An EEG signal processing system for sleep analysis

Abstract

The present invention relates to the field of data processing technology, and in particular to an electroencephalogram (EEG) signal processing system for sleep analysis. First, EEG signals from different areas of the brain are obtained, and the initial interference degree is obtained according to the amplitude change characteristics of the EEG signals within a preset window; the local correlation is obtained according to the difference characteristics of the interference degree at corresponding moments in different areas. The final interference degree and the preliminary pseudo-electrooculogram (EOG) signal are obtained according to the local correlation of different areas and the EEG signal amplitude difference. The removal index is obtained according to the change characteristics of the final interference degree after removing the preliminary pseudo-electrooculogram (EOG) signal, and the final pseudo-electrooculogram (EOG) signal can be adaptively obtained according to the final interference index and the removal index, so that the final pseudo-electrooculogram (EOG) signal value is more accurate, thereby improving the accuracy of obtaining new EEG signals, and can more accurately assist doctors in analyzing sleep conditions through new EEG data.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular to an electroencephalogram signal processing system for sleep analysis. Background Art

[0002] EEG signals are a highly random physiological signal, which is the result of the superposition and amplification of the point activities of a large number of nerve cells in the brain. Currently, the acquisition of EEG signals often uses a non-invasive acquisition method, that is, placing electrodes on the scalp to measure the EEG signals transmitted from the skull to the scalp. However, the EEG signals collected by this method have certain disadvantages, such as low spatial resolution, susceptibility to interference from external noise, and low amplitude of the obtained EEG signals.

[0003] In the process of collecting EEG signals by this method, all eye movements, including eye movements and the electrical signals generated by the physiological behavior of blinking, are collectively referred to as electrooculogram signals. Electrooculogram signals are one of the main artifacts that affect the quality of EEG signals, causing the collected EEG signals to be interfered by noise. When the IMM algorithm is used to automatically and quickly remove electrooculogram artifacts, if the reference electrooculogram signal of the IMM is poor, some large-value EEG signals will be removed, resulting in information loss in the obtained EEG signals, affecting the accuracy of the EEG signals. Summary of the invention

[0004] In order to solve the technical problem that when the above-mentioned traditional IMM algorithm removes the electrooculogram artifacts in the EEG signal, part of the EEG signal is easily eliminated as noise, which affects the accuracy of the EEG signal, the purpose of the present invention is to provide an EEG signal processing system for sleep analysis, and the technical solution adopted is as follows:

[0005] Data acquisition module, used to obtain EEG signals from different brain regions;

[0006] A data processing module is used to obtain the initial interference degree of each area at the target time according to the difference characteristics of the amplitude of the EEG signal in the preset window in the EEG signal of the target area, with the target time as the center of the preset window; and obtain the local correlation at the target time according to the difference characteristics of the initial interference degree between the target area and other areas at the corresponding time in the preset window;

[0007] A feature analysis module is used to obtain the final interference degree at the target moment according to the local correlation at the target moment, the difference characteristics of the EEG signal amplitude between the target area and other areas, and the initial interference degree of the target area; and to obtain a preliminary pseudo-eye signal according to the EEG signal at the target moment and the final interference degree;

[0008] The interference removal module is used to obtain the change characteristics of the final interference degree after removing the preliminary pseudo electrooculogram signal; obtain a removal index according to the change characteristics, iteratively update the preliminary pseudo electrooculogram signal according to the size of the removal index to obtain a final pseudo electrooculogram signal; and remove electrooculogram artifacts according to the final pseudo electrooculogram signal to obtain a new electroencephalogram signal.

[0009] Furthermore, the step of obtaining the initial interference degree includes:

[0010] In the EEG signal, taking the target moment as the center of the preset window, the variance of the EEG signal amplitude, the number of amplitude changes and the range of the amplitude in the preset window are calculated; the product of the variance, the number of changes and the range is calculated to obtain the initial interference degree at the target moment.

[0011] Furthermore, the step of acquiring the local correlation includes:

[0012] In a preset window at the target time, the absolute value of the difference between the initial interference degree of the target area and other areas at the corresponding time is calculated and negative correlation mapping is performed to obtain the local correlation value, and the average value of the local correlation value is calculated to obtain the local correlation at the target time.

