A high-resolution peripheral nerve imaging system based on acoustic-electric coupling

Abstract

The present invention relates to a high-resolution peripheral nerve imaging system based on acoustic-electric coupling, comprising an electrical stimulation induction module, an ultrasonic scanning module, a signal processing module, a signal output module, an electrical stimulation component, an ultrasonic transducer component and an imaging device; the signal processing module includes a signal processing unit and an enhanced extraction unit, the signal processing unit is configured with a signal processing strategy for processing high-frequency ultrasonic coded electrical signals to generate low-frequency bioelectrical signals, and the enhanced extraction unit is configured with an enhanced extraction strategy for processing low-frequency bioelectrical signals to extract target nerve signals; effectively separate mixed bioelectrical signals in the nerve signal enhancement and feature extraction module, especially in the case of containing complex noise and interference signals (such as electromyographic signals), and still be able to accurately extract target nerve signals.

Description

Technical Field

[0001] The present invention relates to the technical field of bioinformatics scanning, and more particularly, to a high-resolution peripheral nerve imaging system based on acousto-electric coupling. Background Art

[0002] Peripheral nerve injury is a common type of neurological disease, often causing movement disorders, sensory abnormalities, and autonomic nervous system dysfunction, seriously affecting the quality of life of patients. Clinically, peripheral nerve imaging technology is a key means for diagnosing and treating peripheral nerve injuries. It can visually present the structure and functional status of the peripheral nervous system, helping doctors evaluate the activity and pathological conditions of nerves.

[0003] In recent years, the main peripheral nerve imaging method in clinical practice is magnetic resonance imaging (MRI). Magnetic resonance neuroimaging (MRI) is a non-invasive medical imaging technology that can be used to observe the structure and function of the peripheral nerves in the human body. MRI uses a strong magnetic field and harmless radio waves to generate high-resolution images. However, the temporal resolution of MRI is between several seconds and dozens of seconds, resulting in the inability to achieve real-time imaging and too long imaging time. Moreover, there are many contraindications for MRI: it is not applicable to critically ill patients, patients with claustrophobia, and pregnant women. MRI scans are also interfered by certain metal substances, such as cardiac pacemakers and artificial joints. Therefore, there is a need to develop a new type of peripheral nerve imaging system with a wider adaptability and high spatio-temporal resolution.

[0004] In 1946, scientists such as Fox studied the effect of ultrasonic waves on the conductivity of salt solutions, which was the first to propose the principle of acousto-electric coupling and laid the foundation for subsequent research on acousto-electric coupling (Fox F E, Herzfeld K F, Rock G D. The effect of ultrasonic waves on the conductivity of salt solutions[J]. Physical review, 1946, 70(5-6): 329.). In 1998 and 1999, Jossinet et al. pointed out that the propagation of ultrasonic waves in electrolyte solutions led to periodic changes in local conductivity, thus generating acousto-electric coupling signals (Jossinet J, Lavandier B, Cathignol D. Impedance modulation by pulsed ultrasound[C]. Annals of the New York Academy of Sciences, New York Acad Sciences: New York, 1999, 396-407.) (Jossinet J, Lavandier B, Cathignol D. The phenomenology of acousto-electric interaction signals in aqueous solutions of electrolytes[J]. Ultrasonics, 1998, 36(1-5): 607-613.). In 2007, R.S. Witte et al. utilized this interaction to image the ionic current injected into the abdominal segment of the lobster nerve cord (Witte R, Olafsson R, Huang S W, et al. Imaging current flow in lobster nerve cord using the acousto-electric effect[J]. Applied physics letters, 2007, 90(16): 163902.), which was the first time they measured ultrasonic space-coded electrophysiological signals based on acousto-electric coupling.In 2008, R. Olafsson et al. proposed ultrasound current source density imaging (UCSDI) based on acousto-electric coupling, which is a direct three-dimensional imaging technique that may endow existing mapping procedures with superior spatial resolution (Olafsson R, Witte R S, Huang S W, et al. Ultrasound current sourcedensity imaging[J]. IEEE Transactions on biomedical engineering, 2008, 55(7): 1840-1848.). In 2015, Y. Qin et al. first proposed measuring the three-dimensional cardiac activation map of a live rabbit heart using only a pair of recording electrodes. They demonstrated that UCSDI has the potential to achieve real-time 3D cardiac activation wave mapping, which will greatly facilitate the ablation treatment of arrhythmia (Qin Y, Li Q, Ingram P, et al. Ultrasound current source density imagingof the cardiac activation wave using a clinical cardiac catheter[J]. IEEE Transactions on Biomedical Engineering, 2015, 62(1): 241-247.). In 2016, R. S. Witte et al. proposed a 4D acousto-electric computer imaging method based on acousto-electric coupling to map current density with the spatial resolution of the focused ultrasound focus, overcoming the defect of low spatial resolution of traditional scalp electroencephalogram (Qin Y, Ingram P, Burton A, et al. 4Dacoustoelectric imaging of current sources in a human head phantom[C] / / 2016IEEE International Ultrasonics Symposium(IUS). IEEE, 2016: 1-4.).

