A method, device and equipment for dynamically sampling electroencephalogram signals
By identifying the abnormal fluctuation segments of EEG signals based on deep learning and dynamically adjusting the sampling frequency, the poor acquisition quality caused by single frequency sampling is solved, and more accurate EEG signal capture and resource optimization are achieved.
Patent Information
- Application Number
- CN202510695195.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2025-05-28
- Publication Date
- 2025-07-25
- Estimated Expiration
- 2045-05-28
AI Technical Summary
When obtaining EEG signals, the prior art cannot accurately capture the details of the rapid changes in the signal using a single sampling frequency, resulting in poor acquisition quality or waste of resources.
Through a convolutional neural network based on deep learning, the abnormal fluctuation segments in the EEG signal are identified, the EEG signal segments are divided, the fluctuation changes and similarity of adjacent extreme points are analyzed, and the sampling frequency is dynamically adjusted to match the fluctuation changes of the EEG signal.
It improves the accuracy of EEG signal acquisition, reduces oversampling when the signal is stable, and improves the acquisition quality.
Smart Images
Figure CN120203600B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of data processing, and particularly relates to a method, device and equipment for dynamically sampling electroencephalogram (EEG) signals. Background Art
[0002] In the field of neurosurgery, accurate acquisition of EEG signals is crucial for formulating subsequent disease treatment plans, ensuring that sufficient details are captured when the signals change rapidly, so that the acquired EEG signals can accurately reflect the patient's neural activities and provide strong support for clinical decision-making.
[0003] When the existing methods acquire EEG signals, a suitable sampling frequency is set to obtain the patient's EEG signals. Since the EEG activities of the patient vary at different times, if the EEG signals are acquired only at a single sampling frequency, the details captured when the signals change rapidly may not be accurate enough, or over-sampling may occur when the patient's EEG signals change relatively smoothly, resulting in poor quality of the acquired EEG signals or waste of more resources. Summary of the Invention
[0004] To solve the above problems, the present invention provides a method, device and equipment for dynamically sampling EEG signals.
[0005] The method, device and equipment for dynamically sampling EEG signals of the present invention adopt the following technical solutions:
[0006] An embodiment of the present invention provides a method for dynamically sampling EEG signals, which includes the following steps:
[0007] Obtain the patient's EEG signals and the initial sampling frequency of the EEG signals;
[0008] Divide the EEG signals to obtain several EEG signal segments; obtain the volatility of each EEG signal segment according to the fluctuation changes of adjacent extreme points in the EEG signal segment;
[0009] Obtain the similarity of volatility between EEG signal segments according to the change of volatility and the change of trend between EEG signal segments; calculate the similarity of volatility between EEG signal segments within different preset local ranges of the EEG signals, and use the cumulative value of the similarity of volatility as the consistency degree of volatility between the predicted EEG signal segment and the target EEG signal segment, and use the volatility of the EEG signal segment corresponding to the maximum value of the consistency degree of volatility as the volatility of the predicted EEG signal segment;
[0010] Take the difference between the volatility of the predicted EEG signal segment and the volatility of the last EEG signal segment of the EEG signals as the volatility change of the predicted EEG signal segment; adjust the initial sampling frequency of the EEG signals according to the volatility change of the predicted EEG signal segment to obtain the sampling frequency of the predicted EEG signal segment;
[0011] Obtain the predicted EEG signal segment according to the sampling frequency of the predicted EEG signal segment, and use the predicted EEG signal segment as the newly acquired EEG signal.
[0012] Further, the steps of dividing the EEG signal into several EEG signal segments are as follows:
[0013] Identify the abnormal fluctuation EEG signal segments in the EEG signal through a convolutional neural network based on deep learning, obtain the length mean of all abnormal fluctuation EEG signal segments, and equally divide the EEG signal except the abnormal fluctuation EEG signal segments according to the length mean to obtain several EEG signal segments, and regard each abnormal fluctuation EEG signal segment as an EEG signal segment.
