Emotion plays crucial roles in human life. Recently, emotion classification from electroencephalogram (EEG) signal has attracted attention by researchers due to the rapid development of brain computer interface (BCI) techniques and machine learning algorithms. However, recent studies on emotion classification show resource utilization because they use the fully-supervised learning methods. Therefore, in this study, we applied the self-supervised learning methods to improve the efficiency of resources usage. We employed a self-supervised approach to train deep multi-task convolutional neural network (CNN) for EEG-based emotion classification. First, six signal transformations were performed on unlabeled EEG data to construct the pretext task. Second, a multi-task CNN was used to perform signal transformation recognition on the transformed signals together with the original signals. After the signal transformation recognition network was trained, the convolutional layer network was frozen and the fully connected layer was reconstructed as emotion recognition network. Finally, the EEG data with affective labels were used to train the emotion recognition network to clarify the emotion. In this paper, we conduct extensive experiments from the data scaling perspective using the SEED, DEAP affective dataset. Results showed that the self-supervised learning methods can learn the internal representation of data and save computation time compared to the fully-supervised learning methods. In conclusion, our study suggests that the self-supervised machine learning model can improve the performance of emotion classification compared to the conventional fully supervised model.

Emotion recognition is important in the application of brain-computer interface (BCI). Building a robust emotion recognition model across subjects and sessions is critical in emotion based BCI systems. Electroencephalogram (EEG) is a widely used tool to recognize different emotion states. However, EEG has disadvantages such as small amplitude, low signal-to-noise ratio, and non-stationary properties, resulting in large differences across subjects. To solve these problems, this paper proposes a new emotion recognition method based on a multi-source associate domain adaptation network, considering both domain invariant and domain-specific features. First, separate branches were constructed for multiple source domains, assuming that different EEG data shared the same low-level features. Secondly, the domain specific features were extracted by using the one-to-one associate domain adaptation. Then, the weighted scores of specific sources were obtained according to the distribution distance, and multiple source classifiers were deduced with the corresponding weighted scores. Finally, EEG emotion recognition experiments were conducted on different datasets, including SEED, DEAP, and SEED-IV dataset. Results indicated that, in the cross-subject experiment, the average accuracy in SEED dataset was 86.16%, DEAP dataset was 65.59%, and SEED-IV was 59.29%. In the cross-session experiment, the accuracies of SEED and SEED-IV datasets were 91.10% and 66.68%, respectively. Our proposed method has achieved better classification results compared to state-of-the-art domain adaptation methods.

The rapid development of the automotive industry has brought great convenience to our life, which also leads to a dramatic increase in the amount of traffic accidents. A large proportion of traffic accidents were caused by driving fatigue. EEG is considered as a direct, effective, and promising modality to detect driving fatigue. In this study, we presented a novel feature extraction strategy based on a deep learning model to achieve high classification accuracy and efficiency in using EEG for driving fatigue detection. EEG signals were recorded from six healthy volunteers in a simulated driving experiment. The feature extraction strategy was developed by integrating the principal component analysis (PCA) and a deep learning model called PCA network (PCANet). In particular, the principal component analysis (PCA) was used to preprocess EEG data to reduce its dimension in order to overcome the limitation of dimension explosion caused by PCANet, making this approach feasible for EEG-based driving fatigue detection. Results demonstrated high and robust performance of the proposed modified PCANet method with classification accuracy up to 95%, which outperformed the conventional feature extraction strategies in the field. We also identified that the parietal and occipital lobes of the brain were strongly associated with driving fatigue. This is the first study, to the best of our knowledge, to practically apply the modified PCANet technique for EEG-based driving fatigue detection.

Support vector machines are powerful tools used to solve the small sample and nonlinear classification problems, but their ultimate classification performance depends heavily upon the selection of appropriate kernel and penalty parameters. In this study, we propose using a particle swarm optimization algorithm to optimize the selection of both the kernel and penalty parameters in order to improve the classification performance of support vector machines. The performance of the optimized classifier was evaluated with motor imagery EEG signals in terms of both classification and prediction. Results show that the optimized classifier can significantly improve the classification accuracy of motor imagery EEG signals.

The current data scarcity problem in EEG-based emotion recognition tasks leads to difficulty in building high-precision models using existing deep learning methods. To tackle this problem, a dual encoder variational autoencoder-generative adversarial network (DEVAE-GAN) incorporating spatiotemporal features is proposed to generate high-quality artificial samples. First, EEG data for different emotions are preprocessed as differential entropy features under five frequency bands and divided into segments with a 5s time window. Secondly, each feature segment is processed in two forms: the temporal morphology data and the spatial morphology data distributed according to the electrode position. Finally, the proposed dual encoder is trained to extract information from these two features, concatenate the two pieces of information as latent variables, and feed them into the decoder to generate artificial samples. To evaluate the effectiveness, a systematic experimental study was conducted in this work on the SEED dataset. First, the original training dataset is augmented with different numbers of generated samples; then, the augmented training datasets are used to train the deep neural network to construct the sentiment model. The results show that the augmented datasets generated by the proposed method have an average accuracy of 97.21% on all subjects, which is a 5% improvement compared to the original dataset, and the similarity between the generated data and the original data distribution is proved. These results demonstrate that our proposed model can effectively learn the distribution of raw data to generate high-quality artificial samples, which can effectively train a high-precision affective model.

For solving the problem of the inevitable decline in the accuracy of cross-subject emotion recognition via Electroencephalograph (EEG) signal transfer learning due to the negative transfer of data in the source domain, this paper offers a new method to dynamically select the data suitable for transfer learning and eliminate the data that may lead to negative transfer. The method which is called cross-subject source domain selection (CSDS) consists of the next three parts. 1) First, a Frank-copula model is established according to Copula function theory to study the correlation between the source domain and the target domain, which is described by the Kendall correlation coefficient. 2) The calculation method for the Maximum Mean Discrepancy is improved to determine the distance between classes in a single source. After normalization, the Kendall correlation coefficient is superimposed, and the threshold is set to identify the source-domain data most suitable for transfer learning. 3) In the process of transfer learning, on the basis of Manifold Embedded Distribution Alignment, the Local Tangent Space Alignment method is used to provide a low-dimensional linear estimation of the local geometry of nonlinear manifolds, which maintains the local characteristics of the sample data after dimensionality reduction. Experimental results show that compared with the traditional methods, the CSDS increases the accuracy of emotion classification by approximately 2.8% and reduces the runtime by approximately 65%.