Electroencephalogram signal preprocessing method and related device thereof

Abstract

The invention relates to an electroencephalogram signal preprocessing method and a related device thereof.The method comprises the steps that initial electroencephalogram signals collected by a multi-channel electrode array and at least one kind of auxiliary data collected synchronously with the initial electroencephalogram signals are obtained, and the auxiliary data comprise electrode impedance values and / or physiological state data; wherein the resistance reactance value indicates an electrode impedance value corresponding to the acquisition channel associated with the initial electroencephalogram signal, and the physiological data at least comprises motion posture data and / or physiological index data; performing multiple filtering processing on the initial electroencephalogram signal to obtain a filtered electroencephalogram signal; inputting the filtered electroencephalogram signals and at least one kind of auxiliary data into a signal restoration model, and optimizing the candidate electroencephalogram signals to obtain optimized target electroencephalogram signals. According to the scheme, the electroencephalogram signals can be preprocessed by dynamically adjusting the filtering parameters and utilizing the impedance value and the auxiliary data, and the quality of the electroencephalogram signals is improved.

Description

Technical Field

[0001] The present application relates to the field of computer technology, and in particular to a method for preprocessing an electroencephalogram (EEG) signal and a related device thereof. Background Art

[0002] Electroencephalography (EEG) refers to the bioelectrical signals generated by brain neuronal activity. Electrodes placed on the scalp can capture and record these subtle voltage changes. These changes reflect the overall activity of a large number of neurons firing synchronously.

[0003] Related technologies usually use fixed-parameter filters for processing during EEG signal acquisition, and are unable to dynamically adjust the filter parameters according to real-time impedance changes, resulting in poor signal processing effects. Summary of the Invention

[0004] In order to solve or partially solve the problems existing in the related technology, the present application provides a method for preprocessing EEG signals and related devices, which can preprocess EEG signals by dynamically adjusting filtering parameters and utilizing impedance values ​​to improve the quality of EEG signals.

[0005] The first aspect of the present application provides a method for preprocessing an EEG signal, comprising: acquiring an initial EEG signal acquired by a multi-channel electrode array, and at least one auxiliary data acquired synchronously with the initial EEG signal, the auxiliary data including an electrode impedance value and / or physiological state data; wherein the resistance reactance value indicates the electrode impedance value corresponding to the acquisition channel associated with the initial EEG signal, and the physiological data includes at least motion posture data and / or physiological indicator data; performing multiple filtering processes on the initial EEG signal to obtain a filtered EEG signal; inputting the filtered EEG signal and at least one of the auxiliary data into a signal restoration model, and optimizing the candidate EEG signal to obtain an optimized target EEG signal.

[0006] In combination with the first aspect, in a possible implementation of the first aspect, the step of obtaining electrode impedance data corresponding to each channel of the initial EEG signal includes: applying a test current signal to the multi-channel electrode array and obtaining a feedback voltage signal of the multi-channel electrode array; and calculating the electrode impedance value corresponding to each channel of the initial EEG signal based on the feedback voltage signal.

[0007] In combination with the first aspect, in a possible implementation of the first aspect, the initial EEG signal is subjected to multiple filtering processes to obtain a filtered EEG signal, including: using an adaptive notch filter to eliminate 50Hz or 60Hz power frequency interference in the initial EEG signal; wherein the filtering parameters are dynamically adjusted according to the real-time impedance value; and the initial EEG signal is sequentially subjected to high-pass filtering with a cutoff frequency of 0.5Hz and low-pass filtering with a cutoff frequency of 70Hz to obtain a filtered EEG signal.

[0008] In combination with the first aspect, in a possible implementation method of the first aspect, it also includes: in the adaptive notch filtering processing stage, adjusting the notch depth of the adaptive notch filter according to the filtering parameters, when the electrode impedance value exceeds 10kΩ, the notch attenuation strength of the acquisition channel corresponding to the electrode impedance value is increased to -30dB; in the high-pass filtering stage, a zero-phase infinite impulse response filter is used to eliminate the baseline drift of the initial EEG signal, the cutoff frequency of the zero-phase infinite impulse response filter is 0.5Hz, the transition band attenuation slope is ≥24dB / oct, and the passband fluctuation is controlled to ±0.05dB; in the low-pass filtering stage, a finite impulse response filter is used to suppress the high-frequency noise of the initial EEG signal, the set cutoff frequency of the finite impulse response filter is 70Hz, the stopband attenuation is greater than 40dB, and the transition band width is less than 10Hz.

[0009] In combination with the first aspect, in a possible implementation of the first aspect, before inputting the filtered EEG signal and the electrode impedance data into the signal restoration model, it also includes: generating a channel quality weight matrix based on the electrode impedance value corresponding to the acquisition channel associated with the initial EEG signal, and the weight of each acquisition channel is positively correlated with the electrode impedance value corresponding to the acquisition channel; wherein the signal restoration model performs differentiated processing on the filtered EEG signal based on the channel quality weight matrix.

[0010] In combination with the first aspect, in a possible implementation method of the first aspect, it also includes: collecting the electrode impedance value corresponding to the collection channel associated with the initial EEG signal within a preset time; based on the electrode impedance value, calculating the impedance fluctuation rate corresponding to different collection channels; if the electrode impedance value within the preset time is greater than a first preset threshold; and / or the impedance fluctuation rate within the preset time is greater than a second preset threshold; then generating an alarm reminder for the collection channel corresponding to the electrode impedance value and / or the impedance fluctuation rate.

