Conditional reinforcers and informative stimuli in a constant environment

Abstract

Five pigeons responded on steady-state concurrent variable-interval variable-interval schedules of food presentation in which half of the foods were removed and replaced with nonfood stimuli. Across conditions, the stimuli were either paired or unpaired with food, and the correlation between the ratio of food deliveries on the two alternatives and the ratio of nonfood stimuli was either -1, 0, or +1. Neither the pairing of stimuli with food, nor the correlation between stimuli and food, affected generalized-matching performance, but paired stimuli had a demonstrable effect at a local level of analysis. This effect was independent of the food-stimulus correlation. These results differ from results previously obtained in a frequently changing environment. We attribute this difference in results to differences in the information value of response-contingent stimuli in frequently changing versus relatively constant environments, as well as to differences between forward pairing and simultaneous pairing of the stimuli with food.

Fig 1

Log (left/right) response ratio after each response-contingent event type as a function of the number of immediately preceding same-alternative, same-type (food or stimulus) events in Condition 3 for each individual subject, as well as the log response ratios calculated after summing the responses of all subjects. The dashed horizontal line is the overall mean response ratio in Condition 3 and the solid horizontal line at 0 represents indifference between the two alternatives.

Fig 2

Log (left/right) response ratio after response-contingent events as a function of the number of immediately preceding same-alternative, same-type (food or stimulus) events, in conditions that arranged stimuli unpaired with food, correlated either −1, 0 or +1 with food. The top row shows results from conditions arranging a 9∶1 reinforcer ratio and the bottom row shows data from 1∶9 conditions. The dashed horizontal lines show extended-level log response ratio values from each condition and the solid horizontal line is a reference point at 0, which corresponds to indifference between the two alternatives.

Fig 3

Log (left/right) response ratio after a response-contingent food or stimulus delivery as a function of the number of preceding food deliveries on the same alternative.

Fig 4

Log (left/right) response ratio in the period after a response-contingent event delivery as a function of the number of preceding food deliveries on either the same or the other alternative.

Fig 5

Log (left/right) response ratio as a function of time since the most recent response-contingent event delivery in Condition 3 for all individual subjects as well as the log response ratio per time bin calculated after summing all responses of all subjects. The horizontal line is at 0 and represents indifference between the two alternatives. Points at +4.5 or −4.5 indicate time bins where the subject responded exclusively to the left or right alternative, respectively.

Fig 6

Log (left/right) response ratio as a function of time since the most recent response-contingent event delivery in conditions arranging stimuli unpaired with food. The top two rows show conditions that arranged a 9∶1 food ratio, and the bottom two rows show conditions arranging a 1∶9 food ratio. Successive panels horizontally represent food–stimulus correlations of −1, 0, and +1 respectively. The unbroken horizontal line shows equal preference between the two alternatives, and the broken line shows the extended-level preference in each condition. Points at +4.5 or −4.5 indicate time bins where responding was exclusive to the left or the right alternative respectively.

Fig 7

Continuations in conditions in which stimuli were paired with food. The top row shows conditions that arranged a 9∶1 reinforcer ratio, with −1, 0, and +1 food–stimulus correlations, respectively. The bottom row shows conditions that arranged a 1∶9 reinforcer ratio. The unbroken horizontal line indicates indifference between the two alternatives and the dashed horizontal line corresponds to the overall preference in that condition.

Fig 8

Log (left/right) response ratio after a food or a stimulus presentation as a function of the number of preceding foods delivered on the same alternative. The unbroken horizontal line depicts indifference between the two alternatives and the dashed horizontal line depicts the overall response ratio in that condition.

Fig 9

Log (left/right) response ratio after response-contingent food or stimulus presentation as a function of the number of preceding foods on either the same, or the other, alternative.

Fig 10

Log (left/right) response ratio as a function of time since the most recent response-contingent event delivery in Condition 19 for all individual subjects, as well as the log response ratio per time bin calculated after summing all responses of all subjects. The horizontal line indicates indifference between the two alternatives. Points at +4.5 or −4.5 indicate time bins where the subject responded exclusively to the left or right alternative, respectively.

Fig 11

Log (left/right) response ratio as a function of time since response-contingent event delivery in all conditions arranging stimuli paired with food. The top two rows depict conditions that arranged a 9∶1 food ratio, with successive horizontal panels showing food–stimulus correlations of −1, 0, and +1 respectively. The bottom two rows depict conditions arranging a 1∶9 food ratio. The unbroken horizontal line represents indifference between the two alternatives and the dashed horizontal line represents the overall response ratio in that condition. Points at +4.5 or −4.5 indicate time bins where responding was exclusive to the left or the right alternative respectively.