control of the relative phase of the two (or more) high-frequency signals. By changing the phase difference, the stimulation focus can also be moved within tiny areas deep within the tissue, providing another technical approach for fine-tuning the focus.

[0089] This embodiment utilizes the physical characteristics of time-domain interferometry to enable effective stimulation signals to penetrate deep into the vagus nerve trunk in the neck, achieving a stimulation depth close to that of an implanted VNS.

[0090] This embodiment combines individualized modeling and simulation with image-guided real-time navigation to precisely focus stimulation energy onto the target nerve, solving the core defects of traditional non-invasive techniques such as current diffusion and poor targeting.

[0091] In this embodiment, the design of the microarray electrode and the multi-channel selection switch enables the system to dynamically adjust the stimulation focus in vivo without physically moving the electrode plate. This not only compensates for attachment errors but also enables scanning stimulation or avoidance of uncomfortable areas, meeting the complex needs of clinical applications.

[0092] This embodiment integrates personalized medical image modeling, computer simulation optimization, dynamic gating of microarray electrodes, and high-precision ultrasound spatial navigation to form a complete and precise stimulation system from virtual planning to physical operation, solving the pain point of "disconnect between planning and execution" in the field of neuromodulation.

[0093] This embodiment uses image-guided navigation to intuitively ensure that the stimulation focus locks onto the target nerve, fundamentally solving the problems of blindness in traditional tcVNS and insufficient targeting of existing TI technology in peripheral applications.

[0094] This embodiment combines the physical advantage of TI technology itself, which is that "high-frequency current does not over-activate superficial tissues," with the technical advantage of avoiding stimulation of non-target tissues through precise navigation. With this dual protection, side effects such as pain and muscle spasms can be eliminated to the greatest extent, laying the foundation for long-term and efficient treatment.

[0095] Example 2 The purpose of this embodiment is to provide a non-invasive deep vagus nerve stimulation system based on temporal interferometry, including: The simulation module is configured to: perform three-dimensional modeling based on the patient's neck medical images to obtain a three-dimensional geometric model of the neck; and, based on the three-dimensional geometric model of the neck, perform time-domain interference electric field simulation by simulating the electric field distribution under different electrode spatial layouts to determine the optimal electrode pair combination scheme library that maximizes the interference field strength in the deep vagus nerve target area. The stimulation modulation module is configured to: extract features from the patient's physiological signals, compare the extracted features with the preset treatment target, dynamically select different electrode pair combinations from the optimal electrode pair combination scheme library based on the comparison results, and achieve in vivo precise spatial adjustment of the stimulation focus based on the coordinate matching between the three-dimensional geometric model of the neck and the real space.

[0096] In further embodiments, the following is also provided: An electronic device includes a memory and a processor, as well as computer instructions stored in the memory and running on the processor. When executed by the processor, the computer instructions perform the method described in Embodiment 1. For brevity, further details are omitted here.

[0097] It should be understood that in this embodiment, the processor can be a central processing unit (CPU), or it can be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor, etc.

[0098] Memory may include read-only memory and random access memory, and provides instructions and data to the processor. A portion of memory may also include non-volatile random access memory. For example, memory may also store information about the device type.

[0099] A computer-readable storage medium for storing computer instructions, which, when executed by a processor, perform the method described in Embodiment 1.

[0100] The method in Embodiment 1 can be directly implemented by a hardware processor, or implemented by a combination of hardware and software modules within the processor. The software modules can reside in readily available storage media in the art, such as random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, or registers. This storage medium is located in memory; the processor reads information from the memory and, in conjunction with its hardware, completes the steps of the above method. To avoid repetition, a detailed description is not provided here.

[0101] A computer program product includes a computer program that, when executed by a processor, implements the method described in Embodiment 1.

[0102] The present invention also provides at least one computer program product tangibly stored on a non-transitory computer-readable storage medium. The computer program product includes computer-executable instructions, such as instructions included in program modules, which execute in a device on a target real or virtual processor to perform the processes / methods described above. Typically, program modules include routines, programs, libraries, objects, classes, components, data structures, etc., that perform specific tasks or implement specific abstract data types. In various embodiments, the functionality of program modules can be combined or divided among program modules as needed. The machine-executable instructions for the program modules can execute within a local or distributed device. In a distributed device, the program modules can reside in both local and remote storage media.

[0103] The computer program code used to implement the methods of the present invention may be written in one or more programming languages. This computer program code may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when executed by the computer or other programmable data processing device, the program code causes the functions / operations specified in the flowcharts and / or block diagrams to be implemented. The program code may be executed entirely on a computer, partially on a computer, as a stand-alone software package, partially on a computer and partially on a remote computer, or entirely on a remote computer or server.

[0104] In the context of this invention, computer program code or related data may be carried by any suitable carrier to enable a device, apparatus, or processor to perform the various processes and operations described above. Examples of carriers include signals, computer-readable media, and the like. Examples of signals may include electrical, optical, radio, sound, or other forms of propagation signals, such as carrier waves, infrared signals, etc.

[0105] Those skilled in the art will recognize that the units and algorithm steps described in conjunction with the embodiments herein can be implemented in electronic hardware or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0106] While the specific embodiments of the present invention have been described above in conjunction with the accompanying drawings, this is not intended to limit the scope of protection of the present invention. Those skilled in the art should understand that various modifications or variations that can be made by those skilled in the art without creative effort based on the technical solutions of the present invention are still within the scope of protection of the present invention.

