Detection method and device for ssvep-based brain-computer interface stimulation paradigm

Abstract

This invention provides a method and apparatus for detecting a brain-computer interface stimulation paradigm based on SSVEP, comprising: generating a brain-computer interface stimulation paradigm based on SSVEP using peripheral vision; acquiring training EEG signals and detection EEG signals from a acquisition channel based on the SSVEP-based brain-computer interface stimulation paradigm; processing the training EEG signals and detection EEG signals to obtain reference signals and detection signals, respectively; determining all gaze patterns of the SSVEP-based brain-computer interface stimulation paradigm, and processing the reference signals to obtain a distinction between different gaze patterns; determining the current gaze pattern of the detection signals based on the reference signals and detection signals, and obtaining the current gaze target based on the current gaze pattern of the detection signals. This invention can achieve higher decoding accuracy and shorter sampling time in application scenarios that do not require much control command output, meaning it achieves high accuracy and fast transmission rate in detecting the current gaze target.

Description

Technical Field

[0001] This invention relates to brain-computer interface technology, and in particular to a method and apparatus for detecting brain-computer interface stimulation paradigms based on SSVEP. Background Technology

[0002] Brain-computer interface (BCI), also known as "brain port" or brain-computer fusion sensing, is a technology that establishes a direct connection between the human or animal brain and an external device, enabling the brain to directly input control commands to the external device. Through this channel, people can express thoughts or manipulate devices directly through their brains without the need for language or movement. This can effectively enhance the ability of patients with severe physical disabilities to communicate with the outside world or control their external environment, thereby improving their quality of life.

[0003] Research on brain-computer interfaces (BCIs) has spanned over 40 years. Based on years of animal experiments, several BCI devices have been designed to assist in the rehabilitation of disabilities caused by congenital or acquired hearing, visual, and motor impairments, and have seen initial applications on a small scale. With continued research progress, the application potential of BCIs is no longer limited to medical rehabilitation; BCI technology also shows great promise in areas such as smart homes and human-computer interaction. More and more companies and research institutions are now turning their attention to the BCI field. The working process of a BCI system can be summarized as follows: acquiring the user's brainwave signals, translating the brainwave signals to generate external commands, and feeding back the external commands to the user.

[0004] Electroencephalography (EEG) is a signal obtained by amplifying the spontaneous bioelectric potentials of the brain through the scalp using sophisticated electronic instruments. It records the spontaneous, rhythmic electrical activity of brain cell groups via electrodes. In the EEG signal acquisition stage of a brain-computer interface (BCI) system, the non-invasiveness and high precision of EEG signals make it the most favored EEG signal method for BCI systems. EEG has extremely wide applications in the medical field, especially in disease monitoring and diagnosis. Because of the extremely high accuracy of EEG recordings of human brain activity, and the availability of many low-cost, portable EEG signal acquisition methods, EEG signals have become the primary EEG signal paradigm in the field of BCI research.

[0005] EEG signal acquisition devices can be broadly categorized into two types based on electrode type. One is the wet electrode acquisition method, which requires applying a conductive gel to the subject's scalp to effectively reduce scalp resistance and facilitate EEG signal acquisition. The other is the dry electrode acquisition method, which uses highly conductive, tentacle-like conductors attached to the electrodes to penetrate hair and increase the contact area between the electrodes and the scalp, thereby reducing resistance at the scalp-electrode contact point. EEG signal acquisition devices obtain EEG signals through evoked potentials.

[0006] There are three main types of evoked potentials: visual evoked potentials, auditory evoked potentials, and tactile evoked potentials. Visual evoked potentials are widely used in the study of electroencephalogram (EEG) signals due to their simplicity and convenience. When the visual system receives stimuli such as light or flashing graphics, the potentials of the EEG signals change; these changes in potential are called visual evoked potentials (VEPs).

[0007] Visual evoked potentials include steady-state visual evoked potentials (SSVEPs), which are EEG signals based on steady-state visual evoked potentials (SSVEPs). These signals are generated when a subject fixates on a visual stimulus that flashes at a specific frequency (e.g., when the subject fixates on a visual stimulus that flashes at a specific frequency). Figure 1 As shown in the figure, the brain will produce a response with the same flashing frequency and corresponding high harmonics. These responses are stable throughout the entire period of the subject's gaze at the stimulus. However, since this type of EEG signal requires external visual stimulation to be generated, prolonged gaze at these flashing visual stimuli can easily cause high visual fatigue for the user. Moreover, previous SSVEP-BCI stimulation paradigms usually focus more on how to further improve the signal-to-noise ratio, with the aim of making the EEG signal characteristics acquired under the new stimulation paradigm more obvious and easier to process.