[0005] Although the above studies have made remarkable progress in the fields of cardiac function imaging and electroencephalogram imaging, they have not been applied to high-resolution peripheral nerve imaging yet. Therefore, there is an urgent need to develop a new type of high-resolution peripheral nerve imaging system based on acousto-electric coupling to meet the actual needs of clinical diagnosis and scientific research, which is of great significance for promoting the technological progress in this field.

[0006] For example, CN107361794B discloses a device and method for detecting motor nerve feedback based on an ultrasonic component and a peripheral nerve stimulator. By combining the peripheral nerve stimulator with an ultrasonic sensor, the feedback of muscles to weak electrical stimulation or deeper stimulation around the body can be effectively, timely and accurately monitored without causing any physical trauma to the body. At the same time, the device can also be widely applied to the field of rehabilitation therapy. The method of the invention accurately monitors and evaluates the motor nerve performance of the monitored object in real time and non-invasively. However, it determines the health state of nerve feedback based on the detection of muscle displacement changes to assist in analyzing the condition, but it cannot obtain a relatively high-precision nerve conduction signal, and the weak nerve conduction signal is affected by the noise of muscle activity signals and other interference signals, so it is difficult to obtain. Summary of the Invention

[0007] In view of this, the object of the present invention is to provide a high-resolution peripheral nerve imaging system based on acoustic-electric coupling.

[0008] To solve the above technical problems, the technical solution of the present invention is: A high-resolution peripheral nerve imaging system based on acoustic-electric coupling, including an electrical stimulation induction module, an ultrasonic scanning module, a signal processing module, a signal output module, an electrical stimulation component, an ultrasonic transducer component, and an imaging device;

[0009] The electrical stimulation induction module is used to control the operation of the electrical stimulation component. The electrical stimulation induction module is configured with a steady-state induction strategy, and the steady-state induction strategy is used to configure corresponding electrical stimulation parameters according to the type information of the peripheral nerve, and control the corresponding electrical stimulation component to work according to the electrical stimulation parameters to make the nerve fibers generate steady-state potentials;

[0010] The ultrasonic scanning module is used to control the operation of the ultrasonic transducer. The ultrasonic scanning module is configured with an ultrasonic configuration strategy and an acoustic-electric reception strategy. The ultrasonic configuration strategy is used to generate an ultrasonic configuration instruction to control the ultrasonic transducer to send focused ultrasonic waves in the target detection area. The acoustic-electric reception strategy is configured with preset reception constraint conditions, and determines high-frequency ultrasonic encoded electrical signals based on acoustic-electric coupling from the feedback waveform signals according to the reception constraint conditions;

[0011] The signal processing module includes a signal processing unit and an enhancement extraction unit. The signal processing unit is configured with a signal processing strategy, and the signal processing strategy is used to process the high-frequency ultrasonic encoded electrical signals to generate low-frequency bioelectrical signals. The enhancement extraction unit is configured with an enhancement extraction strategy, and the enhancement extraction strategy is used to process the low-frequency bioelectrical signals to extract target nerve signals;

[0012] The signal output module is used to generate an imaging output signal to the corresponding imaging device according to the target nerve signal.

[0013] Furthermore, the steady-state induction strategy includes a parameter information configuration table, which stores a number of electrical stimulation parameters and corresponding adjustment instruction sequences. The adjustment instruction sequence includes a number of parameter adjustment instructions for adjusting the electrical stimulation parameters. The electrical stimulation parameters are indexed by the type information of the peripheral nerve. The steady-state induction strategy includes:

[0014] Step A1: Obtain the type information of the peripheral nerve;

[0015] Step A2: Obtain the corresponding electrical stimulation parameters according to the type information;

[0016] Step A3: Control the operation of the electrical stimulation component according to the electrical stimulation parameters;

[0017] Step A4: Judge whether a steady-state potential is generated by the feedback current corresponding to the recording electrode. If a steady-state potential is generated, maintain the electrical stimulation parameters and end the steady-state induction strategy. If a steady-state trigger potential is not generated, enter Step A5;

[0018] Step A5: Adjust and update the electrical stimulation parameters according to the order of the parameter adjustment instructions in the adjustment instruction sequence, and return to Step A3.

[0019] Furthermore, the reception constraint condition is configured such that the frequency of the high-frequency ultrasonic coded electrical signal falls within the constraint frequency range and the amplitude of the high-frequency ultrasonic coded electrical signal falls within the preset constraint amplitude range. The constraint frequency range is generated according to the working frequency of the focused ultrasonic wave, and the constraint amplitude range is generated according to the steady-state potential.

[0020] Furthermore, the signal processing strategy includes:

[0021] Step B1: Configure a preset decimation acquisition frequency to configure the sampling frequency of the high-frequency ultrasonic coded electrical signal;

[0022] Step B2: Configure a filtering frequency to perform filtering processing on the acquired high-frequency ultrasonic coded electrical signal;

[0023] Step B3: Process the filtered high-frequency ultrasonic coded electrical signal through the Hilbert transform algorithm to obtain a low-frequency bioelectrical signal.

