Move to learn: exercise, dance and balance as learning boosters

The Short Version

Across four conversations, the experts converge on one practical conclusion: moving your body is one of the cheapest ways to help your brain learn, and it works through several separate routes. Computational neuroscientist Terry Sejnowski ties exercise to the brain's energy supply through mitochondria and calls it "the best drug you could ever take" [1]. Andrew Huberman links it to blood flow, possible new neurons in the hippocampus and osteocalcin, a hormone released from bone that supports memory circuits [2]. Novel balance challenges, he argues, go further, releasing the very chemicals that open windows for plasticity [4][5]. And Erich Jarvis, whose lab found that speech circuits evolved out of movement circuits, says dancing helps him think and that anyone who wants to stay cognitively sharp into old age "better be moving" [3]. The catch is that routine exercise mostly maintains the brain rather than rewiring it, so movement works best as an addition to focused study, timed and chosen well, not as a replacement for it [2][5].

Table of contents

What They Said

Terry Sejnowski: exercise recharges the brain's energy supply

Sejnowski, a professor at the Salk Institute who directs its computational neurobiology laboratory, is best known for the math of how the brain learns, but Huberman introduces him partly as someone who "uses physical exercise of a particular type in order to enhance his cognition, that is, his brain's ability to learn information and to come up with new ideas" [1].

His explanation starts with energy. Asked why energy declines with age, a question the late Ben Barres once raised, Sejnowski points to mitochondria, the cell's power plants, which become fewer and less efficient with age and can be harmed by some medications [1]. The good news, as he puts it: "That exercise is the best drug you could ever take," benefiting the heart, brain, immune system and every organ [1]. He practices this daily. He runs on the beach below the Salk Institute and climbs back up a 340-foot cliff, doing intervals of about 10 seconds of all-out running followed by jogging, and explains that brief, controlled stress is good for body and brain [1]. Huberman likens this to his own practice of cognitive velocity, reading or listening slightly faster than his default pace while still retaining information, and Sejnowski compares Huberman's reading intervals to his running intervals [1].

Sejnowski also connects movement to new neurons. His colleague Rusty Gage, who discovered adult neurogenesis in the hippocampus, co-authored work with Sejnowski showing that exercise increases it [1]. And he offers a second, subtler way movement helps learning: his best ideas come while walking or jogging without music or podcasts, so he now records voice memos [1]. Psychologists call this mind-wandering, and his free course with Barbara Oakley, Learning How to Learn, teaches students stuck on a problem to walk away, do the dishes or walk around the block [1]. He pairs that with sleep: "the key is to think about it before you go to sleep," rather than watching TV, because the brain works on it overnight [1].

He adds a principle that frames how movement fits into learning. Cognitive learning is cortical, step by step and slower; procedural learning runs through the basal ganglia and becomes automatic, and you can't learn tennis from a book [1]. The key, he says, is that "it has to be active learning," wrestling with a problem yourself rather than being told the answer [1]. He also notes that new recording methods show even simple tasks engage many brain areas at once; in mice, the visual cortex gets more input from the motor system than from the eyes [1].

Andrew Huberman on memory: blood flow, new neurons and hormones from bone

In his episode on memory, Huberman lays out two routes by which exercise supports the hippocampus, the structure that forms new explicit memories [2].

The first is neurogenesis. Cardiovascular exercise appears to increase the birth of new neurons in the dentate gyrus of the hippocampus, but Huberman is careful about how: "To my knowledge, there's no direct relationship between exercise and stimulating the production of new neurons in the brain It seems that it's the improvements in blood flow that also relate to improvements in things like lymphatic flow" that enhance neurogenesis [2]. He also flags the controversy: "there is a lot of debate as to whether or not there Is much if any neurogenesis in the adult human brain" [2]. Either way, he recommends "the 180 to 200 minutes minimum of cardiovascular exercise" per week, zone 2 effort where you can just barely hold a conversation, as important for other health metrics too [2].

The second route is more surprising: "your bones make hormones" [2]. One of them, osteocalcin, has endocrine effects, and work from Eric Kandel's lab at Columbia and others found that cardiovascular exercise "Creates the release of osteocalcin from the bones that travels to the brain and to subregions of the hippocampus and encourages The electrical activity and the formation and maintenance of connections within the hippocampus" [2]. "Load bearing exercise in particular turns out to be important For inducing the release of osteocalcin," Huberman adds [2]. Kandel himself, in his 90s, swims a half mile to a mile a day, and "he credits that exercise as one of the ways in which he keeps his brain sharp," though Huberman notes "Again, this is anecdote" [2].

