Temporal interference-based closed-loop multimodal neural stimulation system and method

Abstract

The present application pertains to the technical field of neural stimulation. Disclosed are a temporal interference-based closed-loop multimodal neural stimulation system and method. The system comprises a temporal interference stimulation system, an electroencephalography-functional near-infrared spectroscopy sampling system, and an upper-level control system. The temporal interference stimulation system utilizes a beat-frequency electric field generated by two sets of electrodes to precisely stimulate a specified brain region. The electroencephalography-functional near-infrared spectroscopy sampling system is a bimodal collector coupling electroencephalography and functional near-infrared spectroscopy, including two parts: signal extraction and correlation analysis, and analyzes stimulation effects and adjusts stimulation schemes by integrating unified brain signal data that combines the temporal precision of EEG and the spatial precision of fNIRS. The upper-level control system includes bimodal fusion model computation, graph convolutional neural network prediction, and stimulation scheme formulation. The present application addresses the problems that traditional stimulation methods lack a closed-loop regulation system, have no means for calibration and optimization, and require a long adaptation period between the stimulation scheme and the user, thus being disadvantageous for applications.

Description

A closed-loop multimodal neural stimulation system and method based on time interference

[0001] This application claims priority to the Chinese patent application filed with the China Patent Office on March 25, 2024, with application number 202410346188.9 and invention name “A closed-loop multimodal neural stimulation system and method based on time interference”, the entire contents of which are incorporated by reference into this application. Technical Field

[0002] The present application belongs to the field of neural stimulation technology, and in particular relates to a closed-loop multimodal neural stimulation system and method based on time interference. Background Art

[0003] Brain activity is primarily manifested in neuronal discharges, so modulating these electrical potentials can induce short-term or long-term changes in brain function. Direct, artificial manipulation of neuronal activity is highly relevant to the treatment of neurological and psychiatric disorders. It holds enormous potential for clinical use in the treatment of neurological disorders and can also contribute to scientific research into the underlying mechanisms of sensory, motor, and cognitive processes in the human brain.

[0004] In some cases, neurostimulation methods mainly include electrical stimulation, magnetic stimulation, optogenetic stimulation, and ultrasonic stimulation. Electrical stimulation activates neurons by delivering electric current to the nervous system. There are two main routes: deep brain stimulation with invasive electrode implants and non-invasive transcranial electrical stimulation. Magnetic stimulation affects brain activity by placing a coil on the scalp to generate a magnetic field. Optogenetic stimulation transfects cells with photosensitive proteins through genetic engineering, and stimulates these proteins with light to cause neuronal excitation. Ultrasonic stimulation stimulates neurons through focused ultrasound. Electrical stimulation and optogenetic stimulation are widely used in brain function research and the treatment of neurological diseases due to their significant effects and high reliability. From the perspective of stimulation method, existing mainstream stimulation methods have obvious drawbacks: deep brain stimulation requires the implantation of electrodes in the brain. This invasive procedure may lead to potential complications such as infection and bleeding, increasing the complexity and risk of the implantation surgery, and also requiring patients to recover for a longer time. In addition, the electrode implantation method has great disadvantages in stimulation selectivity and spatial resolution due to the limitations of implantation depth and electrode positioning. Although transcranial electrical stimulation is non-invasive, transcranial current can only provide subthreshold stimulation. Using this method can increase the activity of neurons and make them easier to activate, but it cannot directly activate nerves. Moreover, transcranial current is non-directional and any area covered by the current will be affected. These two characteristics greatly limit the application of transcranial electrical stimulation. Optogenetic stimulation has high stimulation accuracy and fast response speed, but because it requires surgical implantation of the stimulator and involves transgenic technology, it is also very limited in actual application.

[0005] From the perspective of system structure, traditional stimulation methods do not have a closed-loop control system and lack correction and optimization means. Due to the large individual differences among users, the stimulation scheme and the user have a long running-in period, which is not conducive to its application. Summary of the Invention

[0006] The purpose of this application is to provide a closed-loop multimodal neural stimulation system and method based on time interference, so as to solve the problems existing in the above-mentioned technology that the traditional stimulation method has no closed-loop control system, lacks correction and optimization means, and the stimulation scheme has a long running-in period with the user, which is not conducive to application.

