Taste-aroma interactions as a sugar reduction strategy in flavored milk

Ms. Gloria Wang, M.S. Candidate, Department of Food Science

Date and Location

When (Date/Time)

June 15, 2018, 10:00 AM - 11:00 AM


252 Erickson Food Science Building

Add to calendar


Food manufacturers are actively seeking methods of reducing sugar in products without compromising flavor, functionality, and consumer acceptability. The use of alternative sweeteners is the most common strategy; however, consumers have become wary of these for various reasons. Cross-modal interactions, in particular taste-aroma interactions, has been proposed as another strategy for sugar reduction, as it has been shown that the addition of a congruent aroma to a sweetened matrix may enhance the matrix’s perceived sweetness and vice versa. Currently there are two underlying factors that make this phenomenon difficult to accept as a real occurrence: 1) most studies have been done in model solutions or model foods and not real foods, and 2) enhancement is either present or disappears depending on the sensory testing method and consequently the perceptual strategy used. Enhancements have thus been attributed to what is known as the “dumping effect”: when panelists are given a limited ballot of attributes to rate, they will “dump” any similar perceptions into one category, causing what seems to be an enhancement. When panelists are given an extended ballot of salient product attributes, they are then able to separate these perceptions into individual components, causing enhancements to disappear. However, this method of rating is not representative of a typical eating situation for the average consumer. Instead of this analytical mindset that is employed when rating scales are used, consumers’ perception of food is more often holistic and hedonic. This integration of perceptions is known as the synthetic strategy and methods that encourage such behavior have been proposed as an alternative approach when testing for enhancements in taste-aroma interaction studies. The following seminar outlines studies to answer the following questions: 1) Can enhancements of either aroma or taste be elucidated in a typical rating study of vanilla milk when controlling for responses biases? 2) Can these enhancements be assessed by testing mathematically for synergistic, antagonistic or no interaction effects with the isobole method? And 3) Can sweetness enhancement be found in vanilla milk using a non-scaling testing approach, i.e., a modified ABX test?

The first dose-response study found interaction effects of taste and aroma as measured by the Isobole approach. Vanilla flavor and liking scores were best fit by a second-order model while sweetness, milk flavor, and thickness were best modeled by a first-order model. The additional of vanilla was significant (p<0.05) on the sweetness model and vanilla flavor perception was more pronounced when presented in combination with sugar compared to the samples without. Based on the Isobole method, the skim milk samples containing both vanilla and sugar were found to act synergistically for perceived sweetness at the low to medium sugar
concentrations (I < 1.0).

The second matching experiments, based on a modified ABX test, demonstrated greater interaction between vanilla flavor and sugar in skim milk samples: a greater proportion of panelists matched a vanilla-sucrose milk to a sucrose-milk reference higher in sugar than that of the aroma-spiked sample. This effect was repeatedly observed for both a congruent vanilla flavor and incongruent flavor, although a greater proportion of panelists matched the congruent vanilla-sugar milk to a higher sugar reference than in the case of the incongruent-sugar milk.

Overall, synergistic interactions between vanilla flavor and sugar were found with both scaling (i.e., analytical strategy) and non-scaling (i.e., synthetic strategy) methods in skim milk, with the effect being greater when using a non-scaling strategy. The effect was found in the analytical rating task even after controlling for the dumping effect, over a wide range of concentrations, and using a real food. Thus, taste-aroma interactions seem to be robust and should be further investigated in other real food applications.