Time domain feature extraction method and system of electroencephalogram signal

Abstract

The application relates to the technical field of electroencephalogram signal processing, and discloses a time-domain feature extraction method and system for electroencephalogram signals, which comprises the following steps: receiving an electroencephalogram signal, adopting an adaptive window technology to dynamically segment the electroencephalogram signal, extracting the segmented signal segment by using a deep learning model, applying an improved time-domain feature extraction algorithm to extract final signal features, and outputting a feature extraction result.The method can use a deep learning model iRNN to evaluate the importance of features, identify and select the features that most contribute to classification from complex EEG signals, improve the stability of an EEG signal processing flow through a feature selection and normalization step, and reduce fluctuations of the signal generated under different environments.

Description

Technical Field

[0001] This invention relates to the field of electroencephalogram (EEG) signal processing technology, and in particular to a method, system, computer device, and medium for extracting time-domain features of EEG signals. Background Technology

[0002] Characteristics and Classification of Electroencephalogram (EEG) Signals: Because EEG signals generated by the cerebral cortex are produced by the combined action of multiple nerve cells, and the brain is a highly complex structure, the acquired EEG signals have the following characteristics: 1) Extremely low signal intensity, resulting in weak anti-interference ability, and the signals themselves are prone to fluctuations under different environments. In this case, the primary task is signal denoising, minimizing interference from external factors through various methods. 2) Nonlinearity and non-stationarity of EEG signals: Due to the diverse external stimuli received by the brain, EEG signals are highly varied and complex. 3) EEG signals are not simple signal waves, but rather more complex synthetic waves, thus exhibiting individual differences. The waveform of the brainwave varies from person to person, and not only that, the frequency components also differ, with brainwaves appearing in different forms at different times. Summary of the Invention

[0003] In view of the aforementioned existing problems, the present invention is proposed.

[0004] Therefore, this invention provides a method for extracting temporal features from EEG signals, which can use a deep learning model iRNN to evaluate the importance of features, identify and select the features that contribute most to classification from complex EEG signals, and improve the stability of the EEG signal processing flow and reduce the fluctuations of signals under different environments through feature selection and normalization steps.

[0005] To address the aforementioned technical problems, this invention provides the following technical solution: a method for extracting temporal features from electroencephalogram (EEG) signals, comprising: receiving EEG signals; dynamically segmenting the EEG signals using adaptive windowing technology; extracting the segmented signal segments using a deep learning model; applying an improved temporal feature extraction algorithm to extract the final signal features; and outputting the feature extraction results.

[0006] As a preferred embodiment of the time-domain feature extraction method for electroencephalogram (EEG) signals described in this invention, the method of receiving EEG signals includes capturing the electrical activity of the brain through an EEG electrode cap and receiving the EEG activity recorded by the electrodes, electrode location information, signal quality indicators, timestamps, and sampling rates.

[0007] After receiving the signal, preprocessing is performed to improve signal quality.

[0008] As a preferred embodiment of the time-domain feature extraction method for EEG signals described in this invention, the dynamic segmentation includes dynamically adjusting the window size based on the characteristics of the EEG signal, namely the peak interval and the coefficient of variation of the signal waveform.

[0009] Set an initial window size W0 and a dynamic adjustment factor δ based on signal stability, and extract the following features from the EEG signal within the current window: calculate the peak interval IPI and the coefficient of variation CoV;

[0010] The window size W is dynamically adjusted based on the feature score.

[0011] W new =W old ·(1+δ·S(x)) n

[0012] Among them, W old It is the current window size, W new is the new window size, n is the exponential adjustment coefficient, and S(x) represents the feature score.

[0013] As a preferred embodiment of the time-domain feature extraction method for EEG signals described in this invention, the extraction of segmented signal segments includes: using a deep learning model that is an improved recurrent neural network (iRNN) to capture long-term dependencies in the time series; converting the signal segments into a format suitable for RNN input; and using a variant of the gated recurrent unit to dynamically adjust the weights of the forget gate and the input gate according to the characteristics of the signal segments.

