How to make a mouse smell a smell that doesn’t actually exist

Scientists have implanted an artificial odor directly in the brains of mice. It doesn’t mean that mental Smell-O-Vision technology is coming soon. But the results, published June 18 in Science, deliver clues to how the brain processes information.

Details about the synthetic smell may help answer “fundamental questions in olfaction,” says computational biologist Saket Navlakha of Cold Spring Harbor Laboratory in New York, who wasn’t involved in the study. Studies on the senses offer a window into how brains shape signals from the outside world into perceptions, and how those perceptions can guide behavior (SN: 7/18/19). 

To build artificial smells in mice’s brains, researchers used optogenetics, a technique in which light prods genetically engineered nerve cells to fire signals (SN: 1/15/10). Neuroscientist Dima Rinberg of New York University’s Grossman School of Medicine and colleagues targeted nerve cells in mice’s olfactory bulbs. There, clusters of nerve endings called glomeruli organize the smell signals picked up in the nose.

Like playing a short ditty on a piano, Rinberg and colleagues activated nerve cells in six spots (each of which might include between one and three glomeruli) in a certain order. This neural melody was designed to be a simplified version of how a real odor might play those nerve cells. (It’s not known what the artificial odor actually smells like to a mouse.)

Mice learned to signal the presence of this artificial smell by licking one of two spouts. The synthetic odor didn’t objectively exist, but the mice behaved as though they smelled it anyway, the researchers found. After “smelling” the synthetic odor, mice reliably licked the correct spout. Other scrambled signals, also delivered by optogenetics, didn’t cause the same reaction.

Together, the individual spots the researchers stimulated built the perception of the odor, just as a string of notes makes a melody. Because the smell was completely synthetic, though, researchers could mess with it. By slightly changing some of the signals generating the artificial odor, the researchers could test which qualities of the sequence were important, and which changes rendered the odor unrecognizable.

The beginning of the smell sequence seemed to be key. When the researchers swapped the order of the first few spots’ activity, the mice had more trouble identifying the smell than they did when the spots toward the end of the sequence were changed. And delays near the beginning mattered more than delays toward the end. “If you modify the first few notes, you more easily ruin the song,” Rinberg says.

This result supports an idea called the primacy effect, which holds that the neural signals that come first in a sequence carry more weight, says Tatyana Sharpee, a computational neuroscientist at the Salk Institute for Biological Studies in La Jolla, Calif., who was not involved in the study.

More generally, these results offer an example of how changes in neural activity can affect a perception, Sharpee says. “Ultimately, this hints at the fundamental properties of the neural code.” 

Sharpee suspects that similar properties might apply to other kinds of information processed by the brain, including vision and hearing signals, and perhaps even to more complex tasks such as memory. Such processes all rely on the same basic transformation, she says — “a general mathematical problem of coding inputs to outputs.” In this way, the brain takes incoming information about the world and stitches it into useful perceptions.  

In addition to following the synthetic smells into other parts of the brain, Rinberg and his colleagues want to test whether similar rules apply to real smells. “A synthetic odor is fantastic. It’s a super interesting tool,” Rinberg says. “But at the end of the day, I want to know how odors are formed in the brain.”