An electrophysiological impedance measurement and correction method and system

Abstract

The application discloses an electrophysiological impedance measurement and correction method and system. The method switches a reference electrode driving circuit to output a fixed reference voltage, injects a test current through a digital-to-analog converter to a first input end and a second input end of an instrument amplifier through a current limiting resistor, executes at least twice excitation with different voltage configurations; samples a differential output voltage and a common-mode voltage under each excitation configuration, restores absolute potentials of each input end, determines a current flowing through a contact impedance according to voltage configuration values of each excitation configuration and the restored absolute potentials, constructs a linear equation group for solving, and obtains a contact impedance value of each electrode. The application multiplexes a digital-to-analog converter in an existing acquisition channel to generate a known test current, transfers impedance measurement from a frequency domain to a direct current algebra domain, eliminates dependence on a special alternating current constant current source circuit, and realizes quantitative measurement of a multi-electrode contact impedance at low hardware cost.

Description

Technical Field

[0001] This application relates to the field of electrophysiological signal acquisition technology, and in particular to an electrophysiological impedance measurement and correction method and system. Background Technology

[0002] In electrophysiological signal acquisition systems, the contact impedance between the electrode and the skin is a key factor affecting signal quality. Asymmetric contact impedance converts common-mode interference into differential-mode noise, reducing the signal-to-noise ratio (SNR) of the acquired signal. Existing impedance measurement schemes typically require a dedicated AC constant current source circuit to inject a test signal at a specific frequency and extract impedance information in the frequency domain. This approach not only increases the hardware complexity and cost of the system but also necessitates interrupting normal signal acquisition during measurement or requiring a separate measurement channel. Furthermore, frequency domain measurements rely on signal processing steps such as lock-in amplification or fast Fourier transform, placing additional demands on computational resources. Therefore, achieving independent quantitative measurement of contact impedance without adding dedicated hardware circuitry is a pressing technical problem in this field. Summary of the Invention

[0003] To address the shortcomings of existing technologies that require dedicated constant current source circuits and interrupt normal data acquisition for electrode contact impedance measurement, this application provides a method and system for electrophysiological impedance measurement and correction. This method utilizes a digital-to-analog converter in the existing acquisition channel to generate a known test current via a current-limiting resistor. By combining multiple independent voltage configurations and common-mode voltage restoration, a linear equation system is constructed to analytically determine the contact impedance. This enables independent quantitative measurement of multi-electrode contact impedance without adding dedicated hardware. Specifically, this application provides the following technical solutions: This application provides a method for measuring and correcting electrophysiological impedance, applied to an electrophysiological acquisition system including a first digital-to-analog converter (DAC), a second DAC, a current-limiting resistor, a reference electrode driving circuit, and an instrumentation amplifier. The first DAC and the second DAC are respectively connected to a first input terminal and a second input terminal of the instrumentation amplifier via the current-limiting resistor. The first input terminal and the second input terminal form contact impedances with human skin, respectively. The method includes: switching the reference electrode driving circuit to output a fixed reference voltage, and injecting test current into the first input terminal and the second input terminal of the instrumentation amplifier through the first DAC and the second DAC via the current-limiting resistor, respectively, performing at least two excitations with different voltage configurations; under each excitation configuration, sampling the differential output voltage and common-mode voltage of the instrumentation amplifier, and restoring the absolute potentials of the first input terminal and the second input terminal based on the differential output voltage and the common-mode voltage; determining the current flowing through each contact impedance according to the voltage configuration values ​​of each excitation configuration and the restored absolute potentials, constructing and solving a system of linear equations to obtain the contact impedance value of each electrode.

[0004] Optionally, a calibration step is also included: before performing the excitation, a known voltage is output through the digital-to-analog converter and applied to a precision reference load via the current-limiting resistor to obtain the accurate value of the current-limiting resistor and the compensation coefficient.

[0005] Optionally, obtaining the accurate value of the current-limiting resistor includes: connecting a precision reference load to the first input terminal or the second input terminal, outputting multiple known voltages through the digital-to-analog converter, sampling the corresponding current response, and calculating the measured value of the current-limiting resistor based on Ohm's law.

[0006] Optionally, when determining the current flowing through each contact impedance, the shunting of the test current by the bias resistor at the input of the instrumentation amplifier is included in the current determination.

[0007] Optionally, under each excitation configuration, the differential output voltage and common-mode voltage are sampled multiple times and averaged to reduce the noise of the absolute potential.

[0008] Optionally, the method further includes a signal correction step: based on the contact impedance value and the voltage division relationship between the bias resistor at the input of the instrumentation amplifier and each contact impedance, the common-mode to differential-mode coefficient that converts the common-mode interference into differential-mode components through each contact impedance is determined, and the product of the common-mode to differential-mode coefficient and the real-time common-mode voltage AC component is subtracted from the differential output signal of the instrumentation amplifier to obtain the corrected electrophysiological signal.

[0009] Optionally, in the signal correction step, the excitation, sampling, and solving are periodically re-executed to update the contact impedance value, and the differential output signal is continuously corrected based on the updated contact impedance value.

[0010] Optionally, the step of performing at least two excitations with different voltage configurations further includes adaptive configuration optimization based on the solution results of the previous round: when the condition number of the coefficient matrix of the linear equation system exceeds the optimization threshold, the output offset direction of the second digital-to-analog converter in the next round of excitation is automatically adjusted to maximize the absolute value of the determinant of the coefficient matrix.

[0011] Optionally, the step of constructing and solving the linear equation system further includes an electrode contact quality assessment based on the solved residuals: substituting the solved contact impedance values ​​back into each configuration equation to calculate the residuals, determining the contact quality index based on the normalized root mean square value of the residuals, and performing hierarchical control on the signal correction step according to the contact quality index.

[0012] Optionally, the triggering time for the periodic retest is determined based on impedance drift trend prediction: a linear regression is performed on a sliding window of historical measurement results to obtain the drift slope, and the timing of impedance deviation from the tolerance is predicted based on the drift slope. When the predicted timing is earlier than the planned retest timing, the retest is triggered in advance.

[0013] This application also provides an electrophysiological impedance measurement and correction system, comprising: an excitation configuration module for switching a reference electrode driving circuit to output a fixed reference voltage and injecting test current into the first and second input terminals of an instrumentation amplifier via a digital-to-analog converter and a current-limiting resistor, performing at least two excitations with different voltage configurations; a potential restoration module for sampling the differential output voltage and common-mode voltage of the instrumentation amplifier under each excitation configuration, and restoring the absolute potentials of the first and second input terminals based on the differential output voltage and common-mode voltage; and an impedance solving module for determining the current flowing through each contact impedance according to the voltage configuration value of each excitation configuration and the restored absolute potential, constructing and solving a system of linear equations to obtain the contact impedance value of each electrode.

[0014] Optionally, it also includes a signal correction module, used to determine the common-mode to differential-mode coefficient based on the contact impedance value and the voltage division relationship between the bias resistor at the input of the instrumentation amplifier and each contact impedance, and to subtract the product of the common-mode to differential-mode coefficient and the real-time common-mode voltage AC component from the differential output signal of the instrumentation amplifier.

[0015] The beneficial effects of this application are as follows: By reusing the existing digital-to-analog converter and instrumentation amplifier front-end channels in the acquisition system, the output voltage of the digital-to-analog converter is converted into a known test current using a precision current-limiting resistor. This is combined with a reference electrode drive circuit switching to a fixed reference voltage to establish an equation reference base. The absolute potential of each input terminal is restored using a combination of differential output voltage and common-mode voltage. Through multiple independent voltage configurations, a linear equation system is constructed to directly and analytically obtain the contact impedance of each electrode. This method eliminates the dependence on a dedicated AC constant current source circuit, shifts impedance measurement from the frequency domain to the DC algebraic domain, avoids frequency domain processing steps such as lock-in amplification or fast Fourier transform, and allows normal acquisition to resume after the measurement is completed within a time-division window. This achieves independent quantitative measurement of multi-electrode contact impedance with low hardware cost. Attached Figure Description

[0016] Figure 1 This is a schematic diagram of the overall process of the electrophysiological impedance measurement and correction method provided in the embodiments of this application.