[0007] To achieve the above-mentioned objectives, the present application provides a closed-loop multimodal neural stimulation system based on time interference, including: a time interference stimulation system, an EEG-functional near-infrared sampling system and a host control system; the time interference stimulation system uses the beat electric field generated by two sets of electrodes to precisely stimulate the specified area of ​​the brain; the EEG-functional near-infrared sampling system is a dual-modal collector that couples EEG and functional near-infrared, and includes two parts: signal extraction and correlation analysis. It analyzes the stimulation effect and adjusts the stimulation plan by integrating the integrated brain signal data with EEG time accuracy and fNIRS spatial accuracy; the host control system includes dual-modal fusion model calculation, graph convolutional neural network prediction and stimulation plan formulation.

[0008] In one embodiment, the time interference stimulation system includes: a communication control module, a power supply module and a time interference stimulation module; the communication control module adopts a SoC chip, which internally integrates the functions of a processor and wireless communication, and is used for communication between the time interference stimulation system and the upper control system and control of the stimulation current; the power supply module adopts a design of a voltage regulator with low noise and low voltage drop plus a shielding layer; the time interference stimulation module is composed of two difference beat electrode groups, and two sine wave generators generate two 2KHz AC signals with a phase difference of 10Hz-20Hz. The two AC signals interfere with each other in the skull through the electrode group attached to the surface of the scalp, and the low-frequency envelope current generated by the interference stimulates the nerve cells in the designated area.

[0009] In one embodiment, the EEG-functional near-infrared sampling system includes: an EEG acquisition module and an fNIRS functional near-infrared brain signal acquisition module; the EEG acquisition module uses epidermal electrodes to record signals; the EEG signal generated by time interference stimulation in the EEG acquisition module is filtered and denoised by a preprocessing circuit, and then converted into digital form by an analog-to-digital converter and sent to the communication and control subsystem. The EEG signal is subjected to artifact removal and baseline drift removal by an MCU and then uploaded to a host control system; the fNIRS functional near-infrared brain signal acquisition module measures the activity level of a specific brain area by monitoring the absorption changes of brain hemoglobin and oxyhemoglobin within the near-infrared spectrum. The fNIRS functional near-infrared brain signal acquisition module uses a light-emitting diode as a light source to generate 650nm near-infrared light, and a photodiode as a light signal detector to convert the light signal passing through the brain tissue into an electrical signal for analysis. The obtained electrical signal is denoised and filtered by preprocessing, and then converted into digital form by an analog-to-digital converter and sent to the communication and control subsystem. The MCU calculates brain tissue blood oxygen to obtain a brain activity signal and uploads the activity signal to the host control system.

[0010] In one embodiment, the bimodal fusion model is calculated by extracting features from EEG and fNIRS signal data and combining the two into a fusion model followed by correlation analysis.

[0011] In one embodiment, the graph convolutional neural network prediction predicts the stimulation location based on the desired stimulation effect and formulates a stimulation plan, which is expressed as follows:

[0012] Let G = (V; E) represent a graph network, V represents the node set, E represents the edge set, if two nodes v i ,v j There are edges connected, then (v i ,v j )∈E, the entire graph network is considered to be composed of nodes composed of different brain functional areas connected by connecting nerves, and the whole is a structure composed of subgraphs with a finite number of nodes;

[0013] The four brain waves in a certain brain region node for a fixed period of time are recorded as different dimensions, the number of dimensions is D, and the number of samples is τ. Then the input brain activity feature matrix of N brain network nodes at time t is expressed as follows: χ = (X1, X2, ..., X t ,...,X τ )∈R N×D×τ

[0014] where X t The brain activity feature matrix of N nodes in the figure at time t is expressed as follows:

[0015] For an output brain activity feature matrix with K samples, it is expressed as follows:

[0016] where Y K represents the brain activity matrix of N nodes at the Kth time point;

[0017] The prediction task uses data sampled by N brain network nodes at adjacent T times and expresses the brain activity feature prediction task as a function ψ θ , the expression is as follows: γ=ψ θ (X t-T+1 ,X t-T ,...,X t ; G).

