Sleep and Longevity: A Doctor's 4 Sleep Fixes (That Actually Work on Your Nervous System)

I am not a natural sleeper

I've struggled with sleep for most of my life. For a long time, I didn't notice. Early in my adulthood, in my years as a resident physician, my schedule was so brutal that my body never had a real chance to sleep well anyway. You can't tell you're swimming badly when you're just trying to keep your head above water.

But as I started paying serious attention to the evidence on sleep and longevity, and as my hours finally allowed something resembling a normal life, the problem didn't go away. It had just been masked. Many of my relatives share it: a genetic tendency toward lighter sleep, less deep slow-wave, easier arousal in the middle of the night. I've tried most things. Weighted blankets, every iteration of magnesium, melatonin, prescription options. Things that helped, things that didn't.

I sleep well most days now. But it took real research and real effort to get here. That's the honest version of this guide. Not an academic exercise, but something written by a doctor and a human who has had to actually figure this out, with clinical data to lean on when the "try chamomile tea" crowd runs out of ideas.

The mortality data is hard to argue with

A 2024 meta-analysis in Scientific Reports found that short sleep (under 6 hours) increased all-cause mortality risk by 14%, and long sleep (over 9 hours) by 34%. A 2025 OHSU study confirmed that consistently sleeping fewer than 7 hours per night was associated with measurably decreased life expectancy.

But here's the part that changed how I talk to patients about this. A 2023 study in SLEEP found that sleep *regularity* was a stronger predictor of mortality than total sleep duration. The most irregular sleepers had significantly higher all-cause mortality even when their average hours looked fine. That means the person sleeping 7 hours but at wildly different times may be worse off than someone consistently getting 6.5.

The biological aging data goes further. A 2024 study in Psychoneuroendocrinology showed that short sleep combined with insomnia was associated with accelerated epigenetic aging. Your DNA methylation patterns literally look older. A Mendelian randomization study confirmed a likely causal relationship between sleep traits and epigenetic clock acceleration. And a 2024 SLEEP abstract showed greater sleep regularity was associated with a slower pace of epigenetic aging.

For anyone who thinks of sleep as a passive, optional activity: your epigenetic clock disagrees.

Your nervous system is the problem

Most sleep advice: dark room, cool temperature, no screens. Fine. Also insufficient for the millions of people who do all of that and still don't sleep as well as they should.

The reason is neurological. Sleep requires your autonomic nervous system to shift from sympathetic (fight-or-flight) to parasympathetic (rest-and-repair) dominance. During healthy NREM sleep, heart rate drops, blood pressure falls, cortisol is suppressed, and your body enters a genuine repair state. In people with suboptimal sleep, this switch never fully happens. You might be horizontal for 8 hours while your nervous system runs like you're being chased.

That's why the four interventions below work. They don't address "sleep habits." They target the autonomic nervous system directly.

What happens during a normal night of sleep

Sleep cycles through four stages roughly every 90 minutes. N1 and N2 are light sleep, about 50% of the night. N3 is deep sleep (slow-wave sleep, or SWS). REM rounds out each cycle.

Deep sleep is where the heavy repair happens. This is not optional maintenance. It is the single most physiologically active repair state your body enters.

The glymphatic system, first characterized by Nedergaard's group in 2012, is a brain-wide waste clearance network that operates almost exclusively during NREM slow-wave sleep. During SWS, cortical neurons fire in synchronized slow oscillations (0.5 to 4 Hz). This coordinated firing causes neurons to briefly enter a "down state" where they stop signaling and physically shrink, expanding the interstitial space by roughly 60%. That expansion allows cerebrospinal fluid (CSF) to flow through perivascular channels (Virchow-Robin spaces) along arterioles, mix with interstitial fluid (ISF), and flush soluble waste products out through perivenous drainage pathways.