[0017] Figure 2 This is a schematic diagram of the equivalent circuit for impedance measurement provided in an embodiment of this application.

[0018] Figure 3 This is a schematic diagram of the incentive configuration sub-process provided in an embodiment of this application.

[0019] Figure 4 This is a schematic diagram of the sampling and potential reduction sub-process provided in an embodiment of this application.

[0020] Figure 5 This is a schematic diagram of the equation solving process provided in the embodiments of this application.

[0021] Figure 6 This is a waveform diagram showing the comparison of the correction effect provided in the embodiments of this application.

[0022] Figure 7 This is a schematic diagram of the signal correction and dynamic update process provided in the embodiments of this application.

[0023] Figure 8 This is a schematic diagram of the system module architecture provided in the embodiments of this application. Detailed Implementation

[0024] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below in conjunction with the accompanying drawings and specific embodiments. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0025] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this application, "multiple" means two or more, unless otherwise explicitly specified.

[0026] The electrophysiological impedance measurement and correction method provided in this application is applicable to various electrophysiological signal acquisition devices, such as electromyography (EMG) signal acquisition devices, electrocardiogram (ECG) signal acquisition devices, electroencephalogram (EEG) signal acquisition devices, and multi-channel biopotential recorders. It can also be applied to portable wearable biosensors, implantable neural interface front-end modules, and other device forms. The hardware platform on which the method operates includes an electrophysiological signal acquisition system with a digital-to-analog converter and an instrumentation amplifier front-end. This system picks up weak bioelectrical signals through electrodes attached to the surface of human skin. The contact interface formed between the electrodes and the skin has contact impedance, and the magnitude and symmetry of this contact impedance directly affect the quality of the acquired signal.

[0027] The signal acquisition mode parameters differ for different types of electrophysiological signals: the sampling rate for electromyography (EMG) signals is typically between [value missing]. to Within range, signal bandwidth to signal amplitude to The sampling rate of electrocardiogram (ECG) signals is typically at... to Within range, signal bandwidth to signal amplitude to The sampling rate of electroencephalogram (EEG) signals is typically in the range of... to Within range, signal bandwidth to signal amplitude to .

[0028] The impedance measurement window (approximately) to The sampling interval relative to each signal type (EMG minimum is...) @ ECG approximately @ EEG approximately @ For ECG and EEG scenarios, the single measurement window should account for less than 50%, and the data lost within the measurement window can be recovered by interpolation of the sampling points before and after. For high sampling rate EMG scenarios, the measurement window can be arranged in the resting period between two electromyographic activities to reduce information loss.

[0029] The electrophysiological signal acquisition system includes at least two digital-to-analog converters, connected to the first and second input terminals of an instrumentation amplifier via current-limiting resistors, respectively. The instrumentation amplifier has high input impedance characteristics, and its input terminal is connected to a midpoint voltage source via a bias resistor. A reference electrode drive circuit is connected to the third electrode via an analog switch, forming a negative feedback loop to suppress common-mode interference in normal acquisition mode, and switching to a fixed reference voltage output in impedance measurement mode. For multi-channel electrophysiological acquisition systems, the method can be time-division multiplexed for each channel: each channel shares the same reference electrode and reference electrode drive circuit; excitation configuration, sampling, and solving for each channel are performed sequentially; each channel independently obtains the contact impedance values ​​of its two electrodes and independently performs correction calculations. The total measurement window of the channel system is approximately (Excluding the third configuration), for an 8-channel system, approximately It still satisfies the time-sharing reuse constraint.

[0030] like Figure 1 As shown, the electrophysiological impedance measurement and correction method provided in this embodiment includes the following steps: S000, the known voltage is output through the digital-to-analog converter and applied to the precision reference load through the current-limiting resistor to obtain the accurate value and compensation coefficient of the current-limiting resistor.

[0031] The calibration step is performed when the system is powered on or when environmental conditions change, and is used to eliminate the impact of current-limiting resistor manufacturing tolerances and inherent errors of the analog-to-digital converter on the accuracy of subsequent impedance measurements.

[0032] The calibration process includes two sub-processes: analog-to-digital converter (ADC) linear calibration and precise measurement of the current-limiting resistor. During ADC linear calibration, the ADC sequentially outputs multiple known voltage values ​​covering its full-scale range. The ADC samples each known voltage value to obtain multiple sets of correspondences between ideal input values ​​and actual readings. The number of known voltage points is typically between 5 and 20; this embodiment uses 10 equally spaced voltage points. to by The voltage is increased incrementally. The coverage area must include all voltage values ​​that may occur in subsequent measurements. In this embodiment, the potential range of the current-limiting resistor and the electrode node is [missing information]. Right now to All values ​​are within the calibration range. Using the ideal input value as the independent variable and the actual reading as the dependent variable, a least-squares linear fit is performed to extract the gain error coefficient. and offset error The gain error coefficient Typically, the value ranges from 0.99 to 1.01; in this embodiment, it is 0.99851. The offset error... Usually in to Within the scope, this embodiment takes The same calibration process is performed independently on the auxiliary analog-to-digital converter to obtain its independent calibration coefficients. and All subsequent analog-to-digital converter readings are converted back to the true voltage using the following formula: For example, when the actual reading of the analog-to-digital converter is At that time, the true voltage after calibration and restoration is , with the ideal input at this calibration point A perfect match.

[0033] During the precise measurement of the current-limiting resistor, a precision reference load is connected to either the first or second input terminal. The precision reference load has an accuracy of 0.1%, and its nominal resistance value can be, for example, [missing value]. , In this embodiment, we take... The digital-to-analog converter (DAC) outputs multiple known voltages applied to a series path consisting of a current-limiting resistor and a precision reference load. The DAC samples the voltage drop across the current-limiting resistor and the precision reference load, respectively. The measured value of the current-limiting resistor is calculated using the series voltage divider relationship. in For the known resistance value of the precision reference load, The voltage drop across the current-limiting resistor. This serves as a precision reference for the voltage drop across the load. For example, for the first current-limiting resistor... Perform calibration: when the digital-to-analog converter outputs... At that time, the voltage across the current-limiting resistor after calibration and restoration is The voltage at the connection point between the current-limiting resistor and the precision reference load is The voltage at the lower end of the precision reference load is ,but , , The measured value is relative to the nominal value. The deviation is 0.07%, which meets the measurement accuracy requirements. Regarding the second current-limiting resistor... Perform the same calibration procedure: switch the precision reference load to the second input path when the digital-to-analog converter outputs... At that time, the voltage across the upper end of the second current-limiting resistor after calibration and restoration is The voltage at the connection point is The voltage at the lower end of the precision reference load is ,but , , The measured value is relative to the nominal value. The deviation is 0.14%, which meets the measurement accuracy requirements. It will be used in subsequent S300 steps. (The average value of multiple measurements) is used in accurate current calculations.

[0034] In another embodiment, the precise value of the current-limiting resistor can also be obtained by measuring it with an external precision source meter at the factory and writing it into a one-time programmable memory. This method does not require onboard precision reference load hardware and is suitable for applications with lower requirements for temperature drift tracking. In another embodiment, the current-limiting resistor can also be implemented using a digital potentiometer. The digital potentiometer receives the resistance value setting command from the controller through a serial peripheral interface and can dynamically adjust the resistance value under software control. When using a digital potentiometer, the precise measurement process of the current-limiting resistor in the calibration step becomes directly reading the current code value in the digital potentiometer register and converting it into a resistance value according to its nominal step accuracy, without requiring precision reference load hardware. The resistance accuracy of this method depends on the step accuracy and absolute tolerance of the digital potentiometer and is suitable for scenarios where the excitation current amplitude needs to be adjusted according to the electrode impedance range during operation.

