Odors convey a large amount of information about our environment. Rodents, in particular, use olfactory information to guide social interaction, predator avoidance, and foraging. All olfactory behavior relies on the brain’s ability to identify and discriminate between odorants reliably across a wide range of concentrations. In our recent work, we sought to determine how the brain is able to form qualitatively similar percepts for odors despite fluctuations in concentration (Wilson et al., 2017).

Our work proposes that the brain uses an elegant hack to avoid this complication. As odors are inhaled, the rise of concentration in the nose is gradual, effectively giving olfactory sensory neurons (OSNs) access to a low concentration “snapshot” of odorant at the beginning of every sniff. This is true even when odor concentration is very high outside the nose. As a result, the most sensitive OSNs fire first during this low concentration snapshot followed by their less sensitive counterparts. By relying on this early set of OSNs, the brain can construct a stable representation of odor identity across low and high concentrations.

To test this model, we first demonstrated that animals can identify odors using this “snapshot” that occurs briefly after the start of inhalation. We trained subjects to report the identity of two target odors while monitoring the timing of subjects’ odor inhalation with a pressure sensor. On a small percentage of odor presentations, we created a strong distractor to prevent subjects from identifying the odor using a precisely timed laser light to activate a large population of OSNs optogenetically. As expected, our distractor had a profound effect on subjects’ performance if it was presented before the animal had time to sample the odorant. Conversely, if subjects were given just 100 ms (0.1 s) to accumulate information before the distractor started, they performed as though the distractor was not present. This shows us that animals can rely solely on this early information to identify odors.

Next, we used multichannel silicon probes to recorded mitral/tufted cells (MTCs) in the olfactory bulb during odor presentations. Most of the MTCs we recorded were active only at the highest concentrations we tested. As predicted, we found a small subset of neurons that responded both at low latency compared to the overall population of responses and at all concentrations across a 100-fold change in concentration. While this does not prove that these MTCs are solely used by the brain to resolve this odor’s identity, it does show that this early information is a good candidate for representing odors’ identities across varying concentrations. Together, these experiments begin to provide evidence for a primacy code in olfaction: an elegant, fast, and efficient algorithm for identifying odors across concentrations.

—Christopher D. Wilson

Read the paper “A primacy code for odor identity” in Nature Communications, published November 14, 2017.