A rigorous new study has examined the large-scale brain activity of a number of human subjects while sleeping, presenting one of the most detailed investigations into sleep phases conducted to date. The study suggests that instead of the traditional four sleep stages we generally understand the brain moves through, there are in fact at least 19 different identifiable brain patterns transitioned through while sleeping.

Traditionally scientists have identified four distinct stages our brain transitions through in a general sleep cycle – three non-REM sleep phases (N1-3) that culminate in an REM phase. The four stages have been classically determined and delineated using electroencephalographic (EEG) brainwave recordings.

"This way of dividing sleep into stages is really based on historical conventions, many of which date back to the 1930s," explains Angus Stevner, one of the researchers on the project from the Center for Music in the Brain at Aarhus University. "We've come up with a more precise and detailed description of sleep as a higher number of brain networks which change their communication patterns and dynamic characteristics during sleep."

The new research set out to more comprehensively record whole-brain activity in a number of subjects by using functional magnetic resonance imaging (fMRI). The study began by studying 57 healthy subjects in an fMRI scanner. Each subject was asked to lie in the scanner for 52 minutes with their eyes closed. At the same time, each subject was tracked using an EEG. This allowed the researchers to compare traditional brainwave sleep cycle data with that from the fMRI.

Due to the limited duration of the fMRI data, no subjects were found to enter REM sleep, however, 18 subjects did completely transition from wakefulness through the three non-REM sleep phases according to the EEG data. Highlighting the complexity of brain activity during our wake-to-sleep cycle the researchers confidently chronicled 19 different recurring whole-brain network states.

Mapping these whole-brain states onto traditional EEG-tracked sleep phases revealed a number of compelling correlations. Wakefulness, N2 sleep and N3 sleep all could be represented by specific whole brain states. The range of different fMRI-tracked brain states did reduce as subjects fell into deeper sleep phases, with two different fMRI brain states correlating with N2 sleep, and only one with N3. However, N1 sleep as identified by EEG data, the earliest and least clearly defined sleep phase, did not consistently correspond with any fMRI brain state.

The researchers conclude from this data that N1 is actually a much more complex sleep phase than previously understood. This phase, a strange mix of wakefulness and sleep, seemed to encompass a large range of the 19 different whole-brain network states identified in the fMRI data.

"Our results provide a modern description of human sleep as a function of the brain's complex network activities and we're trying to move on from the somewhat simplified picture that has thus far characterized our understanding of brain activity during sleep," says Stevner.

The researchers suggest this work is only at the very beginning of shedding light on exactly how a human brain transitions through different sleep phases. Further work is set to explore brain activity in subjects suffering from sleep disorders, and delve deeper into brain activity patterns behind different states of consciousness, such as those under the influence of psychedelics or anesthesia.

"Our results can change the way in which we understand sleep and, not least, the way we look at sleep disorders such as insomnia," says Stevner. "This provides a new and potentially revolutionary understanding of brain activity during sleep which can in turn lead to new forms of treatment of the sleep problems that affect far too many people.

The research was published in the journal Nature Communications.