An EEG feedback training method and system based on multi-lead EEG feature analysis

Abstract

The present invention belongs to the technical field of EEG feature analysis, and specifically relates to an EEG feedback training method and system based on multi-lead EEG feature analysis. The invention can effectively classify and evaluate the state of EEG activities through an EEG feedback training method based on multi-lead EEG feature analysis, and when collecting EEG feature spectra, it is not determined according to the collection position, but classified according to the actual collected EEG feature spectra. Specifically, it can be classified into a baseline feature subset and an edge feature subset. The EEG spectrum features in the edge feature subset may correspond to any of the adjacent brain regions, and in practical applications, it can more accurately identify and locate abnormal EEG activity areas. In addition, through the feedback mechanism, the present invention can adjust EEG activities in real time, provide personalized EEG activity adjustment strategies for training subjects, and help training subjects adjust EEG activities in time when abnormal brain regions appear through real-time monitoring and feedback adjustment.

Description

Technical Field

[0001] The present invention belongs to the technical field of electroencephalogram (EEG) feature analysis, and in particular relates to an EEG feedback training method and system based on multi-lead EEG feature analysis. Background Art

[0002] With the rapid development of brain science and neuroscience, people have a deeper understanding of brain activity. Electroencephalogram (EEG), as a non-invasive means of brain function detection, has been widely used in clinical and scientific research fields due to its high temporal resolution and relatively low cost. However, traditional EEG analysis methods often rely on the experience and judgment of experts, lack objective and unified standards, and cannot provide real-time feedback on the state of EEG activity, which limits its application in EEG feedback training.

[0003] In the prior art, although there are some EEG-based feedback training systems, the classification of the EEG feature spectra in various brain regions is often inaccurate. As a result, when evaluating the EEG activity state, there will be large deviations in the output results, thereby affecting the training effect and accuracy. Based on this, the present invention proposes an EEG feedback training method based on multi-lead EEG feature analysis to solve the above problems. Summary of the invention

[0004] The purpose of the present invention is to provide an EEG feedback training method and system based on multi-lead EEG feature analysis, which can more accurately classify and evaluate EEG feature spectra, thereby improving the efficiency and accuracy of EEG feedback training.

[0005] The technical solution adopted by the present invention is as follows:

[0006] An EEG feedback training method based on multi-lead EEG feature analysis, comprising:

[0007] Preprocessing the collected multi-lead EEG data, and extracting features from the preprocessed multi-lead EEG data to extract EEG feature spectra related to EEG activity;

[0008] Classifying and processing the EEG characteristic spectra of different brain regions to obtain multiple benchmark characteristic subsets;

[0009] Screening and processing the EEG characteristic spectra in each of the reference feature subsets to obtain the standard fluctuation amplitude range corresponding to the EEG characteristic spectra under each of the reference feature subsets;

[0010] According to the standard fluctuation amplitude range, identifying the EEG activity state of the target brain area, wherein the EEG activity state includes a normal state and an abnormal state;

[0011] In the normal state, the corresponding brain area is recorded as a normal brain area, and the EEG characteristic spectrum of the normal brain area is monitored in real time, and the training parameters are adjusted in real time according to the monitoring results to maintain the EEG activity within the standard fluctuation range;

[0012] In the abnormal state, the corresponding brain area is recorded as an abnormal brain area, and the training subject is prompted to adjust the EEG activity through a feedback mechanism until the abnormal brain area returns to a normal state.

[0013] In a preferred embodiment, the step of extracting the EEG characteristic spectrum related to EEG activity includes:

[0014] Acquire multi-lead EEG data, and perform filtering on the multi-lead EEG data to remove noise and artifacts;

[0015] Performing dimensionality reduction processing on the filtered multi-lead EEG data, converting the multi-lead EEG data in a multi-dimensional state into multi-lead EEG data in a low-dimensional state;

[0016] The frequency domain analysis of the multi-lead EEG data after dimensionality reduction is performed using Fourier transform, the time domain signal is converted into a frequency domain signal, and the EEG characteristic spectrum is extracted.

[0017] In a preferred embodiment, the step of classifying the EEG feature spectra of different brain regions to obtain a plurality of reference feature subsets includes:

[0018] Acquire all the EEG characteristic spectra, and perform vectorization processing to obtain the EEG characteristic vectors of each of the EEG characteristic spectra;

[0019] Obtaining a classification function, and inputting the EEG feature vectors into the classification function for one-to-one comparison to obtain similarity scores between the EEG feature vectors;

[0020] Obtaining a screening threshold, and comparing the screening threshold with the similarity score, and classifying the EEG feature spectra corresponding to the EEG feature vectors whose similarity scores are greater than the screening threshold as the same type of EEG feature spectra, and then aggregating them into the same benchmark feature subset; conversely, classifying the EEG feature spectra corresponding to the EEG feature vectors as different types of EEG feature spectra, and classifying them into different benchmark feature subsets;

[0021] Each of the reference feature subsets is added with a unique feature identifier, and each of the reference feature subsets corresponds to a brain region.

