The Essence of a Sniff

The following is an article from The Annals of Improbable Research.

by Marc Abrahams

Our most ancient, most reliable sense -- the sense of smell -- may be more simple, more complicated, and more intriguing than most scientists realize.
Several researchers have spent forty years and more learning how to measure exactly what a frog's nose tells the frog's brain. They want to understand how the frog's brain knows what the nose is saying.
They have found things that are different from what anyone expected. And different from what most biologists believe about how noses talk to brains.

Do Frogs and People Smell Differently?

Why a frog? Because it's been easier (and most, but these days not all, would say more palatable) to experiment with a frog's nose in a frog than with a person's in a person.
The only way to tell what's really going on in the system -- the olfactory system -- is to measure what's happening there while a creature is smelling a smell. (You can learn a lot by removing pieces and studying them outside the body. But each piece is in there for a reason -- and can behave differently when it's been yanked out.)
It is possible to monitor individual cells in the business end of a nose, in the frog. People have gotten detailed information about how those cells -- the sensory neurons -- actually behave when they get a whiff of something.

Frogs were the first to reveal their nasal secrets, because they were the easiest to work with. Now rats and other animals are being studied this way, too. How much does the frog nose-and-brain tell us about the human nose-and-brain? Probably quite a lot. The olfactory system seems organized pretty much the same way in amphibia (frogs, salamanders, newts), in mammals (rats, mice, probably humans, too) and even in some of the simplest eukaryotes (a eukaryote, remember, is any creature whose cells each have a nucleus.)

A Little Background About How We Smell

The sensory neurons are the business end of the nose/brain system. Any smell information that reaches the brain comes, directly or indirectly, from them and them alone.
The olfactory system is the only one of the senses in which the sensory nerve cells come directly in contact with chemicals from the outside world, without any other cells physically shielding them from harm. (Even in the taste system, the sensory nerve cells have some safety layer of protection from the things that stimulate them.)
The olfactory sensory neurons are distinct among sensory nerve cells. They have very, very long, very thin axons, with no heavy wrapping of the myelin coat that insulates the thicker nerve fibers. It proved rather difficult to measure their internal electrical activity. The pulsations there are a monologue, by the neuron, about what it is and isn't smelling.
(By the way: of all the neurons, sensory or non, in the vertebrate body, only a tiny percentage have a thick, myelin-wrapping -- yet myelinated neurons are almost the only ones that have been studied. The activity of more than nine tenths of the nervous system is largely a mystery to most scientists.)

"How Does It Work?" vs. "What Are the Parts?"

In the late 1950s and early 1960s a small group at MIT became adept at measuring electrical signals from small nerve cells, first of the eye and then of the olfactory system. They were, and in many respects still are, pioneers at this.
It was a peculiar group of people, electrical engineers who became interested in biology and then turned their unusual skills and much of their lives to poking, measuring, and trying to make sense of nature's most complex electrical system: the brain and the nervous system connected to it.
Robert Gesteland, who is now a professor of neurobiology at the University of Cincinnati, became the guru of olfactory system measurement, developing ways to reliably measure the activity of individual sensory neurons in living, smelling frogs. It is even now just a small community of scientists who do this kind of research.
Other scientists, far more numerous, have been investigating the individual nuts and bolts of the olfactory system. They concentrate on the smell-related genes and on the odor receptors, the specific parts of the sensory neuron surface that respond to particular odor molecules.
The two lines of approach have unfortunately been somewhat disconnected. One measures and describes the workaday activities inside the olfactory system. The other examines the system's design and structure: its DNA blueprints and chemical components.
Gesteland and the rest of the "let's look at how it works" gang have discovered some puzzling activities in the olfactory system. The DNA/biochemical research groups, which are dominant both in size and in funding, have generally dismissed those activities as being puzzling, irrelevant, or apocryphal.
Certainly, these things are puzzling. They are also well documented.

Why Are These Cells Different From All Other Cells?

Here's some of what Gesteland, and his colleagues and competitors around the world, have discovered.

In the frog, olfactory sensory neurons behave differently from all other nerve cells. Here are some peculiar things about them:
<> On average, each olfactory receptor cell lives for only a few weeks, and is then replaced by a new cell. (All other neurons in the olfactory system -- and in the rest of the nervous system -- last for years, in many cases for the lifetime of the animal. Replacement of those dead cells is sluggish, if it occurs at all.)
<> In the frog's nose, every sensory neuron responds to a large number of different stimulus substances, typically to a quarter or a half of all the odors presented to it. (This is very unlike the nerve cells in the eye. The eye has several different kinds of sensory cells, and each kind responds only to some particular, limited range of stimuli -- a certain range of color, for example.)
<> No two of these cells behave the same way. Expose twenty different olfactory sensory neurons to the same odor, and each neuron will produce a different electrical signal. It's as if each nerve cell looks at the world from a different point of view. (This, too, is unlike the visual system.)
<> They are also not very reliable, in the sense that any particular cell does not always react to the same odor in the same way.
By the way: all this variability is not necessarily a bad thing, considering what a nose is expected to do. In general it has to notify the creature when it's smelled something new (something putrid, for example), but not so much when that smell has been lingering ("still smells putrid!"). And because these sensory neurons come in contact with chemicals -- virtually any kind -- from the outside world, they are more at risk than any other nerve cells in the body, and so more often in need of replacement.

The Big Question

There's a big question lurking in the background -- a tantalizing, juicy question that no one seems close to answering.
In a system where the sensing parts are all different from each other, where every one of them gets replaced frequently, and where every replacement part behaves differently from its predecessor, how do smells get remembered and recognized?
To put this another way: now that we know what information the sensory cells are picking up, and we know that the pickup system is composed of varying, ever-changing parts, how is the information processed and used so reliably?
Smell memory is the most permanent and strongest kind of memory, yet it depends on the greatest variation. The solution to this puzzle could open many interesting doors.

Where to Start

Perhaps you could be the person who figures this out. If you want to dig into it, here are some good places to start:
For the basic early work on frog olfactory receptor neurons, see "Speculations on Smell," Jerome Y. Lettvin and Robert C. Gesteland, Cold Spring Harbor Symposium, vol. 30, 1965, pp. 217-25. (For a more technically detailed version of it, see "Chemical Transmission in the Nose of the Frog," Robert C. Gesteland, Jerome Y. Lettvin, andWalter H. Pitts, Journal of Physiology, vol. 181, 1965, pp. 525-9.)
For a summary of recent single receptor neuron physiological studies, see "Peripheral Odor Coding in the Rat and Frog: Quality and Intensity Specification," P. Duchamp-Viret, A. Duchamp and M.A. Chaput, Journal of Neuroscience, vol. 20, 2000, pp. 2383-90.
To see a very different, contrasting view, from one of the leaders of the molecular genetics approach, read "The Molecular Architecture of Odor and Pheromone Sensing in Mammals," L.B. Buck, Cell, vol. 100, 2000, pp. 611-18.