[0011] In combination with the first aspect, in a possible implementation method of the first aspect, it also includes: when the physiological data includes motion posture data, the motion posture data is input into the signal restoration model, and the signal restoration model adjusts the spatial filtering parameters of the signal restoration model based on the motion posture data; when the physiological indicator data includes physiological indicator data, the physiological indicator data is input into the signal restoration model, and the signal restoration model identifies and separates the myoelectric artifacts or electrocardiographic artifacts in the filtered EEG signal based on the physiological indicator data.

[0012] The second aspect of the present application provides a preprocessing device for EEG signals, comprising an acquisition module for acquiring an initial EEG signal acquired by a multi-channel electrode array, and at least one auxiliary data acquired synchronously with the initial EEG signal, the auxiliary data including electrode impedance values ​​and / or physiological state data; wherein the resistance reactance value indicates the electrode impedance value corresponding to the acquisition channel associated with the initial EEG signal, and the physiological data includes at least motion posture data and / or physiological indicator data; a processing module for performing multiple filtering processes on the initial EEG signal to obtain a filtered EEG signal; an input module for inputting the filtered EEG signal and at least one auxiliary data into a signal restoration model, optimizing the candidate EEG signal to obtain an optimized target EEG signal.

[0013] In combination with the second aspect, in a possible implementation of the second aspect, the acquisition module is also used to apply a test current signal to the multi-channel electrode array and obtain a feedback voltage signal of the multi-channel electrode array; based on the feedback voltage signal, calculate the electrode impedance value corresponding to each channel of the initial EEG signal.

[0014] In combination with the second aspect, in a possible implementation of the second aspect, the processing module is also used to use an adaptive notch filter to eliminate 50Hz or 60Hz power frequency interference in the initial EEG signal; wherein the filtering parameters are dynamically adjusted according to the real-time impedance value; and the initial EEG signal is sequentially subjected to high-pass filtering with a cutoff frequency of 0.5Hz and low-pass filtering with a cutoff frequency of 70Hz to obtain a filtered EEG signal.

[0015] In combination with the second aspect, in a possible implementation of the second aspect, the processing module is also used to adjust the notch depth of the adaptive notch filter according to the filtering parameters in the adaptive notch filtering processing stage, and when the electrode impedance value exceeds 10kΩ, the notch attenuation strength of the acquisition channel corresponding to the electrode impedance value is increased to -30dB; in the high-pass filtering stage, a zero-phase infinite impulse response filter is used to eliminate the baseline drift of the initial EEG signal, the cutoff frequency of the zero-phase infinite impulse response filter is 0.5Hz, the transition band attenuation slope is ≥24dB / oct, and the passband fluctuation is controlled to ±0.05dB; in the low-pass filtering stage, a finite impulse response filter is used to suppress the high-frequency noise of the initial EEG signal, the set cutoff frequency of the finite impulse response filter is 70Hz, the stopband attenuation is greater than 40dB, and the transition band width is less than 10Hz.

[0016] In combination with the second aspect, in a possible implementation of the second aspect, the processing module is also used to generate a channel quality weight matrix based on the electrode impedance value corresponding to the acquisition channel associated with the initial EEG signal, and the weight of each acquisition channel is positively correlated with the electrode impedance value corresponding to the acquisition channel; wherein, the signal restoration model performs differentiated processing on the filtered EEG signal based on the channel quality weight matrix.

[0017] In combination with the second aspect, in a possible implementation of the second aspect, the acquisition module is also used to collect the electrode impedance value corresponding to the acquisition channel associated with the initial EEG signal within a preset time; based on the electrode impedance value, the impedance fluctuation rate corresponding to different acquisition channels is calculated; if the electrode impedance value within the preset time is greater than the first preset threshold; and / or the impedance fluctuation rate within the preset time is greater than the second preset threshold; then an alarm reminder is generated for the acquisition channel corresponding to the electrode impedance value and / or the impedance fluctuation rate.

[0018] In conjunction with the second aspect, in a possible implementation of the second aspect, the input module is further configured to, when the physiological data includes motion posture data, input the motion posture data into the signal restoration model, and the signal restoration model adjusts the spatial filtering parameters of the signal restoration model based on the motion posture data; and when the physiological indicator data includes physiological indicator data, input the physiological indicator data into the signal restoration model, and the signal restoration model identifies and separates myoelectric artifacts or electrocardiographic artifacts in the filtered EEG signal based on the physiological indicator data. A third aspect of the present application provides an electronic device, comprising:

[0019] processor; and

[0020] The memory stores executable codes thereon, and when the executable codes are executed by the processor, the processor is caused to execute the method described above.

[0021] A fourth aspect of the present application provides a computer-readable storage medium having executable code stored thereon. When the executable code is executed by a processor of an electronic device, the processor is caused to execute the method described above.

[0022] A fifth aspect of the present application provides a computer program product, comprising a computer program / instruction, which implements the method described above when executed by a processor.

[0023] The technical solution provided by this application may have the following beneficial effects:

[0024] The present application provides a method for preprocessing an EEG signal and a related device, comprising: obtaining an initial EEG signal acquired by a multi-channel electrode array, and at least one auxiliary data acquired synchronously with the initial EEG signal, the auxiliary data including electrode impedance values ​​and / or physiological state data; wherein the resistance reactance value indicates the electrode impedance value corresponding to the acquisition channel associated with the initial EEG signal, and the physiological data includes at least motion posture data and / or physiological indicator data; performing multiple filtering processes on the initial EEG signal to obtain a filtered EEG signal; inputting the filtered EEG signal and at least one auxiliary data into a signal restoration model, optimizing the candidate EEG signal to obtain an optimized target EEG signal, and being able to preprocess the EEG signal by dynamically adjusting the filtering parameters and utilizing the impedance value and the auxiliary data to improve the quality of the EEG signal.

[0025] It should be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] The above and other objects, features and advantages of the present application will become more apparent by describing in more detail exemplary embodiments of the present application in conjunction with the accompanying drawings, wherein the same reference numerals generally represent the same components in the exemplary embodiments of the present application.

