Common Scents


You are here

Prof. Noam Sobel. Physical laws underly smell perception

Most would agree that roses smell good; old sweaty socks do not. What makes one smell pleasant and another odious? Is our perception of smell based on a physical reality, such as the chemical makeup of odor molecules, or is it a subjective, learned experience?
Scientists at the Weizmann Institute of Science and the University of California at Berkeley have now discovered that knowing the molecular structure of a substance can help predict whether we will find its smell delightful or malodorous.
Other senses can be explained by physical laws. When we see red, for instance, it is because light is reflected off an object in a certain wavelength, while hearing is based on the varying frequencies of sound waves. But until now, there was no known physical factor that could explain how the brain senses odors. The new study conducted by Prof. Noam Sobel of the Institute’s Neurobiology Department and his colleagues, published in the Journal of Neuroscience, demonstrates that real physical laws underlie our perception of smell.
To identify the general principles by which our sense of smell is organized, the researchers began with a database of 160 different odors that had been ranked by 150 perfume and smell experts according to a set of 146 characteristics (sweetish, smoky, musty, etc.). These data were then fed into a statistical program that analyzed the variance in perception among the smell experts. The scientists found that the data fell along an axis – described as the “pleasantness rating” of the odors – running from “sweet” and “flowery” at one end to “rancid” and “sickening” at the other. The distribution along this axis, they discovered to their surprise, closely matches the variation in chemical and physical properties from one substance to another. The researchers found they could build a model to predict, from the molecular structure of a substance alone, how pleasing its smell would be perceived to be.
To double-check their model, Sobel and his team had experimental subjects assess for pleasantness 50 odors they had never encountered before. They found that the ratings of their test subjects fit closely with the rankings shown by their model.
In other words, they were able to predict the level of pleasantness quite well, even for unfamiliar smells. They noted that although preferences among smells are commonly supposed to be culturally learned, their study showed that the responses of American subjects, Jewish Israelis and Muslim-Arab Israelis all fit the model’s predictions to the same extent.
Sobel: “Our findings show that the way we perceive smells is at least partially hard-wired in the brain. Although there is a certain amount of flexibility, and our life experience certainly influences our perception of smell, a large part of our sense of whether an odor is pleasant or unpleasant is due to a real order in the physical world. Thus we can now use chemistry to predict how the smells of new substances will be perceived.” 
Prof. Noam Sobel’s research is supported by the Nella and Leon Benoziyo Center for Neurosciences; the J&R Foundation; and the Ben and Joyce Eisenberg Foundation.