The rise and rise of channel rhodopsin-2.

So the first "real use" of ChR2 has just appeared. Karel Svoboda and co-workers have published a beautiful paper in the first issue of Nature in 2008 (Nature again I hear you ask!). In what might be the best piece of work Karel has done so far, they electroporate DNAs for ChR2 and RFP into layer 2/3 neurons in mice in utero and then train them to chose to drink water, associating the choice with light-driven APs in vivo. They can then drive the choice merely by flashing blue light into the brain of a freely moving mouse. They calculate that 60-300 neurons are required for this learning task. In a technically related experiment, Deiseroth and co-workers have a toolbox paper in the Journal of Neuroscience showing how you can make a mouse run (video) by expressing ChR2 in layer 5 neurons. The revolution continues.

Sparse optical microstimulation in barrel cortex drives learned behaviour in freely moving mice. Nature (2008) 451: 61-64.

Targeting and Readout Strategies for Fast Optical Neural Control In Vitro and In Vivo. J. Neurosci. (2007) 27: 14231-8 (so much for the 1000 word 10 reference limit on toolbox papers. When you're hot, you're hot!)