Can your own brainwaves help you sleep?

Follow the process

1. 1Record sleep EEG
2. 2Extract slow-wave patterns
3. 3Translate patterns into sound
4. 4Evaluate the sleep response

Explore the evidence

2025 pilot: personalized sound compared with non-personalized placebo sound. Mean differences reported in the study abstract.

The science behind the sound.

Alain Destexhe and Luc Foubert developed the approach at the Paris-Saclay Institute of Neuroscience (NeuroPSI), within CNRS. Parameters such as each slow wave’s timing, amplitude, rise and decay shape synthesized sound envelopes. This is a translation of recorded patterns into audio, not a claim that an audible tone is itself a brainwave.

Early volunteer findings.

The CNRS project page, updated in March 2023, describes 36 healthy volunteers: 77% rated relaxation above 8/10 and 77% rated the sleep-induction effect above 7/10. These are subjective ratings, not measured reductions in time to fall asleep.

Early insomnia feasibility work.

The same CNRS page describes five people with chronic insomnia at the Clinical Center for Sleep and Vigilance, Hôtel-Dieu, Paris. Two reported help falling asleep. That is 40% of five participants reporting a response, not a 40% reduction in sleep-onset time and not a definitive insomnia-treatment trial.

The 2025 sleep pilot.

Aloulou, Chauvineau, Destexhe and Léger studied 13 adults with subjective insomnia complaints in a randomized, single-blind, at-home crossover study. Each participant listened for 30 minutes at bedtime under two conditions, each lasting 3–5 nights: personalized brainwave-derived sound and non-personalized placebo sound. Sleep was assessed with a dry-EEG headband and a sleep diary. The paper appeared in Sleep Medicine (2025, volume 136, article 106755).

What the results do not establish.

This was a small, short pilot, not evidence of lasting benefit or a replacement for insomnia care. REM latency is the time to first REM sleep, not the time to fall asleep. The abstract does not quantify a 40% reduction in sleep-onset time or establish that every participant rated satisfaction at least 8/10. It does not report an adverse-event analysis. The preprint discloses an author’s patent on the sound-conversion method.