WO2022017202A1 - 脑电信号的动态空间滤波与扩增方法、装置、电子设备以及存储介质 - Google Patents

脑电信号的动态空间滤波与扩增方法、装置、电子设备以及存储介质 Download PDF

Info

Publication number
WO2022017202A1
WO2022017202A1 PCT/CN2021/105590 CN2021105590W WO2022017202A1 WO 2022017202 A1 WO2022017202 A1 WO 2022017202A1 CN 2021105590 W CN2021105590 W CN 2021105590W WO 2022017202 A1 WO2022017202 A1 WO 2022017202A1
Authority
WO
WIPO (PCT)
Prior art keywords
signal
signals
eeg
lead
segmented
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Ceased
Application number
PCT/CN2021/105590
Other languages
English (en)
French (fr)
Inventor
许敏鹏
吴乔逸
罗睿心
明东
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Tianjin University
Original Assignee
Tianjin University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Priority claimed from CN202010728201.9A external-priority patent/CN111820876B/zh
Priority claimed from CN202110426679.0A external-priority patent/CN113509188B/zh
Application filed by Tianjin University filed Critical Tianjin University
Priority to US17/904,790 priority Critical patent/US20230055867A1/en
Publication of WO2022017202A1 publication Critical patent/WO2022017202A1/zh
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Images

Classifications

    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/0002—Remote monitoring of patients using telemetry, e.g. transmission of vital signals via a communication network
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
    • G06F3/01—Input arrangements or combined input and output arrangements for interaction between user and computer
    • G06F3/011—Arrangements for interaction with the human body, e.g. for user immersion in virtual reality
    • G06F3/015—Input arrangements based on nervous system activity detection, e.g. brain waves [EEG] detection, electromyograms [EMG] detection, electrodermal response detection
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/0002—Remote monitoring of patients using telemetry, e.g. transmission of vital signals via a communication network
    • A61B5/0004—Remote monitoring of patients using telemetry, e.g. transmission of vital signals via a communication network characterised by the type of physiological signal transmitted
    • A61B5/0006—ECG or EEG signals
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/02—Detecting, measuring or recording for evaluating the cardiovascular system, e.g. pulse, heart rate, blood pressure or blood flow
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/24—Detecting, measuring or recording bioelectric or biomagnetic signals of the body or parts thereof
    • A61B5/316—Modalities, i.e. specific diagnostic methods
    • A61B5/369—Electroencephalography [EEG]
    • A61B5/372—Analysis of electroencephalograms

