Intelligent early warning bracelet for epilepsy monitoring and computing device

Abstract

The invention discloses an intelligent early warning bracelet for epilepsy monitoring and a computing device, and the bracelet comprises a first processing module which is used for analyzing physiological signal data of a patient, and generating preliminary result data of epileptic seizure; the second processing module is used for identifying a potential epileptic seizure mode in the first data structure and generating a second data structure; the prediction analysis module is used for combining the prediction probability, the epileptic seizure risk score and the event triggering time into a third data structure; the risk grading module is used for carrying out risk grading on the epileptic seizure risk score to generate a fourth data structure; the early warning information generation module is used for generating a fifth data structure according to the size and the resolution ratio of the bracelet display screen; and the bracelet display module is used for sending the fifth data structure to a nurse wearing bracelet for vibration and flickering prompt. Warning of different degrees is given according to different risk levels, nurses can take measures in time, and epilepsy danger of patients is avoided.

Description

Technical Field

[0001] The present invention relates to a field, and in particular to an intelligent early warning bracelet and a computing device for epilepsy monitoring. Background Art

[0002] Traditional epileptic seizure monitoring mainly relies on clinical observation and electroencephalogram (EEG) recordings, but there are still some problems and defects in practical applications, especially in the application on the nurse side. Due to the high reliance on manual interpretation of EEG, it is not only time-consuming and labor-intensive, but also difficult to achieve real-time early warning of epileptic seizures. Even if a remote monitoring system is introduced, it is often limited by problems such as data transmission delays and insufficient data processing capabilities, resulting in the failure of early warning information to be delivered to nurses in time, missing the opportunity for intervention. In addition, existing systems usually only send simple alarm signals and lack a detailed assessment of the risk of epileptic seizures. After receiving the early warning, it is difficult for nurses to quickly judge the patient's condition and take effective measures, which not only increases the risk of patients, but also increases the workload of nurses.

[0003] To solve the above problems, the present invention proposes an intelligent early warning bracelet for epilepsy monitoring, which gives different degrees of warnings according to different risk levels, so that nurses can take timely measures to avoid the risk of epilepsy in patients. Summary of the invention

[0004] In view of the above problems, the present invention provides a smart early warning bracelet and a computing device for epilepsy monitoring.

[0005] According to one aspect of the present invention, there is provided a smart early warning bracelet for epilepsy monitoring, comprising:

[0006] a first processing module, configured to analyze the physiological signal data of the patient collected by the first server to generate preliminary result data of epileptic seizures, wherein the physiological signal data includes electroencephalogram, heart rate variability and activity level; the preliminary result data is stored in the form of a first data structure, wherein the first data structure includes a timestamp, a physiological data feature vector and a preliminary risk assessment value;

[0007] a second processing module, configured to send the first data structure to a second server, identify potential epileptic seizure patterns in the first data structure according to a hidden Markov model, generate an intermediate state sequence, and store the intermediate state sequence as a second data structure, wherein the second data structure includes a power spectral density, a waveform feature, and a time series pattern;

[0008] A prediction analysis module, configured to perform prediction analysis according to the second data structure to obtain a predicted probability of an epileptic seizure in a patient; generate an epileptic seizure risk score according to the predicted probability and clinical experience rules; and combine the predicted probability, the epileptic seizure risk score, and the event trigger time into a third data structure; wherein the third data structure includes an epilepsy risk score level and an epilepsy warning status identifier, and the epilepsy warning status identifier includes an icon code, a color code, and curve data;

[0009] A risk level classification module, used to classify the epileptic seizure risk score into risk levels to generate a fourth data structure; wherein the fourth data structure includes a timestamp, an epileptic seizure risk score level, an early warning level, an event triggering time, and a color code corresponding to the risk level;

[0010] The warning information generation module is used to receive the fourth data structure and generate a fifth data structure according to the size, resolution and color display capability of the wristband display screen; wherein the fifth data structure includes a timestamp, a risk level icon, a vibration mode code and a dynamic warning information curve;

[0011] The bracelet display module is used to send the fifth data structure to the nurse wearing the bracelet, and the bracelet vibrates and flashes according to the fifth data structure.

[0012] In an optional manner, in the first processing module, the EEG signal is decomposed into a plurality of intrinsic mode functions according to an adaptive filtering method of variational mode decomposition;

[0013] For each intrinsic mode function, the cutoff frequency of the filter is adaptively adjusted according to its signal-to-noise ratio.

[0014] In an optional manner, in the second processing module, the state space of the hidden Markov model includes a wake state, a pre-epileptic seizure state, an epileptic seizure state, and a post-epileptic seizure state; wherein the calculation formula for the transition probability between each state is:

[0015]

[0016] in, is the static state transition probability obtained by pre-training; O t is the physiological data feature vector at time t in the first data structure; μ i is the average value of the physiological feature vector of the i-th state; d(O t ,μ i ) is the distance between the observation vector and the average eigenvector of the i-th state; γ and δ are adjustable parameters; Δ(t) is the time interval function, Δ(t) = exp(-k*(tt prev )), tprev is the time of the last state transition, and k is the time decay coefficient.