Why would the brain listen to the bones? Huberman argues that "the vast majority of brain real estate especially in humans is dedicated to two things," vision and movement, and it has long been hypothesized that the maintenance and even improvement of neural circuitry "depends on our body movements And the signal from the body that our brain is still moving" [2]. He credits John Ratey of Harvard, author of Spark, with the idea, and shares Ratey's example of sea creatures that settle permanently on a rock and then digest much of their own brain [2].

Huberman then adds timing. Wendy Suzuki's lab at NYU found learning and memory improved for at least two hours after exercise [2]. His logic: do moderate exercise within one to three hours before learning to boost blood flow and osteocalcin, but place very intense, adrenaline-spiking exercise after a learning bout [2]. That second point comes from the memory research of James McGaugh and Larry Cahill: adrenaline released late in or right after learning stamps in memories, and a hard run is one way to produce that spike [2]. Chronic stress, by contrast, impairs learning and memory; what matters is the change in adrenaline relative to the previous hour or two [2].

Andrew Huberman on balance: novel movement opens the gate

In two episodes on learning through failure, movement and balance, Huberman makes a claim that goes beyond fitness: "movement and balance actually provide windows or portals into our ability to change our nervous system the way we want, even if those changes are not about learning new movements or learning how to balance" [4]. The motor and balance tools, he says, are "for setting a stage or a condition in your brain where you can go learn other things as well" [4].

The mechanism runs through the vestibular system. The inner ear senses head movement in three planes, pitch (nodding), yaw (shaking your head no) and roll (tilting like a puppy), using tiny calcium stones that roll through semicircular canals [4][5]. When you are off balance and must compensate, errors in this system cause the cerebellum to signal deep brain centers that release dopamine, epinephrine and acetylcholine [4]. These circuits evolved to recalibrate movement when our relationship to gravity changes, because "we can't afford to be falling down all the time," and they "tap right into these chemical pathways and those chemical pathways are the gates to plasticity" [4]. He calls this vestibular route "an kind of an amplifier on plasticity" [4].

These are the same chemicals that make any learning stick. Huberman explains that adult brains change only when acetylcholine, epinephrine and dopamine are released at the right time, marking circuits that are then rewired during sleep [4][5]. Errors trigger epinephrine and acetylcholine, and getting something even a little right brings dopamine [4][5]. "And so errors are the basis for neuroplasticity and for learning" [4]. Balance errors are simply an especially direct, hardwired way to produce that chemistry [4][5].

The benefit spills over. "Plasticity is a state of the brain and nervous system," Huberman says, and after an error-filled session you remain in a heightened learning state, he estimates for at least an hour, when you could read, study a language or even work on something in therapy [5]. The same holds for the episode's learning bouts in general: afterward you remain in a heightened state for learning other things for at least an hour [4].

Crucially, he says the key parameter is novelty [5]. A first-time skydiver comes back flooded with neurochemicals; a veteran of tens of thousands of jumps doesn't [5]. A skilled handstander gets essentially no plasticity from half an hour upside down [5]. What helps is being in a new orientation or slightly unstable, including the sensation of nearly falling, done safely: yoga if you're bad at yoga, gymnastics, a new swim stroke, or learning to surf [5]. Stationary bikes with immersive screens don't count, since there's no real vestibular feedback [5]. He firmly warns against dangerous stunts like parkour videos he sees online and says people without the strength or bone density shouldn't take up inversions [5].

He also offers a thought experiment about aging. Children "tend to move in a lot of different relationships to gravity," while "as we get older we tend to get more linear and more regular about specific kinds of movements" [4]. He wonders whether reduced plasticity in older adults, "which includes me," reflects the fact that the chemicals "aren't being deployed because we're not engaging in certain behaviors" rather than the other way around [4], and suspects the relationship runs both ways [5].

Huberman's caveat: routine exercise maintains, it doesn't rewire

Huberman is clear that exercise alone is not a learning technique. Exercise is wonderful for cardiovascular health, strength and bone density, but simply going through your usual workout will maintain the nervous system, not change it [5]. Behavior changes the brain only when it is different enough in specific ways from what you already do well [5]. He dismisses the idea of muscle memory: "Muscles are dumb. The information for motor patterns is stored in the neurons and their firing patterns" [5]. And he rejects flow as a learning state: "flow is an expression of what we already know how to do. It is not a state for learning" [5]. Pushing to muscular failure on the last rep of a gym set, he adds, is a distinct neuromuscular phenomenon with no resemblance to what neuroplasticity requires [5].

