 This paper proposes a novel paradigm and model for detecting dim targets in remote sensing images using brain-computer interfaces, BCIs. The proposed paradigm involves searching for targets while recording EEG signals from participants. This allows for the extraction of event-related brain responses associated with target detection. Additionally, a domain adaptation method was used to reduce the impact of subject differences between different datasets. The proposed model, called the domain adaptive and channel-wise attention-based time-domain convolutional neural network, DCT-CNN, was trained on EEG data collected from participants performing the AVEP task. The DCT-CNN outperformed other models in terms of accuracy and generalizability. Furthermore, visualizations of spatiotemporal features showed the effectiveness and interpretability of the proposed paradigm and model. This article was authored by Liangwei Fan, Huixian, Fengyuxia, and others. We are article.tv, links in the description below.