CAMBRIDGE, Massachusetts—Aging does not affect every part of the brain in the same way. While systems responsible for working memory, attention, decision-making, and problem-solving often become less efficient over time, the brain’s language network appears to follow a strikingly different trajectory. A new brain-imaging study from researchers at MIT and Boston University suggests that the neural architecture supporting language remains remarkably stable across adulthood, even as a more flexible, general-purpose system involved in executive control shows broad signs of age-related decline. The findings offer a detailed explanation for a familiar pattern of healthy aging: older adults may struggle with certain demanding cognitive tasks, yet retain—and sometimes improve—their ability to understand and use language.
The study compared adults between the ages of 17 and 39 with adults between 41 and 80. Rather than treating aging as a uniform process affecting the entire brain, the researchers examined two functionally distinct systems. The first was the language network, a collection of brain regions specialized for processing words, sentences, meaning, and grammatical structure. The second was the multiple demand network, which spans areas of the frontal and parietal lobes and supports tasks requiring cognitive control, including maintaining information in working memory, shifting attention, making decisions, and solving unfamiliar problems. Previous research has shown that this executive system becomes less synchronized and less active with age. The new work asked whether the language network undergoes comparable changes.
To separate the two systems experimentally, the researchers designed tasks that placed different demands on participants’ brains. In one task, volunteers had to remember the locations of squares arranged in a grid, a challenge that primarily engages spatial working memory and the multiple demand network. In other experiments, participants listened to stories and read sentences, activities that activate the language network. The researchers then analyzed functional magnetic resonance imaging data, which measures changes in blood oxygenation associated with neural activity. This allowed them to compare not only how strongly different brain regions responded, but also the spatial organization, synchronization, and selectivity of the networks in younger and older adults.
The spatial memory experiment produced the expected result. In older participants, the multiple demand network was smaller in its apparent functional extent, less synchronized across its component regions, and weaker in its overall activation. These changes are consistent with the idea that aging reduces the efficiency of a system whose job is to flexibly coordinate information across the brain. The network does not store a growing body of specialized knowledge. Instead, it operates as a limited cognitive resource that can be recruited for many different purposes, from suppressing distractions to manipulating information in working memory. Its versatility may also make it particularly vulnerable to the biological effects of aging.
The language experiments revealed a sharply different pattern. Across the younger and older groups, the language network showed highly similar levels and distributions of activity. Its functional topography—the arrangement of regions involved in language—was preserved, as were its hemispheric preferences, its selectivity for linguistic information, and its internal connectivity. Older adults did not show the reduced synchronization observed in the multiple demand network. Nor did their language regions appear to become less specialized by blending more extensively with the executive-control system. In neural terms, the language network looked much more like a stable specialist than a deteriorating generalist.
The researchers tested the network’s sensitivity to linguistic difficulty in greater detail by presenting participants with unfamiliar words and unusual grammatical constructions. These forms of linguistic disruption typically increase activity in language-processing areas because the brain must work harder to interpret them. Younger adults showed this expected response, and older adults did as well. The magnitude and distribution of their responses indicated that the basic mechanisms used to detect and process linguistic difficulty remain intact in later life. According to the researchers, this similarity suggests that older brains are not simply compensating for a damaged language system by relying on entirely different circuitry. Instead, they appear to be processing language through largely preserved neural pathways.
The results may help explain why language abilities often remain stable, or even improve, as people grow older. Vocabulary and reading knowledge can accumulate over decades, provided individuals continue to encounter and use language. Older adults may have richer semantic knowledge, broader experience with grammatical structures, and a larger store of familiar words than younger adults. In that sense, language may benefit from a rare combination of neural specialization and lifelong learning. The brain system that supports it remains comparatively resilient, while the knowledge it draws upon continues to expand. The researchers liken this process to a language model trained on an ever-growing collection of data: continued exposure can increase the available knowledge base even when other forms of cognitive flexibility become less efficient.
The findings also challenge a simple assumption that healthy aging causes all brain networks to lose their distinct identities. Some neuroscientists had proposed that age-related changes might cause the language network to become less synchronized internally and more entangled with the multiple demand network. Such a shift could have indicated that older adults were using executive-control resources to compensate for declining language-specific processing. The imaging data did not support that scenario. Instead, the networks retained their separate functional profiles, with the specialized language system showing no clear evidence of the broad decline seen in the executive network. This distinction emphasizes that the aging brain is not merely becoming globally weaker; different circuits can follow very different biological and functional paths.
The study does not establish that every aspect of language performance improves with age, nor does it suggest that older adults are protected from all language-related problems. Stroke, dementia, neurodegenerative disease, hearing loss, and other medical conditions can disrupt communication profoundly. The participants were healthy adults, and the experiments measured patterns of brain activity rather than providing a comprehensive assessment of language ability in daily life. Even so, the preservation of language-network organization across a wide adult age range provides an important neural foundation for understanding why conversation, reading, comprehension, and vocabulary can remain strong long after other cognitive skills begin to decline. It also raises a central question for future research: what molecular, cellular, and experiential factors protect specialized networks from aging?
The researchers argue that the answer may lie partly in the different jobs performed by the two systems. The multiple demand network must remain adaptable, rapidly reorganizing itself to meet new challenges and learn unfamiliar tasks. That flexibility may come at a cost, making the network more susceptible to age-related changes in connectivity and activation. The language network, by contrast, is specialized and reinforced through constant use, while its knowledge content grows throughout life. Understanding why these systems age differently could eventually guide interventions designed to preserve executive function without disrupting the specialized abilities that remain robust. For now, the study delivers a hopeful message: although aging can weaken the brain’s flexible problem-solving machinery, the neural foundation of language may remain surprisingly durable.
Subject of Research: The effects of healthy aging on the brain’s language-processing network and multiple demand network.
Article Title: Preserved topography, lateralization, selectivity, and functional connectivity of the language network in older brains
News Publication Date: 24-Aug-2026
References: Nature Communications; research funded by the National Institute on Deafness and Other Communication Disorders, MIT’s McGovern Institute, Simons Center for the Social Brain, Poitras Center for Psychiatric Disorders Research, and Quest for Intelligence.
Keywords: aging brain, language processing, neuroscience, cognitive aging, multiple demand network, executive function, working memory, functional connectivity, brain imaging, neurolinguistics, cognition, gerontology