[0035] The calibration process simultaneously reads the midpoint voltage. The actual value is restored to the true value by the calibration coefficient of the analog-to-digital converter, and the midpoint voltage is... Usually in to Within the scope, the system adopted in this embodiment Single power supply The bias resistor and The precise value also needs to be included in the calibration parameters. Since the bias resistor is directly soldered to the printed circuit board and there is no convenient onboard measurement path, this embodiment uses a thin-film resistor with a 0.1% tolerance rating as... and Use its nominal value directly. The calibration parameters are included in the calculation. For higher accuracy, a precision reference load can be temporarily switched into the bias resistor path during the current-limiting resistor calibration process to perform the same ratio measurement. All calibration parameters are written to non-volatile memory for later retrieval. The list of calibration parameters includes: main analog-to-digital converter calibration coefficients. and Auxiliary analog-to-digital converter calibration coefficient and Measured value of the first current-limiting resistor Measured value of the second current-limiting resistor Bias resistor value and Measured value of midpoint voltage The set of calibration parameters, including the pre-calculated code values ​​of the digital-to-analog converters for each excitation configuration, forms the basis for all known quantities in subsequent impedance measurements and signal corrections.

[0036] S100, the reference electrode drive circuit is switched to output a fixed reference voltage, and test current is injected into the first and second input terminals of the instrumentation amplifier through the digital-to-analog converter and the current-limiting resistor, performing at least two excitations with different voltage configurations.

[0037] like Figure 3 As shown, the execution process of the excitation configuration adopts a three-step strategy of "first zeroing, then switching, then excitation" to suppress switching transients. In the first step, the controller presets the output voltage of the digital-to-analog converter to the midpoint voltage. This minimizes the current changes at each node before and after the switching, since the output of the digital-to-analog converter is equal to the DC bias potential at the input of the instrumentation amplifier (which is approximately equal to...). Since no current flows through the current-limiting resistor, the switching will not cause a current jump. In the second step, the controller drives the analog switch to switch the reference electrode drive circuit from negative feedback mode to output a fixed reference voltage. The fixed reference voltage In this embodiment, equal to .because Furthermore, the human body node potential was clamped to negative feedback before the switch. Nearby, the potential difference between the fixed voltage and the human body node is extremely small, and no sudden current will be generated on the third contact impedance. After the switching is completed, wait for the reference electrode drive circuit loop to disconnect and stabilize. In this embodiment, wait... The third step is to transfer the output of the digital-to-analog converter from... Once the target excitation voltage is adjusted, a test current begins to build up in the circuit. This three-step strategy ensures that the current at each node is approximately zero at any switching moment, preventing voltage spikes caused by large current transients through electrode contact impedance from interfering with the instrumentation amplifier.

[0038] Controller configuration for the first stimulus: First digital-to-analog converter output voltage The output voltage of the second digital-to-analog converter ,in The excitation step voltage. Usually in to Within the scope, this embodiment takes The upper limit of this value is determined by the linear operating range of the instrumentation amplifier input, and the lower limit is determined by the quantization resolution of the analog-to-digital converter. The target voltage of the digital-to-analog converter configured for each excitation is pre-calculated as an integer code value and stored in non-volatile memory after the calibration step. The code value calculation formula is as follows: ,in and For the offset and gain calibration coefficients of the digital-to-analog converter, This is the full-scale code value. Pre-calculated code values ​​avoid real-time floating-point operations, ensuring that the switching time of the stimulus configuration depends only on the transmission speed of the serial peripheral interface (typically less than...). The voltage configuration for the first excitation is as follows: , This forms a symmetrical positive excitation, with the test current flowing into the human body from the first digital-to-analog converter side and flowing out from the second digital-to-analog converter side.

[0039] After configuration, wait for circuit setup time. This ensures that the voltage at each node reaches a steady state. The settling time is determined by the equivalent RC time constant from the output of the digital-to-analog converter to the input of the instrumentation amplifier. in This is the parallel value of the current-limiting resistor and the bias resistor. The equivalent capacitance at the input of the instrumentation amplifier includes the parasitic capacitance of the printed circuit board, the electrode-skin interface capacitance, and the input capacitance of the instrumentation amplifier. The equivalent capacitance... Usually in to Within the scope, this embodiment takes Set the waiting time to... At this point, the steady-state error is less than 0.67%. For example, when... , hour, , , .

[0040] After the first excitation is established, the controller sends a sampling ready signal to the sampling and potential restoration steps. After sampling, the controller directly configures the second excitation without first restoring the digital-to-analog converter to the midpoint voltage. The basis for the "no need to return to zero" design is that only the digital-to-analog converter voltage changes between the two configurations; the reference electrode drive circuit still maintains a fixed voltage output mode, the circuit topology remains unchanged, and only the excitation amplitude changes in steps, thus avoiding large transients such as loop disconnection / reconstruction. The second excitation configuration is: the output voltage of the first digital-to-analog converter... The output voltage of the second digital-to-analog converter The second excitation changes the output voltage of the second digital-to-analog converter, making the current vectors generated by the two excitations linearly independent, thus ensuring that the subsequent system of equations has a unique solution. Similarly, a setup time is required. Then a second sampling ready signal is issued.

[0041] The design of the two excitation configurations must satisfy the linear independence condition that the determinant of the coefficient matrix of the equation system is not zero. For example, the current generated by the two configurations is approximately: First configuration , Second configuration , Determinant The linear independence condition holds.

[0042] After all excitation configurations are complete, the controller restores the digital-to-analog converter output to the midpoint voltage, switches the reference electrode drive circuit back to negative feedback mode, and resumes normal signal acquisition after the loop reconstruction stabilizes. The total measurement window for the at least two excitations is approximately [value missing] in this embodiment. This satisfies the short window constraint of time-division multiplexing. The time budget of the total measurement window is decomposed as follows: Preparation stage (reference electrode drive circuit loop disconnected and stabilized). First configuration (establishment) +Sampling ) Second configuration (establishment) + Sampling ) Recovery phase (ring road reconstruction stabilized) ,total (Including marginal margin). When the third configuration is performed to solve for the third contact impedance, its settling time depends on the RC time constant of the third electrode path. ,in For electrode-skin interface capacitance. When , hour, The waiting time for the third configuration setup is (The steady-state error is allowed to be less than 5%, corresponding to) The maximum steal is approximately 5% of the target offset, which is approximately The relative error contribution to the solution of Z3 is approximately For applications where Z3 is primarily used for threshold-level determination of electrode contact quality, its accuracy is sufficient. The total measurement window, including the third configuration, is approximately... still satisfied Time-division multiplexing constraints within a certain range. When the electrode interface capacitance is large, the third configuration can be executed separately from the first two configurations: the first two configurations are executed every sampling period to quickly update Z1 and Z2, while the third configuration is executed only once every few seconds to update Z3.

[0043] In another embodiment, the excitation configuration can also be achieved by a soft-start method where the output of the digital-to-analog converter linearly ramps the voltage from the midpoint to the target value. This method produces smaller switching transients and is suitable for application scenarios with high requirements for signal continuity.

[0044] Optionally, the incentive configuration step further includes adaptive configuration optimization based on the solution results of the previous round. Specifically, the condition number of the coefficient matrix output by the equation solving step... The condition number is passed to the incentive configuration step. Exceeding the optimization threshold In the next round of measurement, the controller automatically adjusts the second digital-to-analog converter output offset of the second excitation configuration. The optimized threshold Typically, the condition number is in the range of 15 to 30; in this embodiment, it is set to 20. The condition number reflects the degree of directional difference between the current vectors generated by the two excitation configurations. A larger condition number indicates that the two vectors are closer to being collinear, leading to a more severe amplification of the solution error. The adaptive adjustment strategy aims to maximize the directional difference of the current vectors in the next two configurations, specifically as follows: in For the first configuration of the previous round, the current flowing through the second contact impedance, To ensure the adjustable upper limit of the excitation step voltage, this embodiment takes... The strategy reverses the direction of the second digital-to-analog converter current in the second configuration compared to the first configuration, thereby maximizing the absolute value of the determinant of the coefficient matrix. When the condition number... Further exceeding the high threshold At the same time, the excitation step voltage is increased to improve the signal-to-noise ratio, wherein the increased excitation step voltage is... The high threshold Typically, the value is in the range of 50 to 80; in this embodiment, it is set to 60.