[0022] In a preferred embodiment, after the benchmark feature subsets are output, if the number of the benchmark feature subsets is higher than the number of brain regions, it indicates that there are redundant classifications in the EEG feature spectrum; if the number of the benchmark feature subsets is lower than the number of brain regions, it indicates that there are missing classifications in the EEG feature spectrum;

[0023] Under the redundant classification, the EEG feature spectra in the benchmark feature subsets under each brain region are merged and processed until the number of the benchmark feature subsets is consistent with the number of brain regions;

[0024] Under the shortage classification, the amount of multi-lead EEG data collected is increased, and feature extraction and classification processing are re-performed until the number of baseline feature subsets matches the number of brain regions.

[0025] In a preferred embodiment, the step of merging the EEG feature spectra in the reference feature subsets of each brain region comprises:

[0026] Obtaining all benchmark feature subsets under the same brain region, and summing and averaging the similarity scores of the EEG feature spectra in each of the benchmark feature subsets to obtain a merging condition parameter;

[0027] Performing a positive offset process on the screening threshold to obtain an allowable merging threshold;

[0028] The reference feature subsets under the merging condition parameters higher than the allowable merging threshold are directly merged, and the EEG feature spectra in the reference feature subsets under the merging condition parameters lower than the allowable merging threshold are recorded as edge feature spectra;

[0029] All the edge feature spectra under adjacent brain regions are merged and recorded as edge feature subsets.

[0030] In a preferred embodiment, the step of screening the EEG feature spectra in each of the reference feature subsets to obtain the standard fluctuation amplitude range corresponding to the EEG feature spectra under each of the reference feature subsets includes:

[0031] Obtaining the EEG feature spectra in the benchmark feature subset and the similarity scores between the EEG feature spectra, and recording them as screening condition parameters;

[0032] Performing bidirectional offset processing on each of the screening condition parameters to obtain a screening condition interval;

[0033] Counting the number of screening condition parameters under each of the screening condition intervals and recording them as condition parameters to be evaluated;

[0034] Arrange the condition parameters to be evaluated in descending order, and record the EEG characteristic spectrum corresponding to the screening condition interval under the condition parameter to be evaluated with the highest ranking as the standard characteristic spectrum;

[0035] The fluctuation amplitude of the standard characteristic spectrum is collected and recorded as the standard fluctuation amplitude range of the EEG characteristic spectrum.

[0036] In a preferred embodiment, the step of identifying the EEG activity state of the target brain area according to the standard fluctuation amplitude range includes:

[0037] Obtain the EEG characteristic spectrum of the target brain area and record it as the characteristic spectrum to be corrected;

[0038] Comparing the characteristic spectrum to be calibrated with the standard fluctuation amplitude range;

[0039] If the fluctuation amplitude of the characteristic spectrum to be corrected is within the standard fluctuation amplitude range, it is determined that the EEG activity state of the target brain area is normal;

[0040] If the fluctuation amplitude of the characteristic spectrum to be corrected exceeds the standard fluctuation amplitude range, it is determined that the EEG activity state of the target brain area is abnormal.

[0041] In a preferred embodiment, the step of prompting the training subject to adjust the EEG activity through a feedback mechanism until the abnormal brain area returns to a normal state comprises:

[0042] Acquiring the EEG characteristic spectrum of the abnormal brain area and recording it as the abnormal characteristic spectrum;

[0043] generating a brain electrical activity adjustment strategy according to the difference between the abnormal characteristic spectrum and the standard characteristic spectrum;

[0044] Feedback the EEG activity adjustment strategy to the training subject through vision, hearing and touch to guide the EEG activity of the target brain area;

[0045] The EEG activity of the target brain area is continuously monitored, and the EEG characteristic spectrum of the abnormal brain area is updated in real time. When the fluctuation amplitude of the EEG characteristic spectrum of the abnormal brain area returns to the standard fluctuation amplitude range, it is confirmed that the EEG activity state has returned to normal and the feedback adjustment is ended.

[0046] The present invention also provides an EEG feedback training system based on multi-lead EEG feature analysis, using the above-mentioned EEG feedback training method based on multi-lead EEG feature analysis, comprising:

[0047] A feature extraction module, which is used to preprocess the collected multi-lead EEG data, and perform feature extraction on the preprocessed multi-lead EEG data to extract EEG feature spectrum related to EEG activity;

[0048] A feature classification module, wherein the feature classification module is used to classify the EEG feature spectra of different brain regions to obtain a plurality of benchmark feature subsets;

[0049] A standard construction module, wherein the standard construction module is used to screen and process the EEG characteristic spectra in each of the reference feature subsets to obtain a standard fluctuation amplitude interval corresponding to the EEG characteristic spectra under each of the reference feature subsets;

[0050] A state evaluation module, the state evaluation module is used to identify the EEG activity state of the target brain area according to the standard fluctuation amplitude interval, wherein the EEG activity state includes a normal state and an abnormal state;

[0051] A first intervention module, wherein the first intervention module is used to record the corresponding brain area as a normal brain area in the normal state, monitor the EEG characteristic spectrum of the normal brain area in real time, and adjust the training parameters in real time according to the monitoring results to maintain the EEG activity within the standard fluctuation range;

[0052] The second intervention module is used to record the corresponding brain area as an abnormal brain area under the abnormal state, and prompt the training subject to adjust the EEG activity through a feedback mechanism until the abnormal brain area returns to a normal state.

