Quantifying animals' perception of environmental predictability using anticipatory behavior

Animals under human care often experience predictable daily husbandry events, which can promote the development of anticipatory behavior. Previous research suggests even short delays in the arrival of a predictable, desired outcome can lead to negative welfare outcomes for animals. As such, providing reliable information to animals regarding the onset of important events may be a simple but useful method to support positive welfare. Here we evaluated the quantitative characteristics of anticipatory behavior of a California sea lion at the San Francisco Zoo in three situations (1) when the animal had accurate information about the occurrence of the event (temporally predictable training sessions), (2) when the information was semi-reliable (unscheduled training session), and (3) when a reliable signal was offered before unscheduled sessions. Results showed that providing a reliable cue resulted in a stronger temporal correlation of anticipation before the beginning of the unscheduled session, similar to the temporally predictable training session. However, providing a reliable cue did not reduce the intensity of the anticipatory behavior. We propose to take into account two aspects of the quantitative characteristics of anticipatory behavior: correlation with time until a desired event (correlation) and intensity, where the correlation indicates the ability of an individual animal to predict the occurrence of an event, and the intensity indicates the degree of sensitivity of the animal to reward. We discuss the implications for animal welfare and husbandry. Anticipatory behavior is expressed by animals ahead of predictable, rewarding events. In zoos or aquariums, however, schedules of husbandry events are rarely fixed in time. Animals expressing anticipation for long periods may be at greater risk for negative welfare states, such as frustration, as a result of not being able to precisely predict when desired events will occur. In this study, we examine how an animal's anticipatory behavior changes in response to providing signaled reliability, and we propose a quantitative measure of anticipation to assess how animals experience the predictability of their environments.image