[0013] Furthermore, the step of obtaining the final interference degree includes:

[0014] At the target moment, the absolute value of the difference between the EEG signals of the target area and other areas is calculated as the signal difference value, and the product of the signal difference and the corresponding local correlation is calculated to obtain the signal correlation difference; the signal correlation differences corresponding to all other areas are accumulated to obtain the signal correlation difference feature;

[0015] Calculate the cumulative value of the absolute value of the difference of the local correlations between other regions to obtain the local correlation difference; calculate the product of the inverse of the local correlation difference and the corresponding signal correlation difference feature and negatively correlate and map them to obtain a first value, calculate the difference between the value 1 and the first value, and obtain a second value;

[0016] The product of the initial interference degree and the corresponding second value is calculated to obtain the interference degree adjustment value, and the sum of the initial interference degree and the corresponding interference degree adjustment value is calculated to obtain the final interference degree.

[0017] Furthermore, the step of obtaining the preliminary pseudo electrooculogram signal includes:

[0018] The product of the final interference degree and the EEG data at the corresponding moment is calculated to obtain the preliminary pseudo-egoculoculographic signal.

[0019] Furthermore, the step of obtaining the removal index includes:

[0020] Calculate the final interference degree of the EEG signal after removing the preliminary pseudo-oculogram signal, as the final interference degree iteration value, calculate the absolute value of the difference between the final interference degree of the EEG signal and the corresponding final interference degree iteration value, as the iteration characterization value; calculate the reciprocal of the final interference degree iteration value and the product of the iteration characterization value, and negatively map them as the removal feature, calculate the difference between the value one and the removal feature, and obtain the removal index.

[0021] Furthermore, the step of obtaining the final pseudo electrooculogram signal includes:

[0022] When the removal index is greater than or equal to a preset removal index threshold, the preliminary pseudo electrooculogram signal is used as the final pseudo electrooculogram signal;

[0023] When the removal index is less than a preset removal index threshold, the sum of the removal index and the preset second value is calculated as the final interference degree secondary iteration value, the product of the final interference degree secondary iteration value and the preliminary pseudo-oculogram signal is calculated to obtain the pseudo-oculogram signal iteration value; the removal index after the pseudo-oculogram signal iteration value is removed from the EEG signal is calculated;

[0024] Until the removal index is greater than or equal to the preset removal index threshold, the final pseudo electrooculogram signal iteration value is output as the final pseudo electrooculogram signal.

[0025] Furthermore, the step of acquiring the new EEG signal includes:

[0026] The final pseudo electrooculogram signal is used as reference electrooculogram data, and the amplitude of the electroencephalogram signal is adjusted through the IMM algorithm and the reference electrooculogram data to complete the reconstruction of the electroencephalogram signal and obtain the new electroencephalogram signal.

[0027] The present invention has the following beneficial effects:

[0028] In an embodiment of the present invention, the initial interference degree can characterize the interference degree of the collected EEG signal by the EO signal, but because the initial interference degree cannot completely and accurately reflect the interference degree of the EO signal, local correlation is obtained based on the propagation characteristics of the EO signal and the difference in the initial interference degree of different brain regions; the local correlation can characterize the interference degree of the EO signal in different regions, and can correct the final interference degree to a certain extent. The final interference degree is obtained according to the local correlation, the amplitude difference characteristics and the initial interference degree, and the interference degree of the EO signal can be accurately characterized in combination with the influence of the EO signal in different regions, so as to improve the accuracy of the subsequent calculation of the pseudo EO signal. The removal index can reflect the effect of removing the pseudo EO signal from the EEG signal. According to the removal index, the preliminary pseudo EO signal is continuously iterated, and finally a more accurate final pseudo EO signal can be obtained, so that the reference EO data of the IMM algorithm is more accurate, thereby improving the removal accuracy of the EO artifacts, making the new EEG signal more accurate, and can assist doctors in analyzing the sleep status more accurately through the new EEG data. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0030] Figure 1 A block diagram of an electroencephalogram signal processing system for sleep analysis provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0031] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following is a detailed description of the specific implementation, structure, features and effects of an EEG signal processing system for sleep analysis proposed by the present invention in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0032] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0033] The following is a detailed description of a specific scheme of an electroencephalogram signal processing system for sleep analysis provided by the present invention in conjunction with the accompanying drawings.

