Enhance Your Learning Speed & Health Using Neuroscience Based Protocols | Dr. Poppy Crum

Guest
Dr. Poppy Crum
Read time
~5 min
Uploaded
29 Sept 2025
Duration
2:35:50
Views
263.7K
OPEN ON YOUTUBE ↗

KEYPOINTS

  • 01

    Every technology you use is reshaping your brain maps

    Crum, a Stanford neuroscientist and former chief scientist at Dolby Laboratories, argues we are more plastic than we realize, and that the statistics of our environments and tools allocate our limited neurons. As we age, that shaping depends increasingly on how attentive and engaged we are rather than on passive consumption.

  • 02

    The textbook homunculus is out of date

    Wilder Penfield's 1940s map came from stimulating the cortex of awake epilepsy patients and asking what they felt. Crum asks students what is wrong with it today: thumbs should be bigger because of phones, and ankle representation may shift as driving changes. The map reflects how the brain allocates resources for success in a given era. Texting did not build a new brain area; it reallocated existing ones.

  • 03

    Your environment sets your sensory thresholds

    Cities have sonic imprints from their vehicles, density and construction, so hearing thresholds and sensitivities can partly predict where someone lives or grew up. Sounds that mattered in childhood, such as a neighbor's rooster, become deeply embedded.

  • 04

    Perception is sensory input filtered through priors

    We combine sensory data, accumulated statistics called Bayesian priors, and expectations, so people experience the same physical information differently. Crum's two-and-a-half-year-old daughter, seeing NASA's image of Earth in a Smithsonian planetarium, shouted 'Minions!' because her prior for that image was the Universal logo. Speed at probabilistic inference is a strong predictor of success in many environments.

  • 05

    Closed-loop feedback accelerates skill learning

    Video games give real-time performance data in a controlled, closed loop, which is why Crum teaches a Stanford course on neuroplasticity and video gaming to musicians, computer scientists and athletes. Real-time feedback, including auditory feedback, sharpens neural representations so you can detect finer differences.

  • 06

    Ask whether a tool makes you smarter or just faster

    Crum sorts tool use into two bins: using technology to gain insight and become cognitively or physically more effective, or using it to replace a cognitive skill for speed. GPS is a replacing tool. London taxi drivers once had more hippocampal gray matter from mental maps; without GPS today, do not expect the same navigation ability. Huberman's worry is that time saved gets filled with things that are neutral to harmful.

  • 07

    Writing with LLMs cuts the load that builds learning

    Cognitive load theory splits effort into intrinsic load from difficult material, extraneous load from poor presentation or a noisy setting, and germane load spent building mental schemas. Germane load is the work of learning. In an MIT EEG study comparing essays written with brain only, search engines or LLMs, the LLM condition reduced engagement and transfer. Crum predicts professionals who learned their field before LLMs will be better at extrapolation and spotting patterns.

  • 08

    Your body broadcasts your state

    Breath carries a shifting mix of CO2, acetone and isoprene that changes with stress or suspense. Pupil size tracks arousal and cognitive load and is now measurable with open-source software on laptops or phones. Huberman notes that changes in lighting can be corrected for by dividing by ambient light measured at the eye. Crum expects smart glasses and environmental sensors to read these signals and adjust rooms to support focus or relaxation.

  • 09

    Plasticity is as fast as the incentive requires

    Eric Knudsen raised owls in prism glasses shifting vision by 15 degrees. Their auditory neurons realigned to form a second map, and much faster when hunting and feeding depended on it. Huberman's takeaway: habits do not take a fixed 29 or 50 days; they form as quickly as necessary, within limits.

EXAMPLES

  • 01

    40 hours of Call of Duty

    Self-identified non-gamers who played 40 hours of Call of Duty developed gamer-level contrast sensitivity that persisted a year later. Work by Alex Pouget and Daphne Bavelier found trained players made the same probabilistic decisions as non-players, but much faster.

  • 02

    A homemade swim-stroke analyzer

    Crum used AI coding tools such as Perplexity Labs and Replit to build a phone-camera computer vision app that tracks her 10-year-old daughter's roll, arm entry, velocity and stroke consistency. She stresses good coaches come first and that she uses the data to support their cues, not lecture her child.

  • 03

    Crum's two absolute-pitch maps

    Crum has absolute pitch and can name frequencies such as 442 Hz. Playing Baroque violin tuned to A at 415 Hz, close to modern G-sharp, she heard every A as wrong until she developed a second absolute-pitch map, letting her switch between A415 and A440 with nothing in between. Reading Knudsen's owl paper led her into neuroscience.

  • 04

    Breath CO2 tracking Free Solo

    At Dolby, CO2 measured from tubes on the floor of a theater showing Free Solo revealed where Alex Honnold summited, abandoned a climb or hurt his ankle, without annotation. Another study found audiences' exhaled chemistry marked when Katniss's dress catches fire in The Hunger Games.

  • 05

    The moth that beats the bat

    Echolocating bats adjust constant-frequency calls so echoes return in their most sensitive range, tracking Doppler shifts for speed, and use FM sweeps for surface detail. Moths escape about 80% of the time: a few neurons trigger random flight, then a drop to the ground as calls get close, while body structures deflect call energy. Orb-weaver spiders tune webs to resonate; when Crum sings near 880 Hz, the spider reacts as if to a threat.

OTHER RELEVANT

  • 01

    Use AI to quiz yourself, not to think for you

    Huberman feeds verbatim text from papers he has already read and annotated into AI, asks it to test him, and asks where he is weak, such as third-order conceptual links. Self-testing away from the material is what drives memory. Crum built a free, no-coding protocol for creating a custom AI coach for any skill or health routine.

  • 02

    Quantify to motivate, then go beyond summary numbers

    Sleep scores and step counts work because they give something to aspire to. Huberman notes 7,000 steps is now considered the threshold, not 10,000. Crum prefers richer feedback on specific components of a skill. Scaring people stops behaviors; making the right ones engaging and measurable starts them.

  • 03

    Choose which maps you let technology change

    Smartphones took about 15 years to integrate and we are still cleaning up harms such as late-night light and constant contact. With AI arriving faster, be deliberate about which skills you hand off. Berkeley students told Crum many peers cannot start a paper without ChatGPT.