The waste products being cleared include beta-amyloid and phosphorylated tau, the proteins that accumulate in Alzheimer's disease. This isn't speculative. A 2019 Science study (Fultz et al.) used fast fMRI to show that slow-wave neural activity directly drives large pulsatile waves of CSF flow through the sleeping brain. The coupling is tight: neural slow waves first, then a hemodynamic shift, then CSF influx within seconds. A single night of sleep deprivation measurably increases beta-amyloid accumulation in the human brain, as shown by Shokri-Kojori et al. (2018) using PET imaging. Tau accumulation follows a similar pattern in longitudinal studies.

The molecular gatekeeper is aquaporin-4 (AQP4), a water channel densely expressed on astrocytic endfeet that wrap around blood vessels. AQP4 facilitates the convective flow of CSF into brain tissue. Genetic variants that reduce AQP4 expression are associated with increased amyloid accumulation and higher Alzheimer's risk. Mouse models with AQP4 knockout show a roughly 70% reduction in glymphatic clearance. Sleep position may even matter: lateral (side) sleeping enhances glymphatic transport compared to supine or prone positions in preclinical studies.

The implication for anyone sleeping 6 to 7 hours: you are likely truncating your deepest SWS periods, which are concentrated in the first half of the night, and meaningfully reducing the window during which your brain clears neurotoxic waste. Over decades, that adds up.

The long-term stakes just got clearer. A 2025 longitudinal study (Cavaillès, Carnethon, Knutson et al.) tracked midlife sleep characteristics over 20 years and found that sleep quality in your 40s and 50s directly predicts Alzheimer's biomarker burden decades later. The mechanism is the same: years of disrupted slow-wave sleep equals years of incomplete glymphatic clearance, amyloid accumulating quietly in the background. The intervention window isn't when you're 70. It's now.

Growth hormone and slow-wave sleep. The largest pulse of growth hormone (GH) secretion in the 24-hour cycle occurs during the first episode of deep slow-wave sleep, typically within the first 90 minutes of sleep onset. In men, approximately 70% of daily GH secretion occurs during sleep, with the first SWS bout triggering a GH pulse that can reach 10 to 20 ng/mL. In women, the pattern is more distributed but SWS remains the dominant trigger.

The relationship is causal, not merely correlational. Pharmacologically suppressing SWS with acoustic stimulation or benzodiazepines reduces GH release proportionally. Shift workers who sleep during the day but at abnormal circadian phases show blunted GH pulses even with equivalent sleep duration, because the circadian gate for GH release aligns with nighttime SWS, not just sleep per se.

Why this matters for longevity: GH drives tissue repair, protein synthesis, lipolysis, and bone density maintenance. The age-related decline in GH secretion ("somatopause") tracks closely with the age-related decline in SWS. By age 40, most adults have lost 60 to 70% of the deep sleep they had at 20. The GH decline follows in parallel. Optimizing SWS through the interventions below is one of the few ways to support endogenous GH secretion without exogenous hormones.

REM sleep handles emotional processing, procedural memory, and synaptic optimization. NREM strengthens neural representations; REM integrates and refines them. Deep sleep dominates the first half of the night. REM dominates the second half.

Circadian biology: why timing is everything

Your circadian clock is not a metaphor. It is a physical molecular oscillator running in nearly every cell of your body, and its disruption is one of the most underappreciated drivers of metabolic disease, immune dysfunction, and accelerated aging.

The master clock lives in the suprachiasmatic nucleus (SCN), a pair of tiny nuclei in the anterior hypothalamus containing roughly 20,000 neurons. The SCN receives direct light input from intrinsically photosensitive retinal ganglion cells (ipRGCs) expressing melanopsin, which are maximally sensitive to blue light at approximately 480 nm. This light signal synchronizes ("entrains") the central clock to the external light-dark cycle.

