Deep-Sleep Sounds Boost Brain Waves, Not Clearly Memory

Not clearly. Sound timed to slow waves does raise slow-wave activity. Whether that improves memory is unsettled. A meta-analysis of 10 studies and 177 participants found a combined effect of g = 0.25 (p = 0.07). Its authors concluded the evidence is not sufficient to recommend commercial devices. A second meta-analysis of 12 studies found effect sizes fell every year, from favoring stimulation in 2013 to slightly favoring no stimulation in 2021.

Does sound during deep sleep raise slow waves?

Yes, in the studies here. Slow oscillations are the large, slow brain waves of deep sleep. Short tones played during those waves can amplify them.

Timing matters. In 133 adolescents on overnight polysomnography, tones played in the upstate of the EEG slow oscillation produced slow oscillations 110% larger in amplitude than tones in the downstate. Slow-wave activity rose 3%.

A home trial of 34 adults also found that closed-loop stimulation raised slow oscillation amplitude. That is a brain-wave result. Whether the extra amplitude helps memory is a separate question.

Does the extra slow-wave activity improve memory?

Possibly a little. A 2021 meta-analysis pooled 10 studies with 11 experiments and 177 healthy participants. The combined effect on overnight episodic memory was g = 0.25 (p = 0.07). That is a small effect that does not reach significance.

Subgroups looked better. Young adults gave g = 0.31 (p = 0.051). Phase-locked stimulation gave 0.36 (p = 0.047). Both together gave 0.44 (p = 0.01). These are subgroup cuts of a small pool.

The authors concluded that the evidence is not sufficient to recommend the use of commercially available devices. See how much deep sleep is needed and how wearables track sleep stages for the context around such claims.

Why do the results shrink over time?

A 2023 meta-analysis of 12 studies, 14 effect sizes and 206 participants found a steady yearly decline in effect size. Publication year alone accounted for 91.8% of the differences between studies. No coded study feature explained more.

The predicted effect on word-pair retention in 2013 was d = 0.99 (95% CI 0.49 to 1.49), favoring stimulation. The predicted effect in 2021 was d = -0.39 (95% CI -0.73 to -0.05), moderately favoring no stimulation.

The authors also estimated that 34% of participants were not truly blind to stimulation because they heard it during sleep. The test-retest reliability of memory retention scores was near zero (0.01). Both weaken small studies. They call for larger samples, true placebo controls and open data.

What about home use and older adults?

Home use has one small negative result here. In 34 adults, one night of closed-loop stimulation at home raised slow oscillation amplitude. It had no effect on declarative memory or vigilance. The authors caution that their sample may have been too small to detect a real effect.

Older adults may need more nights. In 16 older adults with a mean age of 68.9, memory gains appeared only after three nights of real stimulation. The design was within-subject with 9 and 7 people per group. That is a very small sample, so the result is a lead and not a conclusion.

A 2022 meta-analysis reported a memory effect limited to adults under 35. It has since been retracted, and its findings are not used here.

What does this not settle?

The pooled memory effect is small and not significant overall. Effect sizes trend downward across published studies. Participants were mostly healthy young adults tested in labs.

Nothing here tests people with sleep disorders, insomnia or cognitive impairment. Nothing here tests long-term use. Tone timing has been studied more than memory outcomes. A device that raises slow waves has not been shown to improve memory by the same amount.

Common questions