Feedback method and system for brain wave recognition of patient intention based on brain-computer interface

Abstract

The application discloses a feedback method and system for patient intention brain wave recognition based on a brain-computer interface, comprising collecting electroencephalogram signals and associated data of a patient brain-computer interface, and preprocessing the electroencephalogram signals and the associated data; performing mode classification on the electroencephalogram signals to obtain mode data, and performing intention correction on the mode data by using the associated data to obtain corrected data; performing electroencephalogram signal recognition on the corrected data to obtain recognition data, and constructing a patient intention brain wave recognition feedback model according to the recognition data; optimizing the patient intention brain wave recognition feedback model according to a feedback error, inputting to-be-fed-back data into the optimized patient intention brain wave recognition feedback model, and outputting a feedback result. The method can not only improve the accuracy of feedback of patient intention brain wave recognition based on a brain-computer interface, but also has good interpretability, and can be directly applied to a patient intention brain wave recognition feedback system.

Description

Technical Field

[0001] This invention relates to the field of intent recognition, and more particularly to a feedback method and system for recognizing patient intent brainwaves based on brain-computer interfaces. Background Technology

[0002] With the rapid development of neuroscience and information technology, brain-computer interface (BCI) technology has become a new bridge connecting the human brain with external devices or systems, showing great application potential, especially in medical rehabilitation, assisted communication, and neuroscience research. Traditionally, effectively expressing the wishes and intentions of paralyzed patients, those with speech disorders, and those with cognitive impairments has been a major challenge. Therefore, developing technologies that can accurately identify the intentions contained in a patient's electroencephalogram (EEG) signals and provide real-time feedback accordingly is particularly important.

[0003] Currently, brain-computer interface technology primarily decodes an individual's intentions by collecting and analyzing electrical signals generated by the cerebral cortex (i.e., electroencephalogram, EEG). However, EEG signals are complex and variable, susceptible to noise interference, and exhibit significant differences between individuals, all of which increase the difficulty of intention recognition. Existing EEG signal recognition methods often focus on the extraction and classification of signal features, but frequently neglect the importance of related data such as the external environment, emotional state, and historical behavior associated with the intention, resulting in limited recognition accuracy and generalization ability.

[0004] Furthermore, most brain-computer interface systems lack effective feedback mechanisms to optimize and adjust the recognition model, making it difficult for the system to maintain long-term stability and accuracy in the face of complex and ever-changing real-world application scenarios. Therefore, developing a method that can comprehensively consider EEG signals and related data, and continuously optimize the recognition model through feedback mechanisms to achieve high-precision and high-stability patient intention EEG recognition has become a current research hotspot and challenge. Summary of the Invention

[0005] The purpose of this invention is to provide a feedback method for recognizing patient intentions via brainwaves based on a brain-computer interface.

[0006] To achieve the above objectives, the present invention is implemented according to the following technical solution:

[0007] This invention includes the following steps:

[0008] Collect the patient's brain-computer interface EEG signals and related data, and preprocess the EEG signals and related data;

[0009] Pattern data is obtained by performing pattern classification on the electroencephalogram (EEG) signals, and intention correction is performed on the pattern data using the associated data to obtain corrected data.

[0010] The corrected data is subjected to EEG signal recognition to obtain recognition data, and a patient intention EEG recognition feedback model is constructed based on the recognition data.

[0011] The patient intention brainwave recognition feedback model is optimized based on the feedback error. The data to be fed back is input into the optimized patient intention brainwave recognition feedback model, and the feedback result is output.