At the molecular level, the clock runs on a transcription-translation feedback loop that takes approximately 24 hours to complete one cycle. The core loop works like this: the transcription factors CLOCK and BMAL1 heterodimerize and drive expression of their own repressors, the PER (PER1, PER2, PER3) and CRY (CRY1, CRY2) genes. As PER and CRY proteins accumulate, they form a complex that inhibits CLOCK/BMAL1, suppressing their own transcription. As PER/CRY are degraded by casein kinases and ubiquitin ligases, the inhibition lifts and the cycle restarts. Auxiliary loops involving REV-ERBα/β and RORα add stability and fine-tune the oscillation.

This matters because CLOCK/BMAL1 don't just regulate each other. They regulate thousands of downstream genes, roughly 10 to 15% of the transcriptome, including genes controlling glucose metabolism, lipid processing, immune cell activation, DNA repair, and hormone secretion.

Here's the critical part most people miss: the SCN is the master pacemaker, but peripheral clocks in the liver, pancreas, gut, skeletal muscle, adipose tissue, and immune cells run their own molecular oscillators using the same CLOCK/BMAL1/PER/CRY machinery. These peripheral clocks are entrained not primarily by light, but by feeding, temperature, and hormonal signals. When you eat at consistent times, your liver clock synchronizes with the central SCN clock. When you eat erratically, peripheral clocks drift out of phase with the central clock, a state called "internal desynchrony."

Internal desynchrony is measurably harmful. Circadian misalignment studies, where researchers shift feeding or sleep schedules by 12 hours, show rapid deterioration in glucose tolerance (postprandial glucose rises 20 to 30% within days), increased inflammatory markers (CRP, IL-6), disrupted cortisol rhythms, and impaired leptin signaling. Shift workers, who live in chronic circadian misalignment, have significantly higher rates of type 2 diabetes, cardiovascular disease, and certain cancers.

"Social jet lag," the mismatch between weekday and weekend sleep timing, produces a milder version of the same physiology. A study of over 400,000 UK Biobank participants found that each hour of social jet lag was associated with an 11% increase in cardiovascular disease risk. You don't need to be a night-shift nurse to suffer circadian disruption. Sleeping from midnight to 7am on weekdays and 2am to 10am on weekends imposes a 2-hour phase shift twice a week. Over years, that matters.

The biggest circadian disruptors, roughly in order of impact:

1. Irregular light exposure. Blue-enriched light after sunset suppresses melatonin via ipRGCs. You already know this. What most people underestimate is the importance of bright morning light. Bright light (>10,000 lux, roughly equivalent to outdoor daylight) in the first hour after waking is the single strongest signal for entraining the SCN. Without it, your central clock drifts.
2. Irregular meal timing. Peripheral clocks in the liver and pancreas are entrained by food. Eating at random times desynchronizes these from the central clock. Time-restricted eating (consistent 8 to 10-hour feeding window) helps realign peripheral oscillators.
3. Irregular sleep/wake timing. Variable bedtimes create social jet lag. The regularity data tells you what that costs in mortality.
4. Core body temperature disruption. Sleep onset requires a 1 to 2°F drop in core temperature. Late exercise, hot environments, and alcohol all prevent this.

HRV: the best real-time window into sleep quality

Heart rate variability (HRV) is the variation in time between consecutive heartbeats. It is the single best non-invasive, real-time marker of autonomic nervous system balance we have, and it tells you more about your sleep quality than hours logged or subjective "how do I feel" assessments.

During high-quality sleep, parasympathetic tone dominates. The vagus nerve slows heart rate and increases beat-to-beat variability. A night of good sleep shows high HRV (specifically, high RMSSD, the time-domain metric most wearables report), low resting heart rate, and a characteristic pattern: HRV rises through the first half of the night during deep sleep, dips slightly during REM, and remains elevated until the cortisol awakening response begins driving sympathetic activation before you wake.