[0018] The present application also provides a closed-loop multimodal neural stimulation method based on time interference, comprising the following steps:

[0019] Step 1: Determine the stimulation target, the planned stimulation location, and the expected four brainwave characteristic signal responses, including δ, θ, α, and β. The frequency of δ is 0.5Hz-3.5Hz, the frequency of θ is 4Hz-7Hz, the frequency of α is 8Hz-13Hz, and the frequency of β is 14Hz-30Hz.

[0020] Step 2: Setting stimulation parameters. If this is the first stimulation, there is no available graph convolutional neural network feedback results at this time. The placement of the two sets of electrodes and the current frequency need to be calculated based on the characteristics of the envelope electric field. If this is not the first stimulation, the electrode placement and frequency parameter settings are determined based on the results predicted by the graph convolutional neural network.

[0021] Step 3: Execute the stimulation phase, using two sets of electrodes placed on the scalp to generate transcranial alternating current with a slight frequency difference. The two sets of currents will stimulate a low-frequency envelope electric field in a specific area of ​​the brain to activate neurons;

[0022] Step 4: Dual-modal data acquisition and preprocessing: EEG and fNIRS systems jointly collect brain signals, preprocess the collected signals, and package them according to the corresponding sampling time;

[0023] Step 5: Data processing and correlation analysis: The upper control system extracts features from the two modal data and performs correlation analysis to obtain temporal and spatial component features.

[0024] Step 6: Determine whether the stimulation has achieved the expected goal by analyzing the collected four brain wave signals and physiological responses of the stimulated person;

[0025] Step 7: Graph Convolutional Neural Network prediction phase: Use the previous stimulation dataset and the stimulation dataset collected during the stimulation to train the graph convolutional neural network prediction model, and use the four brain waves characteristic of expected brain activity as the starting point to reversely predict the original stimulated points and activation status;

[0026] Step 8: Stimulation scheme optimization phase, using the graph convolutional neural network to predict the stimulated position and activation degree to adjust the electrode placement and the generated envelope electric field frequency.

[0027] In one embodiment, the expression for calculating the placement positions of the two sets of electrodes in step 2 based on the generation characteristics of the envelope electric field is as follows:

[0028] Where E is the electric field intensity, J is the current density, ρ is the spatial current density, and B is the magnetic field intensity. Finite element simulation can be used to determine the distribution of the envelope electric field in the skull and determine the placement of the electrodes. represents the Hamiltonian operator, ε0 represents the vacuum permittivity, j represents an imaginary number, μ0 represents the vacuum permeability, and → represents a vector.

[0029] In one embodiment, the process of dual-modal data collection and preprocessing in step 4 is as follows:

[0030] S41, turning on the near-infrared light source to simultaneously capture electrical and optical signals from the sensor to obtain EEG and fNIRS signals;

[0031] S42, performing analog-to-digital conversion on the acquired signal after removing noise and filtering out clutter;

[0032] S43, collecting the data to the data acquisition system MCU for artifact removal and baseline drift removal;

[0033] S44. Pack the processed EEG and fNIRS data according to a unified clock, and send the data packets in time sequence to the upper control system.

[0034] In one embodiment, the data processing and correlation analysis process in step 5 is as follows:

[0035] S51, extract data features from EEG and fNIRS data;

[0036] S52, using the CCA algorithm to perform correlation analysis on the extracted features to obtain time component features and spatial component features;

[0037] S53. Use the temporal component features and the spatial component features for cross-validation and summarize the bimodal fusion data.

