A nursing monitoring system and method for neurology department

Abstract

The present invention discloses a nursing monitoring system and method for neurology department, specifically related to the field of abnormal pattern recognition of nerve signals, including synchronizing the time axes of MRS and EEG sampling data and establishing a mapping relationship with the stimulation input trigger event; performing feature dimensionality reduction on the set target tracking metabolite concentration data, extracting the energy distribution characteristics of EEG signals through the time-frequency distribution algorithm, and establishing the correlation characteristics between metabolism and nerve electrical activity; establishing the correlation driving path between the real-time metabolite concentration change of the patient and nerve electrical activity and quantifying its causal effect; establishing the abnormal pattern recognition reference for metabolism-nerve electrical activity signals, and judging whether the patient is in the pathological abnormal state by calculating the deviation index of pathological features; when it is determined that the patient is in the abnormal state of metabolism-nerve electrical activity signals, the abnormal pattern of the patient's metabolism-nerve electrical activity signals is identified and classified according to the direction calibration in the causal relationship network.

Description

Technical Field

[0001] The present invention relates to the technical field of abnormal pattern recognition of nerve signals. More specifically, the present invention relates to a nursing monitoring system and method for neurology. Background Art

[0002] In the field of neurology nursing, the dynamic association and causal relationship between brain metabolism and nerve electrical activity signals are important pathological features of many neurological diseases. For example, the sudden increase in lactic acid concentration during epileptic seizures is associated with abnormal δ waves, or the decrease in NAA after stroke is associated with abnormal synchronization of θ waves. The abnormal patterns of these signals can reflect the occurrence and development of diseases and provide important references for clinical intervention. However, traditional analysis methods mostly rely on static correlation analysis, ignoring the dynamic changes of metabolism and nerve electrical activity signals in the time dimension and the causal driving relationship between the two, making it difficult to accurately locate the lesion or identify the pathological driving pattern, resulting in a high missed detection rate and misjudgment rate in the diagnosis of pathological states and pattern classification.

[0003] Therefore, how to implement a comprehensive processing solution based on dynamic analysis of multi-modal signals, causal relationship modeling, and abnormal pattern classification, providing accurate abnormal detection and pattern classification capabilities, so as to support real-time monitoring and personalized intervention in neurology nursing, is an urgent problem to be solved.

[0004] To solve the above problems, a technical solution is provided now. Summary of the Invention

[0005] In order to overcome the above-mentioned defects of the prior art, an embodiment of the present invention provides a nursing monitoring method for neurology to solve the problems raised in the above background art.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] S1: Synchronize the time axes of MRS and EEG sampling data and establish a mapping relationship with the stimulus input trigger event;

[0008] S2: Extract the concentration value changes of the set target tracking metabolite, and at the same time extract the energy distribution characteristics of the EEG signal through the time-frequency distribution algorithm, calculate the correlation coefficient describing the characteristic association between metabolism and nerve electrical activity, and construct an association matrix;

[0009] S3: According to the mapping relationship between the MRS and EEG signals and the stimulus input trigger event, divide the correlation coefficient sequence in the association matrix into multi-segment time window data, establish a time series distribution sequence of the correlation coefficient, bring the correlation coefficient time series distribution sequence into the causal analysis model, establish a real-time metabolite concentration change and nerve electrical activity association driving path of the patient, and quantify its causal effect to construct a causal relationship network;

[0010] S4: Retrieve and analyze historical sample data of healthy individuals, establish a reference correlation matrix for abnormal pattern recognition of metabolic-neural electrical activity signals. By analyzing the deviation between the correlation matrix of real-time metabolite concentration changes and neural electrical signal activity characteristic data of the patient and the reference correlation matrix, calculate the deviation index of the patient's pathological characteristics, and determine whether the patient is in an abnormal state of metabolic-neural electrical activity signals based on the deviation index of the pathological characteristics;

[0011] S5: When it is determined that the patient is in an abnormal state of metabolic-neural electrical activity signals, identify and classify the abnormal patterns of the patient's metabolic-neural electrical activity signals according to the direction calibration in the causal relationship network.

[0012] In a preferred embodiment, in S1, the time-axis synchronization adjustment of MRS and EEG sampling data and the establishment of the mapping relationship with the stimulus input trigger event specifically include:

[0013] Set the sampling frequencies of the MRS and EEG devices respectively, inject calibration pulses into the MRS and EEG devices through the synchronization trigger signal, and synchronize and adjust the signal acquisition time axes of the two groups of devices;

[0014] Real-time monitor and capture the external stimulus input trigger event, bind the time node of the event to the sampling data, and form the mapping relationship between the multi-modal data and the trigger event on the unified time scale;

[0015] Perform timestamp embedding on the synchronized MRS and EEG signals, calibrate the time for each acquisition time point, and form a unified time series data format.

