Reproduce “Recasting Visual Language Models as EEG Foundation Models for Clinical Neurodiagnostics” 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
Xin Zhou, Elena Sergeeva, Nan Bi, Wenyu Song, Lijun Yin, Hao Deng
The authors adapt a vision–language model to classify EEG by rendering task-specific signal segments as waveform images, optionally with spectrograms, and pairing them with prompts that constrain the output to named labels. On four-second vEpiSet discharge detection, Qwen3-VL-4B with LoRA reached 0.835 balanced accuracy, versus 0.763 for BIOT; these are supervised test results, not clinical validation.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.