Safety Analysis Methods for Brain-Computer Interface and Spinal Cord Nerve Modulation Device Implantation

Abstract

This invention discloses a method for analyzing the safety of device implantation under brain-computer interface and spinal cord nerve modulation, involving the interdisciplinary fields of biomedical engineering and neural engineering. This invention constructs a coupled dynamics model describing the pathways of the brain's motor cortex, spinal motor neurons, and muscle effectors. This model takes BCI decoding commands and actual acquired multimodal physiological signals as input, continuously optimizes through parameter identification, calculates the difference between the actual spinal cord response and the expected response generated by the model, compares the predicted and actual values ​​of the coupled dynamics model in real time, dynamically calculates the risk value, and immediately identifies the dominant risk factor once it exceeds a threshold, triggering a highly specific adaptive adjustment mechanism accordingly. This ensures that the BCI-SCN system always operates within a safe range in the complex and ever-changing human neural environment, accurately tracing macroscopically observed abnormalities to specific causes at the microscopic level, providing a basis for subsequent precise regulation.

Description

Technical Field

[0001] This invention relates to the field of interdisciplinary technology of biomedical engineering and neural engineering, specifically to a method for analyzing the safety of implanted devices under brain-computer interface and spinal cord nerve modulation. Background Technology

[0002] With the deep integration of neural engineering and rehabilitation medicine, the combined application of brain-computer interface (BCI) and spinal cord neuromodulation (SCN) has brought revolutionary treatment hope to patients with motor dysfunction caused by spinal cord injury, stroke, etc. BCI technology aims to decode the patient's brain's motor intentions and translate them into control commands, while implanted spinal cord stimulators activate the central pattern generator (CPG) or residual neural pathways through electrical stimulation, thereby triggering gait or motor movements downstream of BCI commands. This "BCI-SCN" constitutes a new type of human-computer interaction neural circuit, which aims to bypass damaged neural pathways and rebuild the functional connection between the brain and limbs.

[0003] However, this deep integration of cross-segmental and multimodal neural interfaces also introduces unprecedented safety challenges. Existing technologies have significant shortcomings. First, current safety analyses are mostly limited to single devices or single signal modalities, such as focusing only on the decoding accuracy of the BCI or the safety threshold of the SCN stimulation itself. There is a lack of methods to model and analyze the brain and spinal cord as a whole coupled system. This makes it impossible to warn of risks caused by conflicts in ascending and descending neural pathway instructions, abnormal signal transmission, or desynchronization of regulation. Second, existing risk assessments are mostly static and offline analyses. They cannot provide dynamic warnings of rapidly evolving neural conflicts (such as abnormal high-frequency oscillations and unexpected reflexes) during the real-time process of device implantation or stimulation application. Once a risk is identified, it cannot automatically and accurately trigger a matching regulatory strategy, resulting in lag and uncertainty.

[0004] Therefore, there is an urgent need in this field for a novel safety analysis method specifically for the "BCI-SCN" integrated system. This method should be able to model the dynamic coupling relationship between the brain and spinal cord, quantify the synergy and conflict of neural function linkages in real time, and on this basis form a safety control system that can make autonomous decisions and intervene dynamically, thereby ensuring the high safety and reliability of this cutting-edge neuromodulation therapy in clinical applications. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of existing technologies and provide a method for analyzing the safety of implanted devices under brain-computer interface and spinal cord nerve modulation. This method constructs a coupled dynamic model describing the pathways of the brain's motor cortex, spinal motor neurons, and muscle effectors. This model takes BCI decoding commands and actual acquired multimodal physiological signals as input, continuously optimizes through parameter identification, calculates the difference between the actual spinal cord response and the expected response generated by the model, compares the predicted and actual values ​​of the coupled dynamic model in real time, dynamically calculates the risk value, and immediately identifies the dominant risk factor once it exceeds a threshold, triggering a highly specific adaptive adjustment mechanism. This ensures that the BCI-SCN system always operates within a safe range in the complex and ever-changing human neural environment, accurately tracing macroscopically observed abnormalities to specific microscopic causes, providing an irreplaceable decision-making basis for subsequent precise regulation.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution: On the one hand, a method for analyzing the safety of device implantation under brain-computer interface and spinal cord nerve modulation, the specific steps of which are as follows:

[0007] S100, Multimodal Physiological Data Acquisition and Synchronization: Brain nerve electrical signals are acquired through an implantable brain-computer interface electrode array, and spinal cord nerve electrical signals and local electromyographic signals are acquired through a modulator electrode implanted in the epidural and subcutaneous regions of the spinal cord. Simultaneously, vital signs such as heart rate, blood pressure, and blood oxygen saturation data are recorded to construct a spatiotemporally synchronized multimodal physiological dataset.

[0008] S200, Neural Function Linkage and Conflict Modeling: Based on the multimodal physiological dataset, establish a coupled dynamic model of the brain-spinal cord neural pathway, quantitatively analyze the synergy and conflict between the expected spinal cord response triggered by brain-computer interface commands and the actual recorded spinal cord response, and identify abnormal linkage patterns.

[0009] S300, Dynamic assessment of security risks: Based on the identification results of the aforementioned collaborative, conflict and abnormal linkage patterns, a risk assessment algorithm is introduced to calculate short-term and long-term risk values ​​in real time.

[0010] S400, Multi-level Early Warning and Adaptive Control Trigger: Set safety thresholds corresponding to the short-term risk value and the long-term risk value. When the risk assessment result exceeds any safety threshold, generate a corresponding abnormal signal and trigger the adaptive adjustment mechanism of the implantable control device accordingly.