[0014] Among them, a variant of the gated recurrent unit is called an adaptive gated unit, which introduces a time-varying weight adjustment factor α. t Dynamic calculation based on the statistical properties of signal segments:

[0015]

[0016] Among them, std(X) t ) is the standard deviation of the signal segment, mean(X) t ) is the mean, and ∈ is a small positive number to avoid division by zero;

[0017] Each signal segment is processed by an iRNN, which outputs a feature sequence H = [h1, h2, ..., h T ], where h t Let be the eigenvector at time t, where t = 1, 2, ..., T.

[0018] As a preferred embodiment of the temporal feature extraction method for electroencephalogram (EEG) signals according to the present invention, the improved temporal feature extraction algorithm includes: further improving the feature quality based on the features extracted by the deep learning model, and initializing the temporal weight vector Λ=[λ1,λ2,...,λ T The temporal weights are dynamically adjusted based on the features extracted by the iRNN; the temporal weights are updated, and the features f extracted by the iRNN are... t Used to calculate time-series weight λ t This is then used to calculate the fusion feature TWFF.

[0019] As a preferred embodiment of the time-domain feature extraction method for EEG signals described in this invention, the improved time-domain feature extraction algorithm further includes selecting the feature subset that contributes the most to classification based on feature importance scoring of a deep learning model.

[0020] An improved iRNN model is used to train the preprocessed data. For each feature, an importance score is calculated using the trained iRNN model. Based on the feature importance score, an adaptive weight W for each feature is calculated. f ;

[0021] Based on the calculated adaptive weight W f We select the K features with the highest weights as the feature subset that contributes the most to the classification.

[0022] As a preferred embodiment of the time-domain feature extraction method for EEG signals described in this invention, the extraction of final signal features includes: performing feature normalization using a method based on signal energy distribution; selecting a subset of features; and then performing energy calculation on each selected feature.

[0023] For each feature, calculate the energy E. f :

[0024]

[0025] Where, x f,i μ is the i-th value of feature f. f and σ f These are the mean and standard deviation of feature f, respectively;

[0026] Using the calculated characteristic energy E f The selected features are normalized to eliminate the influence of dimensions and differences in numerical scale between features.

[0027] Another objective of this invention is to provide a time-domain feature extraction system for electroencephalogram (EEG) signals, which enables the extraction of time-domain features of EEG signals, providing effective feature input for EEG signal analysis.

[0028] As a preferred embodiment of the time-domain feature extraction system for EEG signals described in this invention, it includes: a signal acquisition and preprocessing module, an adaptive signal segmentation module, a deep learning feature extraction module, a feature optimization and weight adjustment module, and a feature selection and normalization module;

[0029] The signal acquisition and preprocessing module is responsible for capturing brain electrical activity through the EEG electrode cap, receiving and recording EEG signals and related information, and performing preliminary signal preprocessing to improve signal quality.

[0030] The adaptive signal segmentation module uses adaptive window technology to dynamically segment the signal based on the characteristics of the EEG signal, ensuring that the signal features within each segmentation window can be effectively extracted.

[0031] The deep learning feature extraction module uses an improved recurrent neural network (iRNN) to process the segmented signal segments, capture long-term dependencies in the time series, output feature sequences, and introduces an adaptive gating unit to dynamically adjust network parameters.

[0032] The feature optimization and weight adjustment module optimizes the feature sequence output by the deep learning model, including dynamically adjusting the temporal weights, calculating the importance scores of the features, and updating the temporal weights based on these scores to further refine the features.

[0033] The feature selection and normalization module selects the subset of features that contribute the most to the classification based on the importance score of the features, and uses a method based on signal energy distribution to normalize the selected features to ensure the consistency and comparability of the features.

[0034] A computer device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the steps of a method for extracting temporal features of electroencephalogram (EEG) signals.

[0035] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of a method for extracting temporal features from electroencephalogram (EEG) signals.