[0024] Furthermore, the enhancement extraction strategy includes:

[0025] Step C1: Initialize the demixing matrix;

[0026] Step C2: Perform mean removal processing and whitening processing on the low-frequency bioelectrical signal;

[0027] Step C3: Determine whether the preset convergence condition is satisfied. If the convergence condition is satisfied, proceed to step C5; if not, proceed to step C4.

[0028] Step C4: Update the unmixing matrix by the gradient ascent method and return to step C3.

[0029] Step C5: Determine each corresponding independent signal component according to the current unmixing matrix, and screen the corresponding independent signal components through a preset screening sub-condition to obtain nerve signal components. The target nerve signal is a set of nerve signal components.

[0030] Furthermore: The update rule of the gradient ascent method is:

[0031] ΔW = η[I+(1 - 2g(Y))Y T W, where ΔW is the updated unmixing matrix, W is the unmixing matrix before update, η is the learning rate, I is the identity matrix, g(Y) is the result after applying a non-linear function to Y, Y is the independent signal component, and Y = WX whitened ,X whitened is the low-frequency bioelectric signal after whitening processing.

[0032] Furthermore: The screening sub-condition is configured such that the signal characteristics of the independent signal component conform to the preset target nerve signal characteristics.

[0033] Furthermore: Step C5 further includes extracting the signal characteristics of the nerve signal component, and the signal characteristics include time domain characteristics and frequency domain characteristics.

[0034] Furthermore: The signal output module further includes a data compensation strategy, and the data compensation strategy is configured with an interpolation sub-strategy to generate the imaging output signal according to the target nerve signal.

[0035] The technical effects of the present invention are mainly reflected in the following aspects:

[0036] 1. Effectively separate the mixed bioelectric signals in the nerve signal enhancement and feature extraction module. Especially in the case of containing complex noise and interference signals (such as electromyographic signals), the target nerve signal can still be accurately extracted.

[0037] 2. Compared with traditional peripheral nerve imaging, the present invention couples ultrasound on the electrical basis. Based on the principle of acoustic-electric coupling and the high targeting of focused ultrasound, the time resolution and spatial resolution of peripheral imaging technology are improved, and high-resolution peripheral nerve imaging is realized.

[0038] 3. The present invention can achieve a time resolution of milliseconds, overcomes the shortcoming of low time resolution of traditional peripheral nerve imaging technology, and has high precision and accuracy in detecting the dynamic changes of nerves. Description of the Drawings

[0039] Figure 1 : Diagram of the interaction relationship of the device of a high-resolution peripheral nerve imaging system based on acoustic-electric coupling according to the present invention;

[0040] Figure 2 : Schematic diagram of the system architecture of a high-resolution peripheral nerve imaging system based on acoustic-electric coupling according to the present invention;

[0041] Figure 3 : Schematic diagram of the imaging principle of the ultrasonic transducer according to the present invention;

[0042] Figure 4 : Flowchart of the implementation of the enhanced extraction strategy algorithm according to the present invention;

[0043] Figure 5 : Flowchart of the signal processing architecture according to the present invention;

[0044] Figure 6 : Diagram of the signal waveform relationship of the experimental example according to the present invention.

[0045] Reference Signs: 1, signal processing device; 2, electrical stimulation component; 3, ultrasonic transducer component; 4, imaging device; 100, electrical stimulation induction module; 200, ultrasonic scanning module; 300, signal processing module; 400, signal output module. Detailed Embodiments

[0046] The following further details the specific embodiments of the present invention in conjunction with the drawings, so that the technical solutions of the present invention are easier to understand and master.

[0047] Refer to Figure 1As shown, the system includes four main modules. First, the electrical stimulation steady-state induction module induces the steady-state potential of the peripheral nerve through a precisely controlled electrical stimulator, providing a stable electrophysiological signal basis for subsequent signal processing and imaging. Subsequently, the ultrasound scanning module 200 is responsible for generating, transmitting, and scanning focused ultrasound to comprehensively cover the target peripheral nerve area, ensuring the acquisition of high-spatial-resolution high-frequency ultrasound-encoded electrical signals. The collected signals are amplified, filtered, and digitally processed by the signal processing module 300 to ensure the integrity and accuracy of the signals. Then, the processed signals are analyzed, and the nerve signals are effectively enhanced through feature extraction algorithms, while the interference of muscle signals is reduced to improve the imaging accuracy. Finally, the signal output module 400 converts these processed signals into high-resolution peripheral nerve imaging results and generates visual images for further analysis and application. Through the collaborative work of these modules, the system can efficiently and accurately achieve peripheral nerve imaging and damage detection. The electrical stimulation induction module 100, the ultrasound scanning module 200, the signal processing module 300, and the signal output module 400 are all configured in the signal processing device 1 for signal processing.

[0048] The working flowchart of the acoustic-electric coupled peripheral nerve imaging system is as Figure 1 shown.