[0034] See also Figure 1, which shows a block diagram of an electroencephalogram signal processing system for sleep analysis provided by an embodiment of the present invention. The system includes: a data acquisition module, a data processing module, a feature analysis module and a sleep analysis module.

[0035] The data acquisition module S1 is used to obtain EEG signals from different areas of the brain.

[0036] In an embodiment of the present invention, the implementation scenario is to remove the electrooculogram artifacts of EEG signals. The IMM algorithm is an interactive multi-model. The basic idea is to use multiple Kalman filters for parallel processing. Each filter corresponds to a different state space model. Different state space models describe different target motion patterns. Since the IMM algorithm is a prior art, the specific principle is not repeated. In order to remove the electrooculogram artifacts of EEG signals, the traditional IMM algorithm passes the EEG signal through a 0-8 Hz filter, filters out the high-frequency EEG information, sets the EEG data with an amplitude less than 10 microvolts to zero, and selects the most noisy lead as the reference electrooculogram. The reference electrooculogram is the data of the electrooculogram artifacts in the EEG information; causing some EEG signals with higher amplitudes to be eliminated as noise, thereby causing some EEG information components to be lost, affecting the accuracy of the EEG signal after denoising; and the quality of the constructed reference electrooculogram is poor. Therefore, in order to make the EEG signal more accurate after removing the electrooculogram artifacts through the IMM algorithm, it is necessary to improve the construction quality of the reference electrooculogram in the IMM algorithm.

[0037] First, the EEG signals of different areas of the brain are obtained and preprocessed. In an embodiment of the present invention, the most important non-invasive acquisition method for obtaining EEG signals in current scientific research experiments is used, that is, electrodes are placed on the scalp to record EEG signals, and multiple electrodes are used to collect EEG signals. The placement of the electrodes needs to be based on the standard electrode 10-20 system placement method specified by the International Electroencephalography Association. The electrodes are placed. Fp represents the prefrontal lobe area of ​​the brain, F represents the frontal lobe area of ​​the brain, and C represents the central lobe area of ​​the brain. These three areas are areas that are severely affected by electrooculographic artifacts. In an embodiment of the present invention, eight-lead EEG signals are obtained, corresponding to Fp1, Fp2, F3, F4, F7, F8, C3, C4 in the international 10-20 system electrode placement method. The sampling frequency of the electrode is one kilohertz, and the collected EEG signals are low-pass filtered to obtain 0.5~100 Hz, and the EEG signals after filtering preprocessing are obtained. It should be noted that the implementer can obtain different numbers of EEG signals from other areas in the international 10-20 system electrode placement method. The implementer can determine the sampling frequency and the filtered frequency according to the implementation scenario, which is not limited here. Low-pass filtering belongs to the prior art, and the specific steps are not repeated. At this point, the EEG signals of different areas of the brain have been obtained and preprocessed. The subsequent steps require analyzing and obtaining the electrooculographic artifacts in the EEG signals.

[0038] The data processing module S2 is used to obtain the initial interference degree of each area at the target moment in the EEG signal of the target area, with the target moment as the center of the preset window, according to the difference characteristics of the EEG signal amplitude in the preset window; and obtain the local correlation at the target moment according to the difference characteristics of the initial interference degree between the target area and other areas at the corresponding moment in the preset window.

[0039] When collecting EEG signals, the potential difference is generated due to the eye movement. The potential difference propagates from front to back through the skull and is captured by the scalp electrodes, causing noise interference to the EEG signals collected by the electrodes and generating electrooculogram artifacts. Therefore, the electrooculogram artifacts need to be filtered out. In the embodiment of the present invention, the EEG signals of different electrodes are analyzed, and a pseudo electrooculogram signal is adaptively constructed based on the degree of interference of the EEG signals, i.e., the reference data of the IMM algorithm. It should be noted that the reason for constructing the pseudo electrooculogram signal is: on the one hand, if the electrooculogram signal is collected synchronously, it is not conducive to the portable development of the EEG signal collection equipment. On the other hand, the electrooculogram signal will also be interfered by the artifacts of the EEG signal during collection. There is a phenomenon of mutual propagation interference between the two, which leads to the instability of the electrooculogram artifact removal effect.

