Self-adaptive scalp electroencephalogram motor imagery signal identification method and related device

Abstract

The invention belongs to the technical field of brain-computer interfaces, relates to an electroencephalogram signal processing technology, and discloses a self-adaptive scalp electroencephalogram motor imagery signal recognition method and a related device. The self-adaptive scalp electroencephalogram motor imagery signal identification method comprises the following steps: acquiring a to-be-identified scalp electroencephalogram motor imagery signal of a target tested object; and based on the scalp electroencephalogram motor imagery signal to be identified, performing signal identification by using a trained adaptive classifier identification model to obtain a scalp electroencephalogram motor imagery signal identification result. According to the technical scheme, the problem of individual differences in EEG signal processing can be better solved through the adaptive classifier recognition model, and motor imagery signals of different individuals can be recognized more accurately.

Description

Technical Field

[0001] The present invention belongs to the technical field of brain-computer interface, relates to electroencephalogram (EEG) signal processing technology, and particularly relates to an adaptive scalp EEG motor imagery signal recognition method and related devices. Background Art

[0002] With the development of science and technology, BCI (Brain Computer Interface) is increasingly used in various fields, especially the BCI paradigm based on motor imagery, which can be used in medical, education and other fields. EEG (Electroencephalogram) motor imagery classification is an important research direction in the field of BCI. It allows users to control external devices by imagining a certain movement (for example, left or right hand movement) by analyzing and identifying motor imagery patterns in EEG signals.

[0003] In recent years, deep learning methods, especially CNN (Convolution neural network) and RNN (Recurrent neural network), have been widely used in EEG motor imagery classification. These methods can automatically extract high-level features of EEG signals and achieve high classification accuracy.

[0004] Deep learning methods have achieved good results in EEG motor imagery classification, but in practical applications, they still have some defects, mainly including: (1) In order to achieve high classification accuracy, existing deep learning methods usually require a large amount of labeled data for training, which leads to a long calibration process; (2) The EEG signal patterns of different individuals may be different. Existing deep learning methods often require training models for each individual separately, which increases the complexity and computational cost of the model. Summary of the invention

[0005] The purpose of the present invention is to provide an adaptive scalp EEG motor imagery signal recognition method and related devices to solve one or more of the above-mentioned technical problems. In the technical solution provided by the present invention, the problem of individual differences in EEG signal processing can be better solved by an adaptive classifier recognition model, and motor imagery signals of different individuals can be more accurately recognized.

[0006] In order to achieve the above object, the present invention adopts the following technical solutions: In a first aspect, the present invention provides an adaptive scalp EEG motor imagery signal recognition method, comprising the following steps: Acquire the scalp EEG motion imagery signal to be identified of the target subject; Based on the scalp EEG motor imagery signal to be identified, signal recognition is performed using a trained adaptive classifier recognition model to obtain a scalp EEG motor imagery signal recognition result; The training and acquisition steps of the trained adaptive classifier recognition model include: Based on the acquired scalp EEG motion imagery signal samples, multiple basic general recognition models are trained by loading the data set with the same paradigm as the recognition target; Among multiple basic universal recognition models, the basic universal recognition model with the lowest verification loss of the scalp EEG motion imagery signal samples of the target subject is used as the pre-training model; The feature extractor module and pre-training parameters of the pre-trained model are read and a model migration fine-tuning strategy is adopted to train an untrained classifier using the scalp EEG motor imagery signal samples of the target subject to obtain a trained adaptive classifier recognition model.

[0007] A further improvement of the present invention is that the step of training and obtaining multiple basic universal recognition models based on the acquired scalp EEG motor imagery signal samples by loading a data set of the same paradigm as the recognition target comprises: Sampling the EEG signals of multiple sample subjects to obtain original EEG signal sample data of each sample subject; The original EEG signal sample data of each sample subject are preprocessed respectively to obtain the training set, validation set and test set of each sample subject; wherein the preprocessing process is, based on the original EEG signal sample data, selecting EEG channel data related to the motor imagery area; filtering and downsampling the selected EEG channel data to obtain two-dimensional data of channel × time; converting the two-dimensional data of channel × time into three-dimensional data of two-dimensional topology map × time, and dividing to obtain the training set, validation set and test set; Based on the training set, validation set and test set of each sample subject, the selected recognition model is trained and updated respectively to obtain multiple basic general recognition models; in the training and updating process, the weight and bias of the recognition model are adjusted by continuously evaluating the loss value of the validation data in each Epoch to complete the parameter optimization and updating of the recognition model.