[0049] A high-resolution peripheral nerve imaging system based on acoustic-electric coupling includes an electrical stimulation induction module 100, an ultrasound scanning module 200, a signal processing module 300, a signal output module 400, an electrical stimulation component 2, an ultrasound transducer component 3, and an imaging device 4; the electrical stimulation component 2 can be configured as a peripheral nerve stimulator, and the ultrasound transducer component can be configured as an ultrasound transducer; the imaging device 4 is configured as any display terminal with an image output function.

[0050] The electrical stimulation induction module 100 is used to control the operation of the electrical stimulation component 2. The electrical stimulation induction module 100 is configured with a steady-state induction strategy, and the steady-state induction strategy is used to configure corresponding electrical stimulation parameters according to the type information of the peripheral nerve. This module induces a steady-state potential in the peripheral nerve by applying appropriate electrical stimulation. This steady-state potential refers to the stable and reproducible electrical signal generated by nerve fibers under continuous electrical stimulation. The generation of the steady-state potential depends on the precise adjustment of electrical stimulation parameters such as voltage intensity, frequency, and stimulation time. By optimizing these parameters, this unit can induce stable electrical signals on different types of peripheral nerves, ensuring the consistency and predictability of nerve responses.

[0051] The steady-state induction strategy includes a parameter information configuration table, which stores a number of electrical stimulation parameters and corresponding adjustment instruction sequences. The adjustment instruction sequences include a number of parameter adjustment instructions for adjusting the electrical stimulation parameters. The electrical stimulation parameters are indexed by the type information of the peripheral nerve. The steady-state induction strategy includes:

[0052] Step A1: Obtain the type information of the peripheral nerve;

[0053] Step A2: Obtain the corresponding electrical stimulation parameters according to the type information;

[0054] Step A3: Control the operation of the electrical stimulation component 2 according to the electrical stimulation parameters;

[0055] Step A4: Determine whether a steady-state potential is generated by judging the feedback current corresponding to the recording electrode. If a steady-state potential is generated, maintain the electrical stimulation parameters and end the steady-state induction strategy. If the steady-state trigger potential is not generated, enter Step A5;

[0056] Step A5: Adjust and update the electrical stimulation parameters according to the order of the parameter adjustment instructions in the adjustment instruction sequence, and return to Step A3. And control the corresponding electrical stimulation component 2 to operate according to the electrical stimulation parameters to make the nerve fibers generate a steady-state potential;

[0057] The ultrasonic scanning module 200 is used to control the operation of the ultrasonic transducer. The ultrasonic scanning module 200 is configured with an ultrasonic configuration strategy and an acoustic-electric reception strategy. The ultrasonic configuration strategy is used to generate ultrasonic configuration instructions to control the ultrasonic transducer to emit focused ultrasonic waves in the target detection area.

[0058] The acoustic-electric reception strategy is configured with preset reception constraint conditions, and determines the high-frequency ultrasonic encoded electrical signal based on acoustic-electric coupling from the feedback waveform signal according to the reception constraint conditions. The configured reception constraint conditions are that the frequency of the high-frequency ultrasonic encoded electrical signal falls within the constraint frequency range and the amplitude of the high-frequency ultrasonic encoded electrical signal falls within the preset constraint amplitude range. The constraint frequency range is generated according to the operating frequency of the focused ultrasonic wave, and the constraint amplitude range is generated according to the steady-state potential.

[0059] The ultrasonic scanning module 200 is the core part of the acoustic-electric coupling peripheral nerve imaging system, responsible for generating, transmitting, and scanning focused ultrasonic waves. This unit focuses the focused ultrasonic waves on the peripheral nerve, and the focal spot is the detection area. The focused ultrasonic waves encode the peripheral nerve electrical signals in the focal spot area. If there are steady-state electrical signals in the peripheral nerve in the focal spot area, a high-frequency ultrasonic encoded electrical signal, that is, an acoustic-electric signal, is generated; if there are no steady-state electrical signals in the peripheral nerve in the focal spot area, disordered noise is generated. The ultrasonic encoded electrical signal has the same frequency as the focused ultrasonic wave and is positively correlated with the amplitude of the peripheral nerve electrical signal.

[0060] The high-frequency ultrasonic encoded electrical signal after the acousto-electric coupling effect is specifically as follows: When the focused ultrasonic wave focuses on the peripheral nerve, the change in the resistivity of the focal spot area of the focused ultrasonic wave causes a change in the electrical signal of the peripheral nerve measured by the recording electrode, generating a high-frequency ultrasonic encoded electrical signal based on acousto-electric coupling. This signal has the same frequency as the focused ultrasonic wave, is positively correlated with the amplitude of the peripheral nerve electrical signal, and carries the spatial position information of the focal spot area of the focused ultrasonic wave. The mathematical relationship is as follows:

[0061]

[0062] Among them, J I is the peripheral nerve current source; is the current density corresponding to the electrode under the condition of unit current injection; K is the acousto-electric coupling interaction coefficient, which depends on the properties of the tissue medium; ρ0 is the original resistivity of the tissue; ΔP is the change in ultrasonic sound pressure; ω is the focal spot area of the focused ultrasonic wave, and the principle is as Figure 3 shown. According to the above formula, the theoretical value of the corresponding signal amplitude can be deduced, and then the corresponding constraint amplitude range can be configured according to the theoretical value.