Definitions

  • the present application relates to various fields such as computer, brain-computer interface, evaluation of cognitive state of brain function, and detection of brain state, in particular to a dynamic spatial filtering and amplification method, device, electronic device and storage medium of EEG signals.
  • Electroencephalogram is the overall reflection of the electrophysiological activity of brain nerve cells in the cerebral cortex, which can be recorded by scalp electrodes (Fig. 1A shows a schematic diagram of electrode placement).
  • EEG signals contain a large amount of physiological information.
  • EEG signals can be used to design and implement Brain-Computer Interface (BCI).
  • BCI Brain-Computer Interface
  • Brain-computer interface is a new type of human-computer interaction system that can acquire and decode physiological signals generated by the human brain to control computers or external devices. Motion Control Pathways. According to the different stimulation paradigms, BCI systems can be divided into three types: active, passive and reactive.
  • Active BCI is characterized by the user actively outputting commands to control external devices, mainly based on motor imagery (MI) signal systems; passive BCI is mostly used to detect brain states, such as mental state and attention level; reactive BCI It is mainly used to detect the response of the brain based on external stimuli and indirectly output control instructions.
  • stimulus-evoked signals such as Event-Related Potential (ERP), Steady State Visual Evoked Potential (SSVEP), Error-related potential (Error-Related Potential, ErrP), event-related desynchronization (Event-Related Desynchronization, ERD) and so on.
  • ERP Event-Related Potential
  • SSVEP Steady State Visual Evoked Potential
  • Error-related potential Error-Related Potential
  • Event-Related Desynchronization ERD
  • the BCI system is especially suitable for the following two application scenarios: (1) patients with impaired basic limb motor function but normal thinking; (2) narrow working space, inconvenient for limb movement (such as aerospace
  • EEG signals are non-stationary, time-varying random signals, and are susceptible to interference from background activity noise, motion artifacts, and electromagnetic noise.
  • noise interference In order to reduce noise interference and improve the signal-to-noise ratio of effective signals, most of the collected EEG signals usually need to undergo various preprocessing before further analysis: for example, downsampling can reduce storage pressure and improve real-time computing speed.
  • the interference of high-frequency noise is suppressed to a certain extent; digital filtering is often used to filter or retain signals in specific frequency bands, and the main types include low-pass filtering, high-pass filtering, band-pass filtering and notch; signal space projection (Signal-Space Projection, SSP) is often used to eliminate electromagnetic noise and electro-ophthalmic interference generated by equipment; Independent Component Analysis (ICA) is often used to separate multiple non-Gaussian statistically independent source signals; Principle Component Analysis (PCA) ) is often used for data dimensionality reduction to extract the main features of the signal; Maxwell Filtering (Maxwell Filtering) and Signal-Space Separation (SSS) are often used to separate and remove external electromagnetic noise (environmental noise) and so on.
  • SSP Signal-Space Projection
  • ICA Independent Component Analysis
  • PCA Principle Component Analysis
  • Maxwell Filtering Maxwell Filtering
  • SSS Signal-Space Separation
  • the traditional signal processing methods for EEG signals mainly include three steps of preprocessing, feature extraction, and pattern recognition (as shown by the dotted box in Figure 1B), which are very important in practical applications.
  • signal preprocessing can suppress noise signals, which is helpful for feature extraction and classification recognition, but cannot increase the effective component information of signals.
  • BCI systems that use EEG signals as a method of brain information acquisition have developed rapidly, and their performance has been continuously optimized.
  • the decoding of EEG signals still has limitations such as low spatial resolution and small amount of data.
  • researchers expect to increase the effective information of the signal through data augmentation methods.
  • EEG signal is a multi-channel dynamic time series, which is not suitable for using traditional image enhancement geometric transformation methods.
  • the EEG signal amplification methods in the current research have limited effect and poor robustness, so they have not been widely used in the construction of actual BCI systems.
  • the present application provides a method, device, electronic device and storage medium for dynamic spatial filtering and amplification of EEG signals, aiming to solve one of the technical problems in the related art at least to a certain extent.
  • a dynamic spatial filtering and amplifying method of an EEG signal including: constructing a spatial filter according to lead information of the EEG signal; and using the spatial filter to amplify the EEG signal .
  • constructing a spatial filter according to the lead information of the EEG signal includes the following steps:
  • step 1) input the current test data and carry out preprocessing according to step 1), and select the lead signal from the preprocessed test data according to the lead set obtained by the model training;
  • step 5 Apply the unified model to the test target lead in conjunction with the signal selected in step 4), dynamically solve the spatial filter applicable to the current environment and perform spatial filtering on the test data.
  • the unified model is specifically:
  • X (k) (i,:) ⁇ R 1 ⁇ m and Y (k) (i,:) ⁇ R 1 ⁇ n represent the target leads in the two time periods t ⁇ t 0 and t>t 0 respectively.
  • the signal of the kth trial of i; Indicates the selected leads from the remaining leads except the target lead i
  • the estimation of ; formula (1) is the spatial filter
  • p represents the p-norm of the vector, argmin represents the variable value to find the minimum value of the objective function; in contrast, argmax represents the variable value to find the maximum value of the objective function ; In order to minimize the corresponding p-norm output value of the spatial filter
  • the estimation of ; formula (2) is to solve the lead set The objective function f of , whose
  • a unified model is applied to the test target leads, a spatial filter suitable for the current environment is dynamically solved, and the test data is spatially filtered, specifically:
  • the signal obtained after noise reduction See formula (6), That is, after the filtering is completed, the test signal in the period of t>t 0; in, To estimate the spatial filter that minimizes the p-norm output value, and are the lead sets in the test data, t ⁇ t 0, respectively signal and single-trial signal of target lead i, and are the lead sets in the test data, t>t 0, respectively Signal vs. single-trial signal for target lead i.
  • constructing a spatial filter according to the lead information of the EEG signal includes: acquiring the EEG signal, determining a first lead from a plurality of leads included in the EEG signal, and obtaining a first lead from the first lead after removing the first lead.
  • the EEG signal is divided into a plurality of segmented EEG signals, Divide the plurality of segmented EEG signals into corresponding signals of a plurality of first time segments and signals of a plurality of corresponding second time segments; respectively determine the signals of the first lead under the plurality of first time segments For the corresponding plurality of first signals, respectively determine the signals of the second lead set under the plurality of first time segments as the corresponding plurality of second signals, and construct them according to the plurality of first signals and the plurality of second signals respectively Corresponding multiple spatial filters.
  • using a spatial filter to process the EEG signal includes: using a plurality of spatial filters to respectively perform spatial filtering processing on the signal of the corresponding first time segment and the signal of the second time segment to obtain an amplified signal; and splicing and integrating a plurality of amplified signals corresponding to the plurality of segmented EEG signals respectively, so as to amplify the EEG signals.
  • the method further includes: updating the current arrangement and combination form, and updating the updated EEG signals.
  • the EEG signals corresponding to the current permutation and combination form are amplified.
  • the EEG signal is divided into a plurality of segmented EEG signals, and the plurality of segmented EEG signals are respectively divided into signals of a plurality of corresponding first time segments, and signals of a plurality of corresponding second time segments.
  • Signals including: using a dynamic time window to divide the EEG signal into a plurality of segmented EEG signals, wherein the dynamic time window is expressed as a time range centered at t in the form of [t- ⁇ t 1 , t+ ⁇ t 2 ], [t- ⁇ t 1 ,t] represents the first time segment, and [t,t+ ⁇ t 2 ] represents the second time segment.
  • a spatial filter is constructed using a target formula, where the target formula is:
  • the target formula represents the constraints of the spatial filter W j corresponding to the segmented EEG signal numbered j,
  • a spatial filter to perform spatial filtering on the signal of the first time segment and the signal of the second time segment of the segmented EEG signal to obtain an amplified signal, including obtaining the amplified signal according to the following formula:
  • an apparatus for amplifying an EEG signal including: a signal acquisition module for acquiring an EEG signal, determining a first lead from a plurality of leads included in the EEG signal, and removing the first lead from the multiple leads included in the EEG signal.
  • the signal division module is used to divide the EEG The signal is divided into a plurality of segmented EEG signals, and the plurality of segmented EEG signals are respectively divided into signals of a plurality of corresponding first time segments and signals of a plurality of corresponding second time segments;
  • the filter building module It is used to respectively determine the signals of the first lead under multiple first time segments as the corresponding multiple first signals, and respectively determine the signals of the second lead set under multiple first time segments as the corresponding multiple first signals.
  • the filtering processing module is used for using the plurality of spatial filters to respectively compare the signals of the corresponding first time segment and the first time segment.
  • the signals of the two time segments are subjected to spatial filtering processing to obtain an amplified signal; and a first amplification module is used for splicing and integrating a plurality of amplified signals corresponding to the plurality of segmented EEG signals respectively, so as to perform the EEG signal analysis. Amplification.
  • an electronic device comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are processed by the at least one processor
  • the processor is executed, so that at least one processor can execute the method for amplifying the electroencephalogram signal disclosed in the present application.
  • a non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are used to cause a computer to execute the method for amplifying an EEG signal disclosed in the present application.
  • the technical solution provided by the present application is to filter and process the EEG signal by constructing and applying a spatial filter to obtain an amplified signal, and further splicing and integrating the amplified signal to amplify the EEG signal. Therefore, the potential EEG information can be effectively discovered from the original EEG signal, reflecting the current EEG characteristics, realizing the purpose of amplifying the EEG signal and improving the reliability and effectiveness of the EEG information. technical effect. Furthermore, it solves the technical problems that the existing EEG signal amplification methods have limited effect and poor robustness, so they have not been widely used in the construction of actual BCI systems.
  • This application can dynamically design spatial filters in the process of EEG signal preprocessing to suppress EEG noise of various non-target features, and has a wide range of applications;
  • the present application can be used for data preprocessing of EEG signals, which can effectively improve the quality of the collected signals, improve the performance of the brain-computer interface system, and is expected to obtain considerable social and economic benefits.
  • 1A is a schematic diagram of electrode placement for collecting EEG signals in the prior art
  • FIG. 1B is a schematic diagram of an EEG signal processing process in the prior art
  • FIG. 1C is a schematic diagram of an EEG signal processing process according to an embodiment of the present application.
  • FIG. 3 is a schematic flowchart according to another embodiment of the present application.
  • FIG. 5 is a schematic diagram of dividing an EEG signal by a dynamic time window according to an embodiment of the present application
  • FIG. 6 is a schematic diagram of an overall flow according to an embodiment of the present application.
  • FIG. 8 is a schematic flowchart according to another embodiment of the present application.
  • FIG. 9 is a block diagram of an electronic device used to implement the method for amplifying an EEG signal according to an embodiment of the present application.
  • the solution provides a method for amplifying an EEG signal. Referring to FIG. 1C , before feature extraction is performed on the EEG signal, the technical solution proposed in this application can amplify the signal. The method will be described below with reference to specific embodiments.
  • the execution subject of the electroencephalographic signal amplification method of this embodiment may be an electroencephalographic signal amplification device, and the device may be implemented by software and/or hardware, and the device may be configured in an electronic In the device, the electronic device may include, but is not limited to, a terminal, a server, and the like.
  • FIG. 2 is a schematic flowchart according to an embodiment of the present application. As shown in FIG. 2 , the method for amplifying an EEG signal includes:
  • the lead information of the EEG signal is obtained first, wherein the EEG signal may include multiple leads, for example, three leads of FP1, C3, and O1, and the information of the multiple leads may be referred to as Lead information of EEG signals.
  • a spatial filter can be constructed according to the lead information.
  • the construction of the spatial filter can be understood as dynamically solving the spatial filter suitable for the current environment.
  • the spatial filter can be constructed in any possible way. limit.
  • S202 Amplify the EEG signal by using a spatial filter.
  • the spatial filter can be used to amplify the EEG signal, which can effectively discover the potential EEG information from the original EEG signal, reflect the current EEG characteristics, and realize the analysis of the EEG signal.
  • the purpose of the amplification is to achieve the technical effect of improving the reliability and validity of the EEG information.
  • FIG. 3 is a schematic flowchart according to another embodiment of the present application. As shown in Figure 3, the method for amplifying the EEG signal includes:
  • S301 Acquire an EEG signal, determine a first lead from multiple leads included in the EEG signal, and determine at least one second lead from the multiple leads after removing the first lead to form a second lead set , take the first lead and the second lead set as the current permutation and combination form.
  • this solution first acquires an EEG signal, wherein the EEG signal is, for example, but not limited to, an epilepsy EEG signal, a steady-state visual evoked potential signal, and the like.
  • the EEG signal may be an originally collected EEG signal, or may be a preprocessed EEG signal.
  • the EEG signal can be in the form of a two-dimensional signal Among them, N c and N t represent the number of channels (ie, the number of leads) and acquisition time points of the EEG signal, and they are both constants, and R represents the set of real numbers.
  • the first lead is determined from the multiple leads included in the EEG signal, and at least one second lead is determined from the multiple leads after removing the first lead to form a second lead set.
  • the epilepsy EEG signal includes, for example, three leads FP1, C3, and O1.
  • the first lead and the second lead set need to be determined from the three leads.
  • the first lead (for example, FP1) can be determined first, and then any number of second leads are extracted from the remaining leads after the first lead is removed to form the second lead set.
  • the second lead set includes ⁇ C3,O1 ⁇ , ⁇ C3 ⁇ , ⁇ O1 ⁇ .
  • the solution also needs to determine the current permutation and combination form corresponding to the EEG signal, wherein the current permutation and combination form is the permutation and combination form of the first lead set and the second lead set.
  • the arrangement and combination of the first lead and the second lead set composed of the three leads FP1, C3, and O1 are shown in Table 1:
  • the permutations and combinations include: FP1 and ⁇ C3,O1 ⁇ , FP1 and ⁇ C3 ⁇ , FP1 and ⁇ O1 ⁇ , C3 and ⁇ FP1,O1 ⁇ , C3 and ⁇ FP1 ⁇ , C3 and ⁇ O1 ⁇ , O1 And ⁇ FP1, C3 ⁇ , O1 and ⁇ FP1 ⁇ , O1 and ⁇ C3 ⁇ nine kinds.
  • This solution needs to determine a permutation and combination form as the current permutation and combination form, for example: the current permutation and combination form is FP1 and ⁇ C3, O1 ⁇ .
  • the current permutation and combination form is FP1 and ⁇ C3, O1 ⁇ .
  • the following will continue to describe the present embodiment by taking FP1 and ⁇ C3, O1 ⁇ as the current arrangement and combination.
  • S302 Divide the EEG signal into a plurality of segmented EEG signals, and divide the plurality of segmented EEG signals into corresponding signals of a plurality of first time segments and signals of a plurality of corresponding second time segments.
  • the technical solution of this embodiment needs to divide the EEG signal into multiple segmented EEG signals, that is, intercept a segment of EEG signal into multiple segmented EEG signals, For example, the EEG signal is divided into: 1-segment EEG signal, 2-segment EEG signal...j-segment EEG signal, j represents the number of the segmented EEG signal.
  • the plurality of segmented EEG signals are respectively divided into signals of a plurality of corresponding first time segments and signals of a plurality of corresponding second time segments.
  • one segmented EEG signal can be arbitrarily extracted from multiple segmented EEG signals, and then the extracted EEG signal is divided again according to time to obtain the signal of the first time segment and the signal of the second time segment.
  • the lengths of the first time segment and the second time segment may be determined according to actual needs, and the lengths of the two time segments may be the same or different, which are not specifically limited in this solution.
  • Each segmented EEG signal is divided again according to the above method, and the division process will not be repeated here. Therefore, for each segmented EEG signal, a corresponding signal of the first time segment and a signal of the second time segment can be obtained.
  • S303 respectively determine that the signals of the first lead under the multiple first time segments are the corresponding multiple first signals, and respectively determine that the signals of the second lead set under the multiple first time segments are the corresponding multiple first signals two signals, and respectively construct a plurality of corresponding spatial filters according to the plurality of first signals and the plurality of second signals.
  • the technical solution of this embodiment respectively determines that the signals of the first lead under multiple first time segments are corresponding to multiple first time segments. For a signal, the signals of the second lead set under the plurality of first time segments are respectively determined as the corresponding plurality of second signals.
  • the signals of the first lead and the second lead set under the first time segment of the segment are respectively determined as the first signal and the second signal.
  • the segmented EEG signal is a j-segmented EEG signal
  • the signal of the first lead FP1 under the first time segment of the j-segmented EEG signal is the first signal
  • the second lead FP1 is the first signal.
  • the signals of C3 and O1 in the first time segment of the j-segmented EEG signal are the second signals.
  • the way of determining the first signal and the second signal corresponding to the other segmented EEG signals is the same as the way of determining the j-segmented EEG signal, and will not be repeated here. Therefore, for each segmented EEG signal (1 segmented EEG signal...j segmented EEG signal), the first and second signals of the first lead and the second lead under the segment can be determined. Second signal.
  • a plurality of corresponding spatial filters are respectively constructed according to the plurality of first signals and the plurality of second signals. That is, a spatial filter is separately constructed according to the first signal and the second signal under each segmented EEG signal, so that each segmented EEG signal corresponds to a separate spatial filter.
  • a spatial filter is separately constructed according to the first signal and the second signal under each segmented EEG signal, so that each segmented EEG signal corresponds to a separate spatial filter.
  • S304 Use a plurality of spatial filters to perform spatial filtering processing on the corresponding signals of the first time segment and the signals of the second time segment, respectively, to obtain an amplified signal.
  • the scheme uses a plurality of spatial filters to perform spatial filtering processing on the signal of the corresponding first time segment and the signal of the second time segment respectively, that is, using each spatial filter.
  • the device filters the first time segment signal and the second time segment signal of the corresponding segmented EEG signal, so as to obtain the amplified signal and the second time segment under the first time segment of each segmented EEG signal. Amplification signal under the fragment.
  • the multiple amplified signals corresponding to the multiple segmented EEG signals are spliced and integrated.
  • the amplification signals under different time segments can be spliced first, and when the splicing of the amplified signals under the first time segment and the second time segment of each segmented EEG signal is completed, the Each segmented EEG signal after splicing is integrated to complete the process of amplifying the EEG signal.
  • an amplified signal is obtained by filtering the EEG signal by constructing and applying a spatial filter, and further splicing and integrating the amplified signal to amplify the EEG signal. Therefore, the potential EEG information can be effectively discovered from the original EEG signal, reflecting the current EEG characteristics, realizing the purpose of amplifying the EEG signal and improving the reliability and effectiveness of the EEG information. technical effect. Furthermore, it solves the technical problems that the existing EEG signal amplification methods have limited effect and poor robustness, so they have not been widely used in the construction of actual BCI systems.
  • FIG. 4 is a schematic diagram according to another embodiment of the present application.
  • the amplification method of the EEG signal includes:
  • S401 Acquire an EEG signal, determine a first lead from multiple leads included in the EEG signal, and determine at least one second lead from the multiple leads after removing the first lead to form a second lead set , take the first lead and the second lead set as the current permutation and combination form.
  • S402 Divide the EEG signal into a plurality of segmented EEG signals, and divide the plurality of segmented EEG signals into corresponding signals of a plurality of first time segments and signals of a plurality of corresponding second time segments.
  • S403 Determining that the signals of the first lead under multiple first time segments are corresponding to multiple first signals, and respectively determining the signals of the second lead set under multiple first time segments are corresponding multiple first signals two signals, and respectively construct a plurality of corresponding spatial filters according to the plurality of first signals and the plurality of second signals.
  • S404 Use a plurality of spatial filters to perform spatial filtering processing on the corresponding signals of the first time segment and the signals of the second time segment, respectively, to obtain an amplified signal.
  • S405 splicing and integrating multiple amplified signals corresponding to the multiple segmented EEG signals to amplify the EEG signals.
  • S406 Update the current permutation and combination form, and amplify the EEG signals corresponding to the updated current permutation and combination form.
  • the present solution can also update the current permutation and combination form.
  • the above nine permutation and combination forms can be traversed, the traversed permutation and combination form can be used as the current permutation and combination form, and the operations of steps S402-S405 are performed.
  • the remaining eight permutations and combinations are traversed, such as traversing to FP1 and ⁇ C3 ⁇ .
  • Permutation and combination form FP1 and ⁇ C3 ⁇ are used as the updated current permutation and combination form.
  • the operations of S402-S405 are performed for FP1 and ⁇ C3 ⁇ , and finally the amplification of the EEG signal in the form of the arrangement and combination of FP1 and ⁇ C3 ⁇ is completed.
  • the permutation and combination forms are traversed in turn, until the EEG signals under the nine permutation and combination forms are all amplified.
  • dividing the EEG signal into a plurality of segmented EEG signals, and dividing the segmented EEG signal into a continuous first time segment signal and a second time segment signal including: using a dynamic time window to divide the EEG signal.
  • the dynamic time window is expressed as a time range centered on t in the form of [t- ⁇ t 1 , t+ ⁇ t 2 ], and [t- ⁇ t 1 , t] represents the first time segment, [t, t+ ⁇ t 2 ] represents the second time segment.
  • This scheme in the operation of dividing the EEG signal into a plurality of segmented EEG signals, and dividing the segmented EEG signal into consecutive first time segment signals and second time segment signals, This scheme can use the dynamic time window to divide the EEG signal.
  • the element t (j) represents each segmented EEG signal after division, and needs to satisfy the condition ⁇ t 1 ⁇ t (j) ⁇ T- ⁇ t 2 , the step size between the centers of different time windows is t s ;
  • j represents the number of segmented EEG signals, namely: t (1) , t (2) ... t (j) correspond to the above-mentioned 1 segment EEG in turn Signal, 2-segment EEG signal...j segmented EEG signal;
  • T represents the total duration of the EEG signal.
  • segmented EEG signals can be divided by this dynamic time window.
  • the interval [t- ⁇ t 1 , t] can be used to represent the first time segment (corresponding to segment 1 in FIG. 5 )
  • the interval [t, t+ ⁇ t 2 ] can be used to represent the second time segment Time segment (corresponding to segment 2 in Figure 5).
  • the EEG signal can be accurately divided to obtain the segmented EEG signal, and the process of dividing the segmented EEG signal into the first time segment and the second time segment is more convenient.
  • a target formula can be used to construct a spatial filter, wherein the target formula is:
  • the target formula represents the constraints of the spatial filter W j corresponding to the segmented EEG signal numbered j,
  • This scheme can use the EEG signal to construct a spatial filter, so that the constructed spatial filter can fully retain the characteristics of the EEG signal, thereby making the EEG signal amplification process more robust, and the amplified EEG The signal can better reflect the current EEG characteristics.
  • the following formula can be used: Get the amplified signal:
  • the signals under each time segment can be amplified separately, so the amplified signals can better reflect the current EEG characteristics.
  • FIG. 6 exemplarily shows a schematic diagram of the overall process of the embodiment of the present application. Referring to FIG. 6, it includes the following step:
  • the target lead corresponds to the first lead in the above embodiment
  • an indefinite number of lead sets corresponding to the above embodiment
  • the preprocessed data a number of segmented data (corresponding to the segmented EEG signal in the above-mentioned embodiment) are sequentially intercepted according to the dynamic time window, and the segmented data is divided into signals of two time segments (corresponding to the above-mentioned embodiment. first time segment and second time segment).
  • the input preprocessed data can be represented as a two-dimensional signal
  • N c and N t represent the number of EEG channels (ie, the number of leads) and the number of data points (acquisition time points), respectively, and they are both constants
  • R represents the set of real numbers.
  • the element t (j) in t (j) ⁇ needs to satisfy the condition ⁇ t 1 ⁇ t (j) ⁇ T- ⁇ t 2 , the step size between the centers of different time windows is t s , where j represents the sequence number of the segmented data, and T represents The total duration of the signal.
  • the dynamic time window can divide the segmented data into: time segment 1 signal in the interval [t- ⁇ t 1 ,t] (corresponding to the first time segment in the above embodiment), which can be represented under the jth segmented data for
  • the time segment 2 signal in the interval [t,t+ ⁇ t 2 ] (corresponding to the second time segment in the above embodiment) can be expressed as the jth segment data Among them, m and n represent the number of sampling points intercepted by the dynamic time window, see formula (1) (2) for details:
  • n [ ⁇ t 2 ⁇ F s ], n is a positive integer (2)
  • Fs is the sampling frequency of the signal
  • [x] is the rounding function, which means that it does not exceed the integer part of the real number x.
  • the construction process of the spatial filter is as follows:
  • Equation (3) represents the constraints of the spatial filter W j ,
  • U j (i,:) is the template signal, representing the signal of the target lead i under the time segment U j; is the fitted signal, representing the set of leads under the time segment U j signal, which It is a set of leads arbitrarily extracted from the remaining N c -1 leads except the target lead i, with a variable number.
  • the application process of the spatial filter is as follows:
  • ⁇ j ⁇ R 1 ⁇ m and ⁇ j ⁇ R 1 ⁇ n represent the amplified signals obtained by U j and V j , respectively, and all possible values of the two are related to the extracted lead set. The number of permutations and combinations is the same.
  • V j (i,:) is the signal of target lead i under time segment V j, is the lower lead group representing the time segment V j signal of.
  • segmented amplified signals ⁇ j ⁇ R 1 ⁇ m and ⁇ j ⁇ R 1 ⁇ n under different segmented data Traverse the segmented data under j different time windows, and obtain segmented amplified signals ⁇ j ⁇ R 1 ⁇ m and ⁇ j ⁇ R 1 ⁇ n under different segmented data.
  • the segmented amplified signals under different time windows are spliced and integrated according to the time window center point set ⁇ t (1) , t (2) ... t (j) ⁇ to obtain the target lead i, lead group new component signal under (ie, amplified signal).
  • the overlapping part can be processed by superposition average; when t s ⁇ t 1 + ⁇ t 2 , the missing part can be filled with zero or interpolated.
  • the time segment 1 in the interval [t- ⁇ t 1 , t] is a task-independent signal
  • the time segment 2 in the interval [t, t+ ⁇ t 2 ] is a task-related signal
  • splicing and summation is performed. Only ⁇ j ⁇ R 1 ⁇ n partial signals can be used for integration.
  • the target lead i and the selected lead group are reselected in turn.
  • y (n) represents the nth new component signal.
  • the set ⁇ y (1) , y (2) ... y (n) ⁇ or its subset and the original signal together form a new EEG component space N s represents the number of components in the amplified space, and its maximum possible value is determined by equation (6).
  • the present application has a wide range of applications in EEG signal processing and analysis, and has considerable practicability.
  • the obtained amplified signal is used as a new EEG component, and the original data is mapped to the new component space, which can effectively explore the potential EEG information.
  • the amplified signal is obtained by constructing and applying a dynamic spatial filter, so the obtained amplified signal can reflect the current EEG characteristics with high reliability and validity.
  • the EEG signal amplification method proposed in this application can generate a large number of amplified signals, and the obtained new component space includes the original signal and the amplified signal.
  • a more detailed description is: under the same target lead, a set of leads arbitrarily selected from the remaining leads of N c -1 with an indefinite number can produce possible combinations. Leads to a number of N c signals, produced by the method The segment amplified signals or subsets together with the original signals constitute a new EEG component space.
  • FIG. 7 is a schematic diagram according to another embodiment of the present application. As shown in FIG. 7 , a dynamic construction method of an EEG spatial filter includes the following steps:
  • Step 101 After inputting the training set data, the front and back data segments are divided according to the set time point, and the target lead is selected from the front and rear data segments;
  • the preprocessing of the training set data is realized through the above step 101 .
  • Step 102 Select the remaining leads, select a part of the leads, respectively determine the target lead and the signal of the selected lead set in the two time periods, and then use the above four parts of the signal to solve the objective function to construct a unified Model; for example: Oz is selected as the target lead, and three leads POz, Pz and FCz are selected in the initial stage to form a lead set ⁇ POz, Pz, FCz ⁇ , and the elements of this set are lead names.
  • the signal of the lead Oz and the signal of the lead set ⁇ POz, Pz, FCz ⁇ are selected respectively in the two time periods before and after, and the above totals four parts of the signal;
  • Step 103 determine whether the output of the objective function satisfies the termination condition, if yes, execute step 104, if not, execute step 102 again;
  • the establishment of the unified model is achieved through the above steps 102 and 103, and if not, the lead selection is performed again.
  • Step 104 Input the current test data and perform preprocessing according to step 101, and select the lead signal from the preprocessed test data through the lead set obtained by the unified model training; for example: if the target lead is selected as Oz, And the lead set obtained by model training is ⁇ POz, Pz, FCz ⁇ , then in the test data, the signals of the three leads of POz, Pz and FCz are similarly extracted as the lead set signal;
  • Step 105 apply the unified model to the target lead in combination with the signals selected in step 104 , dynamically solve the spatial filter suitable for the current environment, and perform spatial filtering on the test data, and the filtering is completed.
  • the signals of all trials under a stimulus condition in the training data can be represented as a three-dimensional tensor N c represents the number of leads included in the collected data, N s represents the total number of trials, and N t represents the number of sampling points of this segment of data.
  • the tensor ⁇ is divided into EEG segments with t ⁇ t 0 and the EEG segment for t>t 0 m and n are the number of data points and both are constants, and R is the set of constants.
  • the relationship between the two EEG segments X and Y is used to model and design a spatial filter, and filtering and noise reduction processing is performed for the EEG segment with t>t 0, which mainly includes the following three parts:
  • X (k) (i,:) ⁇ R 1 ⁇ m and Y (k) (i,:) ⁇ R 1 ⁇ n represent the target leads in the two time periods t ⁇ t 0 and t>t 0 respectively.
  • the signal of the kth trial of i Indicates the selected leads from the remaining leads except the target lead i
  • the signal of the kth trial is
  • Equation (1) is the spatial filter
  • p denotes the p-norm of the vector.
  • argmin means finding the variable value when the objective function takes the minimum value; on the other hand, argmax means finding the variable value when the objective function takes the maximum value.
  • Formula (2) is the solution lead set The objective function f of , whose output is a quantitative index related to the signal quality, with various specific forms, including but not limited to the spectrum, energy, signal-to-noise ratio, etc. of the characteristic signal.
  • the input ⁇ (k) ⁇ R 1 ⁇ m and ⁇ (k) ⁇ R 1 ⁇ n of the objective function are obtained by solving equations (3) and (4).
  • the spatial filter W i suitable for the current EEG environment is dynamically solved by using the formula (1) in the model G, see formula (5).
  • all the leads included in the training data can be used as target leads to perform the above steps and model, and finally the full-lead spatial filtering of the test data can be realized.
  • FIG. 8 is a schematic flowchart according to another embodiment of the present application.
  • the EEG signal amplification device 80 includes: a signal acquisition module 801 for acquiring an EEG signal, determining a first lead from a plurality of leads included in the EEG signal, and removing the first lead from the multiple leads included in the EEG signal. It is determined that at least one second lead constitutes a second lead set among the multiple leads after one lead, and the first lead and the second lead set are used as the current arrangement and combination form;
  • the signal division module 802 is used to divide the EEG signal into a plurality of segmented EEG signals, and divide the plurality of segmented EEG signals into corresponding signals of a plurality of first time segments, and a plurality of corresponding second time segments. time slice signal;
  • the filter construction module 803 is configured to respectively determine the signals of the first lead under multiple first time segments as corresponding multiple first signals, and respectively determine the signals of the second lead set under multiple first time segments for a plurality of corresponding second signals, and respectively construct a plurality of corresponding spatial filters according to the plurality of first signals and the plurality of second signals;
  • a filtering processing module 804 configured to perform spatial filtering processing on the signal of the corresponding first time segment and the signal of the second time segment respectively by using a plurality of spatial filters to obtain an amplified signal
  • the first amplification module 805 is used for splicing and integrating a plurality of amplification signals corresponding to the plurality of segmented EEG signals respectively, so as to amplify the EEG signals.
  • the apparatus 80 further includes: a second amplification module, configured to perform splicing and integration of the multiple amplified signals corresponding to the multiple segmented EEG signals respectively, so as to amplify the EEG signals, to perform a
  • the permutation and combination forms are updated, and the EEG signals corresponding to the updated current permutation and combination forms are amplified.
  • the signal division module 802 includes: a signal division sub-module for dividing the EEG signal into a plurality of segmented EEG signals by using a dynamic time window, wherein the dynamic time window is represented as a time range centered on t as: In the form of [t- ⁇ t 1 ,t+ ⁇ t 2 ], [t- ⁇ t 1 ,t] represents the first time segment, and [t,t+ ⁇ t 2 ] represents the second time segment.
  • the filter building module 803 uses a target formula to build a spatial filter, where the target formula is:
  • the target formula represents the constraints of the spatial filter W j corresponding to the segmented EEG signal numbered j,
  • the filtering processing module 804 obtains the amplified signal according to the following formula:
  • the present application further provides an electronic device and a readable storage medium.
  • FIG. 9 is a block diagram of an electronic device used to implement the method for amplifying an EEG signal according to an embodiment of the present application.
  • Electronic devices are intended to represent various forms of digital computers, such as laptops, desktops, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers.
  • Electronic devices may also represent various forms of mobile devices, such as personal digital processors, cellular phones, smart phones, wearable devices, and other similar computing devices.
  • the components shown herein, their connections and relationships, and their functions are by way of example only, and are not intended to limit implementations of the application described and/or claimed herein.
  • the device 900 includes a computing unit 901 that can be executed according to a computer program stored in a read only memory (ROM) 902 or a computer program loaded from a storage unit 908 into a random access memory (RAM) 903 Various appropriate actions and handling.
  • ROM read only memory
  • RAM random access memory
  • various programs and data necessary for the operation of the device 900 can also be stored.
  • the computing unit 901, the ROM 902, and the RAM 903 are connected to each other through a bus 904.
  • An input/output (I/O) interface 905 is also connected to bus 904 .
  • Various components in the device 900 are connected to the I/O interface 905, including: an input unit 906, such as a keyboard, mouse, etc.; an output unit 907, such as various types of displays, speakers, etc.; a storage unit 908, such as a magnetic disk, an optical disk, etc. ; and a communication unit 909, such as a network card, a modem, a wireless communication transceiver, and the like.
  • the communication unit 909 allows the device 900 to exchange information/data with other devices through a computer network such as the Internet and/or various telecommunication networks.
  • Computing unit 901 may be various general-purpose and/or special-purpose processing components with processing and computing capabilities. Some examples of computing units 901 include, but are not limited to, central processing units (CPUs), graphics processing units (GPUs), various specialized artificial intelligence (AI) computing chips, various computing units that run machine learning model algorithms, digital signal processing processor (DSP), and any suitable processor, controller, microcontroller, etc.
  • CPUs central processing units
  • GPUs graphics processing units
  • AI artificial intelligence
  • DSP digital signal processing processor
  • the computing unit 901 performs the various methods and processes described above, for example, the amplification method of the EEG signal.
  • the method of amplification of EEG signals may be implemented as a computer software program tangibly embodied on a machine-readable medium, such as storage unit 908 .
  • part or all of the computer program may be loaded and/or installed on device 900 via ROM 902 and/or communication unit 909.
  • the computing unit 901 may be configured by any other suitable means (eg, by means of firmware) to perform the amplification method of the EEG signal.
  • Various implementations of the systems and techniques described herein above can be implemented in digital electronic circuitry, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips system (SOC), load programmable logic device (CPLD), computer hardware, firmware, software, and/or combinations thereof.
  • FPGAs field programmable gate arrays
  • ASICs application specific integrated circuits
  • ASSPs application specific standard products
  • SOC systems on chips system
  • CPLD load programmable logic device
  • computer hardware firmware, software, and/or combinations thereof.
  • These various embodiments may include being implemented in one or more computer programs executable and/or interpretable on a programmable system including at least one programmable processor that
  • the processor which may be a special purpose or general-purpose programmable processor, may receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device an output device.
  • the program code for implementing the electroencephalographic signal amplification method of the present application may be written in any combination of one or more programming languages. These program codes may be provided to a processor or controller of a general purpose computer, special purpose computer, or other programmable electroencephalographic signal amplification device, such that the program code, when executed by the processor or controller, causes the flowcharts and/or block diagrams to The specified function/operation is implemented.
  • the program code may execute entirely on the machine, partly on the machine, partly on the machine and partly on a remote machine as a stand-alone software package or entirely on the remote machine or server.
  • a machine-readable medium may be a tangible medium that may contain or store the program for use by or in connection with the instruction execution system, apparatus or device.
  • the machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium.
  • Machine-readable media may include, but are not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices, or devices, or any suitable combination of the foregoing.
  • machine-readable storage media would include one or more wire-based electrical connections, portable computer disks, hard disks, random access memory (RAM), read only memory (ROM), erasable programmable read only memory (EPROM or flash memory), fiber optics, compact disk read only memory (CD-ROM), optical storage, magnetic storage, or any suitable combination of the foregoing.
  • RAM random access memory
  • ROM read only memory
  • EPROM or flash memory erasable programmable read only memory
  • CD-ROM compact disk read only memory
  • magnetic storage or any suitable combination of the foregoing.
  • the systems and techniques described herein may be implemented on a computer having a display device (eg, a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user ); and a keyboard and pointing device (eg, a mouse or trackball) through which a user can provide input to the computer.
  • a display device eg, a CRT (cathode ray tube) or LCD (liquid crystal display) monitor
  • a keyboard and pointing device eg, a mouse or trackball
  • Other kinds of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (eg, visual feedback, auditory feedback, or tactile feedback); and can be in any form (including acoustic input, voice input, or tactile input) to receive input from the user.
  • the systems and techniques described herein may be implemented on a computing system that includes back-end components (eg, as a data server), or a computing system that includes middleware components (eg, an application server), or a computing system that includes front-end components (eg, a user's computer having a graphical user interface or web browser through which a user may interact with implementations of the systems and techniques described herein), or including such backend components, middleware components, Or any combination of front-end components in a computing system.
  • the components of the system may be interconnected by any form or medium of digital data communication (eg, a communication network). Examples of communication networks include: Local Area Networks (LANs), Wide Area Networks (WANs), the Internet, and blockchain networks.
  • a computer system can include clients and servers. Clients and servers are generally remote from each other and usually interact through a communication network. The relationship of client and server arises by virtue of computer programs running on the respective computers and having a client-server relationship to each other.
  • the server can be a cloud server, also known as a cloud computing server or a cloud host. It is a host product in the cloud computing service system to solve the traditional physical host and VPS service ("Virtual Private Server", or "VPS" for short) , there are the defects of difficult management and weak business expansion.
  • the server can also be a server of a distributed system, or a server combined with a blockchain.
  • any description of a process or method in the flowcharts or otherwise described herein may be understood to represent a module, segment or portion of code comprising one or more executable instructions for implementing a specified logical function or step of the process , and the scope of the preferred embodiments of the present application includes alternative implementations in which the functions may be performed out of the order shown or discussed, including performing the functions substantially concurrently or in the reverse order depending upon the functions involved, which should It is understood by those skilled in the art to which the embodiments of the present application belong.
  • each functional unit in each embodiment of the present application may be integrated into one processing module, or each unit may exist physically alone, or two or more units may be integrated into one module.
  • the above-mentioned integrated modules can be implemented in the form of hardware, and can also be implemented in the form of software function modules. If the integrated modules are implemented in the form of software functional modules and sold or used as independent products, they may also be stored in a computer-readable storage medium.
  • the above-mentioned storage medium may be a read-only memory, a magnetic disk or an optical disk, and the like.

