Sleep, Longevity and Cycling
On sleep. What it's actually doing while you're not awake.
The purpose of REM and deep sleep, how much the research says you need, the longevity data, and the layer almost nobody covers: how the menstrual cycle rewrites sleep architecture every month.
Sleep has never been easy in my family. My father spent most of his adult life lying in the dark, waiting. His brain stayed on long after the rest of him wanted to stop — and we now understand why. His melatonin release ran late. The gap between the end of the day and the start of biological night stretched for him in a way it doesn't for most people. My brother has the same thing. I do too.
From my mother I inherited something different. She sleeps lightly — present in her sleep in a way that kept her close to the surface even before she had children. Having children sharpened it into something closer to a biological alarm system: the sensory hypervigilance of a mother that repurposes the nervous system to register small sounds, weight shifts, the quality of silence. It seems to leave a permanent mark. It passed to me in some partial form. The person next to me can turn over in the night without touching me, without making a sound, and I wake up. Something in my nervous system reads the change — a fractional shift in the mattress, the faint difference in air displacement — and brings me back before I have any conscious awareness that anything happened.
I carry both, then: the difficulty falling asleep from my father, the hypervigilant light sleep from my mother. For a long time I had no language for either. I tried most things. Some helped one piece. Nothing addressed the whole picture until I understood that sleep is individual — shaped by genetics, by hormones, by cycle phase, by neurological sensitivity you were born with and what life has layered on top of it.
This post is what I learned. What sleep is actually doing when you're unconscious. What the research says about duration and longevity. And the layer that almost nothing covers: how the menstrual cycle changes sleep architecture across the month.
Sleep is not rest. It's a different programme entirely.
The brain during sleep is not a quiet brain. It is a brain running processes that cannot run during wakefulness — a maintenance and reconstruction cycle that requires the offline state to function. We spent most of human history not knowing this. The modern science of sleep, and the specific discoveries of the last two decades, have dismantled the idea of sleep as passive recovery and replaced it with something considerably more complex.
The glymphatic system is the most striking example. In 2013, Maiken Nedergaard and her colleagues at the University of Rochester published findings showing that the brain's glial cells shrink by up to 60% during sleep, opening channels through which cerebrospinal fluid flows and clears metabolic waste — including amyloid beta and tau proteins, the molecules that accumulate in Alzheimer's disease.¹ This clearance mechanism is the brain's waste disposal system. It runs during sleep. During wakefulness, the channels close and the accumulation continues.
Memory consolidation runs in parallel. During NREM sleep, the hippocampus replays the day's experiences and transfers them to the cortex for long-term storage. During REM sleep, the brain integrates new information with existing memory structures and strips emotional charge from difficult experiences — allowing the memory to persist without the distress that accompanied it. Matthew Walker's research documented this process in detail: sleep after learning improves retention; sleep after emotional experience reduces its acute psychological weight.²
The immune system runs its most significant maintenance during sleep. Natural killer cell activity peaks during deep sleep; cytokine production for pathogen defence concentrates in the sleep window; T-cell homing to lymph nodes is regulated by sleep-dependent circadian signals.³ One night of four hours reduces natural killer cell activity by around 70% compared to a full night. This is a short-term immune impairment with documented consequences for infection susceptibility and, over time, cancer surveillance.
The stages — and what each one is for
Sleep cycles through four stages roughly every 90 minutes. The first half of the night is dominated by deep sleep (slow-wave sleep, SWS). The second half is dominated by REM. This distribution is not arbitrary — it reflects the different functional priorities of each type of sleep.
NREM Stage 1 is the transition from wakefulness: muscle tone drops, theta waves replace alpha waves, the body begins to disengage. It lasts minutes. NREM Stage 2 is where the brain produces sleep spindles — brief bursts of coordinated neural activity that tag memories for consolidation — and K-complexes, which are thought to suppress cortical arousal and maintain sleep continuity. Stage 2 makes up roughly 50% of total sleep time.
Slow-wave sleep is the deepest NREM stage. During SWS, the pituitary releases the largest pulse of growth hormone of the day. The glymphatic clearance is running. DNA repair mechanisms are most active. Cellular recovery from the metabolic demands of waking happens here. The brain generates slow, high-amplitude delta waves. It is difficult to wake someone from SWS, and when you do, they are groggy and disoriented in a way that differs qualitatively from REM awakening.
REM sleep is neurologically unlike any other state. The brain shows activity patterns that resemble wakefulness. Voluntary muscles are paralysed. The limbic system — the emotional brain — is highly active. This is where most vivid dreaming occurs, where emotional memory processing runs, and where the brain makes connections across disparate memory stores, which is one mechanism underlying creativity and insight. REM deprivation specifically degrades emotional regulation, social cognition, and the ability to accurately read other people's emotional states.