[0063] The signal processing module 300 includes a signal processing unit and an enhanced extraction unit. The signal processing unit is configured with a signal processing strategy, which is used to process the high-frequency ultrasonic encoded electrical signal to generate a low-frequency bioelectrical signal. The core function of the signal processing module 300 is to obtain the high-frequency ultrasonic encoded electrical signal based on acousto-electric coupling, that is, the acousto-electric signal, from the peripheral nerve area, and perform downsampling and filtering processing on these signals, and decode the low-frequency signal reflecting nerve activity through Hilbert transform. This unit needs to process complex and weak bioelectrical signals, so it must have extremely high precision and stability in the links of signal acquisition, amplification, filtering, etc.

[0064] The signal processing strategy includes:

[0065] Step B1: Configure a preset downsampling acquisition frequency to configure the sampling frequency of the high-frequency ultrasonic encoded electrical signal;

[0066] Step B2: Configure a filtering frequency to perform filtering processing on the acquired high-frequency ultrasonic encoded electrical signal;

[0067] Step B3: Process the filtered high-frequency ultrasonic encoded electrical signal through the Hilbert transform algorithm to obtain a low-frequency bioelectrical signal.

[0068] The enhanced extraction unit is configured with an enhanced extraction strategy, which is used to process the low-frequency bioelectrical signal to extract the target nerve signal;

[0069] The enhanced extraction strategy includes:

[0070] Step C1: Initialize the demixing matrix;

[0071] Step C2: Perform mean removal and whitening on the low-frequency bioelectric signals;

[0072] Step C3: Determine whether the preset convergence condition is satisfied. If the convergence condition is satisfied, go to Step C5; if not, go to Step C4;

[0073] Step C4: Update the demixing matrix by the gradient ascent method and return to Step C3;

[0074] The update rule of the gradient ascent method is:

[0075] ΔW = η[I+(1 - 2g(Y))Y T W, where ΔW is the updated demixing matrix, W is the demixing matrix before update, η is the learning rate, I is the identity matrix, g(Y) is the result after applying a non-linear function to Y, Y is the independent signal component, and Y = WX whitened , X whitened is the low-frequency bioelectric signal after whitening processing.

[0076] The screening sub-condition is configured such that the signal characteristics of the independent signal component conform to the preset target nerve signal characteristics.

[0077] Step C5 further includes extracting the signal characteristics of the nerve signal component, and the signal characteristics include time domain characteristics and frequency domain characteristics.

[0078] Step C5: Determine each corresponding independent signal component according to the current demixing matrix, and screen the corresponding independent signal components through a preset screening sub-condition to obtain nerve signal components, and the target nerve signal is a set of nerve signal components.

[0079] The nerve signal enhancement and feature extraction module adopts the information maximization independent component analysis algorithm, aiming to separate independent nerve signal components from the mixed bioelectric signals, and enhance and extract the features of these components. By demixing the collected multi-channel signals, this module can effectively suppress noise and electromyogram signal interference, significantly improve the purity and signal-to-noise ratio of nerve signals, and provide a high-quality data basis for subsequent signal analysis and imaging.

[0080] First, the input signal data undergoes mean removal and whitening processing to ensure zero mean and unit covariance matrix of the data. Whitening processing transforms the original signal into uncorrelated components through linear transformation, laying the foundation for the subsequent processing of the algorithm.

[0081] Separate the preprocessed mixed signals into statistically independent signal components. The specific implementation process is as follows, as Figure 4 shown:

[0082] A) Signal Modeling

[0083] Assume the collected mixed signal matrix is \(X\in\mathbb{R}\) Μ×Ν where \(M\) is the number of signal channels and \(N\) is the number of time sampling points. The mixing model is expressed as: \(X = AS\)

[0084] where, \(A\in\mathbb{R}\) Μ×Μ is the unknown mixing matrix, and \(S\in\mathbb{R}\) Μ×Ν is the source signal matrix.

[0085] B) Whitening Processing

[0086] To simplify the separation process and improve the convergence speed of the algorithm, the mixed signal \(X\) is whitened. The whitened signal \(X\) whitened satisfies: where \(V\) and \(D\) are the eigenvector matrix and eigenvalue matrix of the covariance matrix of \(X\), respectively.

[0087] C) Information Maximization Independent Component Analysis

[0088] By maximizing the entropy of the output signal \(Y = WX\) whitened the independent separation of the signal is achieved. The specific steps include:

[0089] Definition of the objective function: Define the objective function as the sum of the entropies of the output signals:

[0090]

[0091] where \(H(Y\) i ) is the entropy of the \(i\)-th independent component.

