Data analysis and early warning method for human health monitoring

Abstract

The invention discloses a data analysis and early warning method for human health monitoring, and belongs to the technical field of electroencephalogram monitoring. The method comprises the following steps: firstly, processing electroencephalogram time domain signals in sleep and waking states by using a sliding window, obtaining frequency domain signals through Fourier transform, and dividing the frequency domain signals into three frequency band signals; then, the deviations of the three frequency band signals in the frequency domain gravity center and energy in the two states are calculated respectively, and a corresponding frequency difference sequence and a corresponding energy difference sequence are obtained; then, features are extracted from the sequences, a feature matrix is constructed, the feature matrix is processed through a multi-channel multi-scale neural network, and an electroencephalogram anomaly evaluation value is obtained. And when the evaluation value exceeds a threshold value, the system triggers risk early warning, so that accurate evaluation and timely early warning of the electroencephalogram abnormality are realized.

Description

Technical Field

[0001] This invention relates to the field of electroencephalogram (EEG) monitoring technology, specifically to a data analysis and early warning method for human health monitoring. Background Technology

[0002] Electroencephalography (EEG), as an important physiological signal, is widely used in human health monitoring, especially for the early diagnosis and risk warning of neurological diseases and sleep disorders. By analyzing EEG signals, it is possible to effectively monitor an individual's neural activity and identify potential health risks, such as epileptic seizures, anxiety disorders, and depression. Therefore, developing efficient EEG signal analysis technologies will greatly enhance the capabilities of health monitoring and early intervention.

[0003] Existing EEG signal analysis techniques have many limitations. Most methods only analyze EEG signals in a single state (such as wakefulness or sleep), ignoring the correlation and differences in EEG characteristics in different states. This makes it difficult to comprehensively reflect the dynamic changes in brain function and fully explore the complex health information contained in EEG signals, resulting in low accuracy in assessing EEG abnormalities. Summary of the Invention

[0004] In view of the above-mentioned shortcomings in the prior art, the present invention provides a data analysis and early warning method for human health monitoring, which solves the problem of low accuracy in the assessment of electroencephalogram (EEG) abnormalities in the prior art.

[0005] To achieve the above-mentioned objectives, the technical solution adopted by this invention is: a data analysis and early warning method for human health monitoring, comprising the following steps:

[0006] A sliding window was used to slide across the EEG time-domain signals in both sleep and wake states. After each slide, a Fourier transform was performed on the signal segment under the sliding window to obtain the EEG frequency-domain signal.

[0007] The EEG frequency domain signal corresponding to each sliding window is divided into the first frequency band signal, the second frequency band signal, and the third frequency band signal;

[0008] Based on the deviations of the two states at the centroids of the frequency domain of the first, second, and third frequency band signals, the first frequency difference sequence, the second frequency difference sequence, and the third frequency difference sequence are obtained.

[0009] Based on the energy deviations of the first, second, and third frequency band signals in the two states, a first energy difference sequence, a second energy difference sequence, and a third energy difference sequence are obtained.

[0010] For each frequency difference sequence, extract frequency difference features and construct three frequency difference feature matrices. For each energy difference sequence, extract energy difference features and construct three energy difference feature matrices.

[0011] A multi-channel, multi-scale neural network is used to process three frequency difference feature matrices and three energy difference feature matrices to obtain an EEG abnormality assessment value. When the EEG abnormality assessment value exceeds the threshold, a risk warning is issued.

[0012] Furthermore, the frequency range of the first frequency band signal is 0~10Hz, the frequency range of the second frequency band signal is 10~20Hz, and the frequency range of the third frequency band signal is 20~40Hz.

[0013] Furthermore, the process of obtaining the first frequency difference sequence, the second frequency difference sequence, and the third frequency difference sequence includes:

[0014] The first frequency difference coefficient is obtained by subtracting the spectral frequency centroid of the first frequency band signal in the awake state from the spectral frequency centroid of the first frequency band signal in the sleep state, and then normalizing the difference in spectral frequency centroids.

[0015] The second frequency difference coefficient is obtained by subtracting the spectral frequency centroid of the second frequency band signal in the sleep state from the spectral frequency centroid of the second frequency band signal in the awake state, and then normalizing the difference in spectral frequency centroids.

[0016] The third frequency difference coefficient is obtained by subtracting the spectral frequency centroid of the third frequency band signal in the sleep state from the spectral frequency centroid of the third frequency band signal in the awake state, and then normalizing the difference in spectral frequency centroids.

[0017] Arrange the first frequency difference coefficients in chronological order of their occurrence to obtain the first frequency difference sequence;

[0018] Arrange the second frequency difference coefficients in chronological order of their occurrence to obtain the second frequency difference sequence;

[0019] Arrange the third frequency difference coefficients in chronological order to obtain the third frequency difference sequence.

