“…At present, the graph connection representation in EEG data based on graph convolution only uses a static connection mode, which is not consistent with the fact that the brain network connection changes from time to time. Especially, for video-induced P3, the neural response can be divided into three stages, i.e., information integration, decision process, and neuronal response ( Desmedt, 1980 ; Song et al, 2021b ). Therefore, this paper proposes a segmentation adaptive spatial-temporal GCN (SAST-GCN) for single-trial video-induced P3 detection based on the spatial-temporal structure of multi-channel EEG signals.…”