Electrode screening method, device and equipment in mobile state and computer readable storage medium

Abstract

The present disclosure relates to a mobile state electrode screening method, device, equipment and computer readable storage medium. Relates to the field of computer technology, including through the band-pass filter to the target electroencephalogram signal is accurate division get multiple preset frequency band, and combined with the relative power calculation realizes feature normalization, overcomes the defect that the traditional method is not enough to protect the subtle electroencephalogram feature. Secondly, based on the relative power data statistics index of normal user and abnormal user is calculated, and the variation coefficient double verification mechanism is introduced, when the statistical index is less than the first threshold value and the variation coefficient is less than the second threshold value, the target electrode is confirmed, this kind of comprehensive screening standard based on statistical significance and feature stability ensures that the electrode has both discriminant ability and robustness. Through this targeted screening mechanism, the electrode redundancy is significantly reduced, which can focus on the electrodes that are really effective for brain injury classification, so as to realize high-precision classification in mobile environment and optimize the calculation efficiency.

Description

Technical Field

[0001] This disclosure relates to the field of computer technology, and in particular to an electrode screening method, apparatus, device, and computer-readable storage medium in a mobile state. Background Technology

[0002] Electroencephalogram (EEG) signals are key physiological signals that reflect brain neural activity and are widely used in fields such as brain injury assessment and neurological disease monitoring.

[0003] In general, electrode screening is easily affected by motion interference while in motion. Multi-channel EEG devices often suffer from electrode redundancy, with some electrodes acquiring signals with high noise and weak correlation with analysis tasks, making it impossible to accurately screen target electrodes that are effective for different brain injury classifications, resulting in a waste of computing resources.

[0004] Therefore, there is an urgent need for an electrode screening method in a mobile state to solve the above problems. Summary of the Invention

[0005] In order to solve the above-mentioned technical problems, or at least partially solve the above-mentioned technical problems, this disclosure provides an electrode screening method, apparatus, device and computer-readable storage medium in a mobile state.

[0006] In a first aspect, embodiments of this disclosure provide an electrode screening method in a mobile state, comprising: By precisely segmenting the target EEG signal using a bandpass filter to obtain multiple preset frequency bands, and combining this with relative power calculation to achieve feature normalization, the shortcomings of traditional methods in protecting subtle EEG features are overcome. Secondly, statistical indicators are calculated based on the relative power data of normal and abnormal users, and a dual verification mechanism using the coefficient of variation is introduced. Target electrodes are only confirmed when the statistical indicators are below a first threshold and the coefficient of variation is below a second threshold. This comprehensive screening criterion based on statistical significance and feature stability ensures that the electrodes possess both discriminative ability and robustness. This targeted screening mechanism significantly reduces electrode redundancy, focusing on electrodes that are truly effective for classifying brain injuries, thereby achieving high-precision classification while optimizing computational efficiency in mobile environments.

[0007] Secondly, embodiments of this disclosure provide an electrode screening device in a mobile state, comprising: The acquisition module is used to acquire the target EEG signal of a user in motion, the user including normal users and abnormal users, and the target EEG signal is obtained through electrodes of a portable EEG device. The segmentation module is used to segment the target EEG signal using a bandpass filter to obtain target EEG signals in multiple preset frequency bands. The calculation module is used to calculate the relative power of the target EEG signal for each preset frequency band; The first determining module is used to determine the statistical indicators corresponding to each preset frequency band based on the relative power of the normal user and the abnormal user in the preset frequency band. The second determining module is used to determine the electrode as the target electrode when the statistical index corresponding to any of the preset frequency bands is less than the first threshold.

[0008] Thirdly, embodiments of this disclosure provide an electronic device, including: Memory; Processor; and Computer programs; The computer program is stored in the memory and configured to be executed by the processor to implement the method as described in the first aspect.

[0009] Fourthly, embodiments of this disclosure provide a computer-readable storage medium having a computer program stored thereon, the computer program being executed by a processor to implement the method described in the first aspect.

[0010] Fifthly, embodiments of this disclosure also provide a computer program product, which includes a computer program or instructions that, when executed by a processor, implement the method described in the first aspect. Attached Figure Description

[0011] The accompanying drawings, which are incorporated in and form a part of this specification, illustrate embodiments consistent with this disclosure and, together with the description, serve to explain the principles of this disclosure.

[0012] To more clearly illustrate the technical solutions in the embodiments of this disclosure or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, those skilled in the art can obtain other drawings based on these drawings without creative effort.

[0013] Figure 1 This is a flowchart of an electrode screening method in a mobile state provided in an embodiment of this disclosure; Figure 2 This is a schematic diagram of the electrode screening device in a moving state provided in an embodiment of the present disclosure; Figure 3 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this disclosure. Detailed Implementation

[0014] To better understand the above-mentioned objectives, features, and advantages of this disclosure, the solutions disclosed herein will be further described below. It should be noted that, unless otherwise specified, the embodiments and features described herein can be combined with each other.

