how to sleep better: 6 tactics, from your first hour awake to 3am

the part that surprises people: you want cortisol high in the morning. on the transcript: "You want this peak in cortisol as high as possible in the first 60 minutes of your day." that morning peak is what sets up a low one at night. the tool is light. viewing sunlight in that first hour drives the peak higher, he says.

the tactic: go outside, not near a window. no sun where you live? he says there is always daylight out, even on a grey December morning. "But you want daylight in your eyes in the first hours of your day." short on options: "Go out on your balcony if that's all you can do. If you're in a car, roll down the window." he says not to do it through a window. and if you skip a day: "The next day, get outside for twice as long."

try this week: for seven mornings, step outside within the first hour of waking, even when it is cloudy. if you miss one, go out twice as long the next day. each night, note how fast you fall asleep.

Huberman cites a huge UK Biobank study for one rule: "you want bright days and you want dark nights, and they're additive." so the evening has two halves.

first, more daylight late in the day: "try and get some daylight in your eyes in the late afternoon or evening". a window, a balcony, a walk around the building. he says it buffers your eyes against artificial light later on. "This is your Netflix inoculation, as I call it." one exception he names: it does not seem to work for teenagers.

the tactic: then go dim, then go dark. "Bright lights stimulate cortisol release. You want your cortisol low at night." his reasonable version for screens: "the last 30 minutes before sleep, no screens." once you are asleep, "try and keep your sleeping environment as dark as possible." "A good eye mask is useful."

try this week: take a 10-minute walk outside in the late afternoon. after dinner, switch to lamps, put your phone outside the bedroom 30 minutes before bed, and cover every standby light in the room.

Huberman likes training in the first hours of the day, but says later is fine: "Your cortisol will come back down." you can help it. "Take a hot shower, which paradoxically lowers your core body temperature and then helps you sleep." warm skin tells a thermostat in your brain to cool your core, and you fall asleep better. long exhales after training help too.

the tactic: do not go zero-carb at dinner. "Many people who don't eat enough starches find that they have sleep issues." his reason: "Starches lower cortisol, and you want your cortisol low at night." his example is a bowl of rice or pasta in your final meal of the day. two limits: "you don't want to eat too close to bedtime", and you should not go to bed with gnawing hunger either.

try this week: on days you train after work, take a hot shower before bed. at dinner, add a normal portion of rice, potatoes or pasta, and finish eating well before you go to bed.

some things get worse when you chase them head on. Greenberg's first example, on the transcript: "Trying to meditate well actually tends to undermine your ability to meditate." then: "Sleep is another one." he describes a study where one group was told to try to sleep and the other was told "we want you to try to stay awake even though you're in bed." as he remembers it, the stay-awake group fell asleep faster.

the tactic: take the pressure off. forcing sleep keeps you alert. so flip the goal: lie still in the dark and calmly try to stay awake. Huberman gets to the same place from the body side: "To fall asleep or to fall back asleep, you need to stop thinking about your body position." and the worries that keep you up get their own slot earlier in the day. that is Greenberg's worry time below, and he says never right before bed.

try this week: set a daily 5pm worry time and write tomorrow's worries down then. in bed, if sleep does not come, stop chasing it: eyes closed, still, and see how long you can stay awake.

Bartlett asks about waking at 2 or 3am and starting to worry. Huberman's first move: "10 long, deliberate exhales." then a trick he admits is odd and that "has indirect scientific support". the idea: to drift off, you need to stop sensing your limbs, and a link between your eyes and your cerebellum can help switch that off.

the tactic: run the sequence. eyes closed. "You're going to move your eyes slowly to one side underneath your eyelids, then to the other side", then up, then down, then a slow circle counterclockwise and one clockwise. last, point them gently toward your nostrils without crossing them, and do a long exhale. two or three rounds. he says it generally works, but maybe not if you are very stressed.

Swart's add-on, on the captions: "if you find yourself awake at night and you're not lying on your side turn yourself onto your side".

try this week: the next time you wake in the night, leave the phone where it is. 10 slow exhales, the eye sequence twice, then roll onto your side.

Swart is a neuroscientist and medical doctor. on the captions: "the ideal is eight hours and 15 minutes" in population studies, and "that doesn't mean it's for everyone but for most people that's the ideal". more is not better: "sleeping more than that can be depressogenic", meaning it can lower your mood. the catch: "you ideally need to be in bed for nine hours to get that amount of sleep".

the tactic: set a floor, not just a goal. Wood, a neuroscientist at the University of Washington, on the transcript, about dementia risk: "the cutoff is really around six hours. So if you're consistently sleeping fewer than six hours, that's where risk starts to increase." Huberman adds that needs vary: "some people, like myself, do fine on six and a half or seven hours".

try this week: count back nine hours from your alarm. that is your lights-out time. hold it for five nights and notice whether you start waking before the alarm.