Work in 90-minute bouts: using ultradian cycles for focus

The Short Version

Across three separate episodes (his focus toolkit, his deep dive on time perception and his essentials episode on learning), Andrew Huberman gives the same basic protocol for hard mental work: start a timer, work for 90 minutes or less, expect the first stretch to be a warm-up, and then deliberately decompress before you try again [1][2][3]. The reason he gives is biological. Waking life, like sleep, runs in roughly 90-minute ultradian cycles, and after about that long the chemicals behind narrow focus and drive (acetylcholine, dopamine and, to some extent, norepinephrine) drop off, so pushing to two or three hours works against the underlying biology [1][2]. The bouts are taxing, so he recommends one a day to start and no more than two or three, spaced hours apart, with 10 to 30 minutes of low-effort, phone-free recovery after each [1][2]. For learning, the same 90-minute window has a shape: 5 to 10 minutes of drifting, focus arriving around 10 to 15 minutes, about an hour of tunnel-vision work, and a stretch of 7 to 30 minutes of deliberate, frustrating errors that tells the brain to change [3]. Working with the cycle instead of against it means fewer, better hours of deep work, with less strain.

Table of contents

What They Said

All three sources are solo episodes by Andrew Huberman, so this is one expert returning to the same idea from three directions: focus as a trainable skill [1], time and entrainment [2], and adult learning and plasticity [3]. Each episode adds pieces the others leave out.

The focus toolkit: 90 minutes or less, with a warm-up and a cool-down

In his focus toolkit, Huberman frames focus with an "arrow model: epinephrine is the shaft that provides alertness, acetylcholine is the arrowhead that spotlights particular neurons, and dopamine is the engine that lets focus keep going for 10 minutes, an hour or longer [1]. He says there is no focus without epinephrine," but that it is necessary rather than sufficient [1]. That model sets up the question of how long a bout of focus should last.

His answer applies broadly. Whether you are sitting down for mental work, "a focused bout of physical exercise or skill learning or musical learning," a conversation, therapy or a class, "the ideal duration is about 90 minutes not exactly 90 minutes but we can reliably say 90 minutes or less" [1]. He is explicit that longer stretches are not the goal: trying "to drop into two hours of focus or three hours of focus while possible is not really in line with what we know about the underlying biology," because "everything from our sleep states or the different stages of sleep and our waking states is divided into these 90 minute cycles" [1].

The warm-up. Huberman compares focus to exercise: no one should expect to drop straight into deep focus, any more than they would sprint or lift a maximal weight without warming up [1]. His own routine is concrete: "what I like to do is set a timer for 90 minutes," and he accepts that "the first five to 10 minutes of that 90 minutes are a transition time it's like the warm up for focus but I do include it in that 90 minutes" [1]. The exception he notes is when he is "really pressed for a deadline" and "optimally caffeinated" [1].

Flicker is normal. Within the bout, focus is not a steady beam. He asks listeners to accept "the fact that occasionally our focus will flicker it will jump out of focus and then a big part of being able to focus is to go back to focusing" [1]. In the arrow picture, the arrowhead suddenly broadens or shifts somewhere else in the room, and "the key is to be able to shift it back and to narrow it once again and that's an active process" [1]. The summary adds that understanding this takes pressure off, and that repeatedly returning to focus is what trains the circuits through neuroplasticity [1].

The cost. Refocusing is expensive. Huberman says it "requires a lot of metabolic energy," that most of the basal metabolic rate is "related to brain function," and, memorably, "your brain is a glutton with respect to caloric need" [1]. So he warns that "at the end of 90 minutes or maybe even after 45 minutes you might feel rather tired or even exhausted" [1].

The decompression. That is why the bout is followed by recovery: "it's very important that after about a focus that you take at least 10 minutes and ideally as long as 30 minutes" of deliberately low-demand activity, focusing "on somewhat menial tasks or things that really don't require a ton of your concentration" [1]. The phone is the main enemy here. He endorses "this idea of not looking at your phone as you walk down the hall to the bathroom certainly not looking your phone in the bathroom," and jokes that longer restroom lines are not digestive: "chances are it's because people are on their phones in the bathroom" [1]. The goal is to "let your mental states idle to not be focused on any one thing," and he insists "that period of idling is essential for your ability to focus much in the same way that rest between sets of resistance training or rest between exercise is vital" [1]. He adds a sleep analogy: "the ultimate decompress the time in which we are not directing our thinking in our action is during sleep," which he calls "the ultimate form of restoration even if we have very intense dreams" [1].

