The brain blueprint spoken words study, published on 19 November in the journal Science, has found that a small region in Wernicke’s area, long thought to handle only the understanding of speech, also prepares the words we are about to say. The discovery rewrites a model of how the brain handles language that has stood for about 150 years.

Researchers at UC San Francisco, led by neurosurgeon Edward Chang, recorded brain activity in 28 people before they had surgery for epilepsy or tumours. They found that neurons in the left posterior superior temporal gyrus, or lpSTG, a small spot inside Wernicke’s area just above the left ear, became active about a quarter of a second before a person began saying a word. The activity was not assessing the word’s meaning; it was encoding the sounds needed to say it, more strongly for a longer word such as “caterpillar” than for a short one such as “cat”.

Brain blueprint spoken words study: what the researchers found

“Every word that you have said today began in that small part above your left ear about a quarter of a second before you said it,” Chang said in a news briefing. The team also tested the region’s role directly: delivering a small electrical pulse to the lpSTG just before a person spoke disrupted their speech, while stimulating nearby areas did not. That points to the region being necessary for producing speech, not merely active at the same time.

The finding answers a question open since 1874, when German neurologist Carl Wernicke described the area as necessary for making sense of speech. People with damage there struggled to understand words, but some also struggled to produce them, with sounds swapped and words replaced by near misses, and nobody could explain why damage to a “hearing” area would break the ability to speak. The new study’s answer is that the same small patch builds a blueprint for the next word before it is spoken.

The work has practical uses. Knowing exactly which patch of cortex prepares speech could help surgeons map the brain more precisely to protect language during operations. It could also improve brain-computer interfaces that restore communication to people who have lost the ability to speak after a brain injury, by reading the intended word earlier and more accurately. The study’s lead author is UCSF bioengineer Alexander B. Silva, and the paper, “A neural code for generative speech production in human Wernicke’s area”, is published in Science (DOI: 10.1126/science.aef8963). The results were also presented on 15 November at the annual meeting of the Society for Neuroscience.

Jeffrey Binder, a neurologist and language researcher at the Medical College of Wisconsin, told Science News the study provides convincing evidence that this brain spot is critical for speech production.

For more science, read about a cancer vaccine that extended survival in a leukaemia trial and Voyager 1 reaching one light-day from Earth.

The study was covered by Medical Xpress and Science News.