[0011] S500, Safety Closed-Loop Verification and Learning Optimization: After performing adaptive regulation, physiological data is collected again, and the characteristic differences of spinal cord response signals and electromyographic signals caused by the same brain-computer interface commands before and after adaptive regulation are compared. When the abnormal linkage pattern disappears or the risk value falls back to the safe range, the relief is deemed effective.

[0012] Furthermore, in S100, the implantable brain-computer interface electrode array is a 16-channel microelectrode array, implanted in the motor cortex and sensory cortex of the brain, for collecting brain neural electrical signals, the brain cortex neural electrical signals including cortical electroencephalogram and neural unit discharge signals;

[0013] The modulator electrode implanted in the epidural space of the spinal cord is an 8-channel cylindrical electrode, and the implantation segment matches the diseased spinal cord segment. The subcutaneous modulator electrode is a 4-channel sheet electrode, which is embedded in the subclavian subcutaneous tissue. The two work together to collect spinal cord nerve electrical signals and local electromyographic signals. The spinal cord nerve electrical signals and local electromyographic signals include composite action potentials and local field potentials recorded in the epidural space of the spinal cord.

[0014] The spatiotemporal synchronization of the multimodal physiological dataset is achieved by assigning a uniform timestamp to all signal sources.

[0015] Furthermore, in S200, the process of constructing the coupled dynamics model is as follows:

[0016] Data preprocessing and feature extraction: The multimodal physiological dataset described in S100 is preprocessed, including using bandpass filtering to extract the effective components of the neural electrical signals, using wavelet transform to remove motion artifacts in the electromyographic signals, and normalizing all signals. Time-domain features and frequency-domain features are extracted from the preprocessed signals. The time-domain features include peak amplitude, latency, and area under the curve, and the frequency-domain features include power spectral density and dominant frequency components.

[0017] Coupled Dynamics Model Construction: Based on preprocessed feature data, a coupled dynamics model describing the interaction between brain and spinal cord signals is established. This model is defined by a set of coupled differential equations, and the model input is directly related to the extracted features. The set of coupled differential equations is as follows: ,in, for The state vector of brain neural signals at any given time is formed by integrating the extracted time-domain and frequency-domain features of the brain's electrical neural signals. for The spinal cord nerve signal state vector at time t is formed by integrating the extracted time-domain and frequency-domain features of the spinal cord nerve electrical signals. The brain signal autodynamic parameter matrix, elements Indicates the first The brain signal features of the first The self-regulation strength of individual brain signal features The spinal cord signal autodynamic parameter matrix, elements Indicates the first The first spinal cord signal feature for the first The self-regulation intensity of individual spinal cord signal features This is the coupling strength matrix from the brain to the spinal cord, with elements... Indicates the first The brain signal features of the first The intensity of modulation of individual spinal cord signal features, This is the coupling strength matrix from the spinal cord to the brain, with elements... Indicates the first The first spinal cord signal feature for the first The feedback intensity of individual brain signal features, This is a nonlinear transformation activation function from the brain to the spinal cord, used to describe the nonlinear modulation effect of brain signals on spinal cord signals. This is a nonlinear transformation activation function from the spinal cord to the brain, used to describe the nonlinear feedback effect of spinal cord signals on brain signals. This is due to the time delay in conduction from the brain to the spinal cord. This is due to the time delay in conduction from the spinal cord to the brain. This is the external control input vector for brain signals, and its value is the command signal strength of the brain-computer interface in S100. This is the external control input vector for spinal cord signals, and its value is the stimulation signal intensity of the spinal cord modulator in S100. The brain signal input gain matrix has elements. Indicates external control input For the The gain coefficient of each brain signal feature ensures that the control input matches the amplitude of the brain signal feature. The spinal cord signal input gain matrix has elements. Indicates external control input For the The gain coefficients of each spinal cord signal feature are used to ensure that the control input matches the amplitude of the spinal cord signal feature. This is noise in the brain signaling system. This is noise in the spinal cord signaling system;

[0018] Model parameter identification and optimization: Parameter matrix in coupled dynamics models Time delay and input gain matrix Online identification and real-time optimization are performed based on the initialization parameter matrix. Time delay and input gain matrix , Using the identity matrix, the extracted real-time feature data is used as the observation value. Through the EKF prediction step and update step, the parameter estimate is updated every 50ms to minimize the mean square error between the model output and the measured data.

[0019] Furthermore, S200 calculates the synergy index CI and the conflict coefficient CC based on the identified and optimized model parameters, wherein the synergy index CI is used to quantify the degree of synchronization and synergy between brain and spinal cord signals. ,in, Brain signal state vector With spinal cord signal state vector The cross-correlation matrix, elements reflect and The degree of linear correlation is represented by Cov(·), where Cov(·) is the covariance, Var(·) is the variance, eig(·) is the eigenvalue of the cross-correlation matrix (larger eigenvalues ​​indicate stronger synergy in that direction), and max(·) is the synergy index, where the largest eigenvalue is used. The synergy was deemed good. The system indicates an abnormality in coordination.

[0020] The conflict coefficient CC is used to quantify the degree of conflict between the actual spinal cord response and the expected response. ,in, This is the feature vector of the actual recorded spinal cord nerve signals. The spinal cord response vector is predicted based on a coupled dynamics model, i.e., through the model. Calculations show that The Euclidean distance between the actual response and the expected response reflects the degree of deviation between the two. The standardization parameter is set to the standard deviation of the spinal cord signal characteristics under the S100 resting state, which is used to eliminate the influence of signal amplitude differences on conflict determination. The closer it is to 1, the more severe the conflict. The system indicates a significant conflict.