The recovery doesn't have to be wasted time. He suggests you "go do something for maybe five 10 maybe even 30 minutes that is functional for your day," doing "things that are rather automatic or reflexive for you," while avoiding focused reading and trying "not to bring your vision into a tight location such as your phone" [1]. Doing this "will allow you to drop into intense bouts of focus again repeatedly and repeatedly throughout the day" [1]. And he repeats that the bouts need not run the full length: "those focus bouts don't have to be full 90 minutes," so if 45 minutes of work is done, "set it down" [1].

How many bouts. Asked how many ultradian cycles one can do per day, Huberman says it "depends on how well you slept how well you are nourished" and "how well trained up your focus capacity is" [1]. Then comes what he calls a paradox: "if you are very trained at focusing if you're very good at dropping into focus you're actually going to need more deliberate decompression and defocus" [1]. He recommends "only doing about two maybe three deep work sessions per day," not one after another [1]. He anticipates the objection ("how many people can afford to just work three hours a day") and clarifies that he is "really talking about the hard mental work" [1]. His explanation comes from training: a stronger lifter or an ultramarathoner is "cutting a deeper cut into your recovery capacity," so "the intensity of the work scales directly with how long you need to rest after that work" [1].

For himself, with his phone away and no emergency to tend to, "I can do three 90 minute focus bouts per day but that's about it," with less demanding work in between [1]. His recommendation for everyone else is more modest: "try doing at least one that is one 90 minute or less bout of focused concentrated work per day and yes that means the weekends too" [1]. On weekends, that bout is often reading a book "with my phone nowhere in sight" or listening to an audiobook on a walk while really concentrating [1]. The summary adds that he suggests starting with one a day and adding more after about four weeks [1].

Where the bouts fit in his day. Huberman places his first bout in a fasted state in the morning: "That's my first focus bout of 90 minutes or less. That is essentially done fasted. And then I'll eat" [1]. After eating he notices "a diminished capacity to focus," but "in the afternoon, I will do another 90 minute bout of focus" [1]. His rule for making the afternoon bout work is not to overeat at lunch, so you don't get "so full that you put sheer nervous system into this parasympathetic mode and make it hard to focus in the afternoon" [1]. The summary describes this as a U-shaped relationship with food: fasting brings clarity, but neurons also thrive on glucose, so he uses both ends [1].

Tools to start a bout. Several of the toolkit's other protocols are explicitly aimed at entering a 90-minute block. He usually listens to 40 Hz binaural beats for about five minutes before work [1]. He uses caffeine, most people doing well on 100 to 200 milligrams before focused work, delayed 90 to 120 minutes after waking and cut off around 1 or 2 PM [1]. When focus is lacking, he layers tools: "you could take a three minute very cold shower or submerge yourself in cold water for three minutes. You might have a cup of coffee as well and then sit down and do that work. Maybe even throw in the 40 Hertz" binaural beats, all of which layer the neurochemicals "that are going to lend themselves to a really terrific 90 minute or less workabout" [1]. The water "should be uncomfortably cold but safe to stay in for one to five minutes" [1].

He also offers a visual drill as a formal warm-up. Because "Your visual focus drives your cognitive focus," he recommends fixing your gaze on one spot for 30 seconds to three minutes [1]. He ties it directly to the warm-up idea: "We don't just drop into a focused state the same way we don't drop into the peak performance of a workout. We warm up. So what I recommend is having a 30 second to three minute period at the beginning of about a focus" anchoring your vision to one location [1]. The summary adds that the same drill can be used mid-session when focus drifts, instead of picking up your phone [1].

Time perception: ultradian cycles as entrainment you choose

In the time-perception episode, the 90-minute block appears as the third layer of entrainment, after the yearly (circannual) and daily (circadian) rhythms [2]. Huberman explains that "Ultradian rhythms are rhythms of about 90 minutes or so and all of our existence is broken up into these 90 minute, Ultradian cycles," starting with sleep, where "early in the night you tend to have more slow wave sleep later in the night you tend to have more REM sleep" [2].

He carries that pattern into waking life: whether you "do math or try and learn a language or do physical work of any kind or work out, the 90 minute time block seems to be the one in which the brain can enter a state of focus and alertness and do hard work," and "then about 90 minutes there's a significant drop in your ability to engage in this mental or physical work" [2]. He acknowledges that the self-help, business and pop-psychology literature has all "tried to leverage these Ultradian cycles," and says, "I am one of those people who's also joined that conversation and indeed I use 90 minute work cycles and I think they are extremely powerful" [2].