[0014] Further, the steps of obtaining the volatility of each EEG signal segment according to the fluctuation change of adjacent extreme points in the EEG signal segment are as follows:
[0015] For any EEG signal segment; take the mean of the amplitude differences of all adjacent extreme points in the EEG signal segment as the first fluctuation change of the EEG signal segment; take the variance of the amplitude differences of all adjacent extreme points in the EEG signal segment as the second fluctuation change of the EEG signal segment; take the variance of the time intervals of all adjacent extreme points in the EEG signal segment as the third fluctuation change of the EEG signal segment;
[0016] Obtain the volatility of the EEG signal segment according to the first fluctuation change, the second fluctuation change and the third fluctuation change.
[0017] Further, the steps of obtaining the volatility of the EEG signal segment according to the first fluctuation change, the second fluctuation change and the third fluctuation change are as follows:
[0018] Fuse the first fluctuation change, the second fluctuation change and the third fluctuation change to obtain the volatility of the EEG signal segment, and the volatility is a normalized value.
[0019] Further, the steps of calculating the similarity of the volatility between EEG signal segments in different preset local ranges of the EEG signal are as follows:
[0020] Denote the local range composed of the last EEG signal segments of the EEG signal as the first local range, which is a preset first numerical value; denote any EEG signal segment in the EEG signal as the target EEG signal segment; in the EEG signal, denote the local range composed of the EEG signal segments before the target EEG signal segment as the second local range;
[0021] Calculate the similarity of the volatility between all EEG signal segments of the first local range and the second local range in the same order.
[0022] Furthermore, the specific steps for obtaining the similarity of the volatility between EEG signal segments according to the change and trend change of the volatility between EEG signal segments are as follows:
[0023] For any two EEG signal segments; take the difference between the volatility of the first EEG signal segment and the volatility of the second EEG signal segment as the change of the volatility between EEG signal segments; take the DTW distance between the first EEG signal segment and the second EEG signal segment as the trend change between EEG signal segments; fuse the change of the volatility and the trend change to obtain the similarity of the volatility between EEG signal segments.
[0024] Furthermore, the specific method for adjusting the initial sampling frequency of the EEG signal by predicting the volatility change of the EEG signal segment is as follows:
[0025] If the predicted volatility of the EEG signal segment increases compared with the volatility of the last EEG signal segment of the EEG signal, increase the initial sampling frequency; if the predicted volatility of the EEG signal segment decreases compared with the volatility of the last EEG signal segment of the EEG signal, decrease the initial sampling frequency; if the predicted volatility of the EEG signal segment remains unchanged compared with the volatility of the last EEG signal segment of the EEG signal, keep the initial sampling frequency unchanged.
[0026] The present invention also proposes an EEG signal dynamic sampling device, including: an electrode cap, a data processor, and a sampling frequency adjustment device; the electrode cap is used to acquire the EEG signal of the patient; the data processor is used to implement the above steps; the sampling frequency adjustment device is used to acquire the initial sampling frequency of the EEG signal and adjust the initial sampling frequency.
[0027] The present invention also proposes an EEG signal dynamic sampling device, including: a data acquisition module, a data processing module, and an EEG signal acquisition module; the data acquisition module is used to acquire the EEG signal of the patient and the initial sampling frequency of the EEG signal; the data processing module is used to divide the EEG signal to obtain a plurality of EEG signal segments; obtain the volatility of each EEG signal segment according to the fluctuation change of adjacent extreme points in the EEG signal segment; obtain the similarity of the volatility between EEG signal segments according to the change and trend change of the volatility between EEG signal segments; obtain the predicted volatility of the EEG signal segment according to the similarity of the volatility between EEG signal segments in different preset local ranges of the EEG signal; adjust the initial sampling frequency according to the predicted volatility change of the EEG signal segment to obtain the sampling frequency of the predicted EEG signal segment; the EEG signal acquisition module is used to acquire the predicted EEG signal segment according to the sampling frequency of the predicted EEG signal segment and use the predicted EEG signal segment as the newly acquired EEG signal.
