Class discrepancy-guided sub-band filter-based common spatial pattern for motor imagery classification

Overlapping filter bank convolutional neural network for multisubject multicategory motor imagery brain-computer interface

BACKGROUND

Motor imagery brain-computer interfaces (BCIs) is a classic and potential BCI technology achieving brain computer integration. In motor imagery BCI, the operational frequency band of the EEG greatly affects the performance of motor imagery EEG recognition model. However, as most algorithms used a broad frequency band, the discrimination from multiple sub-bands were not fully utilized. Thus, using convolutional neural network (CNNs) to extract discriminative features from EEG signals of different frequency components is a promising method in multisubject EEG recognition.

METHODS

This paper presents a novel overlapping filter bank CNN to incorporate discriminative information from multiple frequency components in multisubject motor imagery recognition. Specifically, two overlapping filter banks with fixed low-cut frequency or sliding low-cut frequency are employed to obtain multiple frequency component representations of EEG signals. Then, multiple CNN models are trained separately. Finally, the output probabilities of multiple CNN models are integrated to determine the predicted EEG label.

RESULTS

Experiments were conducted based on four popular CNN backbone models and three public datasets. And the results showed that the overlapping filter bank CNN was efficient and universal in improving multisubject motor imagery BCI performance. Specifically, compared with the original backbone model, the proposed method can improve the average accuracy by 3.69 percentage points, F1 score by 0.04, and AUC by 0.03. In addition, the proposed method performed best among the comparison with the state-of-the-art methods.

CONCLUSION

The proposed overlapping filter bank CNN framework with fixed low-cut frequency is an efficient and universal method to improve the performance of multisubject motor imagery BCI.

Analysis of instantaneous brain interactions contribution to a motor imagery classification task

The purpose of this study is to analyze the contribution of the interactions between electrodes, measured either as correlation or as Jaccard distance, to the classification of two actions in a motor imagery paradigm, namely, left-hand movement and right-hand movement. The analysis is performed in two classifier models, namely, a static (linear discriminant analysis, LDA) model and a dynamic (hidden conditional random field, HCRF) model. The impact of using the sliding window technique (SWT) in the static and dynamic models is also analyzed. The study proved that their combination with temporal features provides significant information to improve the classification in a two-class motor imagery task for LDA (average accuracy: 0.7192 no additional features, 0.7617 by adding correlation, 0.7606 by adding Jaccard distance; p < 0.001) and HCRF (average accuracy: 0.7370 no additional features, 0.7764 by adding correlation, 0.7793 by adding Jaccard distance; p < 0.001). Also, we showed that adding interactions between electrodes improves significantly the performance of each classifier, regarding the nature of the interaction measure or the classifier itself.

A multiscale siamese convolutional neural network with cross-channel fusion for motor imagery decoding

BACKGROUND

Recently, convolutional neural networks (CNN) are widely applied in motor imagery electroencephalography (MI-EEG) signal classification tasks. However, a simple CNN framework is challenging to satisfy the complex MI-EEG signal decoding.

NEW METHOD

In this study, we propose a multiscale Siamese convolutional neural network with cross-channel fusion (MSCCF-Net) for MI-EEG classification tasks. The proposed network consists of three parts: Siamese cross-channel fusion streams, similarity module and classification module. Each Siamese cross-channel fusion stream contains multiple branches, and each branch is supplemented by cross-channel fusion modules to improve multiscale temporal feature representation capability. The similarity module is adopted to measure the feature similarity between multiple branches. At the same time, the classification module provides a strong constraint to classify the features from all Siamese cross-channel fusion streams. The combination of the similarity module and classification module constitutes a new joint training strategy to further optimize the network performance.

RESULTS

The experiment is conducted on the public BCI Competition IV 2a and 2b datasets, and the results show that the proposed network achieves an average accuracy of 87.36% and 87.33%, respectively.

COMPARISON WITH EXISTING METHODS AND CONCLUSIONS

The proposed network adopts cross-channel fusion to learn multiscale temporal characteristics and joint training strategy to optimize the training process. Therefore, the performance outperforms other state-of-the-art MI-EEG signal classification methods.

Bi-dimensional representation of EEGs for BCI classification using CNN architectures

An important challenge when designing Brain Computer Interfaces (BCI) is to create a pipeline (signal conditioning, feature extraction and classification) requiring minimal parameter adjustments for each subject and each run. On the other hand, Convolutional Neural Networks (CNN) have shown outstanding to automatically extract features from images, which may help when distribution of input data is unknown and irregular. To obtain full benefits of a CNN, we propose two meaningful image representations built from multichannel EEG signals. Images are built from spectrograms and scalograms. We evaluated two kinds of classifiers: one based on a CNN-2D and the other built using a CNN-2D combined with a LSTM. Our experiments showed that this pipeline allows to use the same channels and architectures for all subjects, getting competitive accuracy using different datasets: 71.3 ± 11.9% for BCI IV-2a (four classes); 80.7 ± 11.8 % for BCI IV-2a (two classes); 73.8 ± 12.1% for BCI IV-2b; 83.6 ± 1.0% for BCI II-III and 82.10% ± 6.9% for a private database based on mental calculation.