During poor sleep, the pattern inverts. Sympathetic tone stays elevated. HRV is suppressed throughout the night. Resting heart rate stays higher. The result is that you wake up with a low HRV baseline and elevated resting heart rate, often feeling unrested even after "8 hours."

What makes HRV particularly useful for sleep optimization is that it responds to interventions within days. Start a breathing protocol, change your exercise timing, cut alcohol: you will see the effect in overnight HRV within 1 to 2 weeks. This makes it a practical feedback tool, not just a diagnostic metric.

Baseline HRV varies enormously between individuals (age, fitness, genetics all matter), so absolute numbers are less meaningful than your own trend. A 35-year-old athlete might have overnight RMSSD of 80+ ms. A 55-year-old sedentary adult might sit at 25 ms. Both are "normal." What matters is: are your numbers trending up with interventions, or stagnant?

Fix 1: Weighted blankets

This surprises people. A weighted blanket sounds like a comfort item, not medicine. But the mechanism is real, and it starts in the peripheral nervous system.

Deep pressure stimulation activates low-threshold mechanoreceptors in the skin, specifically C-tactile afferents and Aβ fibers in hairy skin. These receptors send afferent signals through the vagus nerve to the nucleus tractus solitarius (NTS) in the brainstem, which is the primary relay station for visceral and somatosensory vagal input. From the NTS, signals project to the dorsal motor nucleus of the vagus, the locus coeruleus (reducing norepinephrine release), and the limbic system. The net effect: parasympathetic output increases, sympathetic output decreases.

The downstream neurochemistry is relevant. Deep pressure has been shown to increase serotonin release (the precursor to melatonin), increase GABA activity (the primary inhibitory neurotransmitter, which reduces neuronal excitability and promotes the thalamocortical oscillations that produce deep sleep), and decrease cortisol. The GABA connection is particularly important. GABA is the neurotransmitter most directly responsible for the transition from wakefulness to sleep. GABAergic neurons in the ventrolateral preoptic area (VLPO) of the hypothalamus actively inhibit the wake-promoting arousal centers. Anything that supports GABAergic tone supports sleep onset and maintenance.

The same mechanism explains why swaddling calms infants, why compression vests reduce anxiety in autism spectrum disorders, and why Temple Grandin's "squeeze machine" works. It's vagal afferent stimulation, translated into pressure.

A 2020 RCT (Ekholm et al., Journal of Clinical Sleep Medicine) found that adults using a weighted blanket (about 10% of body weight) for 4 weeks had significantly improved sleep maintenance, reduced daytime sleepiness, and lower insomnia severity scores compared to controls. A 2024 meta-analysis in the Journal of Psychiatric Research confirmed improvements in sleep quality and reduced anxiety across multiple psychiatric populations.

I should be honest: most weighted blanket studies are in psychiatric populations or have small sample sizes. We don't have large-scale RCTs in the general healthy population yet. The mechanistic rationale is strong, the risk is zero, and I recommend them freely. But I won't pretend the evidence base is airtight.

What to do: Get a blanket that's roughly 10% of your body weight. Use it every night. It's cheap and it can't hurt you.

Fix 2: Paced breathing and the physiological sigh

Controlled breathing is probably the most powerful tool we have for rapid autonomic regulation. Two techniques matter here, and the physiology of why they work is worth understanding.

Paced diaphragmatic breathing at about 6 breaths per minute (5 seconds in, 5 seconds out) maximizes respiratory sinus arrhythmia (RSA), the natural variation in heart rate that occurs with each breath. During inhalation, intrathoracic pressure drops, venous return increases, and heart rate transiently rises. During exhalation, vagal output increases and heart rate slows. At 6 breaths per minute, this oscillation resonates with the baroreflex loop (the blood pressure feedback circuit), creating a maximal amplitude swing in HRV. This is the basis of heart rate variability biofeedback, one of the few interventions that reliably increases both acute and resting HRV.

Higher HRV at bedtime correlates with faster sleep onset, more deep sleep, and lower nocturnal cortisol.