[0045] For example, suppose the previous round of measurements yielded... , (Severely asymmetric), the condition number is calculated. Exceeding the optimization threshold But not achieved The first configuration in the last round (A negative value indicates that the current flows from the human body to the IN- terminal), then The output voltage of the second digital-to-analog converter in the second configuration of the next round is adjusted to... This configuration allows the second digital-to-analog converter to inject a positive current into the second input terminal (opposite to the negative current in the first configuration), maximizing the difference in current vector direction between the two configurations. After this adjustment, the expected condition number is reduced to below approximately 10, improving the solution accuracy. Continuing with this approach... Unchanged (because) In another embodiment, three fixed configuration schemes can be pre-calculated to correspond to low impedance, medium impedance, and high impedance scenarios, respectively, and the corresponding configuration group can be selected based on the impedance value range obtained in the previous round of calculation.

[0046] S200, under each excitation configuration, sample the differential output voltage and common-mode voltage of the instrumentation amplifier, and restore the absolute potentials of the first input terminal and the second input terminal based on the differential output voltage and common-mode voltage.

[0047] like Figure 4 As shown, after each excitation configuration establishes a steady state, the controller synchronously triggers the main analog-to-digital converter (ADC) and the auxiliary ADC to perform sampling. The main ADC samples the differential output of the instrumentation amplifier to obtain the raw differential reading, which is then restored to the true differential voltage using calibration coefficients. The auxiliary analog-to-digital converter samples the common-mode voltage output to obtain the raw common-mode reading, which is then calibrated to restore the true common-mode voltage. The common-mode voltage is provided by taking the midpoint between the first and second input terminals via an equivalent resistor and then isolating it through a buffer amplifier. The resistance value of the equivalent resistor must be much larger than that of the current-limiting resistor and the bias resistor to avoid creating an additional load on the measurement node that could affect measurement accuracy. In this embodiment, the equivalent resistor is... , is a current-limiting resistor The current shunting effect on the node potential is negligible, less than 0.04%, and is 50 times that of the original current. The buffer amplifier uses a unity-gain operational amplifier follower configuration, and its input impedance must be greater than [missing value]. To avoid a load effect on the voltage divider network, the output impedance must be less than [value missing]. To drive the sample-and-hold capacitor of the auxiliary analog-to-digital converter.

[0048] Under the same excitation configuration, the differential output voltage and common-mode voltage are sampled multiple times and averaged to reduce noise in the absolute potential. Number of samplings Typically, the value is in the range of 4 to 16; in this embodiment, it is taken 8 times. The differential gain of the instrumentation amplifier in impedance measurement mode. Set to a low gain value, this embodiment takes... The basis for this value is that, in impedance measurement mode, the potential difference between the two input terminals of the instrumentation amplifier is determined by the excitation current and the difference in contact impedance, typically ranging from several mV to tens of mV (as in the example above). If gain Setting too high (e.g.) Differential output This exceeds the input range of the analog-to-digital converter (in this embodiment) (Full scale) leads to saturation cutoff. hour There is sufficient margin within the linear operating range of the analog-to-digital converter. The gain switching in signal acquisition mode is... Up to 1000 to amplify weak bioelectrical signals. The arithmetic mean of each independent sample: When the noise is Gaussian white noise and the standard deviation is When the averaged noise standard deviation is reduced to For example, when the noise of a single sample is , At that time, the averaged noise decreased to The signal-to-noise ratio is improved by approximately The eight samplings were conducted in... Time consumed at sampling rate Its impact on the overall measurement window is negligible.

[0049] In an alternative implementation, median filtering can be used instead of arithmetic averaging. This method is more robust to pulse interference and is suitable for application environments with intermittent electromagnetic interference. The median filtering... After sorting the sampled values, the median value is taken as the result. No more than [number] samples were collected. It can still output the correct result even when there are outliers. In this embodiment... Median filtering can tolerate up to three outlier sampling points. Its noise suppression performance is slightly lower than the arithmetic mean (its efficiency under Gaussian noise is approximately one-third that of the average). However, in scenarios with impulse interference, its effective signal-to-noise ratio is better than the arithmetic mean. A hybrid strategy of first median and then average can also be used: first... The next sampling performs median filtering to remove extreme values, and then the remaining valid samples are averaged to balance outlier robustness and noise suppression performance.

[0050] The sampling of the main analog-to-digital converter (ADC) and the auxiliary ADC must meet the synchronization requirement. Under steady-state conditions, all nodes in the circuit have DC values, and small differences in sampling time do not affect the accuracy of the result. However, synchronous sampling can eliminate the influence of weak time-varying disturbances. This embodiment uses a dual-channel synchronous successive approximation ADC to achieve synchronous sampling and holding, with the difference in sampling time between the two channels being less than [missing value]. The required synchronization is relatively low under steady-state conditions of impedance measurement: the rate of change of voltage at each node after the excitation is established is less than [a certain value]. , The potential deviation caused by time difference is less than This is much smaller than the quantization noise and can be ignored. In another embodiment, a single analog-to-digital converter can also be used to alternately sample the differential output and common-mode voltage using a multiplexer, with the sampling interval being approximately... (Single conversion time), the corresponding potential deviation is still less than Under steady-state conditions, the accuracy loss is much smaller than the quantization noise, making it suitable for scenarios with limited hardware resources.

[0051] After sampling, the absolute potentials of each input terminal are reconstructed based on the differential output voltage and common-mode voltage. The technical motivation for this simultaneous solution is that the differential output of the instrumentation amplifier only reflects the potential difference between the two input terminals, and cannot individually determine the absolute potential of each input terminal. However, subsequent equation solving requires the absolute potentials of each input terminal to calculate the shunt current on the current-limiting resistor and bias resistor. Introducing common-mode voltage sampling provides a second observation, which, when combined with the differential voltage, can reconstruct the absolute potentials of the two unknowns (…). and The relationship between the differential output voltage and the input potential difference of the instrumentation amplifier is completely determined. The relationship between the common-mode voltage and the input terminal potential is as follows: Solving the two equations simultaneously yields the absolute potentials of the first and second input terminals: in The differential gain of the instrumentation amplifier in impedance measurement mode is, as described above, taken in this embodiment. .

[0052] For example, under the first excitation configuration, the differential output voltage of the instrumentation amplifier Common-mode voltage Gain Then the absolute potential of the first input terminal is restored. absolute potential of the second input terminal Under the second incentive configuration, , , restored , .

[0053] The two sets of absolute potential data, along with the corresponding excitation voltage configuration values, are passed to the equation solving step. The accuracy of the potential restoration is limited by the following factors: analog-to-digital converter quantization noise (in this embodiment, we take...). / time, averaged after 8 times Calibration coefficient The fitting residuals (in this example, the linear error after fitting 10 points is less than) This includes differential output errors caused by the limited common-mode rejection ratio of instrumentation amplifiers. And differential input Under typical operating conditions, the The relative error of the potential reconstruction corresponding to the noise is This satisfies the accuracy requirements for subsequent impedance calculations.

[0054] S300 determines the current flowing through each contact impedance based on the voltage configuration value of each excitation configuration and the restored absolute potential, constructs and solves a system of linear equations, and obtains the contact impedance value of each electrode.

[0055] like Figure 2 As shown, in the impedance measurement equivalent circuit, the first digital-to-analog converter is connected to the first current-limiting resistor. The second digital-to-analog converter is connected to the first input terminal of the instrumentation amplifier via the second current-limiting resistor. Connect to the second input terminal, bias resistor and From the midpoint voltage respectively Connected to the first input terminal and the second input terminal, first contact impedance Second contact resistance Connect each input terminal to the human body node respectively, and the third contact impedance. Connect the human body node to the reference electrode. For example... Figure 5 As shown, the equation solving step receives the absolute potential data obtained from the restoration under each excitation configuration, and combines the known excitation voltage configuration value and circuit parameters to analytically obtain the contact impedance of each electrode through algebraic operations.