Landscapes

  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Engineering & Computer Science (AREA)
  • General Health & Medical Sciences (AREA)
  • Biomedical Technology (AREA)
  • Physics & Mathematics (AREA)
  • Pathology (AREA)
  • Animal Behavior & Ethology (AREA)
  • Biophysics (AREA)
  • Heart & Thoracic Surgery (AREA)
  • Medical Informatics (AREA)
  • Molecular Biology (AREA)
  • Surgery (AREA)
  • Veterinary Medicine (AREA)
  • Public Health (AREA)
  • Physiology (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Cardiology (AREA)
  • General Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Psychiatry (AREA)
  • Neurology (AREA)
  • Neurosurgery (AREA)
  • Dermatology (AREA)
  • Human Computer Interaction (AREA)
  • General Physics & Mathematics (AREA)
  • Psychology (AREA)
  • Measurement And Recording Of Electrical Phenomena And Electrical Characteristics Of The Living Body (AREA)

Abstract

本申请提供一种脑电信号的动态空间滤波与扩增方法、装置、电子设备以及存储介质，方法包括：根据脑电信号的导联信息，构建空间滤波器；以及利用空间滤波器对脑电信号进行扩增处理。从而，可以根据脑电信号构建空间滤波器，并根据空间滤波器从原有的脑电信号中有效发掘潜在的脑电信息，反映当前脑电特性，实现了对脑电信号进行扩增的目的，达到了提高脑电信息的可靠性和有效性的技术效果。