[0020] Furthermore, the process of obtaining the first energy difference sequence, the second energy difference sequence, and the third energy difference sequence includes:

[0021] The energy of the first frequency band signal in the waking state is subtracted from the energy of the first frequency band signal in the sleeping state, and the energy difference is normalized to obtain the first energy difference coefficient.

[0022] The energy of the second frequency band signal in the waking state is subtracted from the energy of the second frequency band signal in the sleeping state, and the energy difference is normalized to obtain the second energy difference coefficient.

[0023] The energy of the third frequency band signal in the waking state is subtracted from the energy of the third frequency band signal in the sleeping state, and the energy difference is normalized to obtain the third energy difference coefficient.

[0024] Arrange the first energy difference coefficients in chronological order to obtain the first energy difference sequence;

[0025] Arrange the second energy difference coefficients in chronological order to obtain the second energy difference sequence;

[0026] Arrange the third energy difference coefficients in chronological order to obtain the third energy difference sequence.

[0027] Furthermore, the process of constructing the three frequency difference feature matrices and the three energy difference feature matrices includes:

[0028] Both the frequency difference sequence and the energy difference sequence are divided into N parts to obtain multiple frequency difference subsequences and multiple energy difference subsequences, where N is a positive integer;

[0029] Extract the mean frequency difference, frequency difference fluctuation value, and frequency difference discrete span from the frequency difference subsequence;

[0030] Extract the mean energy difference, energy difference fluctuation value, and energy difference discrete span from the energy difference subsequences respectively;

[0031] The mean frequency difference, frequency difference fluctuation, and frequency difference discrete span of each frequency difference subsequence of the same frequency difference sequence are used as elements to construct an N×3 frequency difference feature matrix;

[0032] The mean energy difference, fluctuation value, and discrete span of each energy difference subsequence of the same energy difference sequence are used as elements to construct an N×3 energy difference feature matrix.

[0033] Furthermore, the mean frequency difference is the mean of each frequency difference coefficient in the frequency difference subsequence, the frequency difference fluctuation is the variance of each frequency difference coefficient in the frequency difference subsequence, and the frequency difference dispersion span is the difference between the maximum and minimum frequency difference coefficients in the frequency difference subsequence.

[0034] The mean energy difference is the mean of the energy difference coefficients in the energy difference subsequence, the fluctuation of the energy difference is the variance of the energy difference coefficients in the energy difference subsequence, and the discrete span of the energy difference is the difference between the maximum and minimum energy difference coefficients in the energy difference subsequence.

[0035] Furthermore, the multi-channel multi-scale neural network includes: a first feature fusion enhancement channel, a second feature fusion enhancement channel, a third feature fusion enhancement channel, a multi-scale feature extraction unit, a splicing layer, and a fully connected layer;

[0036] The first input terminal of the first feature fusion enhancement channel is used to input the first frequency difference feature matrix, and its second input terminal is used to input the first energy difference feature matrix; the first input terminal of the second feature fusion enhancement channel is used to input the second frequency difference feature matrix, and its second input terminal is used to input the second energy difference feature matrix; the first input terminal of the third feature fusion enhancement channel is used to input the third frequency difference feature matrix, and its second input terminal is used to input the third energy difference feature matrix.

[0037] The first input of the multi-scale feature extraction unit is connected to the output of the first feature fusion enhancement channel, its second input is connected to the output of the second feature fusion enhancement channel, and its third input is connected to the output of the third feature fusion enhancement channel. The input of the splicing layer is connected to the first and second outputs of the multi-scale feature extraction unit, respectively, and its output is connected to the input of the fully connected layer. The output of the fully connected layer serves as the output of the multi-channel multi-scale neural network.

[0038] Furthermore, the first feature fusion enhancement channel, the second feature fusion enhancement channel, and the third feature fusion enhancement channel all include: a feature fusion module and a feature enhancement module;

[0039] The feature fusion module is used to extract features from the frequency difference feature matrix and the energy difference feature matrix, and perform feature fusion to obtain fused features; the feature enhancement module is used to enhance the fused features to obtain enhanced features.

[0040] Furthermore, the feature fusion module includes: a first convolutional layer, a second convolutional layer, a third convolutional layer, a fourth convolutional layer, and an adder A1;

[0041] The input of the first convolutional layer serves as the first input of the feature fusion module, and its output is connected to the input of the second convolutional layer. The input of the third convolutional layer serves as the second input of the feature fusion module, and its output is connected to the input of the fourth convolutional layer. The input of adder A1 is connected to the outputs of the second and fourth convolutional layers, respectively, and its output serves as the output of the feature fusion module.

[0042] Furthermore, the multi-scale feature extraction unit includes: a fifth convolutional layer, a sixth convolutional layer, a seventh convolutional layer, an eighth convolutional layer, a ninth convolutional layer, a tenth convolutional layer, adder A2, and adder A3;

[0043] The input of the fifth convolutional layer is connected to the input of the sixth convolutional layer and serves as the first input of the multi-scale feature extraction unit; the input of the seventh convolutional layer is connected to the input of the eighth convolutional layer and serves as the second input of the multi-scale feature extraction unit; the input of the ninth convolutional layer is connected to the input of the tenth convolutional layer and serves as the third input of the multi-scale feature extraction unit.

