How to Learn Physical Skills Faster: Repetitions, Errors, Idle Time and Visualization

High-Level Overview

In this solo episode, part of a month on physical performance, Andrew Huberman explains the neuroscience of learning motor skills, such as a golf swing, a tennis serve, dance, a musical instrument or a running stride, and turns it into a step-by-step protocol. He first clears up a mix-up from earlier episodes (cooling the palms versus heating the body) and shares a quick fix for side stitches. He then covers how movement is generated in the nervous system, why repetitions matter more than hours, why errors open the door to neuroplasticity, what to do immediately after practice, how to advance with focused attention, slow motion and metronomes, a cerebellum-based trick for range of motion, how well visualization actually works, and whether any supplements help.

The tone is teaching-focused and practical, with study results, simple neuroanatomy and self-deprecating asides about Huberman's lack of tennis, golf and basketball skills. The episode builds steadily from mechanism to a layered, stage-by-stage training protocol.

The Discussion Explores:

• Clarifying cooling, heat and caffeine, plus the double-inhale fix for the side stitch
• Open-loop versus closed-loop skills, and the three parts of any movement skill
• How movement is controlled: central pattern generators, upper motor neurons and lower motor neurons
• Repetitions over hours: the Super Mario effect and the "tube test" of winning
• Errors open plasticity, and why raising dopamine before practice backfires
• The post-practice idle period, when the brain replays the correct movement in reverse
• Advancing skill: focusing on one thing, slow motion, metronomes and a cerebellar trick for range of motion
• Visualization, supplements and session length

Deep Dive into Main Ideas

Housekeeping: Cooling, Heat, Caffeine and the Side Stitch

Huberman first addresses listener confusion. Cooling the palms during exercise (from Craig Heller's lab at Stanford) is a way to cool the core and do more work, such as more pull-ups, dips, bench presses, running or cycling. Heating the body, for example in a sauna, is a separate protocol aimed at growth hormone release. Both are valid, just at separate times, and you should always warm up before exercise. On caffeine: if you regularly consume it (roughly 100 to 300 milligrams, several times a week), having it before training is fine. If you rarely use it, a pre-workout coffee is likely to raise body temperature, change blood flow and hurt performance.

He also explains that the side stitch during running or swimming isn't really a cramp. It is referred pain: the phrenic nerve that controls the diaphragm has a branch to the liver, and shallow breathing can cause pain in your side. The fix is the physiological sigh: a double inhale through the nose followed by a long exhale, repeated two or three times, often while still moving. He does it on runs and mentions unpublished data suggesting it may also help regulate heart rate variability.

What a Skill Is Made Of

Before learning any skill, Huberman says, ask whether it is open loop or closed loop. In open-loop skills you act and then get feedback, like throwing a dart or hitting a tennis serve toward a target. In closed-loop skills you get continuous feedback and can adjust as you go, like a running stride, a swim stroke or keeping a drum rhythm. Any motor skill has three parts: sensory perception (what you see, hear or attend to), the movements themselves, and proprioception, the "sixth sense" of where your limbs are. We only notice proprioception when it's missing, like walking on a leg that has fallen asleep.

Movement comes from three sources. Central pattern generators (CPGs) in the spinal cord run rhythmic, already-learned movements like walking, running, cycling and breathing. In classic experiments, animals without a neocortex could still walk on a treadmill and adjust their speed. Upper motor neurons in the cortex control deliberate movements and anything you are still learning, and they require attention. Lower motor neurons in the spinal cord connect to and fire the muscles. You don't need to control these consciously. Your job is just to answer two questions: open or closed loop, and what to focus on.

Repetitions, Not Hours: The Super Mario Effect

Huberman rejects two myths: Hollywood's instant download of skills, and the "10,000 hours" rule. "It's not about hours it's about repetitions," he says. He describes an online experiment with about 50,000 people learning to order programming commands to move a cursor through a maze. When they made a mistake, half saw "that did not work please try again" and half saw "you just lost five points please continue." Contrary to the popular idea that people work harder to avoid losses, the no-penalty group succeeded 68% of the time versus 52%. The reason was that they made many more attempts per unit of time, while the penalized group gave up sooner.

This matches the "tube test" in mice and rats. Two animals meet in a narrow tube and push until one backs out. Winners tend to keep winning and losers keep losing, even with size, strength and testosterone controlled for. A study in the journal Science found a region in the prefrontal cortex that controls this. Stimulating it turned any animal into a winner, and blocking it turned any animal into a loser. Stimulation didn't make the animals stronger. It made them take more forward steps, more attempts per unit of time. The lesson: when first learning a skill, do as many repetitions per unit of time as you safely can.

Errors Open the Door to Plasticity

Why would repeating something you mostly get wrong help? Huberman explains that errors do two things. They trigger error correction, and they open a window for neuroplasticity. He cites a 2021 paper in Neuron (Norman and colleagues) showing that errors recruit frontal brain areas that direct attention, engaging neuromodulators like dopamine, acetylcholine and epinephrine. Errors also answer the hard question of what to focus on, because they point your attention to what needs to change. When you finally get a repetition right, a spike of dopamine helps lock in that correct pattern. "If you walk away at that point you've made the exact wrong choice," he says, unless the errors are dangerous.

This is why raising your baseline dopamine before practice (he gives the example of a large dose of L-tyrosine) is counterproductive. It may help motivation, but it shrinks the contrast between the reward spike after a correct attempt and the background level. He calls this a "framing effect" and distinguishes it from Carol Dweck's growth mindset, since it's about how learning works, not motivation. He also has a message for coaches: constant corrections from outside can crowd out the learner's own error signals, so leave part of each session for uninterrupted repetitions after a cue has been given.