[0008] A further improvement of the present invention is that, in the step of respectively training and updating the selected recognition model based on the training set, validation set and test set of each sample subject to obtain multiple basic universal recognition models, During the training update process, if the loss of the validation data does not decrease in the preset number of epochs, the early stopping mechanism is triggered and the recognition model stops training.

[0009] A further improvement of the present invention is that the step of reading the feature extractor module and pre-training parameters of the pre-trained model and adopting a model migration fine-tuning strategy to train an untrained classifier using the scalp EEG motor imagery signal samples of the target subject to obtain a trained adaptive classifier recognition model specifically includes: Reading the feature extractor module and pre-training parameters of the pre-trained model and adopting a model migration fine-tuning strategy, using the scalp EEG motor imagery signal samples of the target subject to train an untrained classifier to obtain a trained adaptive classifier recognition model; The performance of the trained adaptive classifier recognition model is evaluated using four indicators: accuracy, specificity, sensitivity, and f1-score. Once the preset performance requirements are met, the trained adaptive classifier recognition model is obtained.

[0010] The second aspect of the present invention provides an adaptive scalp EEG motor imagery signal recognition system, comprising: A signal acquisition module, used for acquiring the scalp EEG motion imagery signal to be identified of the target subject; A signal recognition module, used to perform signal recognition based on the scalp EEG motor imagery signal to be recognized by using a trained adaptive classifier recognition model to obtain a scalp EEG motor imagery signal recognition result; The training and acquisition steps of the trained adaptive classifier recognition model include: Based on the acquired scalp EEG motion imagery signal samples, multiple basic general recognition models are trained by loading the data set with the same paradigm as the recognition target; Among multiple basic universal recognition models, the basic universal recognition model with the lowest verification loss of the scalp EEG motion imagery signal samples of the target subject is used as the pre-training model; The feature extractor module and pre-training parameters of the pre-trained model are read and a model migration fine-tuning strategy is adopted to train an untrained classifier using the scalp EEG motor imagery signal samples of the target subject to obtain a trained adaptive classifier recognition model.

[0011] A further improvement of the present invention is that the step of training and obtaining multiple basic universal recognition models based on the acquired scalp EEG motor imagery signal samples by loading a data set of the same paradigm as the recognition target comprises: Sampling the EEG signals of multiple sample subjects to obtain original EEG signal sample data of each sample subject; The original EEG signal sample data of each sample subject are preprocessed respectively to obtain the training set, validation set and test set of each sample subject; wherein the preprocessing process is, based on the original EEG signal sample data, selecting EEG channel data related to the motor imagery area; filtering and downsampling the selected EEG channel data to obtain two-dimensional data of channel × time; converting the two-dimensional data of channel × time into three-dimensional data of two-dimensional topology map × time, and dividing to obtain the training set, validation set and test set; Based on the training set, validation set and test set of each sample subject, the selected recognition model is trained and updated respectively to obtain multiple basic general recognition models; in the training and updating process, the weight and bias of the recognition model are adjusted by continuously evaluating the loss value of the validation data in each Epoch to complete the parameter optimization and updating of the recognition model.

[0012] A further improvement of the present invention is that, in the step of respectively training and updating the selected recognition model based on the training set, validation set and test set of each sample subject to obtain multiple basic universal recognition models, During the training update process, if the loss of the validation data does not decrease in the preset number of epochs, the early stopping mechanism is triggered and the recognition model stops training.

[0013] A further improvement of the present invention is that the step of reading the feature extractor module and pre-training parameters of the pre-trained model and adopting a model migration fine-tuning strategy to train an untrained classifier using the scalp EEG motor imagery signal samples of the target subject to obtain a trained adaptive classifier recognition model specifically includes: Reading the feature extractor module and pre-training parameters of the pre-trained model and adopting a model migration fine-tuning strategy, using the scalp EEG motor imagery signal samples of the target subject to train an untrained classifier to obtain a trained adaptive classifier recognition model; The performance of the trained adaptive classifier recognition model is evaluated using four indicators: accuracy, specificity, sensitivity, and f1-score. Once the preset performance requirements are met, the trained adaptive classifier recognition model is obtained.