[0038] Therefore, the present application adopts the above-mentioned closed-loop multimodal neural stimulation system and method based on time interference, which can accurately and customizedly determine the actual stimulation plan under different individual conditions; in addition, deep nerve stimulation can be performed without electrode implantation surgery, which greatly reduces costs and greatly improves safety; finally, the complete closed-loop system set up in this application enables the operator to efficiently optimize and adjust the stimulation plan, greatly reducing the time cost in actual use. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] FIG1 is a system diagram of a closed-loop multimodal neural stimulation method based on time interference according to the present application;

[0040] FIG2 is a structural diagram of a closed-loop multimodal neural stimulation system based on time interference in the present application;

[0041] Figure 3 is a flowchart of the correlation analysis of this application;

[0042] Figure 4 is a schematic diagram of the brain network of this application;

[0043] Figure 5 is a diagram of the convolutional neural network structure of this application. DETAILED DESCRIPTION

[0044] The following will be combined with the drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0045] In order to make the above-mentioned purposes, features and advantages of the present application more obvious and easy to understand, the present application is further described in detail below with reference to the accompanying drawings and specific implementation methods.

[0046] Referring to Figures 1 to 5, a closed-loop multimodal neural stimulation system based on temporal interferometry includes: a temporal interferometry stimulation system, an EEG-functional near-infrared sampling system, and a host control system. The temporal interferometry stimulation system utilizes the beat electric field generated by two sets of electrodes to precisely stimulate a specific brain region. The EEG-functional near-infrared sampling system is a dual-modality collector that couples EEG and functional near-infrared signals. It includes signal extraction and correlation analysis, integrating brain signal data with the temporal precision of EEG and the spatial precision of fNIRS to analyze stimulation effects and adjust stimulation protocols. The host control system includes dual-modal fusion model calculation, graph convolutional neural network prediction, and stimulation protocol development. In this embodiment, the host control system is specifically a host computer.

[0047] The temporal interferometric stimulation system includes a communication control module, a power supply module, and a temporal interferometric stimulation module. The communication control module uses a Nordic nRF52832 series SoC chip, which integrates an STM32M4F-level processor and 2.4GHz wireless communication capabilities. It is responsible for communication between the temporal interferometric stimulation system and the host control system and for controlling the stimulation current. The power supply module uses a voltage regulator with low noise (peak-to-peak ripple voltage less than 5mV) and low voltage drop (minimum voltage drop less than 300mV) plus a shielding layer design to ensure that the precise high-frequency signal of the electrode part is not interfered with by noise. The temporal interferometric stimulation module consists of two different beat electrode groups. Two sine wave generators generate two 2kHz AC signals with a phase difference of 10Hz-20Hz. The two AC signals will interfere with each other in the skull through the electrode group attached to the scalp surface. The low-frequency envelope current generated by the interference can stimulate nerve cells in the designated area. This stimulation method can directly stimulate nerve cells in the designated area to produce action potentials while not affecting the adjacent brain areas.

[0048] The EEG-functional near-infrared sampling system includes: an EEG acquisition module and an fNIRS functional near-infrared brain signal acquisition module; the EEG acquisition module uses epidermal electrodes to record signals, and the electrodes are arranged according to the international 10-20 system or a customized layout; in this module, the EEG signals generated by time interference stimulation are filtered and denoised by a preprocessing circuit, and then converted into digital form by an analog-to-digital converter (ADC) and sent to the communication and control subsystem. The communication and control subsystem sends the digitally converted EEG signals to the MCU, and the digitally converted EEG signals are removed by the MCU to remove artifacts and baseline drift before being uploaded to the host computer. The fNIRS functional near-infrared brain signal acquisition module measures the activity level of specific brain areas by monitoring the absorption changes of brain hemoglobin and oxyhemoglobin in the near-infrared spectrum. The module uses a light-emitting diode as a light source to generate 650nm near-infrared light, and a photodiode as a light signal detector to convert the light signal passing through the brain tissue into an electrical signal for analysis. The obtained electrical signal is denoised and filtered through preprocessing, and then digitized by an analog-to-digital converter and sent to the communication and control subsystem. The communication and control subsystem sends the digitized electrical signal to the MCU, which calculates the brain tissue blood oxygen based on the digitized electrical signal to obtain the brain activity signal and upload the activity signal to the host computer. In this application, the signal acquisition system couples EEG and fNIRS to perform multimodal acquisition of brain signals. By combining the high temporal resolution of EEG (which can capture millisecond-level brain activity) and the high spatial resolution of fNIRS (which can identify areas of brain activity at different depths), more detailed and comprehensive brain activity information is obtained in both the temporal and spatial domains. At the same time, this method complements EEG's sensitivity to neural electrical activity and fNIRS's sensitivity to changes in cerebral blood flow and blood oxygenation, providing a more comprehensive range of brain activity information. EEG and fNIRS will be integrated into a hardware system using the same standard clock to resolve the timing accuracy issues caused by the different sampling rates of the two acquisition methods. This solution will reduce the time error of brain activity data collected by the two methods to less than 1.5ms, allowing the data collected from the two modules to be accurately located at a time node. This provides effective and comprehensive brain activity information for subsequent analysis of stimulation effects.

