There are many different types of sensory thresholds, with the two most commonly used in wine defined below.

Detection threshold refers to the lowest concentration or weakest stimulus that can be sensed. So if we were adding an aroma compound, Compound X, to wine, it would be the concentration at which compound X would alter the aroma, or make the wine smell different.

A recognition threshold is the concentration or level at which a compound can not only be detected but can also be recognized. So, for example, if the concentration of Compound X is increased until the smell is described as whatever Compound X smells like.

In wine we use sensory thresholds to interpret chemical information regarding what compounds are influencing the perception, mostly aromas, of wine. However, there are some caveats that need to be understood to fully use this information. The first being that the matrix in which the threshold is determined plays a large role in the concentration that is perceived. For example, vanillin, the main compound from vanilla beans and also found in oak aged wines, has a range of sensory thresholds depending on the matrix it is present in. In air the threshold of vanillin is 100 μg/L (Minematsu and Wu, 2017) while in water the threshold of vanillin is 19 g/L (Powers and Shinholser, 1988). In wine however, the threshold for vanillin is 65 ug/L (Chatonnet et al. 1992). This means that if you were using sensory thresholds to predict whether a particular concentration of vanillin would impact the aroma of a product, you would need to confirm what matrix the sensory threshold was originally determined in.

Threshold measurements are determined using a psychometric curve (Figure 1). A psychometric curve is a mathematical model used to describe the physical stimulus (in our case smell) and the human response to the stimulus at various concentrations. The tests typically used to create this curve involve an individual discriminating between samples with the stimulus from the one without, so a yes or no answer. As the concentration of the stimulus increases, the chances of getting the test correct also increase. The lower limits of the curve are answers due to pure chance while the upper limit is if everyone gets the test right every time (a perfect response). Most threshold testing utilizes a triangle test (Meilgaard et al. 1999) where the chance of getting the test right by guessing is 33%. The psychometric curve is normally S shaped and the inflection point, or middle, of the curve is normally the calculated sensory threshold. In most instances this is where 50% of the population answer the triangle test correctly.

Figure 1. Example of a psychometric curve used to determine thresholds (Strasburger 2001).

So what does this mean for sensory thresholds and interpreting wine aroma chemistry data. Well first, it is important to know what matrix the sensory threshold of interest was determined. Most threshold information can be found in water or ethanol solutions. However, a sensory threshold determined in water, or even an ethanol solution, is not very representative of the perception of that compound in a wine. Previous work done in our lab investigating terpenes showed how perception is altered from a simple model wine solution compared to a wine with only taste and mouthfeel compounds compared to an actual wine (Tomasino et al. 2020). Additionally, the sensory threshold is the point at which 50% of the population can smell it. Therefore, there are individuals that can smell that compound at lower concentrations and individuals that cannot smell the compound until it is at much higher concentrations. This is one of the reasons that individuals can have very different aromatic experiences when they are tasting the same wine, and why in sensory research we try to use as many panelists in sensory panels as we can.

Sensory thresholds can be used in many ways but a current area of research in the Tomasino group is determining the recognition thresholds of smoke compounds in different varietal wines. Not only are we looking at different varietals, we are also determining thresholds in different styles of the same varietal such as a lightly extracted Pinot noir versus a heavier extracted Pinot noir wine. The overall aim is to present a concentration range where individuals describe wines as ashy. This is essential to providing target concentrations for key smoke compounds that can be used to develop mitigation practices in the winery and vineyard. — By Dr. Elizabeth Tomasino, Associate Professor, Department of Food Science & Technology, Oregon State University