[0028] The beneficial effects of the technical solution of the present invention are as follows: after obtaining the EEG signal of the patient and the initial sampling frequency of the EEG signal, the present invention obtains the volatility of each EEG signal segment through the fluctuation change of adjacent extreme points in the EEG signal segment, so as to better analyze the similarity of volatility between EEG signal segments and determine the volatility of the predicted EEG signal segment; by the change of volatility and the change of trend between EEG signal segments, the similarity of volatility between EEG signal segments is obtained, and the influence of similarity on the volatility of the predicted EEG signal segment is reduced; furthermore, the volatility of the predicted EEG signal segment is obtained through the similarity of volatility between EEG signal segments in different preset local ranges of the EEG signal. By analyzing the similarity of volatility between EEG signal segments in different preset local ranges, the volatility of the predicted EEG signal segment is made more accurate, which is convenient for better dynamically adjusting the sampling frequency; finally, the initial sampling frequency is adjusted according to the change of volatility of the predicted EEG signal segment to obtain the sampling frequency of the predicted EEG signal segment, and then the newly collected EEG signal is obtained. By monitoring and adjusting the initial sampling frequency of the EEG signal, the details captured when the signal changes rapidly are more accurate, the problem of oversampling when the EEG signal changes relatively smoothly is reduced, and the acquisition quality of the EEG signal is improved. BRIEF DESCRIPTION OF THE DRAWINGS
[0029] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained according to these drawings.
[0030] Figure 1 It is a flowchart of the steps of a method for dynamically sampling EEG signals provided by an embodiment of the present invention;
[0031] Figure 2 It is a flowchart for obtaining a newly collected EEG signal provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS
[0032] In order to further elaborate on the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following, in combination with the accompanying drawings and preferred embodiments, details the specific embodiments, structures, features and effects of a method, device and equipment for dynamically sampling EEG signals proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, the specific features, structures or characteristics in one or more embodiments can be combined in any suitable form.
[0033] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the technical field to which this invention belongs.
[0034] The following specifically describes the specific solutions of a method, device, and equipment for dynamically sampling electroencephalogram (EEG) signals provided by the present invention with reference to the accompanying drawings.
[0035] Please refer to Figure 1 and Figure 2 , which shows the flowchart of the steps of a method for dynamically sampling EEG signals and the flowchart for obtaining newly acquired EEG signals provided by an embodiment of the present invention. The method includes the following steps:
[0036] Step S001: Obtain the EEG signal of the patient and the initial sampling frequency of the EEG signal.
[0037] It should be noted that the main purpose of this embodiment is to obtain the volatility of the predicted EEG signal segment by the similarity of the volatility between EEG signal segments in different preset local ranges of the EEG signal, and then adjust the initial sampling frequency according to the volatility of the predicted EEG signal segment to obtain the sampling frequency of the predicted EEG signal segment, and finally obtain the predicted EEG signal segment through the sampling frequency of the predicted EEG signal segment; before starting the analysis, the required data is collected first.
[0038] Specifically, to obtain the EEG signal of the patient and the initial sampling frequency of the EEG signal, it should be noted that obtaining the EEG signal of the patient is an existing method. By wearing an electrode cap on the patient, the electrodes in the electrode cap are distributed at different predetermined positions on the scalp to capture the electrical activities of different regions of the brain, thereby obtaining the EEG signal of the patient, which will not be elaborated here; in this embodiment, the initial sampling frequency of the EEG signal is 512 Hz, which is a commonly used sampling frequency, and can capture most EEG activities while retaining certain high-frequency components.
[0039] Thus, the EEG signal of the patient and the initial sampling frequency of the EEG signal are obtained.
[0040] Step S002: Divide the EEG signal to obtain several EEG signal segments; obtain the volatility of each EEG signal segment according to the fluctuation changes of adjacent extreme points in the EEG signal segment.
[0041] It should be noted that the above-mentioned EEG signal is obtained. By analyzing the local fluctuation characteristics of the EEG signal, the sampling frequency of the subsequently collected EEG signal can be adjusted more dynamically. Therefore, it is first necessary to segment the EEG signal so as to analyze the local fluctuation characteristics of the EEG signal through the EEG signal segments later, thereby better adjusting the sampling frequency of the EEG signal.
[0042] Specifically, the EEG signal is divided to obtain several EEG signal segments as follows:
[0043] Identify abnormal fluctuation EEG signal segments in EEG signals through a convolutional neural network based on deep learning, obtain the length mean of all abnormal fluctuation EEG signal segments, and equally divide the EEG signals other than the abnormal fluctuation EEG signal segments in the EEG signals according to the length mean to obtain several EEG signal segments, and regard each abnormal fluctuation EEG signal segment as an EEG signal segment.
