So, actually, the preflash is a two-way dialog between the control flash (in this case, built in to the D7000) and the remote units. There's a reverse-engineered (and several years old, so possibly slightly out of date) explanation of the Nikon optical protocol by Alson van der Meulen. (The site is offline but archived.)

Basically, the control flash fires a minimal pulse, which actually triggers a reduced-power response from the remote flashes. The TTL system in the camera measures this response, and uses it to calculate the right power levels for each. (This may or may not include the control flash.)

Then, the controller flashes again with a reduced-power series of pulses which digitally encode instructions for the remotes, and in response every flash that's included in the exposure fires its main pulse.

All of this happens so quickly that it's impossible (or at least extremely difficult) to observe it unaided.

However, the last control flash actually happens with the shutter open, so even though it's at low power, if you have a wide aperture, or if a reflection of the camera appears in the scene, it can actually be seen. I assume this is a necessity in order to get the timing to work out.

(This works the same way for Pentax as well, although without separate control groups. I assume Canon and Olympus are basically the same but I don't know.)