[0008] It should be noted that the above introduction to the technical background is only for the purpose of providing a clear and complete explanation of the technical solutions of this application and facilitating understanding by those skilled in the art. It should not be assumed that these technical solutions are known to those skilled in the art simply because they have been described in the background section of this application. Summary of the Invention

[0009] In view of the shortcomings of the prior art described above, the purpose of this invention is to provide a method and device for detecting brain-computer interface stimulation paradigms based on SSVEP, in order to solve the problem that prolonged viewing of these flashing visual stimuli in the prior art can easily cause high visual fatigue for users.

[0010] To achieve the above and other related objectives, this invention provides a method for detecting brain-computer interface stimulation paradigms based on SSVEP, comprising at least the following steps:

[0011] Generate a brain-computer interface stimulation paradigm based on SSVEP using peripheral vision;

[0012] A brain-computer interface stimulation paradigm based on SSVEP was used to acquire training EEG signals and detect EEG signals in the acquisition channel.

[0013] The training EEG signal is processed to obtain a reference signal, and the detection EEG signal is processed to obtain a detection signal;

[0014] All gaze patterns of the SSVEP-based brain-computer interface stimulation paradigm are determined, and the reference signal is processed to obtain a way to distinguish different gaze patterns;

[0015] The current gaze pattern of the detection signal is determined based on the reference signal and the detection signal, and the current gaze target is obtained based on the current gaze pattern of the detection signal.

[0016] Preferably, generating a brain-computer interface stimulation paradigm based on SSVEP using peripheral vision includes the following steps:

[0017] Set a certain number of stimulus targets and distribute all stimulus targets according to the set rules;

[0018] Identify the flashing target and its set flashing frequency among the distributed stimulus targets;

[0019] Peripheral vision uses the blinking of the blinking target to identify the non-blinking target.

[0020] Preferably, the setting rule is a uniform distribution in a matrix manner.

[0021] Preferably, the distribution of the stimulus targets is as follows:

[0022] The stimulus targets, which are uniformly distributed in the matrix, are divided into rows of flashing targets and rows of non-flickering targets; the rows of flashing targets and the rows of non-flickering targets are arranged adjacent to each other.

[0023] The flashing target row includes flashing targets and non-flashing targets, and the flashing targets and non-flashing targets are arranged adjacent to each other; the stimulation targets in the non-flashing target row are all non-flashing targets; the flashing targets in different flashing rows are located in the same column.

[0024] Preferably, there are 40 stimulus targets, of which 12 are identified as flashing targets and 28 are identified as non-flashing targets.

[0025] Preferably, all gaze patterns are determined based on the user's location and the number of flashing targets in the peripheral vision.

[0026] Preferably, distance similarity is used to distinguish gaze patterns.

[0027] Preferably, Manhattan distance similarity is used to process the detection signal and the reference signal to determine the current gaze pattern of the detection signal.

[0028] Preferably, the current gaze target is obtained by processing the detection signal of the current gaze mode using a nonlinear correlator.

[0029] To achieve the above and other related objectives, the present invention also provides a detection device for a brain-computer interface stimulation paradigm based on SSVEP, comprising a memory, a processor, and a program stored in the memory and executable on the processor. When the processor executes the program, it implements the steps of the above-described detection method for a brain-computer interface stimulation paradigm based on SSVEP.

[0030] As described above, the detection method and apparatus of the brain-computer interface stimulation paradigm based on SSVEP of the present invention have the following beneficial effects:

[0031] The passive brain-computer interface (BCI) stimulation paradigm detection method proposed in this invention, based on SSVEP, provides a more comfortable user experience while maintaining a certain signal-to-noise ratio for EEG signals. This invention also simultaneously removes redundant flashing stimuli in both the lateral and longitudinal directions of the SSVEP BCI, ensuring that several flashing targets exist near each non-flickering target. This removal of a large amount of redundant flashing stimuli reduces the flashing area, effectively reducing visual fatigue experienced by users when using the SSVEP BCI. Furthermore, the technique of determining the gaze pattern of the detected EEG signal before determining the current gaze target reduces the time required to determine the current gaze target, thus supporting scenarios requiring high information transmission rates, such as brain-computer typing. This invention can achieve higher decoding accuracy and shorter sampling time in other application scenarios that do not require much control command output, meaning higher accuracy and faster transmission rates. Attached Figure Description

[0032] Figure 1 This diagram illustrates a brain-computer interface stimulation paradigm based on SSVEP in the prior art.

