Angelika Rehrig, Louis DiVincenti, Deborah Napolitano, and David McAdam
The psychological well-being of laboratory primates is enhanced when they are allowed to voluntarily participate in their own care through a process called positive reinforcement training (PRT). Through PRT, primates are trained for venipuncture, pole and collar restraint, and experimental testing procedures. PRT relies on rewarding the animals when they perform the desired behavior (e.g., submitting their arm and remaining still for a blood draw). To improve PRT, the process must be both practical and effective at creating the desired behavioral outcomes. One way to accomplish this is to use the right reinforcer, or reward. Often, the reinforcer is chosen by a caregiver’s assumptions of what is reinforcing to a particular animal. This method is subjective and does not take into account fluctuations in preference.
Since food and fluids are often used as reinforcement with primates, a plethora of items could potentially be used during PRT. In contrast to the “best guess” approach, preference assessments systematically and objectively determine preferred items that could function as reinforcers. Our study, which was funded through an AWI Refinement Grant, evaluated the use of the “multiple stimulus without replacement” (MSWO) preference assessments to determine preference hierarchies and stability of food choices for 14 cynomolgus macaques over the course of a month. To investigate whether the most preferred foods functioned as reinforcers, half of the participants completed concurrent-schedule reinforcer evaluations. The practicality of the MSWO was also evaluated in terms of overall time investment.
The MSWO clearly identified food preferences for all 14 macaques and required little time to do so (mean = 3mins 57s). Yogurt-covered raisins/peanuts and grapes were the most preferred items. On average, dried pineapple and banana chips were moderately preferred, while dried apricots and peanuts were the least preferred. All but one of the macaques had unstable preferences across the four assessments, which is consistent with results from similar studies. Following the assessments, the reinforcer evaluations demonstrated that the most preferred item did function as reinforcement for a simple task. The primates were more willing to work for their most preferred item versus the least preferred.
Determining individual preferences can improve PRT by altering a primate’s motivation to learn. Since preference assessments identify a hierarchy of preferred foods, reinforcers can be varied so as not to decrease their potency. If a primate prefers grapes, but always receives them during PRT, their effectiveness may diminish over time. However, if the preference hierarchy is known, reinforcement can be varied such that the most preferred items are used for demanding or novel behaviors, while moderately preferred items are used for less demanding behaviors. Routinely identifying preference may even result in more completed trials and increased data collection.
In conclusion, the effectiveness of PRT relies heavily on the relative power of reinforcers to reward desired behavior. Since preferences change over time, the MSWO provides a quick and accurate method to regularly identify preferred reinforcement for primates.