[0027] Figure 1 1 is a flow chart of a method for preprocessing an EEG signal according to an embodiment of the present application;

[0028] Figure 2 Schematic diagram of the structure of the electroencephalogram signal preprocessing device shown in an embodiment of the present application;

[0029] Figure 3 It is a structural diagram of an electronic device shown in an embodiment of the present application. DETAILED DESCRIPTION

[0030] The following describes embodiments of the present application in more detail with reference to the accompanying drawings. Although the accompanying drawings illustrate embodiments of the present application, it should be understood that the present application can be implemented in various forms and should not be limited by the embodiments described herein. Rather, these embodiments are provided to make the present application more thorough and complete, and to fully convey the scope of the present application to those skilled in the art.

[0031] The terms used in this application are for the purpose of describing specific embodiments only and are not intended to limit this application. The singular forms "a," "an," and "the" used in this application and the appended claims are also intended to include plural forms unless the context clearly indicates otherwise. It should also be understood that the term "and / or" as used herein refers to and includes any or all possible combinations of one or more of the associated listed items.

[0032] It should be understood that although the terms "first", "second", "third", etc. may be used in this application to describe various information, this information should not be limited to these terms. These terms are only used to distinguish information of the same type from each other. For example, without departing from the scope of this application, the first information may also be referred to as the second information, and similarly, the second information may also be referred to as the first information. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features. In the description of this application, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined.

[0033] Epilepsy is a chronic neurological disease caused by sudden abnormal discharges of brain neurons, resulting in temporary brain dysfunction and symptoms such as limb rigidity, abnormal limb twitching, and loss of consciousness. Epilepsy attacks often cause accidental injuries to patients due to loss of consciousness, loss of physical control, and respiratory arrest. If not treated promptly, brain inflammation may aggravate neurological damage, leading to more serious consequences. Epilepsy attacks are sudden and random, affecting patients' normal work and life and causing anxiety. The diagnosis of epilepsy can be made by obtaining the patient's EEG signals.

[0034] In related technologies, during the EEG signal acquisition process, filters with fixed parameters are usually used for processing, and the filter parameters cannot be dynamically adjusted according to real-time impedance changes, resulting in poor signal processing effects.

[0035] In response to the above problems, the embodiments of the present application provide a method for preprocessing EEG signals and related devices, which can preprocess EEG signals by dynamically adjusting filtering parameters and utilizing impedance values ​​to improve the quality of EEG signals.

[0036] The technical solutions of the embodiments of the present application are described in detail below with reference to the accompanying drawings.

[0037] Figure 1 3 is a flow chart of the method for preprocessing EEG signals shown in an embodiment of the present application.

[0038] See also Figure 1 , a method for preprocessing an EEG signal, comprising:

[0039] S110: Acquire an initial EEG signal acquired by a multi-channel electrode array, and at least one auxiliary data acquired synchronously with the initial EEG signal, the auxiliary data including electrode impedance values ​​and / or physiological state data; wherein the resistance reactance value indicates the electrode impedance value corresponding to the acquisition channel associated with the initial EEG signal, and the physiological data includes at least motion posture data and / or physiological indicator data.

[0040] Specifically, the multi-channel electrode array can be an electrode device comprising multiple independent collection points, for example, an octagonal electrode made of flexible material. Each collection point corresponds to the acquisition of signals from a specific brain area, and can collect the initial EEG signal from the patient's head, as well as obtain the electrode impedance value of the collection channel corresponding to the initial EEG signal.

[0041] Specifically, the auxiliary data may include electrode impedance values ​​and the patient's physiological status data. Motion sensors may be used to collect the patient's motion posture data, and physiological monitoring equipment may be used to collect the patient's physiological index data. By obtaining multi-dimensional auxiliary information, a quality assessment basis can be provided for subsequent signal processing. The patient's motion posture data may include the patient's head movement data, etc., and the physiological index data may include the patient's electrocardiogram data, blood pressure data, etc.

[0042] S120: Perform multiple filtering processes on the initial EEG signal to obtain a filtered EEG signal.

[0043] Specifically, the initial EEG signal can be subjected to multiple filtering processes. For example, the original signal collected by the electrode array is sent to an adaptive filter group, and the initial EEG signal is processed through different filtering stages to obtain a filtered EEG signal. Filtering the initial EEG signal can suppress noise in the EEG signal and improve the quality of the EEG signal.

[0044] S130: Inputting the filtered EEG signal and at least one auxiliary data into a signal restoration model, optimizing the candidate EEG signal, and obtaining an optimized target EEG signal.

[0045] Specifically, the filtered EEG signal and the real-time measured electrode impedance data are jointly input into the signal restoration model. The impedance data can represent the credibility of each channel signal in matrix form. The motion posture data can be used to identify and compensate for the artifacts of the head signal caused by body movement. The ECG data can be used to locate and remove ECG interference. The signal restoration model can suppress the noise components of the high-impedance channel through the attention mechanism, while enhancing the effective signal characteristics of the low-impedance channel to obtain the optimized target EEG signal.

[0046] The present application provides a method for preprocessing EEG signals, comprising obtaining an initial EEG signal acquired by a multi-channel electrode array, and at least one auxiliary data acquired synchronously with the initial EEG signal, the auxiliary data including electrode impedance values ​​and / or physiological state data; wherein the resistance reactance value indicates the electrode impedance value corresponding to the acquisition channel associated with the initial EEG signal, and the physiological data includes at least motion posture data and / or physiological index data; performing multiple filtering processes on the initial EEG signal to obtain a filtered EEG signal; inputting the filtered EEG signal and at least one auxiliary data into a signal restoration model, optimizing the candidate EEG signal to obtain an optimized target EEG signal, and being able to improve the quality of the EEG signal by dynamically adjusting the filtering parameters and utilizing the impedance value to preprocess the EEG signal and perform auxiliary data.