[0014] In a third aspect of the present invention, an electronic device is provided, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the adaptive scalp EEG motor imagery signal recognition method as described in any one of the first aspects of the present invention is implemented.

[0015] According to a fourth aspect of the present invention, a non-transitory computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the method for adaptive scalp EEG motor imagery signal recognition as described in any one of the first aspects of the present invention is implemented.

[0016] Compared with the prior art, the present invention has the following beneficial effects: In view of the technical problems existing in the above-mentioned existing deep learning methods, the present invention discloses an adaptive scalp EEG motor imagery signal recognition method, which recognizes scalp EEG motor imagery signals based on an adaptive classifier recognition model, can more accurately recognize motor imagery signals of different individuals, solves the problem of individual differences in EEG signal processing and controls costs. Explanatory, the present invention is based on a transfer learning method, firstly obtains a pre-trained model through basic general model selection, and then performs model migration fine-tuning on it to adapt to new tasks; in addition, the fine-tuning strategy can quickly adapt to new tasks by freezing some layers of the pre-trained model and only updating the weights of a few layers, so as to meet the needs of cross-object scalp EEG motor imagery signal recognition. In summary, the technical solution of the present invention obtains an adaptive classifier recognition model through a model transfer learning strategy. The adaptive model transfer learning strategy can significantly improve the decoding efficiency of EEG signals. Signal recognition using an adaptive classifier recognition model can effectively cope with the differences in EEG signal characteristics between individuals and improve recognition accuracy. In addition, the present invention pre-trains a basic universal recognition model, which can significantly reduce the number of samples required for new target subjects. Only very little sample data is needed for fine-tuning training. The fine-tuned model can adapt to the EEG signal pattern of the new target subject, greatly reducing the time and cost of sample data collection. Furthermore, by using the feature extractor module that has been learned in the pre-trained model, high-quality model fine-tuning can be completed in a relatively short time, which not only improves the adaptability and robustness of the model, but also ensures the stability and accuracy of the model. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below; obviously, the drawings described below are some embodiments of the present invention, and for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0018] Figure 1 It is a flow chart of an adaptive scalp EEG motor imagery signal recognition method in an embodiment of the present invention.

[0019] Figure 2 It is a schematic diagram of the principle of an adaptive scalp EEG motor imagery signal recognition method in a specific embodiment of the present invention.

[0020] Figure 3 It is a schematic diagram of three-dimensional data representation in an embodiment of the present invention.

[0021] Figure 4 is a schematic diagram of a pre-training and fine-tuning migration model framework in an embodiment of the present invention; wherein, Figure 4 (a) is a schematic diagram of the pre-training model. Figure 4 (b) is a schematic diagram of the transfer learning model.

[0022] Figure 5 It is a schematic diagram of an adaptive scalp EEG motor imagery signal recognition system in an embodiment of the present invention. DETAILED DESCRIPTION

[0023] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings in the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0024] See also Figure 1 In an embodiment of the present invention, a method for adaptively recognizing scalp electroencephalogram motor imagery signals is provided, comprising the following steps: Step 1, obtaining the scalp EEG motion imagery signal to be identified of the target subject; Step 2, based on the scalp EEG motor imagery signal to be identified, using the trained adaptive classifier recognition model to perform signal recognition to obtain a scalp EEG motor imagery signal recognition result; The training and acquisition steps of the trained adaptive classifier recognition model include: Based on the acquired scalp EEG motor imagery signal samples, multiple basic general recognition models are trained by loading a data set with the same paradigm as the recognition target; for example, a public data set OpenBMI data set is loaded, and 64 channels of EEG left and right hand motor imagery data are collected according to the 10-20 system configuration. After signal noise reduction and filtering, the EEG data is input into the selected recognition network for training and updating to obtain a basic general recognition model; Among multiple basic universal recognition models, the basic universal recognition model with the lowest verification loss of the scalp EEG motion imagery signal samples of the target subject is used as the pre-training model; The feature extractor module and pre-training parameters of the pre-trained model are read and a model migration fine-tuning strategy is adopted to train an untrained classifier using the scalp EEG motor imagery signal samples of the target subject to obtain a trained adaptive classifier recognition model.