[0017] In an optional manner, in the first processing module, generating preliminary result data of epileptic seizure further includes:

[0018] Segmenting the physiological signal data, wherein each window contains data samples of a preset time length and a preset overlap rate is set between windows;

[0019] The instantaneous frequency and amplitude of each intrinsic mode function in the EEG signal were calculated according to the empirical mode decomposition algorithm to obtain preliminary result data.

[0020] In an optional manner, in the risk level classification module, the epileptic seizure risk score is classified according to an adaptive threshold adjustment method;

[0021] Among them, the epileptic seizure risk score is standardized by the cumulative distribution function, and the risk score is converted into a value between 0 and 1; the adaptive threshold of each risk level is calculated based on the patient's historical risk score data.

[0022] In an optional manner, after receiving the fifth data structure, the bracelet display module performs synchronization according to the timestamp in the fifth data structure to ensure that the time on the bracelet is consistent with that on the server;

[0023] The display screen of the wristband displays icons according to the risk level icon code in the fifth data structure, and the background color of the display screen of the wristband is switched to a color corresponding to the risk level according to the color code in the fifth data structure.

[0024] In an optional manner, in the first processing module, analyzing the physiological signal data further includes:

[0025] Decomposing the EEG signal into a plurality of sub-band signals of different frequencies by a multi-resolution analysis method of wavelet transform;

[0026] Each subband signal is analyzed by Hilbert-Huang transform, the instantaneous frequency, instantaneous amplitude and instantaneous phase of each subband signal are calculated, and its entropy value is extracted as a characteristic parameter;

[0027] The SD1, SD2 parameters, approximate entropy and sample entropy features of the Poincare map of the heart rate variability are calculated.

[0028] In an optional manner, calculating the instantaneous frequency and amplitude of each intrinsic mode function in the electroencephalogram signal according to the empirical mode decomposition algorithm to obtain preliminary result data further includes:

[0029] Take the original signal as the residual signal and add it to the list. When the number of extreme points of the residual signal is greater than 2, find all local maxima and local minima of the current residual signal.

[0030] Fitting the local maximum and the local minimum respectively by cubic spline interpolation to obtain an upper envelope and a lower envelope, and calculating the average values ​​of the upper envelope and the lower envelope;

[0031] Subtracting an average value of the upper envelope and the lower envelope from the current residual signal to obtain a candidate residual signal;

[0032] When the number of extreme value points of the candidate residual signal is less than or equal to 2, the remaining residual signals are added to the list to obtain the preliminary result data.

[0033] In an optional manner, analyzing each subband signal by Hilbert-Huang transform, calculating the instantaneous frequency, instantaneous amplitude and instantaneous phase of each subband signal, and extracting the entropy value as the characteristic parameter further includes:

[0034] Decomposing the original signal into a plurality of sub-band signals, wherein the sub-band signals are obtained by filter banks, wavelet transform or short-time Fourier transform;

[0035] Perform empirical mode decomposition on each sub-band signal to obtain the intrinsic mode function;

[0036] Perform Hilbert transform on each intrinsic mode function to obtain analytical signal, and calculate the instantaneous amplitude, instantaneous phase and instantaneous frequency;

[0037] The Shannon entropy of instantaneous amplitude, instantaneous frequency and instantaneous phase are calculated as characteristic parameters.

[0038] According to another aspect of the present invention, there is provided a computing device, comprising: a processor, a memory, a communication interface and a communication bus, wherein the processor, the memory and the communication interface communicate with each other via the communication bus;

[0039] The memory is used to store at least one executable instruction, and the executable instruction enables the processor to execute the operation corresponding to the above-mentioned smart early warning bracelet for epilepsy monitoring.

[0040] According to the solution provided by the present invention, it includes a first processing module, which is used to analyze the patient's physiological signal data collected by the first server to generate preliminary result data of epileptic seizures, wherein the physiological signal data includes electroencephalogram, heart rate variability and activity level; the preliminary result data is stored in the form of a first data structure, and the first data structure includes a timestamp, a physiological data feature vector and a preliminary risk assessment value; a second processing module, which is used to send the first data structure to a second server, identify potential epileptic seizure patterns in the first data structure according to a hidden Markov model, generate an intermediate state sequence, and store the intermediate state sequence as a second data structure, and the second data structure includes power spectral density, waveform characteristics and time series patterns; a prediction analysis module, which is used to perform prediction analysis based on the second data structure to obtain a predicted probability of epileptic seizures in patients; generate an epileptic seizure risk score based on the predicted probability and clinical experience rules; and store the predicted The probability and the epileptic seizure risk score and the event trigger time are combined into a third data structure; wherein the third data structure includes an epilepsy risk score level and an epilepsy warning state identifier, and the epilepsy warning state identifier includes an icon code, a color code and curve data; a risk level division module is used to divide the epileptic seizure risk score into risk levels to generate a fourth data structure; wherein the fourth data structure includes a timestamp, an epilepsy risk score level, a warning level, an event trigger time and a color code corresponding to the risk level; an early warning information generation module is used to receive the fourth data structure and generate a fifth data structure according to the size, resolution and color display capability of the bracelet display screen; wherein the fifth data structure includes a timestamp, a risk level icon, a vibration mode code and a dynamic early warning information curve; a bracelet display module is used to send the fifth data structure to a nurse wearing a bracelet, and the bracelet vibrates and flashes according to the fifth data structure. The present invention further improves the accuracy and timeliness of early warning by identifying potential epileptic seizure patterns through a hidden Markov model. Predictive analysis is performed based on the intermediate state sequence to generate the predicted probability and risk score of epileptic seizures, and a detailed risk score level is generated in combination with clinical experience rules. After receiving the early warning, the nurse can quickly judge the patient's specific situation and take effective intervention measures. At the same time, it reduces the nurse's manual interpretation of EEG data, reduces the workload, and improves work efficiency.