[0036] The beneficial effects of this invention are as follows: By preprocessing the raw data, the quality of the data in subsequent analysis is ensured, which is crucial for training deep learning models. Training with an improved iRNN model accurately identifies the influence of each feature on the classification result, thereby calculating feature importance scores, which helps in understanding the intrinsic structure of the data and the relationships between features. By calculating adaptive weights and selecting the subset of features with the highest weights, this invention adaptively retains the features most influential on the model's predictions, which not only reduces feature dimensionality but also improves model performance and prediction efficiency. Through a normalization method based on signal energy distribution, this invention ensures feature consistency and comparability, eliminating the influence of dimensionality and numerical scale differences between features, which contributes to a more stable and accurate deep learning model during training. Attached Figure Description

[0037] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. Wherein:

[0038] Figure 1 A schematic flowchart of a method for extracting temporal features of electroencephalogram (EEG) signals according to an embodiment of the present invention is shown.

[0039] Figure 2 This is a schematic diagram of the working module of a time-domain feature extraction system for electroencephalogram (EEG) signals provided in one embodiment of the present invention. Detailed Implementation

[0040] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the protection scope of the present invention.

[0041] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0042] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0043] This invention is described in detail with reference to the schematic diagrams. When detailing the embodiments of this invention, for ease of explanation, the cross-sectional views illustrating the device structure may be partially enlarged, not adhering to the usual scale. Furthermore, the schematic diagrams are merely examples and should not be construed as limiting the scope of protection of this invention. In actual fabrication, the three-dimensional spatial dimensions of length, width, and depth should be included.

[0044] Furthermore, in the description of this invention, it should be noted that the terms "upper," "lower," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. These terms are used solely for the convenience of describing the invention and for simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the invention. In addition, the terms "first," "second," or "third" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0045] Unless otherwise explicitly specified and limited, the terms "installation," "connection," and "joining" in this invention should be interpreted broadly. For example, they can refer to fixed connections, detachable connections, or integral connections; similarly, they can refer to mechanical connections, electrical connections, or direct connections, or indirect connections through an intermediate medium, or internal connections between two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0046] Characteristics and Classification of Electroencephalogram (EEG) Signals: Because EEG signals generated by the cerebral cortex are produced by the combined action of multiple nerve cells, and the brain is a highly complex structure, the acquired EEG signals have the following characteristics: 1) Extremely low signal intensity, resulting in weak anti-interference ability, and the signals themselves are prone to fluctuations under different environments. In this case, the primary task is signal denoising, minimizing interference from external factors through various methods. 2) Nonlinearity and non-stationarity of EEG signals: Due to the diverse external stimuli received by the brain, EEG signals are highly varied and complex. 3) EEG signals are not simple signal waves, but rather more complex synthetic waves, thus exhibiting individual differences. The waveform of the brainwave varies from person to person, and not only that, the frequency components also differ, with brainwaves appearing in different forms at different times.

[0047] Example 1, referring to Figure 1This is the first embodiment of the present invention, which provides a method for extracting time-domain features of electroencephalogram (EEG) signals, including:

[0048] S1: Receives EEG signals and uses adaptive windowing technology to dynamically segment the EEG signals.

[0049] Furthermore, receiving EEG signals includes capturing the brain's electrical activity through an EEG electrode cap, and receiving the EEG activity recorded by the electrodes, electrode location information, signal quality indicators, timestamps, and sampling rate.

[0050] After receiving the signal, preprocessing is performed to improve signal quality.

[0051] The window size is dynamically adjusted based on the characteristics of EEG signals, namely the peak interval and the coefficient of variation of the signal waveform.

[0052] Set an initial window size W0 and a dynamic adjustment factor δ based on signal stability, and extract the following features from the EEG signal within the current window: calculate the peak interval IPI and the coefficient of variation CoV;

[0053] The window size W is dynamically adjusted based on the feature score.

[0054] W new =W old ·(1+δ·S(x)) n

[0055] Among them, W old It is the current window size, W new is the new window size, n is the exponential adjustment coefficient, and S(x) represents the feature score.

[0056] It should also be noted that the feature score S(x) is evaluated using the following formula:

[0057]

[0058] Among them, IPI max σ is the maximum acceptable crest interval, IPI0 is the reference value for the crest interval, CoV0 is the reference value for the coefficient of variation, and CoV is the coefficient of variation.

[0059] S2: Use a deep learning model to extract the segmented signal fragments.