[0053] And, an electronic device, the electronic device comprising:

[0054] at least one processor;

[0055] and a memory communicatively coupled to the at least one processor;

[0056] Among them, the memory stores a computer program that can be executed by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the above-mentioned EEG feedback training method based on multi-lead EEG feature analysis.

[0057] The technical effects achieved by the present invention are:

[0058] The present invention can effectively classify and evaluate the status of EEG activities through an EEG feedback training method based on multi-lead EEG feature analysis, and when collecting EEG feature spectra, it is not determined based on the collection position, but classified based on the actually collected EEG feature spectra. Specifically, it can be classified into a baseline feature subset and an edge feature subset. The EEG spectrum features in the edge feature subset may correspond to any one of the adjacent brain regions. In practical applications, it can more accurately identify and locate abnormal EEG activity areas. In addition, through the feedback mechanism, the present invention can adjust EEG activities in real time and provide personalized EEG activity adjustment strategies for training subjects. Through real-time monitoring and feedback adjustment, it helps training subjects to adjust EEG activities in time when abnormal brain regions appear, promote EEG activities to return to the normal fluctuation amplitude range, and thus improve brain function and promote brain health. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] Figure 1 It is a schematic flow chart of the method of the present invention;

[0060] Figure 2 It is a schematic diagram of the system module of the present invention;

[0061] Figure 3 It is a schematic diagram of the structure of an electronic device of the present invention. DETAILED DESCRIPTION

[0062] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the accompanying drawings.

[0063] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0064] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure or characteristic that may be included in at least one implementation of the present invention. The phrase "in a preferred embodiment" that appears in different places in this specification does not refer to the same embodiment, nor is it a separate or selective embodiment that is mutually exclusive with other embodiments.

[0065] See also Figure 1 As shown, the present invention provides an EEG feedback training method based on multi-lead EEG feature analysis, comprising:

[0066] S1. Preprocessing the collected multi-lead EEG data, and extracting features from the preprocessed multi-lead EEG data to extract EEG feature spectra related to EEG activity;

[0067] S2, classifying and processing the EEG feature spectra of different brain regions to obtain multiple benchmark feature subsets;

[0068] S3, screening and processing the EEG characteristic spectrum in each benchmark characteristic subset to obtain the standard fluctuation amplitude range corresponding to the EEG characteristic spectrum under each benchmark characteristic subset;

[0069] S4. Identify the EEG activity state of the target brain area according to the standard fluctuation amplitude range, wherein the EEG activity state includes a normal state and an abnormal state;

[0070] S5. Under normal conditions, the corresponding brain area is recorded as a normal brain area, and the EEG characteristic spectrum of the normal brain area is monitored in real time. The training parameters are adjusted in real time according to the monitoring results to maintain the EEG activity within the standard fluctuation range;

[0071] S6. Under abnormal conditions, the corresponding brain area is recorded as an abnormal brain area, and the training subject is prompted to adjust the EEG activity through a feedback mechanism until the abnormal brain area returns to normal.

[0072] As described in the above steps S1-S6, with the rapid development of brain science and neuroscience, people's research on brain activities is also becoming more and more advanced. In this embodiment, first, the collected multi-lead EEG data is preprocessed, specifically including noise removal, filtering and other operations to ensure the accuracy and reliability of the data. After the preprocessing is completed, the preprocessed multi-lead EEG data is feature extracted to extract the EEG feature spectrum related to the EEG activity. The EEG feature spectrum can reflect the activity state of the brain. Then, the EEG feature spectra of different brain regions are classified and processed, and output as multiple benchmark feature subsets. Then, the EEG feature spectra in each benchmark feature subset are screened and processed to screen out the standard fluctuation amplitude range corresponding to the EEG feature spectrum under each benchmark feature subset. The standard fluctuation amplitude range The interval is the corresponding basis for judging the state of EEG activity. On this basis, the EEG activity state of the target brain area can be identified according to the standard fluctuation amplitude interval. The EEG activity state mainly includes normal state and abnormal state. Under the normal state, the corresponding brain area is recorded as a normal brain area, and the EEG characteristic spectrum of the normal brain area is monitored in real time. According to the monitoring results, the training parameters can be adjusted in real time to maintain the EEG activity within the standard fluctuation amplitude interval. Under the abnormal state, the corresponding brain area is recorded as an abnormal brain area, and the training object is prompted to adjust the EEG activity through the feedback mechanism. The training object can take corresponding measures according to the feedback information, such as adjusting breathing, relaxation, etc., to prompt the abnormal brain area to return to normal state. This process needs to continue until the EEG characteristic spectrum of the abnormal brain area returns to the standard fluctuation amplitude interval.

[0073] In a preferred embodiment, the step of extracting the EEG characteristic spectrum related to EEG activity includes:

[0074] S101, acquiring multi-lead EEG data, and filtering the multi-lead EEG data to remove noise and artifacts;

[0075] S102, performing dimensionality reduction processing on the filtered multi-lead EEG data, converting the multi-lead EEG data in a multi-dimensional state into multi-lead EEG data in a low-dimensional state;

[0076] S103. Perform frequency domain analysis on the multi-lead EEG data after dimensionality reduction using Fourier transform, convert the time domain signal into a frequency domain signal, and extract the EEG characteristic spectrum.