[0044] The input of adder A2 is connected to the output of the fifth convolutional layer, the output of the seventh convolutional layer, and the output of the ninth convolutional layer, respectively, and its output serves as the first output of the multi-scale feature extraction unit.

[0045] The input of adder A3 is connected to the outputs of the sixth, eighth and tenth convolutional layers, respectively, and its output serves as the second output of the multi-scale feature extraction unit.

[0046] The beneficial effects of this invention are as follows:

[0047] 1. This invention focuses on EEG signals in both sleep and wakefulness states. By calculating the frequency domain centroid deviation and energy deviation in different states at various frequency bands, it fully utilizes the correlation and differences of EEG characteristics in the two states, breaking through the limitations of traditional single-state analysis. It can more comprehensively reflect the dynamic changes of brain neural activity, thereby effectively improving the accuracy of EEG abnormality identification.

[0048] 2. This invention performs Fourier transform on the EEG time-domain signal under each sliding window to obtain the EEG frequency-domain signal, and then divides the EEG frequency-domain signal into three segments. By analyzing the frequency-domain centroid deviation under different states, it can reflect the activity level of the brain in each frequency band and potential functional disorders. The changes in energy reflect the dynamic changes in neural activity and comprehensively reflect the dynamic changes in brain function.

[0049] 3. This invention extracts features from the frequency difference subsequences and energy difference subsequences corresponding to the three frequency bands, constructing three frequency difference feature matrices and three energy difference feature matrices. This enables precise capture of the specific differences of each frequency band in both sleep and wakefulness states. The deviations in the frequency domain centroid and energy of EEG signals in different frequency bands reflect the degree of abnormality in the corresponding neural activity. Processing them separately avoids mutual interference between features of different frequency bands, ensuring that the feature information of each frequency band can be fully extracted.

[0050] 4. This invention uses a multi-channel, multi-scale neural network to process three frequency difference feature matrices and three energy difference feature matrices, fully mining the deep correlation information in the feature matrices, and combining frequency difference features and energy difference features to predict EEG abnormality assessment values, thereby improving the accuracy of EEG abnormality assessment values. Attached Figure Description

[0051] Figure 1 A flowchart for a data analysis and early warning method for human health monitoring;

[0052] Figure 2 This is a schematic diagram of the structure of a multi-channel, multi-scale neural network;

[0053] Figure 3This is a schematic diagram of the structure of the first feature fusion enhancement channel, the second feature fusion enhancement channel, and the third feature fusion enhancement channel;

[0054] Figure 4 This is a schematic diagram of the feature fusion module.

[0055] Figure 5 This is a schematic diagram of the structure of a multi-scale feature extraction unit;

[0056] Figure 6 This is a schematic diagram of the feature enhancement module. Detailed Implementation

[0057] The specific embodiments of the present invention are described below to enable those skilled in the art to understand the present invention. However, it should be understood that the present invention is not limited to the scope of the specific embodiments. For those skilled in the art, various changes are obvious as long as they are within the spirit and scope of the present invention as defined and determined by the appended claims. All inventions utilizing the concept of the present invention are protected.

[0058] like Figure 1 As shown, a data analysis and early warning method for human health monitoring includes the following steps:

[0059] A sliding window was used to slide across the EEG time-domain signals in both sleep and wake states. After each slide, a Fourier transform was performed on the signal segment under the sliding window to obtain the EEG frequency-domain signal.

[0060] The EEG frequency domain signal corresponding to each sliding window is divided into the first frequency band signal, the second frequency band signal, and the third frequency band signal;

[0061] Based on the deviations of the two states at the centroids of the frequency domain of the first, second, and third frequency band signals, the first frequency difference sequence, the second frequency difference sequence, and the third frequency difference sequence are obtained.

[0062] Based on the energy deviations of the first, second, and third frequency band signals in the two states, a first energy difference sequence, a second energy difference sequence, and a third energy difference sequence are obtained.

[0063] For each frequency difference sequence, extract frequency difference features and construct three frequency difference feature matrices. For each energy difference sequence, extract energy difference features and construct three energy difference feature matrices.

[0064] A multi-channel, multi-scale neural network is used to process three frequency difference feature matrices and three energy difference feature matrices to obtain an EEG abnormality assessment value. When the EEG abnormality assessment value exceeds the threshold, a risk warning is issued.

[0065] In this embodiment, a window length of 1 second and a sampling rate of 256 Hz are used, resulting in 256 window points, a sliding step of 0.25 seconds, a window overlap rate of 75%, and a signal length of 30 seconds for both sleep and wakefulness EEG time-domain signals.