[0012] Furthermore, a method for obtaining pattern data by pattern classification of the electroencephalogram (EEG) signals includes:

[0013] Add particles as a search tool for EEG signals, and update the search velocity and position of the particles. The expression is:

[0014]

[0015] The local optimal position of the i-th particle in the t-th iteration is: The global optimal position of the i-th particle in the t-th iteration is The position of the i-th particle in the t-th iteration is The search speed of the i-th particle in the (t+1)-th iteration is The search speed of the i-th particle in the t-th iteration is The position of the i-th particle in the (t+1)-th iteration is The random numbers from 0 to 1 are τ1 and τ2, the acceleration factors are λ1 and λ2 respectively, and the inertia weight factor is η;

[0016] Calculate the similarity between EEG signals:

[0017]

[0018] Among them, the first One EEG signal is The m-th EEG signal is x m EEG signals and EEG signals x m The similarity is EEG signals The probability density is EEG signal x m The probability density is EEG signals and EEG signals x m The weighting factor is

[0019] The EEG signals with the highest similarity are clustered to obtain initial clusters, and the overall silhouette coefficient of the clusters is calculated:

[0020]

[0021] The average distance between the m-th EEG signal and the EEG signals within the initial cluster is l (m), and the average distance between the m-th EEG signal and the EEG signal points between the initial clusters is... The number of EEG signals is

[0022] Calculate the penalty item:

[0023]

[0024] The number of EEG signals with negative silhouette coefficients is N. ng The average distance between the ι-th EEG signal and the EEG signals within the initial cluster is l(ι), and the average distance between the ι-th EEG signal and the EEG signals between the initial clusters is...

[0025] Construct the fitness function, with the following expression:

[0026]

[0027] Where the fitness function is The actual number of clusters is A, the predicted number of clusters is X, the convolution operation is *, and the cutoff distance is φ. o Indicator function is EEG signals and EEG signals x m The distance is No. The center of each cluster is EEG signals and cluster center The distance is EEG signals and cluster center ε m The distance between them is Membership function is

[0028] The initial cluster is updated through continuous iteration of the particle population until the fitness function value is maximized, at which point the iteration stops and the updated initial cluster is output as model data.

[0029] Furthermore, the method for obtaining corrected data by using the associated data to perform intent correction on the pattern data includes:

[0030] Correlation analysis is performed on the associated data to obtain the patient's potential behaviors, and the pattern data is corrected using the potential behaviors to obtain corrected data.

[0031] Furthermore, a method for obtaining recognition data by performing EEG signal recognition on the corrected data includes:

[0032] The profile space and detail space of the corrected data are decomposed step by step. The profile space is transformed by the first conjugate orthogonal mirror filter bank and the second conjugate orthogonal mirror filter bank, respectively, to obtain the transformed profile space and detail space. The expressions are as follows:

[0033]

[0034] β[n]=(-1) 1-n h[1-n]

[0035] The j-th generalization space is E. j The (j+1)th detail space is C j+1 The (j+1)th transformation profile space is E j+1 The first conjugate orthogonal mirror filter group (1-n) is h[1-n], the first conjugate orthogonal mirror filter group is h[n], and the second conjugate orthogonal mirror filter group is β[n].

[0036] The family of functions is defined using equivalent scaling functions and equivalent wavelet functions:

[0037]

[0038] Where the equivalent scaling function is The equivalent wavelet function is Let the integer be R, and the function family be R.

[0039] The function family is decomposed using a multiresolution decomposition algorithm, expressed as follows:

[0040]

[0041] The adjoint function of the first to second k family of functions is h. * [1-2k], the adjoint of the first conjugate orthogonal mirror filter group of the first (1-2k)th is h * [1-2k];

[0042] The number of channels in the EEG signal is used to obtain the reference signal corresponding to the SSVEP visual stimulus frequency:

[0043]

[0044] The b-th reference signal is v b Let the number of target stimuli be b, and the frequency of the b-th target stimulus be f. b The number of sampling points is N s ;

[0045] The correlation coefficient is calculated based on the EEG signal and the reference signal corresponding to the b-th stimulation frequency:

[0046]

[0047] The horizontal weight matrix is ​​D. w The vertical weight matrix is ​​D v If the EEG signal is w, and K is the stimulus frequency data, then the target frequency to be identified is calculated:

[0048]

[0049] The identification data is output based on the target frequency.

