Preference Reversal between Self-Control and Impulsivity

To date there has been no convincing evidence for preference reversal in pigeons using concurrent-chains procedures in which the delay and amount of outcomes of choices are not confounded with their frequency. Three experiments offered pigeons choices between small and large rewards with differing durations of delay from the time of choice. A concurrent-chains procedure with two different sets of choices, signalled by red or green keys was used in all three experiments. During red trials, the pigeon chose between an immediate small reward and a large delayed reward. During green trials, the delays until reinforcement were longer than the delays in red trials by a constant duration. It was hypothesized that the pigeons would initially prefer the small rewards delivered after short delays for both red and green trials, that is, the pigeons would make impulsive choices. However, it was also predicted that once the delay length increased for the green trials the pigeon’s preference would reverse. The pigeons would be more likely to choose the larger later reward, the self-controlled choice. As both red and green trials occurred within the same session, a within-session preference reversal would result. In Experiment 1, the difference between red and green trials was initially short 3 s, and this increased to a longer duration (10-s) for Condition 2 of the experiment. Four of the five pigeons showed within-session preference reversal, acting impulsively in red trials and self-controlled in green trials. In Experiment 2, the delays until reinforcement were increased or decreased by 1, 2, 4, 6, 8, 10, 12, and 15-s. Pigeons in the ascending group did demonstrate an increase in preference for the larger reward as the delay until reinforcement increased on green trials. This increase was a gradual, steady increase, with the birds becoming more self-controlled as the delay to reward was increased. Pigeons in the descending group did not show this pattern on green trials, and did not become more self-controlled with longer delays. Experiment 3 varied both the delays to reinforcement (as in Experiment 1) and the reward magnitude (reward pairs of 1 versus 4.5-s and 2 versus 3.5-s). This experiment aimed to establish the contribution of the magnitude effect to preference reversal. The pigeons initially preferred the smaller sooner reward for both reward pairs, but when the delay to reinforcement increased to 10-s for the reward pair 1 and 4.5-s the pigeons changed their preference to the larger later reward. This did not occur for the 2 and 3.5-s reward pair, showing a reward dependent preference reversal. This shows a bias caused by the delay ratios; increases in the delay increased sensitivity to reward. Experiment 3 suggests that the magnitude effect plays a role in preference reversal. These results show that pigeons can demonstrate preference reversal using a concurrent-chains procedure, also that the magnitude of the reward can affect preference reversal