[0049] The dual-modality fusion model is calculated by extracting features from EEG and fNIRS signal data, combining the two into a fusion model, and then performing correlation analysis, as shown in Figure 3. The steps are as follows:

[0050] The feature extraction of EEG mainly involves extracting the response related to the target stimulus, combining the ERP (event-related potential) of the target stimulus into a data matrix, and then convolving the data matrix with the hemodynamic response function to obtain the EEG data features that can be used for subsequent correlation analysis to form the feature matrix X EEG The feature extraction of fNIRS is to construct a data matrix by sampling the blood oxygen data within 40 seconds after nerve stimulation in time-division order and sorting them by time to obtain the fNIRS data features corresponding to the EEG data features, forming the feature matrix X fNIRS .

[0051] The fusion model is constructed by calculating the maximum correlation matrix A of the two feature data sets through the CCA (canonical correlation analysis) algorithm. EEG and A fNIRS Then, the time component represented by the EEG data and the spatial component C represented by the fNIRS data are fitted by the least squares method. EEG and C fNIRS In this algorithm, the following model formula can be constructed: X i =A i C i ,i=1,2,...,2N

[0052] In this formula A i ∈K T×D , i is the number of modes, T is the number of sampling time points of the feature, V i is the characteristic variable of the characteristic matrix, and D is the minimum rank of the characteristic matrix.

[0053] Therefore, in different correlation matrices A i The multiplication is equal to the correlation coefficient as shown in the following formula:

[0054] Among them A i The vector of the jth column in is represented as Because the model is consistent with the principle of least squares, this method can be used to obtain the time component and spatial component C of the two data sets. i :

[0055] Graph convolutional neural network prediction is to predict the stimulation location based on the desired stimulation effect and formulate a stimulation plan. Graph structured data usually includes nodes and edges, where nodes represent entities and edges represent the relationship between them. In the neural stimulation model, different brain regions can be constructed as nodes, and the different connection states between brain regions can be constructed as edges. Predicting at the node level is to predict the stimulation location. It is expressed as follows:

[0056] Let G = (V; E) represent a graph network, V represents the node set, E represents the edge set, if two nodes v i ,v j There are edges connected, then (v i ,v j )∈E, the entire graph network is regarded as a network composed of nodes from different brain functional areas connected by connecting nerves, and the whole is a structure composed of subgraphs with a finite number of nodes.

[0057] The four brain waves in a certain brain region node for a fixed period of time are recorded as different dimensions, the number of dimensions is D, and the number of samples is τ. Then the input brain activity feature matrix of N brain network nodes at time t is expressed as follows: χ = (X1, X2, ..., X t ,...,X τ )∈R N×D×τ

[0058] where X t The brain activity feature matrix of N nodes in the figure at time t is expressed as follows:

[0059] For an output brain activity feature matrix with K samples, it is expressed as follows:

[0060] where Y K represents the brain activity matrix of N nodes at the Kth time point, x NK represents the brain activity of the Nth node at the Kth time point.

[0061] The prediction task uses data sampled by N brain network nodes at adjacent T times and expresses the brain activity feature prediction task as a function ψ θ , the expression is as follows: γ=ψ θ (X t-T+1 ,X t-T ,...,X t ; G).