The physiological sigh is more targeted, and the gas exchange mechanics explain why it works so well. Two short inhales through the nose, then one long exhale through the mouth.

The double inhale serves a specific purpose. Your lungs contain roughly 500 million alveoli. Over the course of normal breathing, some alveoli gradually collapse (atelectasis), reducing the surface area available for gas exchange. CO2 builds up in the blood, which is an independent driver of anxiety and arousal (CO2 is directly sensed by chemoreceptors in the brainstem reticular formation, and elevated CO2 triggers the suffocation alarm response). The two successive inhales snap collapsed alveoli back open, maximally expanding the gas exchange surface. This allows efficient CO2 offloading during the long exhale.

The extended exhale is where the parasympathetic magic happens. Exhalation activates the vagus nerve through lung stretch receptor feedback and the cardiac vagal neurons in the nucleus ambiguus. A longer exhale relative to inhale shifts the autonomic balance toward parasympathetic dominance within a single breath cycle. The exhale-dominant ratio (short inhale, long exhale) is why sighing is calming and gasping is activating.

A 2023 Stanford RCT published in Cell Reports Medicine (Balban et al.) found that just 5 minutes of daily cyclic sighing significantly reduced anxiety, improved mood, and lowered resting respiratory rate. More effective than mindfulness meditation.

I find this remarkable, honestly. Five minutes. No app, no device, no prescription.

My protocol: 5 minutes of cyclic sighing before bed. If you wake during the night, 2 to 3 minutes of paced breathing (6 breaths per minute) while lying still. Do not reach for your phone.

Fix 3: Diet and the gut-brain axis

The connection between food and sleep goes beyond "don't eat too late." Your gut produces about 95% of your body's serotonin, the direct precursor to melatonin. What you eat determines how much serotonin substrate is available, how inflamed your gut lining is, and how well the gut-brain axis communicates via the vagus nerve.

A 2024 systematic review in Nutrients evaluated 20 studies on the Mediterranean diet and sleep, finding consistent associations between higher adherence and better sleep quality, longer duration, and fewer nighttime awakenings. A 2025 meta-analysis in Sleep Medicine Reviews confirmed this across multiple populations, though it noted that most evidence is observational.

The mechanism makes sense. The Mediterranean diet is rich in tryptophan (fish, poultry, nuts, seeds), which is the amino acid precursor to serotonin and then melatonin. The conversion pathway is: tryptophan → 5-HTP (via tryptophan hydroxylase) → serotonin → N-acetylserotonin → melatonin (via AANAT and ASMT enzymes, which are upregulated by darkness). Tryptophan competes with other large neutral amino acids (LNAAs) for transport across the blood-brain barrier via the LAT1 transporter. This is why timing matters: a carbohydrate-containing meal 3 to 4 hours before bed triggers insulin release, which drives competing LNAAs into muscle, improving tryptophan's relative transport ratio into the brain. A high-protein meal without carbohydrate raises all LNAAs equally and can paradoxically reduce brain tryptophan uptake. The practical takeaway: your dinner should contain both tryptophan-rich protein (salmon, turkey, pumpkin seeds) and some complex carbohydrate (sweet potato, brown rice) rather than pure protein.

Omega-3 fatty acids from fish reduce systemic inflammation and support vagal tone. Magnesium from leafy greens and nuts acts as a GABA receptor co-agonist, promoting neuronal inhibition and muscle relaxation. Polyphenols and fiber support a diverse gut microbiome, which modulates serotonin production.

A note on magnesium testing. Serum magnesium is unreliable. It reflects extracellular magnesium, which is tightly homeostatic and stays in range even when intracellular stores are depleted. RBC magnesium is a far better marker of intracellular status. I check it on every patient with sleep complaints. Optimal range is roughly 5.0 to 6.5 mg/dL. Below 4.5, supplementation makes a measurable difference. The two forms I recommend: magnesium glycinate (for general relaxation and bioavailability) and magnesium L-threonate (which crosses the blood-brain barrier and has shown cognitive benefits in preclinical studies). Standard dose: 200 to 400 mg elemental magnesium before bed.

