Study Shows Humans Have Ability To Track Odors, Much Like Bloodhounds

Student volunteers presented with odors to one nostrilor the other could reliably discern where the odor was coming from, andfunctional magnetic resonance images of their brains showed that thebrain is set up to pay attention to the difference between what theleft and right nostrils sense, much the way it can localize sounds bycontrasting input from the ears.

"It has been very controversialwhether humans can do egocentric localization, that is, keep their headmotionless and say where the spatial source of an odor is," said studycoauthor Noam Sobel, associate professor of psychology at UC Berkeleyand a member of the campus's Helen Wills Neuroscience Institute. "Itseems that we have this ability and that, with practice, you couldbecome really good at it."

In future experiments, UC Berkeleybiophysics graduate student Jess Porter and Sobel plan to trainvolunteers to track odors in the field and test the limits of odorlocalization in humans.

Porter, Sobel and their colleagues reported the results in the August 18 issue of the journal Neuron.

Ina review appearing in the same issue of the journal, Jay A. Gottfriedof the Department of Neurology at Northwestern University's FeinbergSchool of Medicine noted that the UC Berkeley findings open numerousavenues for further research. "Finally, what are the implications forthe Provençal truffle hunt?" he wrote, only partly tongue-in-cheek. "Inthe traditional world of the truffle forests, the dog (or pig) is king.The evidence presented here suggests that humans are every bit as wellequipped to carry out the search."

Forty years ago, Nobel Prizelaureate Georg von Békésy claimed that humans had the ability tolocalize odors, based on experiments in 1964 with human subjects. Hesuggested this was done the same way we locate sounds: by contrastingeither the intensity of the odor or the time of arrival.

Sincethen, however, scientists have had difficulty replicating hisexperiments, according to Sobel. One explanation for this failure wasthat von Békésy used chemicals that stimulate not only the olfactorynerve in the nose, but also a nasal sensory nerve, the trigeminalnerve. Most odors stimulate both, and some, like onions and ammonia,are stinging enough to bring tears to the eyes. Perhaps, somesuggested, von Békésy's subjects were localizing odors based ontrigeminal nerve stimulation, not olfactory nerve stimulation.

Toeliminate this confusion, Porter and Sobel used two odors with minimaltrigeminal stimulation - essence of rose (phenyl ethyl alcohol) andcloves (eugenol) - as well as two trigeminal odorants - propionic acid,which smells like vinegar, and amyl acetate, which smells like banana.They delivered the odors through a specially designed mask with anartificial septum that provided separate air flow to each nostril.

Inaddition, they conducted similar experiments on five volunteers who hadno olfactory nerves and therefore couldn't smell at all, a conditionknown as anosmia.

Normal subjects, 16 in all, were able to tellwhich nostril was receiving a squirt of scent, but anosmic volunteerscould only localize the trigeminal odorants, Sobel said. This showsthat humans are able to localize odors through the olfactory nervesalone.

"One possible objection is that the experimental set-up,with a mask that provides separate air flow to each nostril, isartificial. How behaviorally relevant is that?" said Porter. Subsequentexperiments not yet reported, however, provide additional support fortheir hypothesis that the ability to localize odors to one nostril orthe other is realistic.

The experiments were conducted with thesubjects' heads inside a functional MRI to allow the scientists to seewhich areas of the brain were most active during sniffing and attemptsto identify and localize odors. They found that the left and rightnostrils have separate areas of the primary olfactory cortex - thebrain's smell center - devoted to them, indicating that the brain atleast encodes information that could help it localize an odor. Asuccessful detection of an odor is accompanied by more activity in theregion of the olfactory cortex associated with the particular nostril.

"Whilea subject was doing this task, I could look at the brain and tell youhow accurate he or she would be on every trial and on the taskoverall," Sobel said. "So the fact that we have this predictive valuein the data really suggests that we have actually successfully capturedthe mechanism."

What's more, another area of the brain outsidethe olfactory cortex was very active during successful localization.This area, the superior temporal gyrus, is also involved in thelocalization of sounds and visual objects, Sobel said.

"It'sactually a very nice and elegant convergence of this area, the superiortemporal gyrus, that appears to transform non-spatial information intospatial information," he said. "Together, these results are the firstdescription of the mammalian brain mechanisms for extracting spatialinformation from smell."

One key difference between theirexperiment and previous experiments to replicate the results of vonBékésy is that Porter and Sobel asked their subjects to actively sniff,whereas many previous experiments prevented subjects from sniffing.

"Wethink that most people failed to replicate his results for that reason,that is, the extent to which they enabled natural behavior,specifically sniffing," Sobel said. "In some studies subjects asked tolocalize an odor wouldn't be allowed to sniff. That's almost likestudying auditory localization but having your ears plugged. Weactually enabled natural behavior, we enabled subjects to sniff, and wethink that's a major difference."

In addition to Porter andSobel, other authors of the Neuron paper were UC Berkeley seniorscientist Rehan M. Khan of the Department of Psychology and graduatestudents Tarini Anand and Brad Johnson of the Department ofBioengineering. The work was supported by grants from the NationalInstitutes of Health.