[0066] In this embodiment, the frequency range of the first frequency band signal is 0~10Hz, the frequency range of the second frequency band signal is 10~20Hz, and the frequency range of the third frequency band signal is 20~40Hz. The EEG frequency domain signal belonging to the 0~10Hz frequency domain is classified as the first frequency band signal, the EEG frequency domain signal belonging to the 10~20Hz frequency domain is classified as the second frequency band signal, and the EEG frequency domain signal belonging to the 20~40Hz frequency domain is classified as the third frequency band signal.

[0067] During sleep, brain activity decreases, and signals in the 0-10Hz frequency band dominate, characterized by large amplitude and slow frequency. In contrast, during wakefulness, this frequency band signal changes depending on the state of relaxation or focus.

[0068] During the transition from sleep to wakefulness, the 10-20Hz signal gradually strengthens; in the waking state, this frequency band is associated with concentration and mild cognitive activity; while in the sleep state, the signal in this frequency band is significantly weakened or even suppressed by low-frequency signals.

[0069] In a waking state, especially when engaging in complex thinking or active activities, the 20-40Hz high-frequency signal is significantly enhanced; while in a sleep state (especially non-REM sleep), this frequency band signal is almost suppressed, and only slightly rises during REM sleep (a state similar to wakefulness), but the intensity is much lower than in a waking state.

[0070] In this embodiment, the process of obtaining the first frequency difference sequence, the second frequency difference sequence, and the third frequency difference sequence includes:

[0071] The first frequency difference coefficient is obtained by subtracting the spectral frequency centroid of the first frequency band signal in the awake state from the spectral frequency centroid of the first frequency band signal in the sleep state, and then normalizing the difference in spectral frequency centroids.

[0072] The second frequency difference coefficient is obtained by subtracting the spectral frequency centroid of the second frequency band signal in the sleep state from the spectral frequency centroid of the second frequency band signal in the awake state, and then normalizing the difference in spectral frequency centroids.

[0073] The third frequency difference coefficient is obtained by subtracting the spectral frequency centroid of the third frequency band signal in the sleep state from the spectral frequency centroid of the third frequency band signal in the awake state, and then normalizing the difference in spectral frequency centroids.

[0074] Arrange the first frequency difference coefficients in chronological order of their occurrence to obtain the first frequency difference sequence;

[0075] Arrange the second frequency difference coefficients in chronological order of their occurrence to obtain the second frequency difference sequence;

[0076] Arrange the third frequency difference coefficients in chronological order to obtain the third frequency difference sequence.

[0077] The spectral frequency centroid (the frequency point where the spectral energy is concentrated) is a core indicator reflecting the "dominant frequency" of a signal within a frequency band: the dominant frequency of the same frequency band will undergo a characteristic shift between sleep and wakefulness (e.g., the centroid of the 0~10Hz frequency band shifts towards the low-frequency delta wave during sleep and towards the high-frequency alpha wave when awake).

[0078] The three frequency bands of this invention correspond to the frequency characteristics of different brain activity intensities (low frequency reflects the basal state, mid-to-high frequency reflects the transitional state, and high frequency reflects the active state). This invention obtains the first, second, and third frequency difference sequences through the above process, which can accurately quantify the differences in the spectral frequency centroids of the three frequency band signals during sleep and wakefulness.

[0079] In this embodiment, the formula for calculating the spectral frequency centroid is: Where Z is the spectral frequency centroid, f k Let y be the k-th frequency value in a frequency band signal. k Let K be the amplitude corresponding to the k-th frequency value in a frequency band signal, where K is the number of frequency values ​​in a frequency band signal.

[0080] In this embodiment, the formula for normalizing the spectral frequency centroid difference is: , where μ z Z is the frequency difference coefficient. d For the difference in the centroid of the spectral frequency, Z max The centroid difference represents the maximum spectral frequency.

[0081] In this embodiment, the process of obtaining the first energy difference sequence, the second energy difference sequence, and the third energy difference sequence includes:

[0082] The energy of the first frequency band signal in the waking state is subtracted from the energy of the first frequency band signal in the sleeping state, and the energy difference is normalized to obtain the first energy difference coefficient.

[0083] The energy of the second frequency band signal in the waking state is subtracted from the energy of the second frequency band signal in the sleeping state, and the energy difference is normalized to obtain the second energy difference coefficient.

[0084] The energy of the third frequency band signal in the waking state is subtracted from the energy of the third frequency band signal in the sleeping state, and the energy difference is normalized to obtain the third energy difference coefficient.

[0085] Arrange the first energy difference coefficients in chronological order to obtain the first energy difference sequence;

[0086] Arrange the second energy difference coefficients in chronological order to obtain the second energy difference sequence;

[0087] Arrange the third energy difference coefficients in chronological order to obtain the third energy difference sequence.