[0021] Furthermore, S200 is based on the calculated synergy index. Conflict coefficient Based on the S100 vital signs data, the following abnormal linkage patterns are identified, and these abnormal patterns are directly used as input to the S300 risk assessment algorithm:

[0022] Instruction propagation delay exceeds limit: When the identification time delay exceeds the limit. or Exceeding its normal physiological range threshold When this occurs, it is determined that the instruction transmission delay has exceeded the limit;

[0023] Signal gain anomaly: when the coupling strength matrix or norm Exceeding its normal range When this occurs, it is determined to be an abnormal signal gain;

[0024] Unintended spinal reflexes: when given commands via the S100 brain-computer interface In this case, spinal cord signals Amplitude > The rhythmic discharges were determined to be unexpected spinal reflexes.

[0025] Furthermore, based on the calculated synergy index CI, conflict coefficient CC, and abnormal linkage mode, and combined with the vital sign parameters collected by S100, the S300 constructs a multidimensional risk feature vector F=[CI, CC, Δτ, ΔG, ΔH, ΔR]T, which is used to accurately characterize the risk state in the neural modulation process. Δτ is the maximum deviation of the conduction time delay, defined as Δτ=max(|τ1- |,|τ2- |), The nominal value is within the normal physiological range. The maximum deviation of the norm of the coupling strength matrix is ​​defined as follows: ), Let F be the norm of the matrix. This is the nominal value of the coupling strength. The deviation of heart rate from the baseline is defined as follows: ,in To collect heart rate in real time, This is the baseline value of resting heart rate. The deviation of blood pressure from the baseline is defined as... ,in To collect blood pressure in real time, This is the baseline value for resting blood pressure;

[0026] The risk feature vector F is input into the risk assessment model, and an activation function is applied. Mapping outputs short-term risk values and long-term risk value Quantifying the immediate and chronic risks of neuromodulation, the aforementioned , ,in, , For the weight vector, This is a bias term used to correct the model output offset.

[0027] Furthermore, in S400, the following settings are provided: , To preset a safety threshold, when or For risk feature vectors The contribution of each component in the model is quantitatively analyzed, and the contribution of each feature is calculated. Marginal contribution to risk value Identify the dominant risk factors:

[0028] when CI) CC) contributed the most, identifying neural signal conflict as the dominant risk, generating a diagnostic report of abnormal signal transduction pathways, and providing suggestions to check the integrity of neural pathways:

[0029] when , It has the highest contribution rate, and the dominant risk is determined to be abnormal brain-computer interface commands. It generates a diagnostic report of abnormal command encoding and provides an identifier of command parameters to be adjusted for the adaptive adjustment mechanism.

[0030] when , It has the highest contribution rate, and the dominant risk is determined to be inappropriate spinal cord modulation stimulation parameters. It generates a stimulation parameter optimization diagnostic report and provides the types and directions of stimulation parameters to be adjusted for the adaptive regulation mechanism in S400.

[0031] On the other hand, a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the method for analyzing the safety of device implantation under brain-computer interface and spinal cord nerve modulation.

[0032] Furthermore, an electronic device includes:

[0033] Memory, used to store computer programs;

[0034] A processor is used to execute computer programs stored in the memory to implement a method for analyzing the safety of implantation of devices under brain-computer interface and spinal cord nerve modulation.

[0035] Compared with existing technologies, this method for analyzing the safety of brain-computer interface and spinal cord nerve modulation device implantation has the following advantages:

[0036] I. This invention constructs a coupled dynamics model describing the pathways of the brain's motor cortex, spinal motor neurons, and muscle effectors. This model takes BCI decoding commands and actual acquired multimodal physiological signals as inputs, continuously optimizes through parameter identification, calculates the difference between the actual spinal cord response and the expected response generated by the model, compares the predicted values ​​of the coupled dynamics model with the actual values ​​in real time, dynamically calculates risk values, and immediately locks down the dominant risk factors once a threshold is exceeded, thereby triggering a highly specific adaptive adjustment mechanism. This ensures that the BCI-SCN system always operates within a safe range in the complex and ever-changing human neural environment, accurately tracing macroscopically observed abnormalities to specific causes at the microscopic level, providing an irreplaceable decision-making basis for subsequent precise regulation.

[0037] Other advantages, objectives and features of the invention will be set forth in part in the description which follows, and in part will be apparent to those skilled in the art from the following examination or study, or may be learned from the practice of the invention. Attached Figure Description

[0038] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without any creative effort.

[0039] Figure 1 Flowchart of the operation method for safety analysis of implantation of brain-computer interfaces and spinal cord nerve modulation devices;

[0040] Figure 2 A flowchart illustrating the steps of a method for analyzing the safety of implanted devices under brain-computer interface and spinal cord nerve modulation.

[0041] Figure 3 This is a schematic diagram of the structure of a computer electronic device. Detailed Implementation

[0042] To further illustrate the technical means and effects of the present invention in achieving its intended purpose, the following detailed description of the specific implementation methods, structures, features, and effects of the present invention, in conjunction with the accompanying drawings and preferred embodiments, is provided below.

[0043] Example 1

[0044] This embodiment provides a specific implementation of a method for analyzing the safety of device implantation under brain-computer interface and spinal cord nerve modulation, such as... Figure 2 As shown, this method achieves full-process safety monitoring of the device implantation and control process through multimodal physiological signal synchronous acquisition, coupled neurodynamic modeling, real-time risk quantification assessment, multi-level safety early warning and closed-loop adaptive control, ensuring the stable and safe operation of the implanted device in a complex neural interaction environment.

