Scientists Create a Map of Smell Similarity Based on Neural Activity

A mathematical model untangles the complexities of comparing odor molecules

It is easy to see that red is closer to pink than to blue, but odors are harder to compare: Do almonds smell more like roses or bananas? According to a “smell map” created by researchers at the Weizmann Institute of Science in Israel, almonds smell like roses—and the two scents elicit similar neural activity.

Led by neurobiologist Rafi Haddad, the team identified 40 defin­ing characteristics for odors, such as molecular shape and structure, then translated the resulting profiles of 450 scents into locations on a multidimensional map, as described in a May paper in Nature Methods. In the same way that similar colors are closer together on a rainbow, similar smells are located near one another in the 32-dimensional mathematical model. A scent’s location on the map also predicts the brain activity caused by getting a whiff: previous research in a variety of animals such as fruit flies, honey­bees, mice, rats and tadpoles showed that neighboring odors cause similar patterns of neuron activity. Based on these patterns, the researchers were able to accurately predict the neural signature of formerly untested scents.

The findings could help illuminate the laws that underlie our sense of smell, which are largely unknown and difficult to study, according to Haddad. The smell map might also aid in the study and prediction of animal behavior by illuminating which scents an animal considers good or bad.

Editor's Note: This story was originally printed with the title "Smell Similarity"

or subscribe to access other articles from the August 2008 publication.
Digital Issue $7.95
Digital Subscription $19.99 Subscribe
Share this Article:


You must sign in or register as a member to submit a comment.

The Pi Day Commemorative Package

Get 3 of our best-selling Pi topic issues
Plus a FREE Bonus Issue!

Add to your cart now for just $9.99 >


Email this Article