[0088] This invention uses first, second, and third energy difference sequences to accurately capture the energy differences of three frequency bands in sleep and wakefulness states. By arranging the energy difference coefficients in chronological order to form a sequence, it fully records the dynamic evolution of energy differences during the transition between the two states. The three energy difference sequences comprehensively present the energy distribution differences between sleep and wakefulness states from three spectral dimensions: low, medium, and high.

[0089] In this embodiment, the formula for calculating the energy of the frequency band signal is: Where E is the energy of the frequency band signal.

[0090] In this embodiment, the formula for normalizing the energy difference is: , where μ E E is the energy difference coefficient. d For the energy difference, E max This represents the maximum energy difference.

[0091] In this embodiment, the process of constructing three frequency difference feature matrices and three energy difference feature matrices includes:

[0092] Both the frequency difference sequence and the energy difference sequence are divided into N parts to obtain multiple frequency difference subsequences and multiple energy difference subsequences, where N is a positive integer;

[0093] Extract the mean frequency difference, frequency difference fluctuation value, and frequency difference discrete span from the frequency difference subsequence;

[0094] Extract the mean energy difference, energy difference fluctuation value, and energy difference discrete span from the energy difference subsequences respectively;

[0095] The mean frequency difference, frequency difference fluctuation, and frequency difference discrete span of each frequency difference subsequence of the same frequency difference sequence are used as elements to construct an N×3 frequency difference feature matrix;

[0096] The mean energy difference, fluctuation value, and discrete span of each energy difference subsequence of the same energy difference sequence are used as elements to construct an N×3 energy difference feature matrix.

[0097] The frequency difference feature matrix obtained from the first frequency difference sequence is the first frequency difference feature matrix, the frequency difference feature matrix obtained from the second frequency difference sequence is the second frequency difference feature matrix, and the frequency difference feature matrix obtained from the third frequency difference sequence is the third frequency difference feature matrix.

[0098] The energy difference feature matrix obtained from the first energy difference sequence is the first energy difference feature matrix, the energy difference feature matrix obtained from the second energy difference sequence is the second energy difference feature matrix, and the energy difference feature matrix obtained from the third energy difference sequence is the third energy difference feature matrix.

[0099] In this embodiment, the mean frequency difference is the mean of each frequency difference coefficient in the frequency difference subsequence, the frequency difference fluctuation is the variance of each frequency difference coefficient in the frequency difference subsequence, and the frequency difference dispersion span is the difference between the maximum and minimum frequency difference coefficients in the frequency difference subsequence.

[0100] The mean energy difference is the mean of the energy difference coefficients in the energy difference subsequence, the fluctuation of the energy difference is the variance of the energy difference coefficients in the energy difference subsequence, and the discrete span of the energy difference is the difference between the maximum and minimum energy difference coefficients in the energy difference subsequence.

[0101] This invention divides the frequency difference sequence and energy difference sequence into N parts to obtain subsequences, and then extracts the mean, fluctuation value and discrete span of each subsequence. The mean reflects the overall level of the features in the subsequence, the fluctuation value reflects the stability of the features, and the discrete span shows the distribution range of the features.

[0102] In this embodiment, N is a positive integer greater than or equal to 5. A frequency difference subsequence uses the mean frequency difference, frequency difference fluctuation value, and frequency difference discrete span as elements of each row, resulting in an N×3 frequency difference feature matrix; an energy difference subsequence uses the mean energy difference, energy difference fluctuation value, and energy difference discrete span as elements of each row, resulting in an N×3 energy difference feature matrix.

[0103] In this embodiment, as Figure 2 As shown, the multi-channel multi-scale neural network includes: a first feature fusion enhancement channel, a second feature fusion enhancement channel, a third feature fusion enhancement channel, a multi-scale feature extraction unit, a splicing layer, and a fully connected layer;

[0104] The first input terminal of the first feature fusion enhancement channel is used to input the first frequency difference feature matrix, and its second input terminal is used to input the first energy difference feature matrix; the first input terminal of the second feature fusion enhancement channel is used to input the second frequency difference feature matrix, and its second input terminal is used to input the second energy difference feature matrix; the first input terminal of the third feature fusion enhancement channel is used to input the third frequency difference feature matrix, and its second input terminal is used to input the third energy difference feature matrix.

[0105] The first input of the multi-scale feature extraction unit is connected to the output of the first feature fusion enhancement channel, its second input is connected to the output of the second feature fusion enhancement channel, and its third input is connected to the output of the third feature fusion enhancement channel. The input of the splicing layer is connected to the first and second outputs of the multi-scale feature extraction unit, respectively, and its output is connected to the input of the fully connected layer. The output of the fully connected layer serves as the output of the multi-channel multi-scale neural network.

[0106] This invention employs each feature fusion enhancement channel to process a pair of frequency difference feature matrices and energy difference feature matrices. The frequency difference focuses on the temporal changes of frequency centroid differences, while the energy difference reflects the dynamic laws of energy differences. The fusion of the two allows the channel to simultaneously capture information in both frequency and energy dimensions. Then, a multi-scale feature extraction unit extracts multi-scale features from the fused enhanced features, thereby improving the classification accuracy of the fully connected layer.

