Sleep, a time traditionally associated with rest, actually involves significant brain activity, according to new research. An international team has launched the DREAM database, a comprehensive collection of brain recordings correlated with dream reports, marking a significant advancement in understanding the mysteries of sleep. Published in Nature Communications, the study reveals that dreams can occur in various sleep stages, including deeper, non-REM (NREM) sleep, contradicting decades of belief that confined dreaming primarily to the rapid-eye-movement (REM) stage.
For years, REM sleep, characterized by vivid dreams and rapid eye movements, was thought to be the exclusive realm of dreaming. However, the findings from the DREAM project indicate a broader scope. Participants reported dreaming after approximately 85 percent of REM sleep awakenings and around 40 to 60 percent of NREM sleep awakenings. Even during the deepest sleep stages, dreams were reported, suggesting that the dreaming brain retains a level of awareness akin to wakefulness.
Researchers noted that as sleep deepened, the frequency of dreams decreased but did not vanish completely. Some individuals reported experiencing thoughts or emotions even in the slow-wave stages of sleep. Analysis of brain activity patterns revealed that dreams occurring in NREM sleep exhibited oscillations resembling those seen in quiet wakefulness, indicating that the brain may operate in a “partly awake” state even while the body is at rest.
The DREAM database consolidates data from 2,643 awakenings of 505 volunteers across 20 studies in 13 countries, creating an expansive resource for dream research. Each awakening was categorized as a dream experience, an experience without recall (often referred to as a “white dream”), or no experience at all. By employing electroencephalography (EEG) and magnetoencephalography (MEG), which detects magnetic fields generated by brain activity, the research team was able to analyze brain data just before each awakening and correlate it with subsequent dream reports.
Artificial intelligence played a crucial role in this research, as models were trained to determine if a participant had been dreaming based on the final 30 seconds of brain wave data. The study found that simpler brain-wave characteristics during deep NREM sleep provided moderate accuracy, while complex patterns in REM sleep yielded better results. These findings suggest that dreaming leaves behind a unique neural signature, potentially enabling scientists to detect dreams without waking individuals in the future.
Co-author Giulio Bernardi highlighted the collaborative nature of the project, coordinated by Monash University in Australia, which involved 53 authors from 37 institutions across 13 countries. This research represents a critical move forward in exploring human consciousness by compiling and standardizing decades of dream research into a single, accessible database.
Understanding dreams can unveil insights into consciousness during sleep, as they reflect moments when the brain generates experiences without external stimuli. The DREAM project aims to provide a foundational framework for researchers to investigate the brain”s activities during sleep, uncovering the reasons behind dreaming and variations in dream recall among individuals.