[0092] 2. Nonlinear activation function: Select an appropriate nonlinear function \(g(y)\) to approximate the probability distribution of the output signal. Commonly used functions include the hyperbolic tangent function \(g(Y)=\tanh(Y)\) or the logistic function

[0093] 3. Gradient ascent optimization: Use the gradient ascent method to iteratively update the demixing matrix \(W\). The update rule is:

[0094] \(\Delta W=\eta[I+(1 - 2g(Y))Y\) T W

[0095] where \(\eta\) is the learning rate, \(I\) is the identity matrix, \(Y = WX\) whitened , and \(g(Y)\) is the result after applying the nonlinear function to \(Y\).

[0096] 4. Convergence determination: Repeat the gradient update process until the demixing matrix \(W\) converges to a stable value or reaches the preset number of iterations.

[0097] D) Signal Enhancement and Feature Extraction

[0098] The obtained independent components Y represent the separated source signals. Among these independent components, some correspond to the target nerve signals, while the other components are muscle activity or environmental noise signals. Through further analysis and screening, the target nerve signals can be retained to achieve signal enhancement.

[0099] Key features are extracted from the enhanced nerve signals, including the time-domain waveform features (such as peaks) and frequency-domain features (such as frequency components, power spectral density) of the signals. These features are used for subsequent peripheral nerve imaging analysis.

[0100] The signal output module 400 is used to generate an imaging output signal based on the target nerve signal and send it to the corresponding imaging device 4. The signal output module 400 also includes a data compensation strategy, and the data compensation strategy is configured with an interpolation sub-strategy to generate the imaging output signal according to the target nerve signal.

[0101] The imaging display and data output module is a key part of the system of the present invention for visualizing and outputting the processed nerve signals. The main function of this module is to convert the high-quality nerve signals processed by the nerve signal enhancement and feature extraction module into images. At the same time, this module is also responsible for storing and outputting the generated imaging data, providing convenience for subsequent analysis or scientific research.

[0102] In the present invention, the nerve signals processed and enhanced by the algorithm are sent into the imaging algorithm and converted into high-resolution peripheral nerve imaging images. This module adopts interpolation imaging technology:

[0103] The core idea of the interpolation algorithm is to estimate the unknown values between the known discrete sampling points (such as the signals at electrode positions) to generate a continuous image. For nerve signal imaging, the collected signals are usually limited and discrete, and the interpolation algorithm is used to fill the blank areas between these sampling points to construct a complete nerve activity image. In the present invention, the interpolation process associates the collected discrete nerve signal values with their positions in space, and generates continuous imaging data by calculating the interpolation values between adjacent signal points.

[0104] Linear interpolation is a simple and commonly used interpolation method. By drawing a straight line between adjacent points, the values of the unsampled points are estimated. In two-dimensional or three-dimensional space, linear interpolation determines the values of the interpolation points by calculating the linear relationship between two or more known points. The specific steps are as follows:

[0105] Define the known sampling points: Set (x1, y 1, z1, f(x1, y1, z1)) and (x2, y 2,z2, f(x2, y2, z2)) are two known sampling points, and f is the neural signal value collected at these points.

[0106] Calculate the interpolation points: Calculate the interpolation points between x, y, and z through the interpolation formula. For example, the one-dimensional linear interpolation formula is: f(x) = f(x1)(x2 - x) / (x2 - x1) + f(x2)(x - x1) / (x2 - x1)

[0107] In the two-dimensional or three-dimensional cases, the calculation method can be extended to each dimension, and interpolation is performed through the signal intensity difference between adjacent points.

[0108] First, stimulate the peripheral nerve through the electrically stimulated steady-state evoked unit. This unit includes an electrical stimulator OIympus 5077PR that can adjust the voltage intensity, frequency, and pulse width. The electrical stimulator applies a current signal to the target nerve area to induce the nerve to generate a steady-state potential signal. The generation of the steady-state potential provides a stable electrophysiological signal basis for the subsequent acoustic-electric coupling effect.

[0109] While the electrical stimulation is being carried out, the ultrasonic emission and scanning unit starts to work. This unit uses an Olympus5077PR focused ultrasonic signal generator and an A303s ultrasonic transducer to generate and emit focused ultrasonic signals, and sets ultrasonic characteristic parameters: sets the pulse repetition frequency (100Hz, 200Hz, 500Hz, etc.), ultrasonic oscillation frequency (0.5MHz, 1MHz), excitation pulse intensity (100V, 200V, 300V, 400V), etc. Through the scanning controller, the ultrasonic beam gradually scans the entire target area to ensure that the entire range of the peripheral nerve is covered. The focusing and scanning of the ultrasonic waves enable the acoustic-electric coupling signals in the target area to be accurately excited and captured.

[0110] The acoustic-electric coupling signals are captured by the recording electrodes and transmitted to the signal acquisition and processing unit. This unit includes a high-gain low-noise amplifier for amplifying the weak acoustic-electric signals, then removes noise and interference signals through a filter, and finally converts the processed signals into digital signals through an analog-to-digital converter (ADC) for subsequent analysis and imaging. Downsample the digital signals, extract high-frequency acoustic-electric signals using a band-pass filter with a PRF± of 15Hz, use the Hilbert function to resolve the low-frequency decoded acoustic-electric signals reflecting neural activity, that is, low-frequency bioelectric signals, and finally perform a fast Fourier transform (FFT) on the processed signals to convert the time-domain signals into frequency-domain signals, thereby analyzing their spectral characteristics.