[0044] It should be noted that identifying abnormal fluctuation EEG signal segments in EEG signals through a convolutional neural network based on deep learning is an existing method. The specific network model in this embodiment is a Recursive Neural Network (RNN), which is a well-known technology and will not be elaborated here. When equally dividing the EEG signals other than the abnormal fluctuation EEG signal segments, if the remaining signal length after dividing a part of the signals is less than the length mean of all abnormal fluctuation EEG signal segments, it is directly regarded as an EEG signal segment.
[0045] It should be noted that several EEG signal segments are obtained above. The volatility within an EEG signal segment can reflect the information complexity within the segment signal. The greater the information complexity, the more observable the EEG signal segment is and the more information it contains. Furthermore, the sampling frequency of the predicted EEG signal to be obtained can be adjusted according to the volatility. The predicted EEG signal is a new EEG signal. Therefore, for the convenience of subsequent analysis, it is first necessary to obtain the volatility of each EEG signal segment. Since the volatility of an EEG signal segment is mainly reflected by the fluctuation changes between adjacent extreme points in the EEG signal segment, the volatility of each EEG signal segment is obtained by analyzing the fluctuation changes between adjacent extreme points in the EEG signal segment.
[0046] Preferably, the volatility of each EEG signal segment is obtained according to the fluctuation changes between adjacent extreme points in the EEG signal segment, specifically as follows:
[0047] For any EEG signal segment; take the mean of the amplitude differences between all adjacent extreme points in the EEG signal segment as the first fluctuation change of the EEG signal segment; take the variance of the amplitude differences between all adjacent extreme points in the EEG signal segment as the second fluctuation change of the EEG signal segment; take the variance of the time intervals between all adjacent extreme points in the EEG signal segment as the third fluctuation change of the EEG signal segment; fuse the first fluctuation change, the second fluctuation change, and the third fluctuation change to obtain the volatility of the EEG signal segment, and the volatility is a normalized value.
[0048] As a specific example, the specific method for obtaining the volatility of the EEG signal segment is as follows:
[0049]
[0050] In the formula, is the mean value of the amplitude differences between all adjacent extreme points in the th EEG signal segment, and the amplitude difference is specifically: the absolute value of the difference between the amplitudes of adjacent extreme points; is the variance of the amplitude differences between all adjacent extreme points in the th EEG signal segment; is the variance of the time intervals between all adjacent extreme points in the th EEG signal segment; is the hyperbolic tangent function for normalization; is the volatility of the th EEG signal segment.
[0051] It should be noted that since the fluctuation of the EEG signal segment is mainly related to the fluctuation change of its adjacent extreme points; when the mean value of the amplitude differences between all adjacent extreme points in the EEG signal segment is large, it indicates that the overall extreme value change of the EEG signal segment is large, corresponding to a large volatility of the EEG signal segment. Therefore, by analyzing the mean value of the amplitude differences between all adjacent extreme points in the th EEG signal segment, that is, the first fluctuation change, when the mean value is larger, the volatility of the th EEG signal segment is also larger; when the variance of the amplitude differences between all adjacent extreme points in the th EEG signal segment is larger, that is, the second fluctuation change is larger, it indicates that the th EEG signal segment changes significantly in the amplitude dimension, and the volatility of the th EEG signal segment is larger; when the variance of the time intervals between all adjacent extreme points in the th EEG signal segment is larger, that is, the third fluctuation change is larger, it indicates that the th EEG signal segment changes significantly in the time interval dimension, and the volatility of the th EEG signal segment is also larger. By combining the first fluctuation change, the second fluctuation change, and the third fluctuation change, the volatility of the th EEG signal segment is obtained, making the quantification of its volatility more accurate.
[0052] So far, the volatility of each EEG signal segment is obtained.
[0053] Step S003: Obtain the similarity of the volatility between EEG signal segments according to the change and trend change of the volatility between EEG signal segments; obtain the volatility of the predicted EEG signal segment according to the similarity of the volatility between EEG signal segments within different preset local ranges of the EEG signals.