[0077] As described in the above steps S101-S103, in order to extract the EEG characteristic spectrum related to EEG activity, it is first necessary to obtain multi-lead EEG data. In order to ensure the quality and accuracy of the multi-lead EEG data, it is necessary to filter the multi-lead EEG data. The purpose of filtering is to remove noise and artifacts in the multi-lead EEG data. These noise and artifacts may come from external environmental interference, defects in the equipment itself, or other signals generated by non-EEG activities. Commonly used filtering methods include bandpass filtering, bandstop filtering, and notch filtering. These methods can effectively extract pure EEG signals. After filtering, the obtained multi-lead EEG data is subjected to dimensionality reduction processing. Since the original data is usually high-dimensional and contains a large amount of information and redundant data, the dimensionality reduction is very important. Dimensional processing can help simplify the data structure and convert multi-channel EEG data in a multi-dimensional state into multi-channel EEG data in a low-dimensional state. This step usually involves principal component analysis (PCA), independent component analysis (ICA) or other dimensionality reduction techniques, which aim to retain the most important features in the data while removing unnecessary information. Finally, in order to extract the EEG feature spectrum from the time domain signal, it is necessary to use Fourier transform to perform frequency domain analysis on the multi-channel EEG data after dimensionality reduction. Fourier transform is a mathematical tool that can convert time domain signals into frequency domain signals, thereby revealing the distribution of signals in different frequency components. Through this analysis, characteristic spectra related to EEG activity can be extracted, including different EEG wave bands, such as delta waves, theta waves, alpha waves, beta waves, and gamma waves, etc., providing important data support for further research and application.

[0078] In a preferred embodiment, the step of classifying and processing the EEG feature spectra of different brain regions to obtain a plurality of reference feature subsets includes:

[0079] S201, acquiring all EEG characteristic spectra, and performing vectorization processing to obtain EEG characteristic vectors of each EEG characteristic spectrum;

[0080] S202, obtaining a classification function, and inputting the EEG feature vectors into the classification function for one-to-one comparison to obtain similarity scores between the EEG feature vectors;

[0081] S203, obtaining a screening threshold, and comparing the screening threshold with the similarity score, and classifying the EEG feature spectra corresponding to the EEG feature vectors with similarity scores greater than the screening threshold as the same type of EEG feature spectra, and then summarizing them into the same benchmark feature subset; otherwise, classifying the EEG feature spectra corresponding to the EEG feature vectors as different types of EEG feature spectra, and classifying them into different benchmark feature subsets;

[0082] Among them, each benchmark feature subset is added with a unique feature identifier, and each benchmark feature subset corresponds to a brain region.

[0083] As described in the above steps S201-S203, in order to classify the EEG feature spectra of different brain regions to obtain multiple benchmark feature subsets, it is first necessary to obtain all EEG feature spectra and perform corresponding vectorization processing. This process includes converting each EEG feature spectrum into an EEG feature vector to facilitate subsequent calculation and analysis, ensuring the standardization and comparability of the data. Then, a preset classification function is introduced. The classification function can compare the input EEG feature vectors one by one. Through the classification function, the similarity score between each EEG feature vector can be calculated, wherein the expression of the classification function is: , where represents the similarity score, represents the number of feature points in the EEG feature vector, and Represents different EEG feature vectors. The similarity score reflects the similarity between different EEG feature vectors, which is an important basis for subsequent classification decisions. Then, according to the preset screening threshold, it is judged whether the similarity between EEG feature vectors is high enough. Specifically, the similarity score of each EEG feature vector is compared with the screening threshold. If the similarity score is greater than or equal to the screening threshold, the EEG feature spectrum corresponding to the compared EEG feature vector will be classified as the same type and summarized into the same benchmark feature subset. On the contrary, if the similarity score is less than the screening threshold, the EEG feature spectrum corresponding to the compared EEG feature vector will be classified as non-same type and classified into different benchmark feature subsets. Finally, in order to ensure the uniqueness of each benchmark feature subset, a feature identifier will be added to each subset. The feature identifier is unique and can clearly distinguish different benchmark feature subsets. In addition, each benchmark feature subset corresponds to a specific brain area, which helps to better understand and analyze the EEG activity characteristics of different brain areas. .

[0084] In a preferred embodiment, after the benchmark feature subsets are output, if the number of the benchmark feature subsets is higher than the number of brain regions, it indicates that there are redundant classifications in the EEG feature spectrum, and if the number of the benchmark feature subsets is lower than the number of brain regions, it indicates that there are missing classifications in the EEG feature spectrum;

[0085] Under redundant classification, the EEG feature spectra in the benchmark feature subsets of each brain region are merged until the number of benchmark feature subsets matches the number of brain regions;

[0086] Under the shortage classification, the amount of multi-lead EEG data collected was increased, and feature extraction and classification processing were re-performed until the number of baseline feature subsets matched the number of brain regions.

