Essentials: The Science of Learning & Speaking Languages | Dr. Eddie Chang

Welcome to Huberman Lab Essentials, where we surface the most useful science for mental and physical performance. I’m Andrew Huberman from Stanford, and today I’m joined by Dr. Eddie Chang to explore how the brain manages speech and language, what separates them, and what’s really represented in those circuits.

Speech is the motor signal you hear when I shape breath with my vocal tract, while language is the broader system of meaning, grammar, and intent that you derive from those sounds. Our work focuses on speech production as an acoustic signal, even though language also runs through reading, sign, and other modalities.

For the mechanics, can you separate what the pharynx and larynx do?

Speaking starts by shaping exhaled air as the larynx brings the vocal folds together to vibrate, about 100 to 200 hertz, creating the buzz we hear as voice. The pharynx, tongue, lips, and jaw then sculpt that energy into the consonants and vowels that form words.

What about cries or laughter, are those driven by the same circuits as speech?

Those are vocalizations that still use the larynx but rely on different brain pathways, which is why people with injuries to speech areas can often still moan or laugh, similar to nonhuman primates.

You’ve helped people who are paralyzed translate intended speech into words; how does that work?

Locked-in conditions like brainstem stroke or advanced ALS can leave thinking intact but block movement and speech, so we implant electrodes over speech motor cortex, stream the signals to a computer, and use machine learning with language models to map those patterns to words. In our BRAVO trial, a participant who had used a head‑stick to type for years produced sentences directly from brain activity with a constrained vocabulary and autocorrect, and his delight was obvious even though giggles briefly confused the decoder.

There’s a lot of talk about brain tech and enhancement; where do you stand on augmentation beyond therapy?

Industry is moving these tools toward products, and while humans have long chased performance boosts, the ethics of cognitive enhancement, access, and surgical risk remain underexplored. Progress will likely be incremental because today’s interfaces tap far less neural bandwidth than our evolved communication systems.

You’re also working on blending decoded speech with facial expressions; what’s the goal?

Seeing a face improves comprehension and makes conversation feel natural, so we’re building avatars that mirror a person’s mouth movements and expressions from neural signals to restore richer communication as more social life moves online.

So one day my posts could be spoken by an animated version of me?

Yes, that’s coming soon, and that embodied feedback should also help people learn to control a speech neuroprosthetic more quickly.

Listeners ask about stuttering; what’s going on under the hood?

Stuttering is a speech‑motor coordination issue, not a language deficit, where the finely timed control of larynx, tongue, lips, and jaw can break down, and anxiety can trigger or worsen it without being the root cause.

And what tends to ease it?

Speech therapy focuses on ways to initiate and smooth articulation, often alongside managing anxiety, and adjusting auditory feedback can shift fluency by nudging the loop between what you say and what you hear.

From our first days as friends to now, it’s a joy to sit here with you; your team’s science and clinical impact, from spinal cord to speech restoration, is extraordinary—thank you for sharing it with us.