Dynamic monitoring and early warning system for intracranial pressure after neurosurgical procedures

Abstract

This application discloses a dynamic monitoring and early warning system for intracranial pressure after neurosurgical procedures, relating to the technical field of medical surveillance. The system includes a composite intracranial probe, a postural monitoring unit, a multimodal signal acquisition card, and a computational processing unit. The probe integrates a pressure sensor and a thermistor, and the computational processing unit includes modules for signal preprocessing, attenuation parameter identification, frequency response reconstruction, compliance inversion, and fused early warning. The system removes waveform distortion due to physical obstruction by inverting the mechanical attenuation coefficient of the medium surrounding the probe using a thermomechanical coupling principle and deconvolving and reconstructing the original pressure signal using an inverse system transfer function. It also switches between steady-state respiratory analysis and transient postural disturbance modes depending on the head movement state and calculates an intracranial compliance index. The present invention enables high-fidelity waveform reconstruction and continuous compliance evaluation in multiple scenarios, and can effectively provide early warning of the risk of intracranial compensatory depletion.

Description

Technical Field

[0001] The present invention relates to the technical field of medical monitoring, and more specifically, to the dynamic monitoring and early warning system of intracranial pressure after craniocerebral surgery.

Background Art

[0002] In intensive care for craniocerebral surgery, invasive intracranial pressure monitoring is the gold standard for preventing brain herniation and ensuring cerebral perfusion pressure. However, implanted sensors are likely to be compressed by surrounding brain tissues, wrapped by thrombus, or pressed against the skull inside the skull. As a result, the mechanical attenuation of the pressure transmission path increases, the dynamic response ability of the sensor decreases, and an effect similar to low-pass filtering occurs. This causes distortions such as attenuation of amplitude, smoothing of characteristic peaks, and loss of high-frequency details in the monitored pulse wave waveform. Since existing devices do not have a self-check function for the operating environment of the probe, it is difficult for doctors to distinguish whether the flattening of the waveform is due to physiological attenuation of cerebral pulsation or physical distortion, which affects the judgment of the disease condition.

[0003] Furthermore, conventional monitoring mainly focuses on the average value of pressure. However, based on the relationship between intracranial pressure and volume, intracranial pressure may appear normal despite an increase in the occupancy effect during the compensatory period. The risk of compensatory space depletion cannot be identified only by the absolute value. There are methods for evaluating intracranial compliance based on waveform morphology and slow wave analysis, but these methods depend on the accuracy of the waveform. When waveform distortion occurs due to probe attenuation, compliance evaluation leads to incorrect conclusions and cannot provide an effective early warning during the compensatory period before the onset of increased intracranial pressure.

Summary of the Invention

Problems to be Solved by the Invention

[0004] In contrast to the shortcomings of conventional technology, the present invention provides a dynamic monitoring and early warning system for intracranial pressure after neurosurgical procedures, solving the problem of increased mechanical attenuation and distortion of pressure waveforms due to implantable probes being compressed by brain tissue or becoming entangled with blood clots, which prevents accurate reflection of the characteristics of intracranial pulses. [Means for solving the problem]

[0005] To achieve the above objectives, the present invention is realized by the following technical solutions: A dynamic monitoring and early warning system for intracranial pressure after neurosurgical procedures, A composite intracranial probe integrating a pressure sensor, thermistor, and posture monitoring unit, A multimodal signal acquisition card for collecting original pressure signals, temperature response signals, respiratory signals, and postural signals, A signal preprocessing module that aligns and resamples signals from each channel on the time axis, and calculates head angular velocity based on the body position signal. A damping parameter identification module that controls a thermistor to apply thermal pulse excitation, collects temperature decay data, calculates the thermal relaxation time constant by creating a temperature decay model, and converts the thermal relaxation time constant into a mechanical decay coefficient based on the thermomechanical coupling relationship. A frequency response reconstruction module constructs an inverse system transfer function model using the aforementioned mechanical damping coefficient, performs deconvolution correction on the original pressure signal, and outputs a reconstructed intracranial pressure signal. A compliance inversion module that switches between steady-state analysis mode and transient disturbance mode according to head angular velocity and calculates phase hysteresis features or pressure response ratio features, and The system includes a computing processing unit configured to implement a fused early warning module that verifies the sensor state based on a mechanical damping coefficient and combines the reconstructed intracranial pressure signal with a normalized compliance index to generate stepwise monitoring results.