[0107] like Figure 3 As shown, the first feature fusion enhancement channel, the second feature fusion enhancement channel, and the third feature fusion enhancement channel all include: a feature fusion module and a feature enhancement module;

[0108] The feature fusion module is used to extract features from the frequency difference feature matrix and the energy difference feature matrix, and perform feature fusion to obtain fused features; the feature enhancement module is used to enhance the fused features to obtain enhanced features.

[0109] like Figure 4 As shown, the feature fusion module includes: a first convolutional layer, a second convolutional layer, a third convolutional layer, a fourth convolutional layer, and an adder A1;

[0110] The input of the first convolutional layer serves as the first input of the feature fusion module, and its output is connected to the input of the second convolutional layer. The input of the third convolutional layer serves as the second input of the feature fusion module, and its output is connected to the input of the fourth convolutional layer. The input of adder A1 is connected to the outputs of the second and fourth convolutional layers, respectively, and its output serves as the output of the feature fusion module.

[0111] In this embodiment, the kernel size of the first and third convolutional layers is 1×1, and the kernel size of the second and fourth convolutional layers is 3×3.

[0112] This invention extracts features from the frequency difference feature matrix through the first and second convolutional layers, extracts features from the energy difference feature matrix through the third and fourth convolutional layers, and then performs bitwise addition through adder A1 to obtain the fused features.

[0113] like Figure 5As shown, the multi-scale feature extraction unit includes: a fifth convolutional layer, a sixth convolutional layer, a seventh convolutional layer, an eighth convolutional layer, a ninth convolutional layer, a tenth convolutional layer, adder A2, and adder A3;

[0114] The input of the fifth convolutional layer is connected to the input of the sixth convolutional layer and serves as the first input of the multi-scale feature extraction unit; the input of the seventh convolutional layer is connected to the input of the eighth convolutional layer and serves as the second input of the multi-scale feature extraction unit; the input of the ninth convolutional layer is connected to the input of the tenth convolutional layer and serves as the third input of the multi-scale feature extraction unit.

[0115] The input of adder A2 is connected to the output of the fifth convolutional layer, the output of the seventh convolutional layer, and the output of the ninth convolutional layer, respectively, and its output serves as the first output of the multi-scale feature extraction unit.

[0116] The input of adder A3 is connected to the outputs of the sixth, eighth and tenth convolutional layers, respectively, and its output serves as the second output of the multi-scale feature extraction unit.

[0117] In this embodiment, the kernel size of the fifth, seventh, and ninth convolutional layers is 1×1, and the kernel size of the sixth, eighth, and tenth convolutional layers is 1×3.

[0118] This invention extracts features at different scales using convolutional kernels of different sizes, and adds features at the same scale bit by bit using an adder to achieve the fusion of features at the same scale. At the splicing layer, the two fused scale features are spliced ​​together to improve the classification accuracy of the fully connected layer.

[0119] like Figure 6 As shown, the feature enhancement module includes: a max pooling layer, an average pooling layer, an eleventh convolutional layer, a twelfth convolutional layer, an adder A4, a sigmoid layer, and a multiplier M1;

[0120] The input of the max pooling layer is connected to the input of the average pooling layer and the first input of the multiplier M1, respectively, and serves as the input of the feature enhancement module.

[0121] The input of the eleventh convolutional layer is connected to the output of the max pooling layer; the input of the twelfth convolutional layer is connected to the output of the average pooling layer; the input of adder A4 is connected to the outputs of the eleventh and twelfth convolutional layers respectively, and its output is connected to the input of the sigmoid layer; the output of the sigmoid layer is connected to the second input of multiplier M1; the output of multiplier M1 serves as the output of the feature enhancement module.

[0122] In this embodiment, the kernel size of the eleventh and twelfth convolutional layers is 1×1.

[0123] Max pooling layers focus on salient features, preserving key anomalies or state transition information; average pooling layers capture global trends. The two are processed in parallel, allowing the network to simultaneously acquire "local salient features" and "mean features." After being processed by the eleventh and twelfth convolutional layers, the "local salient features" and "mean features" are fused in adder A4. Attention is generated through the Sigmoid layer to enhance the output features of the feature fusion module, adaptively focusing on important features.

[0124] In this embodiment, the feature enhancement module can also be replaced by an existing channel attention module.

[0125] In this embodiment, the range of the EEG abnormality assessment value can be set from 0 to 10 points (0 points is completely normal, and 10 points is extremely abnormal), and the threshold is set to 5 points (that is, a warning is triggered if the score exceeds 5 points). For example, a score of 1.5 or 2.3 indicates that the EEG rhythm is normal and there is no risk of abnormality.

