Conditioned reinforcing value of the sample stimulus and resistance to change of delayed matching -to -sample performance in pigeons

Pigeons were trained on a two-component multiple schedule with a chain schedule in each component. According to a variable-interval schedule in the initial link, pecks on a center key produced a terminal link with a delayed matching-to-sample trial. A trial began with the onset of a red or green (Experiment 1) or green or blue (Experiment 2) light as a sample stimulus on the center key. The sample stimulus was followed by a retention interval during which all keys were dark, and then by comparison stimuli on the side keys. A response to the comparison stimulus that matched the sample stimulus was reinforced by access to food. The multiple-schedule components differed in terms of parameters that theoretically affect the conditioned reinforcing value of the sample stimulus. In Experiment 1, the value of the sample stimulus in one component was altered by increasing the length of the variable-interval schedule in the initial link of the chain schedule. In Experiment 2, the value was altered by superimposing response-independent food deliveries during the initial link. The conditioned reinforcing value of the sample stimulus was assessed on occasional probes composed of a concurrent-chains schedule. The initial links were equal variable-interval schedules. The terminal links composed of the presentation of the sample stimuli associated with different delays to food reinforcement. To measure resistance to change of delayed matching-to-sample accuracy, stable performance was disrupted by prefeeding and extinction. In both experiments, results showed that there was no systematic relation between the conditioned reinforcing value of the sample stimuli and the resistance to change. Instead, the resistance to change was somewhat greater in the component with the higher rate of primary reinforcement in Experiment 2. This may suggest that, in accord with behavioral momentum theory, the resistance to change of delayed matching-to-sample accuracy is governed by rate of primary reinforcement