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

Demirel and colleagues (2025), full original paper pooled recordings from several laboratories: the largest lucid-dream sample for this analysis at publication, not the world's largest database of all sleep polysomnograms.

The sample

After exclusions there were 26 participants and 44 sleep recordings; one spectral outlier was removed from subsequent analyses, leaving 43. Several people contributed multiple recordings, so 43 recordings do not mean 43 independent people.

The findings

Sensor-level differences from nonlucid REM were small. Source-level analyses showed reduced beta power in right central and parietal regions, increased alpha connectivity and increased gamma connectivity.

Around initial lucidity signaling, gamma1 power at 30–36 Hz increased in right temporo-occipital sources including the precuneus. This is not a universal frontal forty-hertz flash or proof of detachment from the body.

Interpretation limits

The authors describe correlates of lucid dreaming, predominantly associated with REM. ‘A third state’ is a popular framing, not a new sleep stage that excludes REM. Mixed protocols, the modest sample and artifact cleaning limit generalisation.

Correlations do not prove that practice rewires the brain or that all methods work through one rhythm. The paper promises neither consumer headsets in five years, tenfold faster learning nor clinical benefit.

In practice distinguish awareness during a dream, control and vividness. EEG findings supply no home stimulation recipe; MILD has a separate induction evidence base.