[0056] When determining the current flowing through each contact impedance, the shunting of the test current by the bias resistor at the input of the instrumentation amplifier is included in the current determination. At the first input node, according to Kirchhoff's current law, the total current flowing through the first contact impedance (i.e., the contact impedance between the first input terminal and the human body, hereinafter referred to as Z1) includes the current component flowing in through the current-limiting resistor and the current component flowing in through the bias resistor: Similarly, at the second input terminal node, the total current flowing through the second contact impedance (i.e., the contact impedance between the second input terminal and the human body, hereinafter referred to as Z2) is: The accurate current model, compared to the simplified model that only considers the current-limiting resistor (i.e., The additional shunt term due to the bias resistor .when This diversion term can be ignored in some cases; however, in this embodiment... The current shunting contribution of the bias resistor is not negligible. Quantitatively, under the initial configuration, the current shunting contribution through the current-limiting resistor... The current component is via bias resistor The current component is The latter is approximately 10% of the former. If the current shunt by the bias resistor is ignored, the current calculation error is about 10%, which directly leads to a proportional error in the impedance solution, failing to meet engineering accuracy requirements. Therefore, this solution adopts an accurate model that includes the current shunt by the bias resistor.

[0057] in , The measured value for the current-limiting resistor during calibration is... , This is a bias resistor. , For the first The output voltage of the digital-to-analog converter in the secondary excitation configuration , For the first The absolute potential restored under this configuration. When the bias resistor... With current limiting resistor When they are of the same order of magnitude, the shunt current contribution of the bias resistor cannot be ignored. In this embodiment... , The two values ​​are close, and the current shunting contribution of the bias resistor can reach 10% to 50% of the current limiting resistor current under certain configurations (depending on the node potential deviation). The degree of [something] cannot be ignored.

[0058] For example, under the initial incentive configuration, , , Calculate the current flowing through Z1. Similarly... , ,calculate Under the second incentive configuration, , .

[0059] According to Ohm's law, the potential of human nodules The relationship with the potential of each input terminal is as follows: . eliminate The linear equations for Z1 and Z2 are obtained: Write the equations for each of the two excitation configurations, forming a system of linear equations with 2 rows and 2 columns: The system of linear equations is solved analytically using Cramer's rule. The determinant of the coefficient matrix is ​​calculated. ,when The system of equations has a unique solution. The absolute value of the determinant reflects the degree of linear independence between the current vectors generated by the two excitation configurations. A larger value indicates a greater difference in current direction between the two configurations, resulting in a smaller amplification effect on the solution error. The condition number of the coefficient matrix. Quantitatively describe the amplification factor of input data error on the solution result: when the condition number is At that time, the relative error of the input potential data The relative error of the impedance solution will be amplified. In this embodiment, the condition number is approximately 9.4, meaning the analog-to-digital converter... The impact of noise on the impedance solution is amplified by approximately 9.4 times, corresponding to an impedance error of approximately... This meets the engineering accuracy requirements. When When the contact impedance is such that the formulas for solving each contact impedance are: For example, substituting the aforementioned values, . The molecules are , . The molecules are , .

[0060] After obtaining Z1 and Z2, substitute them into any configuration data to calculate the human body node potential. Verify using another configuration set of data. The calculation results from both paths are consistent, verifying the correctness of the solution.

[0061] In the first two excitation configurations, the fixed reference voltage equal to the midpoint voltage Human nodal potential It also approaches the same level under the clamping effect of the reference electrode drive circuit. This results in current flowing through the third contact impedance Z3. Approaching zero, Z3 cannot be solved. To solve Z3 independently, a third excitation configuration is performed: the fixed reference voltage is set to a value different from the midpoint voltage. In this embodiment, The offset The selection is based on: the third contact impedance Typical value Up to dozens Level 1, its clamping capability is strong, small offset (such as ) will be The low resistance cancellation leads to Flowing through The current is extremely small and the signal-to-noise ratio is insufficient; The offset ensures The current on reaches Magnitude, and The path currents are of the same order of magnitude to ensure solution accuracy. Simultaneously, the output of the second digital-to-analog converter is set to... (Breaking current symmetry), the output of the first digital-to-analog converter is maintained. The third configuration causes the reference electrode driving circuit to inject current into the human body node through the third contact impedance, thereby increasing the potential of the human body node and causing it to deviate. This makes both the numerator and denominator of the Z3 formula non-zero.

[0062] After sampling and potential restoration are completed under the third configuration, the current values ​​for the third configuration are calculated using the obtained Z1 and Z2. and and by Calculate the potential of the human body node under the third configuration. According to Kirchhoff's current law for the human body node, the current flowing through the third contact impedance is equal to the negative of the sum of all other currents flowing into the human body node: Combining Ohm's Law Solving for the third contact impedance yields: For example, in the third configuration ( , , ), calculated , (All are negative values, indicating that the current flows out from the human body node and reaches each input terminal through Z1 and Z2). . Substitute into the Z3 formula: Kirchhoff's current law was verified on the third configuration: The law of conservation of current holds true.

[0063] The equation solving steps also include a physical plausibility check: when any of Z1, Z2, or Z3 is negative or exceeds... When the upper limit is reached, the measurement result is marked as abnormal and a remeasurement is triggered; when the condition number of the coefficient matrix exceeds a preset threshold, the confidence level of the marked result is reduced. In this embodiment, the condition number is approximately 9.4, which meets the engineering criterion of less than 100. The remeasurement is executed according to a limited retry strategy: the maximum number of retries allowed within a single impedance measurement cycle. Typically, the number of trials is between 2 and 5; in this embodiment, it is 3. Each retry restarts the complete stimulus-sampling-solution sequence from the stimulus configuration step. When consecutive... If all retries are deemed abnormal, the current measurement is abandoned and an abnormality flag is output. The signal correction step continues to correct using the impedance value from the last successful measurement. If consecutive... If an anomaly is triggered in all measurement cycles (5 cycles in this embodiment), the system enters degraded mode: impedance correction calculation is paused and an electrode contact failure alarm is output, prompting the operator to reattach the electrode or replace the conductive paste. In degraded mode, the system still uses the shortest retest interval. It continuously attempts to measure, and once a valid result is obtained, it automatically exits the degradation mode and resumes the correction calculation.

[0064] In another embodiment, the linear equation system can also be solved using Gaussian elimination, a method equivalent to Cramer's rule in terms of numerical stability, suitable for scenarios where embedded processors lack hardware dividers. When the excitation configuration is executed three or more times, the equation system becomes an overdetermined system, and the least squares method can be used to improve measurement robustness. The least squares solution is... For a 2x2 array The matrix can still be calculated analytically. The noise suppression principle of the overdetermined equation system is similar to that of multiple sampling averaging: Sub-configuration generated Solving an equation for two unknowns is equivalent to performing a process on the result. Double oversampling reduces the noise standard deviation to that of a positive definite system. This represents a multiple. For example, when performing a 4-fold excitation configuration, the standard deviation of the impedance noise of the least-squares solution is reduced to [a factor] compared to a 2-fold positive definite solution. This increases accuracy by approximately 30%, a several-fold increase. The additional time cost for configuration is [per unit]. The total measurement window for the four configurations is approximately Still Within the constraints. The overdetermined system also has outlier detection capability: if the residual of a certain configuration deviates significantly from other configurations, it can be determined that the sampling was affected by transient disturbances and removed from the least squares fitting, further improving the robustness of the solution.

[0065] Optionally, the equation solving step further includes an electrode contact quality assessment based on the solved residuals. Specifically, the solved first contact impedance Z1 and second contact impedance Z2 are substituted back into the equations for each excitation configuration to calculate the residuals for each configuration: in To incentivize configuration frequency, a normalized contact quality index is calculated based on the residuals. in The average differential potential amplitude is used as the normalization benchmark. The contact quality index (CQI) ranges from 0 to 1. A value close to 1 indicates that the equations are sufficiently self-consistent, i.e., the electrode contact is good; a value below the threshold indicates an abnormality at the electrode interface. The contact quality index adopts a dual-threshold, three-level classification: when the CQI is greater than 0.95, it is considered normal, and the signal correction step is executed in standard mode; when the CQI is in the range of 0.80 to 0.95, it is considered a warning level, the impedance retest cycle is shortened, and confidence weighting is applied to the correction result; when the CQI is less than 0.80, it is considered an abnormal level, the correction operation is paused, the previous effective correction coefficient is maintained, and an electrode contact alarm signal is output.