[0087] In this embodiment, after the output of the benchmark feature subsets, the number of benchmark feature subsets is compared with the number of brain regions to determine whether there is redundancy or shortage in classification in the EEG feature spectrum. Specifically, if the number of benchmark feature subsets exceeds the number of brain regions, then this means that there are redundant classifications in the EEG feature spectrum. On the contrary, if the number of benchmark feature subsets is less than the number of brain regions, then this indicates that there are shortages in the EEG feature spectrum. In the case of redundant classification, the EEG feature spectra in the benchmark feature subsets under each brain region need to be carefully merged. This process will continue until the number of benchmark feature subsets is exactly the same as the number of brain regions, and then the merging process is stopped. In the case of shortage classification, it is necessary to increase the amount of multi-lead EEG data collected. By increasing the amount of data, feature extraction and classification are re-performed to ensure that the number of benchmark feature subsets can match the number of brain regions. This process will be repeated until the requirement is met.

[0088] In a preferred embodiment, the step of merging the EEG feature spectra in the reference feature subsets of each brain region includes:

[0089] Obtain all benchmark feature subsets under the same brain region, and sum and average the similarity scores of the EEG feature spectra in each benchmark feature subset to obtain the merging condition parameters;

[0090] Perform positive offset processing on the screening threshold to obtain the permissible merging threshold;

[0091] The baseline feature subsets under the merging condition parameters that are higher than the permissible merging threshold are directly merged, and the EEG feature spectra in the baseline feature subsets under the merging condition parameters that are lower than the permissible merging threshold are recorded as edge feature spectra;

[0092] All edge feature spectra under adjacent brain regions are merged and recorded as edge feature subsets.

[0093] In this embodiment, when processing the EEG feature spectra in the benchmark feature subsets under each brain region, all benchmark feature subsets under the same brain region are first obtained, and then the corresponding similarity scores in the benchmark feature subsets are summed and averaged, so as to obtain a merging condition parameter. After obtaining the merging condition parameter, it is necessary to perform a positive offset processing on the screening threshold to obtain an allowable merging threshold. The allowable merging threshold will be used as a basis for judging whether the benchmark feature subsets can be merged. Specifically, the benchmark feature subsets under the merging condition parameters above the allowable merging threshold will be directly merged, and those below the allowable merging threshold will be directly merged. The baseline feature subset under the merging condition parameter of the value will record the EEG feature spectrum as the edge feature spectrum. Finally, all the edge feature spectra under the adjacent brain areas will be merged and the merged results will be recorded as the edge feature subset. Based on this, the EEG feature spectra of each brain area can be effectively integrated, so as to provide more accurate and comprehensive data support for subsequent analysis and processing. Among them, the edge feature spectrum corresponds to the adjacent brain areas. In actual applications, when the EEG feature spectrum in the baseline feature subset fluctuates normally but the user reacts abnormally, intervention can be made from the direction of the edge feature spectrum to ensure that the user's EEG activity is effectively regulated.

[0094] In a preferred embodiment, the step of screening the EEG feature spectra in each benchmark feature subset to obtain the standard fluctuation amplitude range corresponding to the EEG feature spectra under each benchmark feature subset includes:

[0095] S301, obtaining EEG feature spectra in the benchmark feature subset and similarity scores between the EEG feature spectra, and recording them as screening condition parameters;

[0096] S302, performing bidirectional offset processing on each screening condition parameter to obtain a screening condition interval;

[0097] S303, counting the number of screening condition parameters in each screening condition interval, and recording them as condition parameters to be evaluated;

[0098] S304, arranging the condition parameters to be evaluated in descending order, and recording the EEG characteristic spectrum corresponding to the screening condition interval under the condition parameter to be evaluated with the highest ranking as the standard characteristic spectrum;

[0099] S305, collecting the fluctuation amplitude of the standard characteristic spectrum and recording it as the standard fluctuation amplitude range of the EEG characteristic spectrum.

[0100] As described in the above steps S301-S305, in the process of screening the EEG feature spectra in each benchmark feature subset, it is first necessary to obtain the EEG feature spectra in the benchmark feature subset and the similarity scores between the EEG feature spectra, and record them as screening condition parameters, and then perform bidirectional offset processing on these screening condition parameters to obtain a series of screening condition intervals. After obtaining the screening condition intervals, we need to count the number of screening condition parameters under each screening condition interval and record them as condition parameters to be evaluated. In order to ensure that the selected EEG feature spectrum is representative, all condition parameters to be evaluated are arranged in descending order, and the EEG feature spectrum corresponding to the screening condition interval corresponding to the highest ranked condition parameter to be evaluated is defined as the standard feature spectrum. Finally, according to the fluctuation amplitude of the standard feature spectrum (the fluctuation amplitude of the EEG feature spectrum in the screening condition interval corresponding to the highest ranked condition parameter to be evaluated), it is recorded as the standard fluctuation amplitude interval of the EEG feature spectrum. The standard fluctuation amplitude interval will provide a reference standard for evaluating whether other EEG feature spectra meet the requirements.

[0101] In a preferred embodiment, the step of identifying the EEG activity state of the target brain area according to the standard fluctuation amplitude range includes:

[0102] Obtain the EEG characteristic spectrum of the target brain area and record it as the characteristic spectrum to be corrected;

[0103] Compare the characteristic spectrum to be calibrated with the standard fluctuation amplitude range;

[0104] If the fluctuation amplitude of the characteristic spectrum to be corrected is within the standard fluctuation amplitude range, the EEG activity state of the target brain area is determined to be normal;

[0105] If the fluctuation amplitude of the characteristic spectrum to be corrected exceeds the standard fluctuation amplitude range, the EEG activity state of the target brain area is determined to be abnormal.