[0015] Numerous specific details are set forth in the following description in order to provide a full understanding of this disclosure, but this disclosure may also be implemented in other ways different from those described herein; obviously, the embodiments in the specification are only some, and not all, of the embodiments of this disclosure.

[0016] Electroencephalogram (EEG) signals, as key physiological signals reflecting brain neural activity, have significant application value in medical and health fields such as brain injury assessment and neurological disease monitoring. With the development of portable EEG devices, the demand for EEG signal acquisition in mobile settings is increasing, for example, in rapid screening of suspected brain injury patients in mobile emergency scenarios. However, EEG signals in mobile environments are susceptible to motion artifacts, electromyographic interference, and other factors, posing serious challenges to signal quality and analysis reliability.

[0017] Current processing solutions typically employ multi-channel EEG devices for data acquisition, using conventional filtering methods (such as bandpass filtering) to remove significant noise, and directly utilizing the characteristics of all electrode channels for subsequent analysis. Some techniques remove eye-movement artifacts through independent component analysis (ICA) or use motion sensor data to assist in correcting motion interference. In the feature extraction stage, existing techniques often directly calculate the absolute power of each frequency band or perform simple normalization, and use common statistical tests (such as t-tests) to screen features, ultimately using machine learning algorithms for classification.

[0018] However, existing technologies have significant drawbacks: First, multi-channel EEG devices suffer from prominent electrode redundancy issues, with some electrodes acquiring signals exhibiting high noise levels and weak correlation with specific analysis tasks, resulting in insufficient feature effectiveness. They also cannot accurately identify target electrodes effective for different brain injury classifications, leading to wasted computational resources and limited classification accuracy. More importantly, existing technologies do not adequately consider the quantitative assessment of signal stability under movement conditions and lack comprehensive screening criteria that combine statistical significance and feature stability, thus affecting the robustness and practicality of the final classification model.

[0019] This application uses bandpass filters to precisely segment the target EEG signal into multiple preset frequency bands (such as delta and theta), and combines this with relative power calculation to achieve feature normalization, overcoming the shortcomings of traditional methods in protecting subtle EEG features. Secondly, statistical indicators (such as p-values) are calculated based on the relative power data of normal and abnormal users, and a dual verification mechanism using the coefficient of variation is introduced. Target electrodes are only confirmed when the statistical indicator is less than a first threshold and the coefficient of variation is less than a second threshold. This comprehensive screening criterion based on statistical significance and feature stability ensures that the electrodes possess both discriminative ability and robustness. This targeted screening mechanism significantly reduces electrode redundancy, focusing on electrodes that are truly effective for classifying brain injuries, thereby achieving high-precision classification while optimizing computational efficiency in mobile environments.

[0020] To address this problem, this disclosure provides an electrode screening method in a mobile state, which will be described below with reference to specific embodiments.

[0021] Figure 1 This is a flowchart illustrating an electrode screening method in a mobile state according to an embodiment of this disclosure. The method can be executed by an electrode screening device in a mobile state, which can be implemented in software and / or hardware. This device can be configured in an electronic device, such as a server or terminal, where the terminal specifically includes a mobile phone, computer, or tablet computer.

[0022] The following is about Figure 1 The electrode screening method under the shown mobile state is described below, and the specific steps of this method are as follows: S101. Acquire the target EEG signal of the user in motion.

[0023] The computing device can acquire target EEG signals from a user in motion. The user can include both normal and abnormal users. The target EEG signals are obtained through electrodes of a portable EEG device.

[0024] In some possible implementations, the computing device can record the user's raw EEG signals in real time via electrodes from a portable EEG device, while the user is in a mobile state, and acquire the user's motion data via an auxiliary motion sensor. The raw EEG signals are then preprocessed based on the motion data to obtain the target EEG signal.

[0025] For example, the computing device can record multi-channel raw EEG signals in real time using electrodes from a portable EEG device (such as a dry electrode cap or ear clip electrodes) based on the user's motion state (e.g., while lying down in a vehicle). Simultaneously, it uses auxiliary motion sensors (such as accelerometers) to acquire the user's motion data to detect motion artifacts. Then, based on the motion data, the raw EEG signals are preprocessed (e.g., noise reduction, removal of eye movement, motion artifacts, and electromyography artifacts) to obtain a clean target EEG signal. This process ensures signal reliability even in environments with motion interference, laying the foundation for subsequent bandpass filtering to divide preset frequency bands (e.g., delta: 1-4Hz, theta: 4-8Hz, etc.) and calculating relative power, thereby supporting the determination of statistical indicators (such as p-values) and electrode selection. The entire acquisition process emphasizes real-time performance and interference resistance, optimizing signal quality through preprocessing so that the target EEG signal can be effectively used for brain injury classification analysis between normal and abnormal users.

[0026] S102. The target EEG signal is divided by a bandpass filter to obtain target EEG signals in multiple preset frequency bands.

[0027] The computing device can divide the target EEG signal into multiple preset frequency bands using a bandpass filter.