[0040] Because the EEG signals collected by any electrode contain the original EEG signals without noise, electrooculographic artifacts and Gaussian white noise. In order to obtain more accurate EEG signals, it is necessary to analyze the interference degree of electrooculographic artifacts on EEG signals.

[0041] First, it is necessary to preliminarily analyze the interference of electrooculographic artifacts on EEG signals. In the EEG signals of the target area, the target moment is taken as the center of the preset window, and the initial interference degree at the target moment is obtained according to the difference characteristics of the EEG signal amplitude in the preset window. Specifically, it includes: taking the target moment as the center of the preset window, calculating the variance of the EEG signal amplitude in the preset window, the number of changes of the amplitude, and the range of the amplitude; calculating the product of the variance, the number of changes, and the range to obtain the initial interference degree at the target moment. In one embodiment of the present invention, the formula for obtaining the initial interference degree includes:

[0042]

[0043] In the formula, Indicates the target area Target moment The initial disturbance level, Indicates the target area Target moment The number of changes in the EEG signal amplitude within the preset window, Indicates the target area Target moment The variance of the EEG signal amplitude within the preset window, It means to find the variance of the data in brackets. Indicates the target area Target moment The difference between the maximum and minimum values ​​of the EEG signal amplitude within the preset window. represents an exponential function with a natural constant as the base, and the goal is to perform positive correlation mapping on the formula in the brackets to remove the influence of the dimension. It should be noted that in other embodiments of the present invention, the implementer can remove the dimension by mathematical calculation methods such as calculating the square of the value and other positive correlation mapping means, which is not limited here.

[0044] The meaning and examples of obtaining the initial interference level, where the target area and target time refer to the area collected by any electrode in the brain and any time of collection. Because the collected EEG signals are time series signals and the sampling frequencies of different electrodes are consistent, for the target time , with the target moment as the center of the sliding preset window, in the embodiment of the present invention, the size of the preset window is 25*25, which can be determined by the implementer according to the implementation scenario. The acquisition of the initial interference degree is based on the variance, number of changes and range of the EEG signal amplitude in the window corresponding to the target moment. The amplitude change amplitude of the EEG signal interfered by the electrooculogram signal is larger than the amplitude change amplitude of the EEG signal itself. The range and variance of the amplitude can characterize the amplitude change amplitude of the EEG amplitude in the preset window. The larger the range and variance, the larger the amplitude change amplitude and the greater the degree of interference by the electrooculogram artifact. In addition, because the waveforms of the EEG signal and the electrooculogram signal have certain volatility, when the degree of interference is greater, the superposition phenomenon of the two signals is more serious, and the clutter of the local signal is stronger. The clutter of the signal is reflected in the number of changes of the EEG signal amplitude in the preset window. For the acquisition of the number of changes, the number of different amplitudes at all moments in the preset window and the next moment is counted to obtain the number of changes of the amplitude. The more the number of changes, the stronger the clutter and the greater the degree of interference by the electrooculogram artifact. At this point, the initial interference degree of electrooculographic artifacts on the collected EEG signals was obtained.

[0045] Furthermore, since the frequency of the electrooculogram signal is in the range of 0.1 to 20 Hz, its amplitude is in the range of 50 to 200 microvolts, and the frequency of the electroencephalogram signal is in the range of 0.5 to 40 Hz, its amplitude is in the range of 50 to 100 microvolts. When the amplitude of the electrooculogram signal and the amplitude of the electroencephalogram signal are both low, the amplitude of the signal collected after superimposing the interference is also low. At this time, the initial interference degree value obtained is low, but the interference degree of the electrooculogram signal is actually high. Therefore, it is necessary to optimize the initial interference degree of the electroencephalogram signal based on the changes in electrode signals in different regions in the future, and improve the data accuracy of the interference degree of electrooculogram artifacts on the electroencephalogram signal. Because the collected electrodes are in different regions of the brain, there are electroencephalogram signals in the same region and different regions. For the electroencephalogram signals in the same region, their data features are similar, and the distance between the same region of the brain and the eyeball is also similar, so the interference degree of the electrooculogram signals in the same region is also close, that is, the electroencephalogram signals collected in the same region have certain similarities. Since the electrooculogram signal exhibits the characteristic of continuous attenuation during transmission, for the EEG signal at a certain moment, since different areas are interfered by the electrooculogram signal to different degrees, there will be certain differences in the EEG signals in different areas.