[0025] The significant improvements of the technical solutions disclosed in the embodiments of the present invention include: (1) Solving the problem of individual differences and improving decoding efficiency; Explanatoryally, the technical solution of the embodiment of the present invention obtains an adaptive classifier recognition model through a model transfer fine-tuning strategy, and the subsequent use of the model for signal recognition can effectively deal with the differences in EEG signal characteristics between individuals and improve the accuracy of recognition; the present invention adopts an adaptive model transfer fine-tuning strategy to significantly improve the decoding efficiency of EEG signals; (2) The present invention pre-trains the basic universal recognition model, which can significantly reduce the number of samples required for new target subjects. Only a small amount of sample data is needed for fine-tuning, so that the fine-tuned model can adapt to the EEG signal pattern of the new target subject, greatly reducing the time and cost of sample data collection; (3) The model transfer fine-tuning strategy enables the model to dynamically adjust parameters to adapt to the individual's EEG signal characteristics, thereby improving the adaptability and robustness of the model. The EEG signal of each subject is unique, and the fine-tuning process can capture these characteristics and make effective adjustments. In addition, through transfer learning, the feature extractor module that has been learned in the pre-trained model can be used to complete high-quality model fine-tuning in a shorter time. This not only improves the performance of the model, but also ensures the stability and accuracy of the model.

[0026] See also Figure 2 and Figure 3 In an embodiment of the present invention, an adaptive scalp EEG motor imagery signal recognition method is provided, and the main process is as follows: Sample data collection, including: sampling the EEG signals of the subjects using EEG acquisition equipment of the international 10-20 standard based on the electrode distribution, with a sampling frequency of 1000 Hz; a total of 64 channels of EEG data were collected; The sample data preprocessing includes: after obtaining the original EEG signal sample data of the subject through sample data collection, preprocessing the original EEG signal sample data; in a specific exemplary scheme, the preprocessing step specifically includes: selecting EEG channels related to the motor imagery area (FC5, FC3, FC1, FC2, FC4, FC6, T7, T8, C5, C3, C1, C2, C4, C6, Cz, TP7, TP8, CP5, CP3, CP1, CPz, CP2, CP4, CP6, P 7, P3, P1, Pz, P2, P4, P8), the above EEG channels are named according to the international 10-20 standard EEG channels; the selected 31 channel data are filtered using an 8-30 Hz bandpass filter; the original EEG signal sampling rate is downsampled from 1000 Hz to 400 Hz using the downsampling method; finally, the processed two-dimensional data (channel × time) is converted into three-dimensional data (two-dimensional topology × time), and the data is divided into training set, validation set and test set according to the leave-one-out and five-fold cross-validation methods.

[0027] Basic general recognition model training includes: inputting training data and verification data into the model in a certain ratio (for example, using 5-fold cross validation, 80% for training and 20% for verification); wherein, during the training process, by continuously evaluating the loss value of the verification data in each Epoch (i.e., a cycle in the training process), adjusting the weight and bias of the training model, and optimizing the parameters of the training model; in the preferred scheme of the embodiment of the present invention, in order to reduce the time of model training, an early stopping mechanism is added to the training process, and the early stopping mechanism determines whether the model needs to stop by monitoring the loss change of the verification data; specifically, exemplarily, the tolerance number of the early stopping mechanism is set to 30 times, that is, if the loss of the verification data does not decrease in 30 epochs, the early stopping mechanism is triggered and the model stops training.

