RESEARCH ARTICLE Key odorants or key associations? Insights into elemental and configural odour processing Sébastien Romagny | Gérard Coureaud | Thierry Thomas‐Danguin Centre des Sciences du Goût et de l'Alimentation, INRA, CNRS, AgroSup Dijon, Université Bourgogne Franche‐Comté, F‐ 21000 Dijon, France Correspondence Thierry Thomas‐Danguin, Centre des Sciences du Goût et de l'Alimentation, INRA; 17 rue Sully, 21000 Dijon, France. Email: thierry.thomas‐danguin@inra.fr Funding information French Ministry of Higher Education and Research; FEDER (European Regional Devel- opment Fund); Regional Council of Burgundy Abstract Determining whether odorants can be perceived in an odour mixture or whether the mixture smells different from its components remains challenging. Even in highly complex mixtures, the odour qualities of some elements can be perceived; thus, their identity is conserved within the mixture. Such elements are considered key components and support the elemental perception of the mixture. The concept of key components is also related to elements that do not necessarily carry the odour quality of the mixture but that induce a change in overall mixture perception when they are omitted. In this case, mixture perception often relies on configural processing. To disentangle these multiple aspects of these so‐called key odorants, we sought to study the perceptual role of odorants included in mixtures that are elementally or configurally perceived. Two mixtures, known to be processed configurally and elementally and containing the same 6 odorants in different proportions, were used as references in 4 similarity‐rating experiments. A total of 246 participants evaluated the similarity between the references and single odorants or mixtures of 2 to 6 odorants. This procedure aimed to evaluate whether single odorants or com- binations of odorants can evoke the odour quality of the mixtures. Overall, the results highlighted that elemental perception depended primarily on the odour quality and concentration ratio of many of the mixed odorants, whereas configural perception depended on specific associations of odorants in strict concentration ratios. These results led us to reconsider the impact of key ele- ments in odour mixtures within the framework of a perceptual model stating that elemental per- ception of odour mixtures relies on perception of key odorants, the perceptual features of which are still perceived within the mixture, whereas configural perception relies on key associations of odorants that lose their individual identity when mixed at specific concentration ratios. KEYWORDS human, perception, odour mixture, odour quality, complexity, concentration ratio 1 | INTRODUCTION Our everyday living environment is composed of a myriad of volatile chemicals, many of which provide odour cues, from food, plants and anthropic activities. To process such complexity, organisms can per- ceive mixtures of odorants through two non‐exclusive strategies: ele- mental or configural processing. 1-4 Elemental perception enables an organism to identify the different components of a mixture. Con- versely, the odorants of a mixture can blend and give rise to the configural perception of a novel odour that is distinct from the odour of each individual odorant. 5 These two processing strategies can be influenced by individual‐related factors such as learning, 6-11 adapta- tion, 12 and psychological state. 13,14 However, these processing paths are also driven by the composition of the stimulus, especially the odour quality of each of the odorants and their relative concentrations. 4 The chemical and perceptual natures of an odorant, which lead to its odour quality, are among the most important factors in determining whether an odour mixture is elementally or configurally perceived. These factors are thus prevalent in discrimination and generalisation experiments with animals, where the response to odour mixtures dif- fers from the response to the individual odorants. 3,10,11,15-20 Configural processing was thus found to depend on the specific odor- ants included in the mixture. For instance, a series of experiments with newborn rabbits have shown that animals were not able to generalise their behavioural responsiveness to the odour of ethyl isobutyrate or the odour of ethyl maltol in a binary mixture (30:70 v/v ratio), whereas Received: 21 December 2016 Revised: 4 July 2017 Accepted: 7 September 2017 DOI: 10.1002/ffj.3429 Flavour Fragr J. 2017;1–9. Copyright © 2017 John Wiley & Sons, Ltd. wileyonlinelibrary.com/journal/ffj 1