[0016] In a preferred embodiment, in S2, extract the concentration value changes of the set target tracking metabolite, and at the same time extract the energy distribution characteristics of the EEG signal through the time-frequency distribution algorithm, calculate the correlation coefficient describing the characteristic correlation between metabolism and neural electrical activity, and construct the correlation matrix specifically including:

[0017] Analyze the signal intensity of the MRS signal, extract the concentration value of the set target tracking metabolite, and construct a time series distribution matrix of metabolite concentration changes based on the time series calibration of the concentration change trend and peak characteristics in space-time;

[0018] Perform a fast Fourier transform on the EEG signal, extract the energy distribution characteristics of the set frequency band, remove the high-frequency artifacts and low-frequency drift signals, and construct a neural electrical signal characteristic matrix describing the specific frequency band;

[0019] Combine the concentration change time points of the metabolic signal and the brain region localization of the neural electrical signal, achieve time domain alignment through the dynamic time warping algorithm, and map the neural electrical signal characteristics to the brain region map according to the spatial distribution;

[0020] Perform a correlation analysis on the time series distribution matrix of metabolite concentration changes and the neural electrical signal feature matrix to generate an association matrix R. Each element in the association matrix represents the Pearson correlation coefficient between the dynamic concentration of the metabolite and the energy distribution of the neural electrical signal in a specific frequency band, and the correlation coefficient is used to describe the association strength.

[0021] Use dynamic range compression technology to normalize the values of the association matrix, and uniformly map the associated expression of metabolite concentration and neural electrical signal features to a standardized interval.

[0022] In a preferred embodiment, in S3, according to the mapping relationship between MRS and EEG signals and the stimulus input trigger event, divide the correlation coefficient sequence in the association matrix into multiple time window data, establish a time series distribution sequence of the correlation coefficients, bring the correlation coefficient time series distribution sequence into the causal analysis model, establish the association driving path between the real-time metabolite concentration change and neural electrical activity of the patient, and quantify its causal effect. The construction of the causal relationship network specifically includes:

[0023] Extract the correlation coefficient sequence corresponding to the metabolite concentration and neural electrical signal features from the association matrix R. According to the time start node that captures the stimulus input trigger event, divide the correlation coefficient sequence into multiple time window data based on the mapping relationship of MRS and EEG signals, and construct a time series distribution sequence of the correlation coefficients.

[0024] Use the causal model to analyze the time lag response between metabolite concentration and neural electrical signal activity, quantify the causal strength between metabolite concentration and neural electrical signal activity in each lag time period, and generate a causal sequence matrix. The formula for judging the causal strength is:

[0025]

[0026] In the formula, is the causal value, representing the causal effect quantification strength between variable and variables. is the correlation coefficient value in the i-th row and j-th column of the association matrix R, representing the association strength between the power of the j-th frequency band in the EEG signal corresponding to the i-th target metabolite type. is the concentration change value of the i-th target metabolite extracted from the MRS signal. is the historical sequence of the target variable . is the historical sequence of the target variable and the causal variable . represents the variance of the prediction error residuals of the causal model.

[0027] Summarize the causal strengths within multiple time windows according to the lag time to generate a dynamic causal matrix G. Each element in the matrix represents the cumulative causal strength of metabolite changes on the neural electrical signal activity in specific frequency bands within all segmented time windows. Based on the dynamic causal matrix G, construct a causal relationship network.

[0028] In a preferred embodiment, in S4, retrieve and analyze the historical healthy individual sample data, establish an abnormal pattern recognition reference correlation matrix for the metabolism-neural electrical activity signals. By analyzing the deviation between the correlation matrix of the real-time metabolite concentration changes and the neural electrical signal activity characteristic data of the patient and the reference correlation matrix, calculate the deviation index of the patient's pathological characteristics. Judging whether the patient is in the abnormal state of the metabolism-neural electrical activity signals based on the deviation index of the pathological characteristics specifically includes:

[0029] Retrieve the metabolism and neural electrical activity data of historical healthy individual samples, and generate a reference correlation matrix by calculating the correlation between metabolite concentration changes and neural electrical signal activity characteristics;

[0030] Perform element difference calculation on the correlation matrix of the real-time metabolite concentration changes and the neural electrical signal activity characteristic data of the patient and the reference correlation matrix to generate a correlation deviation matrix, and comprehensively calculate the pathological deviation index through the correlation deviation matrix. The calculation formula is as follows:

[0031]

[0032] In the formula, is the pathological deviation index, is the correlation coefficient of the i-th row and the j-th column of the reference correlation matrix, and m and n are the total number of set target metabolite types and the number of EEG signal frequency bands respectively;

[0033] Set the abnormal threshold of the pathological deviation index. When the pathological deviation index exceeds the abnormal threshold of the pathological deviation index, it is determined that the patient is in the abnormal state of the metabolism-neural electrical activity signals, and vice versa, it is determined that the patient is in the normal state of the metabolism-neural electrical activity signals.

