An Equiratio Mixture Model for non-additive components: a case study for aspartame/acesulfame-K mixtures

The Equiratio Mixture Model predicts the psychophysical function for an equiratio mixture type on the basis of the psychophysical functions for the unmixed components. The model reliably estimates the sweetness of mixtures of sugars and sugar-alcohols, but is unable to predict intensity for aspartame/sucrose mixtures. In this paper, the sweetness of aspartame/acesulfame-K mixtures in aqueous and acidic solutions is investigated. These two intensive sweeteners probably do not comply with the model's original assumption of sensory dependency among components. However, they reveal how the Equiratio Mixture Model could be modified to describe and predict mixture functions for non-additive substances. To predict equiratio functions for all similar tasting substances, a new Equiratio Mixture Model should yield accurate predictions for components eliciting similar intensities at widely differing concentration levels, and for substances exhibiting hypo- or hyperadditivity. In addition, it should be able to correct violations of Stevens's power law. These three problems are resolved in a model that uses equi-intense units as the measure of physical concentration. An interaction index in the formula for the constant accounts for the degree of interaction between mixture components. Deviations from the power law are corrected by a nonlinear response output transformation, assuming a two-stage model of psychophysical judgment.