[0033] Figure 2 The diagram shows the detection method of the brain-computer interface stimulation paradigm based on SSVEP according to the present invention.

[0034] Figure 3 The diagram shows the structure of the brain-computer interface stimulation paradigm based on SSVEP of this invention.

[0035] Figure 4 The diagram shows the identification of the brain-computer interface stimulation paradigm based on SSVEP of this invention.

[0036] Figure 5The diagram shows the generation process of the brain-computer interface stimulation paradigm based on SSVEP according to the present invention.

[0037] Figure 6 The diagram shows the electrode channels selected according to the brain-computer interface stimulation paradigm based on SSVEP of this invention.

[0038] Figure 7 The diagram shows all gaze patterns of the brain-computer interface stimulation paradigm based on SSVEP of this invention.

[0039] Figure 8 The diagram shows the signals corresponding to all gaze patterns in the brain-computer interface stimulation paradigm based on SSVEP of this invention. Detailed Implementation

[0040] The following specific examples illustrate the implementation of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention.

[0041] Please see Figure 2-8 It should be noted that the illustrations provided in this embodiment are only schematic representations of the basic concept of the present invention. Therefore, the drawings only show the components related to the present invention and are not drawn according to the actual number, shape and size of the components in the actual implementation. In the actual implementation, the form, quantity and proportion of each component can be arbitrarily changed, and the layout of the components may also be more complex.

[0042] EEG signals based on steady-state visual evoked potentials (SSVEP) are similar to sine and cosine waves at specific frequencies. When these discrete EEG signals are subjected to discrete Fourier transform (DFT), their characteristics are that stable amplitude values ​​at the flicker frequency and multiple harmonics are significantly higher than other frequency components. These EEG signals have a high signal-to-noise ratio and more stable characteristics, resulting in high accuracy of the acquired EEG signals.

[0043] The main technical concepts of this invention are twofold: first, to design a more comfortable, high signal-to-noise ratio (SNR) brain-computer interface stimulation paradigm based on SSVEP. This paradigm can effectively reduce the flickering stimulation area of ​​SSVEP while ensuring that the EEG signals acquired under this paradigm still have a high SNR; second, to establish a reasonable signal-to-noise model under this stimulation paradigm, and to improve the signal enhancement algorithm under this SNR model. Finally, to optimize the detection algorithm of SSVEP EEG signals to form a complete system, further improving the information transmission efficiency of SSVEP-BCI.

[0044] Based on the above technical concept, this invention proposes a detection method and device for a brain-computer interface stimulation paradigm based on SSVEP. Before detailing the technical concept of this invention, it should be noted that the detection process of this invention involves two stages when the user uses the brain-computer interface proposed in this invention:

[0045] The first stage is called the training stage. In this stage, the user needs to fixate on each target for a few seconds and repeat the fixation on all targets about 10 times. The computer will calibrate the user's EEG data in this stage and complete the training process of the Event Correlation Analysis (TRCA) algorithm.

[0046] The second stage is the normal user stage. In this stage, when the flashing target is flashing, the user can choose to look at the non-flashing target corresponding to the control command or character they want to output. Each time the flashing stops, the computer will output the corresponding control command or character, which is the target to be looked at.

[0047] Figure 2 The diagram shows a flowchart of the brain-computer interface stimulation paradigm detection method based on SSVEP proposed in this invention. The following is a summary of the method. Figure 2 This invention provides a detailed description of the detection method for the brain-computer interface stimulation paradigm based on SSVEP, the detection method comprising:

[0048] S1 utilizes peripheral vision to generate a brain-computer interface stimulation paradigm based on SSVEP;

[0049] This step aims to construct a new paradigm for brain-computer interface stimulation, such as... Figure 3 As shown, this is done to reduce the area of ​​stimulation.

