FSDBN: Foreground-Aware EEG--Visual Alignment via Dynamic Brain Networks

2026-07-20

AI Summary

Existing EEG-based visual decoding approaches often overlook the perceptual asymmetry between foreground and background in complex scenes, rendering them susceptible to background interference, semantic misalignment, and inadequate modeling of the rapid temporal dynamics and non-stationary spatial characteristics of EEG signals. To address these limitations, this work proposes the FSDBN unified framework, which disentangles foreground and background through semantic-consistent saliency alignment, adaptively enhances foreground contributions via a semantic-prior-guided dynamic gating mechanism, and models EEG as a dynamic spatiotemporal brain network to capture neural responses underlying visual attention. This approach achieves the first joint integration of saliency and semantic constraints for foreground-aware EEG–visual alignment, attaining 69.0% top-1 and 92.2% top-5 accuracy in zero-shot brain-to-image retrieval—significantly outperforming current state-of-the-art methods.

Abstract

EEG-based visual decoding provides a non-invasive pathway for interpreting visual semantics. However, existing methods often overlook the perceptual asymmetry between foreground and background in complex scenes, leading to background interference and semantic misalignment. EEG signals also exhibit rapid temporal dynamics and nonstationary spatial patterns, making it difficult to capture the time-varying brain connectivity associated with focal visual attention. To address these limitations, we propose FSDBN, a unified framework for robust EEG-visual decoding. FSDBN introduces Semantic-Consistent Saliency Alignment to separate semantically relevant foreground regions from background noise under joint saliency and semantic constraints. It further employs Semantic-Prior Dynamic Gating Foreground Fusion to adaptively regulate the contributions of foreground and background features. In parallel, EEG signals are modeled as adaptive spatiotemporal brain networks whose functional connectivity dynamically reorganizes to capture neural responses to salient foregrounds. Experiments on zero-shot brain-to-image retrieval demonstrate that FSDBN achieves 69.0 percent top-1 accuracy and 92.2 percent top-5 accuracy, outperforming previous state-of-the-art methods. Code is available at https://github.com/LiuYiheng1/FSDBN-EEG.

Problem

• EEG-based visual decoding
• perceptual asymmetry
• foreground-background interference
• temporal dynamics
• brain connectivity

Innovation

• EEG-visual decoding
• dynamic brain networks
• saliency alignment
• foreground-aware fusion
• zero-shot retrieval