[0046] In order to obtain a more accurate interference degree of electrooculographic artifacts, it is necessary to analyze the correlation of the interference degree of electrooculographic artifacts in different regions. Therefore, the initial interference degree of the target region and other regions at all times within the preset window of the target time can be obtained. The local correlation at the target time is obtained according to the difference characteristics of the initial interference degree of the target region and other regions at the corresponding time in the preset window, which specifically includes:

[0047] In a preset window at the target time, the absolute value of the difference between the initial interference degree of the target area and other areas at the corresponding time is calculated and negative correlation mapping is performed to obtain a local correlation value, and the average value of the local correlation value is calculated to obtain the local correlation at the target time. In one embodiment of the present invention, the formula for obtaining the local correlation includes:

[0048]

[0049] In the formula, Indicates that at the target time When the target area and other areas The local correlation of Indicates the number of moments in the preset window, Indicates different moments within the preset window. Indicates that in the target area China-Israel Target Moment The center of the preset window. The initial disturbance level at time In other areas China-Israel Target Moment The center of the preset window. The initial disturbance level at time. Indicates that the formula in the brackets is negatively correlated and normalized. It should be noted that in other embodiments of the present invention, the implementer may select Other mathematical operation formulas such as are used for negative correlation mapping and normalization, which are not limited here.

[0050] Regarding the meaning of local correlation, and is obtained by The moment is taken as the center of the corresponding preset window, and the amplitude change characteristics of the EEG signal in the preset window are calculated. The initial interference degree of the corresponding area at the moment. The meaning of the target area and other areas is to distinguish different areas. Since the international 10-20 system electrode placement method is used in the acquisition process, the target area or other areas can be judged according to the definition of different areas in the placement method. The more similar the initial interference degree of two different areas at the same moment, the more similar the interference of the electrooculographic artifacts in the two areas, the more similar the waveforms of their EEG signals, and the greater the local correlation.

[0051] At this point, after obtaining the local correlation of EEG signals between different regions, the final interference degree can be calculated based on the local correlation of different regions combined with the initial interference degree analysis.

[0052] The feature analysis module S3 is used to obtain the final interference degree at the target moment according to the local correlation at the target moment, the EEG signal amplitude difference characteristics between the target area and other areas, and the initial interference degree of the target area; and obtain the preliminary pseudo-eye contact signal according to the EEG signal at the target moment and the final interference degree.

[0053] Because the EEG signal characteristics and local correlation of different regions are combined, a more accurate interference degree can be obtained compared to analyzing the EEG signal characteristics of a single region. The final interference degree is obtained according to the local correlation at the target moment, the EEG signal amplitude difference characteristics between the target region and other regions, and the initial interference degree of the target region. Specifically, it includes: at the target moment, the absolute value of the EEG signal difference between the target region and other regions is calculated as the signal difference value, and the product of the signal difference and the corresponding local correlation is calculated to obtain the signal correlation difference; the signal correlation differences corresponding to all other regions are accumulated to obtain the signal correlation difference characteristics; the accumulated value of the absolute value of the difference of the local correlations between other regions is calculated to obtain the local correlation difference; the inverse of the local correlation difference is calculated and the product of the corresponding signal correlation difference characteristics is negatively correlated to obtain the first value, and the difference between the value one and the first value is calculated to obtain the second value; the product of the initial interference degree and the corresponding second value is calculated to obtain the interference degree adjustment value, and the sum of the initial interference degree and the corresponding interference degree adjustment value is calculated to obtain the final interference degree. The formula for obtaining the final interference degree includes:

[0054]

[0055] In the formula, Indicates that at the target time Target area The final level of interference, Indicates the target area Target time The initial disturbance level, Indicates the target time Target area The EEG signal amplitude, Indicates the number of other regions, and Indicates different other areas, Indicates the target time Other areas The amplitude of the EEG signal. Indicates that at the target time When the target area and other areas local correlation. It represents a preset first value, the goal is to prevent the denominator from being zero, and in the embodiment of the present invention, it is 1, which can be determined by the implementer. is the signal-dependent difference, is the signal-related difference feature, is the local correlation difference.