[0028] Model transfer training includes: taking the basic universal recognition model with the lowest validation data loss in the five-fold cross validation of each subject as a pre-training model, and reading the feature extractor module of the pre-training model and its related pre-training parameters, so as to effectively process the motor imagery action of the same paradigm. However, considering the differences of the target subjects, directly using the pre-training model may not be accurate enough. Therefore, using the trained pre-training feature extraction parameters as the initial state, setting the hyperparameter tuning of the model without any specific prior knowledge, loading the target subject data for transfer learning training, and flexibly adapting to the feature distribution of the target domain; through the above operations, the advantages of the basic universal recognition model can be effectively utilized, while the personalized data feature distribution of the target subject can be processed, and finally an adaptive classifier recognition model is generated. It is particularly emphasized that this is the core improvement part of the technical solution of the embodiment of the present invention. The classifier that starts with random initialization has no prior knowledge and is more easily adapted to the data distribution of the target domain, solving the problem of individual differences between different subjects; afterwards, using part of the target subject's data to fine-tune the feature extractor and classifier modules, so that it can better learn the data distribution of the target subject.

[0029] Model evaluation includes: using accuracy, specificity, sensitivity, and f1-score to evaluate the performance of the adaptive classifier recognition model.

[0030] The model application includes: applying the adaptive classifier recognition model to the real-time EEG acquisition system, and when the user performs the corresponding motor imagery behavior, the system performs corresponding feedback through the results predicted by the adaptive classifier recognition model.

[0031] The advantage of the technical solution provided by the embodiment of the present invention is that when the target domain data is used to fine-tune the pre-trained model, the technical solution of the present invention adds a new classifier layer. This design enables the new classifier layer to be optimized specifically for the target subject, so that the features in the EEG signal can be learned and captured more accurately; the traditional transfer learning method has challenges in the individual differences across subjects, because the classification layer of the pre-trained model is designed based on the data of multiple subjects and may not be suitable for specific subjects. The technical solution of the embodiment of the present invention ensures that the recognition model can better adapt to the unique characteristics of each target subject by adding a new classifier layer to each target subject, which can significantly improve the classification accuracy; since the new classifier layer is trained based on the data of the target subject, it can better capture and understand the EEG signal pattern of the subject, which not only enhances the recognition ability of the recognition model for the target subject, but also improves the generalization ability of the model when facing new and unseen data. In the further preferred technical solution, the introduction of the pre-trained model and the combination of the early stopping mechanism significantly reduce the time for model training and fine-tuning; compared with training the model from scratch, the fine-tuning strategy enables the model to achieve ideal performance faster, accelerating the entire development and application process. The fine-tuning strategy is particularly suitable for dealing with situations with large individual differences. There may be significant differences in the EEG signals of each subject. Through fine-tuning, the model can be optimized for specific individuals, solving the problem that general models are difficult to cope with individual differences. The pre-trained model can reuse the data of a large number of subjects. This sharing mechanism effectively utilizes existing data resources, improves data utilization efficiency and the versatility of the model.

[0032] In the specific comparative examples of the present invention, the comparison of different experimental methods is demonstrated, and the comparison results of the specific analysis are shown in Table 1.

[0033] Table 1. Experimental comparison table

[0034] See also Figure 4 and Table 1. The pre-training in Table 1 represents the results of the pre-training model, scheme-1 represents the result of fine-tuning the original classifier, and scheme-2 to scheme-6 represent the adaptive classifier schemes with different numbers of layers added, such as Figure 4 As shown in (b) in the figure; it can be observed from Table 1 that the model performance of the strategy of adding an adaptive classifier is better than the effect of only fine-tuning the original classifier, which proves the effectiveness of the technical solution of the embodiment of the present invention.

[0035] The following are device embodiments of the present invention, which can be used to implement the method embodiments of the present invention. For details not disclosed in the device embodiments, please refer to the method embodiments of the present invention.

[0036] See also Figure 5In an embodiment of the present invention, an adaptive scalp EEG motor imagery signal recognition system is provided, comprising: A signal acquisition module, used for acquiring the scalp EEG motion imagery signal to be identified of the target subject; A signal recognition module, used to perform signal recognition based on the scalp EEG motor imagery signal to be recognized by using a trained adaptive classifier recognition model to obtain a scalp EEG motor imagery signal recognition result; The training and acquisition steps of the trained adaptive classifier recognition model include: Based on the acquired scalp EEG motion imagery signal samples, multiple basic general recognition models are trained by loading the data set with the same paradigm as the recognition target; Among multiple basic universal recognition models, the basic universal recognition model with the lowest verification loss of the scalp EEG motion imagery signal samples of the target subject is used as the pre-training model; The feature extractor module and pre-training parameters of the pre-trained model are read and a model migration fine-tuning strategy is adopted to train an untrained classifier using the scalp EEG motor imagery signal samples of the target subject to obtain a trained adaptive classifier recognition model.