Description

脑电信号的动态空间滤波与扩增方法、装置、电子设备以及存储介质
相关申请的交叉引用
本申请要求天津大学于2020年07月24日提交中国专利局、申请号为202010728201.9、发明名称为“一种脑电空间滤波器的动态构建方法”的中国专利申请的优先权，以及天津大学于2021年04月20日提交中国专利局、申请号为202110426679.0、发明名称为“脑电信号的扩增方法、装置、电子设备以及存储介质”的中国专利申请的优先权。其全部内容通过引用结合在本申请中。
技术领域
本申请涉及计算机、脑机接口、脑功能认知状态评估、脑状态检测等多个领域，特别是指一种脑电信号的动态空间滤波与扩增方法、装置、电子设备以及存储介质。
背景技术
脑电信号(electroencephalogram,EEG)是大脑神经细胞的电生理活动在大脑皮层的总体反映，可由头皮电极进行记录(图1A示出了电极放置示意图)。脑电信号包含了大量的生理信息，在工程应用方面，EEG信号可以用来设计实现脑-机接口(Brain-Computer Interface,BCI)。脑-机接口是一种能够获取并解码人脑产生的生理信号来控制计算机或外部设备的新型人-机交互系统，可以脱离大脑正常的指令输出通路，无需经由外周神经和相关肌肉组织的传统运动控制途径。按照刺激范式的不同，BCI系统可以分为主动式、被动式和反应式三种。主动式BCI的特点是用户主动输出指令控制外部设备，以基于运动想象(Motor Imagery,MI)信号的系统为主；被动式BCI多用于检测大脑状态，如精神状态和 注意力水平等；反应式BCI主要用于检测大脑基于外部刺激的响应并间接输出控制指令，刺激诱发信号种类众多，如事件相关电位(Event-Related Potential,ERP)、稳态视觉诱发电位(Steady State Visual Evoked Potential,SSVEP)、错误相关电位(Error-Related Potential,ErrP)、事件相关去同步(Event-Related Desynchronization,ERD)等。BCI系统尤其适用于以下两种应用场景：(1)基本肢体运动功能受损、但思维正常的患者；(2)工作空间狭窄，不方便进行肢体运动(如航天环境等)。因此，BCI技术在当下越来越受到重视。
EEG信号是非平稳、时变的随机信号，而且容易受到背景活动噪声、运动伪迹、电磁噪声等的干扰。为了降低噪声干扰，提高有效信号的信噪比，大多数采集到的脑电信号通常都需要经过各种预处理之后才能进行下一步的分析：例如降采样可以减少存储压力，提高实时运算速度，同时在一定程度上抑制高频噪声的干扰；数字滤波常用来滤除或保留特定频段的信号，主要种类有低通滤波、高通滤波、带通滤波和陷波等；信号空间投影(Signal-Space Projection,SSP)常用来消除设备产生的电磁噪声和眼电干扰；独立成分分析(Independent Component Analysis,ICA)常用于分离多个非高斯的统计独立的源信号；主成分分析(Principle Component Analysis,PCA)常用于数据降维提取信号的主要特征；麦克斯韦滤波(Maxwell Filtering)与信号空间分离(Signal-Space Separation,SSS)常用于分离并去除外部源的电磁噪声(环境噪声)等等。
最近几年，脑电信号及其应用得到了愈发广泛和深入的研究，脑-机接口系统的指令集数量不断加大，信息传输率(Information Transfer Rate,ITR)逐步提高。然而目前相关研究和发展在更快响应时间和更高准确率两个方向上似乎都达到了瓶颈，其中一个很重要的原因在于前述常规的脑电数据预处理手段不足以进一步提升特征信号的质量，且现有空间滤波器的具体参数大多由训练集数据或相应的先验知识提前固定，因此无法很好地处理具有强随机性、非线性和非平稳性的各类非目标特征的脑电噪声。
此外，针对脑电信号的传统信号处理方法主要包括预处理、特征提取、模 式识别三个步骤(如图1B虚线框所示)，在实际应用中至关重要。其中信号预处理能够抑制噪声信号，有助于特征提取和分类识别，但是无法增加信号的有效成分信息。近年来，以EEG信号作为大脑信息获取方法的BCI系统发展迅速，其性能得到了不断优化。然而，脑电信号的解码仍然存在着空间分辨率低、数据量小等限制。为了有效利用已有的脑电数据，研究者们期望通过数据扩增方法来增加信号的有效信息。EEG信号是多通道的动态时间序列，不适合使用传统图像增强的几何变换方法。目前研究中的脑电信号扩增方法的作用效果有限、鲁棒性较差，因此在实际的BCI系统构建中均未能得到广泛使用。
发明内容
本申请提供了一种脑电信号的动态空间滤波与扩增方法、装置、电子设备以及存储介质，旨在至少在一定程度上解决相关技术中的技术问题之一。
根据第一方面，提供了一种脑电信号的动态空间滤波与扩增方法，包括：根据脑电信号的导联信息，构建空间滤波器；以及利用空间滤波器对脑电信号进行扩增处理。
可选地，根据脑电信号的导联信息，构建空间滤波器，包括以下步骤：
1)输入训练集数据之后按设定时间点划分前后数据段，并从前后数据段中选定目标导联；
2)对剩余导联进行抽选，抽选出一部分导联，分别确定两个时间段下的目标导联与抽选导联集合的信号，然后利用上述四部分信号求解目标函数，构建统一模型；
3)判断目标函数的输出是否满足终止条件，如果是，执行步骤4)，如果否，重新执行步骤2)；
4)输入当前的测试数据并按照步骤1)进行预处理，根据模型训练得出的导联集合从预处理后的测试数据中抽选导联信号；
5)结合步骤4)抽选的信号对测试目标导联应用统一模型，动态求解适用 于当前环境的空间滤波器并对测试数据进行空间滤波。
可选地，按设定时间点划分前后数据段具体为：根据人为设定的起始时刻t＝t 0将张量φ划分为t<t 0的脑电片段
Figure PCTCN2021105590-appb-000001
和t>t 0的脑电片段
Figure PCTCN2021105590-appb-000002
m和n分别为数据点数且均为常数，R表示常数集，N c表示采集数据包含的导联个数，N s表示总试次数。
可选地，统一模型具体为：
Figure PCTCN2021105590-appb-000003
其中，X (k)(i,:)∈R 1×m与Y (k)(i,:)∈R 1×n分别表示t<t 0和t>t 0两个时段内，目标导联i的第k个试次的信号；
Figure PCTCN2021105590-appb-000004
表示从除目标导联i以外的剩余导联中抽选的
Figure PCTCN2021105590-appb-000005
个导联组成的导联集合，
Figure PCTCN2021105590-appb-000006
与
Figure PCTCN2021105590-appb-000007
分别表示t 0前后两个时段内导联集合
Figure PCTCN2021105590-appb-000008
的第k个试次的信号；
Figure PCTCN2021105590-appb-000009
为使得函数f输出值最大的导联集合
Figure PCTCN2021105590-appb-000010
的估计；式(1)为空间滤波器
Figure PCTCN2021105590-appb-000011
的约束条件，||*|| p表示向量的p-范数，argmin表示寻找使得目标函数取最小值时的变量值；与之相对地，argmax表示寻找使得目标函数取最大值时的变量值；
Figure PCTCN2021105590-appb-000012
为使得对应p-范数输出值最小的空间滤波器
Figure PCTCN2021105590-appb-000013
的估计；式(2)为求解导联集合
Figure PCTCN2021105590-appb-000014
的目标函数f，其输出是与信号质量相关的量化指标，目标函数的输入χ (k)∈R 1×m和γ (k)∈R 1×n由式(3)、(4)求解得到。
可选地，对测试目标导联应用统一模型，动态求解适用于当前环境的空间滤波器，并对测试数据进行空间滤波，具体为：
Figure PCTCN2021105590-appb-000015
利用公式(4)与空间滤波器W i对当前测试数据进行空间滤波处理，得到降噪后的信号
Figure PCTCN2021105590-appb-000016
见式(6)，
Figure PCTCN2021105590-appb-000017
即为滤波完成之后、t>t 0时段的测试信号；
Figure PCTCN2021105590-appb-000018
其中，
Figure PCTCN2021105590-appb-000019
为使得p-范数输出值最小的空间滤波器估计，
Figure PCTCN2021105590-appb-000020
和
Figure PCTCN2021105590-appb-000021
分别为测试数据中、t<t 0时段的导联集合
Figure PCTCN2021105590-appb-000022
信号与目标导联i的单试次信号，
Figure PCTCN2021105590-appb-000023
和
Figure PCTCN2021105590-appb-000024
分别为测试数据中、t>t 0时段的导联集合
Figure PCTCN2021105590-appb-000025
信号与目标导联i的单试次信号。
可选地，根据脑电信号的导联信息，构建空间滤波器，包括：获取脑电信号，从脑电信号包含的多个导联中确定第一导联，从去除第一导联后的多个导联中确定至少一个第二导联构成第二导联集合，将第一导联和第二导联集合作为当前排列组合形式；将脑电信号划分为多个分段脑电信号，将多个分段脑电信号分别划分为对应的多个第一时间片段的信号，和对应的多个第二时间片段的信号；分别确定第一导联在多个第一时间片段下的信号为对应的多个第一信号，分别确定第二导联集合在多个第一时间片段下的信号为对应的多个第二信号，并根据多个第一信号和多个第二信号分别构建对应的多个空间滤波器。
可选地，利用空间滤波器对脑电信号进行处理，包括：利用多个空间滤波器分别对对应的第一时间片段的信号和第二时间片段的信号进行空间滤波处理，得到扩增信号；以及将多个分段脑电信号分别对应的多个扩增信号进行拼接整合，以对脑电信号进行扩增。
可选地，在将多个分段脑电信号分别对应的多个扩增信号进行拼接整合， 以对脑电信号进行扩增之后，还包括：对当前排列组合形式进行更新，并对更新后的当前排列组合形式对应的脑电信号进行扩增。
可选地，将脑电信号划分为多个分段脑电信号，将多个分段脑电信号分别划分为对应的多个第一时间片段的信号，和对应的多个第二时间片段的信号，包括：利用动态时间窗将脑电信号划分为多个分段脑电信号，其中动态时间窗表示为以t为中心的时间范围为[t-Δt 1,t+Δt 2]的形式，[t-Δt 1,t]代表第一时间片段，[t,t+Δt 2]代表第二时间片段。
可选地，采用目标公式构建空间滤波器，其中，目标公式为：
Figure PCTCN2021105590-appb-000026
目标公式，表示编号为j的分段脑电信号对应的空间滤波器W j的约束条件，||*|| p为向量的P-范数，argmin函数用于：搜索使得目标最小时的变量值，
Figure PCTCN2021105590-appb-000027
为约束条件下空间滤波器W j的估计，其中U j(i,:)为第一信号，i代表第一导联的编号；
Figure PCTCN2021105590-appb-000028
为第二信号，其中
Figure PCTCN2021105590-appb-000029
代表第二导联集合，
Figure PCTCN2021105590-appb-000030
代表编号为j的分段脑电信号在第一时间片段的信号，其中N c为脑电信号包含的多个导联的数量，m表示动态时间窗截取的采样点数，m＝[Δt 1×F s]，且m取不超过实数的整数部分，Fs为脑电信号的采样频率。
可选地，利用空间滤波器对分段脑电信号的第一时间片段的信号和第二时间片段的信号进行空间滤波处理，得到扩增信号，包括，根据以下公式得到扩增信号：
Figure PCTCN2021105590-appb-000031
Figure PCTCN2021105590-appb-000032
其中，χ j∈R 1×m和γ j∈R 1×n分别代表对U j和V j进行滤波处理得到的扩 增信号，
Figure PCTCN2021105590-appb-000033
代表编号为j的分段脑电信号在第二时间片段的信号，n表示动态时间窗截取的采样点数，n＝[Δt 2×F s]，且n取不超过实数的整数部分，V j(i,:)代表编号为i的第一导联在编号为j的分段脑电信号的第二时间片段的信号，
Figure PCTCN2021105590-appb-000034
代表编号为i的第一导联对应的第二导联集合在编号为j的分段脑电信号的第二时间片段的信号。
根据第二方面，提供了一种脑电信号的扩增装置，包括：信号获取模块，用于获取脑电信号，从脑电信号包含的多个导联中确定第一导联，从去除第一导联后的多个导联中确定至少一个第二导联构成第二导联集合，将第一导联和第二导联集合作为当前排列组合形式；信号划分模块，用于将脑电信号划分为多个分段脑电信号，将多个分段脑电信号分别划分为对应的多个第一时间片段的信号，和对应的多个第二时间片段的信号；滤波器构建模块，用于分别确定第一导联在多个第一时间片段下的信号为对应的多个第一信号，分别确定第二导联集合在多个第一时间片段下的信号为对应的多个第二信号，并根据多个第一信号和多个第二信号分别构建对应的多个空间滤波器；滤波处理模块，用于利用多个空间滤波器分别对对应的第一时间片段的信号和第二时间片段的信号进行空间滤波处理，得到扩增信号；以及第一扩增模块，用于将多个分段脑电信号分别对应的多个扩增信号进行拼接整合，以对脑电信号进行扩增。
根据第三方面，提供了一种电子设备，包括：至少一个处理器；以及与至少一个处理器通信连接的存储器；其中，存储器存储有可被至少一个处理器执行的指令，指令被至少一个处理器执行，以使至少一个处理器能够执行本申请公开的脑电信号的扩增方法。
根据第四方面，提供了一种存储有计算机指令的非瞬时计算机可读存储介质，其中，计算机指令用于使计算机执行本申请公开的脑电信号的扩增方法。
本申请提供的技术方案，通过构建和应用空间滤波器对脑电信号进行滤波处理得到扩增信号，进一步地对扩增信号进行拼接整合，以对脑电信号进行扩 增。从而，可以从原有的脑电信号中有效发掘潜在的脑电信息，反映当前脑电特性，实现了对脑电信号进行扩增的目的，达到了提高脑电信息的可靠性和有效性的技术效果。进而解决了现有的脑电信号扩增方法作用效果有限、鲁棒性较差，因此在实际的BCI系统构建中均未能得到广泛使用的技术问题。
此外，本申请提供的技术方案的有益效果还包括：
1、本申请可在脑电信号预处理过程中动态设计空间滤波器，抑制多种非目标特征的脑电噪声，应用范围广泛；
2、本申请对脑机接口的实验数据分析显示，对于单试次脑电特征信号的信噪比提升显著，并能有效提高后续特征分类的识别准确率，能进一步完善脑电信号的预处理技术，促进该技术向应用成果转化；
3、本申请可用于脑电信号的数据预处理，能有效改善采集信号的质量，改进脑机接口系统性能，有望获得可观的社会效益和经济效益。
附图说明
图1A是现有技术中的采集脑电信号的电极放置示意图；
图1B是现有技术中的脑电信号处理过程的示意图；
图1C是本申请实施例对脑电信号处理过程的示意图；
图2是根据本申请一实施例的流程示意图；
图3是根据本申请另一实施例的流程示意图；
图4是根据本申请另一实施例的流程示意图；
图5是根据本申请实施例通过动态时间窗划分脑电信号的示意图；
图6是根据本申请实施例整体流程示意图；
图7是根据本申请另一实施例的流程示意图；
图8是根据本申请另一实施例的流程示意图；以及
图9是用来实现本申请实施例的脑电信号的扩增方法的电子设备的框图。
具体实施方式
下面详细描述本申请的实施例，所述实施例的示例在附图中示出，其中自始至终相同或类似的标号表示相同或类似的元件或具有相同或类似功能的元件。下面通过参考附图描述的实施例是示例性的，仅用于解释本申请，而不能理解为对本申请的限制。相反，本申请的实施例包括落入所附加权利要求书的精神和内涵范围内的所有变化、修改和等同物。
针对背景技术中提到的现有技术中的脑电信号扩增方法作用效果有限、鲁棒性较差，因此在实际的BCI系统构建中均未能得到广泛使用的技术问题，本实施例技术方案提供了一种脑电信号的扩增方法，参考图1C所示，在对脑电信号进行特征提取之前，本申请提出的技术方案可以对信号进行扩增。下面结合具体的实施例对该方法进行说明。
其中，需要说明的是，本实施例的脑电信号的扩增方法的执行主体可以为脑电信号的扩增装置，该装置可以由软件和/或硬件的方式实现，该装置可以配置在电子设备中，电子设备可以包括但不限于终端、服务器端等。
图2是根据本申请一实施例的流程示意图，如图2所示，该脑电信号的扩增方法包括：
S201：根据脑电信号的导联信息，构建空间滤波器。
本实施例中，首先获取脑电信号的导联信息，其中，脑电信号中可以包括多个导联，例如包括FP1、C3、O1三个导联，多个导联的信息可以被称为脑电信号的导联信息。
获取导联信息后，可以根据导联信息构建空间滤波器，其中，构建空间滤波器可以理解为动态求解适用于当前环境的空间滤波器，可以采用任意可能的方式构建空间滤波器，此处不作限制。
S202：利用空间滤波器对脑电信号进行扩增处理。
上述构建空间滤波器后，可以利用空间滤波器对脑电信号进行扩增处理，可以从原有的脑电信号中有效发掘潜在的脑电信息，反映当前脑电特性，实现了对脑电信号进行扩增的目的，达到了提高脑电信息的可靠性和有效性的技术效果。
图3是根据本申请另一实施例的流程示意图。如图3所示，该脑电信号的扩增方法包括：
S301：获取脑电信号，从脑电信号包含的多个导联中确定第一导联，从去除第一导联后的多个导联中确定至少一个第二导联构成第二导联集合，将第一导联和第二导联集合作为当前排列组合形式。
具体地，本方案首先获取脑电信号，其中该脑电信号例如但不限于是癫痫脑电信号、稳态视觉诱发电位信号等。其中，该脑电信号可以是原始采集的脑电信号，还可以是经过预处理后的脑电信号。并且，脑电信号的形式可以是二维信号
Figure PCTCN2021105590-appb-000035
其中N c和N t分别表示脑电信号的通道数量(即导联数量)、采集时间点，且均为常数，R表示实数集。