[0047] In one possible embodiment, the step of obtaining electrode impedance data corresponding to each channel of the initial EEG signal includes: applying a test current signal to the multi-channel electrode array and obtaining a feedback voltage signal of the multi-channel electrode array; and calculating the electrode impedance value corresponding to each channel of the initial EEG signal based on the feedback voltage signal.

[0048] Specifically, the test current signal is a weak alternating current actively applied by the multi-channel electrode array, such as a sinusoidal wave signal with a frequency of 1kHz. The feedback voltage signal is the voltage response generated by the multi-channel electrode array after the test current is applied. By collecting the feedback voltage signal, it can be used to calculate the electrode impedance value.

[0049] Specifically, within a preset sampling period, the test current signal is synchronously applied to each acquisition channel of the multi-channel electrode array. At this time, the feedback voltage signal corresponding to each channel is independently collected and stored. The frequency of the test current can be adjusted according to actual needs. For example, when there is environmental electromagnetic interference, a higher frequency can be used to reduce the coupling effect. After the feedback voltage signal is amplified and denoised, its amplitude and phase information are extracted through discrete Fourier transform. Combined with the known test current parameters, the real-time impedance value of each channel is obtained using the impedance calculation formula. By dynamically monitoring the impedance changes, poor electrode contact or detachment problems can be discovered in time, thereby ensuring the continuity and reliability of EEG signal acquisition.

[0050] In one possible embodiment, the initial EEG signal is subjected to multiple filtering processes to obtain a filtered EEG signal, including: using an adaptive notch filter to eliminate 50Hz or 60Hz power frequency interference in the EEG signal; wherein the filtering parameters are dynamically adjusted according to the real-time impedance value; and sequentially performing high-pass filtering with a cutoff frequency of 0.1-0.5Hz and low-pass filtering with a cutoff frequency of 30-200Hz on the EEG signal to obtain a filtered EEG signal.

[0051] Specifically, an adaptive notch filter refers to a filter that can dynamically suppress specific frequency interference. Power frequency interference refers to 50Hz or 60Hz fixed frequency noise generated by the power system. The adaptive notch filter can eliminate 50Hz or 60Hz power frequency interference in the EEG signal. The filtering parameters of the adaptive notch filter can be dynamically adjusted according to the real-time impedance value. The real-time impedance value can be input into the adaptive notch filter to dynamically adjust the filtering parameters.

[0052] Specifically, after the initial EEG signal is processed by an adaptive notch filter, it can be subjected to high-pass filtering and low-pass filtering, which can be used to eliminate baseline drift and suppress high-frequency noise. During the filtering process, the initial EEG signal is firstly subjected to an adaptive notch filter to eliminate power frequency interference. The center frequency can be automatically switched to 50Hz or 60Hz according to the actual power frequency noise frequency, and the filter bandwidth can be dynamically adjusted with the real-time impedance value. For example, when the electrode impedance increases and the power frequency interference amplitude increases, the notch depth can be increased accordingly to enhance the suppression effect. Subsequently, a high-pass filter with a cutoff frequency of 0.5Hz is used to eliminate low-frequency baseline drift caused by breathing or movement. The filter avoids introducing time delay through zero-phase design. Finally, a low-pass filter with a cutoff frequency of 70Hz is used to remove high-frequency noise components such as electromyography signals, retaining the effective frequency band of the EEG signal. Multiple filtering steps are performed in the order of notch, high pass, and low pass to avoid the mutual influence of noise in different frequency bands and obtain the filtered EEG signal. This solves the problems of unstable power frequency interference suppression effect in EEG signals and difficulty in synchronous elimination of baseline drift and high-frequency noise, and can reduce phase distortion in the signal processing process.

[0053] In a possible embodiment, it also includes: in the adaptive notch filtering processing stage, the notch depth of the adaptive notch filter is adjusted according to the filtering parameters, which can be when the electrode impedance value exceeds 2-2.5 times the original impedance (the original impedance refers to the impedance value measured in the initial stage of electrode implantation), for example, the electrode impedance value is 10kΩ, and the notch attenuation strength of the acquisition channel corresponding to the electrode impedance value is increased to -30dB; in the high-pass filtering stage, a zero-phase infinite impulse response filter is used to eliminate the baseline drift of the EEG signal, the cutoff frequency of the zero-phase infinite impulse response filter is 0.5Hz, the transition band attenuation slope is ≥24dB / oct, and the passband fluctuation is controlled to ±0.05dB; in the low-pass filtering stage, a finite impulse response filter is used to suppress the high-frequency noise of the EEG signal, the set cutoff frequency of the finite impulse response filter is 70Hz, the stopband attenuation is greater than 40dB, and the transition band width is less than 10Hz.

[0054] Specifically, the initial EEG signal can be high-pass filtered using a zero-phase infinite impulse response filter. The zero-phase infinite impulse response filter is a filter with a symmetrical phase response, which can avoid time shift of the EEG signal during the filtering process, thereby retaining the time series characteristics of the EEG signal; the initial EEG signal can be low-pass filtered using a finite impulse response filter. The finite impulse response filter refers to a filter with a unit impulse response of a finite length, which can avoid signal waveform distortion and suppress high-frequency noise of the initial EEG signal.