Where the idea came from. He traces it to the "basic rest activity cycle," proposed by Nathaniel Kleitman and first established within sleep [2]. "Then it was debated for a long time whether or not these 90 minute cycles also control our ability to focus and perform work in wakeful states. And it turns out that they do. Now there's a lot of literature to support that" [2].

What you are entraining to. He is precise about the mechanism: "What you're in training to is the release of particular neurochemicals. In this case, acetylcholine and dopamine that allow your brain to focus for particular periods of time, 90 minutes or so" [2]. After that, "the amount of those chemicals that can be released tends to drop very low, which is why your ability to focus becomes diminished" [2].

You start the clock. The most practically useful point in this episode answers a question he says he always gets: when does the 90-minute cycle begin? If you wake at 8 a.m. at the end of a sleep cycle, does your work cycle start at 8:01? "No [2]. Unlike sleep cycles, which are hardwired (You don't get the option of making your sleep cycles at 120 minutes or five minutes), waking ultradian cycles you can initiate them whenever you want" [2]. "You can set a clock and decide, okay, now the focus begins. Now the work begins" [2]. He does this "Mid-morning and sometimes twice a day," working in a block where "I limit all distraction as much as possible, put away my phone, often turn off the internet as well" [2].

What you can't negotiate. The start time is flexible, but the ceiling is not: "What you can't negotiate, however, is that at about 100 minutes or 120 minutes, no matter who you are, you're going to see a diminishment in performance" [2]. Those circuits, he says, "are far less willing to engage" after about 90 minutes [2]. He also warns against the obvious workaround: "You can't do one 90 minute cycle, then go right into another 90 minute cycle, then another 90 minute cycle," adding, "You can't cheat these circuits related to aceto-coline and dopamine" and norepinephrine [2].

Dosage. Here his advice is slightly more conservative than in the toolkit. "I suggest that people do no more than three and ideally it would be two or just one of these 90 minute cycles," because "they are very taxing. You are in a very narrow tunnel of focus" [2]. He recommends "that they be spaced by at least two to four hours and most people probably won't be able to handle more than two per day," allowing that "there are probably some mutants out there that could do three or four, but that's exceedingly rare" [2]. Even one, he says, "is going to feel like a significant mental investment," and afterwards "you're going to feel pretty taxed" [2]. He is clear about what the blocks are for: "This is not the kind of work that's like checking email or text messaging or social media. This is very focused, hard work" [2].

Matching the block to the time of day. The episode also suggests what to put in which block. Emerging evidence, he says, suggests dopamine and norepinephrine dominate roughly the first half of the day and serotonin rises toward evening, so that "early in the day you are a much more high resolution camera, so to speak, then you are later in the day" [2]. He recommends putting work with rigid rules and right-or-wrong answers (math, following a recipe, executing musical or physical skills, accounting) early, and brainstorming and creative work later [2]. He adds that poor, short or fragmented sleep blurs these daily phases, which may be part of why concentration suffers after bad nights [2]. Circadian entrainment is the base layer underneath: he recommends 10 to 30 minutes of bright light within an hour of waking, light again later in the day, little light at night, and consistent exercise and meal timing [2]. He also notes that accurate time perception "turns out to be one of the most fundamental ways that predicts how well or poorly you perform that thing or task" [2].

The learning episode: the inside shape of a 90-minute bout

The learning episode uses the same cycle but zooms in on what happens inside it when the goal is to change the brain, not just to get work done [3]. Huberman's premise is that "it's a falsehood that everything that we do in experience changes our brain"; change happens when acetylcholine, epinephrine and dopamine are released at the right time, marking circuits "for change and then the change occurs later during sleep" [3].

He describes ultradian rhythms as "these 90 minute rhythms that break up our 24 hour day," which in waking states "break up our day in ways that allow us to learn best within 90 minute cycles" [3]. Then he lays out the timeline:

• Minutes 0 to 10ish: drift. "For the first five to ten minutes of doing that, your mind is going to drift" [3].
• Around 10 to 15 minutes: focus arrives, "provided that you're visually you're restricting your visual world to the just the material in front of you," which "kick in" somewhere "around the 10 or 15 minute mark" [3].
• The core: about an hour. "At best you're probably going to get about an hour of deliberate kind of tunnel vision learning in there" [3].
• The end: flicker. Toward the end of "what is now an hour and 10 or hour and 20 minute cycle," your brain "will start start to flicker in and out" [3].