[0041] The above description is only an overview of the technical solution of the present invention. In order to more clearly understand the technical means of the present invention, it can be implemented according to the contents of the specification. In order to make the above and other purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Various other advantages and benefits will become apparent to those of ordinary skill in the art by reading the detailed description of the preferred embodiments below. The accompanying drawings are only for the purpose of illustrating the preferred embodiments and are not to be considered as limiting the present invention. Moreover, the same reference symbols are used throughout the accompanying drawings to represent the same components. In the accompanying drawings:

[0043] Figure 1 A schematic diagram of a framework of a smart early warning bracelet for epilepsy monitoring according to an embodiment of the present invention is shown;

[0044] Figure 2a to Figure 2b A schematic diagram showing changes in physiological signal data according to an embodiment of the present invention is shown;

[0045] Figure 3 A schematic diagram of the structure of a computing device according to an embodiment of the present invention is shown. DETAILED DESCRIPTION

[0046] The exemplary embodiments of the present invention will be described in more detail below with reference to the accompanying drawings. Although the exemplary embodiments of the present invention are shown in the accompanying drawings, it should be understood that the present invention can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided in order to enable a more thorough understanding of the present invention and to enable the scope of the present invention to be fully communicated to those skilled in the art.

[0047] Figure 1 FIG. 1 is a schematic diagram showing a framework of a smart early warning bracelet for epilepsy monitoring according to an embodiment of the present invention. Specifically, Figure 1 As shown, including:

[0048] a first processing module, configured to analyze the physiological signal data of the patient collected by the first server to generate preliminary result data of epileptic seizures, wherein the physiological signal data includes electroencephalogram, heart rate variability and activity level; the preliminary result data is stored in the form of a first data structure, wherein the first data structure includes a timestamp, a physiological data feature vector and a preliminary risk assessment value;

[0049] a second processing module, configured to send the first data structure to a second server, identify potential epileptic seizure patterns in the first data structure according to a hidden Markov model, generate an intermediate state sequence, and store the intermediate state sequence as a second data structure, wherein the second data structure includes a power spectral density, a waveform feature, and a time series pattern;

[0050] A prediction analysis module, configured to perform prediction analysis according to the second data structure to obtain a predicted probability of an epileptic seizure in a patient; generate an epileptic seizure risk score according to the predicted probability and clinical experience rules; and combine the predicted probability, the epileptic seizure risk score, and the event trigger time into a third data structure; wherein the third data structure includes an epilepsy risk score level and an epilepsy warning status identifier, and the epilepsy warning status identifier includes an icon code, a color code, and curve data;

[0051] A risk level classification module, used to classify the epileptic seizure risk score into risk levels to generate a fourth data structure; wherein the fourth data structure includes a timestamp, an epileptic seizure risk score level, an early warning level, an event triggering time, and a color code corresponding to the risk level;

[0052] The warning information generation module is used to receive the fourth data structure and generate a fifth data structure according to the size, resolution and color display capability of the wristband display screen; wherein the fifth data structure includes a timestamp, a risk level icon, a vibration mode code and a dynamic warning information curve;

[0053] The bracelet display module is used to send the fifth data structure to the nurse wearing the bracelet, and the bracelet vibrates and flashes according to the fifth data structure.

[0054] In this embodiment, the first processing module generates preliminary result data of epileptic seizures by analyzing the physiological signal data of electroencephalogram, heart rate variability and activity level, and realizes real-time monitoring of the patient's physiological state. The potential epileptic seizure pattern is identified by the hidden Markov model, which further improves the accuracy and timeliness of the early warning. The prediction and analysis module performs predictive analysis based on the intermediate state sequence, generates the predicted probability and risk score of epileptic seizures, and generates a detailed risk score level in combination with clinical experience rules. After receiving the early warning, the nurse can quickly judge the specific situation of the patient and take effective intervention measures. The early warning information generation module generates personalized early warning information according to the size, resolution and color display capability of the bracelet display. Through vibration and flashing prompts, nurses can more intuitively understand the patient's risk level and early warning status, so as to take corresponding intervention measures in time, while reducing the manual interpretation of the EEG data by nurses, reducing the workload and improving work efficiency.