进一步地，从脑电信号包含的多个导联中确定第一导联，从去除第一导联后的多个导联中确定至少一个第二导联构成第二导联集合。在一个具体实例中，以癫痫脑电信号为例，癫痫脑电信号例如包括FP1、C3、O1三个导联，本方案需要从三个导联中确定第一导联和第二导联集合。在实际操作中，可以首先确定第一导联(例如FP1)，然后从去除第一导联后剩余的导联中抽取任意数量的第二导联构成该第二导联集合，本实施例中第二导联集合包括{C3,O1}、{C3}、{O1}。
此外，本方案还需要确定脑电信号对应的当前排列组合形式，其中当前排列组合形式是第一导联和第二导联集合的排列组合形式。在上述实例中，由FP1、C3、O1三个导联构成的第一导联和第二导联集合的排列组合形式如表1所示：
表1
Figure PCTCN2021105590-appb-000036
参考表1，排列组合形式包括：FP1与{C3,O1}、FP1与{C3}、FP1与{O1}、C3与{FP1,O1}、C3与{FP1}、C3与{O1}、O1与{FP1,C3}、O1与{FP1}、O1与{C3}九种。本方案需要确定一种排列组合形式作为该当前排列组合形式，例如：当前排列组合形式为FP1与{C3,O1}。以下将以FP1与{C3,O1}作为该当前排列组合形式对本实施例继续进行说明。
应当理解的是，本方案虽然以脑电信号为例进行说明，但是本领域技术人员还可以将该信号的扩增方法应用到其他应用场景，例如心电信号。关于信号的形式，此处不作具体限定。
S302：将脑电信号划分为多个分段脑电信号，将多个分段脑电信号分别划分为对应的多个第一时间片段的信号，和对应的多个第二时间片段的信号。
具体地，在获取到脑电信号的情况下，本实施例技术方案需要将脑电信号划分为多个分段脑电信号，即：将一段脑电信号截取为多个分段脑电信号，例如将脑电信号划分为：1分段脑电信号、2分段脑电信号......j分段脑电信号，j表示分段脑电信号的编号。
进一步地，将多个分段脑电信号分别划分为对应的多个第一时间片段的信号，和对应的多个第二时间片段的信号。在实际操作中，可以从多个分段脑电 信号中任意抽取一个分段脑电信号，然后将抽取的脑电信号按照时间再次进行划分，得到第一时间片段的信号和第二时间片段的信号。其中，第一时间片段与第二时间片段的长度可以根据实际需要确定，两个时间片段的长度既可以相同，也可以不同，本方案不作具体限定。依照上述方式对每个分段脑电信号再次进行划分，划分过程此处不再赘述。从而，针对每个分段脑电信号都可以得到对应的第一时间片段的信号和第二时间片段的信号。
S303：分别确定第一导联在多个第一时间片段下的信号为对应的多个第一信号，分别确定第二导联集合在多个第一时间片段下的信号为对应的多个第二信号，并根据多个第一信号和多个第二信号分别构建对应的多个空间滤波器。
具体地，在确定第一时间片段的信号和第二时间片段的信号之后，进一步地，本实施例技术方案分别确定第一导联在多个第一时间片段下的信号为对应的多个第一信号，分别确定第二导联集合在多个第一时间片段下的信号为对应的多个第二信号。
在实际操作中，针对每个分段脑电信号，分别确定第一导联和第二导联集合在该分段的第一时间片段下的信号为第一信号和第二信号。
在一个具体实例中，例如：分段脑电信号为j分段脑电信号，则第一导联FP1在j分段脑电信号的第一时间片段下的信号为第一信号，第二导联C3和O1在该j分段脑电信号的第一时间片段下的信号为第二信号。其他分段脑电信号对应的第一信号和第二信号的确定方式同理于j分段脑电信号的确定方式，此处不在赘述。从而，针对每个分段脑电信号(1分段脑电信号...j分段脑电信号)，可以确定第一导联和第二导联在该分段下的第一信号和第二信号。
进一步地，根据多个第一信号和多个第二信号分别构建对应的多个空间滤波器。即：根据每个分段脑电信号下的第一信号和第二信号分别构建空间滤波器，从而每个分段脑电信号都对应有单独的空间滤波器。其中，本实施例技术 方案中的空间滤波器的构建原理例如可以参考现有技术中的空间滤波器构建原理，此处对于空间滤波器不作具体限定。
S304：利用多个空间滤波器分别对对应的第一时间片段的信号和第二时间片段的信号进行空间滤波处理，得到扩增信号。
进一步地，在空间滤波器构建完成的情况下，本方案利用多个空间滤波器分别对对应的第一时间片段的信号和第二时间片段的信号进行空间滤波处理，即：利用每个空间滤波器对对应的分段脑电信号的第一时间片段的信号和第二时间片段的信号进行滤波处理，从而得到每个分段脑电信号的第一时间片段下的扩增信号和第二时间片段下的扩增信号。
S305：将多个分段脑电信号分别对应的多个扩增信号进行拼接整合，以对脑电信号进行扩增。
最终，在得到扩增信号的情况下，本实施例将多个分段脑电信号分别对应的多个扩增信号进行拼接整合。在具体实现上，可以首先对不同的时间片段下的扩增信号进行拼接，在每个分段脑电信号的第一时间片段和第二时间片段下的扩增信号拼接完成的情况下，将拼接后的每个分段脑电信号进行整合，从而完成对脑电信号进行扩增的过程。
根据本实施例技术方案，通过构建和应用空间滤波器对脑电信号进行滤波处理得到扩增信号，进一步地对扩增信号进行拼接整合，以对脑电信号进行扩增。从而，可以从原有的脑电信号中有效发掘潜在的脑电信息，反映当前脑电特性，实现了对脑电信号进行扩增的目的，达到了提高脑电信息的可靠性和有效性的技术效果。进而解决了现有的脑电信号扩增方法作用效果有限、鲁棒性较差，因此在实际的BCI系统构建中均未能得到广泛使用的技术问题。
上述实施例可以实现对一种排列组合形式导联(FP1与{C3,O1})的脑电信号进行信号扩增。但是为了实现对脑电信号进行进一步地扩增，本申请还提出了第二种实施例，图4是根据本申请另一实施例的示意图。如图4所示，该脑 电信号的扩增方法包括：
S401：获取脑电信号，从脑电信号包含的多个导联中确定第一导联，从去除第一导联后的多个导联中确定至少一个第二导联构成第二导联集合，将第一导联和第二导联集合作为当前排列组合形式。
S402：将脑电信号划分为多个分段脑电信号，将多个分段脑电信号分别划分为对应的多个第一时间片段的信号，和对应的多个第二时间片段的信号。
S403：分别确定第一导联在多个第一时间片段下的信号为对应的多个第一信号，分别确定第二导联集合在多个第一时间片段下的信号为对应的多个第二信号，并根据多个第一信号和多个第二信号分别构建对应的多个空间滤波器。
S404：利用多个空间滤波器分别对对应的第一时间片段的信号和第二时间片段的信号进行空间滤波处理，得到扩增信号。
S405：将多个分段脑电信号分别对应的多个扩增信号进行拼接整合，以对脑电信号进行扩增。
S401-S405的描述说明，可以具体参见上述实施例，此处不再赘述。
S406：对当前排列组合形式进行更新，并对更新后的当前排列组合形式对应的脑电信号进行扩增。
具体地，结合上述实施例，在对当前排列组合形式(FP1与{C3,O1})完成信号扩增之后，本方案还可以对当前排列组合形式进行更新。在实际操作中，可以对上述的九种排列组合形式进行遍历，将遍历到的排列组合形式作为当前排列组合形式，并执行步骤S402-S405的操作。
在一个具体实例中，例如在对当前排列组合形式(FP1与{C3,O1})完成信号扩增的情况下，对剩余的八种排列组合形式进行遍历，比如遍历到FP1与{C3}的排列组合形式，则将FP1与{C3}作为更新后的当前排列组合形式。进一步地，针对FP1与{C3}进行S402-S405的操作，最终完成在FP1与{C3}排列组合形 式下对脑电信号的扩增。依次对排列组合形式进行遍历，直至九种排列组合形式下的脑电信号都得到信号扩增。
如果每种排列组合形式下得到的扩增信号用y (n)表示，n表示排列组合的序号，则上述的九种排列组合形式下得到的扩增信号如表2所示：
表2
Figure PCTCN2021105590-appb-000037
从而通过这种方式，可以得到更多的扩增信号，进一步地实现了对脑电信号进行扩增的目的。
可选地，将脑电信号划分为多个分段脑电信号，将分段脑电信号划分为连续的第一时间片段信号和第二时间片段信号，包括：利用动态时间窗将脑电信号划分为多个分段脑电信号，其中动态时间窗表示为以t为中心的时间范围为[t-Δt 1,t+Δt 2]的形式，[t-Δt 1,t]代表第一时间片段，[t,t+Δt 2]代表第二时间片段。
具体地，本申请第三实施例，在将脑电信号划分为多个分段脑电信号，将分段脑电信号划分为连续的第一时间片段信号和第二时间片段信号的操作中，本方案可以利用动态时间窗口对脑电信号进行划分。
在一个具体实例中，参考图5所示，该动态时间窗口例如是[t-Δt 1,t+Δt 2]，表示以t为中心的一个动态时间窗，时间窗中心点集合t＝{t (1),t (2)...t (j)}内元素t (j)代表划分后的每个分段脑电信号，并且需要满足条件Δt 1≤t (j)≤T-Δt 2，不同时间窗中心之间步长为t s；j代表分段脑电信号的编号，即：t (1),t (2)...t (j)依次 对应于上述1分段脑电信号、2分段脑电信号......j分段脑电信号；T代表脑电信号的总时长。
此外，通过该动态时间窗可以对分段脑电信号进行划分。具体地，参考图5所示，可以利用[t-Δt 1,t]区间代表第一时间分段(对应于图5中的①段)，利用[t,t+Δt 2]区间表示第二时间分段(对应于图5中的②段)。
从而通过这种方式，可以准确地对脑电信号进行划分得到分段脑电信号，并且使得对分段脑电信号划分为第一时间片段和第二时间片段的过程更加简便。
可选地，本申请技术方案在构建空间滤波器的过程中，可以采用目标公式构建空间滤波器，其中，目标公式为：
Figure PCTCN2021105590-appb-000038
目标公式，表示编号为j的分段脑电信号对应的空间滤波器W j的约束条件，||*|| p为向量的P-范数，argmin函数用于：搜索使得目标最小时的变量值，
Figure PCTCN2021105590-appb-000039
为约束条件下空间滤波器W j的估计，其中U j(i,:)为第一信号，i代表第一导联的编号；
Figure PCTCN2021105590-appb-000040
为第二信号，其中
Figure PCTCN2021105590-appb-000041
代表第二导联集合，
Figure PCTCN2021105590-appb-000042
代表编号为j的分段脑电信号在第一时间片段的信号，其中N c为脑电信号包含的多个导联的数量，m表示动态时间窗截取的采样点数，m＝[Δt 1×F s]，且m取不超过实数的整数部分，Fs为脑电信号的采样频率。
本方案可以利用脑电信号对空间滤波器进行构建，从而构建的空间滤波器更能充分保留脑电信号特征，进而使得脑电信号扩增过程鲁棒性更强，并且扩增得到的脑电信号更能反应当前脑电特性。
应当理解的是，本方案只是以上述的目标公式为例对构建空间滤波器的过 程进行解释说明，但构建该空间滤波器不限于采用上述的目标公式，还可以采用其他的公式或者其他的形式进行构建，此处不作具体限定。
可选地，本申请技术方案在利用空间滤波器对分段脑电信号的第一时间片段的信号和第二时间片段的信号进行空间滤波处理，得到扩增信号的操作中，可以根据以下公式得到扩增信号：
Figure PCTCN2021105590-appb-000043
Figure PCTCN2021105590-appb-000044
其中，χ j∈R 1×m和γ j∈R 1×n分别代表对U j和V j进行滤波处理得到的扩增信号，
Figure PCTCN2021105590-appb-000045
代表编号为j的分段脑电信号在第二时间片段的信号，n表示动态时间窗截取的采样点数，n＝[Δt 2×F s]，且n取不超过实数的整数部分，V j(i,:)代表编号为i的第一导联在编号为j的分段脑电信号的第二时间片段的信号，
Figure PCTCN2021105590-appb-000046
代表编号为i的第一导联对应的第二导联集合在编号为j的分段脑电信号的第二时间片段的信号。
从而通过这种方式，可以分别对每个时间片段下的信号进行扩增，因此扩增得到的信号更能反应当前脑电特性。
应当理解的是，本方案只是以上述的公式为例对信号的扩增过程进行解释性说明，但信号的扩增过程包括但不限于上述的方式，还可以采用其他的公式或者其他的形式进行信号扩增，此处不作具体限定。
为了对本申请技术方案进行更清楚地描述，将以一个整体流程的实施例对本方案进行进一步地解释说明，6示例性地示出了本申请实施例整体流程示意图，参考图6所示，包括以下步骤：
1)输入预处理后信号(对应于上述实施例中的脑电信号)
2)确定目标导联(该目标导联对应于上述实施例中的第一导联)，并从除目标导联外的剩余导联中任意抽取数量不定的导联集合(对应于上述实施例中的第二导联集合)。针对预处理后的数据，依据动态时间窗依次截取若干分段数据(对应于上述实施例的分段脑电信号)，将分段数据划分为两个时间片段的信号(对应于上述实施例的第一时间片段和第二时间片段)。
输入的预处理后数据可以表示为一个二维信号
Figure PCTCN2021105590-appb-000047
其中N c和N t分别表示脑电信号的通道数量(即导联数量)、数据点数(采集时间点)，且均为常数，R表示实数集。结合图5所示，时间范围[t-Δt 1,t+Δt 2]代表以t为中心的一个动态时间窗，时间窗中心点集合t＝{t (1),t (2)...t (j)}内元素t (j)需要满足条件Δt 1≤t (j)≤T-Δt 2，不同时间窗中心之间步长为t s，其中j代表分段数据的序号，T代表信号的总时长。
该动态时间窗可将分段数据划分为：[t-Δt 1,t]区间内的时间片段①信号(对应于上述实施例的第一时间片段)，在第j个分段数据下可以表示为
Figure PCTCN2021105590-appb-000048
[t,t+Δt 2]区间内的时间片段②信号(对应于上述实施例的第二时间片段)，在第j个分段数据下可以表示为
Figure PCTCN2021105590-appb-000049
其中，m、n表示动态时间窗截取的采样点数，详见式(1)(2)：
m＝[Δt 1×F s]，m为正整数          (1)
n＝[Δt 2×F s]，n为正整数           (2)
其中，Fs为信号的采样频率；[x]为取整函数，表示不超过实数x的整数部分。
3)选取任一分段数据，依据目标导联和抽选导联集合分别确定两个时间片段下的模板信号(模板信号对应于上述实施例的第一信号)和拟合信号(拟合信号对应于上述实施例的第二信号)，并利用前一时间片段的模板信号和拟合 信号构建动态空间滤波器。
空间滤波器的构建过程具体如下：
Figure PCTCN2021105590-appb-000050
式(3)表示空间滤波器W j的约束条件，||*|| p为向量的P-范数，argmin函数的目的是找到使得目标最小时的变量值，
Figure PCTCN2021105590-appb-000051
为该约束条件下空间滤波器W j的估计。
式中，U j(i,:)是模板信号，代表时间片段U j下目标导联i的信号；
Figure PCTCN2021105590-appb-000052
是拟合信号，代表时间片段U j下导联集合
Figure PCTCN2021105590-appb-000053
的信号，其中
Figure PCTCN2021105590-appb-000054
是从除目标导联i外的N c-1个剩余导联中任意抽取、数量不定的导联集合。
4)应用空间滤波器对两个时间片段信号进行空间滤波处理，分别获得该分段数据下不同时间片段的扩增信号。
空间滤波器的应用过程具体如下：
Figure PCTCN2021105590-appb-000055
Figure PCTCN2021105590-appb-000056
式(4)(5)中，χ j∈R 1×m和γ j∈R 1×n分别代表U j和V j得到的扩增信号，两者的所有可能值与抽取导联组
Figure PCTCN2021105590-appb-000057
的排列组合数量一致。其中，V j(i,:)是时间片段V j下目标导联i的信号，
Figure PCTCN2021105590-appb-000058
是代表时间片段V j下导联组
Figure PCTCN2021105590-appb-000059
的信号。
5)重复步骤3)和4)，直至所有分段数据均处理完成。对得到的所有分段扩增信号进行拼接和整合，输出该目标导联和导联组下获得的最终扩增信号。
遍历j个不同时间窗下的分段数据，获得不同分段数据下的分段扩增信号χ j∈R 1×m以及γ j∈R 1×n。对不同时间窗下的分段扩增信号按照时间窗中心点集合{t (1),t (2)...t (j)}进行拼接整合，获得目标导联i、导联组
Figure PCTCN2021105590-appb-000060
下的新成分信号
Figure PCTCN2021105590-appb-000061
(即，扩增信号)。
当t s<Δt 1+Δt 2时，对重叠部分可进行叠加平均处理；当t s≥Δt 1+Δt 2时，对于缺失部分可进行补零或插值处理。需要注意的是，若在[t-Δt 1,t]区间的时间片段①为任务无关信号，在[t,t+Δt 2]区间的时间片段②为任务相关信号，该情况下进行拼接和整合时可以仅利用γ j∈R 1×n部分信号。
重复步骤2)至5)，遍历目标导联和抽选导联组的所有排列组合，获得若干新成分信号。若干新成分信号与原有信号共同构成新的脑电成分空间。
具体地，依次重新选择目标导联i、抽选导联组
Figure PCTCN2021105590-appb-000062
重复步骤2)至5)获得若干新成分信号
Figure PCTCN2021105590-appb-000063
其中，y (n)代表第n个新成分信号。集合{y (1),y (2)...y (n)}或其子集与原有信号共同构成新的脑电成分空间
Figure PCTCN2021105590-appb-000064
N s代表扩增后空间内的成分数量，其最大可能值由式(6)决定。
Figure PCTCN2021105590-appb-000065
本申请在脑电信号处理与分析中应用范围广泛，具有可观的实用性。所得到的扩增信号作为新的脑电成分，将原有数据映射到新的成分空间，能够有效发掘潜在的脑电信息。
此外，通过构建和应用动态空间滤波器来获得扩增信号，因此所得的扩增信号能够反映当前脑电特性，具有较高的可靠性和有效性。