[0006] The thermistor of the composite intracranial probe is located at the end of the probe and adjacent to the pressure sensor diaphragm, and the attenuation parameter identification module is By fitting temperature response data, a thermal relaxation time constant representing the thermal diffusion rate of the medium can be obtained, and A pre-calibrated log-linear mapping function is used to map the thermal relaxation time constant to a mechanical damping coefficient that represents how much the vibration system of the pressure sensor is resisted by the medium.

[0007] The frequency response reconstruction module is To construct a forward transfer function by making the pressure sensor and its media environment equivalent to a quadratic linear time-invariant system. To construct an inverse filter model which is the reciprocal of the forward transfer function, and Numerical differentiation is used to transform the inverse filter model into a discrete difference equation, and then a time-domain deconvolution operation is performed on the original pressure signal to compensate for amplitude-frequency attenuation and phase-frequency delay.

[0008] The compliance inversion module described above is This system is used to monitor head angular velocity in real time, switching to steady-state analysis mode when the absolute value of angular velocity remains below the static threshold, and to transient disturbance mode when the absolute value of angular velocity exceeds the static threshold. In steady-state analysis mode, the compliance inversion module is: To construct a respiratory-intracranial pressure transfer function that takes a respiratory signal as input and outputs a reconstructed intracranial pressure signal, and By analyzing the cross-power spectral density and self-power spectral density, the phase angle at the respiratory-dominant frequency is extracted and used to calculate phase hysteresis features.

[0009] In transient disturbance mode, the compliance inversion module, Based on the principles of fluid hydrostatics, the theoretical change in hydrostatic pressure is calculated using the change in head tilt angle before and after a change in body position, and This is used to obtain pressure response ratio characteristics by calculating the ratio between the observed pressure change of the reconstructed intracranial pressure signal and the theoretical hydrostatic pressure change.

[0010] The aforementioned fusion early warning module is Comparing the mechanical damping coefficient with the physical occlusion threshold, If the mechanical damping coefficient is higher than the physical occlusion threshold, the sensor will be determined to be in a physical failure state, and the pathological evaluation will be stopped. If the mechanical damping coefficient is below the physical occlusion threshold, it is used to execute subsequent stepwise early warning logic.

[0011] The aforementioned fusion early warning module is A weighted normalization algorithm is used to map phase hysteresis features or pressure response ratio features in various modes to a unified intracranial compliance index.

[0012] The aforementioned fusion early warning module is To construct a two-dimensional evaluation model, If the average value of the reconstructed intracranial pressure signal is lower than the high-pressure threshold and the intracranial compliance index is higher than the warning threshold, it is determined that the region is within the physiological steady-state range. If the mean value of the reconstructed intracranial pressure signal is lower than the high-pressure threshold, but the intracranial compliance index is lower than the warning threshold, it should be determined that this is a latent area of ​​compensatory depletion, and If the average value of the reconstructed intracranial pressure signal is higher than the high-pressure threshold and the intracranial compliance index is lower than the warning threshold, it is used to determine that the area is in a pathological decompensation zone.

[0013] The aforementioned signal preprocessing module further, By performing bandpass filtering on the original pressure signal, baseline drift and commercial frequency interference are removed, and By performing time-frequency analysis on the respiration signal, it is used to extract the respiration dominant frequency in real time and serve for restricting the calculation frequency band in the steady analysis mode.

Advantages of the Invention

[0014] The present invention provides a dynamic monitoring and early warning system for intracranial pressure after neurosurgery. It has the following beneficial effects.

[0015] 1. The present invention uses a thermistor to identify the mechanical attenuation coefficient of the medium around the probe in real time and non-invasively, and uses the inverse system deconvolution algorithm to effectively compensate the distortion of the pressure waveform caused by the physical occlusion of the probe, restore the high-frequency details of the intracranial pressure pulse wave, and provide a reliable data basis for the accurate evaluation of subsequent intracranial compliance.