[0050] Furthermore, the method for constructing a patient intention brainwave recognition feedback model based on the recognition data includes:

[0051] Based on the identification data, a target function is constructed, expressed as follows:

[0052]

[0053] The actual identification data is Predictive identification data is Actual identification data and predictive identification data The loss function is The objective function is

[0054] The patient intent EEG recognition feedback model includes empirical mode decomposition, decision tree algorithm, and machine learning pattern recognition network;

[0055] Empirical mode decomposition decomposes the input EEG signal to obtain local amplitude, phase and frequency, and outputs the local amplitude, phase and frequency as the required features;

[0056] The decision tree algorithm recursively divides the demand features into multiple regions, and constructs a tree model based on the decision rules of each region to predict the category and obtain classification data.

[0057] Machine learning pattern recognition networks perform intent matching on classified data to obtain matching results, and output the intents with matching results greater than 0.863 as feedback results.

[0058] Furthermore, the method for optimizing the patient intention brainwave recognition feedback model based on feedback error includes:

[0059] The search subpopulation is optimized by introducing low-dissimilarity sequences, as expressed in the following expression:

[0060]

[0061] The initial position of the search sub in the d-th dimension is: The random number S in the interval 0 to 1 for a low-dissimilarity sequence is...e The maximum value of the search sub-position is Z. max The minimum value of the search sub-position is Z. min ;

[0062] The search terms are randomly divided into four independent groups, and the fitness of the search terms is calculated:

[0063]

[0064] The fitness function of the first group is g1, the current iteration number is t, and the maximum iteration number is t. max The fitness function of the second group is g2, the fitness function of the third group is g3, and the fitness function of the fourth group is g4.

[0065] Calculate the fitness of the four groups of search elements and sort them in descending order. The top four search elements are then designated as attacker, interceptor, chaser, and pursuer, respectively.

[0066] Calculate the correlation coefficient:

[0067]

[0068] The correlation coefficient is The contraction coefficient is a, and the random vector between 0 and 1 is r1;

[0069] The fitness of the search element is updated based on the current iteration number. When the random number p is less than 0.5 and the absolute value of the correlation coefficient is less than 1, the position of the search element is updated to obtain the first position, as expressed by:

[0070]

[0071] The attacker's position is Z. b1 The interceptor's position is Z. b2 The location of the driver is Z. b3 The pursuer's position is Z. b4 The first position of the search sub-sub in the (t+1)th iteration is Z1(t+1), the actual feedback is Q, and the predicted feedback is...

[0072] When the random number p is less than 0.5 and the absolute value of the correlation coefficient is greater than or equal to 1, the search sub-position is updated to obtain the second position, and the expression is:

[0073]

[0074] The second position of the search element in the (t+1)th iteration is Z2(t+1), the random vector between 0 and 1 is r2, and the position of the random search element is Z. random The search subposition in the t-th iteration is Z(t);

[0075] When the random number p is greater than or equal to 0.5, the third position is obtained by updating the search sub-position according to the spiral ascent method, and the expression is:

[0076]

[0077] The third position of the search element in the (t+1)th iteration is Z3(t+1), and the target position of the search element is Z. * (t), where ρ is a random number between -1 and 1, and λ is the spiral control constant;

[0078] Otherwise, continue iterating until the maximum number of iterations is reached, then stop the search.

[0079] Secondly, a feedback system for recognizing patient intentions via brain-computer interfaces includes:

[0080] Data acquisition module: used to acquire EEG signals and related data from the patient's brain-computer interface, and to preprocess the EEG signals and related data;

[0081] Data correction module: used to perform pattern classification on the EEG signals to obtain pattern data, and use the associated data to perform intention correction on the pattern data to obtain corrected data;

[0082] Recognition construction module: used to perform EEG signal recognition on the corrected data to obtain recognition data, and to construct a patient intention EEG recognition feedback model based on the recognition data;

[0083] Optimization feedback module: used to optimize the patient intention brainwave recognition feedback model based on the feedback error, input the data to be fed back into the optimized patient intention brainwave recognition feedback model, and output the feedback result.

