A dream EEG and mentation database (the largest dream database yet) (www.nature.com)

Researchers released the DREAM (Dream EEG and Mentation) database — the largest standardized collection of sleep neurophysiology paired with dream reports to date. The initial release aggregates 20 datasets from 505 participants and 2,643 awakenings, each datum containing at minimum ≥20 s of M/EEG recorded at ≥100 Hz with ≥2 electrodes up to the time of awakening, plus a harmonized, structured dream-report classification and essential PSG metadata (EEG, EOG, EMG). The project provides a unified taxonomy and consistent formatting to make multicenter reuse and large-scale analyses possible, and it is openly accessible and extensible (https://monash.edu/dream-database). Technically, early analyses using the pooled data show that reports of conscious experience can be predicted from objective EEG features in both REM and NREM sleep, supporting previously reported markers such as reduced low-frequency (delta) power during reported dreams. The database also enables more detailed mapping between spectral bands, sleep microstructure (spindles, coherence, network measures) and specific dream contents (e.g., frontal alpha asymmetry for affect; gamma activity for perceptual content). By overcoming single-study sample-size and methodological heterogeneity limits, DREAM should accelerate robust discovery of neural correlates of dreaming, improve reproducibility, and inform clinical problems (e.g., assessing covert awareness in unresponsive patients or intraoperative dreaming) through scalable, multimodal analyses.