Within that window, the key move is to reach the edge of your ability and stay there. "You want to keep making errors for this period of time that I'm saying will last anywhere for about seven to 30 minutes it is exceedingly frustrating but that frustration it liberates the chemical cues that signal that plasticity needs to happen" [3]. The payoff comes later: returning "a day or two later in a learning bout after a nap or a night or two of deep rest," you find that "we get a lot of it right whereas we got it wrong before" [3]. He insists this is not "some little hack or trick"; "it's really about trying to cue the nervous system that something needs to change" [3].

When to schedule it. He answers the questions he says are "the right questions": how often and at what time [3]. "Chances are that you can't focus as well at 4 p.m. as you can at 10 a.m.," so "find the time or times of day when you naturally have the highest mental acuity and that's really when you want to engage in these" learning bouts [3]. The summary adds that afterward you remain in a heightened state for learning other things for at least an hour [3].

Arriving in the right state. The learning episode adds a pre-bout check that the other two treat more loosely. Huberman's term "limbic friction" describes the stress of being in the wrong autonomic state, either too anxious or too tired [3]. "The first gate is to arrive at learning at the appropriate level of autonomic arousal clear and focused is best but don't obsess over being right there it's okay be a little anxious or a little bit tired" [3]. If too alert, he suggests the physiological sigh and panoramic vision; if too tired, good sleep or NSDR, or otherwise caffeine or super-oxygenation breathing [3].

Small pieces, real stakes. The episode's other principles shape what goes into the bout. The prism-glasses experiments from Eric Knudsen's lab showed adults could build large changes when the displacement was stepped up in small increments (about 7 degrees, then 14, then 28), leading to his conclusion: "Put simply incremental learning as an adult is absolutely essential" [3]. In practice, that means short, focused learning bouts on small pieces of information rather than trying to absorb a lot at once [3]. The second route is high contingency: "How badly we need or want the plasticity determines how fast that plasticity will arrive" [3]. And telling yourself that errors are good for learning attaches dopamine to them; he describes the ideal as "almost seeking frustration" and finding "some pleasure in the frustration" [3].

Where They Agree, and Where They Don't

Because all three sources are Huberman, the core message is consistent. They disagree only in emphasis and in some of the numbers, and those differences are worth knowing.

They agree on:

• The length: about 90 minutes, as a ceiling. The toolkit says "90 minutes or less" [1], the time episode says performance declines "at about 100 minutes or 120 minutes, no matter who you are" [2], and the learning episode describes a cycle of "an hour and 10 or hour and 20" minutes in which focus finally flickers [3].
• The biology: the ultradian cycle, shared with sleep. All three root the 90-minute block in the same cycle that structures sleep [1][2][3].
• The warm-up. All three say you should not expect instant depth. The first five to 10 minutes are "a transition time" [1], "One should never expect that you're going to drop immediately into a state of high focus" [2], and "for the first five to ten minutes of doing that, your mind is going to drift" [3].
• The chemistry. Acetylcholine, dopamine and epinephrine/norepinephrine are the chemicals behind focus in the toolkit's arrow model [1], the chemicals whose release drops after about 90 minutes [2], and the chemical cocktail that marks circuits for plasticity [3].
• Restricting inputs. A timer, the phone put away, often the internet off [1][2], and a visual world restricted to "just the material in front of you" [3].
• Morning is prime time. His first bout is in the morning [1][2]; precise, rule-bound work belongs early [2]; and you likely focus better at 10 a.m. than at 4 p.m. [3].

Where they differ:

• How many bouts per day. The toolkit recommends "about two maybe three deep work sessions per day," says he himself can do three, and asks everyone to do at least one [1]. The time episode is stricter: "no more than three and ideally it would be two or just one," with "most people probably won't be able to handle more than two per day" [2]. The difference seems to come from the framing. The toolkit treats focus as a skill that can be trained up over weeks, starting with one a day [1], while the time episode focuses on the non-negotiable decline of the chemicals and on the average listener [2]. The two views are compatible: one is the default, and up to three is for the well-trained on a good day.
• How long to rest between bouts. The toolkit prescribes "at least 10 minutes and ideally as long as 30 minutes" of deliberate decompression after each bout [1]. The time episode says bouts should be "spaced by at least two to four hours" [2]. These describe different things: the immediate cool-down versus the gap before the next deep-work block. In his own day, the toolkit's schedule of a morning bout, a meal, then an afternoon bout [1] fits the two-to-four-hour spacing [2], with lighter work filling the gap. Read together, the two sources suggest a short decompression immediately afterward and a longer interval before another bout.
• How much of the 90 minutes is real deep work. The toolkit counts the warm-up inside the 90 minutes and treats the rest as focused work with normal flickers [1]. The learning episode is more sober: focus arrives at 10 to 15 minutes and "at best" you get "about an hour" of tunnel-vision learning [3]. The toolkit also says you may be exhausted "even after 45 minutes" and can stop there [1]. The honest reading is that 90 minutes is an outer limit, not a promise of 90 productive minutes.
• What fills the bout. The toolkit and time episode treat it as a block for hard work in general [1][2]. The learning episode adds a specific agenda: deliberately seek out errors for 7 to 30 minutes [3]. That is an addition for learning, not a contradiction.

