On the intricacies of control over local choice by time, local food probabilities and brief stimuli

Abstract

The three experiments of the current thesis examined the degree to which choice for two concurrently-available options would correspond with local variations in the availability of reinforcement at these options across time since food delivery. Experiment 1 and Set 1 of Experiment 2 showed that arranging differential reinforcement contingencies on the basis of elapsed time since the most recent reinforcer alone did not produce a pattern of choice that corresponded with the local contingency changes. This might because discrimination and, therefore, control by differential reinforcer contingencies, worsens over elapsed time as predicted by well-established models of discriminative stimulus and temporal control. The modelling approach adopted in the current thesis was consistent with and predicted worsening discrimination over increasing time, therefore, providing a good fit. Control over choice between two concurrently-available options arranging differential reinforcement contingencies across time since the most recent reinforcer can be enhanced by arranging extinction on both keys up until the contingency-change period and increasing the rate of reinforcement during the time at which the contingencies of reinforcement are scheduled to change (Sets 2 and 3 of Experiment 2, respectively). By enhancing the period of the change in local contingencies of reinforcement, choice is driven towards the option with the higher food-delivery probability, particularly when the change in contingencies is a period in which the frequency of food on both keys is extreme. Arranging exteroceptive stimuli to occur at the time of selected variations in the local availability of reinforcement (Experiment 3) equally enhanced preference and drove choice to the option with the higher likelihood of reinforcement during the stimuli. This result suggested that the accuracy of current estimates of time since food delivery and elapsed time’s relation with local food availability at each option can be increased by manipulations arranged to enhance the discrimination of local contingency changes. The effects of exteroceptive stimuli and food-frequency on local choice ratios were also well captured by the modelling approach adopted in the current thesis, although the latter of these manipulations posed some problems for the model. Specifically, some of the model’s assumptions, which are shared with many common models of timing, could not justify the degree of the change in local choice ratios at about the time in which local food frequencies were extreme. This finding suggests a potential limit to the modelling approach adopted in the current thesis.