[0066] The control logic for the signal correction step in the three-level classification is as follows. In the normal level (CQI greater than 0.95), the signal correction step uses the common-mode to differential-mode conversion coefficient. A standard correction operation is performed, and the corrected signal is directly output to the subsequent processing chain. The retest interval maintains the current value determined by the adaptive strategy. At the warning level (CQI between 0.80 and 0.95), the impedance value output from the equation solving step is still used to update... However, the corrected signal is multiplied by the confidence weight factor. Perform weighted mixing: The confidence weight factor Obtained from the CQI linear mapping: When CQI equals 0.95 (Full Trust Correction) When CQI equals 0.80 (No correction). The downweighting and mixing ensures that the correction amount gradually decreases in the early stages of electrode contact degradation, avoiding signal quality degradation caused by correction errors due to inaccurate impedance values. Simultaneously, the retest interval is forcibly shortened to half of the current value, accelerating impedance updates. At the abnormal level (CQI below 0.80), the signal correction step suspends correction calculations based on the current impedance value, and the output signal remains the uncorrected original differential signal. The common-mode transfer coefficient The system locks the CQI to the valid value from the last normal-level measurement, and will continue to correct it from this initial value once the CQI returns to normal. Simultaneously, the system outputs an electrode contact alarm signal to the host computer or display module, prompting the operator to check the electrode adhesion status. If the CQI is below 0.80 for three consecutive remeasurements, the remeasurement interval is forcibly set to the shortest value. And add a priority flag of "recommend replacing the electrode" to the alarm signal.

[0067] For example, when the electrode contact is good, the residual of each configuration In Magnitude In Magnitude This is far higher than the normal threshold of 0.95. The correction is fully effective. When the electrodes begin to loosen and the residual increases to... At 10%, , The correction is reduced to approximately 67% of the standard value, while the retest interval is shortened to obtain a more accurate impedance estimate as quickly as possible. In another embodiment, the potential difference values ​​of the human body nodes calculated separately for each configuration can also be used. Using it directly as a contact quality indicator is simpler to calculate but provides less information.

[0068] S400, based on the contact impedance value and the voltage division relationship between the bias resistor at the input of the instrumentation amplifier and each contact impedance, the common-mode to differential-mode coefficient that converts common-mode interference into differential-mode components through each contact impedance is determined. The product of the common-mode to differential-mode coefficient and the real-time common-mode voltage AC component is subtracted from the differential output signal of the instrumentation amplifier to obtain the corrected electrophysiological signal.

[0069] like Figure 7 As shown, the signal correction step is executed in real time after the system returns to normal signal acquisition mode. The recovery process includes the following synchronous operation: the controller restores the digital-to-analog converter output to the midpoint voltage. The reference electrode drive circuit is switched back to negative feedback mode via an analog switch, and the differential gain of the instrumentation amplifier is simultaneously switched from the measurement mode. Gain when switching to signal acquisition mode The gain switching must be performed synchronously with the mode switching of the reference electrode driving circuit to avoid acquiring residual transients of the measurement mode in a high-gain state. After the switching is completed, wait for the negative feedback loop of the reference electrode driving circuit to rebuild and stabilize. In this embodiment, wait... Afterwards, the signal acquisition mode takes effect, and the correction operation begins execution from the next valid sampling period. The gain... Typically, the range is between 10 and 1000; in this embodiment, we take... ,correspond The EMG signal is amplified to In analog-to-digital converters Sufficient margin is available within the full-scale range.

[0070] The signal correction is based on a common-mode to differential-mode mechanism. When the contact impedances of the first and second input terminals are asymmetrical, the external common-mode interference signal experiences unequal attenuation on the two voltage divider paths formed by the bias resistor and the contact impedance, causing the common-mode interference to be converted into a differential-mode component superimposed on the bioelectrical signal. The common-mode interference component in the actual potential of the first input terminal is then... Attenuation, the common-mode interference component in the actual potential of the second input terminal is attenuated by Attenuation. When At that time, the attenuation coefficients of the two channels are unequal, and after differential operation, a common-mode to differential-mode interference component remains. The correction operation is completed within each sampling period, and its execution time is constrained to a processing window shorter than a single sampling period. In this embodiment, the sampling rate... Corresponding sampling period The correction operation consists of only one multiplication and one subtraction, and its execution time on a 32-bit single-precision floating-point processor is less than [time value missing]. The sampling rate is much smaller than the sampling period, so it will not cause data processing delay. The corrected signal is output to the subsequent digital signal processing chain (such as bandpass filtering, peak detection, feature extraction, etc.) at the same rate as the original sampling. The data format is a corrected voltage value and a currently valid CQI confidence flag for each sampling point.

[0071] Pre-calculate the common-mode and differential-mode coefficients based on the aforementioned contact impedance values: The common-mode transfer coefficient The value is constant between two impedance measurements, pre-calculated and stored in a register-level variable for direct access in each sampling period. The correction operation... This involves one division and one multiplication / subtraction; the required computational accuracy is determined by the corrected residual tolerance. When On the order of 0.02 for When the magnitude is [value], the correction amount is [value]. The order of magnitude requires that the calculation error not exceed 1% of the bioelectric signal amplitude. This precision requirement corresponds to a number of significant digits not less than [number missing]. The required multiplication precision can be achieved with 16-bit fixed-point arithmetic. This embodiment uses 32-bit single-precision floating-point arithmetic to implement the correction calculation, with 23 effective mantissa bits, providing ample precision margin. In resource-constrained embedded scenarios, a 16-bit fixed-point Q15 format can also be used. and After scaling to the Q15 range and performing fixed-point multiplication and addition, the accuracy loss is... Within this range, the project requirements are met.

[0072] For example, when , , hour, .when and hour, Common-mode interference is completely suppressed and requires no correction. The voltage drop across each contact impedance is equivalent to a common-mode slip component when common-mode interference is present. The correction operation eliminates differential-mode interference introduced by contact impedance asymmetry by subtracting this component.

[0073] During each sampling period, the main analog-to-digital converter and the auxiliary analog-to-digital converter synchronously sample to obtain the real-time differential output voltage. and real-time common-mode voltage Extracting the AC component from the common-mode voltage: The AC component This includes residual common-mode interference due to incomplete suppression caused by the finite loop gain of the reference electrode drive circuit. The main sources of this common-mode interference include... Power frequency electromagnetic coupling, static charge changes caused by human movement, and radiated interference from nearby electrical equipment. The calculation is equivalent to using For reference, the DC removal operation, under steady state Since the frequency is constant, the AC component can be extracted without a filter. In another embodiment, a digital high-pass filter (with a cutoff frequency set to a constant) can also be used. (The following) From real time Extracting the AC component from this method allows for automatic tracking. The slow drift is suitable for scenarios where the midpoint voltage exhibits temperature drift. Perform correction calculations: The corrected differential electrophysiological signal The common-mode to differential-mode interference component caused by contact impedance asymmetry has been eliminated.

[0074] like Figure 6 As shown, the comparison of signal waveforms before and after correction demonstrates the effectiveness of the correction method in eliminating common-mode to differential-mode interference. For example, assuming a pure bioelectrical signal... , superimposed Common-mode interference amplitude The uncorrected time-differential signal contains a common-mode to differential-mode component. The ratio of trust to dryness is only After revision The common-mode interference components are completely eliminated.

[0075] The correction method has adaptive boundary characteristics. When the first contact impedance is equal to the second contact impedance and hour, The correction amount is zero, and the signal correction step degenerates into a pass-through operation, introducing no distortion to the original signal. This characteristic ensures that the correction algorithm will not have a negative impact under ideal conditions of good and symmetrical electrode contact. The correction effect is more pronounced when there is severe impedance asymmetry: for example, if , (Asymmetry 5:1) ,same Common-mode interference will generate The differential mode component has exceeded the amplitude of a typical bioelectrical signal ( (e.g., EMG signal), if not corrected, the signal will be completely overwhelmed by interference. After the correction steps described above, the interference component is eliminated. The correction accuracy depends on the impedance measurement accuracy: when the Z measurement error is 1%, The error is approximately 0.5%. The residual correction error of common-mode interference is approximately In contrast The bioelectrical signal is only 1.5%, which meets the signal quality requirements for clinical-grade electrophysiological acquisition.