[0034] In a preferred embodiment, in S5, when it is determined that the patient is in the abnormal state of the metabolism-neural electrical activity signals, identify and classify the abnormal patterns of the patient's metabolism-neural electrical activity signals according to the direction calibration in the causal relationship network. Specifically includes:

[0035] When it is determined that the patient is in the abnormal state of the metabolism-neural electrical activity signals, calculate the deviation rate of each element in the correlation deviation matrix, and screen the elements with a deviation rate greater than the set threshold and label them as dynamic correlation abnormal points;

[0036] Extract matrix elements from the same position in the causal matrix G according to the dynamically associated anomaly points, and generate causal paths for the abnormal driving pattern recognition of metabolic-neural electrical activity signals based on the direction calibration of the nodes in the corresponding causal relationship network;

[0037] If the dynamically associated anomaly point shows that the metabolic signal is the causal starting point in the causal relationship network, then calibrate the abnormal pattern of the patient's metabolic-neural electrical activity signal as the metabolic abnormality-driven pattern;

[0038] If the dynamically associated anomaly point shows that the neural electrical activity signal is the causal starting point in the causal relationship network, then calibrate the abnormal pattern of the patient's metabolic-neural electrical activity signal as the neural activity abnormality-driven pattern;

[0039] If the dynamically associated anomaly point shows a two-way interaction between the metabolic signal and the neural electrical activity signal in the causal relationship network, then calibrate the abnormal pattern of the patient's metabolic-neural electrical activity signal as a metabolic-neural complex abnormality.

[0040] On the other hand, the present invention provides a nursing monitoring system for neurology, including a signal synchronization module, a signal feature analysis module, a causal relationship analysis module, a pathological abnormality judgment module, and an abnormal pattern recognition module:

[0041] Signal synchronization module: Align the sampling time axes of the MRS and EEG devices through a synchronous trigger calibration mechanism, capture external stimulus trigger events in real time and bind time nodes, and generate a unified time series data format;

[0042] Signal feature analysis module: Extract the target metabolite concentration value by analyzing the MRS signal intensity, generate a concentration change time series matrix, extract the spectral features of the EEG signal, construct a spectral feature matrix after removing artifacts and drifts, and use correlation analysis on the two to generate a standardized multi-modal feature correlation expression.

[0043] Causal relationship analysis module: Use a causal model to analyze the lag response of the metabolic signal to the EEG signal, quantify the causal intensity in each time period, and generate a causal sequence matrix. Summarize the causal results of multiple time windows according to the lag time, and construct a dynamic causal matrix to build a causal relationship network;

[0044] Pathological abnormality judgment module: Retrieve the data of healthy individuals to generate a reference association matrix, calculate the difference with the association matrix collected from the patient in real time to generate an association deviation matrix, and comprehensively calculate the pathological deviation index; Set an abnormal threshold, and when the pathological deviation index exceeds the threshold, determine that the patient's metabolic-neural electrical activity signal is abnormal, otherwise determine it to be normal;

[0045] Abnormal pattern recognition module: Calculate the deviation rate of the correlation deviation matrix to screen dynamic correlation abnormal points, generate a causal path in combination with the direction of the causal relationship network, calibrate the metabolic or nerve signal as the causal starting point according to the path, and classify the abnormal pattern as metabolic drive, nerve drive or metabolic-nerve complex abnormality.

[0046] Technical effects and advantages of a nursing monitoring system and method for neurology in the present invention:

[0047] The present invention adopts a causal analysis model to establish an association drive path between the metabolite concentration of a patient and the nerve electrical activity, and quantifies its causal effect. At the same time, an abnormal pattern recognition reference for the metabolic-nerve electrical activity signal is constructed based on the historical data of healthy individuals, and the deviation index of the patient's pathological characteristics is calculated to determine whether the patient is in an abnormal state. When the signal is determined to be abnormal, the abnormal patterns of the patient's metabolic and nerve electrical activity signals are identified through the directional calibration of the causal relationship network and classified as metabolic drive, nerve activity drive or metabolic-nerve complex abnormality.

[0048] In summary, the present invention provides accurate signal analysis and pattern recognition capabilities for neurology nursing, supports personalized diagnosis and treatment decisions, realizes real-time monitoring and dynamic evaluation of complex neuropathological states, and provides a scientific basis for disease intervention. Brief Description of the Drawings

[0049] Figure 1 It is a schematic diagram of a nursing monitoring method for neurology in the present invention;

[0050] Figure 2 It is a structural schematic diagram of a nursing monitoring system for neurology in the present invention. Detailed Embodiments

[0051] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0052] Embodiment 1

[0053] Figure 1 A nursing monitoring method for neurology in the present invention is given, which includes the following steps:

[0054] S1: Synchronize the time axis of the MRS and EEG sampling data and establish a mapping relationship with the stimulus input trigger event;

[0055] S2: Extract the concentration value changes of the set target tracking metabolites, and at the same time extract the energy distribution characteristics of the EEG signals through the time-frequency distribution algorithm, calculate the correlation coefficient describing the characteristic correlation between metabolism and neural electrical activities, and construct a correlation matrix;