[0055] For example, the patient's electroencephalogram, heart rate variability and activity level data are collected and analyzed by the first processing module, and a preliminary risk assessment value of 70 is generated, which is stored as a first data structure. The second processing module sends the first data structure to the second server, uses the hidden Markov model to identify potential epileptic seizure patterns, generates an intermediate state sequence, and stores it as a second data structure. The prediction analysis module performs a prediction analysis based on the second data structure, obtains a predicted probability of epileptic seizure of 85%, and generates a risk score of high risk based on clinical experience rules. The prediction probability and risk score are combined into a third data structure (including a high epilepsy risk score level, a red icon and curve data for the warning state). The risk level classification module divides the high risk score into a high-level warning and generates a fourth data structure. The warning information generation module generates a fifth data structure based on the characteristics of the bracelet display screen and sends it to the bracelet worn by the nurse. The bracelet vibrates and flashes according to the fifth data structure, displays a high-risk icon and a dynamic warning information curve, so that the nurse can understand the patient's risk level and warning state in a timely manner and take corresponding intervention measures.

[0056] In an optional manner, in the first processing module, the EEG signal is decomposed into a plurality of intrinsic mode functions according to an adaptive filtering method of variational mode decomposition;

[0057] For each intrinsic mode function, the cutoff frequency of the filter is adaptively adjusted according to its signal-to-noise ratio.

[0058] In this embodiment, the cutoff frequency of the filter is adaptively adjusted according to the signal-to-noise ratio of each intrinsic mode function (IMF), which can retain useful EEG signal information and solve the problem that the traditional fixed cutoff frequency filtering method cannot effectively handle noise interference in different frequency bands. The variational mode decomposition method adaptively decomposes the EEG signal into multiple intrinsic mode functions, which can more effectively separate EEG activities of different frequencies compared to the traditional empirical mode decomposition method.

[0059] Specifically, the collected EEG signal is input into the variational mode decomposition (VMD) algorithm, and the EEG signal is adaptively decomposed into multiple intrinsic mode functions (IMFs), each of which represents the oscillation mode of the signal in different frequency bands. For each IMF, its signal-to-noise ratio (SNR) is calculated by comparing the energy of the IMF with the noise energy (for example, based on the energy estimation of the known noise frequency range). According to the signal-to-noise ratio of each IMF, the cutoff frequency of the adaptive filter is determined. Among them, the IMF with a high signal-to-noise ratio sets a lower cutoff frequency, and the IMF with a low signal-to-noise ratio sets a higher cutoff frequency. Digital filters such as Butterworth filters and Chebyshev filters can be used to adaptively adjust their cutoff frequencies according to the signal-to-noise ratio, and each IMF is filtered to obtain a filtered IMF. The filtered IMFs are recombined to obtain a denoised EEG signal.

[0060] In an optional manner, in the second processing module, the state space of the hidden Markov model includes a wake state, a pre-epileptic seizure state, an epileptic seizure state, and a post-epileptic seizure state; wherein the calculation formula for the transition probability between each state is:

[0061]

[0062] in, is the static state transition probability obtained by pre-training; O t is the physiological data feature vector at time t in the first data structure; μ i is the average value of the physiological feature vector of the i-th state; d(O t ,μ i ) is the distance between the observation vector and the average eigenvector of the i-th state; γ and δ are adjustable parameters; Δ(t) is the time interval function, Δ(t) = exp(-k*(tt prev )), t prev is the time of the last state transition, and k is the time decay coefficient.

[0063] In this embodiment, through the time interval function and adjustable parameters, the hidden Markov model dynamically adjusts the state transition probability according to real-time data to better adapt to the actual physiological state changes of the patient. The state space of the hidden Markov model includes the awake state, the pre-epileptic state, the epileptic state and the post-epileptic state, which captures each stage of the epileptic attack in more detail.

[0064] In an optional manner, in the first processing module, generating preliminary result data of epileptic seizure further includes:

[0065] Segmenting the physiological signal data, wherein each window contains data samples of a preset time length and a preset overlap rate is set between windows;

[0066] The instantaneous frequency and amplitude of each intrinsic mode function in the EEG signal were calculated according to the empirical mode decomposition algorithm to obtain preliminary result data.

[0067] In this embodiment, Empirical Mode Decomposition (EMD) is an adaptive signal decomposition method that decomposes complex nonlinear and non-stationary signals into a series of intrinsic mode functions (IMF), each of which represents the oscillation mode of the signal on different time scales. By analyzing the instantaneous frequency and amplitude of the IMF, rich time-frequency information of the EEG signal is obtained, which is very beneficial for detecting events with time-dependent and frequency-characteristic changes in epileptic seizures. Compared with traditional time-frequency analysis methods (such as short-time Fourier transform and wavelet transform), EMD does not require pre-setting of basis functions and decomposes according to the characteristics of the signal itself. Therefore, it is more suitable for processing complex physiological signals such as EEG signals with large individual differences. For real-time or near real-time epileptic seizure monitoring systems, computational efficiency is crucial. The EMD algorithm can achieve a faster decomposition speed to meet the real-time requirements of epilepsy prediction.

[0068] In an optional manner, in the risk level classification module, the epileptic seizure risk score is classified according to an adaptive threshold adjustment method;

[0069] Among them, the epileptic seizure risk score is standardized by the cumulative distribution function, and the risk score is converted into a value between 0 and 1; the adaptive threshold of each risk level is calculated based on the patient's historical risk score data.