[0028] In some possible implementations, the computing device can use a bandpass filter to remove DC offset below a first preset value and high-frequency noise above a second preset value from the target EEG signal. Based on preset boundary rules and preset frequency bands, the removed target EEG signal is divided to obtain the target EEG signal in the multiple preset frequency bands.

[0029] For example, a computing device can perform key processing on the target EEG signal using a bandpass filter to eliminate interference and classify frequency bands.

[0030] Specifically, the computing device first uses a bandpass filter to remove DC offsets (such as baseline drift) below a first preset value (e.g., 1 Hz) and high-frequency noise (such as electromyography artifacts) above a second preset value (e.g., 100 Hz) from the target EEG signal, thereby purifying the signal and reducing the impact of motion artifacts in a moving environment. Next, based on preset boundary rules (e.g., non-overlapping division, including the lower limit but excluding the upper limit to avoid information loss) and preset frequency bands (e.g., delta: 1-4 Hz, theta: 4-8 Hz, alpha: 8-12 Hz, beta: 12-30 Hz, gamma: 30-100 Hz), the noise-removed target EEG signal is precisely divided into multiple preset frequency bands, providing standardized input for subsequent calculations of relative power and statistical indicators (e.g., p-value). This process ensures the resolvability of the signal in the frequency domain, supporting the accuracy of electrode selection.

[0031] While bandpass filters, which use fixed preset frequency bands (e.g., delta: 1-4Hz), effectively purify signals, their boundary rules lack adaptability to individual differences and dynamic motion scenarios. This embodiment can also introduce machine learning algorithms to dynamically optimize the frequency boundaries. Specifically, after acquiring the user's target EEG signal in motion using electrodes from a portable EEG device, and performing preprocessing based on motion data collected by auxiliary motion sensors (e.g., accelerometer readings), clustering algorithms (e.g., K-means) can be used to perform pattern recognition on the motion data, automatically clustering noise features corresponding to different motion intensities (e.g., stationary, walking, vehicle bumps), thereby dynamically adjusting the frequency boundaries of the bandpass filter. For example, for high-frequency motion interference scenarios, the theta band can be adaptively widened (e.g., from 4-8Hz to 4-9Hz) to encompass more effective signal components; or reinforcement learning algorithms can be used, with classification accuracy as a reward signal, to optimize the filter boundary parameters in real time. This dynamic adjustment mechanism, combined with the original preset boundary rules, can significantly improve the protection of subtle EEG features while retaining the framework of relative power calculation and statistical indicators (e.g., p-value) verification.

[0032] In some possible implementations, motion sensors can be used in conjunction with filtering to remove motion trajectories. The core of this approach lies in the deep integration of high-precision motion data collected by auxiliary motion sensors (such as accelerometers and gyroscopes) with signal processing procedures to achieve direct identification and compensation of motion trajectories.

[0033] Specifically, while acquiring target EEG signals through electrodes of a portable EEG device, auxiliary motion sensors are used to capture the user's motion trajectory features (such as three-dimensional spatial data like head displacement and rotation angle) in real time. Based on this motion data, a mapping model between the motion trajectory and motion artifacts in the EEG signal is first constructed, and then a corresponding artifact template is generated. Subsequently, based on the division of preset frequency bands (e.g., delta: 1-4Hz) by a bandpass filter, an adaptive filtering algorithm is jointly employed to subtract the motion trajectory-related artifact components from the target EEG signal. This motion sensor-based joint filtering mechanism transforms the motion trajectory, traditionally considered interference, into quantifiable correction parameters, significantly improving the ability to suppress motion artifacts in dynamic motion scenarios. It not only compensates for the insufficient adaptability of fixed frequency boundaries to individual motion differences but also provides a cleaner signal input for subsequent calculations of relative power and statistical indicators (such as p-values), ultimately enhancing the accuracy and robustness of electrode selection in mobile environments.

[0034] In some possible implementations, bandpass filters are divided into fixed preset frequency bands (such as delta: 1-4Hz). While this can effectively purify the signal, it lacks the ability to adapt to individual differences and dynamic motion scenarios.

[0035] This embodiment can also incorporate machine learning algorithms to dynamically optimize the frequency boundaries. Specifically, after acquiring the user's target EEG signal during movement via electrodes from a portable EEG device, and preprocessing the motion data (such as accelerometer readings) collected by auxiliary motion sensors, clustering algorithms (such as K-means) can be used to perform pattern recognition on the motion data. This automatically clusters noise features corresponding to different motion intensities (such as stillness, walking, and vehicle bumps), thereby dynamically adjusting the frequency boundaries of the bandpass filter. For example, for high-frequency motion interference scenarios, the theta band can be widened (e.g., from 4-8Hz to 4-9Hz) to encompass more effective signal components; or, through reinforcement learning algorithms, the filter boundary parameters can be optimized using classification accuracy as a reward signal. This dynamic adjustment mechanism, combined with existing preset boundary rules, can significantly improve the protection capability for subtle EEG features while retaining the framework of relative power calculation and statistical indicators (such as p-values) verification.

