Where maths and language meet - Brain region identified in combining the processing of language and mathematics
en-GBde-DEes-ESfr-FR

Where maths and language meet - Brain region identified in combining the processing of language and mathematics

19/11/2026 University of Tokyo

Researchers, including those at the University of Tokyo, help solve an ongoing debate in the field of neuroscience. Two contradictory positions assume that either maths and language are processed in unique areas of the brain, or they are processed together somewhere. New research says both are correct in some way, as while specialized regions do exist, so does a region newly identified as processing both maths and language.

Unless you’re a mathematician, you’ve probably looked down at a page of maths equations before and thought it’s like looking at a foreign language. But this is not simply a metaphor, and there is undeniably some overlap between maths and language: Both deal with the manipulation, rearrangement and interpretation of symbols, and more things besides. This has led some neuroscientists to believe that maths and language are likely processed by the same region or set of regions in the brain. Conversely, maths is often considered more abstract and independent of language. This and other reasons have led to another school of thought, that they must be processed by entirely different regions with very little overlap.

Associate Professor Tomoya Nakai from the Graduate School of Information Science and Technology at the University of Tokyo and his team are interested in the relation between language and maths, and have even published papers showing overlap of activity relating these domains. But, with access to new analytical tools, the team managed to shed some light on this outstanding issue in neuroscience.

“There has been a long debate about whether language and mathematics share common neural mechanisms or are distinct. Our study shows that both ideas are correct,” said Nakai. “We found some brain areas, including an area called 55b, are shared between language and maths, while other brain areas are either language-specific or math-specific. The shared brain areas showed task-dependent association with language or math-specific areas when we read sentences or calculate arithmetic problems. Our study thus reconciles the opposite views on the independence of these cognitive domains.”

The researchers measured brain activity of 32 participants using fMRI, a type of MRI that tracks blood flow in the brain, while they performed either a language task (reading Japanese sentences) or a mathematics task (solving arithmetic problems). They used a large language model to extract information from sentences and mathematical expressions. They then used this information to build computational models that could predict brain activity during tasks. To investigate whether language and mathematics involve some of the same brain processes, the team tested whether a model trained on language stimuli could predict brain activity during the mathematics task, and vice versa.

The team also examined how different parts of the brain communicate with one another during the two tasks. In particular, they looked at whether the brain regions involved in both language and mathematics changed their connections depending on what participants were doing. They found that, during reading, these shared regions were more closely connected to other areas involved in language processing, while during calculation, they were more closely connected to areas involved in mathematics. This suggests that certain brain regions may help support both activities, while working with different networks depending on the task.

“Previous studies mainly compared brain activity during language and mathematics tasks. We used computational models to compare how the brain represents information in these two domains, testing whether models trained on one task could predict brain activity during the other,” said Nakai. “One important aspect of our study was distinguishing shared neural representations from more general factors, such as visual processing, memory and attention, that could also influence brain activity. To address this, we conducted several control analyses and developed additional models. These analyses confirmed that our findings could not be explained by other mitigating factors, supporting the presence of shared neural representations.”

Nakai and his team initially expected to find shared brain activity in an area called the inferior frontal gyrus, as this was hinted at being relevant to processing of maths and language by some previous studies, but instead they identified area 55b. This may be because this study explored the brain surface in greater detail, distinguishing between nearby regions that earlier studies may have overlooked. Area 55b also has distinct tissue properties and may act as a hub that switches between language- and maths-related brain networks.

“Our findings suggest that people may draw on some of the same abilities when working with language and mathematics, particularly the ability to combine individual elements into structured sequences,” said Nakai. “This raises the possibility that future research could look at things like mathematics teaching strategies. Children could be taught to recognize the underlying structure of mathematical expressions as if they were more linguistic, which could help improve their understanding and prevent common mistakes. As for the next steps of our research, I would like to investigate the relationship between language and logical reasoning. Like mathematics, logical reasoning has long been the subject of debate over whether it relies on the same brain mechanisms as language or involves entirely separate processes.”

Journal:

Tomoya Nakai, Rieko Kubo, and Shinji Nishimoto, “Cross-domain encoding models reveal shared and domain-specific neural representations across language and mathematics”, Science Advances, https://doi.org/10.1126/sciadv.aej5555
Fichiers joints
  • Predicted brain activity. Brain regions associated with language or mathematics, identified by comparing computational models predicted brain activity depending on subjects’ assigned tasks. ©2026 Nakai et al. CC-BY-ND
19/11/2026 University of Tokyo
Regions: Asia, Japan
Keywords: Society, Psychology, Humanities, Education, Linguistics, Applied science, Artificial Intelligence

Disclaimer: AlphaGalileo is not responsible for the accuracy of content posted to AlphaGalileo by contributing institutions or for the use of any information through the AlphaGalileo system.

Témoignages

We have used AlphaGalileo since its foundation but frankly we need it more than ever now to ensure our research news is heard across Europe, Asia and North America. As one of the UK’s leading research universities we want to continue to work with other outstanding researchers in Europe. AlphaGalileo helps us to continue to bring our research story to them and the rest of the world.
Peter Dunn, Director of Press and Media Relations at the University of Warwick
AlphaGalileo has helped us more than double our reach at SciDev.Net. The service has enabled our journalists around the world to reach the mainstream media with articles about the impact of science on people in low- and middle-income countries, leading to big increases in the number of SciDev.Net articles that have been republished.
Ben Deighton, SciDevNet
AlphaGalileo is a great source of global research news. I use it regularly.
Robert Lee Hotz, LA Times

Nous travaillons en étroite collaboration avec...


  • The Research Council of Norway
  • SciDevNet
  • Swiss National Science Foundation
  • iesResearch
Copyright 2026 by DNN Corp Terms Of Use Privacy Statement