How much — and what happens when it falls short
The research consensus for adults is 7 to 9 hours. Below 6 hours, sustained over weeks, the health associations are clear and consistent across multiple conditions: cardiovascular disease, insulin resistance, immune impairment, cognitive decline, and all-cause mortality all show elevated risk.⁵
The critical point is the asymmetry. Cutting sleep from 8 to 7 hours produces modest, often undetectable cognitive and health effects. Cutting it from 7 to 6 produces measurable impairment. Below 6, the evidence suggests the impairment becomes pronounced across multiple systems simultaneously. And because chronic sleep deprivation blunts the subjective experience of sleepiness — the brain stops accurately reporting how impaired it is — people running on 5 to 6 hours often believe themselves to be functioning adequately. The performance data says otherwise.
Weekend "catch-up" sleep helps some metrics but does not fully restore what accumulated deprivation has cost. Sleep debt does not clear the way financial debt clears — with a single large payment. The immune suppression, the glymphatic backlog, the emotional processing that didn't run all leave a trace that a longer Saturday morning does not fully address.
Sleep and longevity — what the data shows
The mortality data on sleep duration follows a U-curve. The risk is highest at the short end — below 6 hours — where it is steep and well-evidenced. The long end of the curve (above 9 to 10 hours) also shows elevated mortality risk, but this association is partly explained by reverse causation: people who sleep long often do so because of underlying illness, depression, or sedentary lifestyle that predates the sleep duration.
The short end requires no such explanatory caveat. Short sleep directly impairs the biological processes — glymphatic clearance, immune function, cardiovascular repair, metabolic regulation — that maintain long-term health. The associations with Alzheimer's disease are particularly well-documented: chronic sleep restriction accelerates amyloid accumulation, and people with sleep disorders show higher rates of dementia onset. The glymphatic system that should be clearing these proteins every night is failing to do so at full capacity when sleep is shortened or fragmented.
For women, a specific longevity finding deserves attention: sleep quality tends to be reported as worse in women than in men across most life stages, and women are twice as likely to have insomnia. Yet sleep research has historically been conducted in male-predominant samples. The sex-specific mechanisms — particularly those involving reproductive hormones — are documented in the literature and largely absent from clinical recommendations.
The layer almost nobody covers: sleep across the cycle
Reproductive hormones modulate sleep architecture. This has been documented in the research for decades.⁴ It has not been translated into clinical sleep recommendations, and it has not, in most cases, been communicated to the cycling women whose sleep changes predictably every month.
Across the follicular phase, from menstruation through to ovulation, sleep is generally at its most efficient. Estrogen has mild pro-sleep effects. Body temperature at night is lower (lower core temperature is associated with faster sleep onset and better sleep maintenance). Many women find their best sleep of the month happens in the week before ovulation.
Progesterone complicates this picture in two distinct ways across the luteal phase. Its metabolite allopregnanolone acts on GABA-A receptors with sedating, anxiolytic effects — which produces initial sedation in the early luteal phase. But as progesterone rises through mid-luteal phase, it also raises basal body temperature, and elevated core body temperature at night interferes with both sleep onset and sleep maintenance. The body needs to drop core temperature to initiate and sustain sleep. Progesterone works against this.
The premenstrual window produces the most disrupted sleep of the cycle for most women. Progesterone and estrogen both fall sharply. The withdrawal of progesterone removes the allopregnanolone-GABA-A sedating effect — producing a rebound insomnia pattern that is physiologically equivalent to benzodiazepine withdrawal. REM sleep specifically is reduced in the late luteal phase. Nighttime waking increases. The sleep that does occur is less restorative.
For women who also experience PMDD, insomnia in the premenstrual window is a diagnostic criterion. For women with ADHD, the late luteal sleep disruption compounds the dopamine depletion that makes that week cognitively the hardest. These systems are not separate. They run on the same hormonal infrastructure, and they respond to the same cycle events simultaneously.
What the research missed by excluding women
The majority of foundational sleep research was conducted in male participants, or in mixed samples where sex was not analysed as a variable. The sex-specific mechanisms — estrogen's effects on sleep spindle density, progesterone's thermoregulatory and GABAergic effects, the cycle-phase variation in REM proportion and sleep efficiency — are documented in the literature but absent from most clinical sleep protocols.
Women present to sleep medicine clinics at higher rates than men. They are more likely to report insomnia. They are more likely to be prescribed sedative-hypnotics without investigation of the hormonal context. A woman presenting with premenstrual insomnia, night sweats interfering with sleep onset, or perimenopausal sleep disruption is offering the clinician a hormonal differential diagnosis that requires specific investigation. Most clinicians have not been trained to take the cycle history that would make that diagnosis possible.