[0106] In this embodiment, in order to accurately identify the EEG activity state of the target brain area, it is first necessary to obtain the EEG characteristic spectrum of the target brain area and record it as the characteristic spectrum to be proofread. Next, the characteristic spectrum to be proofread is carefully compared with the pre-set standard fluctuation amplitude range. During the comparison process, if the fluctuation amplitude of the characteristic spectrum to be proofread falls within the standard fluctuation amplitude range, then it can be determined that the EEG activity state of the target brain area is normal, which means that the EEG activity is within a healthy and normal range and shows no signs of abnormality. However, if the fluctuation amplitude of the characteristic spectrum to be proofread exceeds the standard fluctuation amplitude range, then the EEG activity state of the target brain area will be determined to be abnormal. This situation may indicate that there are some potential neurological problems or abnormalities, which require further inspection and analysis. Through the above steps, we can effectively identify and judge the EEG activity state of the target brain area, thereby providing an important reference basis for subsequent diagnosis and guidance.

[0107] In a preferred embodiment, the step of prompting the training subject to adjust the EEG activity through a feedback mechanism until the abnormal brain area returns to a normal state includes:

[0108] Obtaining the EEG characteristic spectrum of the abnormal brain area and recording it as the abnormal characteristic spectrum;

[0109] Generate a strategy for adjusting brain electrical activity according to the difference between the abnormal characteristic spectrum and the standard characteristic spectrum;

[0110] Feedback the EEG activity adjustment strategy to the training subjects through vision, hearing and touch to guide the EEG activity of the target brain area;

[0111] The EEG activity of the target brain area is continuously monitored, and the EEG characteristic spectrum of the abnormal brain area is updated in real time. When the fluctuation amplitude of the EEG characteristic spectrum of the abnormal brain area returns to the standard fluctuation amplitude range, it is confirmed that the EEG activity state has returned to normal and the feedback adjustment is ended.

[0112] In this embodiment, when the EEG state of the training subject is abnormal, it is necessary to guide the training subject to adjust their EEG activity until the abnormal brain area returns to a normal state. First, it is necessary to obtain the EEG characteristic spectrum of the abnormal brain area and record it as an abnormal characteristic spectrum. Then, the difference between the abnormal characteristic spectrum and the standard characteristic spectrum is compared. Through this comparison, an adjustment strategy for EEG activity can be generated to help the training subject guide their EEG activity to a normal state. Then, the EEG activity adjustment strategy is fed back to the training subject through multiple senses, including vision, hearing and touch. Through these feedbacks, the training subject can better understand how to guide the target brain area to perform appropriate EEG activity. During the training process, the EEG activity of the target brain area is continuously monitored, and the EEG characteristic spectrum of the abnormal brain area is updated in real time to ensure that the adjustment strategy can respond to changes in EEG activity in a timely manner. When the fluctuation amplitude of the EEG characteristic spectrum of the abnormal brain area gradually returns to the standard fluctuation amplitude range, it can be confirmed that the EEG activity state has returned to normal. At this time, the feedback adjustment can be ended, marking the successful completion of the training process.

[0113] See also Figure 2 , an EEG feedback training system based on multi-lead EEG feature analysis, using the above-mentioned EEG feedback training method based on multi-lead EEG feature analysis, comprising:

[0114] A feature extraction module is used to preprocess the collected multi-lead EEG data, and perform feature extraction on the preprocessed multi-lead EEG data to extract EEG feature spectrum related to EEG activity;

[0115] A feature classification module is used to classify the EEG feature spectra of different brain regions to obtain multiple benchmark feature subsets;

[0116] A standard construction module is used to screen and process the EEG feature spectrum in each benchmark feature subset to obtain a standard fluctuation amplitude range corresponding to the EEG feature spectrum under each benchmark feature subset;

[0117] A state evaluation module, which is used to identify the EEG activity state of the target brain area according to the standard fluctuation amplitude range, wherein the EEG activity state includes a normal state and an abnormal state;

[0118] The first intervention module is used to record the corresponding brain area as a normal brain area under normal conditions, monitor the EEG characteristic spectrum of the normal brain area in real time, and adjust the training parameters in real time according to the monitoring results to maintain the EEG activity within the standard fluctuation range;

[0119] The second intervention module is used to record the corresponding brain area as an abnormal brain area under abnormal conditions, and prompt the training subject to adjust the EEG activity through a feedback mechanism until the abnormal brain area returns to normal.

