Choice, time and food: continuous cyclical changes in food probability between reinforcers

Abstract

The current experiment examined the degree to which locally varying food probabilities on two keys across time since food presentations can continue to control choice until the next food delivery. In two sets of conditions, the probability of food delivery being made available on one key relative to the other key varied sinusoidally across a 1-min period following each food delivery. In Set 1, food-probability changes were unsignaled and the number of cycles per min was varied across conditions. In Set 2, there were always two complete cycles of the sinusoid in the 1-min period, and brief key-color changes were arranged at a selection of fixed times since food delivery to signal portions of the sinusoid. In Set 1, control of choice by local probability of food on each key decreased over time since food delivery. Control by local food probabilities was greater in conditions that arranged fewer cycles per min. The onset of stimulus changes in Set 2 led to a transient reinstatement of local control by food probabilities regardless of the portion of the sinusoidal variation in food probabilities signaled by the stimuli. However, in conditions where the same colored stimuli signaled different portions of the sinusoidal variation in food-delivery probabilities, stimulus changes attenuated joint control by elapsed time and food-probability values. These results suggest that, changing relative food probabilities and stimuli can direct preference toward the likely location of the next food delivery across time since a food presentation, although the degree to which control over choice will be maintained across elapsed time depends on how experimenter-arranged contingencies are mapped onto elapsed time.

Control by reinforcers across time and space: A review of recent choice research

“…When the discrimination is more difficult (e.g., same key sooner, other key later, reinforcer), control lasts for only part of the interreinforcer interval. Miranda‐Dukoski, Davison, and Elliffe () demonstrated a similar, but more transient, effect when the interreinforcer contingency changes were more complicated.…”

“…Control is strong where changes in the likely availability of a reinforcer occur soon after the marker event (the reinforcer), as do the changes produced by a changeover delay, and local choice closely matches the local reinforcer ratio. The further in time from a marker event these changes occur, or the more complex the relation between elapsed time and the likely availability of a reinforcer, the weaker the control by the reinforcer differential (e.g., Cowie et al, ; Miranda‐Dukoski et al, ). These discriminative effects of individual reinforcers are consistent with the well‐known results from performance on single schedules (the Fixed‐Interval, FI, scallop, the Fixed Ratio, FR, break‐and‐run pattern, priming the interval or ratio; Ferster & Skinner, ; Schneider, ), in that they highlight the importance of the discriminative properties of the reinforcer as a time marker.…”

Being there on time: Reinforcer effects on timing and locating

“…Categorical stimulus control would not be expected when there are progressive changes in reinforcer availability across time (e.g., Cowie et al, ; Miranda‐Dukoski, Davison, & Elliffe, ), and scalar timing should become more likely under these conditions. Models of timing tend to assume that one or other of these types of discrimination occurs for all discriminations—for example, SET employs a threshold‐based mechanism based on differences between counts in an accumulator and a memory store, whereas LeT employs a continuous process based on the strength of couplings and the flow of activation.…”

Training reinforcement rates, resistance to extinction, and the role of context in reinstatement

“…As a consequence, the reinstating reinforcers in the Non-Target components produced patterns of responding that were not similar to those obtained during the previous baseline condition in the Target components. Such discriminative control over responding has been implied in Pavlovian conditioning (see Rescorla, 1972, for a review), and, more recently, with local effects of reinforcement on operant choice behavior (e.g., Boutros, Elliffe, & Davison, 2011;Cowie, Davison, & Elliffe, 2011;Miranda-Dukoski et al, 2014).…”

“…Component length was exclusive of each 2-s reinforcer. Before detailed analyses of the data we assessed the stability of baseline response rates in the Rich and Lean components by running a version of Davison's (1972) stability analysis (see also MirandaDukoski, Davison, & Elliffe, 2014). We considered response rates in the Rich and Lean components for any pigeon stable once the median rate of responding across successive blocks of three sessions did not differ by more than 5% from the median rate of responding across the previous block of three sessions.…”