[0062] A closed-loop multimodal neural stimulation method based on time interference comprises the following steps:

[0063] Step 1: Determine the stimulation target link, determine the planned stimulation location and the expected four brain wave characteristic signal responses, including δ, θ, α, and β; the frequency of δ is 0.5Hz-3.5Hz, the frequency of θ is 4Hz-7Hz, the frequency of α is 8Hz-13Hz, and the frequency of β is 14Hz-30Hz.

[0064] Step 2: Set the stimulation parameters. If this is the first stimulation, there is no available graph convolutional neural network feedback results at this time. It is necessary to calculate the placement of the two sets of electrodes and the current frequency based on the generation characteristics of the envelope electric field. If this is not the first stimulation, the electrode placement and frequency parameter settings are determined based on the results predicted by the graph convolutional neural network. The expression for calculating the placement of the two sets of electrodes based on the generation characteristics of the envelope electric field is as follows:

[0065] Where E is the electric field intensity, ω represents the angular frequency, J is the current density, ρ is the spatial current density, and B is the magnetic field intensity. Finite element simulation can be used to determine the distribution of the envelope electric field in the skull and determine the placement of the electrodes. represents the Hamiltonian operator, ε0 represents the vacuum permittivity, j represents an imaginary number, μ0 represents the vacuum permeability, and → represents a vector.

[0066] Step 3: Execute the stimulation phase, using two sets of electrodes placed on the scalp to generate transcranial alternating current with a slight frequency difference (such as a 20 kHz and 20.05 kHz signal pair with a 0.05 kHz difference). The two sets of currents will stimulate a low-frequency envelope electric field in a specific area of ​​the brain to activate neurons.

[0067] In step 4, dual-modal data acquisition and preprocessing, the EEG and fNIRS acquisition systems jointly collect brain signals (specifically, using the aforementioned EEG acquisition module and fNIRS functional near-infrared brain signal acquisition module). After preprocessing, the collected signals are packaged according to the corresponding sampling time. The process is as follows:

[0068] S41. Turn on the near-infrared light source and capture electrical signals and optical signals from the sensor at the same time to obtain EEG and fNIRS signals.

[0069] S42, performing analog-to-digital conversion on the acquired signal after removing noise and filtering out clutter.

[0070] S43. The data is aggregated into the data acquisition system MCU (core chip of the data acquisition system) for artifact removal and baseline drift removal processing.

[0071] S44. Pack the processed EEG and fNIRS data according to a unified clock, and send the data packets in time sequence to the upper control system.

[0072] Step 5: Data processing and correlation analysis. The upper control system extracts features from the two modal data and performs correlation analysis to obtain temporal and spatial component features. The process is as follows:

[0073] S51. Extract data features from EEG and fNIRS data.

[0074] S52. Use the CCA algorithm to perform correlation analysis on the extracted features to obtain time component features and spatial component features.

[0075] S53. Use the temporal component features and the spatial component features for cross-validation and summarize the bimodal fusion data.

[0076] Step 6: Determine whether the stimulation has achieved the expected goal by analyzing the four brain wave signals and physiological reactions of the stimulated person.

[0077] Step 7: Graph Convolutional Neural Network prediction phase: Use the previous stimulation dataset and the stimulation dataset collected during the stimulation to train the graph convolutional neural network prediction model, and use the four brain waves that are expected to be the characteristics of brain activity as the starting point to reversely predict the original stimulated points and activation status.

[0078] Step 8: Stimulation scheme optimization phase, using the graph convolutional neural network to predict the stimulated position and activation degree to adjust the electrode placement and the generated envelope electric field frequency.

[0079] Therefore, this application adopts the above-mentioned closed-loop multimodal neural stimulation system and method based on time interference, uses time interference stimulation technology to perform precise non-invasive stimulation on brain nerves; uses EEG coupled fNIRS dual-modal brain signal acquisition scheme to obtain comprehensive and accurate brain activity data; constructs EEG and fNIRS data feature fusion model and uses graph neural network prediction model to guide the formulation of stimulation plan, which can continuously optimize its effect during use.