[0120] In the above, the system includes a feature extraction module, a feature classification module, a standard construction module, a state evaluation module, a first intervention module and a second intervention module. The feature extraction module is responsible for the preliminary preprocessing of the collected multi-lead EEG data, including noise removal, filtering and other steps to ensure the quality of the data. After the preprocessing is completed, the feature extraction module will further analyze the data, extract the characteristic spectrum closely related to the EEG activity, and record it as the EEG characteristic spectrum. The feature classification module will classify the EEG characteristic spectrum of different brain regions. Through this process, the system can divide the EEG data into multiple benchmark feature subsets, each benchmark feature subset represents the activity characteristics of a specific brain region in the brain. The standard construction module further screens and processes the EEG characteristic spectrum in each benchmark feature subset. The goal is to determine the standard fluctuation amplitude range corresponding to the EEG characteristic spectrum under each benchmark feature subset. These standard fluctuation amplitude ranges provide a reference for evaluating the state of brain activity. The system can more accurately determine whether the brain is in a normal state. The state evaluation module identifies and evaluates the EEG activity state of the target brain area according to the standard fluctuation amplitude range. The evaluation results can be divided into normal state and abnormal state. The first intervention module takes effect when the target brain area is in a normal state, and records the corresponding brain area as a normal brain area, and monitors the EEG characteristic spectrum of these brain areas in real time. Through continuous monitoring, the first intervention module can adjust the training parameters in real time according to the monitoring results to ensure that the EEG activity remains within the standard fluctuation amplitude range, which helps to maintain the optimal functional state of the brain. The second intervention module takes effect when the target brain area is in an abnormal state, and records the corresponding brain area as an abnormal brain area, and prompts the training subject to make corresponding adjustments through a feedback mechanism, including changing thinking patterns, adjusting breathing rhythms or other methods that help improve EEG activity. The goal of the second intervention module is to help the training subject gradually restore the abnormal brain area to a normal state, thereby improving overall brain function.

[0121] See also Figure 3 , an electronic device, the electronic device comprising:

[0122] at least one processor;

[0123] and a memory communicatively coupled to the at least one processor;

[0124] Among them, the memory stores a computer program that can be executed by at least one processor, and the computer program is executed by at least one processor so that the at least one processor can execute the above-mentioned EEG feedback training method based on multi-lead EEG feature analysis.

[0125] In the above, the processor of the electronic device may be

[0126] The central processing unit (CPU), graphics processing unit (GPU) or application-specific integrated circuit (ASIC), etc., are responsible for executing the computer program stored in the memory to implement the various steps of the EEG feedback training method. The memory of the electronic device can be a random access memory (RAM), a read-only memory (ROM), a flash memory or other forms of non-volatile memory, which is used to store computer programs and related data. In addition, the electronic device may also include input and output devices and an operator, so that the user can interact with the system and perform necessary computing tasks. The input and output devices may include a keyboard, a mouse, a touch screen, a display, a printer, etc., to provide the user with an operating interface and output results. The operator is responsible for performing more complex mathematical operations to ensure the efficient operation of the system. The interface is used to connect external devices, such as an electroencephalogram (EEG) electrode cap, a display, a speaker, etc., to achieve communication with the user.

[0127] It should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, device, article or method including a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, device, article or method. In the absence of further restrictions, an element defined by the sentence "includes a ..." does not exclude the presence of other identical elements in the process, device, article or method including the element.

[0128] The above is only a preferred embodiment of the present invention. It should be noted that, for those skilled in the art, several improvements and modifications can be made without departing from the principles of the present invention, and these improvements and modifications should also be considered as the protection scope of the present invention. The structures, devices and operating methods not specifically described and explained in the present invention shall be implemented according to the conventional means in the art unless otherwise specified and limited.

Claims

1. An EEG feedback training system based on multi-lead EEG feature analysis, characterized in that: include: A feature extraction module, which is used to preprocess the collected multi-lead EEG data, and perform feature extraction on the preprocessed multi-lead EEG data to extract EEG feature spectrum related to EEG activity; A feature classification module, wherein the feature classification module is used to classify the EEG feature spectra of different brain regions to obtain a plurality of benchmark feature subsets; A standard construction module, wherein the standard construction module is used to screen and process the EEG characteristic spectra in each of the reference feature subsets to obtain a standard fluctuation amplitude interval corresponding to the EEG characteristic spectra under each of the reference feature subsets; A state evaluation module, the state evaluation module is used to identify the EEG activity state of the target brain area according to the standard fluctuation amplitude interval, wherein the EEG activity state includes a normal state and an abnormal state; A first intervention module, wherein the first intervention module is used to record the corresponding brain area as a normal brain area in the normal state, monitor the EEG characteristic spectrum of the normal brain area in real time, and adjust the training parameters in real time according to the monitoring results to maintain the EEG activity within the standard fluctuation range; The second intervention module is used to record the corresponding brain area as an abnormal brain area under the abnormal state, and prompt the training subject to adjust the EEG activity through a feedback mechanism until the abnormal brain area returns to a normal state.

2. The EEG feedback training system based on multi-lead EEG feature analysis according to claim 1, characterized in that: The extracting of the EEG characteristic spectrum related to the EEG activity specifically includes: Acquire multi-lead EEG data, and perform filtering on the multi-lead EEG data to remove noise and artifacts; Performing dimensionality reduction processing on the filtered multi-lead EEG data, converting the multi-lead EEG data in a multi-dimensional state into multi-lead EEG data in a low-dimensional state; The frequency domain analysis of the multi-lead EEG data after dimensionality reduction is performed using Fourier transform, the time domain signal is converted into a frequency domain signal, and the EEG characteristic spectrum is extracted.