[0050] The generation process of the brain-computer interface stimulation paradigm based on SSVEP in this invention is as follows: Figure 5 As shown, it includes the following steps:

[0051] S11, Set a certain number of stimulus targets and distribute all stimulus targets according to the set rules;

[0052] In this embodiment of the invention, 40 stimulus targets are presented on the display, and the rule is set to be a uniform distribution in a matrix manner. The 40 stimulus targets are uniformly distributed in a matrix manner, and the 40 stimulus targets are numbered from 1 to 40 (e.g., ...). Figure 4 As shown in the diagram, the "+" indicates the center of the stimulus target. The 40 stimulus targets can be used for typing or as control commands for other devices.

[0053] As another implementation, the spatial shapes of the 40 different stimulus targets can be marked according to the user's usage habits, such as the 26 English letters and 14 commonly used symbols in the prior art; and the number of stimulus targets is not limited to 40.

[0054] S12, determine the flashing target and its set flashing frequency among the distributed stimulus targets;

[0055] This step aims to reduce the area of ​​flashing stimulation, thereby identifying flashing targets within the stimulation target group; during actual stimulation, different flashing targets flash at different time frequencies. Therefore, the stimulation targets are divided into flashing targets and non-flashing targets.

[0056] In this invention, the distribution of the uniformly distributed stimulus targets in a matrix manner is as follows:

[0057] The stimulus targets, which are uniformly distributed in the matrix, are divided into rows of flashing targets and rows of non-flickering targets; the rows of flashing targets and the rows of non-flickering targets are arranged adjacent to each other.

[0058] The flashing target row includes flashing targets and non-flashing targets, and the flashing targets and non-flashing targets are arranged adjacent to each other; the stimulation targets in the non-flashing target row are all non-flashing targets; the flashing targets in different flashing rows are located in the same column.

[0059] In this embodiment of the invention, 12 of the 40 stimulus targets are identified as flashing targets and 28 are identified as non-flickering targets. The matrix has 5 rows, with 8 stimulus targets in each row. The first, third, and fifth rows are flashing target rows, and the second and fourth rows are non-flickering target rows. The first, third, fifth, and seventh columns of all flashing target rows are flashing targets, and the second, third, sixth, and eighth columns are non-flickering targets. The set flashing frequency distribution of the 12 stimulus targets identified as flashing targets is 8Hz, 8.4Hz, 8.8Hz, 9.2Hz, 11.2Hz, 11.6Hz, 12Hz, 12.4Hz, 14.4Hz, 14.8Hz, 15.2Hz, and 15.6Hz.

[0060] In this embodiment of the invention, the flashing targets are uniformly distributed in a matrix manner, thereby ensuring the uniformity of the flashing light from each flashing target.

[0061] S13, peripheral vision uses the blinking of the blinking target to identify the non-blinking target.

[0062] In this embodiment of the invention, the 28 non-flashing targets are the locations that need to be focused on, and will be identified using the peripheral vision of the 12 adjacent circular flashing stimuli. The non-flashing targets are positioned in the middle of the flashing targets, so that the non-flashing targets can be identified using the peripheral vision of the flashing targets.

[0063] The SSVEP-based brain-computer interface stimulation paradigm developed in this invention reduces target flicker by approximately two-thirds. This paradigm reduces visual fatigue caused by prolonged fixation on flickering stimuli by removing a large amount of redundant flickering stimuli. This paradigm can effectively improve the usability of brain-computer interface paradigms.

[0064] S2, based on the SSVEP brain-computer interface stimulation paradigm, acquires training EEG signals and detects EEG signals in the acquisition channel;

[0065] This invention does not impose specific limitations on the devices used to acquire EEG signals. Currently, most mainstream EEG signal acquisition devices on the market can acquire EEG signals that meet the signal-to-noise ratio requirements of the paradigm proposed in this invention. In this embodiment, the Neuroscan SynAmps2 64-256 channel EEG amplifier is used. This device has 64 channels and can acquire signals from 64 electrodes. Up to four devices can be connected in parallel to acquire EEG signals from 256 electrodes.

[0066] In use, the subject wears a data acquisition headgear, with 64 electrodes corresponding to 64 acquisition points on the scalp, allowing for the acquisition of EEG signals from 64 acquisition channels. In the paradigm proposed in this invention, only electrode channels corresponding to vision-related brain regions are selected. When testing the usability of the invention, we selected... Figure 6 The electrode channel scheme shown is as follows: Figure 6 The head section contains 64 circles representing 64 acquisition points. The solid circles are the electrode channels, or acquisition channels, selected in this embodiment of the invention. EEG signals are acquired through these acquisition channels.

