When he was in need of inspiration, the inventor Thomas Edison used to take a nap in a chair while holding a metal ball in each hand. The moment he dropped off, the balls would drop too and crash to the floor, jolting him awake. Edison claimed that this allowed him to capture creative ideas that had fleetingly bubbled up into his semi-consciousness as he fell asleep.
The state Edison was chasing is known as the sleep-onset period (SOP), a little-studied phase of the sleep-wake cycle. Once seen as merely a brief interlude between wakefulness and slumber, it is now being recognised as a distinct and important stage in its own right. Not only is it involved in orchestrating the shutdown of consciousness, but it may also play a vital role in many of the functions of sleep, including memory-processing and, of course, creativity.
Advertisement
For some people, however, it can be disordered; insomnia and narcolepsy could be the result of it going awry. A better understanding of the SOP could lead to new treatments for these sleep conditions, while also helping anyone who wants to be more alert or creative – so “pretty much everybody”, says , a cognitive neuroscientist at Sorbonne University in Paris, France.
We all undergo the transition from wakefulness to sleep, sometimes several times every 24 hours, and many of us know that it can be an otherworldly journey. That is what inspired Oudiette’s interest in it. “I always see stuff when I fall asleep,” she says. “Faces, scary faces.”
When Oudiette became a sleep researcher, she initially started work on the role of rapid eye movement (REM) sleep in creativity. But she and her colleagues found that the earliest phases of sleep, before REM, also seemed important to creativity. “So we completely changed gear, to sleep onset,” she says. She now leads , a five-year research project with the French National Institute of Health and Medical Research that is trying to understand this evanescent state and learn how to manipulate it.
In sleep research, the SOP has always been a bit of a twilight zone, in part because of the standard methodology used in sleep labs. Polysomnography tracks people’s brain activity, eye movements, muscle tone and sometimes other biological signals as they progress through sleep. Researchers divide the output, a polysomnograph, into 30-second “epochs”, classifying each as one of five states: wakefulness; the three stages of non-REM sleep, N1, N2, and N3, with each stage leading to progressively deeper sleep; and REM sleep. This standardised system treats the SOP as the start of the first epoch that the scorer classifies as sleep. “The sleep-onset period doesn’t exist in the sleep score scoring manual,” says , a neurologist at Brussels University Hospital in Belgium.
Between sleep stages
Slicing sleep into all-or-nothing 30-second chunks is fine for analysing a full night of slumber, but it fails to capture shorter events or transitions between states. For example, an epoch featuring wakefulness and N1 will be scored as one or the other depending on which predominates. But for sleep-onset researchers, this is deeply unsatisfying because, as they are discovering, that transitional state between being awake and sleeping is where the action is.
Researchers – and most people – have long been aware that there is more to the transition into sleep than a simple on-off switch, says Oudiette, in part because of a bizarre, semi-lucid state at the onset of sleep called hypnagogia, featuring hallucinatory, dream-like experiences.
Hypnagogia has been , generally by waking people up soon after they drop off and asking them what was going through their minds. This revealed that hypnagogia is both , ranging from fleeting sensory experiences – images, sounds, touches, tastes and falling sensations – to vivid dreams almost like those of REM. This content often includes snatches of events from the preceding day and can be influenced by external inputs such as sounds. Perceptions of time and space are often distorted and many subjects report that their hypnagogic experiences are scary. Experiments suggest that around 95 per cent of people experience hypnagogia, but only around half recall it upon waking from a full-night’s sleep, says Oudiette.
Not just non-REM
But hypnagogia is just one part of the SOP. In 1994, trying to understand the phenomenon, Tadao Hori and colleagues at Hiroshima University in Japan (EEGs) from people as they fell asleep and sliced the traces into 5-second chunks. After analysing them visually and matching them against participants’ reports, they proposed that hypnagogia is part of a much longer SOP lasting around 10 minutes. They also identified nine distinct SOP brain states, starting with quiet wakefulness (which they called Hori stage 1) and extending to the beginning of the first phase of N2 (Hori stage 9). “The SOP is a unique period that cannot be accurately categorised as either wake or sleep,” they wrote.
Unfortunately, the Hori system turned out to be difficult to implement. “It’s super time-consuming and a little bit ambiguous, you might have a lot of difference between scorers,” says Oudiette. “It’s kind of a pain.” The research didn’t go any further.
But in the past few years, researchers have developed new, automated ways of analysing EEGs, some employing artificial intelligence algorithms to recognise subtle patterns occurring on very short temporal and spatial scales. Other imaging techniques, , are also being used to probe the SOP in ever finer detail. “We now have the tools to really go deeper in the analysis of this period,” says Strauss.
These new techniques reveal . Some areas of the brain appear to be awake while others are in various stages of sleep, and they can flicker from one to the other at the drop of a hat; for several minutes, the brain is neither fully awake nor fully asleep, but a pulsating mosaic of the two. “It’s a succession of different brain states that are different from wakefulness and different from sleep,” says Strauss. “It’s really dynamic.”
Gradually, the fluctuations die down and all regions of the brain settle into a typical N2 pattern of brainwaves. This starts in the thalamus – a sort of neural relay station that passes information between different brain systems – and spreads to the frontal cortex.
Brainwide fluctuations may explain the peculiar perceptual experiences of hypnagogia, says Oudiette, with internally generated dream-like states interacting with external stimuli. It may also explain the enormous variation between individuals’ SOP experiences: some people linger for many minutes while others fall off a cliff straight into deep sleep, and this may vary from one night to the next.