[0054] It should be noted that the volatility of each EEG signal segment is obtained above. Since the change of the EEG signal will last for a relatively long time and there are certain similarities in the volatility change of the EEG signal segments, in this embodiment, the volatility of the predicted EEG signal segment is approximately obtained through the similarity of the volatility change of the existing EEG signal segments, so as to adjust the sampling frequency of the predicted EEG signal by the volatility of the predicted EEG signal segment. Therefore, for the convenience of subsequent analysis, the similarity of the volatility between the EEG signal segments is obtained first. The similarity of the volatility between the EEG signal segments is mainly related to the change of the volatility and the trend change between the EEG signal segments. When the volatility changes between the EEG signal segments are similar and the trend changes are also similar, it indicates that the volatility between the EEG signal segments is more similar. Therefore, by analyzing the change of the volatility and the trend change between the EEG signal segments, the similarity of the volatility between the EEG signal segments is obtained.
[0055] Preferably, according to the change of the volatility and the trend change between the EEG signal segments, the similarity of the volatility between the EEG signal segments is obtained as follows:
[0056] For any two EEG signal segments, the difference between the volatility of the first EEG signal segment and the volatility of the second EEG signal segment is used as the change of the volatility between the EEG signal segments; the DTW distance between the first EEG signal segment and the second EEG signal segment is used as the trend change between the EEG signal segments; the change of the volatility and the trend change are fused to obtain the similarity of the volatility between the EEG signal segments; the change of the volatility, the trend change and the similarity are in an inverse proportion relationship.
[0057] As a specific example, the specific method for obtaining the similarity is as follows:
[0058]
[0059] In the formula, is the volatility of the th EEG signal segment; is the volatility of the th EEG signal segment; is the DTW distance between the th EEG signal segment and the th EEG signal segment; is the exponential function with the natural constant as the base. In this embodiment, the model of is used to present the inverse proportion relationship and normalization processing. is the input of the model. In specific implementation, it can be set as other inverse proportion functions and normalization functions; is the th EEG signal segment and the th EEG signal segment volatility similarity.
[0060] It should be noted that since the similarity of the fluctuations between EEG signal segments is mainly related to the changes and trend changes of the fluctuations between EEG signal segments; therefore, when the fluctuations of the th EEG signal segment and the fluctuations of the th EEG signal segment change less, that is, the smaller, it indicates that the fluctuations of the th EEG signal segment and the th EEG signal segment are relatively similar, and the similarity of the fluctuations between the th EEG signal segment and the th EEG signal segment is higher; when the DTW distance between the th EEG signal segment and the th EEG signal segment is smaller, it indicates that the trend changes of the th EEG signal segment and the th EEG signal segment are relatively similar, and the similarity of the fluctuations between the
[0061] th EEG signal segment and the th EEG signal segment is also higher. Finally, by combining the changes and trend changes of the fluctuations between EEG signal segments, the similarity of the fluctuations between EEG signal segments is obtained.
[0061] It should be noted that the above-mentioned similarity of the fluctuations between EEG signal segments is obtained. Since the changes of EEG signals will continue for a relatively long time, and there are certain similarities in the fluctuations of EEG signal segments; therefore, the similarity of the fluctuations between EEG signal segments in different local ranges of existing EEG signals is used to approximately obtain the fluctuations of the predicted EEG signal segment, so as to adjust the sampling frequency of the predicted EEG signal by predicting the fluctuations of the EEG signal segment in the subsequent stage.
[0062] Specifically, according to the similarity of the fluctuations between EEG signal segments in different preset local ranges of EEG signals, the fluctuations of the predicted EEG signal segment are obtained as follows:
[0063] The local range composed of the last EEG signal segments of the EEG signal is denoted as the first local range. is a preset first value, and in this embodiment, is used for description; any EEG signal segment in the EEG signal is denoted as the target EEG signal segment; in the EEG signal, the local range composed of the EEG signal segments before the target EEG signal segment is denoted as the second local range; the cumulative value of the similarities of the fluctuations between all the EEG signal segments in the same order of the first local range and the second local range is used as the degree of consistency of the fluctuations between the predicted EEG signal segment and the target EEG signal segment.
[0064] Obtain the degree of consistency in volatility between the predicted EEG signal segment and each EEG signal segment; use the volatility of the EEG signal segment corresponding to the maximum degree of consistency as the volatility of the predicted EEG signal segment.