[0045] First, the multimodal physiological data acquisition and synchronization phase (S100) begins. This phase simultaneously acquires multidimensional signals from the cerebral cortex, spinal cord segments, and peripheral physiological systems, assigning them a unified time reference to provide a high-quality, spatiotemporally aligned data foundation for subsequent analysis. Specifically, a 16-channel microelectrode array is implanted in the primary motor cortex and somatosensory cortex of the patient's brain to acquire electrocorticography (ECoG) signals and neuronal firing (spike) signals. Simultaneously, an 8-channel cylindrical electrode is implanted epidurally in the spinal cord segment corresponding to the lesion, and a 4-channel sheet electrode is implanted subcutaneously below the clavicle to jointly acquire spinal compound action potentials (CAP), local field potentials (LFP), and surface electromyography (sEMG) signals. Furthermore, heart rate (HR), blood pressure (BP), and oxygen saturation (SpO2) are simultaneously acquired via a patient monitor. 2）等生命体征信号，所有信号在采集时通过硬件时钟打上统一微秒级精度的时间戳，构成多模态生理数据集，该数据集不仅包含原始的电压时间序列，也初步提取出各信 号的幅值、频率等基础特征，并存入带时间对齐标签的数据缓存区，供下一阶段调用。

[0046] Then, the modeling stage of neural functional linkage and conflict (S200) begins. This stage constructs a coupled dynamic model of the brain-spinal cord-muscle information transmission process and quantitatively analyzes the synergy and conflict in neural functional linkage based on this model. The raw signals from S100 are preprocessed: neural electrical signals undergo bandpass filtering from 0.5 to 3000 Hz to remove low-frequency drift and high-frequency noise; electromyographic signals are processed using wavelet transform to remove motion artifacts. All signals are amplitude normalized to ensure they are of the same magnitude. Time-domain and frequency-domain features for modeling are extracted from the preprocessed signals: time-domain features include peak amplitude, latency, and area under the curve; frequency-domain features include power spectral density of each frequency band and the dominant frequency component of the signal. These features are then combined to form a brain signal state vector. and spinal cord signal state vector Based on eigenvectors, a mathematical model describing the interaction between brain signals and spinal cord signals is established. This model is defined by a system of coupled differential equations, namely... ,in, for The state vector of brain neural signals at any given time is formed by integrating the extracted time-domain and frequency-domain features of the brain's electrical neural signals. for The spinal cord nerve signal state vector at time t is formed by integrating the extracted time-domain and frequency-domain features of the spinal cord nerve electrical signals. The brain signal autodynamic parameter matrix, elements Indicates the first The brain signal features of the first The self-regulation strength of individual brain signal features The spinal cord signal autodynamic parameter matrix, elements Indicates the first The first spinal cord signal feature for the first The self-regulation intensity of individual spinal cord signal features This is the coupling strength matrix from the brain to the spinal cord, with elements... Indicates the first The brain signal features of the first The intensity of modulation of individual spinal cord signal features, This is the coupling strength matrix from the spinal cord to the brain, with elements... Indicates the first The first spinal cord signal feature for the first The feedback intensity of individual brain signal features, This is a nonlinear transformation activation function from the brain to the spinal cord, used to describe the nonlinear modulation effect of brain signals on spinal cord signals. This is a nonlinear transformation activation function from the spinal cord to the brain, used to describe the nonlinear feedback effect of spinal cord signals on brain signals. This is due to the time delay in conduction from the brain to the spinal cord. This is due to the time delay in conduction from the spinal cord to the brain. This is the external control input vector for brain signals, and its value is the command signal strength of the brain-computer interface in S100. This is the external control input vector for spinal cord signals, and its value is the stimulation signal intensity of the spinal cord modulator in S100. The brain signal input gain matrix has elements. Indicates external control input For the The gain coefficient of each brain signal feature ensures that the control input matches the amplitude of the brain signal feature. The spinal cord signal input gain matrix has elements. Indicates external control input For the The gain coefficients of each spinal cord signal feature are used to ensure that the control input matches the amplitude of the spinal cord signal feature. This is noise in the brain signaling system. For noise in the spinal cord signaling system, the parameter matrix in the coupled dynamics model... Time delay and input gain matrix Online identification and real-time optimization are performed based on the initialization parameter matrix. Time delay and input gain matrix , Using the identity matrix, the extracted real-time feature data is used as the observation value. Through EKF prediction and update steps, the parameter estimates are updated every 50ms to minimize the mean square error between the model output and the measured data. Based on the identified and optimized model parameters, the coordination index CI and the conflict coefficient CC are calculated. The coordination index CI is used to quantify the degree of synchronization and coordination between brain and spinal cord signals. Where cross_corr(x, y) is the brain signal state vector. With spinal cord signal state vector The cross-correlation matrix, elements reflect and The degree of linear correlation is represented by Cov(·), where Cov(·) is the covariance, Var(·) is the variance, eig(·) is the eigenvalue of the cross-correlation matrix (larger eigenvalues ​​indicate stronger synergy in that direction), and max(·) is the synergy index, where the largest eigenvalue is used. The synergy was deemed good. The system indicates an abnormality in coordination.

