Imperial College London: Falling asleep follows a brain "tipping point", Imperial-led study shows

Imperial College London and UK Dementia Research Institute researchers have identified a predictable tipping point in the brain as we fall asleep, validating a new way to track the transition to sleep and showing it can be predicted in near real time. The team analysed overnight EEG from more than a thousand people, with findings published in Nature Neuroscience.

Mapping the brain's approach to the tipping point in real time could translate into earlier drowsiness warnings for safer driving, applications in new diagnostics, and better management of sleep onset disorders.

Using a device called a scalp EEG, the researchers represented the moments before sleep as a trajectory in a normalised feature space of brain activity and tracked a single quantity, the 'sleep distance'. In group data, this distance remains relatively stable before dropping abruptly in the final minutes, marking a bifurcation-style tipping point about four and a half minutes before conventional sleep onset. Minutes before that point, the brain shows 'critical slowing', a rise in variance and autocorrelation that often precedes regime shifts in complex systems.

"We discovered that falling asleep is a bifurcation, not a gradual process, with a clear tipping point that can be predicted in real time… The ability to track how individual brains fall asleep has profound implications for our understanding of the sleep process and for developing new treatments for people who struggle with falling asleep," said study lead Dr Nir Grossman (UK DRI at Imperial ).

Read more