[0060] Furthermore, the deep learning model is an improved recurrent neural network (iRNN) that captures long-term dependencies in time series, converts signal segments into a format suitable for RNN input, and uses a variant of gated recurrent units to dynamically adjust the weights of the forget gate and input gate according to the characteristics of the signal segments.

[0061] A variant of the gated loop unit is called an adaptive gated unit:

[0062] h t =(1-z) t )⊙h t-1 +z t ⊙tanh(W h ·[h t-1 ,x t ]+b h )

[0063] Among them, h t It is output in the hidden state, z t It is the update gate of the adaptive gating unit, ⊙ represents element-wise multiplication, W h and b h It consists of weights and biases:

[0064] z t =σ(W z ·[h t-1 ,x t ,α t ]+b z )

[0065] Among them, h t-1 It is the previous hidden state, W z b z It updates the weight matrix and bias vector of the gate;

[0066] Introducing a time-varying weighting adjustment factor α t Dynamic calculation based on the statistical properties of signal segments:

[0067]

[0068] Among them, std(X) t ) is the standard deviation of the signal segment, mean(X) t ) is the mean, and ∈ is a small positive number to avoid division by zero;

[0069] Each signal segment is processed by an iRNN, which outputs a feature sequence H = [h1, h2, ..., h T ], where h t Let be the eigenvector at time t, where t = 1, 2, ..., T.

[0070] S3: Apply an improved time-domain feature extraction algorithm to extract the final signal features and output the feature extraction results.

[0071] Furthermore, based on the features extracted by the deep learning model, the feature quality is further improved by initializing the temporal weight vector Λ=[λ1,λ2,...,λ TThe temporal weights are dynamically adjusted based on the features extracted by the iRNN; the temporal weights are updated as follows:

[0072]

[0073] Where, λ t It is the time series weight at time t, f t Here, γ is the feature extracted by iRNN at time t, γ is the temperature parameter, and LeakyReLU is a linear rectified function with leakage. The fused feature TWFF is calculated as follows:

[0074]

[0075] TWFF is the fused feature vector.

[0076] It should be noted that the improved temporal feature extraction algorithm also includes feature importance scoring based on deep learning models, and selecting the feature subset that contributes the most to classification;

[0077] An improved iRNN model is used to train the preprocessed data. For each feature, an importance score is calculated using the trained iRNN model. Based on the feature importance score, an adaptive weight W for each feature is calculated. f :

[0078]

[0079] Where α is the adjustment parameter and N is the total number of features.

[0080] Based on the calculated adaptive weight W f We select the K features with the highest weights as the feature subset that contributes the most to the classification.

[0081] Furthermore, the extraction of final signal features includes: using a method based on signal energy distribution for feature normalization; selecting a subset of features; and then performing energy calculation on each selected feature.

[0082] For each feature, calculate the energy E. f :

[0083]

[0084] Where, x f,i μ is the i-th value of feature f. f and σ f These are the mean and standard deviation of feature f, respectively;

[0085] Using the calculated characteristic energy E f The selected features are normalized to eliminate the influence of dimensions and differences in numerical scale between features:

[0086]

[0087] Where, x′ f,i These are the normalized eigenvalues.

[0088] Example 2 is an embodiment of the present invention, which provides a method for extracting temporal features of electroencephalogram (EEG) signals. To verify the beneficial effects of the present invention, scientific demonstration is carried out through experiments.

[0089] This study aimed to validate the effectiveness of a novel method for extracting temporal features from electroencephalogram (EEG) signals. Twenty healthy subjects, half male and half female, aged 25 to 35 years, were selected for the experiment. During the experiment, subjects wore EEG electrode caps with 19 electrode points placed according to the international 10-20 system. EEG signals were recorded at a sampling rate of 1,000 Hz for 5 minutes. After signal acquisition, preprocessing was performed, including bandpass filtering from 0.5 to 70 Hz and power line interference removal at 50 Hz.

