Benedict Carey writes: People of a certain age (and we know who we are) don’t spend much leisure time reviewing the research into cognitive performance and aging. The story is grim, for one thing: Memory’s speed and accuracy begin to slip around age 25 and keep on slipping.
The story is familiar, too, for anyone who is over 50 and, having finally learned to live fully in the moment, discovers it’s a senior moment. The finding that the brain slows with age is one of the strongest in all of psychology.
Over the years, some scientists have questioned this dotage curve. But these challenges have had an ornery-old-person slant: that the tests were biased toward the young, for example. Or that older people have learned not to care about clearly trivial things, like memory tests. Or that an older mind must organize information differently from one attached to some 22-year-old who records his every Ultimate Frisbee move on Instagram.
Now comes a new kind of challenge to the evidence of a cognitive decline, from a decidedly digital quarter: data mining, based on theories of information processing. In a paper published in Topics in Cognitive Science, a team of linguistic researchers from the University of Tübingen in Germany used advanced learning models to search enormous databases of words and phrases.
Since educated older people generally know more words than younger people, simply by virtue of having been around longer, the experiment simulates what an older brain has to do to retrieve a word. And when the researchers incorporated that difference into the models, the aging “deficits” largely disappeared.
“What shocked me, to be honest, is that for the first half of the time we were doing this project, I totally bought into the idea of age-related cognitive decline in healthy adults,” the lead author, Michael Ramscar, said by email. But the simulations, he added, “fit so well to human data that it slowly forced me to entertain this idea that I didn’t need to invoke decline at all.” [Continue reading…]