[0080] The technical features of the above embodiments can be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0081] This document uses specific examples to illustrate the principles and implementation methods of this application. The description of the above examples is only intended to help understand the method and core concept of this application. At the same time, for those skilled in the art, based on the concept of this application, there may be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting this application.

Claims

1. A closed-loop multimodal neural stimulation system based on time interference, characterized in that: include: Temporal interferometric stimulation system, EEG-functional near-infrared sampling system and upper control system; the temporal interferometric stimulation system uses the beat electric field generated by two sets of electrodes to precisely stimulate designated areas of the brain; the EEG-functional near-infrared sampling system is a dual-modal collector that couples EEG and functional near-infrared, which includes signal extraction and correlation analysis. It analyzes the stimulation effect and adjusts the stimulation plan by integrating brain signal data with EEG temporal precision and fNIRS spatial precision; the upper control system includes dual-modal fusion model calculation, graph convolutional neural network prediction and stimulation plan formulation.

2. A closed-loop multimodal neural stimulation system based on time interference according to claim 1, characterized in that: The time interference stimulation system includes: a communication control module, a power supply module and a time interference stimulation module; the communication control module adopts a SoC chip, which integrates the functions of a processor and wireless communication internally, and is used for communication between the time interference stimulation system and the upper control system and control of the stimulation current; the power supply module adopts a design with a low-noise and low-voltage drop voltage regulator plus a shielding layer; the time interference stimulation module is composed of two difference beat electrode groups, and two sine wave generators generate two 2KHz AC signals with a phase difference of 10Hz-20Hz. The two AC signals interfere with each other in the skull through the electrode group attached to the surface of the scalp, and the low-frequency envelope current generated by the interference stimulates the nerve cells in the designated area.

3. A closed-loop multimodal neural stimulation system based on time interference according to claim 1, characterized in that: The EEG-functional near-infrared sampling system includes an EEG acquisition module and an fNIRS functional near-infrared brain signal acquisition module. The EEG acquisition module uses epidermal electrodes for signal recording. In the EEG acquisition module, the EEG signals generated by time-interference stimulation are filtered and denoised by a preprocessing circuit, then converted into digital form by an analog-to-digital converter and sent to the communication and control subsystem. The EEG signals are de-artifacted and baseline drifted by an MCU before being uploaded to the host control system. The fNIRS functional near-infrared brain signal acquisition module measures the activity level of specific brain regions by monitoring the absorption changes of brain hemoglobin and oxyhemoglobin within the near-infrared spectrum. The fNIRS functional near-infrared brain signal acquisition module uses a light-emitting diode as a light source to generate 650nm near-infrared light. A photodiode is used as an optical signal detector to convert the optical signal passing through the brain tissue into an electrical signal for analysis. The resulting electrical signal is de-noised and filtered by preprocessing, then converted into digital form by an analog-to-digital converter and sent to the communication and control subsystem. The MCU calculates brain tissue blood oxygen to obtain brain activity signals and uploads the activity signals to the host control system.

4. The closed-loop multimodal neural stimulation system based on time interference according to claim 1, characterized in that: The dual-modal fusion model was calculated by extracting features from EEG and fNIRS signal data and combining the two into a fusion model followed by correlation analysis.

