Follow your nose

Smell triggers memories more readily than other sensory clues. Laura Spinney reports on new research that draws on this primal link to help treat psychological conditions

Aromatherapy is the most widely used complementary therapy in the National Health Service, and doctors use it most often for treating dementia. For elderly patients who have difficulty interacting verbally, and to whom conventional medicine has little to offer, aromatherapy can bring benefits in terms of better sleep, improved motivation, and less disturbed behaviour.

So the thinking goes. But last year, a systematic review of healthcare databases found almost no evidence that aromatherapy is effective in the treatment of dementia. The only trial included in the review that seemed to show a benefit was methodologically flawed. More seriously, a study in 1999 suggested that the use of aromatherapy for people with Alzheimer's and related dementias should be "abandoned or reassessed", given the degeneration of the brain's olfactory bulb and the loss of smell that accompany these conditions. Other findings suggest that aromatherapy works only if you believe it will. In fact, the only research that has unequivocally shown it to have an effect has been carried out on animals.

And yet, touchy-feely placebo or not, aromatherapy is as popular as ever. In January, for instance, the French newspaper Le Monde reported that a prison-hospital in Fresnes, near Orly airport, had given its blessing to the aromatherapist Marie-Thérèse Esneault, whose attempts to stimulate pleasant memories in mentally ill patients had led to dramatic improvements in their behaviour. It is almost as if doctors, patients and therapists know that in smell they have a powerful means of manipulating emotion and memory, but are using it blindly. Perhaps not for much longer, though, because researchers are now revealing links between olfaction, memory and emotion that could suggest far more precise targets for therapy.

Behavioural studies have consistently shown that odours elicit emotional memories far more readily than other sensory cues. And earlier this year, Rachel Herz, of Brown University in Providence, Rhode Island, and colleagues peered into people's heads using functional Magnetic Resonance Imaging (fMRI) to corroborate that. They scanned the brains of five women while they either looked at a photo of a bottle of perfume that evoked a pleasant memory for them, or smelled that perfume. One woman, for instance, remembered how as a child living in Paris, she would watch with excitement as her mother dressed to go out and sprayed herself with that perfume.

The women themselves described the perfume as far more evocative than the photo, and Herz and co-workers found that the scent did indeed activate the amygdala and other brain regions associated with emotion processing far more strongly than the photograph. But the interesting thing was that the memory itself was no better recalled by the odour than by the picture. "People don't remember any more detail or with any more clarity when the memory is recalled with an odour," she says. "However, with the odour, you have this intense emotional feeling that's really visceral."

That's hardly surprising, she thinks, given how the brain has evolved. "The way I like to think about it is that emotion and olfaction are essentially the same thing," she says. "The part of the brain that controls emotion literally grew out of the part of the brain that controls smell." That, she says, probably explains why memories for odours that are associated with intense emotions are so strongly entrenched in us, because smell was initially a survival skill: a signal to approach or to avoid.

Eric Vermetten, a psychiatrist at the University of Utrecht in the Netherlands, says that doctors have long known about the potential of smells to act as traumatic reminders, but the evidence has been largely anecdotal. Last year, while still at Emory University in Atlanta, Georgia, he and others set out to document it by describing three cases of post-traumatic stress disorder (PTSD) in which patients reported either that a certain smell triggered their flashbacks, or that a smell was a feature of the flashback itself. The researchers concluded that odours could be made use of in exposure therapy, or for reconditioning patients' fear responses.

After Vermetten presented his findings at a conference, doctors in the audience told him how they had turned this association around and put it to good use. PTSD patients often undergo group therapy, but the therapy itself can expose them to traumatic reminders. "Some clinicians put a strip of vanilla or a strong, pleasant, everyday odorant such as coffee under their patients' noses, so that they have this continuous olfactory stimulation," says Vermetten. So armed, the patients seem to be better protected against flashbacks. It's purely anecdotal, and nobody knows what's happening in the brain, says Vermetten, but it's possible that the neural pathways by which the odour elicits the pleasant, everyday memory override the fear-conditioned neural pathways that respond to verbal cues.

According to Herz, the therapeutic potential of odours could lie in their very unreliability. She has shown with her perfume-bottle experiment that they don't guarantee any better recall, even if the memories they elicit feel more real. And there's plenty of research to show that our noses can be tricked, because being predominantly visual and verbal creatures, we put more faith in those other modalities.

In 2001, for instance, Gil Morrot, of the National Institute for Agronomic Research in Montpellier, tricked 54 oenology students by secretly colouring a white wine with an odourless red dye just before they were asked to describe the odours of a range of red and white wines. The students described the coloured wine using terms typically reserved for red wines. What's more, just like experts, they used terms alluding to the wine's redness and darkness - visual rather than olfactory qualities. Smell, the researchers concluded, cannot be separated from the other senses.

Last July, Jay Gottfried and Ray Dolan of the Wellcome Department of Imaging Neuroscience in London took that research a step further when they tested people's response times in naming an odour, either when presented with an image that was associated with the odour or one that was not. So, they asked them to sniff vanilla and simultaneously showed them either a picture of ice cream or of cheese, while scanning their brains in an fMRI machine.

People named the smells faster when the picture showed something semantically related to them, and when that happened, a structure called the hippocampus was strongly activated. The researchers' interpretation was that the hippocampus plays a role in integrating information from the senses - information that the brain then uses to decide what it is perceiving.

In December, Herz showed that this kind of hippocampus-mediated trickery can also elicit dramatically different emotional responses. She asked people to sniff inside jars that contained ambiguous odours and had one of two labels attached. One odour was patchouli, which she called either "musty basement" or "incense"; another was a chemical combination she described either as "parmesan cheese" or "vomit". In both cases, the odour stayed the same but people reported either pleasure or disgust depending on which label they saw. They also behaved differently. If they thought the smell was sick, they said they wouldn't go near it; if parmesan, they wanted to eat it.

Not all smells detectable by humans possess this ambiguous quality. But the susceptibility of olfactory perception to distortion by visual or verbal information would seem to suggest an obvious therapeutic intervention, says Herz. Consider a rape victim who suffers a traumatic flashback each time she catches a whiff of the aftershave worn by the rapist - a popular aftershave worn by millions of men. If she could be persuaded by simple exposure to a mislabelled bottle that the smell was not aftershave but a brand of, say, cleaning fluid, it might be possible to recondition her neural pathways and break that association.

One of the problems with assessing the efficacy of aromatherapy is the difficulty of disentangling the effects of the odour from those of the context in which it is given. At the prison-hospital in Fresnes, Esneault receives her patients in a cell decorated to not look like a cell, with colourful fabrics, photos, musical instruments, flowers and ornaments. She herself is a sympathetic presence. Yet listen to her talk and you recognise the same phenomenon that Herz, Vermetten and others describe. "When people smell a particular odour, they are immediately overcome by emotion and pleasant memories from their childhood," says Esneault. "That strengthens their sense of self, and from that stems a sense of responsibility."