[0047] The conflict coefficient CC is used to quantify the degree of conflict between the actual spinal cord response and the expected response. ,in, This is the feature vector of the actual recorded spinal cord nerve signals. The spinal cord response vector is predicted based on a coupled dynamics model, i.e., through the model. Calculations show that The Euclidean distance between the actual response and the expected response reflects the degree of deviation between the two. The standardization parameter is set to the standard deviation of the spinal cord signal characteristics under the S100 resting state, which is used to eliminate the influence of signal amplitude differences on conflict determination. The closer it is to 1, the more severe the conflict. The indicator suggests a significant conflict, based on the calculated synergy index. Conflict coefficient Based on the S100 vital signs data, the following abnormal linkage patterns are identified, and these abnormal patterns are directly used as input to the S300 risk assessment algorithm:

[0048] Instruction propagation delay exceeds limit: When the identification time delay exceeds the limit. or Exceeding its normal physiological range threshold When this occurs, it is determined that the instruction transmission delay has exceeded the limit;

[0049] Signal gain anomaly: when the coupling strength matrix or norm Exceeding its normal range When this occurs, it is determined to be an abnormal signal gain;

[0050] Unintended spinal reflexes: when given commands via the S100 brain-computer interface In this case, spinal cord signals Amplitude > The rhythmic discharges were determined to be unexpected spinal reflexes.

[0051] Next, the dynamic safety risk assessment phase (S300) begins. Based on the coupling model established in S200 and its output, this phase quantifies the real-time risk of the current neural modulation process, constructing a multidimensional risk feature vector F=[CI, CC, Δτ, ΔG, ΔH, ΔR]T, where Δτ is the maximum deviation of the conduction time delay, defined as Δτ=max(|τ1- |,|τ2- |), The nominal value is within the normal physiological range. The maximum deviation of the norm of the coupling strength matrix is ​​defined as follows: ), Let F be the norm of the matrix. This is the nominal value of the coupling strength. The deviation of heart rate from the baseline is defined as follows: ,in To collect heart rate in real time, This is the baseline value of resting heart rate. The deviation of blood pressure from the baseline is defined as... ,in To collect blood pressure in real time, The baseline resting blood pressure value is used as the basis for inputting the risk feature vector F into the risk assessment model, which is then activated by the activation function. Mapping outputs short-term risk values and long-term risk value Quantifying the immediate and chronic risks of neuromodulation, the aforementioned , ,in, , For the weight vector, This is a bias term used to correct the model output offset.

[0052] Secondly, the system enters the multi-level early warning and adaptive control triggering phase (S400), where two levels of risk thresholds are preset: short-term risk threshold. and long-term risk threshold ,once or This triggers risk attribution analysis: calculating the risk feature vector. Marginal contribution of each component To identify the dominant risk factors, when CI) CC) contributes the most, determining the dominant risk as neural signal conflict, generating a diagnostic report of abnormal signal transduction pathways, and providing suggestions to check the integrity of neural pathways. , The highest contribution was made, and the dominant risk was determined to be abnormal brain-computer interface commands. A diagnostic report of command encoding abnormalities was generated, and identifiers of command parameters to be adjusted were provided for the adaptive adjustment mechanism. , It has the highest contribution rate, and the dominant risk is determined to be inappropriate spinal cord modulation stimulation parameters. It generates a stimulation parameter optimization diagnostic report and provides the types and directions of stimulation parameters to be adjusted for the adaptive regulation mechanism in S400.

[0053] Finally, the system enters the safety closed-loop verification and learning optimization stage (S500). After the control action is executed, a new segment of physiological data is collected, features are extracted again, and the risk value is calculated. By comparing the changes in the risk value before and after the control and whether the abnormal pattern has disappeared, the effectiveness of the control measures is verified. All data related to this decision—including the original signal, model parameters, risk value, control action, and verification results—are recorded and stored in the historical database. Using a reinforcement learning framework, with the risk reduction as the reward signal, the model parameter identifier in S200 and the risk assessment network in S300 are fine-tuned and optimized, thereby forming a safety control system with continuous learning capabilities.

[0054] In summary, this embodiment provides a complete implementation plan from signal acquisition, modeling and analysis, risk assessment to decision control. Through multimodal signal fusion and coupled modeling, it achieves a fine characterization of the linkage state of neural functions. Relying on real-time risk quantification and source tracing mechanisms, it achieves precise and adaptive safety control, significantly enhancing the safety, reliability and intelligence of the combined application of brain-computer interface and spinal cord nerve modulation.

[0055] Example 2

[0056] like Figure 1 As shown in Example 1, this example elaborates on the specific steps of the device implantation safety analysis method under brain-computer interface and spinal cord nerve modulation. The specific steps are as follows:

[0057] Multimodal physiological data acquisition and synchronization

[0058] By implanting a 16-channel microelectrode array into the motor and sensory cortex of the brain, cortical electroencephalogram (EEG) signals and neuronal firing signals are collected.

[0059] By using an 8-channel cylindrical electrode implanted in the epidural space of the affected spinal cord segment and a 4-channel sheet electrode implanted under the skin of the subclavian bone, composite action potentials, local field potentials, and local electromyographic signals of the spinal cord are collected in synergy.

[0060] Simultaneously record the patient's vital signs data such as heart rate, blood pressure, and blood oxygen saturation.

[0061] All collected signals are stamped with a unified high-precision timestamp to construct a spatiotemporally aligned multimodal physiological dataset.

[0062] Neural Functional Linkage and Conflict Modeling

[0063] Data preprocessing: Bandpass filtering was performed on the acquired neural electrical signals to extract the effective components; wavelet transform was performed on the electromyographic signals to remove motion artifacts; and normalization was performed on all signals.

[0064] Feature extraction: Extract time-domain features (such as peak amplitude, latency, and area under the curve) and frequency-domain features (such as power spectral density and dominant frequency component) from the preprocessed signal.

