Traditional manual feature-based machine learning methods and deep learning networks have been used for Electroencephalography (EEG)-based emotion recognition in recent years. However, some existing studies ignore the low signal-to-noise ratio and the fact that each subject has unique EEG traits, which suffer from low recognition accuracy and poor robustness. To solve these problems, we propose a novel attention mechanism-based multi-scale feature fusion network (AM-MSFFN) that considers high-level features at different scales to improve the generalization of the model for different subjects. Specifically, we first utilize a spatial-temporal convolutional block to extract temporal and spatial features of EEG signals sequentially. Subsequently, considering the sampling rate of EEG signals, the multi-scale separable convolutions are designed for capturing emotional state-related information, to better combine and output feature mapping relationships. Convolutional module attention mechanism (CBAM) is applied after point-wise convolution, to better handle EEG variations of different subjects and the key information which facilitates classification. In addition, we adopt a preprocessing module based on data augmentation and data alignment to improve the quality of the training samples. Moreover, ablation studies show that the proposed attention mechanism and multiscale separable convolution contribute significant and consistent gain to the performance of our AM-MSFFN model. To verify the effectiveness of the proposed algorithm, we conducted extensive experiments on the DEAP dataset and SEED. The average accuracies achieve 99.479% and 99.297% for arousal and valence, respectively. The results demonstrated the feasibility of the proposed method.

Graph neural networks (GNNs) have demonstrated efficient processing of graph-structured data, making them a promising method for electroencephalogram (EEG) emotion recognition. However, due to dynamic functional connectivity and nonlinear relationships between brain regions, representing EEG as graph data remains a great challenge. To solve this problem, we proposed a multi-domain based graph representation learning (MD $^{2}$ GRL) framework to model EEG signals as graph data. Specifically, MD $^{2}$ GRL leverages gated recurrent units (GRU) and power spectral density (PSD) to construct node features of two subgraphs. Subsequently, the self-attention mechanism is adopted to learn the similarity matrix between nodes and fuse it with the intrinsic spatial matrix of EEG to compute the corresponding adjacency matrix. In addition, we introduced a learnable soft thresholding operator to sparsify the adjacency matrix to reduce noise in the graph structure. In the downstream task, we designed a dual-branch GNN and incorporated spatial asymmetry for graph coarsening. We conducted experiments using the publicly available datasets SEED and DEAP, separately for subject-dependent and subject-independent, to evaluate the performance of our model in emotion classification. Experimental results demonstrated that our method achieved state-of-the-art (SOTA) classification performance in both subject-dependent and subject-independent experiments. Furthermore, the visualization analysis of the learned graph structure reveals EEG channel connections that are significantly related to emotion and suppress irrelevant noise. These findings are consistent with established neuroscience research and demonstrate the potential of our approach in comprehending the neural underpinnings of emotion.

Motor imagery (MI) electroencephalography (EEG) is natural and comfortable for controllers, and has become a research hotspot in the field of the brain–computer interface (BCI). Exploring the inter-subject MI-BCI performance variation is one of the fundamental problems in MI-BCI application. EEG microstates with high spatiotemporal resolution and multichannel information can represent brain cognitive function. In this paper, four EEG microstates (MS1, MS2, MS3, MS4) were used in the analysis of the differences in the subjects’ MI-BCI performance, and the four microstate feature parameters (the mean duration, the occurrences per second, the time coverage ratio, and the transition probability) were calculated. The correlation between the resting-state EEG microstate feature parameters and the subjects’ MI-BCI performance was measured. Based on the negative correlation of the occurrence of MS1 and the positive correlation of the mean duration of MS3, a resting-state microstate predictor was proposed. Twenty-eight subjects were recruited to participate in our MI experiments to assess the performance of our resting-state microstate predictor. The experimental results show that the average area under curve (AUC) value of our resting-state microstate predictor was 0.83, and increased by 17.9% compared with the spectral entropy predictor, representing that the microstate feature parameters can better fit the subjects’ MI-BCI performance than spectral entropy predictor. Moreover, the AUC of microstate predictor is higher than that of spectral entropy predictor at both the single-session level and average level. Overall, our resting-state microstate predictor can help MI-BCI researchers better select subjects, save time, and promote MI-BCI development.