[0055] Specifically, in the adaptive notch filtering stage, when it is detected that the electrode impedance value of a certain channel exceeds 10kΩ, for example, the impedance is increased due to poor electrode contact, the notch attenuation strength of the channel is automatically increased to -30dB, thereby enhancing the ability to suppress power frequency interference. In the high-pass filtering stage, a zero-phase infinite impulse response filter with a cutoff frequency of 0.5Hz is used, such as a Chebyshev filter. The baseline drift is eliminated through bidirectional filtering. At the same time, the transition band attenuation slope is set to not less than 24dB / oct to reduce the transition band width, and the passband fluctuation is controlled within ±0.05dB to avoid signal amplitude distortion. In the low-pass filtering stage, a finite impulse response filter with a cutoff frequency of 70Hz is used. For example, it can be a filter based on a Hamming window design. By setting the stopband attenuation to be greater than 40dB and the transition band width to be less than 10Hz, high-frequency noise such as electromyographic interference can be effectively filtered out. The filtering strength can be adaptively enhanced to address the impedance increase problem caused by poor electrode contact, and power frequency interference can be effectively suppressed. Zero-phase filtering is used to eliminate baseline drift while avoiding signal time shift, ensuring the accuracy of the time characteristics of the EEG signal. Linear phase low-pass filtering is used to remove high-frequency noise while retaining the signal waveform, providing high-quality EEG signals for subsequent signal restoration models.

[0056] In one possible embodiment, before the filtered EEG signal and electrode impedance data are input into the signal restoration model, it also includes: generating a channel quality weight matrix based on the electrode impedance value corresponding to the acquisition channel associated with the initial EEG signal, and the weight of each acquisition channel is positively correlated with the electrode impedance value corresponding to the acquisition channel; wherein the signal restoration model performs differentiated processing on the filtered EEG signal based on the channel quality weight matrix.

[0057] Specifically, the channel quality weight matrix indicates the weight distribution matrix generated by the electrode impedance values ​​of each acquisition channel, which can be obtained by normalizing the electrode impedance values ​​and mapping them into weight coefficients. The weight coefficients are directly proportional to the impedance values. The weight of each acquisition channel is positively correlated with the electrode impedance value corresponding to the acquisition channel. The channel quality weight matrix can be used in the signal restoration model.

[0058] Specifically, in the signal restoration stage, the electrode impedance value of each acquisition channel is converted into a corresponding weight coefficient. For example, a channel with a higher impedance value can be assigned a higher weight, and a channel with a lower impedance value can be assigned a lower weight. After the signal restoration model receives the filtered EEG signal and the channel quality weight matrix, it enhances the signal component of the high-impedance channel and suppresses the noise interference of the low-impedance channel through weighted fusion or dynamic adjustment of the channel contribution. For example, during the model training process, the weight matrix can be used as the input of the attention mechanism, so that the signal restoration model focuses on channels with better signal quality, so that the signal processing process can dynamically distinguish the reliability differences between different channels, thereby improving the accuracy and noise resistance of the restored EEG signal.

[0059] In a possible embodiment, it also includes: collecting the electrode impedance value corresponding to the collection channel associated with the initial EEG signal within a preset time; calculating the impedance fluctuation rate corresponding to different collection channels based on the electrode impedance value; if the electrode impedance value within the preset time is greater than a first preset threshold; and / or the impedance fluctuation rate within the preset time is greater than a second preset threshold; then generating an alarm reminder for the collection channel corresponding to the electrode impedance value and / or the impedance fluctuation rate.

[0060] Specifically, the preset time refers to a continuous time period for monitoring impedance changes, for example, it can be 5 minutes or 10 minutes. The first preset threshold and the second preset threshold can be set according to actual needs. For example, the first preset threshold can be 10kΩ, and the second preset threshold can be 20%. There is no limitation here. The first preset threshold can be used to identify electrode channels with poor contact, and the second preset threshold can be used to identify electrode channels with unstable contact.

[0061] Specifically, during the EEG signal acquisition process, the electrode impedance value of each channel can be recorded in real time, and impedance data can be continuously collected within a preset time window. The impedance fluctuation rate is obtained by calculating the impedance average and standard deviation of each channel. When the impedance average of a channel exceeds the first preset threshold, or its impedance fluctuation rate exceeds the second preset threshold, it is determined that the channel has poor contact or unstable contact problems, and the alarm reminder mechanism is triggered. For example, when the impedance average of a channel reaches 12kΩ and the fluctuation rate reaches 25% within 5 minutes, the system automatically marks the channel in red on the operation interface and issues a prompt sound. It can dynamically monitor the stability of the electrode contact status and issue an alarm in time when the impedance is abnormal or the fluctuation is abnormal, so as to avoid the degradation of EEG signal quality due to electrode contact problems.

[0062] In a possible embodiment, it also includes: when the physiological data includes motion posture data, the motion posture data is input into the signal restoration model, and the signal restoration model adjusts the spatial filtering parameters of the signal restoration model based on the motion posture data; when the physiological indicator data includes physiological indicator data, the physiological indicator data is input into the signal restoration model, and the signal restoration model identifies and separates the electromyographic artifacts or electrocardiographic artifacts in the filtered EEG signal based on the physiological indicator data.

[0063] Specifically, motion posture data can be collected through an inertial measurement unit, which includes three-axis acceleration and angular velocity information, and can quantify the amplitude and direction of the patient's head movement. Physiological indicator data can be collected through physiological monitoring equipment. For example, through multiple sensors installed on the patient, physiological status data including heart rate, respiratory rate and skin conductivity can be obtained.

