Podcast Summary | Dr Matt Walker How To Structure Your Sleep Use Naps Time Caffeine

Huberman Labwith Dr. Matthew Walker· · Episode aired April 17, 2024· 14 min read
Listen to this summary

High-Level Overview

In the third installment of their six-part sleep series, Andrew Huberman and sleep scientist Dr. Matthew Walker of UC Berkeley turn to how we should structure our sleep. They trace how sleep changes from the womb to old age, moving from the many short sleeps of infancy to one consolidated night, and how the mix of REM and non-REM sleep shifts along the way. They then weigh the options for adults: one bout of sleep (monophasic), two (biphasic, including the afternoon nap and the historical "first sleep, second sleep"), or many (the polyphasic schedules popular with biohackers).

The heart of the episode is practical. Dr. Walker explains what naps do for learning, emotion and alertness, who should avoid them, how long and how late to nap, and how to avoid post-nap grogginess. He then explains how caffeine works on the brain's sleep-pressure system, the "nappuccino, and why he has changed his mind about caffeine. The tone is warm, nerdy and often playful, with Huberman's jokes about his post-nap expression and Walker's storytelling (including where the term power nap" came from), all balanced against careful readings of the evidence.

Table of contents

The Discussion Explores:

  • Sleep across the lifespan: from polyphasic infants to monophasic adults, and why babies spend about half their sleep in REM.
  • Biphasic sleep and modern life: afternoon siestas in hunter-gatherer and Mediterranean cultures, and why Walker doubts "first sleep, second sleep" is our natural design.
  • Chronotypes: why they are largely genetic and why human groups benefit from people sleeping at different times.
  • The benefits and risks of naps: better learning, emotional balance and alertness, but lost sleep pressure at night.
  • The optimal nap: about 20 minutes, not after about 3 p.m., and how to avoid sleep inertia.
  • Naps in insomnia and aging: who should skip them, and why daytime napping in older adults may signal poor night sleep.
  • Caffeine: how it blocks adenosine, the caffeine crash, the "nappuccino," dose and timing, and its health benefits.
  • Polyphasic sleep schedules: what the evidence says, and the real danger of drowsy driving.

Deep Dive into Main Ideas

How Sleep Changes From Infancy to Adulthood

Dr. Walker defines the terms: monophasic sleep is one bout of sleep in 24 hours, biphasic is two, and polyphasic is many. Newborns are highly polyphasic, cycling between sleep and waking about every two hours for two reasons: they need to feed roughly that often, and their suprachiasmatic nucleus, the brain's master 24-hour clock, isn't yet fully developed, so they don't know when it's light or dark. By age one, the number of sleep bouts is dropping; by two or three, sleep is consolidating into the night; kindergarteners are typically biphasic (hence nap time, where any teacher will tell you the child who skips the nap becomes the "loose cannon"); and by around five or six, children are monophasic, a pattern that continues through adulthood.

The stages of sleep also change. In the womb, the fetus seems to spend much of its time in a REM-like state, and because the muscle paralysis of REM hasn't yet developed, those dream-state bursts show up as the kicks a mother feels. Newborns sleep 14 to 17 hours a day, and almost 50% of that sleep is REM, compared with about 20% in adults. Walker explains that "REM sleep acts as an electrical fertilizer" for the explosion of new synapses in the infant brain, like an internet provider wiring every house in a new neighborhood. Animal studies depriving young animals of REM sleep stunted brain development. He notes REM abnormalities have been seen in autism and ADHD, but stresses there is no causal evidence yet.

Between six months and two years, total sleep and REM decline while non-REM rises, with a peak in stage 2 sleep spindles that Walker links to the development of motor skills and walking. By about five or six, the ratio stabilizes at roughly 80% non-REM to 20% REM. In the teenage years, deep sleep appears to drive synaptic pruning, removing little-used connections to make the brain more efficient. Along the way, Walker fits in some animal trivia: elephants sleep as little as four hours, the little brown bat sleeps 17 to 18 hours a day, and birds and bats avoid falling from their perches by having REM periods lasting only a few seconds.

