Cosmetics & Personal Care

Odour objects – progress in understanding how we smell

By Thomas Nowotny, University of Sussex

Thomas Nowotny, from the University of Sussex, outlines recent research on how the brain interprets natural odours, which are comp

Thomas Nowotny, from the University of Sussex, outlines recent research on how the brain interprets natural odours, which are complex mixtures of a large number of different compounds
Olfaction – our sense of smell – is arguably the oldest sensory modality, going back as far as simple chemo-reception in single cell organisms. In terms of our understanding of it we are, however, much behind vision and hearing. The most prominent issue is the problem of the vast size of chemical space and the lack of organizing principles in it. While light can be described precisely by wavelength and amplitude, odours consist of volatile chemicals, mostly organic compounds, that can be described by thousands of chemical descriptors. Even though some progress has been made, it is yet unknown which of these descriptors are useful to reflect the nature of odours in order to describe their perceptual qualities.
The second problem, which I will focus on in this article, is that natural odours are typically mixtures of a large number of different chemical compounds and that these mixtures disperse in our environment in very complex spatial-temporal structures, so called odour plumes. A typical natural odour has dozens of components that are above detection threshold for an animal nose. Coffee smell, for instance, has more than 1,000 individual odorants, of which some 20–30 are essential to achieve a realistic aroma.1 When released into the air, this mixture disperses wildly in small filaments,2 much like the smoke above a candle that has been blown out (Figure 1). For animals, this means that smells don’t arrive in an orderly fashion but come occasionally as short whiffs, interspersed with pockets of clean air.
Nowotny Figure 1 2019-05.png

If more than one source of odour is present, which typically is the case, filaments from different sources arrive together or separately, in – for lack of other words – a big mess. This may seem daunting but our collaborators at the University of Konstanz recently discovered that the olfactory perception of insects is so fast (unlike anything previously thought) that the animals can take advantage of the strong temporal structure of odour plumes rather than being flustered by it.
In brief, Szyszka and colleagues3 investigated whether bees could recognise a previously rewarding odour within a mixture with a novel odour. Cunningly, they varied how the odours were mixed, presenting them either at the same time, or delaying one of the odours for a very short time. They found that bees were able to distinguish “synchronous” from “asynchronous” mixtures and recognised the rewarding odour better in the asynchronous situation. Surprisingly, when they started to probe for the minimum time difference that could be resolved, bees were able to distinguish tiny 6 millisecond delays from the synchronous condition. Using computational models of the bee brain, we were subsequently able to show that this observation is consistent with a well-known brain motif of mutual inhibition and so-called winner-take-all dynamics in the antennal lobe, the primal brain centre for olfactory information processing in the bee brain.4
This research has opened up a new view on the complex nature of odour plumes and how they are perceived by animals. Rather than adding complexity, the nature of odour transport in complex odour plumes can be an aid for animals to distinguish odours from separate sources, and so segment the odour percept into separate “odour objects”. We hypothesize that similar processes happen in the human brain and underlie our ability to interpret the chemicals that enter our nose when we go to a coffee shop to be the smell of coffee and bacon, rather than “coffeebacon”, “baconcoffee” or an entirely meaningless mixture of dozens of chemical compounds. Our research into odour object recognition has continued in the “odour object” project (, in which we also discovered that mixtures of odours are more stable percepts than single compounds.5 Our next task will be to investigate the pre-processing of odour percepts in sensory organs, in particular by non-synaptic interactions within the sensilla (sensory hairs) on insect antennae.
  1. Sunarharum WB et al. Food Res Int 2014;62:315–25.
  2. Murlis J et al. Physiol Entomol 2000;25:211–22.
  3. Szyszka P et al. PLoS ONE 2012;7:e36096.
  4. Nowotny T et al. Brain Res 2013;1536:119–34.
  5. Chan HK et al. PLoS Comput Biol 2018;14:e1006536.
Thomas Nowotny, Sussex Neuroscience, School of Engineering and Informatics, University of Sussex, Falmer, Brighton BN1 9QJ, UK