What to avoid matters just as much. High-glycemic meals close to bedtime cause reactive hypoglycemia overnight, triggering cortisol and epinephrine release. Alcohol suppresses REM sleep and fragments sleep architecture even in moderate amounts. Caffeine has a half-life of 5 to 7 hours and blocks the adenosine receptors that build sleep pressure. If you're a slow metabolizer (CYP1A2 *1F variant, present in roughly 50% of the population), caffeine's half-life can exceed 8 hours. A noon coffee might still be affecting adenosine signaling at 10pm.

What to do: Last meal 3+ hours before bed. Include both tryptophan-rich protein and complex carbohydrate at dinner. Consider 200 to 400 mg of magnesium glycinate or threonate before bed. Cut caffeine by noon if you're a slow metabolizer. If you're curious, check your CYP1A2 genotype.

Fix 4: Exercise timing and circadian entrainment

Exercise is the strongest lifestyle intervention for sleep quality. A 2024 network meta-analysis of RCTs in Frontiers in Psychology found that combined exercise (aerobic plus resistance), four times per week at high intensity for 30 minutes or less, had the greatest effect.

But timing matters more than people realize. A 2025 study found that higher exercise strain combined with later timing was dose-dependently associated with delayed sleep onset, shorter sleep duration, lower sleep quality, higher resting heart rate, and lower HRV.

The reason is core body temperature. Sleep onset requires your core temp to drop by 1 to 2°F. Vigorous exercise raises it for 1 to 2 hours afterward. Exercise hard at 8pm, try to sleep at 10pm, and your body temperature hasn't fallen enough. Your sympathetic nervous system is still running hot.

What fewer people know is that morning exercise serves as a powerful circadian zeitgeber (time-giver) through multiple pathways. First, outdoor morning exercise provides bright light exposure, the strongest entrainment signal for the SCN. Second, the acute rise in core body temperature from exercise sends a phase-setting signal to peripheral clocks in muscle, liver, and adipose tissue. Third, exercise-induced cortisol release in the morning reinforces the cortisol awakening response, sharpening the amplitude of the cortisol rhythm (higher morning peak, lower evening trough). A flattened cortisol curve is both a marker and a cause of poor sleep. Morning exercise steepens it.

A systematic review of exercise timing and circadian rhythm (PMC, 2023) found that morning and afternoon exercise enhanced physiological circadian markers and improved subsequent sleep quality. Late-evening high-intensity exercise impaired multiple sleep parameters.

The circadian entrainment effect is why I specifically recommend morning exercise rather than just "not too late." It's not only about avoiding the temperature problem. Morning exercise actively sets the clock, improving both sleep timing and sleep quality that night.

What to do: Finish vigorous exercise at least 3 hours before bedtime. Morning is ideal because outdoor morning exercise gives you bright light exposure, reinforces the cortisol awakening response, and phase-entrains peripheral clocks. If evening is your only option, keep it moderate: yoga, walking, light resistance work. Finish at least 2 hours before bed.

Sleep tracking: what metrics matter and what's noise

Wearable sleep trackers (Oura Ring, Whoop, Apple Watch, and others) have become ubiquitous. They generate a lot of data. Most of it is useful only if you know what to pay attention to and what to ignore.

What the devices actually measure. All consumer wearables use accelerometry (motion detection) and photoplethysmography (optical heart rate from the wrist or finger) to estimate sleep parameters. None of them measure brain waves. Sleep staging (light, deep, REM) is inferred through algorithms that combine movement, heart rate, HRV, and sometimes skin temperature. This distinction matters.