[0090] Next, adaptive windowing technology was applied to dynamically segment the preprocessed EEG signals. The initial window size was set to 256 sampling points, and the dynamic adjustment factor δ was set to 0.1. The window size was dynamically adjusted based on the peak interval and coefficient of variation of the EEG signals within each window. Subsequently, an improved recurrent neural network (iRNN) was used to process the segmented signal segments to capture long-term dependencies in the time series. The iRNN model includes adaptive gating units, which can dynamically adjust the weights of the forget gate and the input gate according to the statistical characteristics of the signal segments. The experimental results are shown in Table 1.

[0091] Table 1

[0092]

[0093]

[0094] Based on the experimental data above, the method of this invention demonstrates significant advantages in extracting temporal features from EEG signals. First, the application of adaptive windowing technology significantly improves the flexibility of signal segmentation. The average adjusted window size is 283 sampling points, which is more adaptable to the dynamic changes in EEG signals compared to the initial window size of 256 sampling points.

[0095] Secondly, the iRNN model effectively captures long-term dependencies in the time series by using adaptive gating units when processing signal segments. The average feature importance score is 0.84, indicating that the extracted features contribute significantly to the classification task. Furthermore, the average signal-to-noise ratio is 21 dB, demonstrating that the preprocessing step effectively improves signal quality.

[0096] Regarding classification accuracy, the average accuracy of this method is 91.5%, which is 6.5 percentage points higher than the average accuracy of existing technologies (approximately 85%). This result demonstrates that the method of this invention has higher accuracy and reliability in EEG signal classification tasks.

[0097] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

[0098] Example 3, the third embodiment of the present invention, differs from the previous two embodiments in that:

[0099] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, essentially, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0100] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0101] More specific examples of computer-readable media (a non-exhaustive list) include: electrical connections (electronic devices) having one or more wires, portable computer disk drives (magnetic devices), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Furthermore, computer-readable media can even be paper or other suitable media on which the program can be printed, because the program can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in computer memory.

[0102] It should be understood that various parts of the present invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0103] Example 4, refer to Figure 2 As an embodiment of the present invention, a temporal feature extraction system for electroencephalogram (EEG) signals is provided, including a signal acquisition and preprocessing module 1, an adaptive signal segmentation module 2, a deep learning feature extraction module 3, a feature optimization and weight adjustment module 4, and a feature selection and normalization module 5.

[0104] The signal acquisition and preprocessing module is responsible for capturing brain electrical activity through the EEG electrode cap, receiving and recording EEG signals and related information, and performing preliminary signal preprocessing to improve signal quality.

[0105] The adaptive signal segmentation module uses adaptive window technology to dynamically segment the signal based on the characteristics of EEG signals, ensuring that the signal features within each segmentation window can be effectively extracted.

[0106] The deep learning feature extraction module uses an improved recurrent neural network (iRNN) to process the segmented signal segments, capture long-term dependencies in the time series, output feature sequences, and introduces an adaptive gating unit to dynamically adjust network parameters.

[0107] The feature optimization and weight adjustment module optimizes the feature sequence output by the deep learning model, including dynamically adjusting the temporal weights, calculating the importance scores of the features, and updating the temporal weights based on these scores to further refine the features.

[0108] The feature selection and normalization module selects the subset of features that contribute the most to the classification based on the importance score of the features, and uses a method based on signal energy distribution to normalize the selected features to ensure the consistency and comparability of the features.

[0109] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for extracting temporal features from electroencephalogram (EEG) signals, characterized in that: include, The system receives EEG signals and uses adaptive windowing technology to dynamically segment the EEG signals. Use deep learning models to extract segmented signal fragments; An improved time-domain feature extraction algorithm is applied to extract the final signal features, and the feature extraction results are output. The process of receiving EEG signals includes capturing the brain's electrical activity through an EEG electrode cap, receiving the EEG activity recorded by the electrodes, electrode location information, signal quality indicators, timestamps, and sampling rate; and performing preprocessing after receiving the signals to improve signal quality. The dynamic segmentation includes dynamically adjusting the window size based on the characteristics of the EEG signal, namely the peak interval of the EEG signal and the coefficient of variation of the signal waveform. Set the initial window size And a dynamic adjustment factor δ based on signal stability, to extract the following features from the EEG signal in the current window: calculate the peak interval IPI, calculate the coefficient of variation CoV; The window size W is dynamically adjusted based on the feature score. in, It is the current window size. This is the new window size, and n is the exponential adjustment factor. The feature score is represented by the expression for the feature score. , It is the maximum acceptable crest interval. It's about adjusting parameters. This is a reference value for the crest interval. This is a reference value for the coefficient of variation. Represents the value of the feature.

