Investigating Temporally Dynamic Models of Choice in Rats and Bumblebees

Temporally dynamic models of choice simulate how animals use past experience to make optimal decisions in variable environments. Chapter 1 briefly reviews a series of such modeling approaches and their mathematical properties. Extension of these models to longer experiments (more than 48 hours) designed to produce spontaneous recovery of choice (a behavioral phenomenon implicated in relapse) previously led to results not described well by current models. Chapter 2 comprises an empirical investigation of different explanations for these inconsistent results using rats, evaluating the effect of different test delays on preference following varied reward conditions. The results of the first two experiments ruled out the possibility of time’s passage alone causing animals to revert to exploratory behavior and suggested such an effect depends on reward variability only at the end of training. A third experiment aimed to distinguish a drift toward exploratory behavior produced by time’s passage can be distinguished from the effect of spontaneous recovery of choice, but produced inconclusive results, leading to a discussion of limitations of the current experimental preparation, alternative modeling approaches, and potential directions for further investigation. Chapter 3 provides a comprehensive review of literature on the suitability of bumblebees for choice research, then details an experimental investigation of the occurrence of spontaneous recovery of choice in B. impatiens. Results from this experiment suggested spontaneous recovery of choice did not occur. Explanations for these results and their generalizability are discussed. Chapter 4 summarizes overall conclusions and directions for future research.