[0036] In some possible implementations, a bandpass filter is used to segment the target EEG signal. First, DC offset below a first preset value and high-frequency noise above a second preset value are removed. To further purify the signal and reduce interference in mobile environments, a notch filter is added to specifically filter out power frequency noise (e.g., 50Hz). This supplementary step, combined with preset boundary rules (e.g., non-overlapping segmentation) and preset frequency bands (e.g., delta: 1-4Hz, theta: 4-8Hz, etc.), ensures the purity and resolvability of the signal in the frequency domain. This provides a more reliable standardized input for subsequent calculations of relative power and statistical indicators, improving the accuracy of electrode selection.

[0037] S103. For each of the preset frequency bands, calculate the relative power of the target EEG signal.

[0038] The computing device can calculate the relative power of the target EEG signal for each preset frequency band.

[0039] For example, the computing device can perform a fast Fourier transform on the target EEG signal for each preset frequency band to obtain the frequency domain coefficients of the target EEG signal. Based on the frequency domain coefficients, the power spectral density of the target EEG signal can be calculated. Then, the relative power of the preset frequency band can be calculated based on the power spectral density and the band boundary of the preset frequency band.

[0040] For each preset frequency band (e.g., delta: 1-4Hz, theta: 4-8Hz, etc.), the computing device first performs a Fast Fourier Transform on the preprocessed target EEG signal, and then applies the standard Discrete Fourier Transform formula to convert the time-domain signal into frequency-domain coefficients to reveal the frequency components of the signal. The standard Discrete Fourier Transform formula is as follows: ; in, It is a time-domain signal sample (length N). These are the frequency domain coefficients. The power spectral density is then... , used to quantify the energy distribution of a signal at different frequency points, where n is the index of the time-domain sample point, representing the specific position of the signal in the time series. It is a discrete integer variable, ranging from 0 to N-1. For example, when n=0, x(0) represents the amplitude of the first sampling point of the target EEG signal. When n=1, x(1) represents the amplitude of the second sampling point, and so on. N: is the total number of sampling points of the signal, that is, the total number of time-domain samples contained in a single data segment participating in the FFT calculation. It determines the frequency resolution; the larger the value of N, the higher the frequency resolution of the frequency domain coefficients X(k) obtained after transformation.

[0041] Based on the power spectral density and the band boundaries of the preset frequency band (e.g., calculating the absolute power of the delta band through integration, the formula is as follows:) ,in This represents the absolute power in the delta band. This represents the band boundary.

[0042] Normalization is performed using a total power benchmark (e.g., base 1-100Hz) to calculate the relative power of the preset frequency band. This process is the core of feature extraction, providing standardized input for subsequent calculation of statistical indicators (e.g., p-value) and electrode selection.

[0043] S104. Based on the relative power of the normal user and the abnormal user in the preset frequency band, determine the statistical index corresponding to each preset frequency band.

[0044] The computing device can determine the statistical indicators corresponding to each preset frequency band based on the relative power of normal users and abnormal users in the preset frequency band.

[0045] For example, the computing device can determine the statistical indicators, such as p-values, corresponding to each preset frequency band based on the relative power data of normal users and abnormal users in preset frequency bands (such as delta: 1-4Hz, theta: 4-8Hz, etc.) through statistical test methods (such as t-test or Wilcoxon rank-sum test) to quantify the significance of differences between groups.

[0046] The computing device first collects the relative power characteristics of the two groups of users (i.e., the normalized power values ​​of each electrode in each band). Then, it performs hypothesis testing for each band: for parametric tests (such as t-tests), it calculates the mean and standard deviation of the differences between paired samples to obtain the t-statistic and p-value. For nonparametric tests (such as Wilcoxon tests), it sorts and ranks the mixed data, calculates the rank sum and U-statistic, and finally derives the p-value matrix. This statistical indicator (e.g., p-value less than 0.05) is used to determine whether the electrode characteristics have discriminative power. This is a crucial step in electrode screening, ensuring that only bands with significant differences are retained for subsequent classification.

[0047] S105. When the statistical index corresponding to any of the preset frequency bands is less than the first threshold, the corresponding electrode is determined to be the target electrode.

[0048] When the statistical index corresponding to any preset frequency band is less than the first threshold, the corresponding electrode is determined to be the target electrode.

[0049] In some possible implementations, the computing device can also perform statistical tests on the relative power characteristics of each electrode, calculate the relative power difference between normal users and abnormal users in a preset frequency band, and calculate the coefficient of variation of the relative power difference based on the standard deviation and mean of the relative power difference. When the statistical index is less than a first threshold and the coefficient of variation is less than a second threshold, the electrode is determined as the target electrode. Here, the coefficient of variation is a relative index used in statistics to measure the degree of data dispersion. It refers to the ratio of the standard deviation to the mean of the relative power difference between normal users and abnormal users in a preset frequency band, and is used to quantify the stability of the characteristics. When this value is less than the second threshold, it indicates that the electrode characteristics have good robustness.