[0111] The processed digital signal enters the neural signal enhancement and feature extraction module. This module uses the Infomax Independent Component Analysis (Infomax ICA) algorithm to demix and separate the mixed signals. The Infomax ICA algorithm effectively separates neural signals from electromyographic signals and other interference signals by maximizing the entropy of the output signals. The separated neural signals are enhanced to further improve the signal-to-noise ratio, and then feature extraction is performed on these signals to generate key parameters for imaging.

[0112] Finally, the enhanced and extracted neural signals are sent to the imaging display and data output module. This module uses an interpolation algorithm to convert discrete neural signal points into continuous two-dimensional or three-dimensional images. Through the interpolation algorithm, the generated images have high resolution and high precision, and can intuitively display the distribution of electrical activities of peripheral nerves.

[0113] Experimental case, in specific implementation, such as Figure 1 As shown, male Wistar rats were anesthetized with isoflurane and 30% urethane (4 - 5 ml / kg). After the rats were in the anesthetized state, surgery was performed and their anesthetized state was maintained throughout the experiment to ensure that the rats were fixed on the experimental platform. First, the hair on the lower limbs of the rats was shaved, then the skin and muscle tissues of the lower limbs were incised to expose the sciatic nerve, and the lower limbs of the rats were fixed on a stereotaxic device to ensure the stability of the surgical area and precise operation. Finally, recording electrodes were placed at the sciatic nerve to ensure good contact between the recording electrodes and the sciatic nerve to obtain stable electrical signal recordings and not damage the nerve sheath.

[0114] We used a device consisting of a 1 MHz single-element focused transducer (A303S, Olympus, USA) and a pulse transmitter / receiver (Olympus, 5077PR, USA) to generate focused ultrasound (FUS). The pulse repetition frequency (PRF) of FUS was controlled by a trigger signal generated by a signal generator (RIGOL DG 4162), which was a pulsed wave with an amplitude of 2.4 V and a duty cycle of 20%. The PRF of FUS was set to 1137 Hz. The trigger signal was fed into the pulse transmitter / receiver to drive the transducer to generate ultrasonic pulses.

[0115] To accurately locate the position of the detection area, the ultrasonic transducer was equipped with a conical collimator. The collimator was filled with a coupling agent to ensure the propagation of focused ultrasonic waves. The height of the collimator was 20 mm, so that the focal area of FUS was located at the sciatic nerve, and the end of the collimator was as close as possible to the recording electrode. In addition, a sine current wave with an amplitude of 800 mV and a frequency of 13 Hz was emitted by a signal generator (RIGOL DG4162) to electrically stimulate the sciatic nerve through a pair of platinum electrodes.

[0116] Referring to Figure 6 as shown, in the digital signal processing unit, first the acquired signal is downsampled to 5000 Hz, and then third-order filtering processing is performed using different filtering parameters according to the filtering purpose. The low-frequency electrical stimulation signal is extracted through band-pass filtering (range 5 to 50 Hz), the high-frequency acoustic-electric signal is extracted through band-pass filtering in the range of PRF ± 15 Hz, and the signal is smoothed by the method of superposition averaging. The Hilbert transform is used to extract the signal envelope to obtain the low-frequency decoded acoustic-electric signal. Finally, fast Fourier transform (FFT) analysis is performed on the processed signal, as Figure 5 shown.

[0117] The processed digital signal enters the neural signal enhancement and feature extraction module, and the enhanced and extracted neural signals are sent to the imaging display and data output module, where the discrete neural signal points are converted into continuous two-dimensional images using the interpolation algorithm.

[0118] In summary, the present invention provides a high-resolution peripheral nerve imaging system based on acoustic-electric coupling, which can achieve high spatio-temporal resolution detection of peripheral nerve electrical activities at the millimeter and millisecond levels. This system integrates advanced electrical stimulation and acoustic-electric coupling technologies, the information-maximization independent component analysis algorithm, and an efficient imaging processing module, and has significant advantages such as high precision and strong adaptability. Its innovative design not only provides a new technical means for the diagnosis and treatment of peripheral nerve diseases, but also lays a solid scientific foundation for basic research in the field of neuroscience, and has broad clinical application prospects and research value.

[0119] Of course, the above are only typical examples of the present invention. In addition, the present invention can also have many other specific implementation manners. Any technical solutions formed by equivalent replacement or equivalent transformation fall within the scope of protection required by the present invention.

