The 21st Century has shown this whole construct to be made of beeswax. Rapid neural growth and pruning happens, sure, but that is nowhere near the end of neural growth and change. However, as a recent report demonstrates, our understanding of the continual creation of new brain cells, the reasons for it and how the new relate to the old, is far from complete.
Dan Cossins, writing for The Scientist, reviews a recent Swedish paper by K.L Spalding et al., “Dynamics of hippocampal neurogenesis in adult
humans.” (Cell, 153:1219-27, 2013). It's a bit hard to follow, but it seems that nuclear tests 50 years ago released carbon 14 into the atmosphere and this somehow made it into human DNA. By comparing the amount of this against the amount of carbon 12, which is naturally occurring (whatever that means), the researchers (n=15) were able to determine the age of individual neurons. They did this in the dentate gyrus, a part of the hippocampus (I always imagine Hippos wearing university colors) and found that roughly a third of the neurons are "renewed" every day. As the term implies, that means a third of the neurons die as well, while new ones take their place.
Why does this happen? The researchers believe it is so that we can form new memories/encode new experiences and keep them distinct from older ones. This would imply that, if the rate of renewal declines as we get older, we should have more and more difficulty with memory, and we do ("Where ARE those car keys?").
So yay. Apparently, the neurogenesis thesis was first proved in human subjects back in 1998 using BrdU injections to label cells, a method that is no longer considered ethical. However, such experiments did continue with mice and rats, as in McDonald and Wojtowicz's strangely similarly titled study from 2005, "Dynamics of neurogenesis in the dentate gyms of adult rats." (Neuroscience letters, 385(1):70-75, 2005). Our information since then has mainly come from studying neurogenesis in mice. Until now we didn't know the rate of new neural growth in humans. "There is surprisingly large amount,” Cossins quotes one of the study authors as saying (italics mine). Their "model" posits that this is not a rotation, but one subpopulation of short-lived neurons that regenerates constantly while the rest stay the same. “It’s welcome to the field as a long-sought confirmation," contributes a learned German commentator (my italics again). Though everyone is happy with activity in the dentate gyrus, we still don't know the distribution or various purposes of regeneration in the brain overall.
I find all this rather disturbing, given the amount that has been written about it with such authority in the past ten years. It's not that I doubt the relevance of rodent brain studies, but it's sure nice to get data from some closer relatives. This experiment and the techniques it introduces provide a new base to build on, which is welcome. But overall, there's still a lot of guesswork.
Luckily, as Tigger would say, guessing is what brains do best!