[0016] 2. The present invention intelligently switches between two physiological disturbance modes using the head angular velocity signal. When the patient is in a quiet state, the spontaneous respiration wave is used as a micro disturbance source to continuously evaluate the compliance, while when the patient's position changes, the hydrostatic change is used as an active disturbance to provide a more robust compliance index. These two modes complement each other, greatly improving the continuity and reliability of compliance monitoring.

[0017] 3. The present invention constructs a two-dimensional evaluation model of pressure compliance that focuses not only on the absolute value of intracranial pressure but also on its compensatory reserve ability. By fusing and analyzing the mechanical attenuation coefficient and reconstructing the intracranial pressure value and the normalized compliance index, it can effectively identify the compensatory ability depletion latent region where the compliance is depleted although the intracranial pressure value is normal, ensure a valuable time margin for clinical intervention, and reduce the risk of harmful events such as sudden intracranial hypertension.

Brief Description of the Drawings

[0018] [Figure 1] This is an architectural diagram of the system of the present invention. [Figure 2] This is a flowchart of data processing of the signal preprocessing module of the present invention.

Embodiments for Carrying out the Invention

[0019] Hereinafter, referring to the drawings in the embodiments of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described. Obviously, the embodiments described here are only a part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative labor shall also be included in the protection scope of the present invention. Embodiment

[0020] Referring to FIGS. 1 to 2, an embodiment of the present invention provides a dynamic monitoring and early warning system for intracranial pressure after neurosurgery, including a composite intracranial probe 10, a body position and posture monitoring unit 20, a multimodal signal acquisition card 30, and a central processing unit 40.

[0021] The composite intracranial probe 10 is configured as an elongated tubular or linear structure suitable for implantation into the human cranial cavity. A pressure sensor and a thermistor are integrated at its detection end. The pressure sensor is used to detect the hydrostatic pressure and fluctuating pressure of the environment inside the skull, and the thermistor has a joule heating function and a temperature detection function. It applies a controlled thermal pulse excitation to the surrounding medium and is used to detect the subsequent temperature decay process. The pressure sensor and the thermistor are spatially closely connected so that both are in contact with the same local physical media environment.

[0022] The posture monitoring unit 20 is an inertial measurement assembly fixed to a bandage or drainage tube fixation device on the patient's head, and is used to measure the tilt angle of the patient's head relative to the gravity vector and the angular velocity information of head movement in real time.

[0023] The multimodal signal acquisition card 30 is electrically connected to both the composite intracranial probe 10 and the posture monitoring unit 20. The multimodal signal acquisition card 30 provides the necessary excitation power to the sensors and is used to synchronously acquire the original pressure analog signal output by the pressure sensor, the temperature analog signal output by the thermistor, and the posture digital signal output by the posture monitoring unit 20, and to convert the above signals into a time-matched digital signal sequence.

[0024] The central processing unit 40 is communicatively connected to the multimodal signal acquisition card 30 and is used to receive digital signal sequences and execute data processing algorithms. The central processing unit 40 has memory and a processor, and its internal operating logic is divided into a signal preprocessing module 41, an attenuation parameter identification module 42, a frequency response reconstruction module 43, a compliance inversion module 44, and a fusion early warning module 45.

[0025] The signal preprocessing module 41 is configured to perform synchronization and feature cleaning processes for physiological and physical multi-source signals, and specifically includes the following processing steps:

[0026] The system performs time-axis alignment and resampling of multi-channel heterogeneous signals. Each channel data collected by the multimodal signal acquisition card 30 has mismatched sampling frequencies. Pressure sensors are typically set to a high sampling rate to satisfy the reconstruction of pulse wave waveforms, while the thermistor and the body position monitoring unit 20 have relatively low sampling rates. The signal preprocessing module 41 constructs a unified discrete-time sequence using the timestamp sequence of the original pressure signal as the reference axis. For probe temperature signals and body position signals with sampling frequencies lower than the reference axis, the system performs upsampling using a linear interpolation algorithm or a polynomial interpolation algorithm to match the data point density of the original pressure signal. In addition, time-shift compensation is performed on each channel data sequence based on the pre-calibrated hardware transmission delay parameters of each sensor to eliminate time phase shifts caused by differences in hardware transmission paths, ensuring that data corresponding to the same time index represents the system state at the same physical time.