5. The closed-loop multimodal neural stimulation system based on time interference according to claim 1, characterized in that: The graph convolutional neural network prediction is to predict the stimulation location according to the desired stimulation effect and formulate a stimulation plan, which is expressed as follows: Let G = (V; E) represent a graph network, V represents the node set, E represents the edge set, if two nodes v i ,v j There are edges connected, then (v i ,v j )∈E, the entire graph network is considered to be composed of nodes composed of different brain functional areas connected by connecting nerves, and the whole is a structure composed of subgraphs with a finite number of nodes; Here, the four brain waves of a certain brain region node in a fixed period of time are recorded as different dimensions, the number of dimensions is D, and the number of samples is τ. Then the input brain activity feature matrix of N brain network nodes at time t is expressed as follows: x=(X1,X2,...,X t ,...,X τ )∈R N×D×τ where X t The brain activity feature matrix of N nodes in the figure at time t is expressed as follows: For an output brain activity feature matrix with K samples, it is expressed as follows: where Y K represents the brain activity matrix of N nodes at the Kth time point; The prediction task uses data sampled by N brain network nodes at adjacent T times and expresses the brain activity feature prediction task as a function ψ θ , the expression is as follows: c = ψ θ (X t-T+1 ,X t-T ,...,X t (G)。 6. A closed-loop multimodal neural stimulation method based on time interference, characterized in that: The following steps are involved: Step 1: Determine the stimulation target, the planned stimulation location, and the expected four brainwave characteristic signal responses, including δ, θ, α, and β. The frequency of δ is 0.5Hz-3.5Hz, the frequency of θ is 4Hz-7Hz, the frequency of α is 8Hz-13Hz, and the frequency of β is 14Hz-30Hz. Step 2: Setting stimulation parameters. If this is the first stimulation, there is no available graph convolutional neural network feedback results at this time. The placement of the two sets of electrodes and the current frequency need to be calculated based on the characteristics of the envelope electric field. If this is not the first stimulation, the electrode placement and frequency parameter settings are determined based on the results predicted by the graph convolutional neural network. Step 3: Execute the stimulation phase, using two sets of electrodes placed on the scalp to generate transcranial alternating current with a slight frequency difference. The two sets of currents will stimulate a low-frequency envelope electric field in a specific area of ​​the brain to activate neurons; Step 4: Dual-modal data acquisition and preprocessing: EEG and fNIRS systems jointly collect brain signals, preprocess the collected signals, and package them according to the corresponding sampling time; Step 5: Data processing and correlation analysis: The upper control system extracts features from the two modal data and performs correlation analysis to obtain temporal and spatial component features. Step 6: Determine whether the stimulation has achieved the expected goal by analyzing the collected four brain wave signals and physiological responses of the stimulated person; Step 7: Graph Convolutional Neural Network prediction phase: Use the previous stimulation dataset and the stimulation dataset collected during the stimulation to train the graph convolutional neural network prediction model, and use the four brain waves characteristic of expected brain activity as the starting point to reversely predict the original stimulated points and activation status; Step 8: Stimulation scheme optimization phase, using the graph convolutional neural network to predict the stimulated position and activation degree to adjust the electrode placement and the generated envelope electric field frequency.

7. The closed-loop multimodal neural stimulation method based on time interference according to claim 6, characterized in that: In step 2, the expression for calculating the placement positions of the two sets of electrodes based on the generation characteristics of the envelope electric field is as follows: Where E is the electric field intensity, J is the current density, ρ is the spatial current density, and B is the magnetic field intensity. Finite element simulation can be used to determine the distribution of the envelope electric field in the skull and determine the placement of the electrodes. represents the Hamiltonian operator, ε0 represents the vacuum permittivity, j represents an imaginary number, μ0 represents the vacuum permeability, and → represents a vector.

8. The closed-loop multimodal neural stimulation method based on time interference according to claim 7, characterized in that: The process of dual-modal data collection and preprocessing in step 4 is as follows: S41, turning on the near-infrared light source to simultaneously capture electrical and optical signals from the sensor to obtain EEG and fNIRS signals; S42, performing analog-to-digital conversion on the acquired signal after removing noise and filtering out clutter; S43, collecting the data to the data acquisition system MCU for artifact removal and baseline drift removal; S44. Pack the processed EEG and fNIRS data according to a unified clock, and send the data packets in time sequence to the upper control system.

9. The closed-loop multimodal neural stimulation method based on time interference according to claim 8, characterized in that: The process of data processing and correlation analysis in step 5 is as follows: S51, extract data features from EEG and fNIRS data; S52, using the CCA algorithm to perform correlation analysis on the extracted features to obtain time component features and spatial component features; S53. Use the temporal component features and the spatial component features for cross-validation and summarize the bimodal fusion data.

Images