HRV: the most reliable and actionable metric. Overnight HRV (RMSSD) is measured directly from the optical heart rate sensor and does not require sleep staging inference. It is the most accurate metric these devices provide, and it is the most useful for tracking sleep quality and intervention response. Watch your overnight HRV trend over 2 to 4 weeks when you change a variable (start a breathing protocol, shift exercise timing, cut alcohol). If overnight HRV goes up, the intervention is likely working.

Resting heart rate: also reliable. Lower overnight resting heart rate reflects stronger parasympathetic tone during sleep. Like HRV, it's a direct measurement, not an estimate. Trend it over weeks.

Sleep stages: take with a grain of salt. This is where consumer devices are weakest. Validation studies comparing wearables to polysomnography (the gold standard, which uses EEG, EMG, and EOG) show that most devices are reasonably accurate at distinguishing sleep from wake and at estimating total sleep time. They are considerably less accurate at distinguishing deep sleep from light sleep, and REM detection varies substantially between devices and even between firmware versions. Oura Gen 3 tends to be among the more accurate for sleep staging; Apple Watch performs well for total sleep time but less consistently for stages.

The practical advice: use sleep stages as a rough directional signal, not as a precise measurement. If your device consistently shows very low deep sleep over weeks, that's worth paying attention to. But don't obsess over night-to-night variation in stage percentages. The algorithms are not accurate enough for that to be meaningful.

Total sleep time: generally accurate. Most devices get this within 15 to 30 minutes of polysomnography readings. Reliable enough to track.

Sleep latency (time to fall asleep): unreliable. Devices can't distinguish lying still and awake from light stage 1 sleep. Don't rely on this number.

What to actually track: Overnight HRV trend (weekly average), overnight resting heart rate trend, total sleep time, and wake time consistency. These four metrics, trended over weeks, give you a genuinely useful picture of whether your sleep is improving. Ignore the nightly "sleep score" composite numbers. They mix reliable and unreliable metrics into a single number that obscures more than it reveals.

One caution: sleep tracking can become counterproductive. "Orthosomnia," the anxiety-driven pursuit of perfect sleep scores, is a real phenomenon documented in sleep medicine literature. If checking your morning data makes you more anxious about sleep, stop checking. The tool should serve you, not the reverse.

Bonus: Temperature therapy and the sleep-onset signal

This one doesn't get enough attention in longevity circles, probably because it sounds more like a spa day than a clinical protocol. The mechanism is real.

Sleep onset requires a drop in core body temperature of roughly 1 to 2°F. Your body achieves this by dilating blood vessels near the skin surface and radiating heat outward. The brain reads the temperature drop as a sleep-initiation signal, triggering the hypothalamic cascades that shift autonomic tone from sympathetic to parasympathetic and begin melatonin secretion. This is why cool rooms help you sleep and why hot environments don't. It's also why your hands and feet get warm before you fall asleep — peripheral vasodilation is how heat leaves the body.

Sauna before bed: counterintuitive, and it works. A sauna session raises core temperature significantly. But 60 to 90 minutes after you exit, your body enters an aggressive cooling phase to compensate — radiating heat rapidly to restore baseline. If timed correctly, this rebound drop hits exactly when you're trying to sleep, amplifying the natural temperature-drop signal. A 2019 meta-analysis in Sleep Medicine Reviews found that passive body heating within 1 to 2 hours of bedtime increased slow-wave sleep and decreased sleep onset latency by about 10 minutes on average. The sweet spot is a 15 to 20 minute sauna session ending 60 to 90 minutes before bed. Too close to bedtime and you're fighting elevated core temp instead of surfing the rebound.

The sauna effect also involves heat shock proteins (HSPs), molecular chaperones upregulated by thermal stress that support cellular repair and have been associated with cardiovascular protection in the Laukkanen cohort studies. Regular sauna use (2 to 3x/week) is associated in Finnish longitudinal data with reduced all-cause mortality, improved cardiovascular outcomes, and subjectively better sleep quality. Whether the sleep benefit is primarily thermoregulatory or also mediated by HSP induction isn't fully resolved.

