MIT Scanned 772 Brains, Drew a New Language Map, Found Language Areas Much Smaller Than Imagined
2026-07-05 17:52:23 Source: DeepTech Deep Tech Henan
In our brain, there is a language map. People have long known that there are areas in the frontal and temporal lobes of the left hemisphere dedicated to processing language. But when the brain processes language, other areas also light up, such as the cerebellum, hippocampus, amygdala, and temporal pole. What exactly these areas do has never been fully understood.
Recently, the team of Evelina Fedorenko from the McGovern Institute for Brain Research at MIT published a paper in the Journal of Neuroscience, attempting to draw a complete version of this language map. They recruited 772 healthy adults and had each person lie in an fMRI scanner to perform a language task.
In this task, real sentences and meaningless strings of characters appeared on the screen. By comparing the differences in blood oxygen signals when the brain processed these two types of materials, they could identify the brain areas specifically responsible for language processing.
With over 700 participants, this is a very large sample for brain imaging research. A larger sample also makes the statistical results more reliable, allowing some accidentally activated patterns to be easily mistaken for real brain areas.
Figure | Evelina Fedorenko (Source: Massachusetts Institute of Technology)
The researchers also added a control experiment, in which some participants performed a high-difficulty spatial working memory task in the scanner. A series of positions appeared on the screen and participants had to remember them, then judge whether a new position was in the set they had just seen.
This task has nothing to do with language but is equally demanding. If a brain area responds to both the language task and the working memory task, it may be doing some general-purpose processing. Only areas that respond to language but not to the working memory task are considered truly dedicated to language.
The researchers also designed it so that they first found the strongest language activation areas in each participant's own brain space, and then aligned these individual areas. This approach avoids misalignment due to differences in brain sulci between individuals. It's like everyone's fingerprint is different; averaging ten fingerprints together would blur the fine features.
Through this, the researchers identified 17 language-responsive areas, including not only the known core regions in the left frontal and temporal lobes but also parts of the temporal pole, medial prefrontal cortex, hippocampus, amygdala, and cerebellum.
Among them, 14 areas showed selectivity for language—they responded to the language task but showed almost no response to the high-difficulty working memory task. This suggests that these areas are indeed doing something related to language, not just processing general cognitive tasks.
Although these areas are scattered throughout the brain, they account for a very small proportion of the total gray matter volume—on average only about 22 cubic centimeters, less than 4% of the total gray matter. Language has only appeared on Earth for tens of thousands of years, while brain evolution spans millions of years. Language has not transformed the entire brain into a language processor; it has only embedded a relatively compact set of modules into the existing architecture.
(Source: Journal of Neuroscience)
This study also challenged some brain areas previously thought to be involved in language. For example, the caudate and putamen in the basal ganglia were previously believed to participate in syntactic processing. However, in this study's data, these two areas showed very weak responses to language tasks but very strong responses to non-language tasks.
That is, language use generally involves auxiliary processes such as attention, memory, and cognitive control. Those brain areas claimed to be related to language might actually be handling these auxiliary processes rather than language itself.
Some cerebellar areas previously thought to be involved in language processing were shown in this study to have selective responses to language, meaning the cerebellum is not only involved in motor coordination but some parts of it are indeed doing something related to language.
For the hippocampus and amygdala, they were also selectively activated during language processing. This is not entirely consistent with the traditional view that the hippocampus is only responsible for memory and the amygdala only for emotion. In other words, language, as a complex cognitive function, recruits a wider range of brain resources than the classic model describes.
As for what these newly discovered language areas specifically do, further experiments are needed. For example, what role does the temporal pole play in semantic integration? What function does the medial prefrontal cortex serve in discourse comprehension? Does the hippocampus play a role in encoding and retrieval during language processing, similar to episodic memory?
Based on this research, future studies can measure the response patterns of these areas to different language tasks, see if they are more sensitive to vocabulary, syntax, or semantics, and record their functional connectivity with the core language network to see if they are receiving signals or sending signals upward.
In terms of application potential, these findings may have reference value for the assessment and intervention of language disorders. The neural basis of aphasia, dyslexia, and specific language impairment may involve not only classical language areas but also these newly discovered extended areas.
Specifically, for patients with cerebellar damage, language fluency issues sometimes occur. Previously, this was often attributed to motor coordination difficulties rather than language processing itself. This study provides a new perspective: besides coordinating the vocal organs, certain parts of the cerebellum may directly participate in language processing.
Similarly, some patients with temporal lobe epilepsy sometimes experience word-finding difficulties, which may be related to involvement of the temporal pole. Previous rehabilitation for language disorders focused on the core areas of the frontal and temporal lobes. Whether these newly discovered extended areas should also be included in intervention is a question worth exploring.
References: Related paper: https://doi.org/10.1523/JNEUROSCI.0638-25.2026 https://bcs.mit.edu/directory/evelina-fedorenko
Disclaimer: This article aims to convey the latest information in life sciences and the medical health industry. It does not represent the platform's stance, does not constitute any investment or advice, and is subject to official/company announcements. This article is also not a treatment recommendation. For treatment guidance, please visit a regular hospital.
Layout: Hu Weiwei
Note: The cover/first image was generated with AI assistance.