The consequence is women receiving general insomnia protocols — sleep hygiene recommendations, cognitive behavioural therapy for insomnia, or sedatives — for a sleep problem that has a specific hormonal mechanism and potentially a specific hormonal intervention. These are not equivalent. CBTI for premenstrual insomnia is not the same clinical response as addressing progesterone withdrawal or stabilising the luteal estrogen floor.
What to do about it
- Track your sleep quality against your cycle for two full cycles. Note sleep latency (how long to fall asleep), waking episodes, and subjective restedness against your cycle day. Two cycles will almost certainly reveal a pattern. The data you collect is clinical information — it belongs in the conversation with any practitioner managing your sleep or your hormonal health.
- Take temperature seriously as a sleep variable. Core body temperature drops are a prerequisite for sleep onset. A cooler bedroom (16–19°C is the range with the most evidence) supports this. In the late luteal phase, when progesterone raises body temperature, this matters more — the gap between external cooling and internal warmth is larger and harder to close. A cool shower before bed, lighter bedding, or a bedroom cooler than usual in the premenstrual week are not arbitrary suggestions. They address a specific physiological obstacle.
- Protect the second half of the night for REM. Alcohol, late eating, early alarms, and SSRI use all disproportionately suppress REM sleep. Because REM dominates the second half of the night, any disruption or shortening of sleep time cuts REM first. The cognitive, emotional, and memory functions that REM runs are not running at full capacity when REM is suppressed. If emotional regulation and memory feel particularly impaired on short or disrupted nights, this is the mechanism.
- For melatonin: timing matters more than dose. For delayed melatonin release — the type my father and I have, where the pineal gland releases melatonin later than the social clock demands — low-dose melatonin (0.5 to 1mg) taken 90 to 120 minutes before the desired sleep time is more effective than higher doses taken close to bedtime. Higher doses do not produce stronger or faster sleep onset. They produce grogginess and can suppress endogenous melatonin production over time. The goal is a signal, not sedation.
- Light before dark. The melatonin system requires a strong morning light signal to calibrate its evening release time. Bright light exposure — ideally sunlight, outside — in the first 30 to 60 minutes after waking sets the circadian clock and determines when melatonin will rise that evening. For people with delayed melatonin release, morning light is often more effective at advancing the sleep phase than any supplement taken at night.
- If sleep worsens at perimenopause, name it specifically. Perimenopausal sleep disruption — night sweats, early morning waking, insomnia onset — is driven by estrogen fluctuation and withdrawal. It is not a general sleep disorder. It responds to hormonal stabilisation, which means the conversation needs to happen with a clinician who will consider HRT as an option and evaluate the hormonal picture rather than treating it as anxiety or ageing.
Sleep is the most consequential thing most of us do every day, and it is the thing we are least likely to have been given a useful framework for. The science of what happens during those hours has changed substantially in the last two decades. The clinical communication of that science has not kept pace.
I sleep better now than I did ten years ago. The melatonin timing is something I can work with. The light sleeping from my mother I have mostly accepted — it is, in some sense, a form of attunement that also has uses. The premenstrual window I now plan around rather than fighting through.
What changed was having the specific information. The mechanisms. The cycle-phase picture. The understanding of what my brain needs to do every night that I was interrupting by not giving it enough time or the right conditions to do it.
Sleep is not passive. You just have to stop being awake long enough to let it run.
There is considerably more on sleep in The Art of Female Health — the specific protocols, supplement timing, the cycle-phase strategies that the general sleep literature doesn't cover because it was built on the wrong population, and the interventions that actually move the needle when the standard advice doesn't fit your body. If this post opened something up, that's where I went further with it. You can find the book here.
– Nina- Xie, L., Kang, H., Xu, Q., Chen, M.J., Liao, Y., Thiyagarajan, M., & Nedergaard, M. (2013). Sleep drives metabolite clearance from the adult brain. Science, 342(6156), 373–377. doi:10.1126/science.1241224
- Walker, M.P. (2009). The role of sleep in cognition and emotion. Annals of the New York Academy of Sciences, 1156(1), 168–197. doi:10.1111/j.1749-6632.2009.04416.x
- Irwin, M.R. (2015). Why sleep is important for health: a psychoneuroimmunology perspective. Annual Review of Psychology, 66, 143–172. doi:10.1146/annurev-psych-010213-115205
- Baker, F.C., & Driver, H.S. (2007). Circadian rhythms, sleep, and the menstrual cycle. Sleep Medicine, 8(6), 613–622. doi:10.1016/j.sleep.2006.09.011
- Cappuccio, F.P., D'Elia, L., Strazzullo, P., & Miller, M.A. (2010). Sleep duration and all-cause mortality: a systematic review and meta-analysis of prospective studies. Sleep, 33(5), 585–592. doi:10.1093/sleep/33.5.585