Claims

1. A non-invasive deep vagus nerve stimulation method based on temporal interferometry, characterized in that, include: Based on the three-dimensional modeling of the patient's neck medical images, a three-dimensional geometric model of the neck is obtained. Based on the three-dimensional geometric model of the neck, the electric field distribution under different electrode spatial layouts is simulated to perform time-domain interference electric field simulation, and the optimal electrode pair combination scheme library that maximizes the interference field strength in the deep vagus nerve target area is determined. The system extracts features from the patient's physiological signals, compares the extracted features with the preset treatment goals, and dynamically selects different electrode pair combinations from the optimal electrode pair combination scheme library based on the comparison results. Based on the coordinate matching between the three-dimensional geometric model of the neck and the real space, it achieves in vivo precise spatial adjustment of the stimulation focus.

2. The non-invasive deep vagus nerve stimulation method based on temporal interference as described in claim 1, characterized in that, A three-dimensional geometric model of the neck is obtained by performing three-dimensional modeling based on the patient's neck medical images. Based on this model, time-domain interference electric field simulation is conducted by simulating the electric field distribution under different electrode spatial layouts. This determines the optimal electrode pair combination scheme library that maximizes the interference field strength in the deep vagus nerve target region. Specifically: Used to reconstruct a three-dimensional geometric and electrical model of the neck, including skin, fat, muscles, blood vessels, and vagus nerve, based on medical imaging data of the neck. Based on the three-dimensional geometric and electrical models of the neck, computer simulation of the time-domain interference electric field was performed. By simulating the electric field distribution under different electrode spatial layouts, the optimal electrode pair combination scheme library that maximizes the interference field strength in the deep vagus nerve target area was calculated and determined.

3. The non-invasive deep vagus nerve stimulation method based on temporal interference as described in claim 2, characterized in that, When constructing the three-dimensional geometric and electrical models of the neck, a Gaussian mixture model is used for tissue segmentation of medical images, specifically including: Acquire the patient's T1 MRI images; Manually label the tissues to be segmented in a portion of the MRI image to form an initial labeled image A; The pixel intensities of different tissues segmented in the initial labeled image A are extracted and substituted into the Gaussian mixture model to calculate the model parameters; The remaining MRI images are segmented using a trained model.

4. The non-invasive deep vagus nerve stimulation method based on temporal interference as described in claim 1, characterized in that, The process of determining the optimal electrode pair combination scheme library through 3D modeling and simulation calculation units specifically includes: The modified NMR image was used to generate a high-quality tetrahedral mesh using CGAL. A coordinate system was established with the cricoid cartilage skin landmark on the ventral side of the neck as the origin, which was used to arrange the array electrodes; By using one electrode as the reference electrode and the other electrodes as moving electrodes, multiple simulation results are obtained and a forward matrix is ​​formed. The optimization aims to maximize the electric field strength of the vagus nerve. The simplex algorithm is used to obtain the optimal electrode parameters and positions, forming the optimal electrode pair combination scheme library.

5. The non-invasive deep vagus nerve stimulation method based on temporal interference as described in claim 1, characterized in that, The extracted features are compared with the preset treatment goals, and different electrode pair combinations are dynamically selected from the optimal electrode pair combination library based on the comparison results. Specifically: Based on the efficacy optimization strategy, when the improvement in heart rate variability does not meet expectations, the intensity of the stimulation current or the difference frequency is gradually increased; If the current intensity reaches the safety limit, then search the optimal electrode pair combination scheme library and switch to an electrode pair combination with similar predicted focus but higher predicted electric field intensity. Based on the side effect suppression strategy, when the amplitude of the neck electromyography signal exceeds the safety threshold, the current intensity is reduced, or the system automatically switches to an electrode pair combination with lower non-target tissue cost in the optimal electrode pair combination scheme library.

6. The non-invasive deep vagus nerve stimulation method based on temporal interference as described in claim 1, characterized in that, Using machine learning algorithms, based on the deviation between the input physiological signals, device status signals, and spatial location information and the preset treatment target, control instructions including stimulation signal parameters and spatial gating parameters are generated.

7. A non-invasive deep vagus nerve stimulation system based on temporal interferometry, characterized in that, include: The simulation module is configured to: perform three-dimensional modeling based on the patient's neck medical images to obtain a three-dimensional geometric model of the neck; and, based on the three-dimensional geometric model of the neck, perform time-domain interference electric field simulation by simulating the electric field distribution under different electrode spatial layouts to determine the optimal electrode pair combination scheme library that maximizes the interference field strength in the deep vagus nerve target area. The stimulation modulation module is configured to: extract features from the patient's physiological signals, compare the extracted features with the preset treatment target, dynamically select different electrode pair combinations from the optimal electrode pair combination scheme library based on the comparison results, and achieve in vivo precise spatial adjustment of the stimulation focus based on the coordinate matching between the three-dimensional geometric model of the neck and the real space.

8. An electronic device, characterized in that, It includes a memory and a processor, as well as computer instructions stored in the memory and running on the processor, which, when executed by the processor, perform the method according to any one of claims 1-6.

9. A computer-readable storage medium, characterized in that, Used to store computer instructions, which, when executed by a processor, perform the method described in any one of claims 1-6.

10. A computer program product, characterized in that, Includes a computer program, which, when executed by a processor, implements the method described in any one of claims 1-6.

Images