A very cool new molecular neuro trick Here's some background and a loose summary. Here is the main reference (surprisingly openly available!). In recent years, neurobiologists have been able to get neurons to make particular proteins by various means (such as filling the cells or in someway with RNA's (such as injection), or integrating code into the filled cells DNA to make proteins, or by genetically transforming whole organisms. In any case some set of cells in the animal has some recombinant code that can express (make) particular engineered genes. Some of these inserted genes produce membranes that insert into outer membranes of neurons. Often they are light activated ion channels, from bacteria or algae that directly effect the neurons electric membrane properties. Some hae their light sensitivities engineered For several years , researchers have used light sensitive proteins (such as rhodopsin) linked (physically or physiologically) to control other proteins such a as ion channels. This allow one to control neuronal activity with light (which is handy in the lab). By shooting light backwards through a microscope, it is possible to apply such light to a much smaller area. Using a confocal microscope even allows you to choose different z-axis levels to intensely illuminate. A two photon confocal microscope has even sharper z-axis resolution since excitation is dependent upon a very high concentration of photos only achieved near the focal point. This is basically optogenetics which has been around for a while. Of course scientists are always thinking of different ways to do things and this paper has an interesting and technically challenging approach. The idea is to stimulate single cells with IR light, by causing local heating, which causes a change in some heat sensitive protein (endogenously added in some way) to cause a change in neural activity. Problem is: you don't want to make too much heat or you could cause general cellular damage instead of seeing the effects of just tweeking membrane physiology (so only small temperature changes are allowable). And you are basically heating a small bit of an aqueous media (the cells and their surrounding tissue environment), so the heat will diffuse away quickly. This requires a protein that is very sensitive to small temperature changes. Since people do not yet know how to code (molecularly) for the construction of proteins like this from scratch (using only basic principles of protein structure), these guys (from Russia) cleverly let Mother Nature do the construction work and just had to find the results among the products of nature (the world, not the publication (until now)). They used a protein (the thermosensitive transient receptor potential (TRP) which open a cation (plus charged ion) channels) that is the sensory protein in a pit viper's pit (a sensory organ pit vipers use to find or "see" their warmer than their surroundings prey in infrared (the target's heat signature). The pit organ is constructed like a poor pinhole camera (big opening --> low image quality), which is structurally similar to a proposed stage of eyeball evolution. To detect IR differences in its environment, these animals have evolved proteins that are very sensitive to temperature differences. This protein was used to supply a highly sensitive heat sensitive controller for their neural processes. They used a fiber optic probe to apply the IR light to a 60 µm spot (about the size of a large neuron cell body). This could activate these channels in 10 ms. Whether shooting light backwards through a microscope, to apply such light to a much smaller area could heat up an area enough to turn on the proteins is not known. This technique was able to drive the zebrafish startle-escape response (AKA the C-start behavior ("C" for the curve shape the body takes) which transmits Rohan-Beard sensory neuron, to large hindbrain interneurons, to spinal cord motorneurons, to muscles). This could technique could allow one to use IR to influence of the ongoing activity in a retina which is being independently driven by illumination with what ever pattern of light your might want to drive it with. Thus, one could observe the at modulation of this "normal pattern" of activity by a specific experimentally controlled physiological effects on particular individual cells. An additional advantage of using IR is that it penetrates tissue better than shorter visible wavelengths and could be used through thicker pieces of tissue. And by using voltage sensitive fluorescent proteins (or proteins that sense intracellular increases in Ca++ which is often increased by neuronal activity) one could observe the activity of one, some, or many cells at once, such as without sticking each cell with microelectrodes (technically difficult). The fluorescent levels can be measured in small time intervals with scanning confocal microscopical scans.