[0084] The beneficial effects of this invention are:

[0085] This invention relates to a feedback method and system for recognizing patient intentions via brain-computer interfaces. Compared with existing technologies, this invention has the following technical advantages:

[0086] This invention improves the feedback accuracy of patient intention EEG recognition in brain-computer interfaces (BCIs) through preprocessing, pattern classification, intent correction, EEG signal recognition, model construction, and model optimization. This enhances the feedback precision of patient intention EEG recognition in BCIs, significantly saving resources and improving work efficiency. It enables intelligent feedback for patient intention EEG recognition in BCIs, allowing for real-time intent correction. This is of great significance for the feedback of patient intention EEG recognition in BCIs, adapting to different standards and varying patient intention EEG recognition needs, thus possessing a certain degree of universality. Attached Figure Description

[0087] Figure 1 This is a flowchart illustrating the steps of the feedback method for patient intention brainwave recognition based on brain-computer interface according to the present invention.

[0088] Figure 2 This is a block diagram of the feedback system for patient intention brainwave recognition based on brain-computer interface according to the present invention. Detailed Implementation

[0089] The present invention will be further described below through specific embodiments. The illustrative embodiments and descriptions herein are used to explain the present invention, but are not intended to limit the present invention.

[0090] The feedback method and system for patient intention brainwave recognition based on brain-computer interface of the present invention includes the following steps:

[0091] like Figure 1 As shown, this embodiment includes the following steps:

[0092] Collect the patient's brain-computer interface EEG signals and related data, and preprocess the EEG signals and related data;

[0093] In actual assessments, relevant data include medical records, medication history, lifestyle habits, emotions, and fatigue levels.

[0094] The list of needs included "wanting to rest", "wanting to remove the tube", "wanting to talk", "thirsty", "wanting to defecate", "wanting to see family", "feeling cold", "feeling hot", "wanting to turn over", "feeling pain", and "having phlegm";

[0095] The study included a 51-year-old male patient with spinal cord injury leading to paralysis. His electroencephalogram (EEG) showed increased beta wave activity in the sensory and motor cortex areas, possibly indicating that the patient was attempting to move or roll over. Theta waves showed slight fluctuations in the occipital lobe, possibly reflecting the patient's discomfort with his current posture. Alpha wave activity was relatively uniform throughout the cerebral cortex, but slightly decreased, possibly indicating that the patient was fatigued from maintaining the same posture for a long time.

[0096] Pattern data is obtained by performing pattern classification on the electroencephalogram (EEG) signals, and intention correction is performed on the pattern data using the associated data to obtain corrected data.

[0097] The corrected data is subjected to EEG signal recognition to obtain recognition data, and a patient intention EEG recognition feedback model is constructed based on the recognition data.

[0098] The patient intention brainwave recognition feedback model is optimized based on the feedback error. The data to be fed back is input into the optimized patient intention brainwave recognition feedback model, and the feedback result is output.

[0099] In the actual assessment, the feedback was that they wanted to turn their lives around.

[0100] In this embodiment, the method for obtaining pattern data by pattern classification of the electroencephalogram (EEG) signals includes:

[0101] Add particles as a search tool for EEG signals, and update the search velocity and position of the particles. The expression is:

[0102]

[0103] The local optimal position of the i-th particle in the t-th iteration is: The global optimal position of the i-th particle in the t-th iteration is The position of the i-th particle in the t-th iteration is The search speed of the i-th particle in the (t+1)-th iteration is The search speed of the i-th particle in the t-th iteration is The position of the i-th particle in the (t+1)-th iteration is The random numbers from 0 to 1 are τ1 and τ2, the acceleration factors are λ1 and λ2 respectively, and the inertia weight factor is η;

[0104] Calculate the similarity between EEG signals:

[0105]

[0106] Among them, the first One EEG signal is The m-th EEG signal is x m EEG signals and EEG signals x m The similarity is EEG signals The probability density is EEG signal x m The probability density is EEG signals and EEG signals x m The weighting factor is

