GitHub

MindScryer: High-Quality EEG-to-Image Reconstruction via Dual-Branch Feature Extraction and Cascaded Conditional Diffusion

Abstract: Reconstructing visual stimuli from EEG signals is a challenging problem at the intersection of neuroscience and computer vision. We propose MindScryer, a novel framework that leverages ...
GitHub

SUP-MCRL: Subject-aware Unified Pseudo-feature Coded Multimodal Contrastive Representation Learning for EEG Visual Decoding

A contrastive learning-based framework for EEG-image cross-modal alignment and retrieval. Designed for the THINGS-EEG dataset, this project achieves precise retrieval from EEG signals to visual ...
Frontiers

MsGCN: a multi-stream graph convolutional network for multiband PLV graph fusion in EEG-based biometric identification

Ju et al., 2023 ). Common spatial pattern (CSP) is also one of the most popular methods for EEG feature extraction ( Becerra et al., 2025 ). However, these representations are generally less effective ...
C&EN

Identifying Sequential Residue Patterns in Bitter and Umami Peptides

A peptide’s amino acid sequence affects its taste, but how? A rigorous structure–property connection is challenging to determine because of both the exponentially growing peptide sequence space and ...