[0050] In some possible implementations, statistical tests may include parametric tests, such as t-tests; specifically, statistical tests are performed on the relative power characteristics of each electrode to calculate the relative power difference between normal and abnormal users in the preset frequency band, which may specifically include: For each electrode, the number of paired samples for relative power characteristics is determined. Based on the number of paired samples, multiple pairs of differences in relative power between normal users and abnormal users in the preset frequency band are calculated. The number of paired samples corresponds to the multiple pairs of differences. The mean and standard deviation of the multiple pairs of differences are calculated. Finally, the relative power difference value can be calculated based on the mean, standard deviation and number of paired samples.

[0051] In some possible implementations, statistical tests may also include nonparametric tests, such as the Wilcoxon rank-sum test.

[0052] Specifically, a statistical test is performed on the relative power characteristics of each electrode to calculate the relative power difference between normal users and the abnormal users in a preset frequency band. For each electrode, the relative power characteristics of normal users and abnormal users in a preset frequency band are mixed to obtain mixed data. The mixed data is sorted according to its size, and each mixed data is assigned a rank. The first rank sum of the relative power of normal users in the preset frequency band is calculated. Based on the first rank sum and the relative power characteristics of normal users and abnormal users in the preset frequency band, statistical values ​​and critical values ​​are calculated. Finally, the relative power difference value can be calculated based on the size of the statistical value and the critical value.

[0053] In some possible implementations, the computing device can perform statistical tests on the relative power characteristics of each electrode to calculate the relative power difference between normal and abnormal users in a preset frequency band (such as the delta band). This process can employ either parametric or non-parametric testing. Parametric testing, represented by the t-test, may include the following steps: First, determine the number of paired samples for relative power characteristics (i.e., the number of comparable sample pairs in the normal and abnormal groups). Based on this number of paired samples, calculate multiple differences in relative power between the two groups of users in a specific band (subtracting the power values ​​from each pair), with a one-to-one correspondence between the number of paired samples and the number of differences. Next, calculate the mean (reflecting the average direction of the difference) and standard deviation (quantifying the degree of fluctuation in the difference) of these differences. Finally, based on the mean, standard deviation, and number of paired samples, calculate the relative power difference value (i.e., the t-value) using the t-statistic formula. The larger the t-value, the more significant the difference between the groups.

[0054] When data does not satisfy the normal distribution assumption, the Wilcoxon rank-sum test (also known as the Mann-Whitney U test) can be used in nonparametric tests. This test first mixes the relative power characteristics of normal and abnormal users to form a mixed dataset. After sorting by numerical value, each data point is assigned a rank (rank number, with duplicates taking the average rank). The first rank sum of the normal user group data (i.e., the sum of the ranks of all samples in this group) is calculated. Then, the U statistic and its critical value are calculated based on the sample sizes of both groups. Finally, by comparing the statistic with the critical value (e.g., rejecting the null hypothesis if the U value is less than the critical value), the relative power difference value (often converted to a p-value) is calculated. The device further calculates the coefficient of variation (CV = standard deviation / mean) based on the standard deviation and mean of the difference value. When the statistical indicator (e.g., p-value) is less than the first threshold (e.g., 0.05) and the coefficient of variation is less than the second threshold (e.g., 0.25), it indicates that the electrode characteristic has both significant discriminative power and stability, and is therefore identified as the target electrode.

[0055] In some possible implementations, the relative power difference between normal and abnormal users in a preset frequency band is calculated based on statistical tests (such as t-tests or Wilcoxon rank-sum tests), and the coefficient of variation is introduced for double verification. While this ensures the discriminative ability and stability of the electrodes, it is still insufficient for capturing nonlinear or temporal EEG features. In this embodiment, a deep learning model can also be integrated as a supplementary verification method to form a hybrid verification framework.

[0056] Specifically, computing devices can input relative power time-series data (such as continuous time series) into a convolutional neural network (CNN) model. The CNN's convolutional layers extract local patterns and multi-scale correlations in the frequency domain. The CNN's output features can be processed in parallel with traditional statistical indicators (such as p-values). Through a fusion module within the CNN (such as weighted averaging or attention mechanisms), two types of features are adaptively weighted: one is the deep features extracted by the CNN from the relative power time-series data, and the other is the statistical indicator (such as the p-value) obtained by traditional methods. This fusion module can achieve optimal fusion by calculating the weights of different features, ultimately outputting a quantified CNN classification confidence score. The higher the classification confidence score, the stronger the model's judgment. Combined with the coefficient of variation (COP) verification results, the target electrode is confirmed only when the statistical indicator is below a first threshold, the COP is below a second threshold, and the CNN classification confidence score is above a third threshold. This hybrid framework retains the advantages of the original statistical significance and feature stability, while enhancing the ability to identify subtle differences through deep learning of complex EEG time-series patterns by the CNN, thereby further improving the accuracy and robustness of brain injury classification in mobile environments.

