CONDITIONED REINFORCEMENT DYNAMICS IN 3-LINK CHAINED SCHEDULES

In two experiments rats were trained on three-link concurrent-chains s chedules of reinforcement. In Experiment 1, additional entries to one terminal link were added during one of the middle links to a baseline schedule that was otherwise equal for the two chains, and, depending o n the condition, these additional terminal-link presentations ended ei ther in food or in no food. When food occurred, preference was always in favor of the chain with the additional terminal-link presentations (which also entailed a higher rate of reinforcement). When no food occ urred at the end of the additional terminal links, the outcome depende d on the nature of the stimuli associated with these additional termin al links. When stimuli different from the reinforced baseline terminal links were used for the no-food terminal links, preference was agains t the choice alternative that led to the extra periods of extinction. When the same stimulus was used for the mio kinds of terminal links, p reference was near indifference, that is, significantly greater than w hen different stimuli were used. Tn Experiment 2, rats learned repeate d reversals of a simultaneous discrimination under a three-link concur rent-chains schedule, in which the food or no-food choice outcomes wer e delayed until the end of the chain. Different conditions were define d by the point in the chain at which differential stimuli occurred. Wh en the middle and terminal links provided no differential stimuli, dis crimination was acquired more slowly than when differential stimuli oc curred in both links. When differential stimuli occurred in the middle but not the terminal links, acquisition rates were intermediate. Both experiments together show that the effects of stimuli in a chain sche dule are due partly to the time to food correlated with the stimuli an d partly to the time to the next conditioned reinforcer in the sequenc e.