Stochastic modelling of odour detection behaviours from trained parasitic wasp using video image tracking software

Microplitis croceipes, a nectar-feeding parasitoid wasp, was found to be able to associatively learn chemical cues using classical conditioning to food. The experiments on M. croceipes were performed and recorded with a Sony camcorder in the USDA-ARS Biological Control Laboratory at Tifton, GA, USA. The experimental videos in this study have shown that M. croceipes can respond to coffee odour when associated with food. The detection capabilities and behaviour of M. croceipes at different levels of coffee odour were studied. First, the data that related to trained M. croceipes behaviour were extracted from the experimental videos and stored in a Microsoft Excel database. The extracted data represented the behaviour of M. croceipes trained to 20 mg of coffee and then, subsequently, exposed to 5, 10, 20 and 40 mg of coffee. Secondly, indices were developed to uniquely characterise the behaviour of trained M. croceipes under these different coffee concentrations. Thirdly, a stochastic mathematical model and its parameters were developed to mimic and predict the responses of trained wasps when exposed to these four different levels of coffee odours. Finally, additional data analyses were performed to investigate the lateralisation and robustness of trained wasps’ conditioned responses. A hypothesis is proposed based on the experimental findings in this research. In summary, this thesis demonstrates the usefulness of mathematical models for interpreting experimental observations, gaining novel insights into the dynamic behaviour of these trained wasps, the development of hypotheses, as well as understanding the design principles of a simple model based on the behaviour of trained wasps