Why It Works

The sources give a layered explanation, from rhythm to chemistry to metabolism to plasticity.

A rhythm that runs day and night. Huberman describes a single roughly 90-minute cycle, first established in sleep as Kleitman's basic rest-activity cycle, which after long debate was found to "also control our ability to focus and perform work in wakeful states" [2]. Sleep is divided into these cycles, with more slow-wave sleep early and more REM later [2], and the toolkit argues that pushing to "two hours of focus or three hours" is "not really in line with what we know about the underlying biology" for exactly this reason [1].

The chemicals run down. Focus depends on acetylcholine (the arrowhead), dopamine (the engine) and epinephrine (the shaft) [1]. In the time episode, Huberman says the 90-minute cycles "are linked to the ability of the neurons that release aceto-coline and dopamine and to some extent, noripinephrine, the things that give us narrow focus, motivation and drive" [2], and after about 90 minutes "the amount of those chemicals that can be released tends to drop very low" [2]. That is why back-to-back blocks fail: "You can't cheat these circuits" [2].

It's metabolically expensive. Narrowing focus again and again is "an active process" that "requires a lot of metabolic energy," and the brain is "a glutton with respect to caloric need" [1]. Tiredness after 45 to 90 minutes is expected, not a character flaw [1].

Rest restores, like rest between sets. The recovery period is part of the method. Idling "is essential for your ability to focus much in the same way that rest between sets of resistance training" is vital [1], and sleep is "the ultimate form of restoration" [1]. The more intense the work, the more recovery is needed: "the intensity of the work scales directly with how long you need to rest after that work" [1]. Learning follows the same logic: the chemicals released during a bout mark circuits, and "the change occurs later during sleep" [3], which is why the gains show up "a day or two later" [3].

Warming up follows how focus actually starts. Focus is not "a drop all the way in and then exit type phenomenon" [1]. The warm-up period and the practice of returning to focus after each flicker are, per the summary, what train the circuits through neuroplasticity [1]. Visual anchoring helps because visual focus drives cognitive focus, engaging the prefrontal cortex and frontal eye fields [1], and restricting your visual world is what lets focus "kick in" at 10 to 15 minutes [3].

Errors inside the bout trigger change. In the learning episode, errors cause epinephrine and acetylcholine release, and dopamine arrives as you begin to succeed [3]. That frustrating 7 to 30 minutes "liberates the chemical cues that signal that plasticity needs to happen" [3], so the 90-minute bout is the container that holds the error stretch.

Time of day changes the brain's "resolution." With dopamine and norepinephrine higher earlier, "early in the day you are a much more high resolution camera" [2], which supports putting precise, rule-based bouts in the morning [2] and learning bouts at your sharpest hours [3].

On evidence, the sources describe the waking ultradian claim as supported by "a lot of literature" and "ample literature" [2] but do not name specific studies for it. Other focus tools in the toolkit are tied to named research (the 2020 Experimental Psychology study on acute stress, Dr. Wendy Suzuki's 13-minute meditation study) [1], but the 90-minute block itself rests on Huberman's summary of the literature and his personal practice.

Where It Applies, and Where It Doesn't

It applies to truly hard work. Huberman says the blocks are for "very focused, hard work," "working on hard problems," "not the kind of work that's like checking email or text messaging or social media" [2]. His three-hours-a-day comment refers to "the hard mental work," not your total working hours [1].

It applies well beyond desk work. The toolkit lists physical exercise, skill learning, musical learning, conversations, therapy and classes [1]; the time episode includes math, learning a language, physical work and workouts [2]; the learning episode says the motor and balance tools help set a stage for learning language, math and emotional coupling [3]. Huberman does his physical work early and mental work both early and late [1].