[0065] It should be noted that since the change of EEG signals lasts for a relatively long time and there are certain similarities in the volatility changes of EEG signal segments; therefore, in this embodiment, the volatility of the predicted EEG signal segment is approximately obtained by the similarity in volatility between EEG signal segments within different local ranges of existing EEG signals; the above first local range is equivalent to the EEG signals for a period of time before the predicted EEG signal segment, and the second local range is equivalent to the EEG signals for another period of time in the EEG signals. If the similarity in volatility between the EEG signal segments of the same order within the first local range and the second local range is greater, it indicates that the volatility changes of the EEG signals for a period of time before the predicted EEG signal segment and the EEG signals for a period of time of the target EEG signal segment are more consistent, that is, the degree of consistency is greater, then the volatility of the predicted EEG signal segment is more likely to be close to the volatility of the target EEG signal segment. Therefore, by selecting the volatility of the EEG signal segment corresponding to the maximum degree of consistency as the volatility of the predicted EEG signal segment.
[0066] So far, the volatility of the predicted EEG signal segment is obtained.
[0067] Step S004: Adjust the initial sampling frequency according to the volatility change of the predicted EEG signal segment to obtain the sampling frequency of the predicted EEG signal segment; obtain the predicted EEG signal segment according to the sampling frequency of the predicted EEG signal segment, and use the predicted EEG signal segment as the newly acquired EEG signal.
[0068] It should be noted that the volatility of the predicted EEG signal segment is obtained above. The initial sampling frequency is adjusted according to the volatility change of the predicted EEG signal segment compared with the previous EEG signal segment to obtain the sampling frequency of the predicted EEG signal segment; if the volatility of the predicted EEG signal segment increases compared with the volatility of the last EEG signal segment of the EEG signal, the sampling frequency needs to be increased to ensure capturing all the fine signal features of rapid changes; if the volatility of the predicted EEG signal segment decreases compared with the volatility of the last EEG signal segment of the EEG signal, the sampling frequency needs to be decreased to save resources.
[0069] Preferably, adjusting the initial sampling frequency according to the volatility change of the predicted EEG signal segment to obtain the sampling frequency of the predicted EEG signal segment is specifically as follows:
[0070] Take the difference between the volatility of the predicted EEG signal segment and the volatility of the last EEG signal segment of the EEG signal as the volatility change of the predicted EEG signal segment; adjust the initial sampling frequency of the EEG signal through the volatility change of the predicted EEG signal segment to obtain the sampling frequency of the predicted EEG signal segment.
[0071] As a specific example, the specific method for obtaining the sampling frequency of the predicted EEG signal segment is as follows:
[0072]
[0073] In the formula, is the volatility of the predicted EEG signal segment; is the volatility of the last EEG signal segment of the EEG signal; is the initial sampling frequency; is the sampling frequency of the predicted EEG signal segment.
[0074] It should be noted that if the volatility of the predicted EEG signal segment increases compared with the volatility of the last EEG signal segment of the EEG signal, the sampling frequency needs to be increased to ensure that all fast-changing subtle signal features are captured. Therefore is greater than 0, then through increase the initial sampling frequency to obtain the sampling frequency of the predicted EEG signal segment; if the volatility of the predicted EEG signal segment decreases compared with the volatility of the last EEG signal segment of the EEG signal, the sampling frequency needs to be reduced to save resources. Therefore is less than 0, then through reduce the initial sampling frequency to obtain the sampling frequency of the predicted EEG signal segment. If , in this embodiment, the initial sampling frequency is directly used as the sampling frequency of the predicted EEG signal segment.
[0075] It should be noted that the sampling frequency of the predicted EEG signal segment is obtained above, and a new EEG signal can be obtained through the obtained sampling frequency.
[0076] Furthermore, the predicted EEG signal segment is obtained according to the sampling frequency of the predicted EEG signal segment, and the predicted EEG signal segment is used as the newly acquired EEG signal. It should be noted that the sampling frequency of the predicted EEG signal segment is obtained above. Through this sampling frequency, the EEG signal in a future period of time, that is, the predicted EEG signal segment, is obtained. In this embodiment, the time series length of the predicted EEG signal segment is the average value of the lengths of all abnormally fluctuating EEG signal segments. Subsequently, the sampling frequency of other predicted EEG signal segments can be dynamically adjusted by the same method as above, and then other newly acquired EEG signals can be obtained, which will not be elaborated here.