[0027] Frequency-selective filtering is performed on the original pressure and respiration signals. The signal preprocessing module 41 performs digital bandpass filtering on the original pressure signal to separate the effective frequency band components for subsequent analysis. The lower limit of the passband frequency range of this digital bandpass filter is set to a cutoff frequency that removes baseline drift and electrode polarization noise, and the upper limit is set to a cutoff frequency that preserves the harmonic features of the heart pulse wave and removes commercial frequency interference and high-frequency electromagnetic noise. After this filtering process, the denoised pressure signal sequence and the respiration signal sequence are output. As for the specific implementation structure of the digital bandpass filter, those skilled in the art can use conventional means such as finite impulse response filters and infinite impulse response filters, and since these are known in the art, the specific transfer function equations will not be described in detail here.

[0028] Head posture calculation based on inertial measurement components: The signal preprocessing module 41 receives the 3-axis acceleration components output from the body posture monitoring unit 20 and calculates the pitch angle of the head relative to the horizontal plane using the projection distribution relationship of the gravitational acceleration vectors on each axis of the sensor coordinate system. The specific calculation logic is as follows.

[0029] The inverse trigonometric function of the ratio of the component of the sensor's sensitivity axis in the direction of gravity to the component perpendicular to the direction of gravity is calculated to obtain a head tilt angle sequence. To remove instantaneous high-frequency acceleration interference caused by non-postural movements such as coughing or tremors, the signal preprocessing module 41 applies low-pass filtering to the calculated head tilt angle sequence and outputs a smoothed attitude angle signal. Next, the signal preprocessing module 41 performs discrete difference calculations on the smoothed attitude angle signal to calculate the angle change per unit time and generate an angular velocity signal of head movement. This angular velocity signal is later used as a reference variable to trigger transient disturbance modes.

[0030] Real-time extraction of respiratory dominance frequency features: The signal preprocessing module 41 performs time-frequency domain analysis on the filtered respiratory signal, captures segments of the respiratory signal using a sliding time window, performs spectral analysis or time-domain peak detection on these signal segments, identifies the reciprocal of the time interval of the frequency point or periodic peak corresponding to the maximum power spectral density, and determines it as the respiratory dominance frequency. This respiratory dominance frequency parameter is transmitted in real time to subsequent modules and used to limit the computational frequency bandwidth of the frequency-domain transfer function analysis.

[0031] The damping parameter identification module 42 is used to reverse the mechanical damping characteristics of the medium surrounding the sensor probe based on its thermodynamic response characteristics. Specifically, it performs the following processing steps:

[0032] Microjoule-level pulse excitation is applied to collect temperature response data. The decay parameter identification module 42 sends a drive signal to a thermistor integrated into the end of the composite intracranial probe 10, controlling it to maintain a constant micropower heating state within a 1-millisecond time window, thereby causing a minute temperature rise on the probe surface. At the moment the heating time window ends, the decay parameter identification module 42 immediately switches the thermistor to temperature measurement operation mode and continuously records a decay data sequence of the probe surface temperature, which naturally decreases over time, at a high sampling frequency.

[0033] Construction of a temperature decay model representing the thermal diffusion characteristics of the medium: The decay parameter identification module 42 constructs a thermal decay model that describes the change in the probe temperature over time after heating is stopped, based on the transient heat conduction physical mechanism. The mathematical formula for this model is as follows:

[0034]

number

number

number

number

number

number

number

[0035] Calculation of thermal relaxation time constant: The damping parameter identification module 42 uses a numerical fitting algorithm to iteratively calculate the thermal relaxation time constant in the current measurement cycle using the collected temperature damping data sequence and the above temperature damping model.

number

[0036] Calculation of mechanical damping coefficient based on thermomechanical coupling relationship: The damping parameter identification module 42 utilizes the physical correlation between the thermal diffusion performance and mechanical viscoelasticity of the medium and converts the calculated thermal relaxation time constant into a mechanical damping coefficient that represents how much the pressure sensor vibration system is resisted by the medium, via a preset thermomechanical coupling mapping function. The expression for this thermomechanical coupling mapping function is as follows:

[0037]

number

number

number

number

number

number

[0038] The frequency response reconstruction module 43 uses the mechanical attenuation coefficient output by the attenuation parameter identification module 42 to perform dynamic compensation based on the inverse system principle on the pre-processed original pressure signal, thereby removing the low-pass filtering effect that occurs in the pressure waveform due to the sealing of the medium. The specific implementation process of this frequency response reconstruction module 43 includes the following steps.