[0056] S3: According to the mapping relationship between the MRS and EEG signals and the stimulus input trigger events, divide the correlation coefficient sequence in the correlation matrix into multi-segment time window data, establish a time series distribution sequence of the correlation coefficients, bring the correlation coefficient time series distribution sequence into the causal analysis model, establish the association drive path between the real-time metabolite concentration changes and neural electrical activities of the patient, and quantify its causal effect to construct a causal relationship network;

[0057] S4: Retrieve and analyze the historical healthy individual sample data, establish an abnormal pattern recognition reference correlation matrix for the metabolism-neural electrical activity signals, calculate the deviation index of the patient's pathological characteristics by analyzing the deviation between the correlation matrix of the patient's real-time metabolite concentration changes and neural electrical signal activity characteristic data and the reference correlation matrix, and judge whether the patient is in the abnormal state of the metabolism-neural electrical activity signals based on the deviation index of the pathological characteristics;

[0058] S5: When it is determined that the patient is in the abnormal state of the metabolism-neural electrical activity signals, identify and classify the abnormal patterns of the patient's metabolism-neural electrical activity signals according to the direction calibration in the causal relationship network.

[0059] In S1, the time-axis synchronization adjustment of the MRS and EEG sampling data and the establishment of the mapping relationship with the stimulus input trigger events specifically include.

[0060] Select the target metabolite types (such as lactate, NAA, glutamate, etc.). In the MRS device, preset the chemical shift range corresponding to the metabolite (for example, lactate at 1.3 ppm), and adjust the spectral resolution to cover the accurate acquisition of the peak signals of the target metabolite. At the same time, in the EEG device, set the sampling frequency to 2000 Hz according to the target frequency band (such as δ wave 0.5 - 4 Hz or γ wave > 30 Hz) to ensure the integrity of the signal characteristics.

[0061] Start the MRS and EEG devices respectively, and set the trigger signal receiving mode. Inject calibration pulses into the MRS and EEG through the external stimulus control device. The calibration signals are triggered at a fixed frequency (such as 1 Hz), and each calibration pulse marks a unique timestamp on the time axis.

[0062] Start the external stimulation device and send trigger events (such as visual and auditory stimuli). Record the time nodes of the trigger events in real time and time-align them with the sampled data of MRS and EEG. Embed the synchronized MRS and EEG signals with unified timestamp information at each sampling point, save the multi-modal data file with embedded timestamps in CSV format, and attach the mapping table of the external stimulation time nodes and signal windows to form a complete time series data format.

[0063] In S2, perform feature dimensionality reduction on the set target-tracking metabolite concentration data, extract the energy distribution features of the EEG signals through time-frequency distribution algorithms, and establish the correlation features between metabolism and neural electrical activities.

[0064] Record the dynamic change signals of the target metabolite through the MRS device, use water suppression technology to remove the water peak interference at 4.7 ppm to ensure the clarity of the metabolite signals. Perform baseline correction on the collected MRS spectra and use smoothing algorithms to eliminate low-frequency drifts. For lactate and NAA, set their chemical shift ranges, extract the corresponding peak regions respectively, and integrate the peak signals to calculate the metabolite concentrations. Calibrate the obtained concentration values on the time axis at each time point to generate the metabolite concentration time series data matrix. 。

[0065] The EEG signals are synchronously collected through a 64-channel device. After collection, use band-pass filtering technology to remove high-frequency artifacts (such as electromyogram interference) and low-frequency drifts (such as electrode offsets). Perform fast Fourier transform (FFT) on the signals to convert the time-domain signals into frequency-domain features, calculate the energy of the corresponding bands (such as δ, γ, θ, etc.) in the set frequency bands (specifically set based on the monitored pathological bands) to form the spectral feature matrix of the EEG signals. 。

[0066] Combine the time points of the concentration changes of the metabolic signals and the brain region localization of the neural electrical signals, achieve time-domain alignment through the dynamic time warping algorithm, and map the neural electrical signal features to the brain region map according to the spatial distribution. Through spatial distribution and feature analysis, reveal the activity patterns, functional states, and possible abnormal behaviors of specific brain regions, providing a visual observation model for nursing monitoring.

[0067] Perform correlation analysis on the metabolite concentration change time series distribution matrix and the neural electrical signal feature matrix to generate the correlation matrix R. Each element in the correlation matrix represents the Pearson correlation coefficient between the dynamic concentration of the metabolite and the energy distribution of the neural electrical signals in a specific band. The correlation coefficient is used to describe the correlation strength and is updated in real time based on the time point calibration of the synchronous time axis. The calculation expression of the correlation coefficient is:

[0068]

[0069] In the formula, is the correlation strength of the power of the j-th frequency band in the EEG signal corresponding to the i-th target metabolite type. and are the means of the concentration of the i-th target metabolite and the power of the j-th frequency band of the EEG signal. and are respectively the concentration of the i-th metabolite at time point t in the time series in the matrix , and the power of the j-th frequency band at time point t in the time series in the matrix . T is the length of the time series.