[0067] During the acquisition or detection of training EEG signals, the subject needs to repeatedly gaze at a non-flashing target m times. Each time, there is a 3-second prompt period during which the user first selects the non-flashing target to gaze at; after 3 seconds, a flashing target will begin to flash for 1-5 seconds. During this period, the user needs to concentrate on the previously selected non-flashing target so that the various acquisition channels can collect EEG signals.

[0068] Preferably, the acquired EEG signals are bandpass filtered to effectively suppress high-frequency components. In one embodiment of the present invention, a 5-45Hz FIR bandpass filter is used for bandpass filtering.

[0069] In this embodiment of the invention, n is 11. The EEG signal is input through the headgear and entered into an amplifier for initial processing before being imported into a processing computer for further processing.

[0070] Specifically, the training EEG signal is the EEG signal acquired in the first stage (training stage); the detection EEG signal is the EEG signal acquired in the second stage (normal use stage); as another implementation method, the detection EEG signal can also be randomly selected from the training stage, that is, a large part of the training EEG signal is used as the training EEG signal for training, and another small part is used as the detection signal for verification and comparison.

[0071] It should be noted that although in this embodiment of the invention, the detection of EEG signals is obtained in step S2, in reality, the detection of EEG signals can also be obtained after training is completed, that is, when the user uses the device normally.

[0072] S3, the training EEG signal is processed to obtain a reference signal, and the detection EEG signal is processed to obtain a detection signal;

[0073] This step aims to minimize the linear differences between channels after superimposing training or detection EEG signals on the electrode channels.

[0074] The Task Related Component Analysis (TRCA) algorithm is used to obtain the weight coefficients of each acquisition channel.

[0075] TRCA is an analytical algorithm for detecting task-related components. This algorithm assumes that the differences between different channels on the EEG electrode cap lie in the varying signal amplitudes of their responses to the same flickering stimulus, as well as the different background noise levels in each channel. This allows for the establishment of a mathematical model for noise suppression of the acquired EEG signals. For the same EEG signal acquisition device, the weighting coefficients of each channel remain constant, allowing for further applications.

[0076] After obtaining the weighting coefficients for each acquisition channel, the product of the mean of the m EEG signals acquired by each acquisition channel and the corresponding weighting coefficient for that channel is used as the detection signal. In other words, TRCA is used to find a set of weighting coefficients that minimize the linear differences between channels in the detection signal obtained after superimposing these acquisition channels.

[0077] The present invention will further elaborate on the above-mentioned process of processing training EEG signals using task correlation analysis to obtain reference signals and processing detection EEG signals using task correlation analysis to obtain detection signals;

[0078] The correlation analysis algorithm was used to calculate the correlation of the EEG signals, and the expression for the highest correlation is as follows:

[0079]

[0080]

[0081] Among them, X i X represents the EEG signal acquired during the i-th experiment in the first phase. j The EEG signal collected in the j-th experiment during the first phase, Cov(X) i X j W represents the covariance of the EEG signal. i W represents the weighting coefficient corresponding to the i-th experiment. j is the weight coefficient corresponding to the j-th repeated experiment.

[0082] The weighting coefficients of each acquisition channel are obtained by solving the expression with the highest correlation of the EEG signals.

[0083] The weighting coefficients of each acquisition channel are calculated using the following formula. :

[0084]

[0085] when The special vector corresponding to the largest eigenvalue is the weight coefficient we are looking for. .

[0086] This invention can effectively enhance signals with minimal training, and at the detection end, it can perform high-precision classification of SSVEP signals through simple correlation analysis. Specifically, it employs signal enhancement algorithms represented by Event Correlation Analysis (TRCA), and beamforming demonstrates powerful performance in MIMIO communication systems, improving the signal-to-noise ratio by appropriately weighting and superimposing signals from multiple channels.

[0087] The optimal weighting system for multiple electrode channels has been obtained using the TRCA algorithm. The reference signal with the maximum signal-to-noise ratio can be obtained using the following formula:

[0088]

[0089] in, It trains brain signals.

