Adaptive attention graph convolution network with normalized embedded Gaussian for rapid serial visualization presentation decoding

• Published in: Engineering applications of artificial intelligence Vol. 161; p. 112093
• Main Authors: Zhao, Mengyuan, Ai, Qingsong, Chen, Kun, Liu, Quan, Xie, Sheng Quan, Ma, Li
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 01.12.2025
• Subjects: Adaptive adjacency matrix; Adaptive attention graph convolution network; Brain computer interface; Convolutional block attention module; Rapid serial visualization presentation; Convolutional block attention module; Adaptive adjacency matrix; Rapid serial visualization presentation; Brain computer interface; Adaptive attention graph convolution network
• ISSN: 0952-1976
• DOI: 10.1016/j.engappai.2025.112093

Graph convolutional networks (GCNs) have been widely used in Brain Computer Interface (BCI) and have shown great prowess in identifying electroencephalogram (EEG) spatiotemporal features. However, most GCNs learn channels topological relationship by fixed adjacency matrix. This lacks connectivity strength information and ignores the data dependency. This paper proposes a data-driven adjacency matrix based on normalized embedded Gaussian function, and constructs a Gaussian-Adaptive Attention Graph Convolution Network (Gaussian-AAGCN). Brain regions connectivity is calculated by normalized embedded Gaussian function, and the topological relationship is adaptively learned by input data in a data-driven manner. This data-driven adaptive adjacency matrix avoids brain activity information loss caused by fixed adjacency matrix and improves the flexibility of graph construction. Convolutional block attention module (CBAM) is introduced to adaptive feature refinement in two independent dimensions, improving model representation ability. Experimental results show that the average area under curve (AUC), true positive rate (TPR) and false positive rate (FPR) of Gaussian-AAGCN on 14 subjects are 93.52 %, 91.59 %, and 4.58 % respectively. Compared to Transformer, Event-Related Potential Capsule Network (ERP-CapsNet), Electroencephalogram Convolutional Neural Network (EEGNet), and Multi-Granularity Information Fusion Network (MGIFNet), the AUC of Gaussian-AAGCN is higher by 18.02 %, 5.32 %, 2.62 %, and 1.82 %, respectively. Using the adaptive adjacency matrix, the model AUC and TPR are increased by about 4.8 % and 7.8 % respectively. After integrating CBAM, the AUC and TPR increased by about 3.5 % and 8 % respectively. [Display omitted] •The Gaussian-AAGCN is proposed, multi-subgraph convolution is designed to extract more complementary features.•An adaptive adjacency matrix is proposed to learn topological relationships of brain regions, avoiding information loss.•CBAM is integrated to adaptive feature refinement, enhancing important spatiotemporal features weights.•Compared with other advanced models, the proposed model achieves superior performance.