[0064] Specifically, after the motion state data and physiological state data are input into the signal restoration model, during the signal restoration process of the filtered EEG signal, when it is detected that a certain axial acceleration in the motion posture data exceeds the threshold, the spatial filtering parameter dynamic adjustment module automatically increases the filter attenuation coefficient of the corresponding motion direction channel. For example, when it is detected that the vertical acceleration of the electrodes in the frontal lobe area continuously exceeds 0.5g, the weight coefficient of the channel in this area in the spatial filtering is reduced. At the same time, the synchronization between the physiological indicator data and the filtered EEG signal can be analyzed to identify the interference caused by specific physiological activities. For example, when it is detected that the heart rate data has a large correlation ratio with the energy fluctuation in the 1-5Hz frequency band in the filtered EEG signal, it is determined that there is an ECG artifact, and the artifact separation algorithm based on wavelet transform is started. The myoelectric artifact is identified by analyzing the time domain correlation between the skin conductivity change and the 20-200Hz high-frequency noise component, and then the myoelectric artifact in the EEG signal is separated and removed by the adaptive filtering algorithm. This can achieve effective identification and removal of artifacts caused by head movement, heartbeat and muscle activity in the EEG signal, effectively compensate for the influence of head movement on the spatial distribution of the EEG signal, and improve the purity and reliability of the EEG signal.

[0065] In a possible implementation, the signal restoration model further includes a plurality of neural networks, and the neural network is one or more of a convolutional neural network, a recurrent neural network, and a long short-term memory network.

[0066] Specifically, the signal restoration model can optimize the processing of filtered EEG signals by combining different neural network structures. When using a convolutional neural network, the model can extract the spatial correlation characteristics between multi-channel electrodes; when using a recurrent neural network or a long short-term memory network, the model can capture the dynamic pattern of signal changes over time. During implementation, the network structure can be dynamically selected based on the fluctuation of impedance data. For example, when the impedance is stable, the convolutional neural network is used to extract spatial features, and when the impedance fluctuates, the long short-term memory network is switched to process timing anomalies to improve the accuracy of signal restoration.

[0067] The EEG signal preprocessing method of the present application can effectively improve the targeted noise suppression during the EEG signal preprocessing process, reduce signal distortion caused by poor electrode contact or impedance changes, and at the same time realize differentiated processing of multi-channel data through the signal restoration model, and finally output the target EEG signal that is closer to the real neural activity.

[0068] Corresponding to the aforementioned embodiment of the method for realizing the application function, the present application also provides an EEG signal preprocessing device, an electronic device and corresponding embodiments.

[0069] Figure 2 Schematic diagram of the structure of the EEG signal preprocessing device shown in the embodiment of the present application.

[0070] See also Figure 2 , a pre-processing device 200 for an electroencephalogram signal, comprising:

[0071] The acquisition module 210 is used to acquire the initial EEG signal collected by the multi-channel electrode array, and at least one auxiliary data collected synchronously with the initial EEG signal, the auxiliary data including electrode impedance value and / or physiological state data; wherein the resistive reactance value indicates the electrode impedance value corresponding to the collection channel associated with the initial EEG signal, and the physiological data includes at least motion posture data and / or physiological indicator data.

[0072] The processing module 220 is used to perform multiple filtering processes on the initial EEG signal to obtain a filtered EEG signal.

[0073] The input module 230 is used to input the filtered EEG signal and at least one auxiliary data into the signal restoration model, optimize the candidate EEG signal, and obtain an optimized target EEG signal.

[0074] The present application provides a preprocessing device for EEG signals, comprising an acquisition module for acquiring an initial EEG signal acquired by a multi-channel electrode array, and at least one auxiliary data acquired synchronously with the initial EEG signal, the auxiliary data including electrode impedance values ​​and / or physiological state data; wherein the resistance reactance value indicates the electrode impedance value corresponding to the acquisition channel associated with the initial EEG signal, and the physiological data includes at least motion posture data and / or physiological indicator data; a processing module for performing multiple filtering processes on the initial EEG signal to obtain a filtered EEG signal; an input module for inputting the filtered EEG signal and at least one auxiliary data into a signal restoration model, optimizing the candidate EEG signal to obtain an optimized target EEG signal, and being able to preprocess the EEG signal by dynamically adjusting the filtering parameters and utilizing the impedance value to improve the quality of the EEG signal.

[0075] In a possible embodiment, the acquisition module 210 is further used to apply a test current signal to the multi-channel electrode array and obtain a feedback voltage signal of the multi-channel electrode array; and calculate the electrode impedance value corresponding to each channel of the initial EEG signal based on the feedback voltage signal.

[0076] In one possible embodiment, the processing module 220 is also used to use an adaptive notch filter to eliminate 50Hz or 60Hz power frequency interference in the EEG signal; wherein the filtering parameters are dynamically adjusted according to the real-time impedance value; and the EEG signal is sequentially subjected to high-pass filtering with a cutoff frequency of 0.5Hz and low-pass filtering with a cutoff frequency of 70Hz to obtain a filtered EEG signal.

[0077] In one possible embodiment, the processing module 220 is also used to adjust the notch depth of the adaptive notch filter according to the filtering parameters in the adaptive notch filtering processing stage, and when the electrode impedance value exceeds 10kΩ, the notch attenuation strength of the acquisition channel corresponding to the electrode impedance value is increased to -30dB; in the high-pass filtering stage, a zero-phase infinite impulse response filter is used to eliminate the baseline drift of the EEG signal, the cutoff frequency of the zero-phase infinite impulse response filter is 0.5Hz, the transition band attenuation slope is ≥24dB / oct, and the passband fluctuation is controlled to ±0.05dB; in the low-pass filtering stage, a finite impulse response filter is used to suppress the high-frequency noise of the EEG signal, the set cutoff frequency of the finite impulse response filter is 70Hz, the stopband attenuation is greater than 40dB, and the transition band width is less than 10Hz.

[0078] In one possible embodiment, the processing module 220 is also used to generate a channel quality weight matrix based on the electrode impedance value corresponding to the acquisition channel associated with the initial EEG signal, and the weight of each acquisition channel is positively correlated with the electrode impedance value corresponding to the acquisition channel; wherein, the signal restoration model performs differentiated processing on the filtered EEG signal based on the channel quality weight matrix.

