Aging does not affect every mental ability in the same way. While remembering information for a few seconds or quickly solving a new problem may become harder, understanding language can remain surprisingly strong.
Many older adults also know more words than they did when they were younger. A lifetime of reading, listening and conversation can continue adding to vocabulary even when some other thinking skills begin to slow.
Researchers from MIT and Boston University have now found a striking difference inside the brain that mirrors this pattern.
Their imaging study showed that a network devoted to language remained stable across age, while another network used for general mental control changed substantially.
The findings were published in Nature Communications. Lead authors Anne Billot and Niharika Jhingan worked with senior authors Evelina Fedorenko at MIT and Swathi Kiran at Boston University.
The researchers compared adults aged 17 to 39 with adults aged 41 to 80. Instead of looking at the brain as one system, they focused on two networks with very different functions.
The first was the language network. This network contains areas that respond strongly when people listen to meaningful speech, read sentences and work out how words fit together.
The second was the multiple demand network, which is recruited for many difficult mental tasks. People rely on it when they focus attention, keep information in working memory, make decisions, solve unfamiliar problems or control competing responses.
Previous research has shown that executive abilities supported by this general-purpose network tend to weaken with age. Scientists have also observed changes in how strongly and consistently its different brain areas communicate.
Language presented a puzzle because everyday performance often remains stable. The researchers wanted to know whether the language network quietly deteriorates anyway or whether its brain activity is genuinely preserved.
To examine the multiple demand network, participants completed a task involving the locations of squares in a grid. Remembering those positions required attention and working memory.
Brain scans showed clear age-related differences. In older adults, the executive network was smaller in its active extent, produced weaker responses and showed less synchronized activity across its regions.
These findings fit with the common experience that mentally juggling new information can require more effort with age. They also agree with decades of research showing gradual declines in several executive functions during healthy aging.
The researchers then examined language. Participants listened to stories and read sentences while their brain activity was recorded.
This time, the age difference largely disappeared. Older adults activated the same language system with a similar strength and pattern to younger adults.
The team did more than test ordinary sentences. They also included unfamiliar words and unusual grammar, which make language harder for the brain to process.
When younger people encounter these difficult parts of language, their language regions normally work harder. Older adults showed the same increase in activity, suggesting that the system still recognized and responded to linguistic difficulty in a similar way.
Another question concerned coordination within the network. Aging can cause some brain networks to become less synchronized, meaning their different regions no longer work together as closely.
The language network showed no clear evidence of this decline. Its regions remained coordinated, and the researchers also found little evidence that language processing had begun relying more heavily on the separate executive-control network.
Why might language be so resilient? One possibility is that it is a highly specialized system strengthened by decades of repeated use.
People encounter language constantly from childhood onward. Unlike some abilities that depend heavily on quickly manipulating unfamiliar information, language also draws on a large store of knowledge accumulated over many years.
Vocabulary is a good example. Older adults have had more time to encounter words, meanings, expressions and sentence patterns, and this knowledge can continue increasing through reading and other experiences.
Fedorenko compared the process with a large language model receiving more training data. The longer people interact with language, the richer the store of information available to the brain may become.
The multiple demand network works differently. It is a flexible mental resource rather than a storehouse for one specialized kind of knowledge, allowing people to tackle many new tasks and rapidly adjust to changing demands.
The study raises the possibility that this flexibility makes the system more sensitive to aging. Specialized networks that repeatedly perform the same broad type of task may be better able to maintain their organization.
That idea remains a hypothesis rather than a proven rule. Researchers will need to examine other specialized brain networks to see whether they show the same resistance to aging.
There are other limitations as well. The study compared people across a broad age range at one point in time rather than following every participant for decades, so future long-term research could provide a clearer picture of how individual brains change.
Preserved brain responses also do not mean that aging has no effect on communication. Hearing difficulties, slower processing and neurological diseases can all interfere with language, and the study focused on healthy aging rather than dementia or stroke.
Still, the results challenge the idea that the entire brain simply becomes less capable with age. Some general thinking systems clearly change, while the core machinery used to understand language may remain remarkably robust.
The study’s strongest message is therefore not that aging leaves the brain unchanged. It is that aging is selective, and a lifetime of language experience may be supported by a neural system that holds up far better than some of the brain’s more flexible mental machinery.
