Association for Behavior Analysis International

The ability to time accurately is important in a range of circumstances, from choosing in situations where time since a past event dictates where reinforcers are likely to be in the present, to choosing between reinforcers that will occur in the future. Time exerts relatively strong control over behaviour, even when other stimuli provide more reliable information about reinforcer availability. However, accuracy and precision of timing depends on a variety of environmental factors. This symposium explores factors that enhance or disrupt timing. We discuss timing processes underlying behavior in a variety of procedures, when timing is the most reliable stimulus signaling the likely availability of reinforcers, and when stimuli other than time are more reliable cues for the likely availability of reinforcers. Talks investigate how timing processes depend on other cues such as location, reinforcer rate, and how timing in the present is affected by past experience under time-based contingencies. We discuss the implications of findings for theories of timing, and for understanding the role of time in self control.

In a midsession reversal task animals are required to do a simultaneous color discrimination between S+ and S-, and half way through the session contingencies are reversed. Typical performance of anticipation and perseveration errors around the reversal point, suggests that the effect of reinforcement is modulated by the time elapsed from the beginning of the session. In this task, some manipulations of the reinforcement rate produce biases in pigeons' performance consistent with the idea of time-regulated behavior. Other manipulations of the reinforcement rate seem to disrupt timing and enhance the control of local cues. These results can only be accounted by hybrid models of learning that take into consideration the combined effect of reinforcement and time over behavior. The Learning-to-Time model accounts for performance in this task when behavior is under temporal control and accurately predicts the observed biases. Altogether, model simulations and experimental data suggest important implications for current learning theories.

Recent research has supported a notion of time as a stimulus dimension that competes for control over behavior with other dimensions, such as location. In this experiment, pigeons were exposed to a temporal bisection task, with side keys associated to short and long stimulus durations (i.e., 3 s- left, 12 s- right). Then, intermediate durations were also presented without reinforcement. Keys were spatially set apart in a long operant chamber, and floor panels recorded location throughout trials. After acquisition, movement patterns were stereotypical in long-duration trials: Birds would go to the short side and, after about 4.5 s, leave for the long side. When differential reinforcement was provided for correct "long" and "short" key pecks, animals switched to the long side earlier when the "long" response produced more reinforcement, and later when it produced less. These manipulations also disrupted the stereotypical pattern: after leaving for the long side, pigeons returned to the short side depending on their location in the moment the duration ended (i.e., stimulus turns off). It is likely that location gained behavioral control because the differential payoff reduced the probability that the animal engaged in timing. However, the mechanisms behind this effect are still unclear.

Timing processes have been implicated as potential mechanisms of interventions hypothesized to improve self-controlled choice. Specifically, exposure to delayed reinforcement and interval-based schedules has been shown to reduce subsequent impulsive choice while also improving temporal discrimination. However, the mechanism behind these interventions remains unclear. If improvement in timing processes is an underlying factor in previous intervention studies, then it is likely that similar timing improvements will occur when timing is assessed in isolation. In the current study, timing was assessed using a peak procedure prior to and following exposure to a differential-reinforcement-of-low-rate (DRL) schedule in mice. Across three treatment groups, mice experienced a DRL schedule that was the same value, half the value, or one and a half the value of the fixed-interval schedule used in the peak procedure. A control group remained on the peak procedure. Following the DRL intervention, timing was reassessed using the same peak procedure. In contrast to previous reports, the DRL intervention produced a transient degradation in timing precision via increased peak spread. It is possible that previous DRL interventions may have functioned to decrease sensitivity to immediacy in some circumstances leading to alterations in impulsive choice.