[0079] In one possible embodiment, the acquisition module 210 is also used to collect the electrode impedance value corresponding to the acquisition channel associated with the initial EEG signal within a preset time; based on the electrode impedance value, the impedance fluctuation rate corresponding to different acquisition channels is calculated; if the electrode impedance value within the preset time is greater than the first preset threshold; and / or the impedance fluctuation rate within the preset time is greater than the second preset threshold; then an alarm reminder for the acquisition channel corresponding to the electrode impedance value and / or the impedance fluctuation rate is generated.

[0080] In one possible embodiment, the input module 230 is also used to input the motion posture data into the signal restoration model when the physiological data includes motion posture data, and the signal restoration model adjusts the spatial filtering parameters of the signal restoration model based on the motion posture data; when the physiological indicator data includes physiological indicator data, the physiological indicator data is input into the signal restoration model, and the signal restoration model identifies and separates the electromyographic artifacts or electrocardiographic artifacts in the filtered EEG signal based on the physiological indicator data.

[0081] In a possible implementation, the input module 230 is further used for the signal restoration model to include several neural networks, where the neural network is one or more of a convolutional neural network, a recurrent neural network, and a long short-term memory network.

[0082] Regarding the apparatus in the above embodiment, the specific manner in which each module performs operations has been described in detail in the embodiment of the method, and will not be elaborated again here.

[0083] An embodiment of the present application also provides an electronic device. Figure 3 This is a hardware structure diagram of an embodiment of an electronic device of the present application. The electronic device includes a memory 320 and at least one processor 310. The memory 320 is electrically connected to the at least one processor 310. The memory 320 stores instructions. The at least one processor 310 calls the instructions in the memory 320, causing the electronic device to execute the EEG signal preprocessing method according to any of the aforementioned embodiments of the present application.

[0084] Specifically, the processor 310 may include a central processing unit (CPU), or an application specific integrated circuit (ASIC), or may be configured to implement one or more integrated circuits of the embodiments of the present application.

[0085] The memory 320 may include a large capacity memory 320 for data or instructions. By way of example and not limitation, the memory 320 may include a hard disk drive (HDD), a floppy disk drive, a flash memory, an optical disk, a magneto-optical disk, a magnetic tape, or a universal serial bus (USB) drive, or a combination of two or more of these. Where appropriate, the memory 320 may include a removable or non-removable (or fixed) medium. Where appropriate, the memory 320 may be inside or outside the integrated gateway disaster recovery device. In a specific embodiment, the memory 320 is a non-volatile solid-state memory. In a specific embodiment, the memory 320 includes a read-only memory (ROM). Where appropriate, the ROM may be a mask-programmed ROM, a programmable ROM (PROM), an erasable PROM (EPROM), an electrically erasable PROM (EEPROM), an electrically rewritable ROM (EAROM), or a flash memory, or a combination of two or more of these.

[0086] In one example, the control device may further include a communication interface 330 and a bus 340. The processor 310, the memory 320, and the communication interface 330 are connected via the bus 340 and communicate with each other.

[0087] The communication interface 330 is mainly used to implement communication between various modules, devices, units and / or equipment in the embodiments of the present application.

[0088] Bus 340 includes hardware, software or both, and the parts of online data flow billing equipment are coupled to each other. For example, but not limitation, bus can include accelerated graphics port (AGP) or other graphics bus, enhanced industry standard architecture (EISA) bus, front side bus (FSB), hypertransport (HT) interconnection, industry standard architecture (ISA) bus, infinite bandwidth interconnection, low pin count (LPC) bus, memory bus, micro channel architecture (MCA) bus, peripheral component interconnection (PCI) bus, PCI-Express (PCI-X) bus, serial advanced technology attachment (SATA) bus, video electronics standard association local (VLB) bus or other suitable bus or two or more of these combinations. In appropriate cases, bus 340 can include one or more buses. Although the present application embodiment describes and shows specific bus, the application considers any suitable bus or interconnection.

[0089] In addition, in conjunction with the EEG signal preprocessing method in the above embodiments, the present application embodiment may provide a computer-readable storage medium for implementation. The computer-readable storage medium stores instructions that, when executed by a processor, implement any of the EEG signal preprocessing methods in the above embodiments.

[0090] The present application is not limited to the specific configurations and processes described above and shown in the figures. For the sake of brevity, a detailed description of known methods is omitted here. In the above embodiments, several specific steps are described and shown as examples. However, the method process of the present application is not limited to the specific steps described and shown. Those skilled in the art can make various changes, modifications and additions, or change the order of the steps after understanding the spirit of the present application.

[0091] The functional blocks shown in the above block diagram can be implemented as hardware, software, firmware or a combination thereof. When implemented in hardware, they can be, for example, electronic circuits, application specific integrated circuits (ASICs), appropriate firmware, plug-ins, function cards, etc. When implemented in software, the elements of the present application are programs or code segments that are used to perform the required tasks. Programs or code segments can be stored in machine-readable media, or transmitted on a transmission medium or a communication link by a data signal carried in a carrier wave. "Machine-readable media" can include any medium capable of storing or transmitting information. Examples of machine-readable media include electronic circuits, semiconductor memory devices, ROMs, flash memories, erasable ROMs (EROMs), floppy disks, CD-ROMs, optical disks, hard disks, optical fiber media, radio frequency (RF) links, etc. The code segments can be downloaded via computer networks such as the Internet, intranets, etc.

[0092] It should also be noted that the exemplary embodiments mentioned in this application describe some methods or systems based on a series of steps or devices. However, this application is not limited to the order of the above steps. In other words, the steps can be performed in the order mentioned in the embodiments, or in a different order, or several steps can be performed simultaneously.

