Older, Slower—But Wiser?


by Jim Schnabel

March 17, 2014

Without a doubt the brain changes as it ages from early to late adulthood. The physical changes include alterations in gene expression, as well as reductions in neurons and nerve fibers. Some regions of the brain, such as the dentate gyrus of the hippocampus, shrink significantly.

Along with these physical changes come functional changes: Older people tend to perform cognitive tasks more slowly, for example. Even their automatic brain responses to new stimuli become slower and weaker.

Just when these changes become apparent is still a matter of debate. Some studies suggest after 60; others the mid-40s or even earlier. Whenever they occur, these changes are typically portrayed as manifestations of age-related cognitive decline-a decline that occurs even in the absence of disease, and is sometimes called "healthy cognitive aging."

However, two new studies have put a different spin on age-related cognitive changes, concluding that some of them could be signs of increasing mental maturity and experience, not decline. The authors of one study, led by Michael Ramscar of the University of Tübingen in Germany, even suggested that "many of the assumptions scientists currently make about 'cognitive decline' are seriously flawed and, for the most part, formally invalid."

Bigger database, slower search

The Ramscar study, reported in the January issue of Topics in Cognitive Science, reproduced some of the classic signs of age-related cognitive decline using standard computer models of human cognitive processing. How? Simply by adding new information to the models' memory-banks, to simulate the accumulating experience of the average person as he or she ages.

For example, Ramscar and his colleagues showed that adding new words to the vocabulary of a word-recognition model caused the model to take longer during recognition tasks-even recognitions of words familiar to all-in a way that resembles human age-related slowing on word recognition tasks. In effect, the expansion of the model's memory bank increased the time needed to sift through it to recognize any given word. "[W]e suggest"-wrote the researchers-"that older adults' changing performance reflects memory search demands, which escalate as experience grows."

The Tübingen scientists buttressed their case by showing, firstly, that the cognitive models they used already closely match experimental results from cognitive tests of humans, and secondly, that the average human store of words does keep getting larger even in old age.

Moreover, they argued that the effect of accumulating experience isn't limited to task-slowing, and indeed would be hard to eliminate from most if not all types of cognitive test. One example they cited involves a standard cognitive measure called paired-associate learning (PAL). In a PAL test, the subject initially is shown two paired words, and later is shown one and asked to recall the other. Older people tend to perform not only more slowly but also less accurately on such tests. However, Ramscar and his colleagues pointed out that older people will tend to have more experience of real-world, logical word pairings (e.g., baby - cries) and thus should find it harder to ignore, in effect, such learned pairings when asked to learn new, random ones (e.g., baby - Pluto). That will give them lower scores on PAL tests overall, even though-as actual test data show-they will do about as well as younger subjects for more logical word pairs. Thus, the scientists imply, younger subjects tend to do "better" on this type of test only because their brains tend to be emptier.

What about the well established physical changes with age, in the dentate gyrus for example, that have been taken as evidence for cognitive decline? Ramscar and his colleagues point out that "neurobiological studies can reveal only that the structure and/or biology of neural processing changes" with aging-not that there is some unequivocal "decline" in function.

In short, they conclude, "current studies of aging systematically fail to control for the way vocabulary (and other forms of) knowledge continues to increase throughout adulthood"-and until they do, the notion of age-related cognitive decline should be considered a "myth."

[P]opulation aging is seen as a problem because of the fear that older adults will be a burden on society; what is more likely is that the myth of cognitive decline is leading to an absurd waste of human potential and human capital.

Old dogs don't need new tricks?

One type of cognition-related test doesn't really depend on verbal knowledge at all, but instead involves changes in brain activity in response to simple sensory cues, such as auditory tones. Among these "evoked-response" measures is an increasingly popular one known as the mismatch negativity (MMN) test.

In a typical MMN test, a person listens to a standard tone that occurs again and again so that his or her brain quickly learns to expect it. But then without warning a new, "deviant" tone is played, and this triggers a characteristic "surprise" or "novelty" pattern of activity in the subject's auditory cortex and connected regions, which researchers can detect with precise timing, using electroencephalography (EEG) or magnetoencephalography (MEG).

In older people, the MMN response is slower and weaker, and this has been taken as another indicator of cognitive decline. But once again, cognitive theorists and their models now suggest that this "decline" may be only an artefact of accumulating experience.

In the January issue of PLoS Computational Biology, researcher Rosalyn Moran of Virginia Polytechnic Institute, and her colleagues, reported a study of the MMN test in 97 subjects ages 20 to 83.

