Predictive Modeling of Flavor Compound Formation in the Maillard Reaction: A SWOT Analysis

The importance of Maillard flavor compounds for food quality is undisputed, but we are far from being able to control such formation quantitatively in food processing. Kinetic models attempt to predict rates of formation as a function of temperature, pH, water activity/content, and chemical reactivity. There is, however, a gap between mechanistic models for well-defined model systems, and their application to food, where we are left mainly with empirical models. This paper addresses the strengths and weaknesses of the current models in their ability to predict Maillard flavor compounds quantitatively, the opportunities to progress with these models to make the step to real foods, and the threats that models are simplifications by their very nature. Topics addressed are the usefulness of uniresponse versus multiresponse models, mechanistic versus empirical models, model systems versus real foods. A combination of kinetics and thermodynamics is proposed as direction for future research