[0056] It should be noted that in the embodiment of the present invention, the number of other regions is 2. Since the embodiment of the present invention collects eight-channel EEG signals from three areas of the brain, and the EEG signals in the same area are similar, the EEG signals in the target area and other areas in the embodiment of the present invention are calculated as any EEG signal in the corresponding area. The implementer can select other areas and EEG signals according to the implementation scenario.

[0057] Regarding the meaning of the final interference degree, since there is no interference from the electrooculogram signal to the EEG signal, the difference in the EEG signal values ​​in different regions will not be large. The greater the difference in the amplitude of the two EEG signals in different regions, that is, The larger the value of , the greater the interference of the EEG signal by the EO signal, resulting in greater differences in the EEG signals in different regions. Because the local correlation can characterize the correlation of the interference degree of the EEG signals in two regions, the interference degree of the EO signal is characterized by the product of the signal difference value and the corresponding local correlation. When the signal correlation difference feature is larger, it means that the interference degree of the EO signal on the EEG signal is greater. Further, because the embodiment of the present invention optimizes the initial interference degree based on the propagation characteristics of the EO signal, and when there are certain differences in the EEG signals in different regions, the propagation characteristics of the EO signal will be interfered with, making the characterization of the interference situation inaccurate, when there are differences in the EEG signals in two other regions, the difference in their local correlation is large, resulting in a larger final interference degree. Therefore, the result of the signal correlation difference feature characterizing the interference degree is corrected by the local correlation difference, eliminating the influence of the differences in the EEG signals in different regions on the final interference degree result. When the local correlation difference is smaller and the signal correlation difference feature is larger, it means that the interference degree of the EO signal is greater, the calculated interference degree adjustment value is larger, and the final interference degree obtained is greater.

[0058] At this point, the final interference degree is obtained, which can accurately characterize the impact of electrooculogram artifacts on the collected EEG signals compared to the initial interference degree. The preliminary pseudo electrooculogram signal can be obtained by multiplying the final interference degree with the EEG signal at the corresponding moment, and the final pseudo electrooculogram signal is subsequently determined based on the preliminary pseudo electrooculogram signal.

[0059] The interference removal module S4 is used to obtain the change characteristics of the final interference degree after removing the preliminary pseudo electrooculogram signal; obtain the removal index according to the change characteristics, iteratively update the preliminary pseudo electrooculogram signal according to the size of the removal index, and obtain the final pseudo electrooculogram signal; remove electrooculogram artifacts according to the final pseudo electrooculogram signal to obtain a new EEG signal.

[0060] The greater the final interference degree, the greater the current EEG signal amplitude noise. Therefore, in order to obtain a more accurate EEG signal, a preliminary pseudo-eog signal can be obtained based on the collected EEG signal and the final interference degree. The removal index is obtained according to the change characteristics of the final interference degree after removing the preliminary pseudo-eog signal. The preliminary pseudo-eog signal is iteratively updated according to the size of the removal index to obtain the final pseudo-eog signal.

[0061] First, the steps for obtaining the removal index specifically include: calculating the final interference degree of the EEG signal after removing the preliminary pseudo-eye signal as the final interference degree iteration value, calculating the absolute value of the difference between the final interference degree of the EEG signal and the corresponding final interference degree iteration value as the iteration representation value; calculating the product of the inverse of the final interference degree iteration value and the iteration representation value, and negatively mapping them as the removal feature, calculating the difference between the value one and the removal feature, and obtaining the removal index. The formula for obtaining the removal index includes:

[0062]

[0063] In the formula, Indicates that at the target time Target area The removal index, Indicates the target time Target area The EEG signal is based on the corresponding final interference level The final interference degree calculated again after removing the pseudo electrooculogram signal is used as the final interference degree iteration value.

[0064] Regarding the meaning of the removal index, the removal index can reflect the effect of removing pseudo-eye contact signals according to the final interference level. First, the preliminary pseudo-eye contact signal is obtained by multiplying the final interference level and the corresponding EEG signal. After removing the preliminary pseudo-eye contact signal from the collected EEG signal, the final interference level is calculated again, and the final interference level is represented by the iterative value of the final interference level. When the iterative value of the final interference level is less than the corresponding final interference level, it means that there is a certain removal effect. When the iterative value of the final interference level is smaller, it means that the pseudo-eye contact signal removal effect is better. Therefore, by removing the feature Characterizes the effect of removing pseudo electrooculogram signals. The smaller the removed eigenvalue is, the better the removal effect is.