[0070] In this embodiment, compared with the fixed threshold, the adaptive threshold is more robust to noise and abnormal data, and even if some abnormal risk scores appear in the short term, it will not have a significant impact on the risk level classification. The traditional fixed threshold method uses the same risk level classification standard for all patients, ignoring the individual differences between patients.

[0071] For example, the historical risk score data of 5 patients are shown in Table 1:

[0072] Table 1

[0073] Patient ID Historical risk score data 1 [0.2,0.3,0.4,0.5,0.6] 2 [0.1,0.2,0.3,0.4,0.5] 3 [0.4,0.5,0.6,0.7,0.8] 4 [0.3,0.4,0.5,0.6,0.7] 5 [0.5,0.6,0.7,0.8,0.9]

[0074] Assume that the current risk score of patient 1 is 0.45. Based on the historical data of patient 1, the probability that the risk score is less than or equal to 0.45 is calculated to be 60%. Therefore, the standardized value is 0.6. Assume that the threshold for defining the low risk level is the 30% quantile and the threshold for the medium risk level is the 70% quantile. Since 0.35<=0.6<0.55 does not hold, and 0.6>=0.55 holds, the risk level is high risk.

[0075] In an optional manner, after receiving the fifth data structure, the bracelet display module performs synchronization according to the timestamp in the fifth data structure to ensure that the time on the bracelet is consistent with that on the server;

[0076] The display screen of the wristband displays icons according to the risk level icon code in the fifth data structure, and the background color of the display screen of the wristband is switched to a color corresponding to the risk level according to the color code in the fifth data structure.

[0077] In this embodiment, by synchronizing with the server-side timestamp, the wristband time is consistent with the server time, ensuring the timeliness and accuracy of the risk level information display, and avoiding information misjudgment caused by time asynchrony. The risk level is displayed in a combination of icons and colors, so that you can quickly understand your current risk status.

[0078] In an optional manner, in the first processing module, analyzing the physiological signal data further includes:

[0079] Decomposing the EEG signal into a plurality of sub-band signals of different frequencies by a multi-resolution analysis method of wavelet transform;

[0080] Each subband signal is analyzed by Hilbert-Huang transform, the instantaneous frequency, instantaneous amplitude and instantaneous phase of each subband signal are calculated, and its entropy value is extracted as a characteristic parameter;

[0081] The SD1, SD2 parameters, approximate entropy and sample entropy features of the Poincare map of the heart rate variability are calculated.

[0082] In this embodiment, the EEG signal usually contains multiple frequency components, reflecting different brain activities. Wavelet transform decomposes the EEG signal into sub-band signals of different frequencies to achieve multi-resolution analysis. The Hilbert-Huang transform analyzes non-stationary signals, extracts the instantaneous frequency, instantaneous amplitude and instantaneous phase of the signal, and uses the entropy value to extract the nonlinear characteristics of the signal to reflect the dynamic changes of the EEG signal. The Poincare map parameters (SD1, SD2), approximate entropy and sample entropy features reflect the complexity and regularity of heart rate variability and indirectly reflect the physiological state.

[0083] Specifically, wavelet basis functions such as Daubechies wavelet and Symlets wavelet and the number of decomposition layers are selected to perform wavelet decomposition on the preprocessed EEG signal to obtain sub-band signals of different frequency bands. For example, the EEG signal is decomposed into frequency bands such as Delta (0.5-4Hz), Theta (4-8Hz), Alpha (8-13Hz), Beta (13-30Hz) and Gamma (>30Hz). The Hilbert-Huang transform performs empirical mode decomposition (EMD) on each sub-band signal to obtain a series of intrinsic mode functions (IMFs), and each IMF is Hilbert transformed to calculate its instantaneous frequency, instantaneous amplitude and instantaneous phase. The entropy value of each sub-band signal (such as Shannon entropy, approximate entropy, sample entropy, etc.) is calculated as a characteristic parameter, and the entropy values ​​are calculated for the instantaneous frequency, instantaneous amplitude and instantaneous phase respectively. Extract the RR interval (the time interval between two adjacent R waves) from the electrocardiogram (ECG) or photoplethysmography (PPG) signal, construct the Poincare map of the RR interval, and calculate the SD1 and SD2 parameters of the Poincare map. Among them, SD1 reflects short-term heart rate variability, and SD2 reflects long-term heart rate variability. Calculate the approximate entropy and sample entropy of the RR interval to reflect the complexity and regularity of heart rate variability.

[0084] In an optional manner, calculating the instantaneous frequency and amplitude of each intrinsic mode function in the electroencephalogram signal according to the empirical mode decomposition algorithm to obtain preliminary result data further includes:

[0085] Take the original signal as the residual signal and add it to the list. When the number of extreme points of the residual signal is greater than 2, find all local maxima and local minima of the current residual signal.

[0086] Fitting the local maximum and the local minimum respectively by cubic spline interpolation to obtain an upper envelope and a lower envelope, and calculating the average values ​​of the upper envelope and the lower envelope;

[0087] Subtracting an average value of the upper envelope and the lower envelope from the current residual signal to obtain a candidate residual signal;

[0088] When the number of extreme value points of the candidate residual signal is less than or equal to 2, the remaining residual signals are added to the list to obtain the preliminary result data.