Moran took the MEG-measured brain-activity data from the test and fit them to a model, known as a dynamic causal model, of the major brain connections that are thought to mediate the MMN response. From this she determined a likely reason for the slowing and weakening of the response: In older brains, the facility for short-term learning of novel sensory information (such as the MMN-test tones) is effectively dialed down. Thus, while the older brain learns to expect the standard MMN tone, it does so less quickly and less intensely, and its "surprise" at the new, deviant tone is more muted.

These conclusions are again consistent with the notion that signs of "cognitive aging" may reflect accumulating experience rather than deterioration. In this case, the findings add weight to a relatively new concept, according to which the aging brain becomes less reliant on new information as its model of the world becomes more complex and more accurate.

In other words, the aging brain comes to need "new" information less, and thus-by design, to make the most efficient use of its resources-lowers its sensitivity to the newness of sensory experience. As it navigates through the world, it can rely more on the memory-based expectations generated by the higher reaches of its neural hierarchy. Indeed, studies suggest that this "top down" processing does become more active with aging, although researchers had supposed that this is a form of compensation for weaker cognition overall.

"The idea is that as the environment become less novel-through experience-the prediction pathways along top-down axonal connections can adequately signal much of what a younger brain finds surprising," Moran says.

This implies a loss of flexibility and sensitivity to new experiences-traits long associated with aging. But for a species that evolved to live for only a few decades, it is arguably an optimization: maintaining adequate function while reducing resource-usage. In this view, the gradual, age-associated disappearance of cells and nerve fibers in various parts of the brain is not necessarily a bad thing, but instead may be, for the most part, a removal of infrastructure that isn't needed. In any case, Moran suspects that the weakening of MMN responses comes not from this loss of cells and nerve fibers, but from a programmed dialing down of sensitivity to novel sensory stimuli.

Unanswered questions

So is the notion of age-related cognitive decline just a myth, as Ramscar and his colleagues have argued?

"I don't think they've convinced everyone-that's for sure," says Thomas Hills, a researcher at the University of Warwick (UK) who co-authored a commentary on the Ramscar et al study.

Hills points out that the Ramscar study in particular tries to link standard cognitive task-performance slowing or even worsening to increasing experience, which seems to imply that one would perform faster and better by learning less. Yet-at the far end of life, at least-the opposite seems to be true."People who are smarter in general tend to show the least amount of age-related cognitive decline," Hills says. "And people who retire earlier, and thus stop putting so much new stuff into their heads earlier, tend to get dementia faster."

Also, the performance of the rest of the body, including brain areas that control non-cognitive functions such as muscle movements and balance, clearly tends to decline from middle age-professional athletes older than 40 are rare-and why should cognition be so different? There are, moreover, specific neurobiological changes with aging that are hard to see as anything other than degeneration-for example, the slow breakdown of the myelin sheathing on nerve fibers. Factors such as physical exercise and calorie restriction appear to slow signs of cognitive decline and degeneration and in some animals significantly extend lifespan, suggesting that they are countering processes of true decline. Finally, dementia conditions such as Alzheimer's, which are not improvements in any sense, bring about faster forms of the same cognitive-task-performance changes seen with ordinary aging-which again suggests that the latter are based on degeneration and decline.

As Hills points out, any supposed driver of cognitive changes in middle-aged and elderly people needs to be viewed in the full context of aging, in which many things are changing, not just time and experience. Amid this complexity, it may be very difficult to separate any pure "aging" effect from other effects that are usually associated with age-which is why it may be hard to resolve debates over these effects' relevance. For example, common "late-onset" Alzheimer's is now thought to cause detectable brain lesions and other changes, and measurably diminish cognition, in higher-risk people starting in their 40s, long before they would be diagnosed with dementia. "That's been a problem with the aging literature all along: we've had real difficulty separating out the people who have non-healthy aging from the people who have healthy aging," Hills says.

In general, there are many conditions, from obesity and diabetes to systemic inflammation and decreasing physical mobility, that are age-related and can independently affect cognition.

Still, neuroscientists are not closed to the possibility that some age-related changes in cognitive function aren't marks of decline. Carol Barnes, a University of Arizona cognitive aging researcher (and member of the Dana Alliance for Brain Initiatives) notes that "this dichotomy in the way age changes can be interpreted has been around a while." In one of her earliest papers, she says, she suggested that the decreased flexibility of aging rat synapses could be considered a deficit or alternatively "an adaptive mechanism to protect important old information from overwriting or interference with new information."

Thus the arguments presented in the Ramscar and Moran papers, she says, are "perhaps not as heretical as they sound."