[0107] The EEG signals with the highest similarity are clustered to obtain initial clusters, and the overall silhouette coefficient of the clusters is calculated:

[0108]

[0109] The average distance between the m-th EEG signal and the EEG signals within the initial cluster is l (m), and the average distance between the m-th EEG signal and the EEG signal points between the initial clusters is... The number of EEG signals is

[0110] Calculate the penalty item:

[0111]

[0112] The number of EEG signals with negative silhouette coefficients is N. ng The average distance between the ι-th EEG signal and the EEG signals within the initial cluster is l(ι), and the average distance between the ι-th EEG signal and the EEG signals between the initial clusters is...

[0113] Construct the fitness function, with the following expression:

[0114]

[0115] Where the fitness function is The actual number of clusters is A, the predicted number of clusters is X, the convolution operation is *, and the cutoff distance is φ. o Indicator function is EEG signals and EEG signals x m The distance is No. The center of each cluster is EEG signals and cluster center The distance is EEG signals and cluster center ε m The distance between them is Membership function is

[0116] The initial cluster is updated through continuous iteration of the particle population until the fitness function value is maximized, at which point the iteration stops and the updated initial cluster is output as model data.

[0117] In this embodiment, the method for obtaining corrected data by modifying the pattern data using the associated data includes:

[0118] Correlation analysis is performed on the associated data to obtain the patient's potential behaviors, and the pattern data is corrected using the potential behaviors to obtain corrected data.

[0119] In this embodiment, the method for obtaining recognition data by performing EEG signal recognition on the corrected data includes:

[0120] The profile space and detail space of the corrected data are decomposed step by step. The profile space is transformed by the first conjugate orthogonal mirror filter bank and the second conjugate orthogonal mirror filter bank, respectively, to obtain the transformed profile space and detail space. The expressions are as follows:

[0121]

[0122] β[n]=(-1) 1-n h[1-n]

[0123] The j-th generalization space is E. j The (j+1)th detail space is C j+1 The (j+1)th transformation profile space is E j+1 The first conjugate orthogonal mirror filter group (1-n) is h[1-n], the first conjugate orthogonal mirror filter group is h[n], and the second conjugate orthogonal mirror filter group is β[n].

[0124] The family of functions is defined using equivalent scaling functions and equivalent wavelet functions:

[0125]

[0126] Where the equivalent scaling function is The equivalent wavelet function is Let the integer be R, and the function family be R.

[0127] The function family is decomposed using a multiresolution decomposition algorithm, expressed as follows:

[0128]

[0129] The adjoint function of the first to second k family of functions is h. * [1-2k], the adjoint of the first conjugate orthogonal mirror filter group of the first (1-2k)th is h * [1-2k];

[0130] The number of channels in the EEG signal is used to obtain the reference signal corresponding to the SSVEP visual stimulus frequency:

[0131]

[0132] The b-th reference signal is v b Let the number of target stimuli be b, and the frequency of the b-th target stimulus be f. b The number of sampling points is N s ;

[0133] The correlation coefficient is calculated based on the EEG signal and the reference signal corresponding to the b-th stimulation frequency:

[0134]

[0135] The horizontal weight matrix is ​​D. w The vertical weight matrix is ​​D v If the EEG signal is w, and K is the stimulus frequency data, then the target frequency to be identified is calculated:

[0136]

[0137] The identification data is output based on the target frequency.

[0138] In this embodiment, the method for constructing a patient intention brainwave recognition feedback model based on the recognition data includes:

[0139] Based on the identification data, a target function is constructed, expressed as follows:

[0140]

[0141] The actual identification data is Predictive identification data is Actual identification data and predictive identification data The loss function is The objective function is

[0142] The patient intent EEG recognition feedback model includes empirical mode decomposition, decision tree algorithm, and machine learning pattern recognition network;

[0143] Empirical mode decomposition decomposes the input EEG signal to obtain local amplitude, phase and frequency, and outputs the local amplitude, phase and frequency as the required features;

[0144] The decision tree algorithm recursively divides the demand features into multiple regions, and constructs a tree model based on the decision rules of each region to predict the category and obtain classification data.

