Science / Medicine : Scientists Trying to Sniff Out Secrets of the Sense of Smell
If you are like most people, you might have a hard time describing just what fresh-brewed coffee, burning rubber or pine needles smell like. But you would have no trouble identifying one of these odors and countless others if they happened to waft your way.
Just how we distinguish among them is a question most of us probably would take for granted. But such questions have baffled scientists for centuries and still do to a surprisingly large degree.
Basically, smelling things is believed to involve sensing and analyzing the molecules of odors--a subtle and complex task that is apparently accomplished with ease at extremely low concentrations of molecules.
“Until about two years ago, I used to always say that olfactory science was pre-Galilean. There has been so much happening that perhaps we may have moved up to a stage comparable to Galileo’s time in the understanding of physical phenomena,” said Charles Wysocki, an olfactory scientist at the Monell Chemical Sense Center of Philadelphia. “But still, it is safe to say that olfaction is the least understood of the senses and for researchers is really pretty much virgin territory.”
Yet at the same time, scientists like Wysocki have derived a fair amount of observation-based data about our sense of smell.
Women, for instance, generally have a more acute sense of smell than men. This has even been shown to hold true in 2-day-old infants, lending some credence to the hypothesis that this sense has an innate basis rather than simply being culturally influenced.
It is known that the sense of smell tends to diminish as people age. Yet this observation would seem to fly in the face of the fact that our olfactory neurons--the nerve cells that send the information about aromas in the air from our noses to our brains--are the only nerve cells in our body that regenerate.
Scientists have also discovered that there seems to be a good deal of variation in what people smell and how well they can smell it. Many people have been shown, for instance, to suffer from a condition called specific anosmias, the nasal equivalent of color blindness.
In addition, studies have shown that as many as 45% of all people cannot smell the chemical androstenone--a musky-smelling odor found in underarm sweat and pork products--no matter how strongly it is present. Other people can smell this particular odor when it is present in concentrations as low as two parts per trillion.
John Kauer, a neurobiologist at the New England Medical Center in Boston, is among a small but growing number of researchers trying to unravel the mysteries of how we smell by studying what are known as olfactory neural networks.
As Kauer emphasizes, researchers do not even know exactly what a smell is. Unlike light and sound, the odor quality at the molecular level cannot be characterized easily along a continuum.
“For the olfactory system,” he said, “we don’t have any good idea about what the critical properties are that are being encoded.”
Aside from not knowing exactly what makes a molecule smell the way it does, another related and formidable problem is how the receptors in our nose and olfactory bulb work to make sense of incoming odors.
Each individual olfactory nerve cell seems able to respond to many different types of odors, and yet there appears to be a good deal of specialization among the receptors as well. The way our olfactory neurons work together to categorize the aromatic information is not known with any precision.
Kauer put it this way: “When we smell bananas, how is it that we can immediately recognize their scent? Does it mean that there is some grand banana neuron somewhere in our brain to which the new aroma is compared?”
Kauer’s notion of a “banana neuron” may sound farfetched, but this, essentially, was the working theory of smell--or olfaction--that scientists used until as recently as 20 years ago.
Put simply, the hypothesis was that somehow odorants worked with receptors much like keys open a lock, and that it was just a matter for the given smell to find the receptor that would open to the smell.
Now, the working analogy seems to have switched from keys in locks to keys on a piano, as several olfactory researchers explained it.
The idea is that there seems to be some large but finite number of receptors that look for a particular piece of a molecule as it is sniffed in through the nose. Out of these, a tremendously large variation of “chords” can be discerned by matching together the information over arrays of receptors.
The important change is that researchers have come to hypothesize that our identification and recognition of a smell happen only from a “chord” or pattern of information across thousands or even millions of nerve cells. But how many distinct keys are on the piano--whether there are 10 different types of receptors or 1,000--remains a mystery.
Contrast this to research into vision, for example, where scientists have discerned two major types of receptors in the retina: rods and cones, which take in all the information needed to enable us to see.
Using video cameras and image-enhancement techniques developed in the space program for satellite reconnaissance, Kauer has recorded what actually happens to the nerve cells in the olfactory bulb of live salamanders when an odor passes their way.
What Kauer and others have demonstrated is that one odorant, say, the smell of bananas, excites certain neurons, forming a distinct pattern over time, while another, like coffee, makes a distinct and different pattern.
Interestingly, however, a familiar smell seems to make a different pattern after multiple exposures, suggesting that even at this stage of processing, some recognition or even memory is taking place.
Kauer’s work fits in neatly with theoretical work by other researchers on information-processing in neural networks. Some of this work comes from computer scientists who are trying to model neural networks as examples of what they call “parallel processing”--systems where information is processed by many individual components simultaneously. By contrast, most standard computers have many-thousand bits of memory but only one processor.
Because the research by Kauer and others has shown that recognition of a smell comes only through a group of receptors working together at once to encode the information, it is a good candidate for this type of modeling.
One such ingenious model comes via a collaboration between neurobiologist Gary Lynch and computer scientist Richard Granger at UC Irvine. Lynch and Granger have designed a computer simulation model that is based on the workings of olfactory receptors in rats.
Simulating individual odors as bar codes (much like those used at the supermarket), these researchers found that the system seems to use the cells closest to the entrance of the nose to look for similarities between molecules and the cells farther downstream in the system to tease out the more subtle differences between the odor inputs.
Kauer said the work by Lynch and Granger shows promise, but that it is still too simplified. But he also acknowledges the importance of the theoretical models like Lynch and Granger’s.
“Without them,” he said, “it is a little like having a radio and trying to understand how it works by sticking a probe into it and listening to what you hear from each little transistor. The real understanding comes from the connectivity of the parts.”
