When the N.I.M.H. study began in 1991, Giedd said he and his colleagues expected to stop when the subjects turned 16. “We figured that by 16 their bodies were pretty big physically,” he said. But every time the children returned, their brains were found still to be changing. The scientists extended the end date of the study to age 18, then 20, then 22. The subjects’ brains were still changing even then. Tellingly, the most significant changes took place in the prefrontal cortex and cerebellum, the regions involved in emotional control and higher-order cognitive function。

　　As the brain matures, one thing that happens is the pruning of the synapses. Synaptic pruning does not occur willy-nilly; it depends largely on how any one brain pathway is used. By cutting off unused pathways, the brain eventually settles into a structure that’s most efficient for the owner of that brain, creating well-worn grooves for the pathways that person uses most. Synaptic pruning intensifies after rapid brain-cell proliferation during childhood and again in the period that encompasses adolescence and the 20s. It is the mechanism of “use it or lose it”： the brains we have are shaped largely in response to the demands made of them。

　　We have come to accept the idea that environmental influences in the first three years of life have long-term consequences for cognition, emotional control, attention and the like. Is it time to place a similar emphasis, with hopes for a similar outcome, on enriching the cognitive environment of people in their 20s?

　　N.I.M.H. scientists also found a time lag between the growth of the limbic system, where emotions originate, and of the prefrontal cortex, which manages those emotions. The limbic system explodes during puberty, but the prefrontal cortex keeps maturing for another 10 years. Giedd said it is logical to suppose — and for now, neuroscientists have to make a lot of logical suppositions — that when the limbic system is fully active but the cortex is still being built, emotions might outweigh ration-ality. “The prefrontal part is the part that allows you to control your impulses, come up with a long-range strategy, answer the question ‘What am I going to do with my life?’ ” he told me. “That weighing of the future keeps changing into the 20s and 30s。”

　　Among study subjects who enrolled as children, M.R.I. scans have been done so far only to age 25, so scientists have to make another logical supposition about what happens to the brain in the late 20s, the 30s and beyond. Is it possible that the brain just keeps changing and pruning, for years and years? “Guessing from the shape of the growth curves we have,” Giedd’s colleague Philip Shaw wrote in an e-mail message, “it does seem that much of the gray matter,” where synaptic pruning takes place, “seems to have completed its most dramatic structural change” by age 25. For white matter, where insulation that helps impulses travel faster continues to form, “it does look as if the curves are still going up, suggesting continued growth” after age 25, he wrote, though at a slower rate than before。

　　None of this is new, of course; the brains of young people have always been works in progress, even when we didn’t have sophisticated scanning machinery to chart it precisely. Why, then, is the youthful brain only now arising as an explanation for why people in their 20s are seeming a bit unfinished? Maybe there’s an analogy to be found in the hierarchy of needs, a theory put forth in the 1940s by the psychologist Abraham Maslow. According to Maslow, people can pursue more elevated goals only after their basic needs of food, shelter and sex have been met. What if the brain has its own hierarchy of needs? When people are forced to adopt adult responsibilities early, maybe they just do what they have to do, whether or not their brains are ready. Maybe it’s only now, when young people are allowed to forestall adult obligations without fear of public censure, that the rate of societal maturation can finally fall into better sync with the maturation of the brain。

　　Cultural expectations might also reinforce the delay. The “changing timetable for adulthood” has, in many ways, become internalized by 20-somethings and their parents alike. Today young people don’t expect to marry until their late 20s, don’t expect to start a family until their 30s, don’t expect to be on track for a rewarding career until much later than their parents were. So they make decisions about their futures that reflect this wider time horizon. Many of them would not be ready to take on the trappings of adulthood any earlier even if the opportunity arose; they haven’t braced themselves for it。

　　Nor do parents expect their children to grow up right away — and they might not even want them to. Parents might regret having themselves jumped into marriage or a career and hope for more considered choices for their children. Or they might want to hold on to a reassuring connection with their children as the kids leave home. If they were “helicopter parents” — a term that describes heavily invested parents who hover over their children, swooping down to take charge and solve problems at a moment’s notice — they might keep hovering and problem-solving long past the time when their children should be solving problems on their own. This might, in a strange way, be part of what keeps their grown children in the limbo between adolescence and adulthood. It can be hard sometimes to tease out to what extent a child doesn’t quite want to grow up and to what extent a parent doesn’t quite want to let go。