Behavioral variability produced by two reinforcement schedules

Abstract

The aim of this study was to examine how different irenforcement contingencies affect behavioral variability. Fourteen rats were required to press a right or lever to complete four-response sequences. Some animals were reinforced if the sequence pattern was different from the patterns of each of the last four sequences (LAG) For other rats, the probability of reinforcement increased with the number of switching responses within a sequence (ALT). Finally, another group of rats received the same reinforcement distribution as the other animals, regardless of their sequence pattern or the number of switching responses in a sequence (YOKE). After 40 sessions, both the LAG and the ALT contingencies produced high behavioral variability levels, but only the LAG contingency maintained these high levels until the last session. The Yoke contingency produced repetitive sequences. The results confirm the idea that sequence variability is not produced by intermittent reinforcement: They also suggest that for variability to be maintained it must be directly reinforced; it is not enough to reinforce switching behavior. The data are consistent with the hypothesis that behavioral variability is an operant.