For people who speak two languages, the seamless switch between "Earth" and "tierra" feels intuitive — but what actually happens in the brain during that translation? Researchers at Rice University and Baylor College of Medicine in Houston have uncovered something remarkable: the brain maintains a shared geometric map that keeps languages organized and accessible, no matter which one you're using.
The study, published in the journal Cell, found that bilingual brains don't rely on identical neurons for each language. Instead, they preserve the relationships between concepts — like how "dog" relates to "animal" — across languages, while allowing each language to use its own neural "readout axes." The researchers call this phenomenon "shared semantic geometry."
"Our findings suggest that the brain may store meaning in a language-independent format," said Dr. Sameer Sheth, professor of neurosurgery at Baylor College of Medicine and co-senior author of the study. "Different languages appear to access a shared conceptual map rather than creating entirely separate representations of the world."
To arrive at this conclusion, the team studied four fully bilingual English-Spanish speakers who were already undergoing neurosurgical procedures for epilepsy treatment at Houston Methodist Hospital. Using ultra-high-resolution recording technologies — including microelectrodes and Neuropixels probes — the researchers measured the activity of hundreds of individual neurons in the hippocampus as participants listened to stories, read phrases aloud, and engaged in spontaneous conversations in both languages.
The findings challenge the idea that translation relies on specialized "dictionary neurons" that map directly from one language to another. "Our results show that bilingual meaning is an emergent property of neural populations," said lead author Xinyuan Yan, a postdoctoral scholar. "The brain does not appear to rely on one-to-one translation cells. Instead, it preserves patterns of relationships among concepts across languages."
Perhaps most striking, the researchers compared their human brain data to multilingual artificial intelligence language models, including multilingual BERT, and found striking similarities in how meaning is organized. "Large language models and the human brain may be converging on similar computational solutions for representing meaning," said Benjamin Hayden, co-senior author and adjunct professor at Rice. "That does not mean AI works exactly like the brain, but it suggests there may be universal principles for organizing knowledge."
The discovery opens new doors for developing brain-computer interfaces, language rehabilitation therapies, and AI systems that communicate more naturally with humans. For the more than half of the world's population that speaks multiple languages, the research validates what bilingual speakers have long suspected: their brain isn't managing two separate worlds — it's holding a single, elegant map of meaning that every language can read from.