What tempted me to use “replacing” rather than “growing” was that the neurons giving the progenitor cells input were the same cell type (anatomically) that the progenitors were differentiating into. Let me explain: The article indicates that stimulation of dentate gyrus (or, more specifically, theta burst stimulation of the perforant pathway, which goes from cortex to dentate gyrus) leads to release of GABA on the progenitor cells. (It is my understanding that the progenitor cells start out in the sub-granular zone (SGZ) of dentate and then, upon maturation, migrate into the granule cell layer of dentate — ie. they don’t go too far over the course of their life cycle.)

So, if I’m understanding the experiment correctly, we have stimulation of DG GABAergic cells releasing GABA onto SGZ progenitors, which then themselves become dentate cells. What’s unclear is if the progenitors become GABAergic or glutamateric adult neurons. (There is some evidence that DG cells actually secrete both neurotransmitters, too. But that is for another discussion…) This might already be known (whether, in vivo, progenitors become excitatory/inhibitory/both neurons)… I am not sure and my casual PubMed search couldn’t dig up a straight answer.

The story in vitro is somewhat clearer. There is evidence (for example, here) that SGZ progenitors can be coaxed into becoming either excitatory or inhibitory cells, although it is not clear how natural such in vitro conditions (eg. use of particular neurotrophins) are.

Of course, the best evidence for “growing” over “replacing” is the recent work showing that some of our neurons as almost as old as we are. That is, there are neurons with us since almost birth that stay with us until death.

Still, given the limited volume of our skulls and the inherent frailty of wetware, it seems likely that at least some of the progenitors might be “replacing” — and to me, this is an especially tantalizing idea if the “replace me” signal is coming from the mature dentate gyrus interneurons themselves!