E-noses

Humans have always enhanced their senses with technology. We can’t hear as well as cats, or see as well as hawks, but with parabolic microphones and binoculars, we make up for it.

One sense has lagged in technological enhancement: smell. In the mammalian nose, smelling takes place in tissue called the olfactory epithelium. We have four or five square centimetres of this tissue. Dogs have 18 square centimetres, and cats 21.

Can the electronic nose—the e-nose—help us catch up?

E-noses analyze gas samples by passing the gas over electrodes coated with a substance that changes resistance in the presence of certain molecules. Simple versions, designed to detect just a handful of specific gases, are already in use in the food industry (to sniff out rotting food) and the chemical industry (to help find leaks).

But more complex versions are beginning to take on more challenging tasks, helped along not just by more sensitive sensors but by neural networks, computer software that, by mimicking the way the brain works, can be taught to discriminate among various smells.

For example, in England researchers are trying to teach an electronic nose to recognize the distinctive smell of methicillin-resistant Staphylococcus aureus (MRSA), one of the dangerous antibiotic-resistant “superbugs” increasingly found in hospital settings. By having it “sniff” swabs taken from infected patients, hey’ve successfully trained it to recognize the smell of Staphylococcus aureus, but not to distinguish between the ordinary strain and MRSA. Even being able to detect who has been exposed to some kind of the bug, though, would be helpful in controlling outbreaks by limiting the number of people who would have to be tested by current means. Culturing MSRA takes one or two days, and even a new DNA-based test coming soon takes a couple of hours. The e-nose reaches its verdict in 15 minutes.

In just a few years, we may see electronic noses so sensitive they’ll be able to do something that sounds like magic (or at least science fiction): diagnose disease by sniffing someone’s breath.

University of Buffalo researchers are trying to scale up and scale down existing e-noses at the same time: scale up their sensitivity, and scale down their size.

Led by chemist Frank Bright, they’ve engineered an e-nose with 100 sensors so far, and their goal is a million. When a patient breathes into the device, different molecules in his or her breath will cause different sensors to respond. Researchers have already identified “biomarkers” in the breath linked to specific diseases, ranging from asthma to diabetes to AIDS to schizophrenia. Right now, analyzing breath for those biomarkers requires a bulky gas chromatograph. The new device, which they expect to have a prototype of, ready to enter clinical trials, within a year, will be small enough to work anywhere: “in the field, at the mall, in the parking lot, and at home,” says Bright.

Meanwhile in Europe, researchers from universities in Spain, France and Italy are creating an electronic nose that combines human engineering with nature, by coating a gold microelectrode with a layer of the olfactory receptor proteins found in the mammalian nose. The microelectrode is mounted on a two-millimetre-long computer chip. The other end of the chip is immersed in a liquid cell containing additional microelectrodes connected to an instrument that measures electrochemical changes. Currently, the researchers are pumping odor-causing chemicals into the liquid cell, then recording the response from the olfactory receptor proteins. Even this initial version can detect odorants at levels imperceptible to humans.

So far the researchers have only used one olfactory receptor protein from rats and one from humans. A full-scale nose would have hundreds of proteins (the human nose uses about 1,000 proteins, which enables the brain to distinguish about 10,000 different smells). The European researchers expect their electronic nose to reach prototype stage in five to 10 years.

Of course, that’s 35 to 70 dog years which I mention only because dogs have already done what we’re trying to create e-noses to do. In a study earlier this year, five household dogs were trained within three weeks to detect lung or breast cancer by sniffing the breath of cancer patients, with an accuracy ranging from 88 percent for breast cancer to 97 percent or better for lung cancer.

Once again, we’re playing catch-up with our furry friends, using technology to augment our relatively poor senses.

But then, we’ve had a lot of practice.

Permanent link to this article: https://edwardwillett.com/2006/06/e-noses/

Leave a Reply

Your email address will not be published.

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Easy AdSense Pro by Unreal