Biphasic Sleep, "First Sleep, Second Sleep" and Why Lying Down Makes Us Sleepy

Adults are mostly monophasic, but Walker says we may not be sleeping as we were designed to. We all have a post-lunch dip in alertness somewhere between about 1 and 4 p.m. Hunter-gatherer tribes often nap in the afternoon, especially in the hot, dry season, and Latin and Mediterranean cultures have the siesta. These groups also go to sleep, on average, about two hours after sundown and wake just before sunrise, triggered by the drop in temperature, so for them midnight really is the middle of the night. Huberman mentions Jocko Willink's famous 4:30 a.m. watch photos as an example of an extreme morning type.

Walker is skeptical of the popular idea that humans naturally sleep in two nightly chunks with a wakeful period in between. People in Britain and elsewhere did this between roughly the 15th and 19th centuries, but the one study often cited as biological evidence has, in his view, three problems: participants were kept in bed for 14 hours a night, the middle-of-the-night waking was only a probability rather than a clear, shared awakening, and it involved just seven healthy men. He remains open to new evidence but doesn't see strong support for it now. Asked why lying down makes us sleepy, Walker points to temperature: lying flat helps the body move warm blood from the core to the skin, releasing heat, and a drop in core body temperature is needed to fall asleep.

Chronotypes: Genetic, Adaptive and Partly Flexible

Why don't we all sleep at the same time? Walker argues that chronotype variation is protective. If everyone slept at once, the whole group would be vulnerable for eight hours. With some people sleeping from 8 p.m. to 4 a.m. and others from midnight to 8 a.m., each person gets a full night while the clan's vulnerable window shrinks by about 50%. Huberman compares this to the circadian clock itself running slightly longer than 24 hours, which allows it to adjust to seasonal changes in daylight.

Chronotype is highly, though not entirely, genetic. Walker says it is very unlikely for two extreme morning-type parents to have a night-owl child, and people tend to drift back to their inherited pattern when life allows, for example after retiring from touring in a rock band. But environment matters too: electric light at night, too much caffeine, and screens and social media can push a natural "neutral" type like Walker toward a 1 a.m. to 9 a.m. schedule.

What Naps Can Do for You

Walker's lab has shown striking benefits from naps. In one study, participants learned lists of facts at midday; one group then took a 90-minute nap opportunity while the other stayed awake watching a nature documentary. At 5 p.m., both learned new facts. The group that stayed awake saw their learning capacity steadily decline, while the nappers maintained or even improved theirs, a difference of about 20%. The learning benefit tracked the number of sleep spindles during the nap.

In another study, people who didn't nap rated fearful and angry faces as more fearful and angry by 5 p.m. The nappers showed a reduced fear response, blunted the usual rise in anger sensitivity, and rated happy faces more positively. As Walker puts it, a nap can "reset the magnetic north of your emotional compass." This benefit depended on REM sleep, not spindles. Naps also improve attention, concentration, energy, decision-making, and measures like blood pressure, cardiovascular health and immunity. Huberman adds the example of a highly productive Stanford colleague, a Howard Hughes investigator and National Academy member, who insists on a scheduled post-lunch nap even when visiting other universities.

The Downside: Sleep Pressure, Insomnia and Late Naps

The catch is adenosine, the chemical that builds up the longer we're awake and creates healthy sleepiness after about 16 hours. A nap is like the pressure valve on a steam cooker, releasing some of that sleep pressure. For people with insomnia, the advice is clear: don't nap, because it lowers your chances of sleeping well at night. Walker notes that avoiding naps is part of cognitive behavioral therapy for insomnia (CBT-I), and admits the advice is hard to follow, having had two bouts of reactive insomnia himself.

For people who sleep well, naps are fine, as long as they aren't too late. A late nap is like snacking before a main meal: it takes the edge off your appetite for sleep. And it doesn't only delay falling asleep; it can make night sleep shallower and more fragile, so you wake more often and struggle to get back to sleep. If you notice this, Walker suggests moving the nap earlier, such as right after lunch, and testing it with an "on-off-on" experiment: change the routine for about two weeks, then return to the old one. If things get worse again when you stop the intervention, he says, that is when you can start believing the change worked. Huberman calls this a proper negative and positive control.

Get new Huberman Lab summaries by email

The GIST of every new Huberman Lab summary in your inbox, daily or weekly. Free, and you can unsubscribe any time.

