Reproduce “Inter-subject Deep Transfer Learning for Motor Imagery EEG Decoding” 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
Xiaoxi Wei, Pablo Ortega, A. Aldo Faisal
The SCSN uses subject-specific feature extractors around shared layers to train multi-subject motor-imagery EEG decoders. It is designed to limit negative transfer, when pooled subjects’ data harms a shared model. On the BCI Competition IV 2a benchmark, it reaches 81.8% accuracy versus 73.4% for the multi-subject CNN, near the single-subject CNN’s 82%, recovering lost accuracy but not exceeding that baseline.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.