[0070] The dynamic range compression technology is used to normalize the values of the correlation matrix, and the correlation expression of metabolite concentration and neuroelectrical signal characteristics is uniformly mapped to the standardized interval.

[0071] In S3, based on the causal analysis model, an association-driven path between the real-time metabolite concentration change of the patient and neuroelectrical activity is established, and its causal effect is quantified.

[0072] Extract the correlation coefficient sequence corresponding to metabolite concentration and neuroelectrical signal characteristics from the correlation matrix R. According to the time start node that captures the stimulus input trigger event, the correlation coefficient sequence is divided into multiple time window data according to the mapping relationship of the signal, and a correlation coefficient time series distribution sequence is constructed.

[0073] Use a causal model (such as the Granger causality analysis model) to analyze the time-lag response of metabolite concentration and neuroelectrical signal activity, quantify the causal strength of metabolite concentration and neuroelectrical signal activity in each lag time period (the lag period uses multiple time windows as the time axis reference benchmark), and generate a causal sequence matrix. The causal strength judgment formula is:

[0074]

[0075] In the formula, is the causal value, indicating the causal effect quantification strength of variable and variables. is the correlation coefficient value of the i-th row and j-th column of the correlation matrix R, indicating the correlation strength of the power of the j-th frequency band in the EEG signal corresponding to the i-th target metabolite type. is the concentration change value of the i-th target metabolite extracted from the MRS signal. is the historical sequence of the target variable . is the historical sequence of the target variable and the causal variable . represents the variance of the prediction error residuals of the causal model.

[0076] If , it shows that after introducing , the prediction error is significantly reduced, indicating a causal relationship. If , it shows that after introducing , the prediction is not significantly improved, indicating an insignificant causal relationship.

[0077] In the model with the causal variable added, the residuals are recalculated for each time point based on the historical sequence data. The specific process is as follows:

[0078]

[0079] In the formula, is the predicted target variable of the model at the a-th time point (the latest time point) in the time series, is the lag order of the target variable, that is, the number of lag time steps when predicting using its own historical sequence values (the step size is determined by the time axis scale of multiple time windows), is the lag order of the causal variable, that is, the causal variable the number of lag time steps when predicting using historical values, , , in the time series model, the linear influence coefficient of historical values on the current value, the error between the predicted value and the actual value of the target variable by the model at time t. Adjust the lag orders p and q to minimize the model error and calculate the variance of .

[0080] The expression form of the causal sequence matrix is:

[0081]

[0082] is the causal intensity distribution within the k-th time window, is the complete expression form at z time points within the time window.

[0083] The causal intensities within multiple time windows are aggregated according to the lag time to obtain an average matrix, which serves as the dynamic causal matrix G. Each element in the matrix represents the cumulative causal intensity of metabolite changes on the neural electrical signal activity in a specific frequency band within all segmented time windows. Based on the dynamic causal matrix G, a causal relationship network is constructed, and the weight of the edge is determined by the causal intensity of the dynamic causal matrix.

[0084] In S4, an abnormal pattern recognition reference for the metabolism-neural electrical activity signal is established based on historical healthy individual sample data. By calculating the deviation index of pathological features, it is determined whether the patient is in a pathologically abnormal state.

[0085] Retrieve the metabolic and neuroelectrical activity data of healthy individual samples, and generate a reference correlation matrix (consistent with the process in S2) by calculating the correlation between the metabolite concentration changes and the characteristics of neuroelectrical signals.

[0086] Calculate the element difference between the correlation matrix of the real-time metabolite concentration changes and the characteristics of neuroelectrical signal activities of the patient and the reference correlation matrix to generate a correlation deviation matrix, and comprehensively calculate the pathological deviation index through the correlation deviation matrix. The calculation formula is as follows:

[0087]

[0088] In the formula, is the pathological deviation index, is the correlation coefficient of the i-th row and j-th column of the reference correlation matrix, and m and n are the total number of set target metabolite types and the number of EEG signal frequency bands respectively.

[0089] Set the abnormal threshold of the pathological deviation index (comprehensively set by medical guardians based on the patient's pathological data level and monitoring sensitivity requirements). When the pathological deviation index exceeds the abnormal threshold of the pathological deviation index, it is determined that the patient is in an abnormal state of metabolic-neuroelectrical activity signals, otherwise it is determined that the patient is in a normal state of metabolic-neuroelectrical activity signals.

[0090] In S5, when it is determined that the patient is in an abnormal state of metabolic-neuroelectrical activity signals, identify and classify the abnormal patterns of the patient's metabolic-neuroelectrical activity signals according to the direction calibration in the causal relationship network.

[0091] When it is determined that the patient is in an abnormal state of metabolic-neuroelectrical activity signals, calculate the deviation rate (numerical deviation ratio) of each element in the correlation deviation matrix, and screen the elements with a deviation rate greater than the set threshold and mark them as dynamic correlation abnormal points. The dynamic correlation abnormal points include significantly abnormal metabolite-neuroelectrical activity signal feature pairs.