[0093] Alternatively, the present application also provides a computer program product that can implement part or all of the steps of the method in the above embodiments. The computer program product includes a computer program / instructions, which implement part or all of the steps of the method in the above embodiments when executed by a processor.

[0094] The above is only a specific implementation method of the present application. Those skilled in the art can clearly understand that for the convenience and brevity of description, the specific working processes of the systems, modules and units described above can refer to the corresponding processes in the aforementioned method embodiments, and will not be repeated here. It should be understood that the scope of protection of the present application is not limited to this. Any technician familiar with this technical field can easily think of various equivalent modifications or replacements within the technical scope disclosed in this application, and these modifications or replacements should be included in the scope of protection of this application.

Claims

1. A method for preprocessing an EEG signal, characterized in that: include: Acquiring an initial EEG signal acquired by a multi-channel electrode array and at least one auxiliary data acquired synchronously with the initial EEG signal, the auxiliary data including an electrode impedance value and / or physiological state data; wherein the resistance reactance value indicates the electrode impedance value corresponding to the acquisition channel associated with the initial EEG signal, and the physiological data includes at least motion posture data and / or physiological indicator data; Performing multiple filtering processes on the initial EEG signal to obtain a filtered EEG signal; The filtered EEG signal and at least one of the auxiliary data are input into a signal restoration model, and the candidate EEG signal is optimized to obtain an optimized target EEG signal.

2. The method according to claim 1, characterized in that The step of obtaining electrode impedance data corresponding to each channel of the initial EEG signal includes: applying a test current signal to the multi-channel electrode array and obtaining a feedback voltage signal of the multi-channel electrode array; The electrode impedance values ​​corresponding to each channel of the initial EEG signal are calculated according to the feedback voltage signal.

3. The method according to claim 1, characterized in that The performing multiple filtering processes on the initial EEG signal to obtain a filtered EEG signal includes: Adopting an adaptive notch filter to eliminate 50Hz or 60Hz power frequency interference in the EEG signal; wherein the filter parameters are dynamically adjusted according to the real-time impedance value; The initial EEG signal is sequentially subjected to high-pass filtering with a cutoff frequency of 0.5 Hz and low-pass filtering with a cutoff frequency of 70 Hz to obtain a filtered EEG signal.

4. The method according to claim 3, characterized in that Also includes: In the adaptive notch filter processing stage, the notch depth of the adaptive notch filter is adjusted according to the filtering parameters. When the electrode impedance value exceeds 10kΩ, the notch attenuation strength of the acquisition channel corresponding to the electrode impedance value is increased to -30dB; In the high-pass filtering stage, a zero-phase infinite impulse response filter is used to eliminate the baseline drift of the initial EEG signal. The cutoff frequency of the zero-phase infinite impulse response filter is 0.5 Hz, the transition band attenuation slope is ≥24 dB / oct, and the passband fluctuation is controlled to ±0.05 dB; In the low-pass filtering stage, a finite impulse response filter is used to suppress high-frequency noise of the initial EEG signal. The set cutoff frequency of the finite impulse response filter is 70 Hz, the stopband attenuation is greater than 40 dB, and the transition band width is less than 10 Hz.

5. The method according to claim 1, wherein Before inputting the filtered EEG signal and the electrode impedance data into the signal restoration model, the method further includes: Generating a channel quality weight matrix based on the electrode impedance value corresponding to the acquisition channel associated with the initial EEG signal, wherein the weight of each acquisition channel is positively correlated to the electrode impedance value corresponding to the acquisition channel; The signal restoration model performs differential processing on the filtered EEG signal based on the channel quality weight matrix.

6. The method according to claim 1, wherein Also includes: Collecting the electrode impedance value corresponding to the collection channel associated with the initial EEG signal within a preset time; Based on the electrode impedance value, the impedance fluctuation rate corresponding to different acquisition channels is calculated; If the electrode impedance value within the preset time is greater than a first preset threshold; And / or, the impedance fluctuation rate within the preset time is greater than a second preset threshold; An alarm reminder for the collection channel corresponding to the electrode impedance value and / or the impedance fluctuation rate is generated.

7. The method according to claim 1, characterized in that Also includes: When the physiological data includes motion posture data, the motion posture data is input into the signal restoration model, and the signal restoration model adjusts the spatial filtering parameters of the signal restoration model based on the motion posture data; When the physiological indicator data includes physiological indicator data, the physiological indicator data is input into the signal restoration model, and the signal restoration model identifies and separates the myoelectric artifacts or electrocardiographic artifacts in the filtered electroencephalogram signal based on the physiological indicator data.

8. A preprocessing device for electroencephalogram signals, characterized in that: include: an acquisition module, configured to acquire an initial EEG signal acquired by a multi-channel electrode array, and at least one auxiliary data acquired synchronously with the initial EEG signal, the auxiliary data including an electrode impedance value and / or physiological state data; wherein the resistance reactance value indicates the electrode impedance value corresponding to the acquisition channel associated with the initial EEG signal, and the physiological data includes at least motion posture data and / or physiological indicator data; A processing module, configured to perform multiple filtering processes on the initial EEG signal to obtain a filtered EEG signal; An input module is used to input the filtered EEG signal and at least one of the auxiliary data into a signal restoration model, optimize the candidate EEG signal, and obtain an optimized target EEG signal.

9. An electronic device, characterized in that: include: processor; as well as A memory having executable codes stored thereon, which, when executed by the processor, causes the processor to execute the method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: An executable code is stored thereon, and when the executable code is executed by a processor of an electronic device, the processor is caused to execute the method according to any one of claims 1 to 7.

Images