In the new study, the researchers demonstrated that the human brain falls asleep abruptly, rather than gradually, with a ‘tipping point’ marking the transition from wakefulness into sleep. They were then able to predict the momentary progression into sleep with unprecedented precision.
The findings could be used to develop new ways to diagnose and treat sleep disorders, such as insomnia, and as a marker of brain health in the context of ageing and neurodegenerative disease, and even to improve how we monitor anaesthesia during surgical procedures.
Led by the UK Dementia Research Institute (UK DRI) at Imperial College London and the UK DRI Centre for Care Research & Technology at the University of Surrey, and published in the journal Nature Neuroscience, the study analysed electroencephalogram (EEG) recordings taken from more than a thousand people who wore electrodes to record their brain activity during the night.
The researchers used a novel computational method that represented the changes in the brain as movement in a multi-dimensional space (called a feature space), and mapped out how each individual moved from bedtime toward sleep.
They found that no matter how long participants were in bed, the transition from wakefulness to sleep occurred abruptly in the final few minutes at a clear tipping point. Such a dynamic is known as a ‘bifurcation’, another example of which is the gradual bending of a stick until it eventually snaps. It also resembles the movement of a falling object, thereby supporting the subjective sensation of ‘falling asleep’.
In a different experiment, involving recording over multiple nights, they discovered that each participant had a unique location (in the multi-dimensional space) at which sleep began, which was consistent across multiple nights of sleep.
Armed with this discovery and the bifurcation model, the researchers showed that they could predict with 98% accuracy the exact progression (second-by-second) by which each participant fell asleep.
Previous definitions of sleep onset have relied on subjective, discrete annotations of brain activity from EEG readings, or indirect measures such as heart rate, breathing rate, and movement. This new approach is the first time scientists have been able to objectively infer the precise momentary progression and exact point at which the brain falls asleep.