A third state of consciousness: what the largest EEG study shows

For decades the consensus was that the brain exists in two discrete modes: waking and sleep, with sleep further divided into slow-wave and REM phases. In 2024 Çağatay Demirel's group at the Radboud University Medical Center analysed the largest sleep EEG database in the world and demonstrated that lucid dreaming is not a subtype of REM — it is a separate, third operating mode of the cortex.

The key biomarker is selective reactivation of the fronto-parietal network while muscular atonia is preserved. In ordinary REM the dorsolateral prefrontal cortex is suppressed, which is why absurd scenes feel normal. At the moment of lucidity this region jumps back online, restoring metacognition — the ability to think about your own thoughts. The body, meanwhile, stays paralysed.

The electrophysiology is even more interesting. In a lucid dream beta-wave power drops in the temporo-parietal cortex — the brain physically disconnects from sensory processing of its own body. Simultaneously a synchronised gamma rhythm around 40 Hz fires in the prefrontal cortex — the frequency band associated in waking subjects with insight and the 'eureka' moment.

The Wisconsin Institute for Sleep and Consciousness uncovered long-term neuroanatomical correlates: people who regularly practise lucid dreaming show increased grey-matter volume in the prefrontal cortex and stronger functional connectivity between the frontopolar and temporo-parietal regions. Practice literally rewires the brain.

LaBerge's Stanford experiments in the 80s added another finding. When subjects performed arithmetic inside a lucid dream, activity shifted left; when they sang, it shifted right — the same pattern seen in waking subjects. A dream with metacognition is, by neural signature, closer to a waking task than to an ordinary dream.

For a practitioner this means something concrete: a lucid dream is not 'weaker' or 'less real' than a daytime thought. It is a full cognitive mode with measurable characteristics. And like any mode, it can be trained.

What's next. Within the next five years consumer-grade portable EEG headsets that detect REM onset and deliver a soft auditory cue for lucid induction are expected to reach the market. Prototypes exist; the bottleneck is mass production and regulation. Once they spread, the speed of skill acquisition is likely to grow by an order of magnitude.