[0089] In this embodiment, EMD decomposes the signal into a series of IMFs with different frequency scales, each of which represents the main oscillation mode of the signal in a certain time period, and well captures the local time-frequency characteristics of the signal.

[0090] The step of analyzing each subband signal by Hilbert-Huang transform, calculating the instantaneous frequency, instantaneous amplitude and instantaneous phase of each subband signal, and extracting the entropy value thereof as a characteristic parameter further includes:

[0091] Decomposing the original signal into a plurality of sub-band signals, wherein the sub-band signals are obtained by filter banks, wavelet transform or short-time Fourier transform;

[0092] Perform empirical mode decomposition on each sub-band signal to obtain the intrinsic mode function;

[0093] Perform Hilbert transform on each intrinsic mode function to obtain analytical signal, and calculate the instantaneous amplitude, instantaneous phase and instantaneous frequency;

[0094] The Shannon entropy of instantaneous amplitude, instantaneous frequency and instantaneous phase are calculated as characteristic parameters.

[0095] Specifically, the original signal is used as the residual signal, and an empty list is created to store the extracted IMF. As long as the number of extreme points of the current residual signal is greater than 2, all local maximum points and local minimum points of the current residual signal are found. All local maximum points are fitted using cubic spline interpolation to obtain the upper envelope, and all local minimum points are fitted using cubic spline interpolation to obtain the lower envelope. The average values ​​of the upper envelope and the lower envelope are calculated. The average envelope is subtracted from the current residual signal to obtain the candidate residual signal. If the number of extreme points of the candidate residual signal is less than or equal to 2, it means that the fluctuation of the signal is very small, and it can be considered that an IMF has been extracted. The candidate residual signal is added to the empty list. The remaining residual signal is used as the new residual signal to start the next round of iteration. The iteration stops when the number of extreme points of the residual signal is less than or equal to 2. The residual signal left at the end is called the residual, which represents the trend term or DC component of the signal.

[0096] According to the solution provided by the present invention, it includes a first processing module, which is used to analyze the patient's physiological signal data collected by the first server to generate preliminary result data of epileptic seizures, wherein the physiological signal data includes electroencephalogram, heart rate variability and activity level; the preliminary result data is stored in the form of a first data structure, and the first data structure includes a timestamp, a physiological data feature vector and a preliminary risk assessment value; a second processing module, which is used to send the first data structure to a second server, identify potential epileptic seizure patterns in the first data structure according to a hidden Markov model, generate an intermediate state sequence, and store the intermediate state sequence as a second data structure, and the second data structure includes power spectral density, waveform characteristics and time series patterns; a prediction analysis module, which is used to perform prediction analysis based on the second data structure to obtain a predicted probability of epileptic seizures in patients; generate an epileptic seizure risk score based on the predicted probability and clinical experience rules; and store the predicted The probability and the epileptic seizure risk score and the event trigger time are combined into a third data structure; wherein the third data structure includes an epilepsy risk score level and an epilepsy warning state identifier, and the epilepsy warning state identifier includes an icon code, a color code and curve data; a risk level division module is used to divide the epileptic seizure risk score into risk levels to generate a fourth data structure; wherein the fourth data structure includes a timestamp, an epilepsy risk score level, a warning level, an event trigger time and a color code corresponding to the risk level; an early warning information generation module is used to receive the fourth data structure and generate a fifth data structure according to the size, resolution and color display capability of the bracelet display screen; wherein the fifth data structure includes a timestamp, a risk level icon, a vibration mode code and a dynamic early warning information curve; a bracelet display module is used to send the fifth data structure to a nurse wearing a bracelet, and the bracelet vibrates and flashes according to the fifth data structure. The present invention further improves the accuracy and timeliness of early warning by identifying potential epileptic seizure patterns through a hidden Markov model. Predictive analysis is performed based on the intermediate state sequence to generate the predicted probability and risk score of epileptic seizures, and a detailed risk score level is generated in combination with clinical experience rules. After receiving the early warning, the nurse can quickly judge the patient's specific situation and take effective intervention measures. At the same time, it reduces the nurse's manual interpretation of EEG data, reduces the workload, and improves work efficiency.

[0097] Figure 3 The schematic diagram of the structure of the computing device embodiment of the present invention is shown. The specific embodiment of the present invention does not limit the specific implementation of the computing device.

[0098] like Figure 3As shown, the computing device may include: a processor (processor) 302 , a communication interface (Communications Interface) 304 , a memory (memory) 306 , and a communication bus 308 .

[0099] The processor 302, the communication interface 304, and the memory 306 communicate with each other via the communication bus 308. The communication interface 304 is used to communicate with other devices such as a client or other server network elements. The processor 302 is used to execute the program 310, which can specifically execute the relevant steps in the above-mentioned embodiment of the smart early warning bracelet for epilepsy monitoring.

[0100] Specifically, the program 310 may include program codes, which include computer operation instructions.

[0101] The processor 302 may be a central processing unit (CPU), or an application-specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present invention. The one or more processors included in the computing device may be processors of the same type, such as one or more CPUs; or processors of different types, such as one or more CPUs and one or more ASICs.

