Lucid dreaming can be learned by some people, but new studies suggest that realising you are dreaming and steering the dream are separate skills, and that success is far from guaranteed
Learning to recognise a dream while it is happening, known as lucid dreaming, is possible for some people with practice. New research suggests it is harder to master, and to steer, than popular advice implies. A review of 19 studies found the best evidence supports a technique called mnemonic induction, but success rates ranged from under 20% to over 40%. A September 2026 paper in the Journal of Sleep Research argues that knowing you are dreaming and controlling the dream are separate abilities.
Which techniques have evidence behind them
The systematic review, also in the Journal of Sleep Research, assessed 14 techniques. The strongest evidence was for the Mnemonic Induction of Lucid Dreams (MILD), a method based on setting an intention to notice you are dreaming. Six of the eight studies of it reported positive results. The reviewers could not fully explain why success rates varied so much between studies.
Other methods fared less well:
- Senses-Initiated Lucid Dreams (SSILD): one study of one, which the review called promising but not replicated.
- Reality testing: one positive result in six studies. The review found insufficient evidence that repeatedly checking whether you are dreaming works by itself.
- Dream diaries: one positive result in two studies, inadequate as a stand-alone method.
- Sound and light stimulation: inconclusive.
- Galantamine: a medication, with preliminary evidence and medication-related risks.
- Transcranial alternating-current stimulation: not shown to work.
A separate study of 355 volunteers found MILD and SSILD were similarly effective, and that combining them did no better than using either alone. Remembering more dreams and falling back asleep within 10 minutes of the exercises predicted success.
Technology and the lab
Researchers at Northwestern University developed Targeted Lucidity Reactivation (TLR). Participants learn to link certain sounds with the realisation that they are dreaming, and the sounds are then played during sleep. Results published in Consciousness and Cognition in 2024 showed more lucid dream reports, and the sounds worked best when they matched those used in training. That does not show an ordinary smartphone app can reliably detect REM sleep or produce lucid dreams on demand.
A July 2026 experiment by scientists at the University of Bern and Radboud University gave 28 participants two weeks of training and then stimulated their balance system during REM sleep. That was linked to higher lucidity scores, while electrical muscle stimulation was not. The researchers could not separate the effect of the stimulation from the training and participants’ expectations, so it is not evidence that commercial devices work.
Awareness is not control
The September paper, by Clarita Bonamino and Emma Peters, proposes two kinds of control. Deliberate control means consciously trying to change the dream, such as choosing where to go. Expectation-driven control means the dream changes because of what the sleeper anticipates. The authors say reliable strategies for controlling the setting, events, dream characters and the dreamer’s own actions have not been established in standardised experiments. Individual differences, emotions and expectations may all play a part.
Studies highlighted by the American Psychological Association (APA) point the same way. Dream characters often behaved unpredictably when lucid dreamers asked them to draw or compose poetry. Adolescents reported feelings of empowerment, but there is not enough evidence that practising skills in dreams improves real-world performance.
What happens in the brain
A May 2026 study using EEG recordings found distinct patterns of brain activity in lucid REM sleep compared with ordinary REM dreams, with more involvement from systems linked to visual experience, self-awareness and executive processing. It does not identify a single brain region responsible.
Researchers have also shown limited two-way communication with dreamers. In research published in Current Biology, four laboratories tested 36 participants, and six gave 29 correct answers using eye movements and facial-muscle signals. A February 2026 study of 13 frequent lucid dreamers found they could signal with pre-arranged sniffs. Communication was uncommon and needed controlled laboratory conditions.
Risks to weigh
Not every lucid dream is pleasant. Research highlighted by the APA in February 2026 classed about 20% of lucid dreams as lucid nightmares, and people who already had frequent nightmares were more likely to have them. A 2025 review looked at possible links with dissociation, psychosis, anxiety and sleep paralysis. It did not conclude that lucid dreaming generally causes any of them.
Cleveland Clinic psychologist Alicia Roth recommends protecting normal sleep and avoiding practices that leave people unrested. Wake-back-to-bed methods deliberately interrupt sleep. Anyone troubled by distressing dreams or poor sleep should speak to a doctor.