[0037] In one embodiment of the present invention, a computer device is provided, the computer device comprising a processor and a memory, the memory being used to store a computer program, the computer program comprising program instructions, and the processor being used to execute the program instructions stored in the computer storage medium. The processor may be a central processing unit (CPU), or may be other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc., which is the computing core and control core of the terminal, which is suitable for implementing one or more instructions, and is specifically suitable for loading and executing one or more instructions in the computer storage medium to implement the corresponding method flow or corresponding function; the processor described in the embodiment of the present invention can be used to perform the operation of the adaptive scalp electroencephalogram motor imagery signal recognition method.

[0038] In one embodiment of the present invention, a storage medium is provided, specifically a computer-readable storage medium (Memory), which is a memory device in a computer device for storing programs and data. It is understandable that the computer-readable storage medium here can include both built-in storage media in the computer device and, of course, extended storage media supported by the computer device. The computer-readable storage medium provides a storage space, which stores the operating system of the terminal. In addition, one or more instructions suitable for being loaded and executed by the processor are also stored in the storage space, and these instructions can be one or more computer programs (including program codes). It should be noted that the computer-readable storage medium here can be a high-speed RAM memory or a non-volatile memory, such as at least one disk memory. The processor can load and execute one or more instructions stored in the computer-readable storage medium to implement the corresponding steps of the adaptive scalp electroencephalogram motion imagery signal recognition method in the above embodiment.

[0039] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.

[0040] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0041] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0042] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process in the computer or other programmable device. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0043] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the relevant field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents. Any modification or equivalent replacement that does not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. An adaptive scalp EEG motor imagery signal recognition method, characterized in that: The following steps are involved: Acquire the scalp EEG motion imagery signal to be identified of the target subject; Based on the scalp EEG motor imagery signal to be identified, signal recognition is performed using a trained adaptive classifier recognition model to obtain a scalp EEG motor imagery signal recognition result; The training and acquisition steps of the trained adaptive classifier recognition model include: Based on the acquired scalp EEG motion imagery signal samples, multiple basic general recognition models are trained by loading the data set with the same paradigm as the recognition target; Among multiple basic universal recognition models, the basic universal recognition model with the lowest verification loss of the scalp EEG motion imagery signal samples of the target subject is used as the pre-training model; The feature extractor module and pre-training parameters of the pre-trained model are read and a model migration fine-tuning strategy is adopted to train an untrained classifier using the scalp EEG motor imagery signal samples of the target subject to obtain a trained adaptive classifier recognition model.

2. The adaptive scalp EEG motor imagery signal recognition method according to claim 1, characterized in that: The step of training and obtaining multiple basic universal recognition models based on the acquired scalp EEG motor imagery signal samples by loading a data set of the same paradigm as the recognition target comprises: Sampling the EEG signals of multiple sample subjects to obtain original EEG signal sample data of each sample subject; The original EEG signal sample data of each sample subject are preprocessed respectively to obtain the training set, validation set and test set of each sample subject; wherein the preprocessing process is to select the EEG channel data related to the motor imagery area based on the original EEG signal sample data; filter and downsample the selected EEG channel data to obtain two-dimensional data of channel × time; convert the two-dimensional data of channel × time into three-dimensional data of two-dimensional topology map × time, and divide them to obtain the training set, validation set and test set; Based on the training set, validation set and test set of each sample subject, the selected recognition model is trained and updated respectively to obtain multiple basic general recognition models; in the training and updating process, the weight and bias of the recognition model are adjusted by continuously evaluating the loss value of the validation data in each Epoch to complete the parameter optimization and updating of the recognition model.

3. The adaptive scalp EEG motor imagery signal recognition method according to claim 2, characterized in that: In the step of training and updating the selected recognition models based on the training set, validation set and test set of each sample subject to obtain multiple basic universal recognition models, During the training update process, if the loss of the validation data does not decrease in the preset number of epochs, the early stopping mechanism is triggered and the recognition model stops training.