A hotbed of creativity
The SOP may also be doing important things. The purpose of sleep in general is unclear, although a lot of evidence suggests that it plays a role in memory-processing and generating creativity. And there are hints that the SOP is involved in both.
Memory consolidation – the processing of newly encoded memories into longer-term formats – is . However, there hasn’t been much research on the role of the early stages of sleep in that process, says Oudiette, although what little there is suggests the SOP may be important. Some of the first research into this area found that people who engage in a new physical or mental task during the day often experience a kind of hallucinatory re-enactment of those tasks as they fall asleep, a phenomenon popularly known as the Tetris effect after people who played the game found they couldn’t stop seeing falling bricks as they dozed off.
In 2000, , a psychiatrist studying sleep and cognition at Harvard Medical School, to investigate the role of hypnagogia in memory, finding that even people who were unable to form new memories still reported seeing Tetris-like blocks as they drifted off. The results implied that in sleep onset, people are recalling the day’s events and extracting what’s meaningful. “It’s possible that it is a stage that initiates the memory consolidation that will occur later,” says Oudiette.
The fact that hypnagogia sometimes replays experiences from the day is suggestive. So is , which found that people who were woken from N1 were worse at recalling the locations of 48 objects on a grid that they had memorised before falling asleep, compared with those who entered N2. Exactly what this demonstrates is unclear. However, Oudiette says it strongly suggests that the SOP may be an important part of memory processing – either by pruning irrelevant information and making room in the synaptic network for memory consolidation, or by tagging, through reactivation, memories that will be reinforced in later sleep stages. In preventing sleepers from descending into N2, researchers may have aborted the tagging process, resulting in those memories being lost rather than reinforced. “I think we still need to have more evidence,” says Strauss, who wasn’t involved in this research. “But it’s very interesting to study.”
The creativity sweet spot occurs when you start to lose control of your thoughts
Hypnagogic memory replay may also be involved in creativity. In the early 2020s, Oudiette was researching the effect of REM sleep on this elusive phenomenon. Volunteers were given a task that required them to try to memorise the position of objects and were told there was a hidden rule that would allow them to solve it instantly. They attempted the task, slept on it, then tried again. “Unfortunately, this experiment failed miserably,” says Oudiette. “The rule was too difficult and almost no participants found it.” But one participant who did said he discovered it as he was drifting off to sleep. This suggests to Oudiette that the early part of sleep plays a role in creative problem-solving.
In 2018, a team led by Michael Craig at Heriot-Watt University in Edinburgh, UK, had shown that in a dimly lit room with minimal stimulation significantly boosted people’s ability to discover a hidden rule in a different task. But the researchers didn’t look at whether the participants had fallen asleep, leaving open the question of whether that state was involved in their success. So Oudiette’s group dropped the REM research and reran Craig’s experiment with this added step of quiet rest and potential sleep.
They got a group of 103 volunteers to solve the same problem as in Craig’s research, called the number reduction task. Participants were asked to transform a given string of eight digits, in this case a series of 1s, 4s and 9s, into a new “solution string” of seven digits using two simple rules, applied digit-by-digit, that were provided by the experimenter. Although it only takes seconds to solve conventionally, there is an even quicker route to solving the puzzle: the last digit of the solution string is always the same as the second digit of the solution string. People who grasp this can finish the task almost instantaneously.
A cure for insomnia?
The volunteers completed 60 strings, then – having eliminated 16 who had spotted the hidden rule – the researchers gave them a 20-minute Edison-style rest, reclining in a chair with their eyes closed and a bottle in hand. About half stayed fully awake and a quarter drifted briefly into N1, the first stage of non-REM sleep, before being roused by the falling bottle. The rest fell into N2.
When they returned to the task, who entered N1 spotted the hidden rule, compared with 31 per cent of the people who stayed awake and just 14 per cent of the deeper sleepers. The result suggests that a brief foray into N1 can inspire a flash of creativity. “It fits with the definition of creativity, which is producing something that is both original and adapted to the context,” says Oudiette. “It’s original because we didn’t tell them there was a rule and it’s adapted because they go faster when they find it.”
Other researchers have since made similar findings using different tasks. Last year, Stickgold’s lab found that people who had had a short period of N1 were . As they fell asleep (or didn’t), the subjects were verbally prompted to think about a tree. Then, after being woken, they were asked to write a creative story about a tree, think of alternative uses for a tree and respond to nouns with the first verb that popped into their heads. The sleepers did better in all three tasks. “There’s clearly a form of creativity that is massively enhanced by the first few moments of sleep,” says Stickgold.
This makes sense: two of the brain networks that are , which generates spontaneous thoughts, and the cognitive control network of higher functions such as attention and planning. According to Oudiette, creativity relies on the , with the former generating random ideas and the latter critically evaluating them.
“I think it’s a sweet spot when you start losing the control of your thoughts, you can let it go and have more spontaneous exploration of ideas,” she says. “But at the same time, you can still be a bit conscious and evaluate them.”
Even if the search for a creative sweet spot in the SOP disintegrates like a half-recalled dream, there are likely to be practical applications. Around a fifth of people have conditions such as narcolepsy or insomnia, where the SOP happens too quickly or takes too long, says Strauss.
A detailed analysis of their individual brain activity could pinpoint where things go awry. It may be possible to employ neurofeedback – using EEG to allow people to watch their brainwaves in real time and learn to control them – to fix the problem. “I can try to teach you how to modify your brain activity to either accelerate or resist sleep, depending on your needs,” says Oudiette. Similar techniques could be used to train people to take restorative power naps or foster creativity, she says. “To arrive to that at the end of my project, that is my dream.”
