Most adults, when asked how they’re sleeping, will give you a number in hours. Seven and a half. Six and a half. They will not give you a number in minutes of N3 — and yet that number is the one that determines whether they wake up actually rested.
The data on this is uncomfortable. By age 40, the average healthy adult has lost roughly 40% of the slow-wave sleep they had in their twenties [1]. By 60, the loss approaches 70% — even when total time-in-bed and self-reported sleep quality stay constant. The architecture quietly disassembles itself, one decade at a time, and most people only notice the consequences: the slower memory consolidation, the dimmer mornings, the recovery that takes a day longer than it used to.
What “deep sleep” actually is
Polysomnography divides the night into roughly 90-minute cycles, each containing four progressively deeper non-REM stages and one REM stage. Stage three — what sleep researchers call N3, or slow-wave sleep — is the one that does the bulk of physical recovery: glymphatic clearance, growth hormone release, immune modulation, memory consolidation [2].
“You don’t lose sleep with age. You lose the kind of sleep that does the work.” — Dr. Matthew Walker, UC Berkeley
The mechanism isn’t fully understood, but the candidates are well-characterized: declining slow-wave-promoting neurons in the preoptic area, age-related shifts in adenosine receptor density, and progressive flattening of the circadian temperature curve.
The three levers with the largest measured effect
If you read the meta-analyses rather than the supplement marketing, three interventions repeatedly rise above the noise:
1. Core temperature manipulation in the 90 minutes before bed
A 2019 systematic review of 17 RCTs found that warming the body — paradoxically — increases N3 by an average of 11 minutes per night, with effect sizes that hold up across age cohorts [3]. The mechanism is the post-bath cooling curve: getting warm forces a steeper drop in core body temperature once you exit, and that drop is what triggers sleep-onset slow waves.
2. Strict, dim, consistent evening light
Light intensity above 10 lux in the two hours before bed flattens the melatonin onset curve by an average of 71% [4]. This is the single largest behavioral lever in the literature, and it costs nothing. Dim lamps, no overhead lights, and a phone in night mode is not the same as no phone — but it is meaningfully better than nothing.
3. The morning, not the night
What you do in the first 30 minutes after waking — specifically, whether you get bright outdoor light into your eyes — sets the timing of melatonin release that night. The effect size is genuinely large, and most adults skip it entirely.
Where supplements fit (and don’t)
The evidence base for sleep supplements is thinner than the marketing implies. Magnesium glycinate has the cleanest data for sleep latency, with effect sizes in the 5–10 minute range; melatonin is excellent for circadian shifting but mediocre for sleep maintenance; valerian and L-theanine have mixed-to-positive but mostly underpowered trials.
If you’re going to take one thing, the magnesium data is the most defensible. If you’re going to do one thing, fix the light.
[1] Ohayon MM et al. Sleep, 2004. [2] Léger D et al. Sleep Med Reviews, 2018. [3] Haghayegh S et al. Sleep Med Reviews, 2019. [4] Gooley JJ et al. J Clin Endocrinol Metab, 2011.