[0065] Coupled Dynamics Model Construction: Utilizing extracted features, a mathematical model is constructed to describe the interaction between brain and spinal cord signals. This model can simulate the process of brain commands being transmitted to the spinal cord and triggering responses, as well as the feedback from the spinal cord to the brain.

[0066] Model parameter identification and optimization: An estimation algorithm is used to identify and dynamically optimize key parameters in the model (such as connection strength and signal transmission delay) in real time, so that the model's predicted output is as close as possible to the actual observed physiological signals.

[0067] Quantitative Analysis of Collaboration and Conflict

[0068] Calculate the synergy index: By analyzing the statistical correlation between brain signals and spinal cord signals, a synergy index is calculated to quantify the degree of synchronization and coordination between their neural activities.

[0069] Calculate the conflict coefficient: Compare the ideal spinal cord response predicted by the model with the actual recorded response to calculate a conflict coefficient, which is used to quantify the degree of deviation or anomaly between the two.

[0070] Abnormal linkage pattern recognition

[0071] Based on the calculated synergy index, conflict coefficient, and vital sign data, the system automatically identifies whether a preset abnormal pattern exists.

[0072] The main abnormal patterns identified include: excessive command transmission delay (signal transmission is too slow), abnormal signal gain (modulation is too strong or too weak), and unexpected spinal reflexes (abnormal activity of the spinal cord without brain command).

[0073] Dynamic assessment of security risks

[0074] Constructing a risk feature vector: Combining the synergy index, conflict coefficient, conduction delay bias, coupling strength bias, and real-time changes in heart rate and blood pressure into a comprehensive risk feature vector.

[0075] Calculate the risk value: Input the risk feature vector into an assessment model to calculate a value representing short-term acute risk and a value representing long-term chronic risk in real time.

[0076] Multi-level early warning and adaptive control

[0077] Set safety thresholds: Set warning thresholds and danger thresholds for short-term and long-term risk values ​​respectively.

[0078] Risk source tracing and decision-making: When any risk value exceeds the threshold, the system automatically analyzes which risk feature contributes the most, thereby determining the root cause of the risk (whether it is abnormal instruction, inappropriate stimulus parameters, or neural pathway problems).

[0079] Triggered regulation: Based on the risk tracing results, the system automatically generates a diagnostic report and triggers corresponding adaptive adjustment mechanisms. For example, it may fine-tune or interrupt the command output of the brain-computer interface or adjust the stimulation parameters of the spinal cord stimulator (such as amplitude and frequency).

[0080] Security closed-loop verification and learning optimization

[0081] Verify the effectiveness of the control measures: After implementing the control measures, physiological data are collected again, the risk value is recalculated, and the abnormal pattern is checked to verify whether the control measures are effective.

[0082] System learning and optimization: Record the complete data of this event (from signal acquisition to control results) into the database, and use it to regularly optimize and update the parameters of the coupling model and risk assessment model, so that the system can continuously evolve and become more and more accurate as it is used.

[0083] To achieve the above Figure 1 The steps of the method for analyzing the safety of brain-computer interface and spinal cord nerve modulation device implantation are as follows: Figure 3 As shown in the diagram, this embodiment provides a structural schematic of a computer electronic device, which may include the aforementioned components. Figure 1 The steps of the method for analyzing the safety of brain-computer interface and spinal cord nerve modulation device implantation are shown. Optionally, the electronic device 410 may include a first processor 2001.

[0084] Optionally, the electronic device 410 may also include a memory 2002 and a transceiver 2003.

[0085] The first processor 2001, memory 2002, and transceiver 2003 can be connected via a communication bus.

[0086] The following is combined Figure 3 A detailed description of each component of electronic device 410 is provided below:

[0087] The first processor 2001 is the control center of the electronic device 410. It can be a single processor or a collective term for multiple processing elements. For example, the first processor 2001 can be one or more central processing units (CPUs), application-specific integrated circuits (ASICs), or one or more integrated circuits configured to implement embodiments of the present invention, such as one or more digital signal processors (DSPs), or one or more field-programmable gate arrays (FPGAs).

[0088] Optionally, the first processor 2001 can perform various functions of the electronic device 410 by running or executing software programs stored in the memory 2002 and calling data stored in the memory 2002.

[0089] In a specific implementation, as one example, the first processor 2001 may include one or more CPUs, for example... Figure 3 CPU0 and CPU1 are shown in the diagram.

[0090] In a specific implementation, as one example, the electronic device 410 may also include multiple processors, for example... Figure 3 The first processor 2001 and the second processor 2004 are shown in the diagram. Each of these processors can be a single-core processor or a multi-core processor. Here, a processor can refer to one or more devices, circuits, and / or processing cores used to process data (such as computer program instructions).

[0091] The memory 2002 is used to store the software program that executes the present invention, and is controlled by the first processor 2001 to execute it. The specific implementation method can be referred to the above method embodiment, and will not be repeated here.

[0092] Optionally, the memory 2002 may be a read-only memory (ROM) or other type of static storage device capable of storing static information and instructions, random access memory (RAM) or other type of dynamic storage device capable of storing information and instructions, or electrically erasable programmable read-only memory (EEPROM), compact disc read-only memory (CD-ROM) or other optical disc storage, optical disc storage (including compressed optical discs, laser discs, optical discs, digital universal optical discs, Blu-ray discs, etc.), magnetic disk storage media or other magnetic storage devices, or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and accessible by a computer, but not limited thereto. The memory 2002 may be integrated with the first processor 2001 or may exist independently and be accessed through the interface circuit of the electronic device 410. Figure 3 (Not shown in the image) is coupled to the first processor 2001, and this embodiment of the invention does not specifically limit this.

