Preference for a stimulus that follows a relatively aversive event: contrast or delay reduction?

Abstract

Several types of contrast effects have been identified including incentive contrast, anticipatory contrast, and behavioral contrast. Clement, Feltus, Kaiser, and Zentall (2000) proposed a type of contrast that appears to be different from these others and called it within-trial contrast. In this form of contrast the relative value of a reinforcer depends on the events that occur immediately prior to the reinforcer. Reinforcers that follow relatively aversive events are preferred over those that follow less aversive events. In many cases the delay reduction hypothesis proposed by Fantino (1969) also can account for such effects. The current experiments provide a direct test of the delay reduction and contrast hypotheses by manipulating the schedule of reinforcement while holding trial duration constant. In Experiment 1, preference for fixed-interval (FI) versus differential-reinforcement-of-other-behavior (DRO) schedules of reinforcement was assessed. Some pigeons preferred one schedule over the other while others demonstrated a position (side) preference. Thus, no systematic preference was found. In Experiment 2, a simultaneous color discrimination followed the FI or DRO schedule, and following training, preference was assessed by presenting the two positive stimuli simultaneously. Consistent with the contrast hypothesis, pigeons showed a significant preference for the positive stimulus that in training had followed their less preferred schedule.

Fig 1. Model of within-trial contrast based on relative hedonic states.

Each trial begins at a relative value of 0. Key pecking (or another nonpreferred event such as delay) causes a negative change in value, and reinforcement results in a positive change in value. The FR 20 response requirement results in a relatively more negative change in value than does the FR 1 requirement and hence in a greater positive change in value upon obtaining the reinforcer. Adapted from Zentall (2005).

Fig 2. Absolute (top) and relative (bottom) delay reduction as applied to the procedure used by Clement et al. (2000).

The absolute time between the presentation of the discriminative stimuli and food reinforcement (top) is the same on FR 20 and FR 1 trials but the discriminative stimuli represent a smaller proportion of the total duration of the FR 20 trial than the FR 1 trial (bottom). Thus, the discriminative stimuli should be more predictive of reinforcement on FR 20 trials.

Fig 3. Schedule training procedure used in Experiment 1.

One peck to the lit left key initiated vertical lines on the center key indicating that a DRO schedule was in effect. One peck to the lit right key initiated horizontal lines on the center key indicating that a FI schedule was in effect. The schedule associated with the left and right keys was counterbalanced over subjects.

Fig 4. Design of Experiment 2.

The left panel shows a DRO trial with completion of the DRO requirement followed by one simultaneous discrimination (red+ yellow−). The right panel shows a FI trial with completion of the FI requirement followed by a different simultaneous discrimination (green+ blue−). Colors associated with the two simultaneous discriminations were counterbalanced.

Fig 5. Experiment 2: Preference of each pigeon for the S+ that followed the less preferred schedule or side on Test Session 4.

Fig 6. Mean percentage preference above chance for the S+ that followed the less preferred schedule or side over the four testing sessions in Experiment 2.

Error bars represent 1 standard error of the mean.

Fig 7. Experiment 2: Preference of each pigeon for the S− that followed the less preferred schedule or side on Test Session 4.