Relevant psychological dimensions in the perceptual space of perfumery odors

Odor databases are compilations of odor descriptions based on semantic or numeric methods. One of the largest databases of numeric odor profiles was obtained by Boelens and Haring [Boelens, H., & Haring, H. G. (1981). Molecular structure and olfactive quality. Bussum, The Netherlands: Internal Report, Naarden International] from a panel of perfumers. Each panelist smelled 309 compounds and rated the smell similarity to 30 standards. Most of these references were perfume raw materials. Hence, this database contains valuable information to understand the underlying psychological dimensions in the perceptual space of perfumery scents. A principal component analysis was applied to this database. The first principal component (PC1) was interpreted as a dimension of freshness, and the second component (PC2) discriminated feminine versus masculine cosmetic odors. The hedonic dimension did not clearly show up, though different studies have revealed that it is the most salient dimension if a wide range of odors is assessed. The loading plot corresponding to PC1–2 explains 32% of the data variance and provides a low-dimensional representation of olfactory perception space that may lead to the development of meaningful sensory maps for perfumery odors.