[0092] Extract matrix elements from the causal matrix G at the same position according to the dynamic correlation abnormal points, and generate a causal path for the recognition of the abnormal driving mode of the metabolic-neuroelectrical activity signals according to the direction calibration of the nodes in the corresponding causal relationship network;

[0093] If the dynamic correlation abnormal point shows that the metabolic signal is the causal starting point in the causal relationship network, then mark the abnormal pattern of the patient's metabolic-neuroelectrical activity signals as the metabolic abnormal driving mode. For example, if the lactic acid concentration of a patient increases abnormally (metabolic signal) resulting in a significant abnormality in the δ wave intensity, then mark this abnormal pattern as metabolic abnormal driving.

[0094] If the dynamic correlation anomaly point shows the neural electrical activity signal as the causal starting point in the causal relationship network, then calibrate the abnormal pattern of the patient's metabolism-neural electrical activity signal as the neural activity abnormal drive mode. For example, if the patient's gamma wave frequency is too high (abnormal neural activity) causing glutamate metabolism disorder, then calibrate this abnormal pattern as neural activity abnormal drive.

[0095] If the dynamic correlation anomaly point shows a two-way interaction between the metabolic signal and the neural electrical activity signal in the causal relationship network, then calibrate the abnormal pattern of the patient's metabolism-neural electrical activity signal as metabolism-neural complex abnormality. Through this method, the driving mechanism of the patient's metabolism and neural electrical activity abnormalities can be effectively analyzed, providing a basis for precise intervention.

[0096] Example 2

[0097] The difference between Example 2 and Example 1 of the present invention is that this example introduces a nursing monitoring method for neurology department.

[0098] Figure 2 The structural schematic diagram of a nursing monitoring system for neurology department of the present invention is given. A nursing monitoring system for neurology department includes a signal synchronization module, a signal feature analysis module, a causal relationship analysis module, a pathological abnormality judgment module, and an abnormal pattern recognition module:

[0099] Signal synchronization module: Through the synchronous trigger calibration mechanism, align the sampling time axes of the MRS and EEG devices, capture external stimulus trigger events in real time and bind time nodes to generate a unified time series data format;

[0100] Signal feature analysis module: Extract the target metabolite concentration value by analyzing the MRS signal intensity to generate a concentration change time series matrix, extract the spectral features of the EEG signal, and construct a spectral feature matrix after removing artifacts and drifts. Then, use correlation analysis on the two to generate a standardized multi-modal feature correlation expression.

[0101] Causal relationship analysis module: Use the causal model to analyze the lag response of the metabolic signal to the EEG signal, quantify the causal intensity in each time period, and generate a causal sequence matrix. Summarize the causal results of multiple time windows according to the lag time, and construct a dynamic causal matrix to build a causal relationship network;

[0102] Pathological abnormality judgment module: Retrieve the healthy individual data to generate a reference correlation matrix, calculate the difference with the correlation matrix collected from the patient in real time to generate a correlation deviation matrix, and comprehensively calculate the pathological deviation index; Set an abnormal threshold. When the pathological deviation index exceeds the threshold, it is determined that the patient's metabolism-neural electrical activity signal is abnormal, otherwise it is determined to be normal;

[0103] Abnormal pattern recognition module: Calculate the deviation rate of the correlation deviation matrix to screen dynamic correlation abnormal points, generate causal paths in combination with the direction of the causal relationship network, calibrate metabolic or neural signals as causal starting points according to the paths, and classify abnormal patterns as metabolic-driven, neural-driven, or metabolic-neural complex abnormalities.

[0104] All the above formulas are calculated by removing the dimension and taking their numerical values. The formulas are obtained by collecting a large amount of data for software simulation to get a formula closest to the actual situation. The preset parameters and threshold selection in the formulas are set by those skilled in the art according to the actual situation.

[0105] The above embodiments can be implemented in whole or in part by software, hardware, firmware, or any other combination. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the processes or functions described in the embodiments of the present application are generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center by wire (such as infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that the computer can access or a data storage device such as a server or data center that contains one or more collections of available media. The available media can be magnetic media (such as floppy disks, hard disks, magnetic tapes), optical media (such as DVDs), or semiconductor media. The semiconductor media can be a solid-state drive.

[0106] Those of ordinary skill in the art can realize that the modules and algorithm steps of each example described in combination with the embodiments disclosed herein can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution. Professional technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of this application.

[0107] Those skilled in the art can clearly understand that for the convenience and conciseness of description, the specific working processes of the systems, devices, and modules described above can refer to the corresponding processes in the foregoing method embodiments and will not be repeated here.

[0108] In several embodiments provided in the present application, it should be understood that the disclosed systems, devices, and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the modules is only a logical function division. In actual implementation, there may be other division methods. For example, multiple modules or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed couplings or direct couplings or communication connections to each other can be through some interfaces. The indirect couplings or communication connections of the devices or modules can be in electrical, mechanical, or other forms.

