Sleep

Dream database aims to answer burning questions

New paper describes an organisational effort to reduce the variability of methods used in dream research — an approach which could supercharge investigations.

12 September 2025

By Emma Young



Why do some people wake up every morning with detailed memories of their dreams, while others can't remember dreaming at all? Why do so few people have 'lucid dreams', which they can control? And, more fundamentally, by looking at brain activity alone, how can we spot when someone is dreaming?

Despite decades of research into dreams, we still don't have answers to these big questions. That's partly down to the inconsistent methods and, because of the high costs involved, the often small numbers of participants included in experiments, write a team led by William Wong at Monash University, Australia, in Nature Communications. In a bid to overcome these two stumbling blocks, the team has now established the DREAM database — a new and expanding collection of standardised datasets of both measures of brain activity during sleep and accompanying dream reports.

The initial database, described in their paper, consists of 20 datasets on 505 participants, and includes measures of electrical activity in these participants' brains just prior to 2,463 awakenings, after which they were questioned about whether they'd been dreaming.

In earlier dream studies, these subjective reports have been categorised in all kinds of different ways, making comparisons between studies difficult. In this database, they are allocated to one of three main categories. One is 'experience' — having a specific memory of having any kind of conscious experience, which could mean a complex dream, or a simpler sensory impression, which the team also classify as being a dream. Then there's 'experience without recall' — the participant was sure they had just been dreaming but couldn't remember any detail. The final category is 'no experience' — no awareness of having any conscious experience just before they were woken.

The team also established minimum technical standards for the EEG aspect of studies that can be included in the database — with at least 20 seconds of continuous sleep recorded up until the moment of an awakening, for example.

In a bid to demonstrate that the database could be useful for research, the team analysed their combined EEG and dream report data. They concluded that, as expected from earlier research, dreams happened in both REM (rapid eye movement) and non-REM sleep. They also found that, during both REM and non-REM sleep, there were clear differences in the EEG pattern for someone who reported an 'experience' rather than 'no experience'. "These results demonstrate that the DREAM database can be exploited to effectively investigate the neural correlates of dreaming," the authors write.

In theory, this new database may help researchers to get a better understanding of the neural underpinnings of dreams, and provide a more solid footing by which to answer outstanding questions about both dreaming and consciousness. For example, if we can learn how to identify instances of conscious experience (ie. a dream of some sort) in a sleeping person based on brain activity alone, this may allow us to spot consciousness in other situations in which a person is unable to communicate, the team writes. At the moment, it's not possible to be certain that in these situations, the individual is not conscious at all. "By contrasting dreamless, unconscious sleep with dreamful, conscious sleep, the distinction between levels of consciousness within the same brain state becomes possible," the team writes.

It will, however, take more contributions for the database to grow into a truly robust resource. If you're a researcher interested in learning more, the database (which is supported by volunteers) can be found at https://monash.edu/dream-database, and is open to contributions from labs worldwide.

Read the paper in full:
Wong, W., Herzog, R., Andrade, K. C., Andrillon, T., Barros, D., Arnulf, I., Somayeh Ataei, Giulia Avvenuti, Baird, B., Bellesi, M., Bergamo, D., Bernardi, G., Blagrove, M., Decat, N., Çağatay Demirel, Dresler, M., Eichenlaub, J.-B., Elce, V., Steffen Gais, & Gennaro, L. D. (2025). A dream EEG and mentation database. Nature Communications, 16(1). https://doi.org/10.1038/s41467-025-61945-1

SOURCE:

https://www.bps.org.uk/research-digest/dream-database-aims-answer-burning-questions(accessed 19.09.25)