[0065] After obtaining the removal index, the final pseudo electrooculogram signal can be determined according to the size of the removal index, specifically including: when the removal index is greater than or equal to the preset removal index threshold, the preliminary pseudo electrooculogram signal is used as the final pseudo electrooculogram signal; when the removal index is less than the preset removal index threshold, the sum of the removal index and the preset second value is calculated as the final interference degree secondary iteration value, the product of the final interference degree secondary iteration value and the preliminary pseudo electrooculogram signal is calculated, and the pseudo electrooculogram signal iteration value is obtained; the removal index after the pseudo electrooculogram signal iteration value is removed from the EEG signal is calculated; until the removal index is greater than or equal to the preset removal index threshold, the final pseudo electrooculogram signal iteration value is output as the final pseudo electrooculogram signal. In the embodiment of the present invention, the preset removal index threshold is 0.8, and the preset second value is 1, which can be determined by the implementer according to the implementation scenario. When the removal index is less than the preset removal index threshold, it means that the current electrooculogram artifact removal effect is not good, and further removal is needed. Therefore, the pseudo-electrooculogram signal iteration value and the removal index iteration value are obtained according to the product of the secondary iteration value of the final interference degree and the preliminary pseudo-electrooculogram signal, and the relationship between the iteration value of the removal index and the preset removal index threshold is judged again. If it is still less than the preset removal index threshold, the electrooculogram signal iteration value and the removal index iteration value are continued to be iterated until the iteration value of the removal index is greater than or equal to the preset removal index threshold, and the pseudo-electrooculogram signal iteration value at this time is output as the final pseudo-electrooculogram signal.

[0066] At this point, the final pseudo-eog signal of the collected EEG signal is obtained. The final pseudo-eog signal is adaptively adjusted according to the final interference degree of the collected EEG signal, which improves the accuracy of the final pseudo-eog signal. The final pseudo-eog signal at different times is used as the reference EEG data in the IMM algorithm, thereby improving the accuracy of removing EEG artifacts. The amplitude of the corresponding EEG signal is adjusted through the IMM algorithm and the reference EEG data to complete the reconstruction of the EEG signal and obtain a new EEG signal. It should be noted that the IMM algorithm belongs to the existing technology, and the specific steps will not be repeated. The new EEG signal can more accurately assist doctors in completing the analysis of sleep conditions.

[0067] In summary, an embodiment of the present invention provides an EEG signal processing system for sleep analysis, which first obtains EEG signals from different areas of the brain, obtains the initial interference degree according to the amplitude change characteristics of the EEG signals within a preset window, and obtains the local correlation according to the difference characteristics of the interference degrees at corresponding moments in different areas. The final interference degree and the preliminary pseudo-eoglycan signal are obtained according to the local correlation of different areas and the difference in EEG signal amplitude. The removal index is obtained according to the change characteristics of the final interference degree after removing the preliminary pseudo-eoglycan signal, and the final pseudo-eoglycan signal can be adaptively obtained according to the final interference index and the removal index, so that the final pseudo-eoglycan signal value is more accurate, thereby improving the accuracy of obtaining new EEG signals, and can more accurately assist doctors in analyzing sleep conditions through new EEG data.

[0068] It should be noted that the sequence of the above embodiments of the present invention is only for description and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0069] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.

