In a joint research effort led by Dana M Small of The John B Pierce Laboratory and Yale University and Thomas Hummel of the University of Dresden Medical School, the researchers launched their exploration into the brain's possible dual response to odors because of the well-known phenomenon that sensing an odor "orthonasally" through the nose triggers the perception that it is coming from the outside world, while sensing it through the mouth--or "retronasally"--causes the perception that it arises from the mouth.
"The illusion that retronasally perceived odors are localized to the mouth is so powerful that people routinely mistake retronasal olfaction for 'taste,'" they wrote. "For example, we may say that we like the 'taste' of a wine, because of its fruity or spicy notes. However, gustation refers only to the sensations of sweet, sour, salty, savory, and bitter, and thus the pleasant 'taste' to which we refer is actually a pleasant odor sensed retronasally."
"The role of olfaction in taste is powerful," they said. For example, they pointed out that pinching the nose while eating or drinking--which blocks airflow from the mouth through the olfactory system--blocks flavor perception. Releasing the nose restores the sense of flavor in the mouth.
"The fact that the olfactory referral illusion is maintained even though the subject is now aware that the experience is related to an event in the nose demonstrates that olfactory referral is robust and cognitively impenetrable."
To begin to penetrate the neural cause of olfaction's duality, the researchers devised the first experiments to directly compare the same odorants introduced through the nose and the mouth.
"Although several studies have examined brain responses to retronasal olfactory stimulation, none have directly compared orthonasal and retronasal stimulation in the same subjects or considered the possibility that the effects of route of stimulation depend on the way that odors are typically sensed," wrote Small and her colleagues. "For example, food odors are normally experienced both orthonasally and retronasally, whereas nonfood odors are perceived only orthonasally. Therefore, it is possible that the route of stimulation may have different effects for food versus nonfood odors."
In their experiments, they inserted small tubes into the noses of volunteers such that one tube ended at the nostrils and the other ended farther back in the nasal passage near the throat, where odors from the mouth would originate. As they introduced odors into one tube or the other, they scanned the subject's brains using functional magnetic resonance imaging, a technique in which harmless magnetic fields and radio waves detect increased blood flow to brain areas, which reflects increased activity.
They used four odorants: chocolate odor represented a food odor, and lavender represented a non-food odor that was similarly pleasant as chocolate. They also chose two odorant chemicals--butanol and farnesol--to test a theory that the olfactory system distinguishes molecules according to whether they are more water soluble (butanol) or oily (farnesol).
The researchers found that the chocolate odor, indeed, activated different brain regions according to the route of administration, supporting the duality of olfaction. The lavender odor did appear to activate different regions, but to a far lesser extent.
"The effect of route of delivery was greatest for the chocolate odor, rais ing the possibility that odorant administration interacts with experience to engage unique brain regions and that olfactory referral induced by retronasal stimulation creates a differential reward context for food but not for non-food odors by signaling availability versus receipt of food," they concluded. "Because the current study tested only one food, future experiments are needed to determine whether other food odors produce the same differential brain activations," they wrote.
However, the researchers found that the two different chemicals butanol and farnesol did not elicit significantly different brain responses according to the route of delivery, indicating that the properties of the molecules do not play a role in the response.