[0126] This invention focuses on EEG signals in both sleep and wakefulness states. By calculating the frequency domain centroid deviation and energy deviation in different states across various frequency bands, it fully utilizes the correlation and differences in EEG characteristics between the two states, breaking through the limitations of traditional single-state analysis. It can more comprehensively reflect the dynamic changes in brain neural activity, thereby effectively improving the accuracy of EEG abnormality identification.

[0127] This invention performs Fourier transform on the EEG time-domain signal under each sliding window to obtain the EEG frequency-domain signal, and then divides the EEG frequency-domain signal into three segments. By analyzing the frequency-domain centroid deviation under different states, it can reflect the activity level of the brain in each frequency band and potential functional disorders. The changes in energy reflect the dynamic changes in neural activity and comprehensively reflect the dynamic changes in brain function.

[0128] This invention extracts features from frequency difference subsequences and energy difference subsequences corresponding to three frequency bands, constructing three frequency difference feature matrices and three energy difference feature matrices. This enables precise capture of the specific differences of each frequency band in both sleep and wakefulness states. The deviations in the frequency domain centroid and energy of EEG signals in different frequency bands reflect the degree of abnormality in the corresponding neural activity. Processing them separately avoids mutual interference between features of different frequency bands, ensuring that the feature information of each frequency band can be fully extracted.

[0129] This invention employs a multi-channel, multi-scale neural network to process three frequency difference feature matrices and three energy difference feature matrices, fully mining the deep correlation information in the feature matrices, and combining frequency difference features and energy difference features to predict EEG abnormality assessment values, thereby improving the accuracy of EEG abnormality assessment values.

[0130] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A data analysis and early warning method for human health monitoring, characterized in that, Includes the following steps: A sliding window was used to slide across the EEG time-domain signals in both sleep and wake states. After each slide, a Fourier transform was performed on the signal segment under the sliding window to obtain the EEG frequency-domain signal. The EEG frequency domain signal corresponding to each sliding window is divided into the first frequency band signal, the second frequency band signal, and the third frequency band signal; Based on the deviations of the two states at the centroids of the frequency domain of the first, second, and third frequency band signals, the first frequency difference sequence, the second frequency difference sequence, and the third frequency difference sequence are obtained. Based on the energy deviations of the first, second, and third frequency band signals in the two states, a first energy difference sequence, a second energy difference sequence, and a third energy difference sequence are obtained. For each frequency difference sequence, extract frequency difference features and construct three frequency difference feature matrices. For each energy difference sequence, extract energy difference features and construct three energy difference feature matrices. A multi-channel, multi-scale neural network is used to process three frequency difference feature matrices and three energy difference feature matrices to obtain an EEG abnormality assessment value. When the EEG abnormality assessment value exceeds the threshold, a risk warning is issued.

2. The data analysis and early warning method for human health monitoring according to claim 1, characterized in that, The frequency range of the first frequency band signal is 0~10Hz, the frequency range of the second frequency band signal is 10~20Hz, and the frequency range of the third frequency band signal is 20~40Hz.

3. The data analysis and early warning method for human health monitoring according to claim 1, characterized in that, The process of obtaining the first frequency difference sequence, the second frequency difference sequence, and the third frequency difference sequence includes: The first frequency difference coefficient is obtained by subtracting the spectral frequency centroid of the first frequency band signal in the awake state from the spectral frequency centroid of the first frequency band signal in the sleep state, and then normalizing the difference in spectral frequency centroids. The second frequency difference coefficient is obtained by subtracting the spectral frequency centroid of the second frequency band signal in the sleep state from the spectral frequency centroid of the second frequency band signal in the awake state, and then normalizing the difference in spectral frequency centroids. The third frequency difference coefficient is obtained by subtracting the spectral frequency centroid of the third frequency band signal in the sleep state from the spectral frequency centroid of the third frequency band signal in the awake state, and then normalizing the difference in spectral frequency centroids. Arrange the first frequency difference coefficients in chronological order of their occurrence to obtain the first frequency difference sequence; Arrange the second frequency difference coefficients in chronological order of their occurrence to obtain the second frequency difference sequence; Arrange the third frequency difference coefficients in chronological order to obtain the third frequency difference sequence.

4. The data analysis and early warning method for human health monitoring according to claim 1, characterized in that, The process of obtaining the first energy difference sequence, the second energy difference sequence, and the third energy difference sequence includes: The energy of the first frequency band signal in the waking state is subtracted from the energy of the first frequency band signal in the sleeping state, and the energy difference is normalized to obtain the first energy difference coefficient. The energy of the second frequency band signal in the waking state is subtracted from the energy of the second frequency band signal in the sleeping state, and the energy difference is normalized to obtain the second energy difference coefficient. The energy of the third frequency band signal in the waking state is subtracted from the energy of the third frequency band signal in the sleeping state, and the energy difference is normalized to obtain the third energy difference coefficient. Arrange the first energy difference coefficients in chronological order to obtain the first energy difference sequence; Arrange the second energy difference coefficients in chronological order to obtain the second energy difference sequence; Arrange the third energy difference coefficients in chronological order to obtain the third energy difference sequence.