[0057] In some possible implementations, electrode screening primarily relies on individualized verification of statistical indicators and coefficients of variation based on the relative power data of individual electrodes within a preset frequency band. While this ensures the independent discriminative ability of each electrode, it fails to consider the functional synergistic relationships between electrodes. To enhance the classification ability of brain network abnormalities, graph theory and network analysis can be introduced beyond the coefficient of variation verification based on the relative power characteristics of individual electrodes. This allows for the construction of a multi-electrode-based brain functional connectivity network, thereby extending the assessment of feature stability from individual electrodes to the dimension of electrode group synergy.

[0058] Specifically, after dividing the target EEG signal into multiple preset frequency bands (such as delta and theta bands) using a bandpass filter and calculating the relative power of each electrode, the functional connectivity strength between any two electrodes can be calculated using phase synchronization or coherence algorithms based on the relative power time-series data of normal and abnormal users, forming an adjacency matrix of the brain functional connectivity network. On this basis, the graph theory features of each electrode in the network (such as node degree centrality and participation coefficient) are calculated, and the variation of these graph theory features within the user group is further analyzed, thereby defining the correlation coefficient of variation for the electrode subset. This correlation coefficient of variation quantifies the cooperative stability of the electrode group in functional connectivity patterns. The final electrode selection criteria can be extended to: an electrode is confirmed as a target electrode only when its statistical index in the preset frequency band is less than a first threshold, its individual relative power coefficient of variation is less than a second threshold, and the correlation coefficient of variation of its functional subnetwork is less than a third threshold. This multi-dimensional verification mechanism ensures that the selected electrodes not only possess individual-level saliency and stability but also exhibit good group synergy at the brain network level, thereby significantly improving the classification accuracy of abnormal patterns in brain injury-related networks.

[0059] This application uses bandpass filters to precisely segment the target EEG signal into multiple preset frequency bands (such as delta and theta), and combines this with relative power calculation to achieve feature normalization, overcoming the shortcomings of traditional methods in protecting subtle EEG features. Secondly, statistical indicators (such as p-values) are calculated based on the relative power data of normal and abnormal users, and a dual verification mechanism using the coefficient of variation is introduced. Target electrodes are only confirmed when the statistical indicator is less than a first threshold and the coefficient of variation is less than a second threshold. This comprehensive screening criterion based on statistical significance and feature stability ensures that the electrodes possess both discriminative ability and robustness. This targeted screening mechanism significantly reduces electrode redundancy, focusing on electrodes that are truly effective for classifying brain injuries, thereby achieving high-precision classification while optimizing computational efficiency in mobile environments.

[0060] Figure 2 This is a schematic diagram of the structure of an electrode screening device in a mobile state provided in an embodiment of this disclosure. The electrode screening device in a mobile state can be a terminal as described in the above embodiment, or it can be a component or assembly within the terminal. The electrode screening device in a mobile state provided in this disclosure can execute the processing flow provided in the embodiments of the electrode screening method in a mobile state, such as... Figure 2 As shown, the electrode screening device 20 in the moving state includes: Acquisition module 21 is used to acquire target EEG signals of a user in motion state, wherein the user includes normal users and abnormal users, and the target EEG signals are obtained through electrodes of a portable EEG device. The segmentation module 22 is used to segment the target EEG signal using a bandpass filter to obtain target EEG signals in multiple preset frequency bands; Calculation module 23 is used to calculate the relative power of the target EEG signal for each preset frequency band; The first determining module 24 is used to determine the statistical indicators corresponding to each preset frequency band based on the relative power of the normal user and the abnormal user in the preset frequency band. The second determining module 25 is used to determine the electrode as the target electrode when the statistical index corresponding to any of the preset frequency bands is less than the first threshold.

[0061] Optional, the acquisition module is specifically used for: Based on the user being in a mobile state, the user's original brainwave signals are recorded in real time through the electrodes of the portable EEG device, and the user's motion data is acquired through an auxiliary motion sensor; The original EEG signal is preprocessed based on the motion data to obtain the target EEG signal.

[0062] Optionally, divide into modules, specifically for: The bandpass filter removes DC offset below a first preset value and high-frequency noise above a second preset value from the target EEG signal. Based on preset boundary rules and preset frequency bands, the removed target EEG signals are divided to obtain target EEG signals in multiple preset frequency bands.

[0063] Optional, a calculation module, specifically used for: For each of the preset frequency bands, a fast Fourier transform is performed on the target EEG signal to obtain the frequency domain coefficients of the target EEG signal; Based on the frequency domain coefficients, the power spectral density of the target EEG signal is calculated; The relative power of the preset frequency band is calculated based on the power spectral density and the band boundary of the preset frequency band.

[0064] Optional, the second determining module is specifically used for: Statistical tests are performed on the relative power characteristics of each electrode to calculate the relative power difference between the normal user and the abnormal user in the preset frequency band. Calculate the coefficient of variation of the relative power difference based on the standard deviation and mean of the relative power difference values; When the statistical index is less than a first threshold and the coefficient of variation is less than a second threshold, the electrode is determined to be the target electrode.

