Experimental control of neural activity has become more and more sophisticated, with glutamate uncaging and multineuron patch clamping allowing the targeting of a specific neural area. But there is a catch. “You can't target cell types in that way,” says Deisseroth. Usually the individual cell types “are sparsely embedded in the networks.”
The obvious solution to this is genetics. Promoters to drive expression in specific cell types are available, as many of the interneuron types express unique neuropeptides or other markers. Initial attempts have met with partial success, but the complexity of the introduced signal cascades has meant that control has been on the order of seconds and minutes rather than milliseconds.
The Stanford group used rapid optical switches plus the single component channelrhodopsin-2 from the green alga Chlamydomonas reinhardtii. After lentivirus infection of rat neurons with their gene construct, blue light resulted in rapid depolarizing currents. Repeated light pulses could elicit spike trains typical of active neurons, with patterns that were reproducible in either the same or different neurons. In the absence of light, the resting potential, response to injected current, and cell health were unaffected.
Deisseroth is interested in how certain cell types connect with others, and what specific function each one provides. He plans to apply the new technique in the mouse or rat hippocampus to get at some of the more enigmatic functions—such as mood—controlled by this brain region. Experiments with brain sections may be followed by experiments in live animals using optical fibers and 2-photon excitation.