[0039] Creation of a second-order dynamic response model of the sensor system: The frequency response reconstruction module 43 treats the sensor-medium coupled system as a second-order linear time-invariant system. At the physical level, when the probe is contained in a viscoelastic medium, the response characteristics to intracranial pressure fluctuations are determined by both the system's natural frequency and its current decay state. The frequency response reconstruction module 43 constructs a transfer function model of this coupled system in the Laplace domain, and the expression of this model is as follows:

[0040]

number

number

number

number

number

number

number

[0041] Construction of an inverse filter transfer function for signal restoration: To restore the true pressure signal from the distorted original pressure signal, the frequency response reconstruction module 43 constructs an inverse system model based on the forward model described above, and this inverse system model is mathematically the reciprocal of the forward transfer function, i.e., immediately

number

[0042]

number

number

[0043] Execution of deconvolution reconstruction operation in the discrete-time domain: Since the signal processing is performed in the discrete-time domain, the frequency response reconstruction module 43 uses the finite difference method to convert the continuous-domain inverse transfer function described above into a discrete-time difference equation. This difference equation uses the current and past original pressure sampling points to calculate the reconstructed pressure value at the present time. The specific reconstruction calculation formula is as follows:

[0044]

number

number

number

number

number

number

number

[0045] The compliance inversion module 44 dynamically switches between steady-state respiratory analysis mode and transient positional disturbance mode depending on the patient's head movement state, and inverts the intracranial compliance features using the reconstructed intracranial pressure signal output by the frequency response reconstruction module 43, specifically performing the following processing steps.

[0046] Determination of mode switching logic based on head movement state: The compliance inversion module 44 continuously reads the head angular velocity signal output by the signal preprocessing module 41 and compares the absolute value of this angular velocity signal with a preset static determination threshold. If the absolute value of the angular velocity signal remains below the static determination threshold within a preset time window, the compliance inversion module 44 determines that the system is in steady-state analysis mode and activates a calculation logic based on the respiratory transfer function. If the absolute value of the angular velocity signal exceeds the static determination threshold, the compliance inversion module 44 determines that the system is transitioning to transient disturbance mode and activates a calculation logic based on the hydrostatic step response.

[0047] Calculation of Phase Hysteresis Features of Respiratory Intracranial Pressure Transfer Function in Steady-State Analysis Mode: The compliance inversion module 44 uses the synchronously collected respiratory signals as the system's input excitation variable and the reconstructed intracranial pressure signals output by the frequency response reconstruction module 43 as the system's output response variable. The compliance inversion module 44 uses a digital signal processing algorithm to calculate the cross-power spectral density and self-power spectral density between the above input excitation variable and output response variable, and based on this, constructs a frequency-domain transfer function to describe the conduction characteristics of respiratory waves to intracranial pressure. The calculation process for the specific numerical values ​​of the cross-power spectral density and self-power spectral density can be implemented using the periodogram method or the Welch method, which are known techniques in the art and will not be described in detail here. Based on the constructed transfer function, the compliance inversion module 44 extracts phase angle information at the respiratory dominant frequency and calculates a phase lag parameter representing intracranial compensatory capacity. The calculation formula is as follows:

[0048]

number

number

number

number

number

number

[0049] Calculation of theoretical pressure input based on hydrostatics in transient disturbance mode: When it is determined that the body is transitioning to transient disturbance mode, the compliance inversion module 44 fixes the initial steady tilt angle at the start time of the positional change motion and the final steady tilt angle at the end time of the motion. Based on a hydrostatic physical model, the compliance inversion module 44 calculates the theoretical change in hydrostatic pressure due to the position of the positional change sensor probe, and uses this as the step input reference value for the evaluated intracranial compliance. The calculation formula is as follows.