Claims

1. An electroencephalogram signal processing system for sleep analysis, characterized in that: The system includes the following modules: Data acquisition module, used to obtain EEG signals from different areas of the brain; A data processing module is used to obtain the initial interference degree of each area at the target time according to the difference characteristics of the amplitude of the EEG signal in the preset window in the EEG signal of the target area, with the target time as the center of the preset window; and obtain the local correlation at the target time according to the difference characteristics of the initial interference degree between the target area and other areas at the corresponding time in the preset window; A feature analysis module is used to obtain the final interference degree at the target moment according to the local correlation at the target moment, the difference characteristics of the EEG signal amplitude between the target area and other areas, and the initial interference degree of the target area; and to obtain a preliminary pseudo-eye signal according to the EEG signal at the target moment and the final interference degree; An interference removal module is used to obtain the final interference degree change characteristics after removing the preliminary pseudo-oculogram signal; Obtaining a removal index according to the change feature, iteratively updating the preliminary pseudo electrooculogram signal according to the size of the removal index, and obtaining a final pseudo electrooculogram signal; According to the final pseudo electrooculogram signal, electrooculogram artifacts are removed to obtain a new electroencephalogram signal; The step of obtaining the removal index comprises: Calculate the final interference degree of the EEG signal after removing the preliminary pseudo-oculogram signal, as the final interference degree iteration value, and calculate the absolute value of the difference between the final interference degree of the EEG signal and the corresponding final interference degree iteration value, as the iteration representation value; The product of the inverse of the final interference degree iteration value and the iteration characterization value is calculated, and negatively correlated with the mapping, which is used as a removal feature. The difference between the value one and the removal feature is calculated to obtain a removal index.

2. The electroencephalogram signal processing system for sleep analysis according to claim 1, characterized in that: The step of obtaining the initial interference degree comprises: In the EEG signal, taking the target moment as the center of the preset window, the variance of the EEG signal amplitude, the number of amplitude changes and the range of the amplitude in the preset window are calculated; the product of the variance, the number of changes and the range is calculated to obtain the initial interference degree at the target moment.

3. The electroencephalogram signal processing system for sleep analysis according to claim 1, characterized in that: The step of obtaining the local correlation comprises: In a preset window at the target time, the absolute value of the difference between the initial interference degree of the target area and other areas at the corresponding time is calculated and negative correlation mapping is performed to obtain the local correlation value, and the average value of the local correlation value is calculated to obtain the local correlation at the target time.

4. The electroencephalogram signal processing system for sleep analysis according to claim 1, characterized in that: The step of obtaining the final interference degree comprises: At the target moment, the absolute value of the difference between the EEG signals of the target area and other areas is calculated as the signal difference value, and the product of the signal difference and the corresponding local correlation is calculated to obtain the signal correlation difference; the signal correlation differences corresponding to all other areas are accumulated to obtain the signal correlation difference feature; Calculate the cumulative value of the absolute value of the difference of the local correlations between other regions to obtain the local correlation difference; calculate the product of the inverse of the local correlation difference and the corresponding signal correlation difference feature and negatively correlate and map them to obtain a first value, calculate the difference between the value 1 and the first value, and obtain a second value; The product of the initial interference degree and the corresponding second value is calculated to obtain the interference degree adjustment value, and the sum of the initial interference degree and the corresponding interference degree adjustment value is calculated to obtain the final interference degree.

5. The electroencephalogram signal processing system for sleep analysis according to claim 1, characterized in that: The step of obtaining the preliminary pseudo electrooculogram signal comprises: The product of the final interference degree and the EEG data at the corresponding moment is calculated to obtain the preliminary pseudo-egoculoculographic signal.

6. The electroencephalogram signal processing system for sleep analysis according to claim 1, characterized in that: The step of obtaining the final pseudo electrooculogram signal comprises: When the removal index is greater than or equal to a preset removal index threshold, the preliminary pseudo electrooculogram signal is used as the final pseudo electrooculogram signal; When the removal index is less than a preset removal index threshold, the sum of the removal index and the preset second value is calculated as the final interference degree secondary iteration value, the product of the final interference degree secondary iteration value and the preliminary pseudo-oculogram signal is calculated to obtain the pseudo-oculogram signal iteration value; the removal index after the pseudo-oculogram signal iteration value is removed from the EEG signal is calculated; Until the removal index is greater than or equal to the preset removal index threshold, the final pseudo electrooculogram signal iteration value is output as the final pseudo electrooculogram signal.

7. The electroencephalogram signal processing system for sleep analysis according to claim 1, characterized in that: The step of acquiring the new EEG signal comprises: The final pseudo electrooculogram signal is used as reference electrooculogram data, and the amplitude of the electroencephalogram signal is adjusted through the IMM algorithm and the reference electrooculogram data to complete the reconstruction of the electroencephalogram signal and obtain the new electroencephalogram signal.

Images