Odorant Chemical Structure Can Predict Changes in Olfactory Perception Ratings
Mentor 1
Adam Greenberg
Start Date
10-5-2022 10:00 AM
Description
The human olfactory system uses chemosensation whereby chemical odorants bind to G-protein coupled receptors on odor receptor neurons (ORNs). Individual odor percepts are represented by the pattern of activity across the population of ORNs. Thus, many models of olfaction link the chemical structure of odorants with olfactory perception. Despite this link, direct evidence for a causal role of odorant chemical structure leading to perceptual changes has not yet been established. One problem in the extant literature is a somewhat oversimplification of the physicochemical-perceptual relationship. This study aims to explore more complex and accurate linkages between odor perception and chemical structure using exploratory computational methods. An extensive list of commonly used odors was compiled and 14 chemical features were identified for each odor. These data were then analyzed using principal component analysis (PCA) to identify latent variables that may better explain the relationship between chemical structure and olfactory perception. Using two large bodies (N1 = 180, N2 = 100) of behavioral olfactory perception data, ratings of odorants were compared to PCA output. Results showed that the first principal component was strongly related to molecular size and weight (r = 0.91), while the second component was related to ring structure (r = 0.87; i.e. molecular complexity). Analysis of Euclidean distances between pairs of odors in the multidimensional principal component space also revealed significant positive correlations with pleasantness and intensity ratings. This analysis provides a basis upon which further predictions can be made regarding how olfactory perception is related to odorant chemical structure.
Odorant Chemical Structure Can Predict Changes in Olfactory Perception Ratings
The human olfactory system uses chemosensation whereby chemical odorants bind to G-protein coupled receptors on odor receptor neurons (ORNs). Individual odor percepts are represented by the pattern of activity across the population of ORNs. Thus, many models of olfaction link the chemical structure of odorants with olfactory perception. Despite this link, direct evidence for a causal role of odorant chemical structure leading to perceptual changes has not yet been established. One problem in the extant literature is a somewhat oversimplification of the physicochemical-perceptual relationship. This study aims to explore more complex and accurate linkages between odor perception and chemical structure using exploratory computational methods. An extensive list of commonly used odors was compiled and 14 chemical features were identified for each odor. These data were then analyzed using principal component analysis (PCA) to identify latent variables that may better explain the relationship between chemical structure and olfactory perception. Using two large bodies (N1 = 180, N2 = 100) of behavioral olfactory perception data, ratings of odorants were compared to PCA output. Results showed that the first principal component was strongly related to molecular size and weight (r = 0.91), while the second component was related to ring structure (r = 0.87; i.e. molecular complexity). Analysis of Euclidean distances between pairs of odors in the multidimensional principal component space also revealed significant positive correlations with pleasantness and intensity ratings. This analysis provides a basis upon which further predictions can be made regarding how olfactory perception is related to odorant chemical structure.
