Explaining taste association of odorants by multivariate statistical analysis - INRAE - Institut national de recherche pour l’agriculture, l’alimentation et l’environnement Access content directly
Conference Poster Year : 2019

Explaining taste association of odorants by multivariate statistical analysis


Odour taste association relies on a cognitive process based on previous experience and associative memory [1]. This association between odour and taste has been successfully applied to enhance taste perception in foods with low sugar [2] or low salt [3] content. However, the selection of odour-active molecules able to enhance taste perception remains a crucial step. Previous results have shown that the associated taste of odorants smelled from a fruit juice extract can be determined using the new concept of gas chromatography/olfactometry-associated taste (GC/O-AT) [2]. Using this technic, a total of 18 molecules from a multifruit juice extract have been described with odour descriptors by classical GC/O and then with taste associated descriptions (sweet, salty, bitter, sour) by GC/O-AT, with their detection frequencies (DF%). Here, we searched for links between the taste descriptors and the odour descriptors given by the 12 panellists during the GC runs. In order to analyse the relationships between odour and taste description, we performed a computational analysis to build a network based on co-occurrence matrix of odours and tastes. We applied a multidimensional scaling (MDS) approach previously developed for odours and odorants [4], to project the 18 molecules in the odour space reduced to the main dimensions. The results showed that sweetness was mainly associated with fruity odour and, to a lesser extend, to sweet or floral odours. Some odorants with floral odour were also associated with bitterness, while some odorants described as sweet were associated with sourness and others described as fruity to saltiness. MDS projection of the odorants on the three main dimensions (V1-V3) showed that V1 separated odorants according to fruity odour, while V2 and V3 separated odorants according to sweet and floral odour respectively. Seven molecules with both fruity and sweet odours, which have the highest DF (>42%) for associated sweetness perception, were close to each other in the V1/V2 plan. However, three other molecules with high DF for sweetness were outside this group: (E)-β-ionone (42%) described as floral-fruity, not sweet, and associated to saltiness (33%), probably due to its spicy-plastic odours; (E)-β-ocimene (75%) described as fruity-floral not sweet and furaneol (42%) described only as sweet. Overall, our approach provided a visualisation tool for a better understanding of the relationships between odour descriptors and associated taste description and could be used to predict a potential application of odour-active molecules that enhance taste perception.
Fichier principal
Vignette du fichier
2019_guichard_wartburg_symp_1.pdf (1.17 Mo) Télécharger le fichier
Origin : Files produced by the author(s)

Dates and versions

hal-02786016 , version 1 (04-06-2020)


  • HAL Id : hal-02786016 , version 1
  • PRODINRA : 474491


Elisabeth Guichard, Carmen Barba, Thierry Thomas-Danguin, Anne Tromelin. Explaining taste association of odorants by multivariate statistical analysis. 12. Wartburg symposium on flavour chemistry and biology, May 2019, Eisenach, Germany. 1 p., 2019. ⟨hal-02786016⟩
20 View
9 Download


Gmail Facebook Twitter LinkedIn More