[0093] The transceiver 2003 is used to communicate with network devices or with terminal devices.

[0094] Alternatively, transceiver 2003 may include a receiver and a transmitter. Figure 3 (Not shown separately). The receiver is used to implement the receiving function, and the transmitter is used to implement the transmitting function.

[0095] Optionally, the transceiver 2003 can be integrated with the first processor 2001, or it can exist independently and be connected via the interface circuit of the electronic device 410. Figure 3 (Not shown in the image) is coupled to the first processor 2001, and this embodiment of the invention does not specifically limit this.

[0096] It should be noted that, Figure 3 The structure of the electronic device 410 shown does not constitute a limitation on the router. Actual knowledge structure identification devices may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0097] Furthermore, the technical effects of the electronic device 410 can be referred to the technical effects of the brain-computer interface and spinal cord nerve modulation device implantation safety analysis method described in the above method embodiments, and will not be repeated here.

[0098] It should be understood that the first processor 2001 in this embodiment of the invention may be a central processing unit (CPU), or it may be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor.

[0099] It should also be understood that the memory in the embodiments of the present invention can be volatile memory or non-volatile memory, or may include both volatile and non-volatile memory. The non-volatile memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. The volatile memory can be random access memory (RAM), which is used as an external cache. By way of example, but not limitation, many forms of random access memory (RAM) are available, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate synchronous DRAM (DDRSDRAM), enhanced synchronous DRAM (ESDRAM), synchronous linked DRAM (SLDRAM), and direct rambus RAM (DRRAM).

[0100] The above embodiments can be implemented, in whole or in part, by software, hardware (such as circuits), firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of the present invention are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more sets of available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. A semiconductor medium can be a solid-state drive.

[0101] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any simple modifications, equivalent changes and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. A method for analyzing the safety of implanted devices under brain-computer interface and spinal cord nerve modulation, characterized in that, The specific steps of this method are as follows: S100, Multimodal Physiological Data Acquisition and Synchronization: Brain nerve electrical signals are acquired through an implantable brain-computer interface electrode array, and spinal cord nerve electrical signals and local electromyographic signals are acquired through a modulator electrode implanted in the epidural and subcutaneous regions of the spinal cord. Simultaneously, vital signs such as heart rate, blood pressure, and blood oxygen saturation data are recorded to construct a spatiotemporally synchronized multimodal physiological dataset. S200, Neural Function Linkage and Conflict Modeling: Based on the multimodal physiological dataset, establish a coupling dynamic model of the brain-spinal cord neural pathway, quantitatively analyze the synergy index and conflict coefficient between the expected spinal cord response triggered by brain-computer interface commands and the actual recorded spinal cord response, and identify abnormal linkage patterns. In S200, the process of constructing the coupled dynamics model is as follows: Data preprocessing and feature extraction: The multimodal physiological dataset described in S100 is preprocessed, including using bandpass filtering to extract the effective components of the neural electrical signals, using wavelet transform to remove motion artifacts in the electromyographic signals, and normalizing all signals. Time-domain features and frequency-domain features are extracted from the preprocessed signals. The time-domain features include peak amplitude, latency, and area under the curve, and the frequency-domain features include power spectral density and dominant frequency components. Coupled Dynamics Model Construction: Based on preprocessed feature data, a coupled dynamics model describing the interaction between brain and spinal cord signals is established. This model is defined by a set of coupled differential equations, and the model input is directly related to the extracted features. The set of coupled differential equations is as follows: ,in, for The state vector of brain neural signals at time t, for The spinal cord nerve signal state vector at time t, The brain signal autodynamic parameter matrix, elements Indicates the first The brain signal features of the first The self-regulation strength of individual brain signal features The spinal cord signal autodynamic parameter matrix, elements Indicates the first The first spinal cord signal feature for the first The self-regulation intensity of individual spinal cord signal features This is the coupling strength matrix from the brain to the spinal cord, with elements... Indicates the first The brain signal features of the first The intensity of modulation of individual spinal cord signal features, This is the coupling strength matrix from the spinal cord to the brain, with elements... Indicates the first The first spinal cord signal feature for the first The feedback intensity of individual brain signal features, For the nonlinear transformation activation function from brain to spinal cord, This is the activation function for the nonlinear transformation from the spinal cord to the brain. This is due to the time delay in conduction from the brain to the spinal cord. This is due to the time delay in conduction from the spinal cord to the brain. This is the external control input vector for brain signals, and its value is the command signal strength of the brain-computer interface in S100. This is the external control input vector for spinal cord signals, and its value is the stimulation signal intensity of the spinal cord modulator in S100. The brain signal input gain matrix has elements. Indicates external control input For the Gain coefficient of individual brain signal features The spinal cord signal input gain matrix has elements. Indicates external control input For the Gain coefficients for individual spinal cord signal characteristics This is noise in the brain signaling system. This is noise in the spinal cord signaling system; Model parameter identification and optimization: Parameter matrix in coupled dynamics models Time delay and input gain matrix Online identification and real-time optimization are performed to minimize the mean square error between the model output and the measured data; Based on the model parameters after online identification and real-time optimization, the synergy index CI and the conflict coefficient CC are calculated. The synergy index CI is used to quantify the degree of synchronization and synergy between brain and spinal cord signals. Where cross_corr(x, y) is the brain signal state vector. With spinal cord signal state vector The cross-correlation matrix is ​​given by max(·), where the largest eigenvalue is taken as the synergy index. The synergy was deemed good. The system indicates an abnormality in coordination. The conflict coefficient CC is used to quantify the degree of conflict between the actual spinal cord response and the expected response. ,in, This is the feature vector of the actual recorded spinal cord nerve signals. This is the spinal cord response vector predicted based on the coupled dynamics model. The Euclidean distance between the actual response and the expected response. The standardized parameter is the standard deviation of the spinal cord signal characteristics under the S100 resting state. The closer it is to 1, the more severe the conflict. The system indicates a significant conflict. S300, Dynamic assessment of security risks: Based on the identification results of the aforementioned synergy index, conflict coefficient, and abnormal linkage mode, a risk assessment algorithm is introduced to calculate short-term and long-term risk values ​​in real time. S400, Multi-level Early Warning and Adaptive Control Trigger: Set safety thresholds corresponding to the short-term risk value and the long-term risk value. When the risk assessment result exceeds any safety threshold, generate a corresponding abnormal signal and trigger the adaptive adjustment mechanism of the implantable control device accordingly. S500, Safety Closed-Loop Verification and Learning Optimization: After performing adaptive regulation, physiological data is collected again, and the characteristic differences of spinal cord response signals and electromyographic signals caused by the same brain-computer interface commands before and after adaptive regulation are compared. When the abnormal linkage pattern disappears or the risk value falls back to the safe range, the relief is deemed effective.