[0090] S4, determine all gaze patterns of the SSVEP-based brain-computer interface stimulation paradigm, and process the reference signal to obtain the differentiation method of different gaze patterns;

[0091] This step aims to process the reference signal obtained after merging the training EEG signals from multiple acquisition channels to obtain all gaze targets.

[0092] All gaze patterns are determined based on the user's location and the number of flashing targets in the peripheral vision. For the results of this invention, based on the uniform distribution of the matrix of stimulus targets in the stimulus paradigm, the gaze patterns, influenced by the user's location, fall into two categories: one is gazing at flashing targets (such as...). Figure 7 (a) Another type is focusing on non-flashing targets (such as...) Figure 7 (b) Figure 7 (c) Figure 7 (d) Figure 7 (e) and Figure 7 (f)); further, under the influence of the number of flashing targets in the peripheral vision, focusing on a flashing target includes having one flashing target in the peripheral vision (e.g., Figure 7 (a)); focusing on a non-flickering target includes having a flickering target within peripheral vision (e.g., Figure 7 (c) There are two flashing targets in the peripheral vision (e.g.) Figure 7 (b) such as Figure 7 (d) and Figure 7 (f) There are four flashing targets in the peripheral vision (such as...) Figure 7 (e)); Therefore, the gaze patterns of the brain-computer interface stimulation paradigm based on SSVEP in this invention are of 6 types.

[0093] The method for distinguishing different gaze patterns by processing the reference signal includes: processing and analyzing the reference signal to obtain analysis results, and determining the gaze pattern based on the analysis results;

[0094] Performing a Fourier transform on the reference signal yields the following result: Figure 8 The signal diagrams of the six gaze patterns shown can be analyzed to reveal that the reference signals obtained by combining and processing training EEG signals have significant amplitude differences among these six gaze patterns. Therefore, this invention uses distance similarity to distinguish gaze patterns.

[0095] S5, determine the current gaze pattern of the detection signal based on the reference signal and the detection signal, and obtain the current gaze target based on the current gaze pattern of the detection signal.

[0096] Under the same gaze pattern, the main difference in EEG signals when a user gazes at different stimuli is in frequency. By using a nonlinear correlation coefficient to determine which stimulus the user is specifically gazing at, high-precision decoding of the user's EEG signal (detection signal) can be achieved, significantly improving the decoding accuracy of weak-feature EEG signals under user-friendly brain-computer interface stimulation paradigms. Therefore, this invention first determines the current gaze pattern of the detection signal, enabling faster and more accurate analysis and judgment under the current gaze pattern, thus obtaining the current gaze target more efficiently.

[0097] The current gaze pattern of the detected signal is determined based on the reference signal and the detected signal. Specifically, the Manhattan distance similarity is used to process the detected signal and the reference signal. When the Manhattan distance similarity between the detected signal and a certain reference signal is the smallest, it means that the gaze patterns corresponding to the detected signal and the reference signal are the same. At this time, the gaze pattern corresponding to the reference signal is the current gaze pattern of the detected signal.

[0098] Manhattan distance similarity is represented as:

[0099]

[0100] The current gaze mode is determined as follows:

[0101]

[0102] In the formula, This is the superimposed average of all training EEG signals corresponding to the m-th stimulus target. The detected EEG signal corresponds to the currently gazed target. To calculate the Manhattan distance between two signals; τ represents the current gaze mode.

[0103] After determining the current gaze pattern of the detected signal, the current gaze target is obtained based on the current gaze pattern of the detected signal. Specifically, the gaze target is obtained by processing the detected signal with the current gaze pattern using a nonlinear correlator; when the Manhattan distance similarity between the detected signal and a certain reference signal is the smallest, it means that the gaze pattern corresponding to the detected signal and the reference signal is the same, and at this time the gaze pattern corresponding to the reference signal is the current gaze pattern of the detected signal.

[0104] The correlation coefficient of the Spearman nonlinear correlation classifier is calculated as follows:

[0105]

[0106] The current gaze target is determined as follows:

[0107]

[0108] in, Indicates the user's current gaze target under the current gaze mode. This is the superimposed average of all training EEG signals under the current gaze mode. To calculate the correlation coefficient of a nonlinear correlation classifier (Spearman) between two sets of signals. The user's current gaze target.

[0109] This invention employs a combination of distance similarity decoding and nonlinear correlation decoding to achieve high-precision decoding of weak EEG signals.