本申请中提出的脑电信号扩增方法能产生大量扩增信号，获得的新成分空 间中包含原有信号和扩增信号。更详细地描述是：同一个目标导联下，从N c-1个剩余导联中任意抽取、数量不定的导联集合
Figure PCTCN2021105590-appb-000066
能产生
Figure PCTCN2021105590-appb-000067
种可能组合。对导联数为N c的信号来说，该方法产生的
Figure PCTCN2021105590-appb-000068
段扩增信号或子集与原有信号共同构成了新的脑电成分空间。
图7是根据本申请另一实施例的示意图，如图7所示，一种脑电空间滤波器的动态构建方法，该方法包括以下步骤：
步骤101：输入训练集数据之后按设定时间点划分前后数据段，从前后数据段中选定目标导联；
即通过上述步骤101实现对训练集数据的预处理。
步骤102：对剩余导联进行抽选，抽选出一部分导联，分别确定两个时间段下的目标导联与抽选导联集合的信号，然后利用上述四部分信号求解目标函数，构建统一模型；例如：目标导联选为Oz，在初始阶段抽选POz、Pz和FCz三个导联组成导联集合{POz、Pz、FCz}，该集合元素为导联名称。在前后两个时间段均分别抽选导联Oz的信号与导联集合{POz、Pz、FCz}的信号，以上共计四部分信号；
步骤103：判断目标函数的输出是否满足终止条件，如果是，执行步骤104，如果否，重新执行步骤102；
即，通过上述步骤102和103实现了统一模型的建立，如果否，则重新进行导联抽选。
步骤104：输入当前的测试数据并按照步骤101进行预处理，通过统一模型训练得出的导联集合从预处理后的测试数据中抽选导联信号；例如：若目标导联选为Oz，且模型训练得出的导联集合为{POz、Pz、FCz}，则在测试数据中，同样地抽取POz、Pz和FCz这三个导联的信号作为导联集合信号；
步骤105：结合步骤104抽选的信号对目标导联应用统一模型，动态求解适用于当前环境的空间滤波器并对测试数据进行空间滤波，至此滤波完成。
综上所述，通过上述步骤101-步骤105实现了对脑电空间滤波器的动态构建，满足了实际应用中的多种需要。
下面结合具体的计算公式、实例、图7对上述方案进行进一步地扩展、细化，详见下文描述：
训练数据中某一刺激条件下所有试次的信号可以表示为一个三维张量
Figure PCTCN2021105590-appb-000069
N c表示采集数据包含的导联个数，N s表示总试次数，N t表示该段数据的采样点数。根据人为设定的起始时刻(t＝t 0)将张量φ划分为t<t 0的脑电片段
Figure PCTCN2021105590-appb-000070
和t>t 0的脑电片段
Figure PCTCN2021105590-appb-000071
m和n分别为数据点数且均为常数，R表示常数集。
本申请即利用X、Y两个脑电片段之间的关系来建模与设计空间滤波器，针对t>t 0的脑电片段进行滤波降噪处理，主要包括以下三个部分：
(1)由训练数据建立求解动态滤波器的统一模型G，详见式(1)-(4)。
Figure PCTCN2021105590-appb-000072
其中，X (k)(i,:)∈R 1×m与Y (k)(i,:)∈R 1×n分别表示t<t 0和t>t 0两个时段内，目标导联i的第k个试次的信号；
Figure PCTCN2021105590-appb-000073
表示从除目标导联i以外的剩余导联中 抽选的
Figure PCTCN2021105590-appb-000074
个导联组成的导联集合(
Figure PCTCN2021105590-appb-000075
的数值不固定)，
Figure PCTCN2021105590-appb-000076
与
Figure PCTCN2021105590-appb-000077
分别表示t 0前后两个时段内导联集合
Figure PCTCN2021105590-appb-000078
的第k个试次的信号。
式(1)为空间滤波器
Figure PCTCN2021105590-appb-000079
的约束条件，||*|| p表示向量的p-范数。argmin表示寻找使得目标函数取最小值时的变量值；与之相对地，argmax表示寻找使得目标函数取最大值时的变量值。式(2)为求解导联集合
Figure PCTCN2021105590-appb-000080
的目标函数f，其输出是与信号质量相关的量化指标，具体形式多样，包括但不限于特征信号的频谱、能量、信噪比等等。目标函数的输入χ (k)∈R 1×m和γ (k)∈R 1×n由式(3)、(4)求解得到。
(2)对测试数据进行预处理，按同样的时间点(t＝t 0)划分前后数据段得到
Figure PCTCN2021105590-appb-000081
和
Figure PCTCN2021105590-appb-000082
利用模型G训练得到的导联集合
Figure PCTCN2021105590-appb-000083
与目标导联i抽选信号
Figure PCTCN2021105590-appb-000084
和
Figure PCTCN2021105590-appb-000085
其中，
Figure PCTCN2021105590-appb-000086
分别表示t<t 0时段内，测试数据中目标导联i与导联集合
Figure PCTCN2021105590-appb-000087
的单试次信号；
Figure PCTCN2021105590-appb-000088
与
Figure PCTCN2021105590-appb-000089
分别表示t>t 0时段内，测试数据中目标导联i与导联集合
Figure PCTCN2021105590-appb-000090
的单试次信号。
利用模型G中的公式(1)动态求解适用于当前脑电环境的空间滤波器W i，见式(5)。
Figure PCTCN2021105590-appb-000091
(3)利用模型G中的公式(4)与空间滤波器W i对当前测试数据进行空间滤波处理，得到降噪后的信号
Figure PCTCN2021105590-appb-000092
见式(6)。
Figure PCTCN2021105590-appb-000093
即为滤波完成之后、t>t 0时段的测试信号。
Figure PCTCN2021105590-appb-000094
理论上训练数据中包含的所有导联都可以作为目标导联执行上述步骤并建模，最终可以实现测试数据的全导联空间滤波。
图8是根据本申请另一实施例的流程示意图。
如图8所示，该脑电信号的扩增装置80，包括：信号获取模块801，用于获取脑电信号，从脑电信号包含的多个导联中确定第一导联，从去除第一导联后的多个导联中确定至少一个第二导联构成第二导联集合，将第一导联和第二导联集合作为当前排列组合形式；
信号划分模块802，用于将脑电信号划分为多个分段脑电信号，将多个分段脑电信号分别划分为对应的多个第一时间片段的信号，和对应的多个第二时间片段的信号；
滤波器构建模块803，用于分别确定第一导联在多个第一时间片段下的信号为对应的多个第一信号，分别确定第二导联集合在多个第一时间片段下的信号为对应的多个第二信号，并根据多个第一信号和多个第二信号分别构建对应的多个空间滤波器；
滤波处理模块804，用于利用多个空间滤波器分别对对应的第一时间片段的信号和第二时间片段的信号进行空间滤波处理，得到扩增信号；以及
第一扩增模块805，用于将多个分段脑电信号分别对应的多个扩增信号进行拼接整合，以对脑电信号进行扩增。
可选地，装置80还包括：第二扩增模块，用于在将多个分段脑电信号分别对应的多个扩增信号进行拼接整合，以对脑电信号进行扩增之后，对当前排列组合形式进行更新，并对更新后的当前排列组合形式对应的脑电信号进行扩增。
可选地，信号划分模块802，包括：信号划分子模块，用于利用动态时间窗将脑电信号划分为多个分段脑电信号，其中动态时间窗表示为以t为中心的时间范围为[t-Δt 1,t+Δt 2]的形式，[t-Δt 1,t]代表第一时间片段，[t,t+Δt 2]代表第二时间片段。
可选地，滤波器构建模块803采用目标公式构建空间滤波器，其中，目标公式为：
Figure PCTCN2021105590-appb-000095
目标公式，表示编号为j的分段脑电信号对应的空间滤波器W j的约束条件，||*|| p为向量的P-范数，argmin函数用于：搜索使得目标最小时的变量值，
Figure PCTCN2021105590-appb-000096
为约束条件下空间滤波器W j的估计，其中U j(i,:)为第一信号，i代表第一导联的编号；
Figure PCTCN2021105590-appb-000097
为第二信号，其中
Figure PCTCN2021105590-appb-000098
代表第二导联集合，
Figure PCTCN2021105590-appb-000099
代表编号为j的分段脑电信号在第一时间片段的信号，其中N c为脑电信号包含的多个导联的数量，m表示动态时间窗截取的采样点数，m＝[Δt 1×F s]，且m取不超过实数的整数部分，Fs为脑电信号的采样频率。
可选地，滤波处理模块804根据以下公式得到扩增信号：
Figure PCTCN2021105590-appb-000100
Figure PCTCN2021105590-appb-000101
其中，χ j∈R 1×m和γ j∈R 1×n分别代表对U j和V j进行滤波处理得到的扩增信号，
Figure PCTCN2021105590-appb-000102
代表编号为j的分段脑电信号在第二时间片段的信号，n表 示动态时间窗截取的采样点数，n＝[Δt 2×F s]，且n取不超过实数的整数部分，V j(i,:)代表编号为i的第一导联在编号为j的分段脑电信号的第二时间片段的信号，
Figure PCTCN2021105590-appb-000103
代表编号为i的第一导联对应的第二导联集合在编号为j的分段脑电信号的第二时间片段的信号。
需要说明的是，前述对脑电信号的扩增方法的解释说明也适用于本实施例的装置，此处不再赘述。
根据本申请的实施例，本申请还提供了一种电子设备和一种可读存储介质。
图9是用来实现本申请实施例的脑电信号的扩增方法的电子设备的框图。电子设备旨在表示各种形式的数字计算机，诸如，膝上型计算机、台式计算机、工作台、个人数字助理、服务器、刀片式服务器、大型计算机、和其它适合的计算机。电子设备还可以表示各种形式的移动装置，诸如，个人数字处理、蜂窝电话、智能电话、可穿戴设备和其它类似的计算装置。本文所示的部件、它们的连接和关系、以及它们的功能仅仅作为示例，并且不意在限制本文中描述的和/或者要求的本申请的实现。
如图9所示，设备900包括计算单元901，其可以根据存储在只读存储器(ROM)902中的计算机程序或者从存储单元908加载到随机访问存储器(RAM)903中的计算机程序，来执行各种适当的动作和处理。在RAM 903中，还可存储设备900操作所需的各种程序和数据。计算单元901、ROM 902以及RAM 903通过总线904彼此相连。输入/输出(I/O)接口905也连接至总线904。
设备900中的多个部件连接至I/O接口905，包括：输入单元906，例如键盘、鼠标等；输出单元907，例如各种类型的显示器、扬声器等；存储单元908，例如磁盘、光盘等；以及通信单元909，例如网卡、调制解调器、无线通信收发机等。通信单元909允许设备900通过诸如因特网的计算机网络和/或各种电信网络与其他设备交换信息/数据。
计算单元901可以是各种具有处理和计算能力的通用和/或专用处理组件。计算单元901的一些示例包括但不限于中央处理单元(CPU)、图形处理单元(GPU)、各种专用的人工智能(AI)计算芯片、各种运行机器学习模型算法的计算单元、数字信号处理器(DSP)、以及任何适当的处理器、控制器、微控制器等。计算单元901执行上文所描述的各个方法和处理，例如，脑电信号的扩增方法。
例如，在一些实施例中，脑电信号的扩增方法可被实现为计算机软件程序，其被有形地包含于机器可读介质，例如存储单元908。在一些实施例中，计算机程序的部分或者全部可以经由ROM 902和/或通信单元909而被载入和/或安装到设备900上。当计算机程序加载到RAM 903并由计算单元901执行时，可以执行上文描述的脑电信号的扩增方法的一个或多个步骤。备选地，在其他实施例中，计算单元901可以通过其他任何适当的方式(例如，借助于固件)而被配置为执行脑电信号的扩增方法。
本文中以上描述的系统和技术的各种实施方式可以在数字电子电路系统、集成电路系统、现场可编程门阵列(FPGA)、专用集成电路(ASIC)、专用标准产品(ASSP)、芯片上系统的系统(SOC)、负载可编程逻辑设备(CPLD)、计算机硬件、固件、软件、和/或它们的组合中实现。这些各种实施方式可以包括：实施在一个或者多个计算机程序中，该一个或者多个计算机程序可在包括至少一个可编程处理器的可编程系统上执行和/或解释，该可编程处理器可以是专用或者通用可编程处理器，可以从存储系统、至少一个输入装置、和至少一个输出装置接收数据和指令，并且将数据和指令传输至该存储系统、该至少一个输入装置、和该至少一个输出装置。
用于实施本申请的脑电信号的扩增方法的程序代码可以采用一个或多个编程语言的任何组合来编写。这些程序代码可以提供给通用计算机、专用计算机或其他可编程脑电信号的扩增装置的处理器或控制器，使得程序代码当由处理器或控制器执行时使流程图和/或框图中所规定的功能/操作被实施。程序代码 可以完全在机器上执行、部分地在机器上执行，作为独立软件包部分地在机器上执行且部分地在远程机器上执行或完全在远程机器或服务器上执行。
在本申请的上下文中，机器可读介质可以是有形的介质，其可以包含或存储以供指令执行系统、装置或设备使用或与指令执行系统、装置或设备结合地使用的程序。机器可读介质可以是机器可读信号介质或机器可读储存介质。机器可读介质可以包括但不限于电子的、磁性的、光学的、电磁的、红外的、或半导体系统、装置或设备，或者上述内容的任何合适组合。机器可读存储介质的更具体示例会包括基于一个或多个线的电气连接、便携式计算机盘、硬盘、随机存取存储器(RAM)、只读存储器(ROM)、可擦除可编程只读存储器(EPROM或快闪存储器)、光纤、便捷式紧凑盘只读存储器(CD-ROM)、光学储存设备、磁储存设备、或上述内容的任何合适组合。
为了提供与用户的交互，可以在计算机上实施此处描述的系统和技术，该计算机具有：用于向用户显示信息的显示装置(例如，CRT(阴极射线管)或者LCD(液晶显示器)监视器)；以及键盘和指向装置(例如，鼠标或者轨迹球)，用户可以通过该键盘和该指向装置来将输入提供给计算机。其它种类的装置还可以用于提供与用户的交互；例如，提供给用户的反馈可以是任何形式的传感反馈(例如，视觉反馈、听觉反馈、或者触觉反馈)；并且可以用任何形式(包括声输入、语音输入或者、触觉输入)来接收来自用户的输入。
可以将此处描述的系统和技术实施在包括后台部件的计算系统(例如，作为数据服务器)、或者包括中间件部件的计算系统(例如，应用服务器)、或者包括前端部件的计算系统(例如，具有图形用户界面或者网络浏览器的用户计算机，用户可以通过该图形用户界面或者该网络浏览器来与此处描述的系统和技术的实施方式交互)、或者包括这种后台部件、中间件部件、或者前端部件的任何组合的计算系统中。可以通过任何形式或者介质的数字数据通信(例如，通信网络)来将系统的部件相互连接。通信网络的示例包括：局域网(LAN)、广域网(WAN)、互联网及区块链网络。
计算机系统可以包括客户端和服务器。客户端和服务器一般远离彼此并且通常通过通信网络进行交互。通过在相应的计算机上运行并且彼此具有客户端-服务器关系的计算机程序来产生客户端和服务器的关系。服务器可以是云服务器，又称为云计算服务器或云主机，是云计算服务体系中的一项主机产品，以解决了传统物理主机与VPS服务("Virtual Private Server"，或简称"VPS")中，存在的管理难度大，业务扩展性弱的缺陷。服务器也可以为分布式系统的服务器，或者是结合了区块链的服务器。
需要说明的是，在本申请的描述中，术语“第一”、“第二”等仅用于描述目的，而不能理解为指示或暗示相对重要性。此外，在本申请的描述中，除非另有说明，“多个”的含义是两个或两个以上。
流程图中或在此以其他方式描述的任何过程或方法描述可以被理解为，表示包括一个或更多个用于实现特定逻辑功能或过程的步骤的可执行指令的代码的模块、片段或部分，并且本申请的优选实施方式的范围包括另外的实现，其中可以不按所示出或讨论的顺序，包括根据所涉及的功能按基本同时的方式或按相反的顺序，来执行功能，这应被本申请的实施例所属技术领域的技术人员所理解。
应当理解，本申请的各部分可以用硬件、软件、固件或它们的组合来实现。在上述实施方式中，多个步骤或方法可以用存储在存储器中且由合适的指令执行系统执行的软件或固件来实现。例如，如果用硬件来实现，和在另一实施方式中一样，可用本领域公知的下列技术中的任一项或他们的组合来实现：具有用于对数据信号实现逻辑功能的逻辑门电路的离散逻辑电路，具有合适的组合逻辑门电路的专用集成电路，可编程门阵列(PGA)，现场可编程门阵列(FPGA)等。
本技术领域的普通技术人员可以理解实现上述实施例方法携带的全部或部分步骤是可以通过程序来指令相关的硬件完成，所述的程序可以存储于一种计算机可读存储介质中，该程序在执行时，包括方法实施例的步骤之一或其组合。
此外，在本申请各个实施例中的各功能单元可以集成在一个处理模块中，也可以是各个单元单独物理存在，也可以两个或两个以上单元集成在一个模块中。上述集成的模块既可以采用硬件的形式实现，也可以采用软件功能模块的形式实现。所述集成的模块如果以软件功能模块的形式实现并作为独立的产品销售或使用时，也可以存储在一个计算机可读取存储介质中。
上述提到的存储介质可以是只读存储器，磁盘或光盘等。
在本说明书的描述中，参考术语“一个实施例”、“一些实施例”、“示例”、“具体示例”、或“一些示例”等的描述意指结合该实施例或示例描述的具体特征、结构、材料或者特点包含于本申请的至少一个实施例或示例中。在本说明书中，对上述术语的示意性表述不一定指的是相同的实施例或示例。而且，描述的具体特征、结构、材料或者特点可以在任何的一个或多个实施例或示例中以合适的方式结合。
尽管上面已经示出和描述了本申请的实施例，可以理解的是，上述实施例是示例性的，不能理解为对本申请的限制，本领域的普通技术人员在本申请的范围内可以对上述实施例进行变化、修改、替换和变型。