The attentional processes are conceptualized as a system of anatomical brain areas involving three specialized networks of alerting, orienting and executive control, each of which has been proven to have a relation with specified time-frequency oscillations through electrophysiological techniques. Nevertheless, at present, it is still unclear how the idea of these three independent attention networks is reflected in the specific short-time topology propagation of the brain, assembled with complexity and precision. In this study, we investigated the temporal patterns of dynamic information flow in each attention network via electroencephalograph (EEG)-based analysis. A modified version of the attention network test (ANT) with an EEG recording was adopted to probe the dynamic topology propagation in the three attention networks. First, the event-related potentials (ERP) analysis was used to extract sub-stage networks corresponding to the role of each attention network. Then, the dynamic network model of each attention network was constructed by post hoc test between conditions followed by the short-time-windows fitting model and brain network construction. We found that the alerting involved long-range interaction among the prefrontal cortex and posterior cortex of brain. The orienting elicited more sparse information flow after the target onset in the frequency band 1–30 Hz, and the executive control contained complex top-down control originating from the frontal cortex of the brain. Moreover, the switch of the activated regions in the associated time courses was elicited in attention networks contributing to diverse processing stages, which further extends our knowledge of the mechanism of attention networks.

Objective. Rapid serial visual presentation (RSVP) based on electroencephalography (EEG) has been widely used in the target detection field, which distinguishes target and non-target by detecting event-related potential (ERP) components. However, the classification performance of the RSVP task is limited by the variability of ERP components, which is a great challenge in developing RSVP for real-life applications. Approach. To tackle this issue, a classification framework based on the ERP feature enhancement to offset the negative impact of the variability of ERP components for RSVP task classification named latency detection and EEG reconstruction was proposed in this paper. First, a spatial-temporal similarity measurement approach was proposed for latency detection. Subsequently, we constructed a single-trial EEG signal model containing ERP latency information. Then, according to the latency information detected in the first step, the model can be solved to obtain the corrected ERP signal and realize the enhancement of ERP features. Finally, the EEG signal after ERP enhancement can be processed by most of the existing feature extraction and classification methods of the RSVP task in this framework. Main results. Nine subjects were recruited to participate in the RSVP experiment on vehicle detection. Four popular algorithms (spatially weighted Fisher linear discrimination-principal component analysis (PCA), hierarchical discriminant PCA, hierarchical discriminant component analysis, and spatial-temporal hybrid common spatial pattern-PCA) in RSVP-based brain–computer interface for feature extraction were selected to verify the performance of our proposed framework. Experimental results showed that our proposed framework significantly outperforms the conventional classification framework in terms of area under curve, balanced accuracy, true positive rate, and false positive rate in four feature extraction methods. Additionally, statistical results showed that our proposed framework enables better performance with fewer training samples, channel numbers, and shorter temporal window sizes. Significance. As a result, the classification performance of the RSVP task was significantly improved by using our proposed framework. Our proposed classification framework will significantly promote the practical application of the RSVP task.

Background: Motor imagery-based brain-computer interfaces (MI-BCIs) enable individuals who are unable to perform physical movements to interact with the external world by imagining movements. Users are typically classified as good performers or BCI-illiterate based on the classification accuracy of distinct EEG patterns (e.g., 60% or 70%). Yet, studies show that approximately 70% of users fall within intermediate accuracies between 60% and 80%, and although exceed the chance level, they often fail to achieve reliable MI-BCI control. Intermediate users often exhibit asymmetric motor imagery abilities between left and right hands, highlighting the need for refined early assessment and stratified training approaches. Methods: We employed ICA to decompose each participant’s EEG data and extract independent ERD/ERS components as indicators using a rule-based automated framework. This framework integrated dipole localization, ERD/ERS characteristics, and frequency-band power features of ICs. Importantly, we applied a power spectral parameterization approach to remove the 1/f-like background activity in power estimation and used statistical methods to precisely estimate the latency and duration of ERD. The extracted indicators were subsequently subjected to clustering analysis to categorize participants into four groups. Results: In addition to good performers (24.8%) and poor performers (35.8%), two groups were identified: LgoodRpoor (27.5%), who performed well in left-hand MI but poorly in right-hand MI, and LpoorRgood (11.9%), who showed the opposite pattern. Notably, these unilateral performers did not show significant differences in contralateral ERD but exhibited substantial differences in ipsilateral ERS. Conclusions: The proposed independent event-related brain dynamics model enables more refined stratification of MI-BCI users. Findings from this characterization study may inform the design of graded training protocols, especially for users demonstrating unilateral motor imagery proficiency.