5. The data analysis and early warning method for human health monitoring according to claim 1, characterized in that, The process of constructing three frequency difference feature matrices and three energy difference feature matrices includes: Both the frequency difference sequence and the energy difference sequence are divided into N parts to obtain multiple frequency difference subsequences and multiple energy difference subsequences, where N is a positive integer; Extract the mean frequency difference, frequency difference fluctuation value, and frequency difference discrete span from the frequency difference subsequence; Extract the mean energy difference, energy difference fluctuation value, and energy difference discrete span from the energy difference subsequences respectively; The mean frequency difference, frequency difference fluctuation, and frequency difference discrete span of each frequency difference subsequence of the same frequency difference sequence are used as elements to construct an N×3 frequency difference feature matrix; The mean energy difference, fluctuation value, and discrete span of each energy difference subsequence of the same energy difference sequence are used as elements to construct an N×3 energy difference feature matrix.

6. The data analysis and early warning method for human health monitoring according to claim 5, characterized in that, The mean frequency difference is the mean of each frequency difference coefficient in the frequency difference subsequence, the frequency difference fluctuation is the variance of each frequency difference coefficient in the frequency difference subsequence, and the frequency difference dispersion span is the difference between the maximum and minimum frequency difference coefficients in the frequency difference subsequence. The mean energy difference is the mean of the energy difference coefficients in the energy difference subsequence, the fluctuation of the energy difference is the variance of the energy difference coefficients in the energy difference subsequence, and the discrete span of the energy difference is the difference between the maximum and minimum energy difference coefficients in the energy difference subsequence.

7. The data analysis and early warning method for human health monitoring according to claim 1, characterized in that, The multi-channel multi-scale neural network includes: a first feature fusion enhancement channel, a second feature fusion enhancement channel, a third feature fusion enhancement channel, a multi-scale feature extraction unit, a splicing layer, and a fully connected layer; The first input terminal of the first feature fusion enhancement channel is used to input the first frequency difference feature matrix, and its second input terminal is used to input the first energy difference feature matrix; the first input terminal of the second feature fusion enhancement channel is used to input the second frequency difference feature matrix, and its second input terminal is used to input the second energy difference feature matrix; the first input terminal of the third feature fusion enhancement channel is used to input the third frequency difference feature matrix, and its second input terminal is used to input the third energy difference feature matrix. The first input of the multi-scale feature extraction unit is connected to the output of the first feature fusion enhancement channel, its second input is connected to the output of the second feature fusion enhancement channel, and its third input is connected to the output of the third feature fusion enhancement channel. The input of the splicing layer is connected to the first and second outputs of the multi-scale feature extraction unit, respectively, and its output is connected to the input of the fully connected layer. The output of the fully connected layer serves as the output of the multi-channel multi-scale neural network.

8. The data analysis and early warning method for human health monitoring according to claim 7, characterized in that, The first feature fusion enhancement channel, the second feature fusion enhancement channel, and the third feature fusion enhancement channel all include: a feature fusion module and a feature enhancement module; The feature fusion module is used to extract features from the frequency difference feature matrix and the energy difference feature matrix, and perform feature fusion to obtain fused features; the feature enhancement module is used to enhance the fused features to obtain enhanced features.

9. The data analysis and early warning method for human health monitoring according to claim 8, characterized in that, The feature fusion module includes: a first convolutional layer, a second convolutional layer, a third convolutional layer, a fourth convolutional layer, and an adder A1; The input of the first convolutional layer serves as the first input of the feature fusion module, and its output is connected to the input of the second convolutional layer. The input of the third convolutional layer serves as the second input of the feature fusion module, and its output is connected to the input of the fourth convolutional layer. The input of adder A1 is connected to the outputs of the second and fourth convolutional layers, respectively, and its output serves as the output of the feature fusion module.

10. The data analysis and early warning method for human health monitoring according to claim 8, characterized in that, The multi-scale feature extraction unit includes: a fifth convolutional layer, a sixth convolutional layer, a seventh convolutional layer, an eighth convolutional layer, a ninth convolutional layer, a tenth convolutional layer, adder A2, and adder A3; The input of the fifth convolutional layer is connected to the input of the sixth convolutional layer and serves as the first input of the multi-scale feature extraction unit; the input of the seventh convolutional layer is connected to the input of the eighth convolutional layer and serves as the second input of the multi-scale feature extraction unit; the input of the ninth convolutional layer is connected to the input of the tenth convolutional layer and serves as the third input of the multi-scale feature extraction unit. The input of adder A2 is connected to the output of the fifth convolutional layer, the output of the seventh convolutional layer, and the output of the ninth convolutional layer, respectively, and its output serves as the first output of the multi-scale feature extraction unit. The input of adder A3 is connected to the outputs of the sixth, eighth and tenth convolutional layers, respectively, and its output serves as the second output of the multi-scale feature extraction unit.

Images