Claims (16)

  1. 一种脑电信号的动态空间滤波与扩增方法，其特征在于，所述方法包括：
    根据脑电信号的导联信息，构建空间滤波器；以及
    利用所述空间滤波器对所述脑电信号进行扩增处理。
  2. 如权利要求1所述的方法，其特征在于，根据脑电信号的导联信息，构建空间滤波器，包括以下步骤：
    1)输入训练集数据之后按设定时间点划分前后数据段，并从前后数据段中选定目标导联；
    2)对剩余导联进行抽选，抽选出一部分导联，分别确定两个时间段下的目标导联与抽选导联集合的信号，然后利用上述四部分信号求解目标函数，构建统一模型；
    3)判断目标函数的输出是否满足终止条件，如果是，执行步骤4)，如果否，重新执行步骤2)；
    4)输入当前的测试数据并按照步骤1)进行预处理，根据模型训练得出的导联集合从预处理后的测试数据中抽选导联信号；
    5)结合步骤4)抽选的信号对测试目标导联应用统一模型，动态求解适用于当前环境的空间滤波器并对测试数据进行空间滤波。
  3. 如权利要求2所述的方法，其特征在于，所述按设定时间点划分前后数据段具体为：
    根据人为设定的起始时刻t＝t 0将张量φ划分为t<t 0的脑电片段
    Figure PCTCN2021105590-appb-100001
    和t>t 0的脑电片段
    Figure PCTCN2021105590-appb-100002
    m和n分别为数据点数且均为常数，R表示常数集，N c表示采集数据包含的导联个数，N s表示总试次数。
  4. 如权利要求3所述的方法，其特征在于，其特征在于，所述统一模型具体为：
    Figure PCTCN2021105590-appb-100003
    其中，X (k)(i,:)∈R 1×m与Y (k)(i,:)∈R 1×n分别表示t<t 0和t>t 0两个时段内，目标导联i的第k个试次的信号；
    Figure PCTCN2021105590-appb-100004
    表示从除目标导联i以外的剩余导联中抽选的
    Figure PCTCN2021105590-appb-100005
    个导联组成的导联集合，
    Figure PCTCN2021105590-appb-100006
    与
    Figure PCTCN2021105590-appb-100007
    分别表示t 0前后两个时段内导联集合
    Figure PCTCN2021105590-appb-100008
    的第k个试次的信号；
    Figure PCTCN2021105590-appb-100009
    为使得函数f输出值最大的导联集合
    Figure PCTCN2021105590-appb-100010
    的估计；
    式(1)为空间滤波器
    Figure PCTCN2021105590-appb-100011
    的约束条件，||*|| p表示向量的p-范数，arg min表示寻找使得目标函数取最小值时的变量值；与之相对地，arg max表示寻找使得目标函数取最大值时的变量值；
    Figure PCTCN2021105590-appb-100012
    为使得对应p-范数输出值最小的空间滤波器
    Figure PCTCN2021105590-appb-100013
    的估计；
    式(2)为求解导联集合
    Figure PCTCN2021105590-appb-100014
    的目标函数f，其输出是与信号质量相关的量化指标，目标函数的输入χ (k)∈R 1×m和γ (k)∈R 1×n由式(3)、(4)求解得到。
  5. 如权利要求4所述的方法，其特征在于，所述对测试目标导联应用统一模型，动态求解适用于当前环境的空间滤波器，并对测试数据进行空间滤波，具体为：
    Figure PCTCN2021105590-appb-100015
    利用公式(4)与空间滤波器W i对当前测试数据进行空间滤波处理，得到降噪后的信号
    Figure PCTCN2021105590-appb-100016
    见式(6)，
    Figure PCTCN2021105590-appb-100017
    即为滤波完成之后、t>t 0时段的测试信号；
    Figure PCTCN2021105590-appb-100018
    其中，
    Figure PCTCN2021105590-appb-100019
    为使得p-范数输出值最小的空间滤波器估计，
    Figure PCTCN2021105590-appb-100020
    和
    Figure PCTCN2021105590-appb-100021
    分别为测试数据中、t<t 0时段的导联集合
    Figure PCTCN2021105590-appb-100022
    信号与目标导联i的单试次信号，
    Figure PCTCN2021105590-appb-100023
    和
    Figure PCTCN2021105590-appb-100024
    分别为测试数据中、t>t 0时段的导联集合
    Figure PCTCN2021105590-appb-100025
    信号与目标导联i的单试次信号。
  6. 如权利要求1所述的方法，其特征在于，根据脑电信号的导联信息，构建空间滤波器，包括：
    获取脑电信号，从所述脑电信号包含的多个导联中确定第一导联，从去除所述第一导联后的所述多个导联中确定至少一个第二导联构成第二导联集合，将所述第一导联和第二导联集合作为当前排列组合形式；
    将所述脑电信号划分为多个分段脑电信号，将所述多个分段脑电信号分别划分为对应的多个第一时间片段的信号，和对应的多个第二时间片段的信号；
    分别确定所述第一导联在所述多个第一时间片段下的信号为对应的多个第一信号，分别确定所述第二导联集合在所述多个第一时间片段下的信号为对应的多个第二信号，并根据所述多个第一信号和所述多个第二信号分别构建对应的多个空间滤波器。
  7. 如权利要求6所述的方法，其特征在于，利用所述空间滤波器对所述脑电信号进行处理，包括：
    利用所述多个空间滤波器分别对对应的所述第一时间片段的信号和所述第二时间片段的信号进行空间滤波处理，得到扩增信号；以及
    将所述多个分段脑电信号分别对应的多个扩增信号进行拼接整合，以对所 述脑电信号进行扩增。
  8. 如权利要求7所述的方法，其特征在于，在所述将所述多个分段脑电信号分别对应的多个扩增信号进行拼接整合，以对所述脑电信号进行扩增之后，还包括：
    对所述当前排列组合形式进行更新，并对更新后的当前排列组合形式对应的脑电信号进行扩增。
  9. 如权利要求7所述的方法，其特征在于，将所述脑电信号划分为多个分段脑电信号，将所述多个分段脑电信号分别划分为对应的多个第一时间片段的信号，和对应的多个第二时间片段的信号，包括：
    利用动态时间窗将所述脑电信号划分为多个分段脑电信号，其中所述动态时间窗表示为以t为中心的时间范围为[t-Δt 1,t+Δt 2]的形式，[t-Δt 1,t]代表第一时间片段，[t,t+Δt 2]代表第二时间片段。
  10. 如权利要求9所述的方法，其特征在于，采用目标公式构建所述空间滤波器，其中，所述目标公式为：
    Figure PCTCN2021105590-appb-100026
    所述目标公式，表示编号为j的分段脑电信号对应的空间滤波器W j的约束条件，||*|| p为向量的P-范数，argmin函数用于：搜索使得目标最小时的变量值，
    Figure PCTCN2021105590-appb-100027
    为所述约束条件下空间滤波器W j的估计，其中U j(i,:)为所述第一信号，i代表所述第一导联的编号；
    Figure PCTCN2021105590-appb-100028
    为所述第二信号，其中
    Figure PCTCN2021105590-appb-100029
    代表所述第二导联集合，
    Figure PCTCN2021105590-appb-100030
    代表编号为j的分段脑电信号在第一时间片段的信号，其中N c为所述脑电信号包含的多个导联的数量，m表示所述动态时间窗截取的采样点数，m＝[Δt 1×F s]，且m取不超过实数的整数部分，Fs为所述脑电信号的采样频率。
  11. 如权利要求10所述的方法，其特征在于，利用所述空间滤波器对所述分段脑电信号的第一时间片段的信号和第二时间片段的信号进行空间滤波处理，得到扩增信号，包括，根据以下公式得到所述扩增信号：
    Figure PCTCN2021105590-appb-100031
    Figure PCTCN2021105590-appb-100032
    其中，χ j∈R 1×m和γ j∈R 1×n分别代表对U j和V j进行滤波处理得到的扩增信号，
    Figure PCTCN2021105590-appb-100033
    代表编号为j的分段脑电信号在第二时间片段的信号，n表示所述动态时间窗截取的采样点数，n＝[Δt 2×F s]，且n取不超过实数的整数部分，V j(i,:)代表编号为i的第一导联在编号为j的分段脑电信号的第二时间片段的信号，
    Figure PCTCN2021105590-appb-100034
    代表编号为i的第一导联对应的第二导联集合在编号为j的分段脑电信号的第二时间片段的信号。
  12. 一种脑电信号的扩增装置，包括：
    信号获取模块，用于从所述脑电信号包含的多个导联中确定第一导联，从去除所述第一导联后的所述多个导联中确定至少一个第二导联构成第二导联集合，将所述第一导联和第二导联集合作为当前排列组合形式；
    信号划分模块，用于将所述脑电信号划分为多个分段脑电信号，将所述多个分段脑电信号分别划分为对应的多个第一时间片段的信号，和对应的多个第二时间片段的信号；
    滤波器构建模块，用于分别确定所述第一导联在所述多个第一时间片段下的信号为对应的多个第一信号，分别确定所述第二导联集合在所述多个第一时间片段下的信号为对应的多个第二信号，并根据所述多个第一信号和所述多个第二信号分别构建对应的多个空间滤波器；
    滤波处理模块，用于利用所述多个空间滤波器分别对对应的所述第一时间片段的信号和所述第二时间片段的信号进行空间滤波处理，得到扩增信号；以 及
    第一扩增模块，用于将所述多个分段脑电信号分别对应的多个扩增信号进行拼接整合，以对所述脑电信号进行扩增。
  13. 如权利要求12所述的脑电信号的扩增装置，其特征在于，还包括：第二扩增模块，用于在所述将所述多个分段脑电信号分别对应的多个扩增信号进行拼接整合，以对所述脑电信号进行扩增之后，对所述当前排列组合形式进行更新，并对更新后的当前排列组合形式对应的脑电信号进行扩增。
  14. 如权利要求12所述的脑电信号的扩增装置，其特征在于，所述信号划分模块，包括：
    信号划分子模块，用于利用动态时间窗将所述脑电信号划分为多个分段脑电信号，其中所述动态时间窗表示为以t为中心的时间范围为[t-Δt 1,t+Δt 2]的形式，[t-Δt 1,t]代表第一时间片段，[t,t+Δt 2]代表第二时间片段。
  15. 一种电子设备，包括：
    至少一个处理器；以及
    与所述至少一个处理器通信连接的存储器；其中，
    所述存储器存储有可被所述至少一个处理器执行的指令，所述指令被所述至少一个处理器执行，以使所述至少一个处理器能够执行权利要求1-11中任一项所述的方法。
  16. 一种存储有计算机指令的非瞬时计算机可读存储介质，其中，所述计算机指令用于使所述计算机执行根据权利要求1-11中任一项所述的方法。
PCT/CN2021/105590 2020-07-24 2021-07-09 脑电信号的动态空间滤波与扩增方法、装置、电子设备以及存储介质 Ceased WO2022017202A1 (zh)