[0076] In the signal correction step, the excitation, sampling, and solution are periodically re-executed to update the contact impedance value, and the differential output signal is continuously corrected based on the updated contact impedance value. The retest period... Usually in to Within the scope, this embodiment initially takes After each retest, record the newly measured impedance value. Compared with the baseline value of the last retest Compare and calculate the relative change: The adaptive update strategy is based on continuous The relative change after the second retest Adjust the retest interval for the next cycle. The number of consecutive determinations. Typically, the range is 3 to 10; in this embodiment, 5 times are used. When consecutive... Second measurement All are less than the stability threshold When the electrode contact is determined to be in a steady state, the retest interval is extended to twice the current interval, i.e. ,in As the upper limit of the longest retest interval, this embodiment takes... When any measurement Exceeding the change threshold If the impedance shows a significant drift, the retest interval will be shortened to half of the current interval. ,in As the lower limit of the shortest retest interval, this embodiment takes The stability threshold Typically, the range is between 0.5% and 2%, and in this embodiment, it is taken as 1%; the change threshold Typically, it ranges from 3% to 10%; in this embodiment, it is 5%. In and During this period, the current retest interval remains unchanged.

[0077] During each retest, a brief measurement window is inserted into the normal acquisition sequence. The duration of this measurement window is equal to the total time consumed by the excitation configuration step and the sampling step; in this embodiment, it is approximately [time value missing]. Immediately after the window ends, signal acquisition resumes and the common-mode to differential-mode coefficient is recalculated using the new impedance value. The updated coefficients take effect in the next sampling period. Data acquired during retesting is marked as invalid frames and discarded by subsequent signal processing modules. For example, when electrode contact is stable ( (5 consecutive times), retest interval from the initial... Gradually extended to , , (Capped) to reduce system power consumption and the impact of the measurement window on signal continuity. When the wearer begins to move, it causes a sudden change in impedance ( When, the interval is from Gradually shorten to , , Until (Minimum guarantee) ensures that the correction factor tracks impedance changes.

[0078] In another embodiment, the dynamic tracking of the contact impedance can also be achieved through a prediction method based on a first-order exponential decay model. The impedance change trend is predicted based on historical measurements to reduce the actual measurement frequency, which is suitable for long-term monitoring scenarios where impedance drift is highly regular.

[0079] Optionally, the triggering time for the periodic retest can also be determined based on impedance drift trend prediction. The technical motivation for this predictive triggering is that the impedance drift caused by electrode paste drying is a slow, approximately monotonically increasing process, and its short-term behavior can be well approximated by a first-order linear model. Compared to the exponential decay model, linear regression requires only two fitting parameters and has direct extrapolation prediction capabilities, making it suitable for the limited computing resources of embedded processors. Specifically, maintaining the equation solution steps is the most recent... Historical sliding window composed of measurement results The drift slope is obtained by performing a first-order linear regression fit on the data within the window. : in The unit is , The fitting intercept. The size of the sliding window. Typically, the window size is in the range of 5 to 20; in this embodiment, it is 10. The choice of window size balances fitting accuracy and response speed. When the value is 5, the response to trend changes is rapid but the fitting variance is large, making it suitable for motion scenarios; When the value is 20, the fitting variance is small, but the response delay to mutations is approximately [missing value]. This method is suitable for long-term resting monitoring. The predicted time of exceeding the limit is calculated based on the drift slope. in For drift tolerance, this embodiment uses 3% of the current impedance value. This is the baseline value after the last successful retest. When the predicted time exceeds the limit... If the next retest time is later than the currently planned time, the retest will be triggered in advance to ensure that the update is completed before the impedance deviation exceeds the correction accuracy tolerance.

[0080] The prediction mechanism includes failure detection and rollback strategies: when the determination coefficient of linear regression... When the value falls below the fit validity threshold, the linear model is deemed unsuitable for the current drift pattern. In this case, predictive triggering is abandoned, and the system reverts to the adaptive periodic retest strategy. The fit validity threshold is typically in the range of 0.6 to 0.8; in this embodiment, it is set to 0.7. Typical failure scenarios include brief electrode contact causing a non-monotonic impedance jump, and impedance fluctuating randomly around a certain value without a clear trend. After the revert, the system maintains the aforementioned... The driver adjusts the adaptive interval, and the prediction function is implemented in subsequent windows. It will automatically reactivate when the value recovers to above the effective threshold.

[0081] For example, suppose the current In 10 consecutive retests from linear growth to ( Window span ), fitted to obtain Drift tolerance Current deviation ,but The tolerance limit has been reached; a retest should be triggered immediately with a new one. Update baseline value If the next retest planned is in ,and (Still available) (margin), then the system in Perform retests in advance instead of waiting The period has expired.

[0082] The predictive triggering allows the system to proactively intervene before the signal correction accuracy substantially deteriorates during the impedance monotonic drift caused by electrode paste drying, offering better correction continuity compared to passive threshold comparison. The same drift prediction is performed on the second contact impedance Z2, and the earlier of the two exceeding the limit is taken as the actual triggering time. In another embodiment, an exponentially weighted moving average can be used instead of linear regression to calculate the drift trend. This method requires less computation but lacks extrapolation prediction capability, making it suitable for scenarios with lower prediction accuracy requirements.

[0083] Based on the same inventive concept, embodiments of this application also provide an electrophysiological impedance measurement and correction system. For example... Figure 8 As shown, the system includes an excitation configuration module, a potential restoration module, an impedance calculation module, and a signal correction module.

[0084] The excitation configuration module is used to switch the reference electrode drive circuit to output a fixed reference voltage and inject test current into the first and second input terminals of the instrumentation amplifier through a digital-to-analog converter and a current-limiting resistor, performing at least two excitations with different voltage configurations. The excitation configuration module includes a mode switching subunit and a voltage configuration subunit. After receiving the measurement trigger signal, the mode switching subunit first presets the output of each digital-to-analog converter to the midpoint voltage, then controls the analog switch to switch the reference electrode drive circuit from negative feedback mode to fixed voltage output mode, and simultaneously switches the instrumentation amplifier gain from high gain in signal acquisition mode to low gain in measurement mode. The voltage configuration subunit reads the target code values ​​of the digital-to-analog converters for each excitation configuration from the pre-calculated parameter storage area, writes them sequentially into the digital-to-analog converter register, and starts a hardware timer after each configuration to wait for the setup time to reach a steady state. After the timer expires, it sends a sampling ready signal to the potential restoration module. After all configurations are completed, the excitation configuration module performs a recovery operation: restoring the digital-to-analog converter output to the midpoint voltage, switching the reference electrode drive circuit back to negative feedback mode, and restoring the instrumentation amplifier gain to the signal acquisition mode value.

[0085] The potential restoration module is used to sample the differential output voltage and common-mode voltage of the instrumentation amplifier under each excitation configuration, and restore the absolute potentials of the first and second input terminals based on the differential output voltage and common-mode voltage. The potential restoration module includes a synchronous sampling subunit and a potential calculation subunit. After receiving the sampling ready signal from the excitation configuration module, the synchronous sampling subunit drives the main analog-to-digital converter and the auxiliary analog-to-digital converter to synchronously start sample-and-hold with the same hardware trigger signal. Under the same excitation configuration, it continuously performs a preset number of samplings and stores the original code values ​​in a circular buffer. After sampling is completed, it sends a data ready signal to the potential calculation subunit. The potential calculation subunit reads the multiple sampled code values ​​in the buffer, calculates the arithmetic mean of the differential channel and the common-mode channel respectively, restores the voltage to the true voltage using the analog-to-digital converter calibration coefficient, and then solves for the absolute potentials of each input terminal based on the differential output voltage, common-mode voltage, and instrumentation amplifier gain. The calculation results are written into a measurement data structure for the impedance calculation module to read.