[0050]

number

number

number

number

number

number

number

[0051] Calculation of pressure response ratio features in transient disturbance modes: The compliance inversion module 44 extracts the mean pressure difference between two steady-state phases before and after a change in body position from the reconstructed intracranial pressure signal and defines this as the observed pressure change. Next, the compliance inversion module 44 calculates the ratio of this observed pressure change to the calculated theoretical hydrostatic pressure change and generates pressure response ratio features. The calculation formula is as follows:

[0052]

number

number

number

[0053] The fused early warning module 45 is configured to construct a multidimensional state evaluation matrix by combining the mechanical attenuation coefficient output by the attenuation parameter identification module 42, the reconstructed intracranial pressure signal values ​​output by the frequency response reconstruction module 43, and the compliance feature parameters output by the compliance inversion module 44, and to generate stepwise monitoring results based on this matrix. Specifically, the following processing steps are performed.

[0054] Calculation of the normalized intracranial compliance index: Since the compliance inversion module 44 outputs feature parameters of various dimensions depending on the head movement state, the fused early warning module 45 uses a weighted normalization algorithm to map the phase lag parameter or pressure response ratio feature to a unified intracranial compliance index, thereby achieving continuity of the evaluation index over time. The formula for this weighted normalization algorithm is as follows:

[0055]

number

number

number

number

number

number

number

number

number

[0056] Checking the reliability of the sensor's physical state based on the mechanical damping coefficient: The fusion early warning module 45 compares the mechanical damping coefficient, which is input in real time, with a preset physical occlusion threshold to determine the physical validity of the current pressure data. If the mechanical damping coefficient is higher than the physical occlusion threshold, it indicates that the sensor probe surface is occluded due to adhesion of a high-viscosity medium or to a wall surface. In this case, the pressure sensor's response to pressure is significantly distorted. Based on this determination, the fusion early warning module 45 generates a physical failure status indicator for the sensor, blocks subsequent clinicopathological evaluation logic, and marks the reliability of the current data as invalid or low in the monitoring results, thereby eliminating misjudgments caused by false low-pressure waveforms due to the sensor adhering to a wall surface.

[0057] To construct a two-dimensional quadrant evaluation model of pressure compliance and generate stepwise early warnings: If the mechanical attenuation coefficient is below the physical occlusion threshold, the fused early warning module 45 calculates the mean of the reconstructed intracranial pressure signal, combines it with the calculated intracranial compliance index, and generates a stepwise evaluation result of the clinical condition based on a pre-set decision boundary logic. The specific judgment logic is as follows:

[0058] If the reconstructed average value of the intracranial pressure signal is lower than a preset high-pressure threshold, and the intracranial compliance index is higher than a preset compliance warning threshold, it is determined to be in the physiological steady-state region, and a normal monitoring signal is output. If the average value of the reconstructed intracranial pressure signal is lower than a preset high-pressure threshold, but the intracranial compliance index is lower than a preset compliance warning threshold, it is determined to be a latent compensatory capacity depletion zone. In this case, although the intracranial pressure measurement is still within the normal range, the volume compensatory reserve of the craniospinal cavity is approaching its limit, and the system outputs a warning signal indicating that the compensatory reserve has been depleted.

[0059] If the average value of the reconstructed intracranial pressure signal is higher than a preset high-pressure threshold, and the intracranial compliance index is lower than a preset compliance warning threshold, it is determined to be a pathological decompensation zone. In this case, intracranial hypertension has formed and cannot be alleviated by physiological regulation, so the system outputs an alarm signal for the adverse event of intracranial hypertension. [Explanation of Symbols]

[0060] 10. Compound intracranial probe 20 Posture Monitoring Unit 30 Multimodal Signal Acquisition Cards 40 Central Processing Units 41 Signal Preprocessing Module 42 Attenuation Parameter Identification Module 43 Frequency Response Reconstruction Module 44 Compliance Inversion Module 45. Fusion Early Warning Module

Claims

1. A dynamic monitoring and early warning system for intracranial pressure after neurosurgical procedures, A composite intracranial probe integrating a pressure sensor, thermistor, and posture monitoring unit, A multimodal signal acquisition card for collecting original pressure signals, temperature response signals, respiratory signals, and postural signals, A signal preprocessing module that aligns and resamples signals from each channel on the time axis, and calculates head angular velocity based on the body position signal. A damping parameter identification module that controls a thermistor to apply thermal pulse excitation, collects temperature decay data, calculates the thermal relaxation time constant by creating a temperature decay model, and converts the thermal relaxation time constant into a mechanical decay coefficient based on the thermomechanical coupling relationship. A frequency response reconstruction module constructs an inverse system transfer function model using the aforementioned mechanical damping coefficient, performs deconvolution correction on the original pressure signal, and outputs a reconstructed intracranial pressure signal. A compliance inversion module that switches between steady-state analysis mode and transient disturbance mode according to head angular velocity and calculates phase hysteresis features or pressure response ratio features, and A dynamic monitoring and early warning system for intracranial pressure after neurosurgical procedures, comprising: a computing processing unit configured to realize a fused early warning module that verifies the sensor state based on a mechanical attenuation coefficient and combines the reconstructed intracranial pressure signal with a normalized compliance index to generate stepwise monitoring results.