That is why his full recipe for adult learning combines four elements: arrive at the right level of arousal, make errors in short incremental bouts, use novel vestibular-motor challenges as an amplifier, and raise the contingency so the learning matters to you [4][5]. Movement is one of four levers, not the whole machine.

Erich Jarvis: the brain's thinking circuits sit next to its movement circuits

Jarvis, a Rockefeller University neuroscientist who trained in ballet at the Alvin Ailey and Joffrey schools before choosing science, brings an evolutionary argument [3]. His lab found that vocal learning pathways in songbirds, parrots and humans sit embedded within circuits for learning movement, which led to the motor theory of vocal learning origin: speech pathways arose from a duplication of surrounding motor circuits [3]. Work by Ani Patel and others showed that only vocal learning species can learn to dance, and Jarvis calls this discovery "bringing my life full circle" [3]. Huberman's summary: we're speaking with our bodies when we dance [3].

Jarvis challenges a tidy division in science: "we like to think of a separation between movement and action and cognition," he says, framing cognition as perception and production as moving [3]. But he sees them as neighbors. He kept dancing after choosing science: "I thought one day I would stop dancing, but I haven't because I find it fulfilling" [3]. When he slowed down, as during the pandemic, he noticed changes in muscle tone [3]. His conclusion is that "by dancing, it is helping me think," and because speech circuits sit next to movement circuits, "if you want to stay cognitively intact until your old age, you better be moving" [3]. He recommends consistent dancing, walking or running, plus practicing speech, oratory or singing, to keep cognitive circuits in tune, while being clear that this conviction comes from personal experience [3].

Huberman adds his own version: long slow runs help him prepare for long solo recordings. "And without those long slow runs, at least the day before or even the morning of, I don't think I could do it, at least not as well" [3]. Jarvis notes other evidence of the link between movement and communication: singing or music helps Parkinson's patients move and speak, likely because singing is the more ancestral use of these circuits [3].

Where They Agree, and Where They Don't

They agree movement helps the brain learn and stay sharp. Sejnowski calls exercise the best drug available and uses intervals daily [1]; Huberman links it to hippocampal function through blood flow and osteocalcin [2] and to plasticity through balance [4][5]; Jarvis dances and runs to keep thinking sharp [3]. Kandel's daily swim [2], Sejnowski's cliff runs [1], Jarvis's dancing [3] and Huberman's long runs [3] are all the same habit, practiced by people who think for a living.

They agree the mechanisms are multiple. Energy supply through mitochondria [1], new neurons in the hippocampus [1][2], blood and lymphatic flow [2], bone-derived osteocalcin [2], the cerebellum's release of dopamine, epinephrine and acetylcholine [4][5], and the shared wiring of movement and speech [3] are separate routes, which is part of why movement is such a good bargain.

They differ on neurogenesis. Sejnowski's summary presents exercise increasing hippocampal neurogenesis as a finding he co-authored with Rusty Gage [1]. Huberman is more cautious, saying there is a lot of debate about whether adult humans make much if any new neurons, and that exercise likely acts indirectly through blood flow [2]. The sources agree the effect exists in some form but disagree on how large it is in adult humans. Huberman's point is that exercise is worth doing regardless, because osteocalcin and other routes don't depend on neurogenesis [2].

They differ on whether routine exercise itself drives learning. Jarvis's view that consistent movement keeps cognitive circuits in tune [3] and Sejnowski's daily intervals [1] suggest regular exercise helps cognition broadly. Huberman agrees exercise maintains the brain and supports memory systems [2], but insists that going through your familiar workout will not change the nervous system; novelty and errors are required for actual rewiring [5]. The difference is between supporting the hardware (energy, blood flow, hormones) and opening a plasticity window (novel balance errors), and the sources are compatible once that distinction is made.

They differ in the evidence behind each claim. Huberman cites specific labs and experiments for osteocalcin, adrenaline timing and post-exercise learning [2]. Jarvis says plainly his dancing advice comes from personal experience [3]. Huberman's balance-and-aging idea is framed as a question he wonders about [4][5], and Kandel's swim is labeled anecdote [2].

Why It Works

Energy. Sejnowski points to mitochondria, which decline in number and efficiency with age, and to exercise as the remedy; brief, controlled stress such as 10-second sprints is good for body and brain [1]. Synapses are energetically expensive, which is why the brain prunes them [1], so an energy boost is a direct aid to a learning brain.

Blood flow and the hippocampus. Cardiovascular exercise improves blood and lymphatic flow, which appears to support new neurons in the dentate gyrus, though their extent in adult humans is debated [2].

