Enrichment protocol: Certainty of reward

In Lab Note #1, we answered the question, "What is enrichment?" and in Lab Note #2, our video shows how an Arduino UNO board programmed by Arduino software will allow us to change the time of day that our 'smart feeder' releases food for the lemurs. This lab note explains why it is important to alter the predictability with which food is released by the smart feeder, otherwise known as the 'certainty of reward'.

Captive animals often have low activity levels because they are unmotivated to perform behaviors that are unnecessary. By giving animals something to do (like foraging with enrichment), their activity levels are expected to increase because they must 'work' to achieve a reward. The goal of our project is to increase the amount of time that the lemurs at Oakland Zoo spend actively searching for food. To accomplish this goal, we will alter when the smart feeders release the food to examine the level of predictability (aka: certainty of reward) necessary to encourage the greatest activity levels in the lemurs.

Our ideas are based on previous experiments on contra freeloading (where animals work for food even in the presence of free food), which show that when a reward has some level of uncertainty, animals will work harder for it. Jason Watters' research from the San Francisco Zoo has shown that changing the rate of reward in different parts of the exhibit encourages movement around the enclosure. However, when the probability of reward is 100%, or completely predictable, animal's motivation to obtain said reward diminishes quickly (Figure 1) as it is too easy. An animal's motivation to seek the reward is also impeded when the probability of reward is 0%, or entirely unpredictable, as it is difficult and frustrating. Animal trainers take advantage of this simple observation and promote more motivated responses through partial or varied reinforcement.

Animal motivation to perform a behavior is maintained when the payoff for performing that behavior exceeds that of other behaviors (Figure 2). If motivation to perform a behavior is maintained by high payoffs, and control is equal to certainty of reward, then Figures 1 and 2 can be combined to create Figure 3 (Watters, 2009). This suggests that if animal motivation is highest at uncertain payoffs then there is something more rewarding or stimulating about an uncertain payoff than a certain one.

Our project aims to build on this previous research by better understanding how the activity level and enclosure use of captive lemurs is affected by the deployment of environmental enrichment with different levels of certaintyof reward.

 

Figure 1: Hypothesized relationship between animal control and motivation. This is based on interpretation of observations on animal learning and response to training.(Reprinted from Watters, 2009 with permission from the author).

 

Figure 2: Relationship between payoff and motivation, according to optimal foraging theory. (Reprinted from Watters, 2009 with permission from the author).

 

Figure 3: Model for the relationship of certainty of reward and motivation. Points A, B, C, D, and E are equally spaced along the x-axis. An animal foraging in a patch that offers a reward at rate A is expected to have a relatively low motivation to forage there. When the patch's probability of reward shifts from A to B, the result is a large shift in animal motivation.(Reprinted from Watters, 2009 with permission from the author).

References

Watters, J.V. (2009). Toward a predictive theory for environmental enrichment. Zoo Biology 28:609-622.