Sensory Analysis and Mathematical Modelling Drive Improved Stevia Taste

Dr. John Fry, international-acknowledged expert on high-potency sweeteners, presented “Application and Innovation in Stevia and Taste Development: Improved leaf extracts from advanced sensory study” at the International Union of Nutritional Sciences (IUNS) International Congress of Nutrition (ICN), held in Buenos Aires, Argentina October 15-20, 2017. Discoveries he described are now helping to create today’s stevia sweeteners with greatly improved taste. Read on as Dr. Fry explains important research findings, in addition to more on advances made in the taste profile of one of the fastest growing sweeteners today.

BY: Dr. John Fry

Key Takeaways:
• Multiple ways to characterize and develop unique steviol glycoside blends for superior taste
• Using sensory analysis and mathematical modelling as tools to guide blend choice
• Optimized blends deliver superior taste with deep sugar reduction in both model systems and key applications

Charting earlier uses of stevia in consumer products

In 2008, the first commercial steviol glycoside sweetener in Western markets was high purity rebaudioside A (reb A). This was widely thought at the time to be the best-tasting of the main leaf glycosides. It was soon apparent that reb A – and other steviol glycosides also had non-sweet side tastes. Typically described as bitter or liquorice, these were particularly noticeable at higher usage levels. Such challenging taste qualities, coupled with the relative expense of reb A, threatened to limit the use of this new ingredient.

Meanwhile, in using lower purity stevia leaf extracts such as RebA-80 (80% purity) manufacturers discovered that they could taste better than pure Reb A (e.g., RebA-99, Reb 97). This lead to the investigation and greater understanding of the minor steviol glycosides and their contribution to the taste of stevia.

Approach I

The first study used highly purified individual steviol glycosides. Eleven compounds were assessed for attributes such as taste recognition threshold and sweet and bitter concentration-response curves.

Ultimately, three concentration-response curves were measured for each glycoside: sweetness referred to sucrose, bitterness referred to caffeine, and liquorice referenced to a standardised liquorice extract – the first time the latter two calibrations had been attempted for steviol glycosides.

The enormous program of tasting demanded high-throughput sensory methods. These involved a body of about 100 trained panellists who, working over two years, carried out difference, descriptive, threshold and quantitative work.

Identifying a superior taste profile

The resulting concentration-response data were the first input to a predictive mathematical model. Once individual glycosides had been assessed, binary mixtures were created and tested again, this time particularly looking for synergistic interactions that might enhance sweetness and/or reduce the undesirable side tastes. Ternary and higher order mixtures were similarly investigated.

The resultant refined model helps identify glycoside mixtures of superior taste. For example, one output is color-coded “maps” showing all possible combinations of various glycosides, and highlighting those areas where sweetness is enhanced or undesirable side tastes reduced.

These model visualisations have further uses. Similar to geographical mapping, taste contour lines can be drawn. The “finding the sweet spot” chart shows three such plots, one each for sweetness, bitterness and liquorice, overlaid on each other. The easily-seen highlighted area predicts the glycoside compositions with maximum sweetness intensity and minimum side tastes.

Source: Carlson et al Cargill Inc, US Patent application 20150237898

The approach is not limited to three-way blends, but the picture becomes more complicated as more components are added. Ultimately, two-dimensional visualisation of more than four-component blends becomes impossible, and the output of the model is then still as valuable but purely mathematical.

By combining contour plots in this way, the very large and complex array of potential mixtures could be reduced to a small number of blends likely to exhibit the very best taste properties. The model was first validated by comparing its predictions with the properties of some known blends. For example, the model correctly predicted the identical sweetness concentration response curves of reb A and RA80. Other glycoside blends also performed as predicted.

Source: Carlson et al Cargill Inc, US Patent application 20150237898

Helping Reduce Sugar by 75% While Keeping Taste Quality

Subsequently, using the indications of the model, glycoside blends of potentially superior taste quality were identified and their taste properties verified. Several high-performance quaternary blends were found.

Some of these synergistic extracts have been commercialised. To produce them, there is no need to isolate the individual glycosides as was done for the research. Instead, different leaf extracts are carefully analysed and combined to give the key glycosides in the correct ratios.

Source: Connect Counseling based on data from Carlson et al Cargill Inc, US Patent application 20150237898

Such extracts have much reduced side tastes – so they can be used in higher concentrations, permitting greater levels of sugar reduction. For example, in a lemon lime carbonated soft drink, the maximum acceptable sugar reduction was about 50% with reb A. In contrast, a synergistic leaf extract could be used to achieve 75% sugar reduction with little change in quality.

Source: Carlson et al Cargill Inc, US Patent application 20150237898

Approach II

In an alternative approach, a design of experiment platform was used to first screen leaf extracts of single and combination steviol glycosides in specific applications. The data from this screening was applied to determine the optimal combination of glycosides to achieve the sweetness and sensory attributes closest to target taste profile. A descriptive analysis panel provided a description of key sensory attributes of the experimental design prototypes in finished food and beverage application and statistical significance of attribute difference was determined. Principle component analysis (PCA) was used to visually compare the control prototype to the optimal and design samples.

The optimized blend of glycosides identified by the design of experiment platform outperformed Reb A as predicted in the model. In the examples below, the deep sugar reduced chocolate milk and no sugar added yogurt performed best with the optimized blends. These solutions both showed a significant improvement in bitterness and overall liking compared to the single glycoside, Reb A.

Caption: PureCircle Proprietary Research

In Conclusion

These studies show substantial investment in sensory and mathematical analysis has driven the development of today’s synergistic mixtures of steviol glycosides with superior taste that allows greater sugar reduction than ever.

For more information about stevia, contact the International Stevia Council or the Calorie Control Council.

About Dr. John Fry

John Fry is an internationally-acknowledged expert on high-potency sweeteners. Since 1997 he has directed Connect Consulting, one of the world’s foremost technical resources for sweetener manufacturers and users. He speaks and trains widely on sweeteners, sweetness and calorie-control.

Previously, John was Director of Scientific & Technical Services at Holland Sweetener Company, before which he managed the Science Group at Leatherhead Food Research.

John has a BSc and PhD in Food Science from Leeds University. He is also a Chartered Chemist and holds Fellowships of the Royal Society of Chemistry, the Institute of Food Science & Technology and the British Society of Flavourists.

The International Stevia Council and the Calorie Control Council sponsored Dr. Fry’s presentation.
@Copyright 2017 – 2018

The post Sensory Analysis and Mathematical Modelling Drive Improved Stevia Taste appeared first on

faq2Do you have questions about low-calorie sweeteners? Want to learn more about maintaining a healthy lifestyle? You asked and we listened. Our resident Registered Dietitians answered the most popular questions about low-calorie sweeteners.