[0102] The memory 306 is used to store the program 310. The memory 306 may include a high-speed RAM memory, and may also include a non-volatile memory (non-volatile memory), such as at least one disk memory.

[0103] According to the solution provided by the present invention, it includes a first processing module, which is used to analyze the patient's physiological signal data collected by the first server to generate preliminary result data of epileptic seizures, wherein the physiological signal data includes electroencephalogram, heart rate variability and activity level; the preliminary result data is stored in the form of a first data structure, and the first data structure includes a timestamp, a physiological data feature vector and a preliminary risk assessment value; a second processing module, which is used to send the first data structure to a second server, identify potential epileptic seizure patterns in the first data structure according to a hidden Markov model, generate an intermediate state sequence, and store the intermediate state sequence as a second data structure, and the second data structure includes power spectral density, waveform characteristics and time series patterns; a prediction analysis module, which is used to perform prediction analysis based on the second data structure to obtain a predicted probability of epileptic seizures in patients; generate an epileptic seizure risk score based on the predicted probability and clinical experience rules; and store the predicted The probability and the epileptic seizure risk score and the event trigger time are combined into a third data structure; wherein the third data structure includes an epilepsy risk score level and an epilepsy warning state identifier, and the epilepsy warning state identifier includes an icon code, a color code and curve data; a risk level division module is used to divide the epileptic seizure risk score into risk levels to generate a fourth data structure; wherein the fourth data structure includes a timestamp, an epilepsy risk score level, a warning level, an event trigger time and a color code corresponding to the risk level; an early warning information generation module is used to receive the fourth data structure and generate a fifth data structure according to the size, resolution and color display capability of the bracelet display screen; wherein the fifth data structure includes a timestamp, a risk level icon, a vibration mode code and a dynamic early warning information curve; a bracelet display module is used to send the fifth data structure to a nurse wearing a bracelet, and the bracelet vibrates and flashes according to the fifth data structure. The present invention further improves the accuracy and timeliness of early warning by identifying potential epileptic seizure patterns through a hidden Markov model. Predictive analysis is performed based on the intermediate state sequence to generate the predicted probability and risk score of epileptic seizures, and a detailed risk score level is generated in combination with clinical experience rules. After receiving the early warning, the nurse can quickly judge the patient's specific situation and take effective intervention measures. At the same time, it reduces the nurse's manual interpretation of EEG data, reduces the workload, and improves work efficiency.

[0104] Those skilled in the art will appreciate that the modules in the devices in the embodiments can be adaptively changed and set in one or more devices different from the embodiments. The modules or units or components in the embodiments can be combined into one module or unit or component, and in addition they can be divided into multiple submodules or subunits or subcomponents. Except that at least some of such features and / or processes or units are mutually exclusive, all features disclosed in this specification (including the accompanying claims, abstracts and drawings) and all processes or units of any method or device disclosed in this manner can be combined in any combination. Unless otherwise explicitly stated, each feature disclosed in this specification (including the accompanying claims, abstracts and drawings) can be replaced by an alternative feature that provides the same, equivalent or similar purpose. In addition, those skilled in the art will appreciate that although some embodiments herein include certain features included in other embodiments but not other features, the combination of features of different embodiments means being within the scope of the present invention and forming different embodiments. For example, in the following claims, any one of the claimed embodiments can be used in any combination. The present invention can be implemented by hardware including several different elements and by appropriately programmed computers. In a unit claim that lists several means, several of these means may be embodied by the same hardware item. The steps in the above embodiments should not be understood as limiting the execution order unless otherwise specified.

Claims

1. An intelligent early warning bracelet for epilepsy monitoring, characterized in that: include: a first processing module, configured to analyze the physiological signal data of the patient collected by the first server to generate preliminary result data of epileptic seizures, wherein the physiological signal data includes electroencephalogram, heart rate variability and activity level; the preliminary result data is stored in the form of a first data structure, wherein the first data structure includes a timestamp, a physiological data feature vector and a preliminary risk assessment value; a second processing module, configured to send the first data structure to a second server, identify potential epileptic seizure patterns in the first data structure according to a hidden Markov model, generate an intermediate state sequence, and store the intermediate state sequence as a second data structure, wherein the second data structure includes a power spectral density, a waveform feature, and a time series pattern; A prediction analysis module, configured to perform prediction analysis according to the second data structure to obtain a predicted probability of an epileptic seizure in a patient; generate an epileptic seizure risk score according to the predicted probability and clinical experience rules; and combine the predicted probability, the epileptic seizure risk score, and the event trigger time into a third data structure; wherein the third data structure includes an epilepsy risk score level and an epilepsy warning status identifier, and the epilepsy warning status identifier includes an icon code, a color code, and curve data; A risk level classification module, used to classify the epileptic seizure risk score into risk levels to generate a fourth data structure; wherein the fourth data structure includes a timestamp, an epileptic seizure risk score level, an early warning level, an event triggering time, and a color code corresponding to the risk level; The warning information generation module is used to receive the fourth data structure and generate a fifth data structure according to the size, resolution and color display capability of the wristband display screen; wherein the fifth data structure includes a timestamp, a risk level icon, a vibration mode code and a dynamic warning information curve; The bracelet display module is used to send the fifth data structure to the nurse wearing the bracelet, and the bracelet vibrates and flashes according to the fifth data structure.