[0110] Device Example:

[0111] The present invention also provides a detection device for a brain-computer interface stimulation paradigm based on SSVEP, comprising a memory, a processor, and a program stored in the memory and executable on the processor. When the processor executes the program, it implements the steps of the above-described detection method for a brain-computer interface stimulation paradigm based on SSVEP.

[0112] The detailed steps of the method for detecting brain-computer interface stimulation paradigms based on SSVEP have been described in detail in the method embodiments and will not be repeated here.

[0113] In summary, the detection method and apparatus for the SSVEP-based brain-computer interface stimulation paradigm of this invention first removes a large number of redundant flickering stimuli, reducing the flickering area. Because the SSVEP-based brain-computer interface stimulation paradigm of this invention reduces the dependence of the brain-computer interface on external stimuli, it can effectively reduce visual fatigue experienced by users when using the SSVEP brain-computer interface. When the signal sampling window length is set to 4.6 seconds, the decoding accuracy of the EEG signal under this stimulation paradigm reaches 77.12%. For a brain-computer interface system capable of outputting 40 control commands, such decoding accuracy is highly usable. Furthermore, the technique of first determining the gaze pattern of the detected EEG signal and then determining the current gaze target reduces the time spent judging the current gaze target, thus supporting scenarios requiring high information transmission rates, such as brain-computer typing. This means that in other application scenarios that do not require so many control command outputs, the solution of this invention can achieve higher decoding accuracy and shorter sampling time. Therefore, this invention effectively overcomes the various shortcomings of the prior art and has high industrial applicability.

[0114] The above embodiments are merely illustrative of the principles and effects of the present invention and are not intended to limit the invention. Any person skilled in the art can modify or alter the above embodiments without departing from the spirit and scope of the present invention. Therefore, all equivalent modifications or alterations made by those skilled in the art without departing from the spirit and technical concept disclosed in the present invention should still be covered by the claims of the present invention.

Claims

1. A method for detecting brain-computer interface stimulation paradigms based on SSVEP, characterized in that, At least the following steps are included: A brain-computer interface stimulation paradigm based on SSVEP is generated using peripheral vision; including: setting a certain number of stimulation targets and distributing all stimulation targets according to a set rule; identifying flashing targets and their set flashing frequencies among the distributed stimulation targets; and using peripheral vision to identify non-flickering targets by means of the flashing of the flashing targets. The setting rule is a uniform distribution in a matrix manner; the distribution method of the stimulus targets is as follows: the uniformly distributed stimulus targets in the matrix are divided into rows of flashing targets and rows of non-flickering targets; the rows of flashing targets and the rows of non-flickering targets are arranged adjacent to each other; wherein, the rows of flashing targets include flashing targets and non-flickering targets, and the flashing targets and the non-flickering targets are arranged adjacent to each other; the stimulus targets in the rows of non-flickering targets are all non-flickering targets; the flashing targets in different flashing rows are located in the same column; A brain-computer interface stimulation paradigm based on SSVEP was used to acquire training EEG signals and detect EEG signals in the acquisition channel. Task-related component analysis is performed on the training EEG signal and the detection EEG signal to obtain the weight coefficients of each acquisition channel. Based on the weight coefficients, the training EEG signal is processed to obtain a reference signal, and the detection EEG signal is processed to obtain a detection signal. All gaze patterns of the brain-computer interface stimulation paradigm based on SSVEP are determined, and the reference signal is processed to obtain the differentiation method of different gaze patterns. The reference signal and the detection signal are processed based on the Manhattan distance similarity to determine the current gaze pattern of the detection signal, and the detection signal of the current gaze pattern is processed based on the nonlinear correlator to obtain the current gaze target.

2. The method for detecting brain-computer interface stimulation paradigms based on SSVEP according to claim 1, characterized in that, There were 40 stimulus targets, of which 12 were identified as flashing targets and 28 as non-flashing targets.

3. The method for detecting brain-computer interface stimulation paradigms based on SSVEP according to claim 1, characterized in that, All gaze patterns are determined based on the user's location and the number of flashing targets in their peripheral vision.

4. A detection device based on the SSVEP brain-computer interface stimulation paradigm, characterized in that, The method includes a memory, a processor, and a program stored in the memory and executable on the processor, wherein when the processor executes the program, it implements the steps of the detection method for the brain-computer interface stimulation paradigm based on any one of claims 1-3.

Images