2. The method for safety analysis of brain-computer interface and spinal cord nerve modulation device implantation according to claim 1, characterized in that, In S100, the implantable brain-computer interface electrode array is a 16-channel microelectrode array, which is implanted into the motor cortex and sensory cortex of the brain to collect brain neural electrical signals, including cortical electroencephalography and neural unit discharge signals. The modulator electrode implanted in the epidural space of the spinal cord is an 8-channel cylindrical electrode, and the implantation segment matches the diseased spinal cord segment. The subcutaneous modulator electrode is a 4-channel sheet electrode, which is embedded in the subclavian subcutaneous tissue. The two work together to collect spinal cord nerve electrical signals and local electromyographic signals. The spinal cord nerve electrical signals and local electromyographic signals include composite action potentials and local field potentials recorded in the epidural space of the spinal cord. The spatiotemporal synchronization of the multimodal physiological dataset is achieved by assigning a uniform timestamp to all signal sources.

3. The method for safety analysis of brain-computer interface and spinal cord nerve modulation device implantation according to claim 1, characterized in that, The S200 is based on the calculated synergy index. Conflict coefficient Based on the S100 vital signs data, the following abnormal linkage patterns are identified, and these abnormal patterns are directly used as input to the S300 risk assessment algorithm: Instruction propagation delay exceeds limit: When the identification time delay exceeds the limit. or Exceeding its normal physiological range threshold When this occurs, it is determined that the instruction transmission delay has exceeded the limit; Signal gain anomaly: when the coupling strength matrix or norm Exceeding its normal range When this occurs, it is determined to be an abnormal signal gain; Unintended spinal reflexes: when given commands via the S100 brain-computer interface In this case, spinal cord signals Amplitude > The rhythmic discharges were determined to be unexpected spinal reflexes.

4. The method for safety analysis of brain-computer interface and spinal cord nerve modulation device implantation according to claim 1, characterized in that, The S300, based on the calculated synergy index CI, conflict coefficient CC, and abnormal linkage mode, combined with the vital sign parameters collected by the S100, constructs a multidimensional risk feature vector F=[CI, CC, Δτ, ΔG, ΔH, ΔR]T, where Δτ is the maximum deviation of the transmission time delay, defined as Δτ=max(| - |,| - |), The nominal value is within the normal physiological range. The maximum deviation of the norm of the coupling strength matrix is ​​defined as follows: ), Let F be the norm of the matrix. This is the nominal value of the coupling strength. The deviation of heart rate from the baseline is defined as follows: ,in To collect heart rate in real time, This is the baseline value of resting heart rate. The deviation of blood pressure from the baseline is defined as... ,in To collect blood pressure in real time, This is the baseline value for resting blood pressure; The risk feature vector F is input into the risk assessment model, and an activation function is applied. Mapping outputs short-term risk values and long-term risk value Quantifying the immediate and chronic risks of neuromodulation, the aforementioned , ,in, , For the weight vector, This is a bias term used to correct the model output offset.

5. The method for analyzing the safety of brain-computer interface and spinal cord nerve modulation device implantation according to claim 4, characterized in that, In the S400, the following settings are provided: , To preset a safety threshold, when or For risk feature vectors The contribution of each component in the model is quantitatively analyzed, and the contribution of each feature is calculated. Marginal contribution to risk value Identify the dominant risk factors: when , The component with the highest contribution is identified as having a dominant risk factor of neural signal conflict. It generates a diagnostic report of abnormal signal transduction pathways and provides suggestions to check the integrity of neural pathways. when , It has the highest contribution rate, and the dominant risk is determined to be abnormal brain-computer interface commands. It generates a diagnostic report of abnormal command encoding and provides an identifier of command parameters to be adjusted for the adaptive adjustment mechanism. when , It has the highest contribution rate, and the dominant risk is determined to be inappropriate spinal cord modulation stimulation parameters. It generates a stimulation parameter optimization diagnostic report and provides the types and directions of stimulation parameters to be adjusted for the adaptive regulation mechanism in S400.

6. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by the processor, the program implements the steps of the method for analyzing the safety of device implantation under brain-computer interface and spinal cord nerve modulation as described in any one of claims 1-5.

7. A computer electronic device, characterized in that, include: Memory, used to store computer programs; A processor is configured to execute a computer program stored in the memory to implement the method for safety analysis of brain-computer interface and spinal cord nerve modulation device implantation as described in any one of claims 1-5.

Images