Most people will learn one or two languages in their lives. But Vaughn Smith, a 47-year-old carpet cleaner from Washington, D.C., speaks 24. Smith is a hyperpolyglot—a rare individual who speaks more than 10 languages.

In a new brain imaging study, researchers peered inside the minds of polyglots like Smith to tease out how language-specific regions in their brains respond to hearing different languages. Familiar languages elicited a stronger reaction than unfamiliar ones, they found, with one important exception: native languages, which provoked relatively little brain activity. This, the authors note, suggests there’s something special about the languages we learn early in life.

This study “contributes to our understanding of how our brain learns new things,” says Augusto Buchweitz, a cognitive neuroscientist at the University of Connecticut, Storrs, who was not involved in the work. “The earlier you learn something, the more your brain [adapts] and probably uses less resources.”

Scientists have largely ignored what’s going on inside the brains of polyglots—people who speak more than five languages—says Ev Fedorenko, a cognitive neuroscientist at the Massachusetts Institute of Technology who led the new study. “There’s oodles of work on individuals whose language systems are not functioning properly,” she says, but almost none on people with advanced language skills. That’s partly because they account for only 1% of people globally, making it difficult to find enough participants for research.

But studying this group can help linguists understand the human “language network,” a set of specialized brain areas located in the left frontal and temporal lobes. These areas help humans with the most basic aspect of understanding language: connecting sounds with meaning, Fedorenko says.

To find out how the brain processes five or more languages, Fedorenko teamed up with Saima Malik-Moraleda—a graduate student at Harvard University and a polyglot herself—and a team of other researchers. They scanned the brains of 25 polyglots, 16 of whom were hyperpolyglots, including one who spoke more than 50 different languages. They used a brain imaging technique called functional magnetic resonance imaging (fMRI), which measures blood flow in the brain, to map out these language networks.

Inside the fMRI machine, the polyglots listened to a series of 16-second-long recordings in one of eight different languages. Each recording was selected from a random chunk of the Bible or Alice’s Adventures in Wonderland, which they or other groups had previously translated into 25 and 46 languages, respectively. The eight languages included each participant’s native language, three others they learned later in life, and four unfamiliar languages. Two of the unfamiliar languages were closely related to the participant’s native language—for instance, Spanish for a native Italian speaker. The other two unfamiliar languages came from unrelated language families.

The researchers found that when participants heard any of the nine languages, blood always rushed to the same brain regions. Instead of using different parts of the brain, the participants’ brains appeared to use the same basic network as monolinguals to try to make sense of the sounds, regardless of which language they heard.

The activity in the brain’s language networks fluctuated based on how well participants understood a language. The more familiar the language, the larger the response. Brain activity particularly revved up when participants heard unfamiliar languages that were closely related to ones they knew well. This might have happened as brain areas worked overtime to puzzle out the meanings based on similarities between the languages.

There was one exception to the rule: When participants heard their native tongue, their language networks were actually quieter than when they heard other familiar languages, the researchers reported last month in a preprint uploaded to the server bioRxiv, which hasn’t been peer reviewed. This trend held even when participants were fluent in their other familiar languages, suggesting less brain power is needed to process languages learned early in life.

That could be because expertise reduces the amount of brain power needed for a task, the researchers note. Previous studies have shown similar results in birders and used car salespeople asked to talk about familiar and unfamiliar topics. “When you become a specialist at something, you use fewer resources,” Malik-Moraleda says. The study suggests reaching peak cognitive efficiency may be more likely when it’s learned at a young age.

No previous study has studied so many polyglots. “It’s a very significant piece of work about polyglots,” Buchweitz says. But because the results are purely descriptive, any conclusions about the work are still tentative.

Many polyglots and hyperpolyglots deny any talent for language learning, Fedorenko says. Still, she wants to investigate how polyglots’ brains pull off a trick that so many others find next to impossible, and whether they have an innate talent or just an interest or opportunity. Understanding what it takes for a brain to learn languages, she says, could one day lead to better tools to help people relearn languages more easily after a stroke or brain damage.