3. The EEG feedback training system based on multi-lead EEG feature analysis according to claim 1, characterized in that: The classification and processing of the EEG characteristic spectra of different brain regions to obtain a plurality of benchmark characteristic subsets specifically includes: Acquire all the EEG characteristic spectra, and perform vectorization processing to obtain the EEG characteristic vectors of each of the EEG characteristic spectra; Obtaining a classification function, and inputting the EEG feature vectors into the classification function for one-to-one comparison to obtain similarity scores between the EEG feature vectors; Obtaining a screening threshold, and comparing the screening threshold with the similarity score, and classifying the EEG feature spectra corresponding to the EEG feature vectors whose similarity scores are greater than the screening threshold as the same type of EEG feature spectra, and then aggregating them into the same benchmark feature subset; conversely, classifying the EEG feature spectra corresponding to the EEG feature vectors as different types of EEG feature spectra, and classifying them into different benchmark feature subsets; Each of the reference feature subsets is added with a unique feature identifier, and each of the reference feature subsets corresponds to a brain region.

4. The EEG feedback training system based on multi-lead EEG feature analysis according to claim 3, characterized in that: After the benchmark feature subsets are output, if the number of the benchmark feature subsets is higher than the number of brain regions, it indicates that there are redundant classifications in the EEG feature spectrum; if the number of the benchmark feature subsets is lower than the number of brain regions, it indicates that there are missing classifications in the EEG feature spectrum; Under the redundant classification, the EEG feature spectra in the benchmark feature subsets under each brain region are merged and processed until the number of the benchmark feature subsets is consistent with the number of brain regions; Under the shortage classification, the amount of multi-lead EEG data collected is increased, and feature extraction and classification processing are re-performed until the number of baseline feature subsets matches the number of brain regions.

5. The EEG feedback training system based on multi-lead EEG feature analysis according to claim 4, characterized in that: The merging process of the EEG feature spectra in the reference feature subsets under each brain region specifically includes: Obtaining all benchmark feature subsets under the same brain region, and summing and averaging the similarity scores of the EEG feature spectra in each of the benchmark feature subsets to obtain a merging condition parameter; Performing a positive offset process on the screening threshold to obtain an allowable merging threshold; The reference feature subsets under the merging condition parameters higher than the allowable merging threshold are directly merged, and the EEG feature spectra in the reference feature subsets under the merging condition parameters lower than the allowable merging threshold are recorded as edge feature spectra; All the edge feature spectra under adjacent brain regions are merged and recorded as edge feature subsets.

6. The EEG feedback training system based on multi-lead EEG feature analysis according to claim 1, characterized in that: The EEG feature spectra in each of the reference feature subsets are screened to obtain the standard fluctuation amplitude range corresponding to the EEG feature spectra under each of the reference feature subsets, specifically including: Obtaining the EEG feature spectra in the benchmark feature subset and the similarity scores between the EEG feature spectra, and recording them as screening condition parameters; Performing bidirectional offset processing on each of the screening condition parameters to obtain a screening condition interval; Counting the number of screening condition parameters under each of the screening condition intervals and recording them as condition parameters to be evaluated; Arrange the condition parameters to be evaluated in descending order, and record the EEG characteristic spectrum corresponding to the screening condition interval under the condition parameter to be evaluated with the highest ranking as the standard characteristic spectrum; The fluctuation amplitude of the standard characteristic spectrum is collected and recorded as the standard fluctuation amplitude range of the EEG characteristic spectrum.

7. The EEG feedback training system based on multi-lead EEG feature analysis according to claim 1, characterized in that: The step of identifying the EEG activity state of the target brain area according to the standard fluctuation amplitude range specifically includes: Obtain the EEG characteristic spectrum of the target brain area and record it as the characteristic spectrum to be corrected; Comparing the characteristic spectrum to be calibrated with the standard fluctuation amplitude range; If the fluctuation amplitude of the characteristic spectrum to be corrected is within the standard fluctuation amplitude range, it is determined that the EEG activity state of the target brain area is normal; If the fluctuation amplitude of the characteristic spectrum to be corrected exceeds the standard fluctuation amplitude range, it is determined that the EEG activity state of the target brain area is abnormal.

8. The EEG feedback training system based on multi-lead EEG feature analysis according to claim 1, characterized in that: The step of prompting the training subject to adjust the EEG activity through the feedback mechanism until the abnormal brain area returns to a normal state specifically includes: Acquiring the EEG characteristic spectrum of the abnormal brain area and recording it as the abnormal characteristic spectrum; generating a brain electrical activity adjustment strategy according to the difference between the abnormal characteristic spectrum and the standard characteristic spectrum; Feedback the EEG activity adjustment strategy to the training subject through vision, hearing and touch to guide the EEG activity of the target brain area; The EEG activity of the target brain area is continuously monitored, and the EEG characteristic spectrum of the abnormal brain area is updated in real time. When the fluctuation amplitude of the EEG characteristic spectrum of the abnormal brain area returns to the standard fluctuation amplitude range, it is confirmed that the EEG activity state has returned to normal and the feedback adjustment is ended.

9. An electronic device, characterized in that: The electronic device comprises: at least one processor; and a memory communicatively coupled to the at least one processor; Wherein, the memory stores a computer program that can be executed by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the EEG feedback training system based on multi-lead EEG feature analysis as described in any one of claims 1 to 8.

Images