2. The smart early warning bracelet for epilepsy monitoring according to claim 1, characterized in that: In the first processing module, the EEG signal is decomposed into a plurality of intrinsic mode functions according to an adaptive filtering method of variational mode decomposition; For each intrinsic mode function, the cutoff frequency of the filter is adaptively adjusted according to its signal-to-noise ratio.

3. The smart early warning bracelet for epilepsy monitoring according to claim 1, characterized in that: In the second processing module, the state space of the hidden Markov model includes a wakeful state, a pre-epileptic seizure state, an epileptic seizure state, and a post-epileptic seizure state; wherein the calculation formula for the transition probability between the states is: in, is the static state transition probability obtained by pre-training; O t is the physiological data feature vector at time t in the first data structure; μ i is the average value of the physiological feature vector of the i-th state; d(O t ,μ i ) is the distance between the observation vector and the average eigenvector of the i-th state; γ and δ are adjustable parameters; Δ(t) is the time interval function, Δ(t) = exp(-k*(tt prev )), t prev is the time of the last state transition, and k is the time decay coefficient.

4. The smart early warning bracelet for epilepsy monitoring according to claim 1, characterized in that: In the first processing module, generating preliminary result data of epileptic seizures further comprises: Segmenting the physiological signal data, wherein each window contains data samples of a preset time length and a preset overlap rate is set between windows; The instantaneous frequency and amplitude of each intrinsic mode function in the EEG signal were calculated according to the empirical mode decomposition algorithm to obtain preliminary result data.

5. The smart early warning bracelet for epilepsy monitoring according to claim 1, characterized in that: In the risk grading module, epileptic seizure risk scores are graded according to an adaptive threshold adjustment method; Among them, the epileptic seizure risk score is standardized by the cumulative distribution function, and the risk score is converted into a value between 0 and 1; the adaptive threshold of each risk level is calculated based on the patient's historical risk score data.

6. The smart early warning bracelet for epilepsy monitoring according to claim 1, characterized in that: After receiving the fifth data structure, the bracelet display module performs synchronization according to the timestamp in the fifth data structure to ensure that the time on the bracelet is consistent with that on the server; The display screen of the wristband displays icons according to the risk level icon code in the fifth data structure, and the background color of the display screen of the wristband is switched to a color corresponding to the risk level according to the color code in the fifth data structure.

7. The smart early warning bracelet for epilepsy monitoring according to claim 1, characterized in that: In the first processing module, analyzing the physiological signal data further includes: Decomposing the EEG signal into a plurality of sub-band signals of different frequencies by a multi-resolution analysis method of wavelet transform; Each subband signal is analyzed by Hilbert-Huang transform, the instantaneous frequency, instantaneous amplitude and instantaneous phase of each subband signal are calculated, and its entropy value is extracted as a characteristic parameter; The SD1, SD2 parameters, approximate entropy and sample entropy features of the Poincare map of the heart rate variability are calculated.

8. The smart early warning bracelet for epilepsy monitoring according to claim 4, characterized in that: The step of calculating the instantaneous frequency and amplitude of each intrinsic mode function in the electroencephalogram signal according to the empirical mode decomposition algorithm to obtain preliminary result data further includes: Take the original signal as the residual signal and add it to the list. When the number of extreme points of the residual signal is greater than 2, find all local maxima and local minima of the current residual signal. Fitting the local maximum and the local minimum respectively by cubic spline interpolation to obtain an upper envelope and a lower envelope, and calculating the average values ​​of the upper envelope and the lower envelope; Subtracting an average value of the upper envelope and the lower envelope from the current residual signal to obtain a candidate residual signal; When the number of extreme value points of the candidate residual signal is less than or equal to 2, the remaining residual signals are added to the list to obtain the preliminary result data.

9. The smart early warning bracelet for epilepsy monitoring according to claim 7, characterized in that: The step of analyzing each sub-band signal by Hilbert-Huang transform, calculating the instantaneous frequency, instantaneous amplitude and instantaneous phase of each sub-band signal, and extracting the entropy value thereof as a characteristic parameter further comprises: Decomposing the original signal into a plurality of sub-band signals, wherein the sub-band signals are obtained by filter banks, wavelet transform or short-time Fourier transform; Perform empirical mode decomposition on each sub-band signal to obtain the intrinsic mode function; Perform Hilbert transform on each intrinsic mode function to obtain analytical signal, and calculate the instantaneous amplitude, instantaneous phase and instantaneous frequency; The Shannon entropy of instantaneous amplitude, instantaneous frequency and instantaneous phase are calculated as characteristic parameters.

10. A computing device comprising: A processor, a memory, a communication interface and a communication bus, wherein the processor, the memory and the communication interface communicate with each other via the communication bus; The memory is used to store at least one executable instruction, and the executable instruction enables the processor to execute the operation corresponding to the above-mentioned smart early warning bracelet for epilepsy monitoring.

Images