As we age, attention fades, working memory weakens, and learning new skills becomes harder. But one cognitive ability appears remarkably resistant to time: language processing. A large-scale study by researchers at MIT and Boston University, published in Nature Communications, found that the brain’s language network functions nearly identically in 20-year-olds and 70-year-olds — even as other neural systems show clear signs of deterioration.

What you need to know

  • Brain scans of participants aged 17 to 80 showed that the “multiple demand network” — responsible for working memory, attention, and novel problem-solving — loses synchronization and activation strength noticeably after age 40.
  • In contrast, language-network activation patterns were virtually identical between young and older adults during reading and listening tasks, with equally sharp responses to unfamiliar words and unusual grammar.
  • The researchers suggest that lifelong accumulation of linguistic data — more books read, more words heard — acts as a continuous training stimulus that keeps language circuits in shape over decades.
One brain region is best preserved with age. Photo.
One brain region is best preserved with age

The brain’s flexible problem-solving system wears down first

The study examined how the brain handles different types of cognitive tasks across the lifespan. The “multiple demand network” is a general-purpose computational system that activates whenever we face something new and mentally demanding — it underpins flexible thinking, working memory, and attention.

In the experiments, participants aged 17 to 80 performed spatial memory tasks such as remembering the positions of squares in a grid. Brain scans revealed that in people over 40, this network operated differently from younger participants: neural connections lost synchronization, activation zones narrowed, and overall response strength dropped noticeably.

Language networks show striking resilience

Neuroscientists had long assumed that the age-related desynchronization of neurons would affect language areas as well. The MIT and Boston University study directly tested this assumption — and overturned it.

Participants read sentences and listened to stories while researchers recorded activity in the brain’s language network. The findings were striking:

  • Brain activation patterns during language tasks were virtually identical in 20-year-old and 70-year-old participants.
  • Responses to unfamiliar words or unusual grammatical constructions remained equally sharp regardless of age.
  • Language zones did not merge with other networks and did not lose their internal synchronization over time.

In other words, an older adult’s brain processes linguistic information in essentially the same way, and with comparable efficiency, as a young student’s brain.

Colorful functional MRI brain scans showing active language processing areas on a dark background
Researchers found common features in the brains of young and older adults

Why language may be protected: data accumulation over speed

The key difference between the aging decision-making network and the resilient language network lies in their specialization. The multiple demand system is a flexible resource spent on adapting to novel situations. Language centers, by contrast, are a highly specialized tool that continuously accumulates a database of experience.

Over a lifetime, people read more books, hear more words, and steadily expand their linguistic repertoire. Evelina Fedorenko, a co-author of the study from MIT’s McGovern Institute for Brain Research, compares this process to training neural networks: the human mind resembles a large language model that trains on an ever-growing volume of data. This constant influx of new linguistic input, the researchers suggest, provides a stimulus that keeps the language system functioning well across decades.

What remains uncertain

The findings demonstrate that brain aging is not a uniform decline but a complex process in which some regions can successfully resist the passage of time. While centers responsible for adapting to the unfamiliar wear down, areas built on accumulated experience remain largely intact.

However, the researchers acknowledge that they have not yet found a precise answer to why the flexible multiple demand network is so vulnerable to aging. Understanding why some neurons become depleted over time while others continue to function effectively throughout life could, they suggest, eventually lead to new methods for protecting cognitive health in old age.