4. The method for adaptive scalp EEG motor imagery signal recognition according to claim 1, characterized in that: The step of reading the feature extractor module and pre-training parameters of the pre-trained model and adopting a model migration fine-tuning strategy to train an untrained classifier using the scalp EEG motor imagery signal samples of the target subject to obtain a trained adaptive classifier recognition model specifically includes: Reading the feature extractor module and pre-training parameters of the pre-trained model and adopting a model migration fine-tuning strategy, using the scalp EEG motor imagery signal samples of the target subject to train an untrained classifier to obtain a trained adaptive classifier recognition model; The performance of the trained adaptive classifier recognition model is evaluated using four indicators: accuracy, specificity, sensitivity, and f1-score. Once the preset performance requirements are met, the trained adaptive classifier recognition model is obtained.

5. An adaptive scalp EEG motor imagery signal recognition system, characterized in that: include: A signal acquisition module, used for acquiring the scalp EEG motion imagery signal to be identified of the target subject; A signal recognition module, used to perform signal recognition based on the scalp EEG motor imagery signal to be recognized by using a trained adaptive classifier recognition model to obtain a scalp EEG motor imagery signal recognition result; The training and acquisition steps of the trained adaptive classifier recognition model include: Based on the acquired scalp EEG motion imagery signal samples, multiple basic general recognition models are trained by loading the data set with the same paradigm as the recognition target; Among multiple basic universal recognition models, the basic universal recognition model with the lowest verification loss of the scalp EEG motion imagery signal samples of the target subject is used as the pre-training model; The feature extractor module and pre-training parameters of the pre-trained model are read and a model migration fine-tuning strategy is adopted to train an untrained classifier using the scalp EEG motor imagery signal samples of the target subject to obtain a trained adaptive classifier recognition model.

6. The adaptive scalp EEG motor imagery signal recognition system according to claim 5, characterized in that: The step of training and obtaining multiple basic universal recognition models based on the acquired scalp EEG motor imagery signal samples by loading a data set of the same paradigm as the recognition target comprises: Sampling the EEG signals of multiple sample subjects to obtain original EEG signal sample data of each sample subject; The original EEG signal sample data of each sample subject are preprocessed respectively to obtain the training set, validation set and test set of each sample subject; wherein the preprocessing process is, based on the original EEG signal sample data, selecting EEG channel data related to the motor imagery area; filtering and downsampling the selected EEG channel data to obtain two-dimensional data of channel × time; converting the two-dimensional data of channel × time into three-dimensional data of two-dimensional topology map × time, and dividing to obtain the training set, validation set and test set; Based on the training set, validation set and test set of each sample subject, the selected recognition model is trained and updated respectively to obtain multiple basic general recognition models; in the training and updating process, the weight and bias of the recognition model are adjusted by continuously evaluating the loss value of the validation data in each Epoch to complete the parameter optimization and updating of the recognition model.

7. The adaptive scalp EEG motor imagery signal recognition system according to claim 6, characterized in that: In the step of training and updating the selected recognition models based on the training set, validation set and test set of each sample subject to obtain multiple basic universal recognition models, During the training update process, if the loss of the validation data does not decrease in the preset number of epochs, the early stopping mechanism is triggered and the recognition model stops training.

8. The adaptive scalp EEG motor imagery signal recognition system according to claim 5, characterized in that: The step of reading the feature extractor module and pre-training parameters of the pre-trained model and adopting a model migration fine-tuning strategy to train an untrained classifier using the scalp EEG motor imagery signal samples of the target subject to obtain a trained adaptive classifier recognition model specifically includes: Reading the feature extractor module and pre-training parameters of the pre-trained model and adopting a model migration fine-tuning strategy, using the scalp EEG motor imagery signal samples of the target subject to train an untrained classifier to obtain a trained adaptive classifier recognition model; The performance of the trained adaptive classifier recognition model is evaluated using four indicators: accuracy, specificity, sensitivity, and f1-score. Once the preset performance requirements are met, the trained adaptive classifier recognition model is obtained.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the adaptive scalp EEG motor imagery signal recognition method according to any one of claims 1 to 4 is implemented.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the adaptive scalp EEG motor imagery signal recognition method according to any one of claims 1 to 4 is implemented.

Images