[0077] Through the above steps, a dynamic EEG signal sampling method is completed.
[0078] An embodiment of the present invention further provides an electroencephalogram (EEG) signal dynamic sampling device, including: an electrode cap, a data processor, and a sampling frequency adjustment device; the electrode cap is used to acquire the EEG signal of a patient; the data processor is used to implement the above steps; the sampling frequency adjustment device is used to acquire the initial sampling frequency of the EEG signal and adjust the initial sampling frequency.
[0079] An embodiment of the present invention further provides an EEG signal dynamic sampling device, including: a data acquisition module, a data processing module, and an EEG signal acquisition module; the data acquisition module is used to acquire the EEG signal of a patient and the initial sampling frequency of the EEG signal; the data processing module is used to divide the EEG signal to obtain a plurality of EEG signal segments; according to the fluctuation change of adjacent extreme points in the EEG signal segment, the volatility of each EEG signal segment is obtained; according to the change of volatility and the trend change between EEG signal segments, the similarity of volatility between EEG signal segments is obtained; according to the similarity of volatility between EEG signal segments within different preset local ranges of the EEG signal, the volatility of the predicted EEG signal segment is obtained; according to the change of volatility of the predicted EEG signal segment, the initial sampling frequency is adjusted to obtain the sampling frequency of the predicted EEG signal segment; the EEG signal acquisition module is used to acquire the predicted EEG signal segment according to the sampling frequency of the predicted EEG signal segment and use the predicted EEG signal segment as the newly acquired EEG signal.
[0080] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the principle of the present invention shall be included in the protection scope of the present invention.
Claims
1. A method for dynamically sampling electroencephalogram signals, characterized in that, The method includes the following steps: Obtain the electroencephalogram (EEG) signal of a patient and the initial sampling frequency of the EEG signal; Divide the EEG signal to obtain a number of EEG signal segments; according to the fluctuation changes of adjacent extreme points in the EEG signal segments, obtain the volatility of each EEG signal segment; According to the change in volatility and the change in trend between EEG signal segments, obtain the similarity of volatility between EEG signal segments; calculate the similarity of volatility between EEG signal segments within different preset local ranges of the EEG signal, and take the cumulative value of the similarity of volatility as the degree of consistency of volatility between the predicted EEG signal segment and the target EEG signal segment, and take the volatility of the EEG signal segment corresponding to the maximum value of the degree of consistency of volatility as the volatility of the predicted EEG signal segment; Among them, the method for obtaining the similarity of volatility between EEG signal segments is as follows: for any two EEG signal segments; take the difference between the volatility of the first EEG signal segment and the volatility of the second EEG signal segment as the change in volatility between EEG signal segments; take the DTW distance between the first EEG signal segment and the second EEG signal segment as the change in trend between EEG signal segments; fuse the change in volatility and the change in trend to obtain the similarity of volatility between EEG signal segments; Take the difference between the volatility of the predicted EEG signal segment and the volatility of the last EEG signal segment of the EEG signal as the change in volatility of the predicted EEG signal segment; adjust the initial sampling frequency of the EEG signal through the change in volatility of the predicted EEG signal segment to obtain the sampling frequency of the predicted EEG signal segment; Obtain the predicted EEG signal segment according to the sampling frequency of the predicted EEG signal segment, and take the predicted EEG signal segment as the newly collected EEG signal.
2. The method for dynamically sampling EEG signals according to claim 1, wherein The step of dividing the EEG signal to obtain a number of EEG signal segments includes the following specific steps: Identify the abnormal fluctuation EEG signal segments in the EEG signal through a convolutional neural network based on deep learning, obtain the length mean of all abnormal fluctuation EEG signal segments, and evenly divide the EEG signal except the abnormal fluctuation EEG signal segments according to the length mean to obtain a number of EEG signal segments, and take each abnormal fluctuation EEG signal segment as an EEG signal segment.
