Speech sounds are fuzzy: Here’s how your brain sorts them

Speech gets slurred unless you know the language; scientists have discovered the brain signal that separates words

By Elise Cutts edited by Sarah Lewin Frasier

Join our community of science lovers!

Speech seems to be made of words, but this impression has more to do with what is in our heads than with what comes out of our mouths. In natural language, there are no clear acoustic boundaries separating words; we stop as many times within the words as between them. This is particularly evident when listening an unknown language being spoken: words often seem to “blur” into a single spread out stream of sound.

So how does the brain cut speech into recognizable pieces? Recent research by neurologist and neurosurgeon Edward Chang of the University of California, San Francisco and colleagues reveals a clue. In a study published In Neuronresearchers looked at fast brain waves that flicker about 70 to 150 times per second in a part of the brain involved in speech perception. They realized that the power of these “high gamma” waves systematically drops about 100 milliseconds after a word boundary. Like a blank space in printed text, the pronounced drop marks the end of a word for fluent speakers of the language.

“To my knowledge, this is the first time we have a direct neural correlate of words,” says Chang. “It’s a big deal.”

On supporting science journalism

If you enjoy this article, please consider supporting our award-winning journalism by subscribe. By purchasing a subscription, you are helping to ensure the future of impactful stories about the discoveries and ideas shaping our world today.

In another study, published In Naturethe scientists reported that native speakers of English, Spanish, or Mandarin all showed these elevated gamma responses to their native language, but that listening to foreign speech did not trigger the declines as strongly or consistently. Bilingual people showed native-like patterns in both of their languages, and the brain activity of adult English learners listening to English seemed more native the more proficient they were.

“This is a first big foray into the question” of how the brain marks word boundaries, says Evelina Fedorenko, a neuroscientist at the Massachusetts Institute of Technology, who was not involved in any of this work. She adds, however, that it is not yet clear whether actual understanding of a language is necessary for recognition of hyphenations. Perhaps the brain simply picks up the sound patterns it often hears, regardless of understanding. Or maybe meaning matters, as with muffled speech in a movie that suddenly seems clearer when the subtitles are turned on. Even though speech sounds and higher-level language structures are processed differently in the brain, the two can impact each other. Experiments with artificial language that mimics natural speech sounds could tease out the details, Fedorenko says.

When it comes to deciphering words, Chang suspects there may not be a clear distinction between these different types of processing; The signal he and his colleagues linked to word boundaries occurs in a region of the brain that also recognizes speech sounds. Historically, Chang says, researchers imagined that different levels of language structure, from sounds to words to meaning, would be processed in dedicated brain regions. These new findings, he adds, “kind of explode that. All of that is actually happening in the same place. When we calculate sounds, we calculate words.”