Improving the performance of multisubject motor imagery-based BCIs using twin cascaded softmax CNNs

OBJECTIVE

Motor imagery (MI) EEG signals vary greatly among subjects, so scholarly research on motor imagery-based brain-computer interfaces (BCIs) has mainly focused on single-subject systems or subject-dependent systems. However, the single-subject model is applicable only to the target subject, and the small sample number greatly limits the performance of the model. This paper aims to study a convolutional neural network to achieve an adaptable MI-BCI that is applicable to multiple subjects.

APPROACH

In this paper, a twin cascaded softmax convolutional neural network (TCSCNN) is proposed for multisubject MI-BCIs. The proposed TCSCNN is independent and can be applied to any single-subject MI classification CNN model. First, to reduce the influence of individual differences, subject recognition and MI recognition are accomplished simultaneously. A cascaded softmax structure consisting of two softmax layers, related to subject recognition and MI recognition, is subsequently applied. Second, to improve the MI classification precision, a twin network structure is proposed on the basis of ensemble learning. TCSCNN is built by combining a cascaded softmax structure and twin network structure.

MAIN RESULTS

Experiments were conducted on three popular CNN models (EEGNet and Shallow ConvNet and Deep ConvNet from EEGDecoding) and three public datasets (BCI Competition IV datasets 2a and 2b and the High-Gamma dataset) to verify the performance of the proposed TCSCNN. The results show that compared with the state-of-the-art CNN model, the proposed TCSCNN obviously improves the precision and convergence of multisubject MI recognition.

SIGNIFICANCE

This study provides a promising scheme for multisubject MI-BCI, reflecting the progress made in the development and application of MI-BCIs.

MI-EEGNET: A novel convolutional neural network for motor imagery classification

BACKGROUND

Brain-computer interfaces (BCI) permits humans to interact with machines by decoding brainwaves to command for a variety of purposes. Convolutional neural networks (ConvNet) have improved the state-of-the-art of motor imagery decoding in an end-to-end approach. However, shallow ConvNets usually perform better than their deep counterparts. Thus, we aim to design a novel ConvNet that is deeper than the existing models, with an increase in terms of performances, and with optimal complexity.

NEW METHOD

We develop a ConvNet based on Inception and Xception architectures that uses convolutional layers to extract temporal and spatial features. We adopt separable convolutions and depthwise convolutions to enable faster and efficient ConvNet. Then, we introduce a new block that is inspired by Inception to learn more rich features to improve the classification performances.

RESULTS

The obtained results are comparable with other state-of-the-art techniques. Also, the weights of the convolutional layers give us some insights onto the learned features and reveal the most relevant ones.

COMPARISON WITH EXISTING METHOD(S)

We show that our model significantly outperforms Filter Bank Common Spatial Pattern (FBCSP), Riemannian Geometry (RG) approaches, and ShallowConvNet (p < 0.05).

CONCLUSIONS

The obtained results prove that motor imagery decoding is possible without handcrafted features.

Motor imagery EEG classification based on ensemble support vector learning

BACKGROUND AND OBJECTIVE

Brain-computer interfaces build a communication pathway from the human brain to a computer. Motor imagery-based electroencephalogram (EEG) classification is a widely applied paradigm in brain-computer interfaces. The common spatial pattern, based on the event-related desynchronization (ERD)/event-related synchronization (ERS) phenomenon, is one of the most popular algorithms for motor imagery-based EEG classification. Moreover, the spatiotemporal discrepancy feature based on the event-related potential phenomenon has been demonstrated to provide complementary information to ERD/ERS-based features. In this paper, aiming to improve the performance of motor imagery-based EEG classification in a few-channel situation, an ensemble support vector learning (ESVL)-based approach is proposed to combine the advantages of the ERD/ERS-based features and the event-related potential-based features in motor imagery-based EEG classification.

METHODS

ESVL is an ensemble learning algorithm based on support vector machine classifier. Specifically, the decision boundary with the largest interclass margin is obtained using the support vector machine algorithm, and the distances between sample points and the decision boundary are mapped to posterior probabilities. The probabilities obtained from different support vector machine classifiers are combined to make prediction. Thus, ESVL leverages the advantages of multiple trained support vector machine classifiers and makes a better prediction based on the posterior probabilities. The class discrepancy-guided sub-band-based common spatial pattern and the spatiotemporal discrepancy feature are applied to extract discriminative features, and then, the extracted features are used to train the ESVL classifier and make predictions.

RESULTS

The BCI Competition IV datasets 2a and 2b are employed to evaluate the performance of the proposed ESVL algorithm. Experimental comparisons with the state-of-the-art methods are performed, and the proposed ESVL-based approach achieves an average max kappa value of 0.60 and 0.71 on BCI Competition IV datasets 2a and 2b respectively. The results show that the proposed ESVL-based approach improves the performance of motor imagery-based brain-computer interfaces.

CONCLUSION

The proposed ESVL classifier could use the posterior probabilities to realize ensemble learning and the ESVL-based motor imagery classification approach takes advantage of the merits of ERD/ERS based feature and event-related potential based feature to improve the experimental performance.