[0065] Optionally, the statistical test includes a parametric test, which includes a t-test; The second determining module is specifically used for: For each of the electrodes, determine the number of paired samples for the relative power characteristics; Based on the number of paired samples, multiple pairs of differences in the relative power of the normal user and the abnormal user in the preset frequency band are calculated, and the number of paired samples and the multiple pairs of differences have a corresponding relationship. Calculate the mean and standard deviation of the multiple pairs of differences; The relative power difference value is calculated based on the mean, the standard deviation, and the number of paired samples.

[0066] Optionally, the statistical test may also include a nonparametric test, which may include the Wilcoxon rank-sum test. The second determining module is specifically used for: For each electrode, the relative power characteristics of the normal user and the abnormal user in the preset frequency band are mixed to obtain mixed data; The mixed data is sorted according to its size, and each mixed data item is assigned a rank. Calculate the first rank sum of the relative power of the normal user in the preset frequency band; Statistical values ​​and critical values ​​are calculated based on the first rank sum, the relative power characteristics of the normal user and the abnormal user in the preset frequency band; The relative power difference value is calculated based on the statistical value and the magnitude of the critical value.

[0067] Figure 2 The electrode screening device in the moving state shown in the embodiment can be used to execute the technical solution of the electrode screening method embodiment in the moving state described above. Its implementation principle and technical effect are similar, and will not be repeated here.

[0068] Figure 3 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this disclosure. The electronic device can be a terminal as described in the above embodiments. The electronic device provided in this disclosure can execute the processing flow provided in the embodiments of the electrode screening method in a mobile state, such as... Figure 3 As shown, the electronic device 70 includes: a memory 71, a processor 72, a computer program, and a communication interface 73; wherein the computer program is stored in the memory 71 and is configured to be executed by the processor 72 as described above in the mobile state electrode screening method.

[0069] In addition, this disclosure also provides a computer-readable storage medium having a computer program stored thereon, the computer program being executed by a processor to implement the electrode screening method in a mobile state as described in the above embodiments.

[0070] Furthermore, this disclosure also provides a computer program product, which includes a computer program or instructions that, when executed by a processor, implement the electrode screening method in a mobile state as described above.

[0071] It should be noted that the computer-readable medium described in this disclosure can be a computer-readable signal medium or a computer-readable storage medium, or any combination thereof. A computer-readable storage medium can be, for example,—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this disclosure, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In this disclosure, a computer-readable signal medium can include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium can be any computer-readable medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to: wires, optical fibers, RF (radio frequency), etc., or any suitable combination thereof.

[0072] In some implementations, clients and servers can communicate using any currently known or future-developed network protocol such as HTTP (Hypertext Transfer Protocol) and can interconnect with digital data communication (e.g., communication networks) of any form or medium. Examples of communication networks include local area networks (“LANs”), wide area networks (“WANs”), the Internet (e.g., the Internet of Things), and peer-to-peer networks (e.g., ad hoc peer-to-peer networks), as well as any currently known or future-developed networks.

[0073] The aforementioned computer-readable medium may be included in the aforementioned electronic device; or it may exist independently and not assembled into the electronic device.

[0074] The aforementioned computer-readable medium carries one or more programs that, when executed by the electronic device, cause the electronic device to: This application uses bandpass filters to precisely segment the target EEG signal into multiple preset frequency bands (such as delta and theta), and combines this with relative power calculation to achieve feature normalization, overcoming the shortcomings of traditional methods in protecting subtle EEG features. Secondly, statistical indicators (such as p-values) are calculated based on the relative power data of normal and abnormal users, and a dual verification mechanism using the coefficient of variation is introduced. Target electrodes are only confirmed when the statistical indicator is less than a first threshold and the coefficient of variation is less than a second threshold. This comprehensive screening criterion based on statistical significance and feature stability ensures that the electrodes possess both discriminative ability and robustness. This targeted screening mechanism significantly reduces electrode redundancy, focusing on electrodes that are truly effective for classifying brain injuries, thereby achieving high-precision classification while optimizing computational efficiency in mobile environments.

[0075] In addition, the electronic device can also perform other steps in the electrode screening method in the moving state as described above.

[0076] Computer program code for performing the operations of this disclosure can be written in one or more programming languages ​​or a combination thereof, including but not limited to object-oriented programming languages ​​such as Java, Smalltalk, and C++, as well as conventional procedural programming languages ​​such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0077] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0078] The units described in the embodiments of this disclosure can be implemented in software or hardware. The names of the units are not, in some cases, intended to limit the specific unit.

[0079] The functions described above in this document can be performed at least in part by one or more hardware logic components. For example, exemplary types of hardware logic components that can be used, without limitation, include: field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), system-on-a-chip (SoCs), complex programmable logic devices (CPLDs), and so on.

