Reward Shifts in Forced-Choice and Free-Choice Autoshaping With Rats

Successive negative contrast (SNC) involves a disruption of behavior when the paired reward is unexpectedly reduced from a large to a small amount, relative to a control always receiving the small amount. Five experiments with rats explored SNC in the Pavlovian autoshaping procedure in which a retractable lever is paired with the delivery of food pellets. In Experiment 1, a 12-to-2 pellet downshift either early in training (after 3 sessions) or late in training (after 20 sessions) yielded no significant evidence of SNC either in terms of lever pressing or goal entries. Experiment 2 showed that presession feeding (another form of reward devaluation) suppressed lever pressing in nonreinforced tests early in training. However, no statistical evidence of lever pressing suppression was found late in training. Presession feeding also suppressed lever pressing late in training if the test session included reinforcements. Experiment 3, using reward downshift, showed that adding a nontarget lever produced no statistical evidence of response suppression to the target lever during the downshift. Lever pressing to the target lever increased and goal entries tended to decrease after a 12-to-2 pellet downshift. Using a within-subject design and two target levers with distinct reward values (Experiment 4), reward downshift produced evidence of lever pressing enhancement in single-lever trials, but lever pressing suppression and a switch in preference to the unshifted lever in nonreinforced free-choice trials. Experiment 5 replicated these within-subject SNC effects, but found only modest evidence for a successive positive contrast effect in free-choice behavior. These results suggest that autoshaping in rats may induce response invigoration in forced-choice situations, but response suppression in free-choice situations.