Reproduce “A 1000-hour EEG-EMG-audio dataset of Japanese speech production” 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
Motoshige Sato, Ilya Horiguchi, Masakazu Inoue, Kenichi Tomeoka, Eri Hatakeyama, Yuya Kita, Atsushi Yamamoto, Ippei Fujisawa and 1 more
Researchers at Araya Inc. constructed JapanEEG, a 1020-hour multimodal dataset of scalp EEG, facial EMG, and speech audio from three Japanese speakers recorded across three distinct EEG systems. This resource facilitates research into non-invasive speech decoding, artifact modeling, and model generalizability across diverse recording conditions.
Key insights
- The multimodal, longitudinal, and cross-device design of JapanEEG is critical for developing robust non-invasive speech decoding methods and improving model generalization.
- Technical validation via Power Spectral Density (PSD) and Event-Related Potential (ERP) analyses confirmed the dataset's high signal quality and physiological relevance.
- Consistent ERP patterns observed across different EEG systems and tasks (listening, overt, covert speech) validate the dataset's utility for studying neural mechanisms of speech production and perception.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.