Researchers Make Mice Smell Odors that Aren’t Really There

In the case of smell, for example, it is well-known that odorous molecules traveling up the nose bind to receptors on cells that then transmit signals along their axons to bundles of nerve endings—glomeruli—in a brain area called the olfactory bulb. A single molecule can cause a whole array of different glomeruli to fire in quick succession, explains neurobiologist Kevin Franks of Duke University who also did not participate in the research. And because these activity patterns “have many different spatial and temporal features,” he says, “it is difficult to know which of those features is actually most relevant [for perception].”

To find out, neuroscientist Dmitry Rinberg of New York University and colleagues bypassed the nose entirely. “The clever part of their approach is to gain direct control of these neurons with light, rather than by sending odors up the animal’s nose,” Caltech neurobiologist Markus Meister, who was not involved in the work, writes in an email to The Scientist.

See “How Manipulating Rodent Memories Can Elucidate Neurological Function”

The team used mice genetically engineered to produce light-sensitive ion channels in their olfactory bulb cells. They then used precisely focused lasers to activate a specific pattern of glomeruli in the region of the bulb closest to the top of the animal’s head, through a surgically implanted window in the skull. The mice were trained to associate this activation pattern with a reward—water, delivered via a lick-tube. The same mice did not associate random activation patterns with the reward, suggesting they had learned to distinguish the reward-associated pattern, or synthetic smell, from others.

Although the activation patterns were not based on any particular odors, they were designed to be as life-like as possible. For example, the glomeruli were activated one after the other within the space of 300 milliseconds from the time at which the mouse sniffed—detected by a sensor. “But, I’ll be honest with you, I have no idea if it stinks [or] it is pleasant” for the mouse, Rinberg says.

Once the mice were thoroughly trained, the team made methodical alterations to the activity pattern—changing the order in which the glomeruli were activated, switching out individual activation sites for alternatives, and changing the timing of the activation relative to the sniff. They tried "hundreds of different combinations," Rinberg says. He likened it to altering the notes in a tune. “If you change the notes, or the timing of the notes, does the song remain the same?” he asks. That is, would the mice still be able to recognize the induced scent?

From these experiments, a general picture emerged: alterations to the earliest-activated regions caused the most significant impairment to the animal’s ability to recognize the scent. “What they showed is that, even though an odor will [induce] a very complex pattern of activity, really it is just the earliest inputs, the first few glomeruli that are activated that are really important for perception,” says Franks.

Rinberg says he thinks these early glomeruli most likely represent the receptors to which an odorant binds most strongly.

With these insights into the importance of glomeruli firing times for scent recognition, “the obvious next question,” says Franks, is to go deeper into the brain to where the olfactory bulb neurons project and ask, “ How does the cortex make sense of this?”

E. Chong et al., “Manipulating synthetic optogenetic odors reveals the coding logic of olfactory perception,” Science, 368:eaba2357, 2020.