It applies differently depending on your state. How many cycles you can do "depends on how well you slept how well you are nourished" and how trained your focus is [1]. Sleep is the master modulator: aim for quality sleep on about 80 percent of nights, since "there's simply no tool" that overcomes chronic sleep deprivation [1]. Poor sleep also blurs the daily dopamine-serotonin phases [2]. If you arrive too anxious or too tired, the learning episode says you won't reach the starting line until you fix your state [3].

It doesn't mean you must hit 90 minutes. It is a ceiling, not a quota. "It doesn't have to be the full 90 minutes" [1], and 45 minutes is a legitimate bout [1].

It doesn't let you stack cycles. Back-to-back 90-minute blocks don't work, and no one gets past the decline at 100 to 120 minutes [2]. The rare person who can do three or four is, in his word, one of "some mutants" [2].

It doesn't guarantee smooth focus. Flicker within the bout is expected [1], and even Huberman says "We all struggle to varying degrees to achieve focus and motivation and drive within those 90 minute cycles" [2].

It depends on your life circumstances. Huberman's three bouts a day come "provided the landscape of my life is right," with no burning emergency and the phone off [1]. Deadlines change the warm-up [1], and not everyone keeps his schedule; the learning episode notes that everyone's peak depends on "when you're sleeping and you're kind of natural chemistry and rhythms" [3].

Supporting tools are optional and personal. Caffeine, cold exposure, supplements and prescription drugs appear in the toolkit as layers, not requirements. He says no one needs all of them, advises combining no more than two or three at once, and puts "behavioral tools" first, then nutrition and sleep, then supplements, and prescription drugs last [1]. People prone to anxiety or panic should be cautious with caffeine [1]. If you are considering stimulants, supplements, cold exposure or ADHD medication to support focus, talk with a doctor first, especially if you have heart, anxiety or sleep problems.

What To Do With It

1. Pick one daily block and protect it, weekends included. Start with a single bout of "90 minute or less" focused work per day, "and yes that means the weekends too" [1]. Stay at one a day for about four weeks before adding more [1].

2. Put it at your sharpest time, usually mid-morning. Huberman does his mid-morning [2]; you likely focus better "at 10 a.m." than "at 4 p.m." [3]. Give that slot to precise, rule-bound work like math, accounting or technical skill, and save brainstorming for later in the day [2].

3. Start the cycle on purpose. Set a timer for 90 minutes and decide, "now the focus begins" [1][2]. Put the phone away and consider turning off the internet [2]. Choose one small, specific piece of the problem, since incremental learning is "absolutely essential" for adults [3].

4. Check your state before you start. If you are wired, try a physiological sigh (two inhales through the nose, one long exhale through the mouth) or widen your gaze; if you are flat, try NSDR, caffeine (100 to 200 milligrams works for most people, not within 90 to 120 minutes of waking and not after early afternoon) or a three-minute cold shower [1][3]. Aim for "clear and focused," but "don't obsess" [3].

5. Warm up for 30 seconds to three minutes. Fix your gaze on one spot, refocusing when your eyes drift, as a bridge into the work; about five minutes of 40 Hz binaural beats is another option [1]. Count the warm-up inside your 90 minutes and expect the first five to 10 minutes to feel scattered [1][3].

6. Treat drift as reps, not failure. When the arrowhead widens, narrow it again; that act of returning is the training [1]. If you drift mid-bout, repeat the visual drill instead of reaching for your phone [1].

7. If you're learning, spend 7 to 30 minutes in the error zone. Push to the point where you are getting it wrong and "keep making errors," telling yourself the mistakes are what drive the learning [3]. Expect the improvement a day or two later, after sleep [3].

8. Stop at the ceiling, or before it. End by 90 minutes; performance drops at 100 to 120 minutes no matter who you are [2]. If you are spent at 45 minutes, "set it down" [1].

9. Decompress for 10 to 30 minutes with no phone. Do menial, automatic, useful tasks; walk down the hall without your phone; avoid focused reading and anything that pulls your eyes to a small screen [1].

10. Space further bouts by two to four hours and cap them. Don't chain cycles back to back [2]. Most people should stop at two per day, three at most, and even Huberman stops at three [1][2]. Protect sleep, the "ultimate form of restoration" [1], because that is when the changes you set up during the bout are made [3].

Sources

1. Focus Toolkit Tools To Improve Your Focus Concentration · Huberman Lab
2. Time Perception & Entrainment by Dopamine, Serotonin & Hormones · Huberman Lab
3. How to Learn Faster by Using Failures, Movement & Balance | Huberman Lab Essentials · Huberman Lab