[0145] Machine learning pattern recognition networks perform intent matching on classified data to obtain matching results, and output the intents with matching results greater than 0.863 as feedback results.

[0146] In this embodiment, the method for optimizing the patient intention brainwave recognition feedback model based on feedback error includes:

[0147] The search subpopulation is optimized by introducing low-dissimilarity sequences, as expressed in the following expression:

[0148]

[0149] The initial position of the search sub in the d-th dimension is: The random number S in the interval 0 to 1 for a low-dissimilarity sequence is... e The maximum value of the search sub-position is Z. max The minimum value of the search sub-position is Z. min ;

[0150] The search terms are randomly divided into four independent groups, and the fitness of the search terms is calculated:

[0151]

[0152] The fitness function of the first group is g1, the current iteration number is t, and the maximum iteration number is t. max The fitness function of the second group is g2, the fitness function of the third group is g3, and the fitness function of the fourth group is g4.

[0153] Calculate the fitness of the four groups of search elements and sort them in descending order. The top four search elements are then designated as attacker, interceptor, chaser, and pursuer, respectively.

[0154] Calculate the correlation coefficient:

[0155]

[0156] The correlation coefficient is The contraction coefficient is a, and the random vector between 0 and 1 is r1;

[0157] The fitness of the search element is updated based on the current iteration number. When the random number p is less than 0.5 and the absolute value of the correlation coefficient is less than 1, the position of the search element is updated to obtain the first position, as expressed by:

[0158]

[0159] The attacker's position is Z. b1 The interceptor's position is Z. b2 The location of the driver is Z. b3 The pursuer's position is Z. b4 The first position of the search sub-sub in the (t+1)th iteration is Z1(t+1), the actual feedback is Q, and the predicted feedback is...

[0160] When the random number p is less than 0.5 and the absolute value of the correlation coefficient is greater than or equal to 1, the search sub-position is updated to obtain the second position, and the expression is:

[0161]

[0162] The second position of the search element in the (t+1)th iteration is Z2(t+1), the random vector between 0 and 1 is r2, and the position of the random search element is Z. random The search subposition in the t-th iteration is Z(t);

[0163] When the random number p is greater than or equal to 0.5, the third position is obtained by updating the search sub-position according to the spiral ascent method, and the expression is:

[0164]

[0165] The third position of the search element in the (t+1)th iteration is Z3(t+1), and the target position of the search element is Z. * (t), where ρ is a random number between -1 and 1, and λ is the spiral control constant;

[0166] Otherwise, continue iterating until the maximum number of iterations is reached, then stop the search.

[0167] Secondly, a feedback system for recognizing patient intentions via brain-computer interfaces includes:

[0168] Data acquisition module: used to acquire EEG signals and related data from the patient's brain-computer interface, and to preprocess the EEG signals and related data;

[0169] Data correction module: used to perform pattern classification on the EEG signals to obtain pattern data, and use the associated data to perform intention correction on the pattern data to obtain corrected data;

[0170] Recognition construction module: used to perform EEG signal recognition on the corrected data to obtain recognition data, and to construct a patient intention EEG recognition feedback model based on the recognition data;

[0171] Optimization feedback module: used to optimize the patient intention brainwave recognition feedback model based on the feedback error, input the data to be fed back into the optimized patient intention brainwave recognition feedback model, and output the feedback result.

[0172] In actual assessments, the feedback system for patient intention brainwave recognition based on brain-computer interfaces includes software and hardware components. The hardware component includes a data acquisition module and a patient support and connection module, as well as a data correction module, a recognition construction module, and an optimized feedback module.