[0080] In the context of this disclosure, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can be, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0081] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0082] The above description is merely a specific embodiment of this disclosure, enabling those skilled in the art to understand or implement it. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this disclosure. Therefore, this disclosure is not to be limited to the embodiments described herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for electrode screening in a moving state, characterized in that, The method includes: Acquire target EEG signals of a user in motion state, the user including normal users and abnormal users, the target EEG signals are obtained through electrodes of a portable EEG device; The target EEG signal is divided by a bandpass filter to obtain target EEG signals in multiple preset frequency bands; For each of the preset frequency bands, the relative power of the target EEG signal is calculated; Based on the relative power of the normal users and the abnormal users in the preset frequency band, determine the statistical indicators corresponding to each preset frequency band. When the statistical index corresponding to any of the preset frequency bands is less than the first threshold, the corresponding electrode is determined to be the target electrode.

2. The method according to claim 1, characterized in that, The acquisition of the target EEG signal of the user in motion includes: Based on the user being in a mobile state, the user's original brainwave signals are recorded in real time through the electrodes of the portable EEG device, and the user's motion data is acquired through an auxiliary motion sensor; The original EEG signal is preprocessed based on the motion data to obtain the target EEG signal.

3. The method according to claim 1, characterized in that, The step of dividing the target EEG signal using a bandpass filter to obtain target EEG signals in multiple preset frequency bands includes: The bandpass filter removes DC offset below a first preset value and high-frequency noise above a second preset value from the target EEG signal. Based on preset boundary rules and preset frequency bands, the removed target EEG signals are divided to obtain target EEG signals in multiple preset frequency bands.

4. The method according to claim 1, characterized in that, The step of calculating the relative power of the target EEG signal for each of the preset frequency bands includes: For each of the preset frequency bands, a fast Fourier transform is performed on the target EEG signal to obtain the frequency domain coefficients of the target EEG signal; Based on the frequency domain coefficients, the power spectral density of the target EEG signal is calculated; The relative power of the preset frequency band is calculated based on the power spectral density and the band boundary of the preset frequency band.

5. The method according to claim 1, characterized in that, When the statistical index corresponding to any of the preset frequency bands is less than a first threshold, the corresponding electrode is determined to be the target electrode, including: Statistical tests are performed on the relative power characteristics of each electrode to calculate the relative power difference between the normal user and the abnormal user in the preset frequency band. Calculate the coefficient of variation of the relative power difference based on the standard deviation and mean of the relative power difference values; When the statistical index is less than a first threshold and the coefficient of variation is less than a second threshold, the electrode is determined to be the target electrode.

6. The method according to claim 5, characterized in that, The statistical tests include parametric tests, and the parametric tests include t-tests; Specifically, statistical tests are performed on the relative power characteristics of each electrode to calculate the relative power difference between the normal user and the abnormal user in the preset frequency band, including: For each of the electrodes, determine the number of paired samples for the relative power characteristics; Based on the number of paired samples, multiple pairs of differences in the relative power of the normal user and the abnormal user in the preset frequency band are calculated, and the number of paired samples and the multiple pairs of differences have a corresponding relationship. Calculate the mean and standard deviation of the multiple pairs of differences; The relative power difference value is calculated based on the mean, the standard deviation, and the number of paired samples.

7. The method according to claim 5, characterized in that, The statistical tests also include nonparametric tests, including the Wilcoxon rank-sum test. Specifically, statistical tests are performed on the relative power characteristics of each electrode to calculate the relative power difference between the normal user and the abnormal user in the preset frequency band, including: For each electrode, the relative power characteristics of the normal user and the abnormal user in the preset frequency band are mixed to obtain mixed data; The mixed data is sorted according to its size, and each mixed data item is assigned a rank. Calculate the first rank sum of the relative power of the normal user in the preset frequency band; Statistical values ​​and critical values ​​are calculated based on the first rank sum, the relative power characteristics of the normal user and the abnormal user in the preset frequency band; The relative power difference value is calculated based on the statistical value and the magnitude of the critical value.

8. An electrode screening device in a moving state, characterized in that, The device includes: The acquisition module is used to acquire the target EEG signal of a user in motion, the user including normal users and abnormal users, and the target EEG signal is obtained through electrodes of a portable EEG device. The segmentation module is used to segment the target EEG signal using a bandpass filter to obtain target EEG signals in multiple preset frequency bands. The calculation module is used to calculate the relative power of the target EEG signal for each of the preset frequency bands; The first determining module is used to determine the statistical indicators corresponding to each preset frequency band based on the relative power of the normal user and the abnormal user in the preset frequency band. The second determining module is used to determine the electrode as the target electrode when the statistical index corresponding to any of the preset frequency bands is less than the first threshold.

9. An electronic device, characterized in that, include: Memory; processor; as well as Computer programs; The computer program is stored in the memory and configured to be executed by the processor to implement the method as described in any one of claims 1-7.

10. 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 method as described in any one of claims 1-7.

Images