[0086] The impedance solving module is used to determine the current flowing through each contact impedance based on the voltage configuration value and the restored absolute potential of each excitation configuration, construct and solve a system of linear equations to obtain the contact impedance value of each electrode. The impedance solving module includes a current calculation subunit, an equation construction subunit, and a rationality check subunit. The current calculation subunit reads the absolute potential of each configuration and the corresponding digital-to-analog converter voltage configuration value from the measurement data structure, combines the calibration values ​​of the current-limiting resistor and bias resistor, and calculates the total current flowing through each contact impedance under each configuration according to the accurate current model. The equation construction subunit receives the current calculation results and potential data, assembles the coefficient matrix and constant vector, calculates the determinant of the coefficient matrix, and analytically solves for each contact impedance value and the human body node potential using Cramer's rule. The rationality check subunit performs positive value checks, value range checks, and consistency checks on the solution results sequentially. Based on the check results, it sets the confidence flag to pass, warning, or failure, and writes the impedance value and confidence flag into the new value area of ​​the double-buffered structure, replacing the current valid value through an atomic exchange operation for the signal correction module to read.

[0087] The signal correction module determines the common-mode to differential-mode coefficient based on the contact impedance value and the voltage division relationship between the bias resistor at the input of the instrumentation amplifier and each contact impedance. It subtracts the product of the common-mode to differential-mode coefficient and the real-time common-mode voltage AC component from the differential output signal of the instrumentation amplifier, outputting the corrected electrophysiological signal. Upon receiving the updated effective impedance value from the impedance solving module, the signal correction module pre-calculates the common-mode to differential-mode coefficient and stores it in a register-level variable. During each sampling period in the signal acquisition mode, the signal correction module reads the real-time sampled values ​​from the main analog-to-digital converter and the auxiliary analog-to-digital converter, performs calibration and restoration, subtracts the common-mode interference to differential-mode component from the differential signal, and outputs the corrected electrophysiological signal to the subsequent digital signal processing chain. The signal correction module also maintains a retest timer, determines the trigger time for the next measurement based on the adaptive update strategy and drift prediction results, and sends a measurement trigger signal to the excitation configuration module upon expiration to initiate a new round of impedance measurement.

[0088] The excitation configuration module, potential restoration module, impedance calculation module, and signal correction module can be implemented by a processor executing a computer program stored in memory. The processor is a microcontroller or digital signal processor with floating-point arithmetic capabilities. The floating-point arithmetic capability requirement of the processor stems from the correction calculation... The calculation involves subtracting two fractions close to 1. Fixed-point arithmetic in this scenario is prone to losing significant bits, while 32-bit single-precision floating-point arithmetic naturally guarantees 23 bits of mantissa precision, avoiding this problem. The processor's clock frequency requirement is determined by the most intensive computational task: the correction operation performs a multiplication-subtraction operation once per sampling period. (Level), the impedance solution performs matrix operations (approximately) on each retest. (level), both are related Processors with the above clock frequencies can operate without issue. The memory includes non-volatile memory and random access memory: the non-volatile memory is used to store calibration parameters and pre-calculated code values, with a capacity requirement of approximately [missing information]. Random access memory is used to store the real-time sampling buffer, drift prediction sliding window, and impedance value double-buffered structure, with a capacity requirement of approximately [missing information]. The aforementioned resource requirements can be met by the currently mainstream... The microcontroller meets the requirements without the need for an external coprocessor or large-capacity storage expansion.

[0089] This application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the above-described electrophysiological impedance measurement and correction method.

[0090] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working process of the system and modules described above can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0091] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit described above can be implemented in hardware.

[0092] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions within the technical scope disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for measuring and correcting electrophysiological impedance, applied to an electrophysiological acquisition system including a first digital-to-analog converter (DAC), a second DAC, a current-limiting resistor, a reference electrode driving circuit, and an instrumentation amplifier, wherein the first DAC and the second DAC are respectively connected to a first input terminal and a second input terminal of the instrumentation amplifier via the current-limiting resistor, and the first input terminal and the second input terminal respectively form contact impedances with human skin, characterized in that, The method includes: The reference electrode drive circuit is switched to output a fixed reference voltage, and test current is injected into the first and second input terminals of the instrumentation amplifier through the first digital-to-analog converter and the second digital-to-analog converter respectively via the current-limiting resistor, performing at least two excitations with different voltage configurations. Under each excitation configuration, the differential output voltage and common-mode voltage of the instrumentation amplifier are sampled, and the absolute potentials of the first and second input terminals are reconstructed based on the differential output voltage and common-mode voltage. Based on the voltage configuration value and the restored absolute potential of each excitation configuration, the current flowing through each contact impedance is determined, a system of linear equations is constructed and solved to obtain the contact impedance value of each electrode.

2. The method according to claim 1, characterized in that, It also includes a calibration step: before performing the excitation, a known voltage is output through the digital-to-analog converter and applied to a precision reference load through the current-limiting resistor to obtain the accurate value of the current-limiting resistor and the compensation coefficient.

3. The method according to claim 2, characterized in that, The process of obtaining the accurate value of the current-limiting resistor includes: connecting a precision reference load to the first input terminal or the second input terminal, outputting multiple known voltages through the digital-to-analog converter, sampling the corresponding current response, and calculating the measured value of the current-limiting resistor based on Ohm's law.

4. The method according to claim 1, characterized in that, When determining the current flowing through each contact impedance, the shunting of the test current by the bias resistor at the input of the instrumentation amplifier is included in the current determination.

5. The method according to claim 4, characterized in that, Under each excitation configuration, the differential output voltage and common-mode voltage are sampled multiple times and averaged to reduce the noise of the absolute potential.

6. The method according to claim 1, characterized in that, It also includes signal correction steps: Based on the contact impedance value and the voltage division relationship between the bias resistor at the input of the instrumentation amplifier and each contact impedance, the common-mode to differential-mode coefficient that converts common-mode interference into differential-mode components through each contact impedance is determined. The product of the common-mode to differential-mode coefficient and the real-time common-mode voltage AC component is subtracted from the differential output signal of the instrumentation amplifier to obtain the corrected electrophysiological signal.

7. The method according to claim 6, characterized in that, In the signal correction step, the excitation, sampling, and solution are periodically re-executed to update the contact impedance value, and the differential output signal is continuously corrected based on the updated contact impedance value.

8. The method according to claim 1, characterized in that, The step of performing at least two excitations with different voltage configurations also includes adaptive configuration optimization based on the solution results of the previous round: when the condition number of the coefficient matrix of the linear equation system exceeds the optimization threshold, the output offset direction of the second digital-to-analog converter in the next round of excitation is automatically adjusted to maximize the absolute value of the determinant of the coefficient matrix.

9. The method according to claim 6, characterized in that, The step of constructing and solving a system of linear equations further includes an electrode contact quality assessment based on the residuals: substituting the solved contact impedance values ​​back into each configuration equation to calculate the residuals, determining the contact quality index based on the normalized root mean square value of the residuals, and performing hierarchical control on the signal correction step according to the contact quality index.

10. An electrophysiological impedance measurement and correction system, applied to electrophysiological acquisition scenarios where contact impedance is formed between the electrode and human skin, wherein the input terminal of the instrumentation amplifier in the system is connected to a midpoint voltage source via a bias resistor, characterized in that... The system includes: The excitation configuration module is used to switch the reference electrode drive circuit to output a fixed reference voltage and inject test current into the first and second input terminals of the instrumentation amplifier through a digital-to-analog converter and a current-limiting resistor to perform at least two excitations with different voltage configurations. The potential restoration module is used to sample the differential output voltage and common-mode voltage of the instrumentation amplifier under each excitation configuration, and restore the absolute potential of the first input terminal and the second input terminal based on the differential output voltage and common-mode voltage; The impedance solving module is used to determine the current flowing through each contact impedance based on the voltage configuration value of each excitation configuration and the restored absolute potential, construct a system of linear equations and solve them to obtain the contact impedance value of each electrode.

Images