[0173] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A feedback method for patient intention brainwave recognition based on brain-computer interface, characterized in that, Includes the following steps: Collect the patient's brain-computer interface EEG signals and related data, and preprocess the EEG signals and related data; Pattern data is obtained by performing pattern classification on the electroencephalogram (EEG) signals, and intention correction is performed on the pattern data using the associated data to obtain corrected data. The corrected data is subjected to EEG signal recognition to obtain recognition data, and a patient intention EEG recognition feedback model is constructed based on the recognition data. The patient intention brainwave recognition feedback model is optimized based on the feedback error. The data to be fed back is input into the optimized patient intention brainwave recognition feedback model, and the feedback result is output.

2. The feedback method for patient intention brainwave recognition based on brain-computer interface according to claim 1, characterized in that, A method for obtaining pattern data by pattern classification of the electroencephalogram (EEG) signals includes: Add particles as a search tool for EEG signals, and update the search velocity and position of the particles. The expression is: The local optimal position of the i-th particle in the t-th iteration is: The global optimal position of the i-th particle in the t-th iteration is The position of the i-th particle in the t-th iteration is The search speed of the i-th particle in the (t+1)-th iteration is The search speed of the i-th particle in the t-th iteration is The position of the i-th particle in the (t+1)-th iteration is A random number between 0 and 1 is , The acceleration factors are respectively , The inertia weighting factor is ; Calculate the similarity between EEG signals: Among them, the first One EEG signal is The m-th EEG signal is EEG signals and EEG signals The similarity is EEG signals The probability density is EEG signals The probability density is EEG signals and EEG signals The weighting factor is ; The EEG signals with the highest similarity are clustered to obtain initial clusters, and the overall silhouette coefficient of the clusters is calculated: The average distance between the m-th EEG signal and the initial cluster EEG signals is . The average distance between the m-th EEG signal and the initial inter-cluster EEG signal points is The number of brain signals is ; Calculate the penalty item: The number of EEG signals with negative silhouette coefficients is . , No. The average distance between each EEG signal and the initial EEG signal within the cluster is , No. The average distance between each EEG signal and the initial inter-cluster EEG signal is ; Construct the fitness function, with the following expression: Where the fitness function is The actual number of clusters is Predict the number of clusters in a cluster. Convolution operation is The cutoff distance is Indicator function is EEG signals and EEG signals The distance is , No. The center of each cluster is EEG signals and cluster center The distance is EEG signals and cluster center The distance between them is The membership function is ; The initial cluster is updated through continuous iteration of the particle population until the fitness function value is maximized, at which point the iteration stops and the updated initial cluster is output as model data.

3. The feedback method for patient intention brainwave recognition based on brain-computer interface according to claim 1, characterized in that, A method for obtaining corrected data by performing intent correction on the pattern data using the associated data includes: Correlation analysis is performed on the associated data to obtain the patient's potential behaviors, and the pattern data is corrected using the potential behaviors to obtain corrected data.

4. The feedback method for patient intention brainwave recognition based on brain-computer interface according to claim 1, characterized in that, A method for obtaining recognition data by performing EEG signal recognition on the corrected data includes: The general space and detail space of the corrected data are decomposed step by step. Perform the first conjugate orthogonal mirror filter bank transformation on the overview space respectively. The transformation of the second conjugate orthogonal mirror filter bank yields the transformation overview space and detail space. The expression is: The j-th generalization space is The (j+1)th detail space is The (j+1)th transformation profile space is The first (1-n)th first conjugate orthogonal mirror filter bank is The first conjugate orthogonal mirror filter bank is The second conjugate orthogonal mirror filter bank is ; The family of functions is defined using equivalent scaling functions and equivalent wavelet functions: Where the equivalent scaling function is The equivalent wavelet function is Let the integer be R, and the function family be R. ; The function family is decomposed using a multiresolution decomposition algorithm, expressed as follows: Among them, the first The adjoint of a family of functions is , No. The adjoint of the first conjugate orthogonal mirror filter bank is ; The number of channels in the EEG signal is used to obtain the reference signal corresponding to the SSVEP visual stimulus frequency: The b-th reference signal is Let the number of target stimuli be b, and the frequency of the b-th target stimulus be... The number of sampling points is ; The correlation coefficient is calculated based on the EEG signal and the reference signal corresponding to the b-th stimulation frequency: The horizontal weight matrix is: The vertical weight matrix is If the EEG signal is w, and K is the stimulus frequency data, then the target frequency to be identified is calculated: The identification data is output based on the target frequency.