The 20-Minute Nap, Sleep Inertia and How to Nap

For most people who just want a quick reboot of alertness, concentration and motivation, Walker recommends about 20 minutes. Dose-response studies show that 5 to 10 minutes gives only a brief bump that fades quickly, while beyond about 15 to 17 minutes the benefits become meaningful and last into the evening. Stopping at 20 minutes also avoids sleep inertia, the groggy "sleep hangover" (Huberman's "post-nap expression) that comes from waking out of deep non-REM sleep, which typically begins after 30 to 40 minutes. Longer naps bring bigger and longer-lasting benefits, but you may feel worse for up to an hour before they kick in. His rule of thumb is that the dose and the timing make the poison": about 20 minutes, and not after about 3 p.m. for the average adult. What counts as optimal depends on what you want from the nap; getting the REM-dependent emotional benefits, for example, requires a longer nap.

For people who want to try napping, Walker's advice is to mimic nighttime conditions: block out light with an eye mask, block out noise with earplugs or a sound machine, take off your shoes, get under a blanket, set an alarm, and time the nap to your own afternoon dip. He also discusses liminal states such as yoga nidra and non-sleep deep rest (NSDR), a term Huberman coined for a shorter version of yoga nidra without intentions. Walker suspects these states produce "local sleep," where parts of the brain briefly show deep, slow-wave activity while the person remains awake, and the two discuss studying this together with high-density EEG.

NASA, Pilots and the Origin of the "Power Nap"

Walker says NASA worked this out in the 1980s. Astronauts in orbit may see 10 to 20 sunrises and sunsets a day, and the human is the weak link in a mission. NASA found that short naps produced almost a 20% boost in alertness and almost a 50% boost in task productivity, and the practice spread to its ground staff as a "NASA nap culture."

The term "power nap has a funnier origin. Sleep researchers David Dinges and Mark Rosekind studied napping in long-haul pilots, for whom landing is the riskiest phase. Surprisingly, a nap placed early in the flight worked better than one near the end. When they presented the findings to the Federal Aviation Administration, their proposed name, prophylactic napping," drew chuckles and was deemed a nonstarter for a mostly male pilot culture, so they renamed it the power nap.

Naps and Aging: A Signal of Poor Night Sleep

Large studies of adults over about 65 found that napping predicted worse health outcomes and a higher likelihood of early death. Walker thinks the naps are mainly a marker of poor sleep at night rather than a cause. Deep sleep begins declining in our mid to late 30s; by age 50 we have about 50% of the deep non-REM sleep we had at 17 or 18, and by 65 to 75 about 5%. Older adults' sleep also becomes much more fragmented, so napping may be an attempt to compensate. He says the field should not yet tell all older adults to stop napping, and he remains open to evidence either way.

How Caffeine Works, the Caffeine Crash and the "Nappuccino"

Caffeine works by blocking adenosine receptors. Walker's analogy is a room full of chairs: adenosine arrives to take the seats and build up sleepiness, but caffeine rushes in and occupies them first. Adenosine keeps building up, and caffeine has simply "hit the mute button on your sleepiness." When caffeine wears off, you feel all the adenosine at once, including what built up meanwhile, which is the caffeine crash that drives many people to more caffeine. Caffeine reaches peak concentration after about 12 to 17 minutes, so the lift people feel in the first five minutes of a morning coffee comes from something else: partly conditioning and anticipation, as Huberman suggests, and partly the warm drink raising core body temperature. Walker summarizes the temperature rules as: warm up to cool down to fall asleep, stay cool to stay asleep, and warm up to wake up.

The caffeine nap, or "nappuccino," takes advantage of that timing. Drink a quick espresso right before a 20-minute nap; you fall asleep before the caffeine takes effect, and wake as it peaks, getting the nap's benefits without the grogginess. Walker adds that apart from sleep, and especially non-REM sleep, very little clears adenosine. Walks, daylight and exercise raise alertness by overriding adenosine, not removing it. He explains that during deep sleep the less active brain simply produces less adenosine, which lets normal clearance catch up, so anesthesia or deep NSDR-like states might clear some too.