Claims

1. A high-resolution peripheral nerve imaging system based on acoustic-electric coupling, characterized in that: It includes an electrical stimulation induction module, an ultrasonic scanning module, a signal processing module, a signal output module, an electrical stimulation component, an ultrasonic transducer component, and an imaging device; The electrical stimulation induction module is used to control the operation of the electrical stimulation component. The electrical stimulation induction module is configured with a steady-state induction strategy. The steady-state induction strategy is used to configure corresponding electrical stimulation parameters according to the type information of the peripheral nerve, and control the corresponding electrical stimulation component to operate according to the electrical stimulation parameters so as to generate a steady-state potential in the nerve fiber; The ultrasonic scanning module is used to control the operation of the ultrasonic transducer. The ultrasonic scanning module is configured with an ultrasonic configuration strategy and an acoustic-electric reception strategy. The ultrasonic configuration strategy is used to generate an ultrasonic configuration instruction to control the ultrasonic transducer to transmit focused ultrasonic waves in the target detection area. The acoustic-electric reception strategy is configured with preset reception constraint conditions, and determines a high-frequency ultrasonic encoded electrical signal based on acoustic-electric coupling from the feedback waveform signal according to the reception constraint conditions; The signal processing module includes a signal processing unit and an enhancement extraction unit. The signal processing unit is configured with a signal processing strategy. The signal processing strategy is used to process the high-frequency ultrasonic encoded electrical signal to generate a low-frequency bioelectrical signal. The enhancement extraction unit is configured with an enhancement extraction strategy. The enhancement extraction strategy is used to process the low-frequency bioelectrical signal to extract the target nerve signal; The signal output module is used to generate an imaging output signal according to the target nerve signal to the corresponding imaging device; The enhancement extraction strategy includes: Step C1, initialize the demixing matrix; Step C2, perform mean removal processing and whitening processing on the low-frequency bioelectrical signal; Step C3, determine whether the preset convergence condition is satisfied. If the convergence condition is satisfied, enter step C5. If the convergence condition is not satisfied, enter step C4; Step C4, update the demixing matrix by the gradient ascent method, and return to step C3; Step C5, determine each corresponding independent signal component according to the current demixing matrix, and screen the corresponding independent signal components through a preset screening sub-condition to obtain a nerve signal component. The target nerve signal is a set of nerve signal components; The update rule of the gradient ascent method is: , where is the updated demixing matrix, is the demixing matrix before update, and is the identity matrix, is the result of applying a non-linear function to , is the independent signal component, and , is the whitened low-frequency bioelectric signal.

2. The high-resolution peripheral nerve imaging system based on acoustic-electric coupling according to claim 1, wherein: The steady-state induction strategy includes a parameter information configuration table. The parameter information configuration table stores a number of electrical stimulation parameters and corresponding adjustment instruction sequences. The adjustment instruction sequence includes a number of parameter adjustment instructions for adjusting the electrical stimulation parameters. The electrical stimulation parameters are indexed by the type information of the peripheral nerve. The steady-state induction strategy includes: Step A1, obtain the type information of the peripheral nerve; Step A2, obtain the corresponding electrical stimulation parameters according to the type information; Step A3, control the operation of the electrical stimulation component according to the electrical stimulation parameters; Step A4, judge whether a steady-state potential is generated through the feedback current corresponding to the recording electrode. If a steady-state potential is generated, maintain the electrical stimulation parameters and end the steady-state induction strategy. If a steady-state trigger potential is not generated, enter step A5; Step A5, adjust and update the electrical stimulation parameters according to the order of the parameter adjustment instructions in the adjustment instruction sequence, and return to step A3.

3. The high-resolution peripheral nerve imaging system based on acoustic-electric coupling according to claim 1, wherein: The described reception constraint condition is configured such that the frequency of the high-frequency ultrasonic encoded electrical signal falls within a constraint frequency range and the amplitude of the high-frequency ultrasonic encoded electrical signal falls within a preset constraint amplitude range. The constraint frequency range is generated according to the working frequency of the focused ultrasonic wave, and the constraint amplitude range is generated according to the steady-state potential.

4. The high-resolution peripheral nerve imaging system based on acousto-electric coupling according to claim 1, wherein: The described signal processing strategy includes: Step B1, configuring a preset decimation acquisition frequency to configure the sampling frequency of the high-frequency ultrasonic encoded electrical signal; Step B2, configuring a filtering frequency to perform filtering processing on the acquired high-frequency ultrasonic encoded electrical signal; Step B3, processing the filtered high-frequency ultrasonic encoded electrical signal through a Hilbert transform algorithm to obtain a low-frequency bioelectrical signal.

5. The high-resolution peripheral nerve imaging system based on acoustic-electric coupling according to claim 1, characterized in that: The described screening sub-condition is configured such that the signal characteristics of the independent signal component conform to the preset target nerve signal characteristics.

6. The high-resolution peripheral nerve imaging system based on acoustic-electric coupling according to claim 1, wherein: Step C5 further includes extracting the signal characteristics of the nerve signal component, and the signal characteristics include time-domain characteristics and frequency-domain characteristics.

7. The high-resolution peripheral nerve imaging system based on acoustic-electric coupling according to claim 1, wherein: The signal output module further includes a data compensation strategy, and the data compensation strategy is configured with an interpolation sub-strategy to generate the imaging output signal according to the target nerve signal.

Images