Cold plunge: a morning tool, not a sleep tool. Acute cold exposure drives a norepinephrine surge of 200 to 300%. That's acutely alerting and sympathetically activating — the opposite of what you want at bedtime. Morning cold exposure, however, is genuinely useful for circadian entrainment. It reinforces the cortisol awakening response, sharpens the amplitude of the cortisol rhythm, and combined with bright morning light, sends a strong zeitgeber signal to peripheral clocks. The post-cold parasympathetic rebound also improves HRV over time with consistent practice.

Contrast therapy (alternating heat and cold): The autonomic oscillation from alternating heat and cold — sympathetic activation during cold immersion, parasympathetic rebound after — may produce a stronger net parasympathetic shift than either modality alone, though direct sleep data here is limited. If you're doing contrast therapy, morning is the better timing. Evening contrast therapy likely carries the same risks as evening high-intensity exercise: you're activating the sympathetic nervous system at the wrong point in the circadian cycle.

What to do: If you have sauna access, use it 3 to 4x/week. For sleep optimization, finish 60 to 90 minutes before bed and let the thermoregulatory rebound work for you. For cold exposure, do it in the morning — it's a circadian tool, not a sleep tool.

Lab markers worth checking

If you've optimized the interventions above and your sleep is still not where you want it, it's worth looking at the underlying biology. Here's what I order:

If you've optimized the four fixes above and your sleep is still mediocre, it's time to look at the biology. Here's what I order:

• AM cortisol (7 to 9am draw): elevated levels suggest HPA axis hyperactivation
• PM cortisol (3 to 5pm draw): should be much lower than AM. If AM and PM are similar (flat cortisol curve), that's chronic stress physiology
• DHEA-S: the cortisol-to-DHEA ratio reflects adrenal stress reserve
• Thyroid panel (TSH, free T4, free T3): both hypo- and hyperthyroidism disrupt sleep. Subclinical hyperthyroidism is a common, overlooked cause of fragmented sleep
• Ferritin: below 30 to 50 ng/mL is associated with restless legs syndrome, which is frequently missed
• Fasting insulin and HbA1c: insulin resistance causes nocturnal hypoglycemia and reactive cortisol spikes
• Magnesium RBC: serum magnesium is unreliable (as noted above). RBC magnesium is a better marker of intracellular status. Optimal 5.0 to 6.5 mg/dL. Deficiency is common and directly impairs GABA-mediated neuronal inhibition

My honest take

I'm more confident about sleep than almost anything else in longevity medicine. The mortality data is consistent across populations. The epigenetic clock data is compelling. The mechanisms, from glymphatic clearance to GH secretion to circadian gene expression, are well characterized. And unlike rapamycin or metformin, optimizing sleep carries no meaningful risk.

I should be transparent about what we don't know. Most mortality data is observational. You can't randomize people to years of sleep deprivation (ethics committees tend to frown on that). The epigenetic clock studies are still sorting out whether the acceleration they measure is truly causal or a biomarker of broader physiological stress. The weighted blanket and breathing literature is still maturing.

What I tell patients: these interventions are low-risk, low-cost, and physiologically sound. They target the nervous system rather than treating symptoms. Even if individual effect sizes from any one study are modest, the cumulative benefit of better autonomic regulation, better sleep architecture, and more consistent circadian rhythm compounds over decades. Sleep is the foundation. Everything else in longevity medicine works better on top of it.

A brief note: if you snore loudly, gasp during sleep, or sleep 7 to 8 hours and wake consistently unrefreshed, get evaluated for obstructive sleep apnea. An estimated 80% of moderate-to-severe cases remain undiagnosed, and it independently increases cardiovascular mortality and accelerates cognitive decline. A home sleep test is a reasonable first screen.