A neural model of time to toxin production by non-proteolytic clostridium botulinum

Clostridium botulinum is a bacterium present in the raw ingredients of many foods. It produces a powerful neurotoxin as part of its growth process, that can prove fatal when doses as small as a few micrograms are consumed. It is therefore vital to be able to accurately determine the food processing and storage conditions where toxin production is prevented. This paper describes a model of the effect of heat treatment and subsequent incubation temperature on time to toxin production by non-proteolytic C. botulinum. A multi-layer perceptron network is trained, using a cross-entropy error metric, to estimate the a posteriori probability of toxin production given the heat treatment applied and incubation time and temperature. The results obtained compare favourably with those obtained using a conventional model based on a quadratic response surface