Every day and with every interaction, animals learn how to respond to the people in their research environment. Positive reinforcement training (PRT) techniques can improve animals’ level of compliance with research tasks, as well as their physiological response to their environment. It’s critical that IACUC members have a working knowledge of classical and operant conditioning in order to critically assess whether research proposals include appropriate PRT.
On April 17, PRIM&R hosted a webinar to instruct IACUC professionals and members about PRT. Kelly Morrisroe, a research scientist and primate trainer at the Washington National Primate Research Center (WaNPRC) at the University of Washington, served as the speaker.
After the webinar, the presenter responded to some of the attendee questions that time didn’t permit us to address live. We’re pleased to share those responses with the readers of Ampersand.
How do you shape a very inhibited/cautious animal that does not offer behaviors?
Animals are always behaving, but they don’t always offer behaviors you are looking for. If you start with a behavior they are already doing, even if it’s as small as a head turn or even an eye movement that they get rewarded for, you can teach them that what they do affects the delivery of a reward. This is really the key element: the animal learns that the consequence follows the behavior and that it has control of the behavior and therefore the consequence. Often people get stuck focusing on their training plan which calls for specific behaviors and then don’t know how to proceed when the animal doesn’t offer a behavior that would lead to their goal. The first goal is always that the animal learns how to learn.
This is a different question, however, than the animal who is afraid to take a treat or any other sort of positive reinforcement. These animals are too affected by our presence to be able to learn. If our presence is too aversive for the animal to stay under stress threshold and be able to eat, or interact with us, then we must start with desensitization. You begin with some testing. Put a favored treat in their food bowl, walk away from the cage, and see how far away you get before they are comfortable coming forward for the treat. This is the distance you start your training from, and your first behavioral goal is for them to take the food. You present the food at the full distance multiple times, then take one fewer step away and see if they will come with you that one step closer. Gradually you reduce the distance, until you can feed the animal while you’re next to the cage.
What if animal finds the clicker aversive or is afraid of it?
A clicker is simply a tool. There is nothing particularly magic about it. If the animal finds it aversive, use something else! While voice markers aren’t ideal because they are variable, some animals do better with them. Or you can find another type of mechanical marker such as a whistle or bell that may be less aversive to that animal.
What if animal is not interested in food treats, i.e., food treats are not rewarding to them?
All animals eat. If the animal is not taking food it may be that either you are not offering the right kind of food, or that your presence is more aversive than the food is rewarding. I always start sessions with a new animal by just doing some preference testing. Do they prefer sweet or salty? Do they have texture preferences? A piece of apple may be more rewarding than applesauce or vice-versa. I had one neuroscience monkey who would only work for raisin juice. Not grape juice—it had to be raisins run through a Vitamix blender until it fit through the reward tube. I would encourage you to be creative!
The timing of when you are offering rewards is also important. When you have just eaten a big meal, you aren’t likely to want to eat something that isn’t a preferred food. If you schedule your training session for before an animal is normally fed, you are likely to get more motivation for the food than when they’ve just gorged on biscuits. But remember, even after Thanksgiving dinner, most everyone makes room for pie!
You can also use social interactions as reward. Many high drive dogs work for toys instead of food. If you have an animal that likes physical contact you can use that as well, or access to a preferred location in their enclosure. The important thing is to find what the animal likes. What are they willing to offer behaviors for in order to get?
You mention the importance of a high reward rate. What is the range of reward rate that you find to be effective? Suppose you start with a very high rate of reward and reduce the rate. What is the longest amount of time that you can wait to reward an animal? A couple minutes? Five minutes or more? Does it vary by species of animal?
There are several different layers to this question that involve a few terms other than rate of reinforcement. When I mentioned rate of reinforcement being high at the beginning of training, I am talking about number of rewards I can give in a minute. I need the animal to be engaged and wanting to work with me so I make things very easy allowing them to be successful in achieving reward quickly and often.
When you start talking about “waiting to reward” an animal, you are actually talking about building duration. Duration can be built using a number of tools, most of which involve varying types of schedules of reinforcement. The duration an animal can maintain a behavior is only limited by the effort needed to maintain the behavior, and the animal’s motivation to obtain the reward.
Schedules of reinforcement are their own topic that require more time and space than I can cover here. This blog post offers a good basic explanation of the different schedules and their effect.
What are your thoughts on the statement, “some animals cannot be trained”?
We are training our animals every day with every interaction. Every animal can be trained, but not every animal can be trained for every job. You cannot train a fish to climb a tree, but that doesn’t mean fish can’t be trained. Knowing your species’ physical and cognitive strengths and limitations will help you recognize what tasks are possible and what is going to motivate your animal to do those tasks.
Individual variability also has a huge impact on the time and effort it takes to train an animal. At WaNPRC, we pre-select animals for certain types of studies based on traits such as food drive, resiliency, and response to novelty. Selecting animals with traits we know will help them be successful shortens our timelines and increases our success rates.
Time, money, and personnel are often the limiting factors for training animals in research. These are more often the reason animals fail training than the animal’s actual limitations.
I am also interested to hear your thoughts on the length of time for training. How long is the typical time for pole and chair training? Is there a length of time when you say it’s not working?
This is the number one question I get from every research team. The length of time for training depends on multiple factors.
- Animals: Where are your animals starting from? A shy animal who doesn’t take treats will take much longer than a bold animal who is excited to work with you.
- Personnel: How much training experience do they have? What is their rapport with the animals?
- Training style: How much aversion is used? What’s your desensitization strategy or are you doing an all positive approach?
- Equipment: How does the animal have to maneuver into the chair? Is it a simple walk in or do they have to step down into a well area?
I never give timelines for training; the amount of time to train any individual is dependent on all the above factors which vary greatly with every lab. Decisions about when to cut your losses with any given animal and start over with a new one will depend on the team working with the animal. The behavioral factors I consider when deciding whether to proceed with a particular animal are:
- Motivation: Is the animal easy to reward?
- Stress: Can the animal stay under a stress threshold that allows it to learn?
- Relationship: Have the people decided an animal is untrainable?
- Options: Is there another animal available that would do better in this situation?
PRIM&R thanks Ms. Morrisroe for sharing her expertise.
The recording of this webinar is available for individuals to purchase in PRIM&R’s online store. If you would like to purchase the webinar for group viewing, please download the order form (PDF) and send it to firstname.lastname@example.org.