Behavioral adjustment to modifications in the temporal parameters of the environment

We study the dynamics of behavioral transitions when European starlings (Sturnus vulgaris) experience stepwise changes in the value of a meaningful time interval. Subjects were primed to respond at a certain time T1. After extensive training, the primed time changed to a new value T2. In Experiment 1 subjects were reinforced on 40% of the trials and they experienced a single transition which lasted until asymptotic behavior was reached. Starlings showed a progressive adjustment to T2, with no obvious discontinuities. In Experiment 2, probability of reinforcement was initially 20%, and the schedule switched to extinction after a varied number of trials were reinforced at the post-transition time. The number of post-transition reinforcements was used as independent variable. Behavior was examined in extinction to judge the state of temporal performance after a controlled amount of experience. Under these conditions, adjustment to T2 took place in two stages, and there was an intermediate phase when behavior changed little. These results are consistent with the hypotheses that animals continuously update the subjective probabilities that reinforcement comes at any given time and that responding occurs when the current estimate is above a certain threshold. We show that in spite of the continuous updating of time estimates, responding can show either continuous or discontinuous adjustments depending on the vicinity of the pre- and post-transition times and the probability of reinforcement.