[0109] The modules described as separate components may or may not be physically separated. The components shown as modules may or may not be physical modules. They can be located in one place or distributed to multiple network modules. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0110] In addition, in each embodiment of the present application, the functional modules can be integrated in a processing module, or each module can exist physically alone, or two or more modules can be integrated in one module.

[0111] If the functions are implemented in the form of software function modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, in essence, or the part that contributes to the prior art, or a part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in each embodiment of the present application. The foregoing storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical discs that can store program codes.

[0112] As described above, the above are only the specific implementation manners of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art within the technical scope disclosed in the present application can easily think of changes or substitutions, which should all be covered within the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

[0113] Finally, the above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A nursing monitoring method for neurology department, characterized in that, The steps include: S1: Synchronize the time axis of MRS and EEG sampling data and establish a mapping relationship with the stimulus input trigger event; S2: Extract the concentration changes of the set target metabolites, and extract the energy distribution characteristics of the EEG signal through the time-frequency distribution algorithm. Calculate the correlation coefficient describing the characteristic correlation between metabolism and neural electrical activity, and construct the correlation matrix. Specifically, it includes: Analyze the signal intensity of the MRS signal, extract the concentration value of the set target metabolite, and calibrate the concentration change trend and the spatiotemporal distribution of the peak characteristics based on the time series to construct the time series distribution matrix of the metabolite concentration change; Perform fast Fourier transform on the EEG signal to extract the energy distribution characteristics of the set frequency band, remove high-frequency artifacts and low-frequency drift signals, and construct a neural electrical signal feature matrix describing the specific band; Combining the concentration change time points of metabolic signals with the brain region localization of neural electrical signals, the dynamic time warping algorithm is used to achieve time domain alignment, and the neural electrical signal characteristics are mapped to the brain region map according to their spatial distribution; The correlation matrix R is generated by performing correlation analysis on the metabolite concentration change time series distribution matrix and the neural electrical signal feature matrix. Each element in the correlation matrix represents the Pearson correlation coefficient between the dynamic concentration of the metabolite and the neural electrical signal energy distribution in a specific band. The correlation coefficient is used to describe the correlation strength. S3: Based on the mapping relationship between MRS and EEG signals and stimulus input triggering events, the correlation coefficient sequence in the correlation matrix is divided into multiple time window data segments, and a time series distribution sequence of the correlation coefficient is established. The time series distribution sequence of the correlation coefficient is introduced into the causal analysis model to establish the association driving path between the patient's real-time metabolite concentration changes and neural electrical activity, and quantify its causal effect. The construction of the causal relationship network specifically includes: Extract the correlation coefficient sequence corresponding to metabolite concentration and neural electrical signal characteristics from the correlation matrix R. According to the time starting node of the triggering event of the captured stimulus input, the correlation coefficient sequence is divided into multiple time window data based on the mapping relationship between MRS and EEG signals, and the time series distribution sequence of the correlation coefficient is constructed. The causal model was used to analyze the time-lagged response of metabolite concentrations and neural electrical signal activity, quantify the causal strength of metabolite concentrations and neural electrical signal activity in each lag time period, and generate a causal sequence matrix, where the causal strength judgment formula is: Wherein, is the causal value, representing the variable and the quantitative intensity of the causal effect between variables, is the correlation coefficient value of the i-th row and j-th column of the correlation matrix R, representing the correlation intensity between the power of the j-th frequency band in the EEG signal corresponding to the i-th target metabolite type, is the concentration change value of the i-th target metabolite extracted from the MRS signal, is the target variable 's historical sequence, is the target variable and the causal variable 's historical sequence, represents the variance of the prediction error residuals of the causal model; The causal strengths within multiple time windows are summarized by lag time to generate a dynamic causal matrix G. Each element in the matrix represents the cumulative causal strength of metabolite changes on neural electrical signal activity in a specific band within all segmented time windows. The causal relationship network is constructed based on the dynamic causal matrix G. S4: Retrieve and analyze historical healthy individual sample data to establish a reference correlation matrix for abnormal pattern recognition of metabolic-neuroelectric activity signals. Calculate the deviation index of the patient's pathological characteristics by analyzing the deviation between the correlation matrix of the patient's real-time metabolite concentration changes and the neural electrical signal activity characteristic data and the reference correlation matrix. Determine whether the patient is in an abnormal metabolic-neuroelectric activity signal state based on the deviation index of the pathological characteristics. Specifically, the following steps are included: Retrieve the metabolic and neuroelectrical activity data of healthy individual samples in the historical records. By calculating the correlation between the changes in metabolite concentrations and the characteristics of neuroelectrical signals, a reference correlation matrix is generated. Calculate the element differences between the correlation matrix of the real-time metabolite concentration changes and neuroelectrical signal activity characteristics data of the patient and the reference correlation matrix to generate a correlation deviation matrix. Calculate the pathological deviation index comprehensively through the correlation deviation matrix. The calculation formula is as follows: wherein is the pathological deviation index is the correlation coefficient of the i-th row and j-th column of the reference correlation matrix, and m and n are the total number of preset target metabolite types and the number of EEG signal frequency bands, respectively; Set the abnormal threshold of the pathological deviation index. When the pathological deviation index exceeds the abnormal threshold of the pathological deviation index, it is determined that the patient is in an abnormal state of metabolic-neuroelectrical activity signals; otherwise, it is determined that the patient is in a normal state of metabolic-neuroelectrical activity signals. S5: When it is determined that the patient is in an abnormal state of metabolic-neuroelectrical activity signals, according to the direction calibration in the causal relationship network, identify and classify the abnormal patterns of the patient's metabolic-neuroelectrical activity signals.