Hormones from bone. Cardio and especially load-bearing exercise release osteocalcin, which travels to the hippocampus and supports its electrical activity and connections [2]. Huberman's interpretation is that this signal tells the brain the body is still moving, which matters because so much brain real estate is devoted to movement [2].

Adrenaline timing. Adrenaline released late in or right after learning stamps in memories; intense exercise after a study session is one way to produce that spike [2]. Moderate exercise before learning improves the state you learn in, with benefits lasting at least two hours in Suzuki's work [2].

Balance errors and the plasticity gate. Being off balance produces vestibular errors; the cerebellum responds by signaling deep brain centers that release dopamine, epinephrine and acetylcholine, the chemical gates of plasticity [4][5]. Novelty is essential because familiar movements no longer produce errors [5]. The resulting brain state persists for about an hour and can be used to learn unrelated material [4][5].

Shared circuits. Speech and language pathways evolved from, and remain embedded in, motor-learning circuits, which Jarvis sees as the reason movement, singing, dancing and cognition are intertwined [3]. Even simple tasks recruit many brain areas, including motor input into visual cortex [1].

Mind-wandering. Low-focus walking or jogging without audio gives the brain space to work on problems; Sejnowski gets his best ideas that way [1].

Where It Applies, and Where It Doesn't

Movement as a learning aid applies to almost anyone who studies or works with their mind: students, adults learning new skills, people learning a language, and older adults who want to stay sharp [1][2][3][5]. Huberman stresses the balance-driven plasticity state helps with language, math and emotional learning, not only motor skills [4].

It doesn't replace the learning itself. Routine workouts maintain rather than rewire [5]; flow and muscular failure aren't learning states [5]; and the heightened state after novel balance work has to be used for focused, error-filled study to pay off [4][5]. Novelty wears off: a skill you've mastered, like a practiced handstand or your hundredth skydive, no longer opens the window [5]. Virtual movement without real vestibular feedback doesn't count [5].

Timing matters. Intense, adrenaline-heavy exercise before learning may be less useful than after it, while moderate exercise before learning helps [2]. Chronic stress, including constantly spiking yourself, blunts the benefit [2].

Safety limits apply. Huberman warns against dangerous stunts and against inversions for people without adequate strength or bone density [5]. Some of the evidence is personal or anecdotal, including Jarvis's dancing [3] and Kandel's swimming [2], and the extent of adult human neurogenesis is contested [2].

If you have heart, bone or balance problems, check with a doctor or physical therapist before adding intense intervals, load-bearing work or balance challenges.

What To Do With It

1. Build a cardio base. Aim for at least 180 to 200 minutes a week of zone 2 cardio, steady enough that you can just barely hold a conversation [2].

2. Add load-bearing exercise. Include weight-bearing or resistance work to trigger osteocalcin release from bone, which supports the hippocampus [2].

3. Try short, intense intervals. Follow Sejnowski's pattern of roughly 10 seconds all-out followed by jogging, as brief, controlled stress for body and brain [1].

4. Time exercise around study. Do moderate exercise one to three hours before learning, and save intense, adrenaline-spiking effort, like a hard run or cold water, for right after a learning bout [2].

5. Add something new and slightly unstable. Pick a safe balance challenge you're bad at, such as yoga as a beginner, a new swim stroke, gymnastics or surfing, and do it before a study session to open the plasticity window [5].

6. Use the hour after. Follow a novel balance or error-filled practice session with focused study, reading or language practice within the next hour, while the brain is primed [4][5].

7. Walk without audio when stuck. Take a walk or jog without music or podcasts to let your mind wander on a problem, and capture ideas with voice memos [1].

8. Dance, sing or move to rhythm. Take Jarvis's advice to keep dancing, walking or running consistently, and practice speaking or singing, to keep neighboring movement and cognitive circuits in use [3].

9. Vary your movement as you age. Break linear routines like the same treadmill or the same walk; move in different relationships to gravity, safely, the way children do [4][5].

10. Think about the problem before bed. Pair daytime movement with Sejnowski's habit of reviewing a hard problem before sleep instead of watching TV, since sleep is when rewiring happens [1][4].

Sources

1. Dr Terry Sejnowski How To Improve At Learning Using Neuroscience Ai · Huberman Lab
2. Understand & Improve Memory Using Science-Based Tools · Huberman Lab
3. Dr. Erich Jarvis: The Neuroscience of Speech, Language & Music · Huberman Lab
4. How to Learn Faster by Using Failures, Movement & Balance | Huberman Lab Essentials · Huberman Lab
5. Learn Faster Using Failures, Movement & Balance · Huberman Lab