Reproduce “OneEEG: A Unified Signal-Visual Multimodal Language Model with Task-Aware Mixture of Projectors for Multi-Task EEG Understanding” on OpenResearch

Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.

Universal app for Apple silicon and Intel. Requires macOS 11 or later.

About this paper

Jiacheng Zhang, Xueyi Zhang, Xin Lu, Siqi Cai, Haizhou Li

OneEEG combines numerical EEG signals and waveform images, routing signal features through task-aware projectors for EEG question-answering. For standard-channel recordings it uses both inputs, while non-standard-channel recordings use images alone. On the benchmark’s closed-set questions, it reaches 65.67% Overall and 68.24% Macro accuracy, versus GPT-5.5’s 50.39% and 53.01%. The comparison tests the complete system, not MoP alone or clinical interpretation.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.