A man lay in a sleep lab at Northwestern University while a researcher asked him, out loud, what eight minus six was. He got it right, signalling the answer with his eyeballs, and he was sound asleep the entire time.
That experiment was published in Current Biology in 2021 by Karen Konkoly and a team spanning four labs in the United States, Germany, France and the Netherlands. Thirty-six volunteers took part. Six of them answered correctly on 29 occasions, using deliberate eye movements and small twitches of facial muscle to signal back. They solved simple arithmetic and told different sensory cues apart. One of them had narcolepsy and lucid dreams most nights, which turned out to be a considerable professional advantage.
So I understand why the follow-up claim spreads so easily. Scientists have supposedly confirmed that the prefrontal cortex fires up during a lucid dream in a pattern nearly identical to being awake, meaning part of you is properly conscious while the rest of you sleeps. It sounds like the obvious next chapter. The research has gone somewhere considerably weirder.
The scan everyone cites came from one sleeping man
Behind almost every version of that headline sits a 2012 paper in the journal Sleep from Martin Dresler and colleagues at the Max Planck Institute of Psychiatry in Munich. They recruited four experienced lucid dreamers and put them through combined EEG and fMRI overnight. Out of those four, exactly one produced two verified lucid episodes long enough to analyse.
One person. Two episodes.
During them, the right dorsolateral prefrontal cortex lit up, along with frontopolar areas, the precuneus and occipito-temporal cortex. It is a careful piece of work and the researchers were honest about its limits, putting the words “case study” in their own title. The comparison, though, ran lucid REM against ordinary REM sleep, with waking never entering the equation. Nobody measured a waking-identical pattern, because nobody was looking for one. That flourish got attached somewhere between the journal and your phone.
The 40 Hz signature had an eye muscle hiding in it
The second pillar is EEG. In 2009, Ursula Voss and Allan Hobson reported elevated gamma power around 40 Hz over frontal and frontolateral scalp sites during lucid dreaming, based on three participants with one verified lucid dream each. Voss described lucidity as a hybrid state, something perched between REM and waking. Even in that original paper, gamma during lucidity sat higher than plain REM sleep and still lower than wakefulness.
Then Benjamin Baird, Giulio Tononi and Stephen LaBerge went back over the numbers. In 2022 they reanalysed 14 signal-verified lucid dreams from the Stanford database. Most of the frontolateral 40 Hz bump, they reported, was attributable to the saccadic spike potential: a tiny electrical blip thrown off by the eye muscles a fraction of a second before a rapid eye movement.
Lucid dreams come with more eye movements. More eye movements mean more spikes, and those spikes shrug off the standard cleaning routines researchers had been trusting for years. The artefact is worst at electrodes F7 and F8, which happen to be precisely where Voss recorded her strongest effect. Their paper’s title puts it flatly: lucid dreaming occurs in activated REM sleep, not a mixture of sleep and wakefulness.
What has actually survived
Something real is still going on, just pointing in the other direction. A 2025 analysis in the Journal of Neuroscience by Demirel and colleagues pooled lucid dreaming EEG recordings across laboratories using a preprocessing pipeline built specifically to strip out those eye artefacts. Compared with wakefulness, they found broad reductions in power from low alpha through gamma, with beta activity dropping in right central and parietal regions. Sensor-level differences between lucid and non-lucid REM were minimal.
Less wake-like, in other words, not more.
The prefrontal story does hang on in a different form. Baird’s group scanned 14 people who lucid dream three or more times a week and found stronger resting connectivity between frontopolar cortex and temporoparietal areas than in matched controls. Elisa Filevich and colleagues, in a separate imaging study, tied lucid dreaming frequency to grey matter in anterior prefrontal cortex and to performance on waking metacognition tasks. Both were measured while participants were awake. They describe a kind of brain, not a kind of dream.
Why the tidy version is the boring one
I have had perhaps eight lucid dreams in my life, all of them squandered on flying badly. The idea that a slice of my frontal lobe had quietly clocked on for the shift is a comforting image, and comfortable images are exactly the ones I have learned to check.
Strip the mythology out and the residue is stranger than the myth.
Sleep does not need to borrow wakefulness to support reflection, arithmetic or a considered answer to a question from a stranger in a lab coat. It can host all of that on its own terms, in its own electrical signature, while the body stays switched off at the neck. The tidy version of the story keeps trying to explain consciousness during sleep by smuggling in a scrap of waking, and the data keeps declining the offer.