Priority Applications (1)

Application Number Priority Date Filing Date Title
US17/904,790 US20230055867A1 (en) 2020-07-24 2021-07-09 Method and apparatus for performing spatial filtering and augmenting electroencephalogram signal, electronic device, and storage medium

Applications Claiming Priority (4)

Application Number Priority Date Filing Date Title
CN202010728201.9A CN111820876B (zh) 2020-07-24 2020-07-24 一种脑电空间滤波器的动态构建方法
CN202010728201.9 2020-07-24
CN202110426679.0 2021-04-20
CN202110426679.0A CN113509188B (zh) 2021-04-20 2021-04-20 脑电信号的扩增方法、装置、电子设备以及存储介质

Publications (1)

Publication Number Publication Date
WO2022017202A1 true WO2022017202A1 (zh) 2022-01-27

Family

ID=79728519

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2021/105590 Ceased WO2022017202A1 (zh) 2020-07-24 2021-07-09 脑电信号的动态空间滤波与扩增方法、装置、电子设备以及存储介质

Country Status (2)

Country Link
US (1) US20230055867A1 (zh)
WO (1) WO2022017202A1 (zh)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2022241578A1 (en) * 2021-05-21 2022-11-24 Interaxon Inc. Systems and methods for neural networks and dynamic spatial filters to reweigh channels
WO2023184771A1 (zh) * 2022-04-01 2023-10-05 之江实验室 一种基于位点等效增强的脑机接口解码方法及装置

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116595455B (zh) * 2023-05-30 2023-11-10 江南大学 基于时空频特征提取的运动想象脑电信号分类方法及系统
CN120234059B (zh) * 2025-03-21 2026-03-17 普译生物科技(绍兴)有限公司 多通道数据显示方法、装置和存储介质

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20120059273A1 (en) * 2010-09-03 2012-03-08 Faculdades Catolicas, a nonprofit association, Maintainer of the Pontificia Universidade Cotolica Process and device for brain computer interface
CN102722255A (zh) * 2012-06-29 2012-10-10 上海海事大学 基于神经反馈的运动想象脑-机接口交互训练系统及方法
CN102715911A (zh) * 2012-06-15 2012-10-10 天津大学 基于脑电特征的情绪状态识别方法
CN111820876A (zh) * 2020-07-24 2020-10-27 天津大学 一种脑电空间滤波器的动态构建方法

Family Cites Families (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US10531806B2 (en) * 2013-12-17 2020-01-14 University Of Florida Research Foundation, Inc. Brain state advisory system using calibrated metrics and optimal time-series decomposition
WO2017205734A1 (en) * 2016-05-26 2017-11-30 University Of Washington Reducing sensor noise in multichannel arrays using oversampled temporal projection and associated systems and methods
WO2020042511A1 (zh) * 2018-08-28 2020-03-05 天津大学 基于空间滤波与模版匹配的运动电位脑机接口编解码方法

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20120059273A1 (en) * 2010-09-03 2012-03-08 Faculdades Catolicas, a nonprofit association, Maintainer of the Pontificia Universidade Cotolica Process and device for brain computer interface
CN102715911A (zh) * 2012-06-15 2012-10-10 天津大学 基于脑电特征的情绪状态识别方法
CN102722255A (zh) * 2012-06-29 2012-10-10 上海海事大学 基于神经反馈的运动想象脑-机接口交互训练系统及方法
CN111820876A (zh) * 2020-07-24 2020-10-27 天津大学 一种脑电空间滤波器的动态构建方法

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
CHENG LONGLONG, QIU SHUANG;XU RUI;XU MINPENG;YI WEIBO;MING DONG;QI HONGZHI;WAN BAIKUN: "Research on EEG Feature Enhancement Based on Extreme Energy Difference and Common Spatial Pattern Algorithms", GAOJISHU TONGXUN - HIGH TECHNOLOGY LETTERS, BEIJING, CN, vol. 23, no. 9, 15 September 2013 (2013-09-15), CN , pages 900 - 907, XP055889171, ISSN: 1002-0470, DOI: 10.3772/j.issn.1002-0470.2013.09.016 *

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2022241578A1 (en) * 2021-05-21 2022-11-24 Interaxon Inc. Systems and methods for neural networks and dynamic spatial filters to reweigh channels
WO2023184771A1 (zh) * 2022-04-01 2023-10-05 之江实验室 一种基于位点等效增强的脑机接口解码方法及装置

Also Published As

Publication number Publication date
US20230055867A1 (en) 2023-02-23

Similar Documents

Publication Publication Date Title
WO2022017202A1 (zh) 脑电信号的动态空间滤波与扩增方法、装置、电子设备以及存储介质
Deng et al. TRCA-Net: using TRCA filters to boost the SSVEP classification with convolutional neural network
CN111797804A (zh) 一种基于深度学习的信道状态信息人类活动识别方法及系统
WO2022160676A1 (zh) 热力图生成模型的训练方法、装置、电子设备和存储介质
CN112183477A (zh) 一种基于持续同调的脑电信号持续特征提取方法
CN111820876B (zh) 一种脑电空间滤波器的动态构建方法
CN114424940B (zh) 基于多模态时空特征融合的情绪识别方法及系统
CN104503592A (zh) 头部动作确定方法和装置
CN107958214A (zh) Ecg信号的并行分析装置、方法和移动终端
WO2021159571A1 (zh) 一种有向动态脑功能网络多类情绪识别构建方法及其装置
CN116035593B (zh) 一种基于生成对抗式并行神经网络的脑电降噪方法
CN115363599A (zh) 一种用于心房颤动识别的心电信号处理方法及系统
CN110477865A (zh) 一种癫痫发作检测装置、终端设备及存储介质
CN108470182B (zh) 一种用于非对称脑电特征增强与识别的脑-机接口方法
CN116196017B (zh) 脑电信号分类模型训练方法、脑电信号分类方法及装置
CN118986286A (zh) 基于图结构的睡眠分期方法和装置、电子设备及存储介质
CN114795247A (zh) 脑电信号分析方法、装置、电子设备和存储介质
CN113509188B (zh) 脑电信号的扩增方法、装置、电子设备以及存储介质
CN117493974A (zh) 基于时空特征的TSMNet网络及SSVEP分类方法
CN114529945B (zh) 一种情感识别方法、装置、设备及存储介质
Hag et al. A wearable single EEG channel analysis for mental stress state detection
CN112990025B (zh) 用于处理数据的方法、装置、设备以及存储介质
CN112150463A (zh) 用于确定黄斑中心凹位置的方法及装置
CN115691799A (zh) 猝死风险预测模型的训练方法、猝死风险预测方法和装置
CN111427450A (zh) 一种情绪识别的方法、系统、设备及可读存储介质

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 21846104

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 21846104

Country of ref document: EP

Kind code of ref document: A1