2. The thermistor of the composite intracranial probe is located at the end of the probe and adjacent to the diaphragm of the pressure sensor, and the attenuation parameter identification module is By fitting temperature response data, a thermal relaxation time constant representing the thermal diffusion rate of the medium can be obtained, and The system according to claim 1, characterized in that a pre-calibrated log-linear mapping function is used to map the thermal relaxation time constant as a mechanical damping coefficient representing how much the vibration system of the pressure sensor is resisted by the medium.

3. The frequency response reconstruction module is To construct a forward transfer function by making the pressure sensor and its media environment equivalent to a quadratic linear time-invariant system. To construct an inverse filter model which is the reciprocal of the forward transfer function, and The system according to claim 1, characterized in that it is used to compensate for amplitude-frequency attenuation and phase-frequency delay by converting an inverse filter model into a discrete difference equation using a numerical differentiation method and performing a time-domain deconvolution operation on the original pressure signal.

4. The compliance inversion module described above is The system according to claim 1, characterized in that it is used to monitor head angular velocity in real time, to switch to a steady-state analysis mode when the absolute value of the angular velocity remains below a static threshold, and to switch to a transient disturbance mode when the absolute value of the angular velocity exceeds a static threshold.

5. In steady-state analysis mode, the compliance inversion module is: To construct a respiratory-intracranial pressure transfer function that takes a respiratory signal as input and outputs a reconstructed intracranial pressure signal, and The system according to claim 4, characterized in that it is used to extract the phase angle at the respiratory dominant frequency by analyzing the cross power spectral density and the self power spectral density, and to calculate the phase hysteresis features.

6. In transient disturbance mode, the compliance inversion module, Based on the principles of fluid hydrostatics, the theoretical change in hydrostatic pressure is calculated using the change in head tilt angle before and after a change in body position, and The system according to claim 4, characterized in that it is used to obtain pressure response ratio characteristics by calculating the ratio between the observed pressure change of the reconstructed intracranial pressure signal and the theoretical hydrostatic pressure change.

7. The aforementioned fusion early warning module is Comparing the mechanical damping coefficient with the physical occlusion threshold, If the mechanical damping coefficient is higher than the physical occlusion threshold, the sensor will be determined to be in a physical failure state, and the pathological evaluation will be stopped. The system according to claim 1, characterized in that it is used to execute a subsequent stepwise early warning logic when the mechanical damping coefficient is below a physical occlusion threshold.

8. The aforementioned fusion early warning module is The system according to claim 1, characterized in that it is used to map phase hysteresis features or pressure response ratio features in various modes to a unified intracranial compliance index using a weighted normalization algorithm.

9. The aforementioned fusion early warning module is To construct a two-dimensional evaluation model, If the average value of the reconstructed intracranial pressure signal is lower than the high-pressure threshold and the intracranial compliance index is higher than the warning threshold, it is determined that the region is within the physiological steady-state range. If the mean value of the reconstructed intracranial pressure signal is lower than the high-pressure threshold, but the intracranial compliance index is lower than the warning threshold, it should be determined that this is a latent area of ​​compensatory depletion, and The system according to claim 8, characterized in that it is used to determine a pathological decompensation area when the average value of the reconstructed intracranial pressure signal is higher than the high-pressure threshold and the intracranial compliance index is lower than the warning threshold.

10. The aforementioned signal preprocessing module further, By performing bandpass filtering on the original pressure signal, baseline drift and commercial frequency interference are removed, and The system according to claim 1, characterized in that it is used to extract respiratory dominant frequencies in real time by performing time-frequency analysis on respiratory signals, and to limit the computational frequency band in steady-state analysis mode.

Images