5. The feedback method for patient intention brainwave recognition based on brain-computer interface according to claim 1, characterized in that, A method for constructing a patient intention brainwave recognition feedback model based on the recognition data includes: Based on the identification data, a target function is constructed, expressed as follows: The actual identification data is The predicted identification data is Actual identification data and predictive identification data The loss function is The objective function is ; The patient intent EEG recognition feedback model includes empirical mode decomposition, decision tree algorithm, and machine learning pattern recognition network; Empirical mode decomposition decomposes the input EEG signal to obtain local amplitude, phase and frequency, and outputs the local amplitude, phase and frequency as the required features; The decision tree algorithm recursively divides the demand features into multiple regions, and constructs a tree model based on the decision rules of each region to predict the category and obtain classification data. Machine learning pattern recognition networks perform intent matching on classified data to obtain matching results, and output the intents with matching results greater than 0.863 as feedback results.

6. The feedback method for patient intention brainwave recognition based on brain-computer interface according to claim 1, characterized in that, A method for optimizing the patient intention EEG recognition feedback model based on feedback error includes: The search subpopulation is optimized by introducing low-dissimilarity sequences, as expressed in the following expression: The initial position of the search sub in the d-th dimension is: The random number for low-dissimilarity sequences in the interval between 0 and 1 is The maximum value of the search sub-position is The minimum value of the search sub-position is ; The search terms are randomly divided into four independent groups, and the fitness of the search terms is calculated: The fitness function for the first group is: The current iteration number is t, and the maximum iteration number is t. The fitness function of the second group is The fitness function of the third group is The fitness function of the fourth group is ; Calculate the fitness of the four groups of search elements and sort them in descending order. The top four search elements are then designated as attacker, interceptor, chaser, and pursuer, respectively. Calculate the correlation coefficient: The correlation coefficient is The contraction coefficient is a, and the random vector between 0 and 1 is... ; The fitness of the search element is updated based on the current iteration number. When the random number p is less than 0.5 and the absolute value of the correlation coefficient is less than 1, the position of the search element is updated to obtain the first position, as expressed by: The attacker's location is The interceptor's location is The location of the driver is The pursuer's position is The first position of the search sub-sub in the (t+1)th iteration is The actual feedback is Q, and the predicted feedback is ; When the random number p is less than 0.5 and the absolute value of the correlation coefficient is greater than or equal to 1, the search sub-position is updated to obtain the second position, and the expression is: The second position of the search sub in the (t+1)th iteration is A random vector between 0 and 1 is The position of the random search sub is The search subposition in the t-th iteration is ; When the random number p is greater than or equal to 0.5, the third position is obtained by updating the search sub-position according to the spiral ascent method, and the expression is: The third position of the search sub in the (t+1)th iteration is... The target location of the search sub is A random number between -1 and 1 is The screw control constant is ; Otherwise, continue iterating until the maximum number of iterations is reached, then stop the search.

7. A feedback system for patient intention brainwave recognition based on a brain-computer interface, used to perform the method according to any one of claims 1-6, characterized in that, include: Data acquisition module: used to acquire EEG signals and related data from the patient's brain-computer interface, and to preprocess the EEG signals and related data; Data correction module: used to perform pattern classification on the EEG signals to obtain pattern data, and use the associated data to perform intention correction on the pattern data to obtain corrected data; Recognition construction module: used to perform EEG signal recognition on the corrected data to obtain recognition data, and to construct a patient intention EEG recognition feedback model based on the recognition data; Optimization feedback module: used to optimize the patient intention brainwave recognition feedback model based on feedback error, input the data to be fed back into the optimized patient intention brainwave recognition feedback model, and output the feedback result.

Images