Maximizing the Effectiveness of Environmental Enrichment: Suggestions from the Experimental Analysis of Behavior

Environmental enrichment programs provide benefits to both captive animals and the facilities that house them, but cost time and resources to design, implement, and maintain. As yet, there have been few theoretically based guidelines to assist animal care staff in establishing cost-efficient enrichment methods that both elicit the desired behavioral changes and maintain their success over time. We describe several well-studied principles from the field of experimental analysis of behavior, including intrinsic reinforcement, extrinsic reinforcement, habituation, extinction, and schedules of reinforcement that could be very useful for evaluating the short- and long-term effectiveness of enrichment. We use this theoretical framework to generate testable hypotheses and provide examples of enrichment studies relevant to our predictions. In particular, we suggest that enrichment devices that offer extrinsic reinforcement (food, social access, etc. as a result of performing behaviors) should produce greater and more prolonged changes in behavior than devices that rely on the behavior itself being reinforcing to the animal. For techniques that provide no extrinsic reinforcement, using stimuli that are novel, are more different from the environment, have been withheld or altered in some way, or are presented less frequently may help reduce habituation. For techniques that provide extrinsic reinforcement, making reinforcement more difficult to obtain and providing more or higher quality reinforcers may increase the long-term success of the enrichment program. In addition, enrichment may be more effective if animal care staff avoid continuously reinforcing behaviors after they are established, enriching immediately after feeding, and exposing animals to enrichment when reinforcement is no longer available. While the current enrichment literature supports the application of behavior analytic theory, empirical evaluation of many of our predictions is still needed