Modeling of Sensory Characteristics Based on the Growth of Food Spoilage Bacteria

During last years theoretical works shed new light and proposed new hypothesis on the mechanisms which regulate the time behaviour of biological populations in different natural systems. Despite of this, a relevant physical and biological issue such as the role of environmental variables in ecological systems is still an open question. Filling this gap of knowledge is a crucial task for a deeper comprehension of the dynamics of biological populations in real ecosystems. The aim of this work is to study how dynamics of food spoilage bacteria influences the sensory characteristics of fresh fish specimens. This topic is worth of investigation in view of a better understanding of the role played by the bacterial growth on the organoleptic properties, and becomes crucial in the context of quality evaluation and risk assessment of food products. We therefore analyze and reproduce the time behaviour, in fresh fish specimens, of sensory characteristics starting from the growth curves of two spoilage bacterial communities. The theoretical study, initially based on a deterministic model, is performed by using the temperature profiles obtained during the experimental analysis. As a first step, a model of predictive microbiology is used to reproduce the experimental behaviour of the two bacterial populations. Afterwards, the theoretical bacterial growths are converted, through suitable differential equations, into "sensory" scores, based on the Quality Index Method (QIM), a scoring system for freshness and quality sensory estimation of fishery products. As a third step, the theoretical curves of QIM scores are compared with the experimental data obtained by sensory analysis. Finally, the differential equations for QIM scores are modified by adding terms of multiplicative white noise, which mimics the effects of uncertainty and variability in sensory analysis. A better agreement between experimental and theoretical QIM scores is observed, in some cases, in the presence of suitable values of noise intensity respect to the deterministic analysis.