2. The nursing monitoring method for neurology department according to claim 1, characterized in that, In S1, synchronize the time axes of the MRS and EEG sampling data and establish a mapping relationship with the stimulus input trigger event, which specifically includes: Set the target metabolite types to be tracked, and at the same time set the sampling frequencies of the MRS and EEG devices respectively. Set the sampling frequencies of the MRS and EEG devices respectively. Inject calibration pulses into the MRS and EEG devices through the synchronization trigger signal to synchronize and adjust the signal acquisition time axes of the two groups of devices. Monitor and capture the external stimulus input trigger event in real time, bind the time nodes of the event to the sampling data, and form a mapping relationship between the multimodal data and the trigger event on the unified time scale. Embed time stamps into the synchronized MRS and EEG signals, calibrate the time for each acquisition time point, and form a unified time series data format.

3. The nursing monitoring method for neurology department according to claim 2, wherein In S2, it is also necessary to use dynamic range compression technology to normalize the values of the correlation matrix, and uniformly map the correlation expressions of metabolite concentrations and neuroelectrical signal characteristics to the standardized interval.

4. The nursing monitoring method for neurology department according to claim 3, wherein, In S5, when it is determined that the patient is in an abnormal state of metabolic-neuroelectrical activity signals, according to the direction calibration in the causal relationship network, identify and classify the abnormal patterns of the patient's metabolic-neuroelectrical activity signals, which specifically includes: When it is determined that the patient is in an abnormal state of metabolic-neuroelectrical activity signals, calculate the deviation rate of each element in the correlation deviation matrix, and screen the elements with a deviation rate greater than the set threshold and label them as dynamic correlation abnormal points. Extract matrix elements from the causal matrix G at the same positions according to the dynamic correlation abnormal points, and generate a causal path for the recognition of the abnormal driving mode of the metabolic-neuroelectrical activity signals according to the direction calibration of the nodes corresponding to these elements in the causal relationship network. If the dynamic correlation abnormal point shows that the metabolic signal is the causal starting point in the causal relationship network, label the abnormal pattern of the patient's metabolic-neuroelectrical activity signals as the metabolic abnormal driving mode. If the dynamic correlation abnormal point shows that the neuroelectrical activity signal is the causal starting point in the causal relationship network, label the abnormal pattern of the patient's metabolic-neuroelectrical activity signals as the neuroactivity abnormal driving mode. If the dynamic correlation abnormal point shows a two-way interaction between the metabolic signal and the neuroelectrical activity signal in the causal relationship network, label the abnormal pattern of the patient's metabolic-neuroelectrical activity signals as metabolic-neuro complex abnormality.

5. A nursing monitoring system for neurology department, which is used to implement a nursing monitoring method for neurology department described in any one of claims 1-4, characterized in that, It includes a signal synchronization module, a signal feature analysis module, a causal relationship analysis module, a pathological abnormality judgment module, and an abnormal pattern recognition module: Signal synchronization module: By means of a synchronous trigger calibration mechanism, align the sampling time axes of the MRS and EEG devices, capture external stimulus trigger events in real time and bind time nodes to generate a unified time series data format; Signal feature analysis module: Extract the target metabolite concentration value by analyzing the MRS signal intensity, generate a concentration change time series matrix, extract the spectral features of the EEG signal, construct a spectral feature matrix after removing artifacts and drifts, and use correlation analysis on the two to generate a standardized multi-modal feature correlation expression; Causal relationship analysis module: Use a causal model to analyze the lag response of metabolic signals to EEG signals, quantify the causal intensity in each time period, and generate a causal sequence matrix; Summarize the causal results of multiple time windows according to the lag time, construct a dynamic causal matrix to build a causal relationship network; Pathological abnormality judgment module: Retrieve healthy individual data to generate a reference association matrix, calculate the difference with the association matrix collected in real time from the patient to generate an association deviation matrix, and comprehensively calculate the pathological deviation index; Set an abnormal threshold. When the pathological deviation index exceeds the threshold, it is determined that the patient's metabolic-neural electrical activity signal is abnormal, otherwise it is determined to be normal; Abnormal pattern recognition module: Calculate the deviation rate of the association deviation matrix to screen dynamic association abnormal points, generate a causal path in combination with the direction of the causal relationship network, calibrate the metabolic or neural signal as the causal starting point according to the path, and classify the abnormal pattern as metabolic-driven, neural-driven, or metabolic-neural complex abnormality.

Images