Delaying Morning Caffeine and Changing His Mind on Caffeine

Huberman clarifies that his suggestion to delay caffeine 90 to 120 minutes after waking was never a mandate; he often drinks it right away before exercising. Walker thinks the delay is worth testing for two reasons: caffeine makes the brain more metabolically active, building up more adenosine and a harsher afternoon dip, and early caffeine can mask poor sleep. He suggests a test: after getting through about two weeks of withdrawal, hold off caffeine in the morning and ask whether, after the first 90 minutes of normal sleep inertia, you feel rested and sharp. If not, your sleep deserves a closer look.

Walker also says he has changed his mind about caffeine. When his book came out he was, in his words, dictatorial and binary about it. But the health data for coffee are striking, with a U-shaped curve that turns downward beyond about three or four cups. The benefits appear to come largely from coffee's antioxidants, since decaf shows many of the same benefits: "It's not the caffeine, it's the coffee itself." His guidelines: no more than about three cups a day, and stop caffeine roughly 8 hours before bed if you're not sensitive, 10 to 12 hours for most, and 12 to 15 hours (and one cup) if you're very sensitive, which is partly genetic. Even people who fall asleep fine after an evening espresso can lose up to 20% of their deep sleep without knowing it. His recent study of Wall Street traders also found a cycle of caffeine as an "upper by day and alcohol as a downer" at night.

The "Nap Plus Plus" Stack

Walker describes a Japanese study with five groups: no nap, nap, nap plus caffeine, nap plus washing the face and hands with cold water, and nap plus bright light (2,000 lux) on waking. Napping alone improved alertness and cognition and reduced sleepiness, and each add-on gave an extra, though smaller, boost. Cold water on the face and hands constricts surface blood vessels and pushes blood to the core, slightly raising core temperature, plus a shot of adrenaline. The study didn't test all the add-ons together, but Walker suspects they'd be additive: espresso, a 20-minute nap, cold water on the hands and face, then five to 10 minutes of daylight. Huberman notes the same tools, light, cold water and caffeine timing, apply to waking up in the morning, and Walker points out that we instinctively do the reverse at night by washing with warm water.

Polyphasic Schedules and the Real Danger of Drowsy Driving

Walker traces polyphasic sleep in adults to the biohacking and quantified-self movements of the late 1990s and 2000s, with named schedules such as the Uberman (nothing to do with Huberman, they joke), the Everyman and the triphasic schedule, promoted online by groups like the Polyphasic Society. The earliest description he found is a 1943 Time Magazine account of Buckminster Fuller's Dymaxion sleep schedule. A review by Harvard researchers found no evidence that these schedules improve mood, cognition, productivity or health. Instead, they cut total sleep, produced poor-quality sleep, reduced REM, and impaired cognition, decision-making, mood and glucose regulation.

Walker's philosophy is "do as you wish, but know what you're doing," as long as you're not hurting others. And that is where drowsy driving comes in. In a simulator study, people getting less than six hours of sleep for several nights had a 30% higher risk of crashing. AAA data suggest that at five hours the risk is two to three times higher, and at four hours close to ten times higher, so risk rises exponentially, not linearly. He asks listeners to imagine choosing between two taxis, one with a 30% higher crash risk, for their family.


Conclusion

This episode gives a clear framework for how to structure sleep across a lifetime. Babies need many sleeps rich in REM to wire their brains, children gradually consolidate into one night, and adults do best with a solid night, optionally plus a short afternoon nap. For good sleepers, naps boost learning, mood and alertness when kept to about 20 minutes and taken before about 3 p.m.; people with insomnia should skip them, and older adults' naps may be a warning sign of poor night sleep. Caffeine masks adenosine rather than clearing it, so timing and dose matter, even as coffee itself carries real health benefits. Polyphasic schedules have little evidence behind them, and short sleep carries real risks on the road. Throughout, Walker and Huberman offer experiments to test on yourself rather than rigid rules, which makes the full conversation well worth hearing.


This is a summary of "Dr Matt Walker How To Structure Your Sleep Use Naps Time Caffeine" from Huberman Lab, giving you the GIST so you can decide if you want the REST. Listen to the full episode.

Get new Huberman Lab summaries by email

The GIST of every new Huberman Lab summary in your inbox, daily or weekly. Free, and you can unsubscribe any time.

Want the REST?

This is an independent summary. Listen to the full conversation from Huberman Lab.

Full episode

More from Huberman Lab

Report a problem

Something wrong on this page? A mistake, a broken link, audio that won't play? Tell us and we'll fix it.