What to Do Right After Practice: Nothing

Huberman's most distinctive recommendation is what to do after a learning session: nothing. Research on the hippocampus shows that during sleep, the brain replays sequences of neural activity from the day. After motor skill practice, if you don't take in lots of new input, the brain also replays the correct motor sequence and drops the incorrect ones, and oddly, it replays it backward right after practice and forward later during sleep. Why it runs backward is unknown, but it appears important for consolidating the skill.

The protocol is to sit or lie quietly with your eyes closed for anywhere from one to five to ten minutes after practice, with no phone, email or conversation, and then get good sleep. He points to a 2011 review in Neuron by Dayan and Cohen, "Neuroplasticity Subserving Motor Skill Learning," for more detail. Most people jump straight to their phones after jiu-jitsu, yoga or dance class, and he suggests that wastes a valuable learning window.

Advancing: Focus on One Thing, and Uncertainty for Experts

He describes a progression from uncertainty, to skill, to mastery, to virtuosity. Learners try to reduce uncertainty, while virtuosos, like big-wave surfers Laird Hamilton and Luca Padua, seek it out as a chance to perform beyond themselves. Once you're somewhat familiar with a skill, you can start choosing where to put your attention in later sessions. What you focus on matters less than keeping it on one specific thing for the whole session. Research from Claudia Lappe and colleagues (2018) had people learn piano key sequences. Learning rates were the same whether the keys played the correct notes or one constant tone, and much worse when the tones were random. Early learning, in other words, is mostly about the motor commands, not the sensory feedback. Only once you're making fewer errors should you switch focus from attempt to attempt (landing spot, speed, grip, stance).

Super-slow-motion practice can help, but only after some skill is in place, roughly once you're succeeding 25 to 30% of the time. Very slow movement gives inaccurate proprioceptive feedback and is too accurate to generate errors, so the window for plasticity never opens. Huberman jokes that he nails free throws in slow motion every time, except there's no ball. For intermediate and advanced learners, a metronome set slightly faster than your current pace (using free apps, or devices for swimmers and runners) forces more repetitions and more errors, and seems to speed up learning more than the same number of reps without an external pace. It also trains central pattern generators to work at the top of their speed range. He gives examples from speed walking, horses changing gait, and competitive cup stacking.

A Cerebellar Trick for Instant Range of Motion

The cerebellum, or "mini brain," combines visual information about head and eye movement (pitch, yaw and roll), body position and timing, and handles much of the timing in learned movements. Much of our flexibility, Huberman explains, is limited by the nervous system, which stops a muscle from stretching further, rather than by the length of tendons or muscles. Because vision and body movement come together in the cerebellum, widening your visual range can widen your physical range.

The protocol is to measure your range first (for example, arms out in a T pressed back as far as possible). Then, keeping your head and body still, move your eyes to the far left, far right, up and down. Retest, and you typically gain about 5 to 15 degrees of range of motion. It works for the legs too. Huberman uses it as part of his warm-up before exercise or skill practice, especially since he dislikes stretching.

Visualization: Useful, but Not a Replacement

Mental rehearsal does work, he says, but it is not as good as physical practice, despite popular claims. In a 2004 study, subjects imagined moving against heavy resistance for 15 minutes a day, five days a week, for 12 weeks. Imagined training increased finger strength by about 35% and elbow flexion strength by about 13.5%, while actual physical training produced about 53%. Visualization activates the upper motor neurons that issue movement commands, but not the lower motor neurons and CPGs that carry them out, and it lacks the proprioceptive feedback that tells the nervous system something is really happening. To those who claim imagining an experience is the same as having it, he says, "your bubble is made of myths." Visualization is best as an add-on, or a fallback when you can't practice, like on a plane or in a thunderstorm. He also recommends a review in Neuron by Herholz and Zatorre on musical training and its carryover to physical skills.

Supplements, Timing and Session Length

There is no pill that gives you more learning from fewer repetitions, Huberman stresses. The real question is what conditions let you focus and do more reps, whether that means a cup of coffee or avoiding one. He highlights alpha-GPC (alpha-glycerophosphocholine). At about 300 to 600 milligrams before training, it has been associated with greater power output (one study found a 14% increase). At up to 1,200 milligrams a day split into three 400-milligram doses, studies show modest benefits against cognitive decline in older people and improved cognition in Alzheimer's among supplements. It also increases fat oxidation and slightly promotes growth hormone release. For physical learning, he advises taking anything stimulating before training (the opposite of his advice to raise adrenaline after cognitive learning), then taking the idle period and protecting your sleep.

Finally, he says the 90-minute ultradian cycle is not a useful limit for skill practice. Ten minutes of focused, high-repetition work beats two hours of distracted effort, because what matters is "the density of training inside of a session." Progress won't always be steady, he warns, but it will come.

Conclusion

Huberman's message is that physical skills are built from repetitions and errors, not hours or wishful thinking. In early sessions, do as many safe attempts as you can, let your mistakes direct your attention and open the brain to change, don't raise dopamine beforehand, and then rest quietly for a few minutes after practice and sleep well so the brain can replay and lock in the correct pattern. As you improve, focus on one element per session, add slow-motion practice once you're succeeding about a quarter of the time, and use a metronome to push speed. Useful extras include the eye-movement trick for range of motion, visualization as a supplement to real practice, and possibly alpha-GPC for power. The overall lesson is that training density and your own errors are what drive learning.

This is a summary of "How to Learn Skills Faster" from Huberman Lab, giving you the GIST so you can decide if you want the REST. Listen to the full episode.