2. The method for extracting time-domain features of electroencephalogram (EEG) signals as described in claim 1, characterized in that: The extracted segmented signal fragments include a deep learning model that is an improved recurrent neural network (iRNN) to capture long-term dependencies in the time series, convert the signal fragments into a format suitable for RNN input, and use a variant of the gated recurrent unit to dynamically adjust the weights of the forget gate and the input gate according to the characteristics of the signal fragments. Among them, a variant of the gated recurrent unit is called an adaptive gated unit, which introduces a time-varying weight adjustment factor. Dynamic calculation based on the statistical properties of signal segments: ； in, It is the standard deviation of the signal segment. It is the mean. It is a positive number to avoid division by zero. The signal segment at time t represents the signal segment at time t; Each signal segment is processed by an iRNN, which outputs a feature sequence. ,in, It is the eigenvector at time t. T represents the total number of time steps of the feature sequence.

3. The method for extracting temporal features of electroencephalogram (EEG) signals as described in claim 2, characterized in that: The improved temporal feature extraction algorithm includes further improving the feature quality based on the features extracted by the deep learning model and initializing the temporal weight vector. The temporal weights are dynamically adjusted based on the features extracted by iRNN; For time-series weight updates, the features extracted by iRNN Used to calculate time-series weights This is then used to calculate the fusion feature TWFF, where, It is the time series weight at time t. These are the features extracted by iRNN at time t.

4. The method for extracting temporal features of electroencephalogram (EEG) signals as described in claim 3, characterized in that: The improved temporal feature extraction algorithm also includes selecting the feature subset that contributes the most to classification based on feature importance scoring of a deep learning model. An improved iRNN model is used to train the preprocessed data. For each feature, an importance score is calculated using the trained iRNN model, and an adaptive weight is calculated for each feature based on the feature importance score. ; Based on the calculated adaptive weights We select the K features with the highest weights as the feature subset that contributes the most to the classification.

5. The method for extracting temporal features of electroencephalogram (EEG) signals as described in claim 4, characterized in that: The extraction of final signal features includes: normalizing features using a method based on signal energy distribution; selecting a subset of features; and calculating the energy of each selected feature. For each feature, calculate the energy. : ； in, It is the i-th value of feature f. and These are the mean and standard deviation of feature f, respectively, and M represents the total number of sampling points of feature f in a single signal segment; Using the calculated characteristic energy The selected features are normalized to eliminate the influence of dimensions and differences in numerical scale between features.

6. A system employing the time-domain feature extraction method for electroencephalogram (EEG) signals as described in any one of claims 1 to 5, characterized in that: It includes a signal acquisition and preprocessing module, an adaptive signal segmentation module, a deep learning feature extraction module, a feature optimization and weight adjustment module, and a feature selection and normalization module; The signal acquisition and preprocessing module is responsible for capturing brain electrical activity through the EEG electrode cap, receiving and recording EEG signals and related information, and performing preliminary signal preprocessing to improve signal quality. The adaptive signal segmentation module uses adaptive window technology to dynamically segment the signal based on the characteristics of the EEG signal, ensuring that the signal features within each segmentation window can be effectively extracted. The deep learning feature extraction module uses an improved recurrent neural network (iRNN) to process the segmented signal segments, capture long-term dependencies in the time series, output feature sequences, and introduces an adaptive gating unit to dynamically adjust network parameters. The feature optimization and weight adjustment module optimizes the feature sequence output by the deep learning model, including dynamically adjusting the temporal weights, calculating the importance scores of the features, and updating the temporal weights based on these scores to further refine the features. The feature selection and normalization module selects the subset of features that contribute the most to the classification based on the importance score of the features, and uses a method based on signal energy distribution to normalize the selected features to ensure the consistency and comparability of the features.

7. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 5.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 5.

Images