3. The method for dynamically sampling electroencephalogram signals according to claim 1, characterized in that The step of obtaining the volatility of each EEG signal segment according to the fluctuation changes of adjacent extreme points in the EEG signal segment includes the following specific steps: For any one EEG signal segment; take the mean of the amplitude differences of all adjacent extreme points in the EEG signal segment as the first fluctuation change of the EEG signal segment; take the variance of the amplitude differences of all adjacent extreme points in the EEG signal segment as the second fluctuation change of the EEG signal segment; take the variance of the time intervals of all adjacent extreme points in the EEG signal segment as the third fluctuation change of the EEG signal segment; Obtain the volatility of the EEG signal segment according to the first fluctuation change, the second fluctuation change and the third fluctuation change.
4. The method for dynamically sampling electroencephalogram signals according to claim 3, wherein The step of obtaining the volatility of the EEG signal segment according to the first fluctuation change, the second fluctuation change and the third fluctuation change includes the following specific steps: Fuse the first fluctuation change, the second fluctuation change and the third fluctuation change to obtain the volatility of the EEG signal segment, and the volatility is a normalized value.
5. The method for dynamically sampling electroencephalogram signals according to claim 1, wherein Calculating the similarity of the volatility between EEG signal segments within different preset local ranges of the EEG signal includes the following specific steps: The local range formed by the last EEG signal segment is denoted as the first local range, which is a preset first value; any EEG signal segment in the EEG signal is denoted as the target EEG signal segment; In the electroencephalogram (EEG) signal, the local range composed of EEG signal segments before the target EEG signal segment is denoted as the second local range; Calculating the similarity of the volatility between EEG signal segments of the same order in the first local range and the second local range.
6. The method for dynamically sampling EEG signals according to claim 1, characterized in that, The specific method for adjusting the initial sampling frequency of the EEG signal by predicting the volatility change of the EEG signal segment is as follows: If the volatility of the predicted EEG signal segment increases compared with the volatility of the last EEG signal segment of the EEG signal, increase the initial sampling frequency; if the volatility of the predicted EEG signal segment decreases compared with the volatility of the last EEG signal segment of the EEG signal, decrease the initial sampling frequency; if the volatility of the predicted EEG signal segment remains unchanged compared with the volatility of the last EEG signal segment of the EEG signal, keep the initial sampling frequency unchanged.
7. A dynamic electroencephalogram signal sampling device, comprising: An electrode cap, a data processor, and a sampling frequency adjustment device, characterized in that the electrode cap is used to acquire the EEG signal of a patient; the data processor is used to implement the steps of the EEG signal dynamic sampling method described in any one of claims 1 to 6.
8. The electroencephalogram signal dynamic sampling device according to claim 7, wherein The sampling frequency adjustment device is used to acquire the initial sampling frequency of the EEG signal and adjust the initial sampling frequency.
9. A dynamic electroencephalogram signal sampling device, comprising: A data acquisition module, a data processing module, and an EEG signal acquisition module, characterized in that the data acquisition module is used to acquire the EEG signal of a patient and the initial sampling frequency of the EEG signal; the data processing module is used to divide the EEG signal to obtain a plurality of EEG signal segments; according to the fluctuation change of adjacent extreme points in the EEG signal segment, obtain the volatility of each EEG signal segment; according to the change and trend change of the volatility between EEG signal segments, obtain the similarity of the volatility between EEG signal segments; wherein, the method for obtaining the similarity of the volatility between EEG signal segments is: for any two EEG signal segments; taking the difference between the volatility of the first EEG signal segment and the volatility of the second EEG signal segment as the change of the volatility between EEG signal segments; taking the DTW distance between the first EEG signal segment and the second EEG signal segment as the trend change between EEG signal segments; fusing the change of the volatility and the trend change to obtain the similarity of the volatility between EEG signal segments; according to the similarity of the volatility between EEG signal segments within different preset local ranges of the EEG signal, obtain the volatility of the predicted EEG signal segment; adjusting the initial sampling frequency according to the volatility change of the predicted EEG signal segment to obtain the sampling frequency of the predicted EEG signal segment; the EEG signal acquisition module is used to obtain the predicted EEG signal segment according to the sampling frequency of the predicted EEG signal segment, and take the predicted EEG signal segment as the newly acquired EEG signal.
Citation Information
Patent Citations
Multi-scale quantitative analysis method of electroencephalogram signal
CN106821318A
CNV electroencephalogram lie detection method based on multifractal detrended fluctuation analysis
CN108498106A