This is the third in a series of posts based on my notes from the 2020 Art and Science of Animal Training Conference that was held in Hurst, Texas on February 22-23, 2020. To learn more about the conference, you can visit the conference website.
While I try to take accurate notes, it is possible that there are errors or that some detail is lacking. If you post a comment or email me, I can try to clarify or provide some additional information. Many thanks to the speakers and organizers who allow me to share.
Dr. Jesús Rosales-Ruiz: C.A.T. Rules
Dr. Jesús Rosales-Ruiz spent the first part of his presentation talking about the origin, development, and early implementation of CAT – Constructional Aggression Treatment.
In 2005, Dr. Jesús Rosales-Ruiz met Alexandra Kurland when they were both asked to be on the ClickerExpo faculty. He attended her presentation on rope handling and was intrigued by Alex’s use of negative reinforcement in rope handling and through physical contact (touch). After meeting her, he found he had to rethink some of his ideas about negative reinforcement. She was interested in talking about the applications of negative reinforcement with someone who knew more about the science of behavior analysis. So, they started a discussion which, as Jesús said, continued on for several ClickerExpos and resulted in late nights, shared taxi rides, and regular email communication. CAT emerged out of these discussions.
At first, Jesús thought that the difference between their ideas about negative reinforcement was due to differences in their language or definitions. But, as they talked more and more, he realized that Alex used negative reinforcement as communication, whereas he tended to think about it in terms of fear. He started thinking about possible applications for negative reinforcement and he threw the idea out to his graduate students. Would one of them like to do some work on using negative reinforcement to change behavior?
Kelly Snider took on the project. She had a friend whose dog was reactive to other dogs and she decided to see if she could shape calmer behavior by using distance as a reinforcer. It worked! After that, she worked with a number of other dogs who showed varying levels of aggression towards dogs or people. It took her a while to work out the mechanics of the procedure and she wrote up her results as her masters thesis. If you want to read the thesis, you can find it here.
At this point, they thought that all they had to do was put CAT out there and people would be able to do it. She and Jesús taught several seminars on the procedure and a DVD was produced. They also formed an email list to support people using CAT. On that list, there were 239 reported cases of success, but there were also some difficulties that people had implementing it. One of the problems was that people were using going too quickly, flooding the dogs, and putting them over threshold. The current emphasis on free shaping did not help either. Instead of setting up conditions where an appropriate behavior was likely to occur and shaping it with negative reinforcement, people were trying to select an appropriate behavior out of whatever the dog was doing when it became stressed by the presence of the stimulus. This didn’t work very well.
The principles behind CAT were sound, but the application didn’t work and it didn’t have any “social validity.” People didn’t want to use it if the dogs became more stressed out in the process. Some improvement needed to be done.
Back to the lab….
The next major work on CAT was done by Angie Rentfro and was done with feral cats. You can read her thesis here. She took the original CAT protocol and refined it. Working with a large number of cats, she identified a number of problems with the original procedure. Some of the things she changed were how to find the starting point, how quickly to progress, and what kinds of behaviors to reinforce. She realized that she needed to reinforce a large variety of “normal” behaviors, ones that cats would normally perform in the absence of the stimulus. She was very successful and her work contributed to the “new rules for CAT.”
New Rules for CAT
Basic Characteristics: CAT …
- Is constructional
- Bases treatment on a functional analysis of fearful and aggressive behavior
- Uses shaping
Constructional Programs focus on the behavior you want, rather than the problem you want to eliminate. CAT is an operant procedure based on the functional analysis of behavior. You are always working with what the animal wants. It is not based on classical conditioning. CAT uses shaping to create organism-environment friendly interactions.
Jesús had a slide comparing the different approaches to resolving fear and aggression issues.
The Three Stages of CAT
- Undesirable behavior (e.g. attack or retreat) maintained by negative reinforcement.
- Desirable behavior (e.g. approach, sniff, staying in place, and other desirable behavior) maintained by negative reinforcement
- Desirable behavior (e.g. approach, sniff, staying in place, and other desirable behavior) maintained by positive reinforcement.
It’s important to note that prior to CAT, no one was talking about how aggression was tied to stimulus control and no one was talking about “switchover” which is when the animal changes from seeking distance to choosing interaction.
The CAT Program
- Step 1: Understand your dog
- Step 2: Imagine the perfect dog
- Step 3: Manage the current environment
- Step 4: Get ready for training
- Step 5: Determine the starting point
- Step 6: Shaping new behavior
- Step 7: Interaction
- Step 8: Training for generalization
Let’s look at each step in more detail.
Step 1: Understand your dog
Determine exactly where and when the aggressive behavior occurs
- Does your dog engage in these behaviors around certain people, while in certain places, or when certain types of activities are going on?
- Is your dog engaging in aggressive behaviors in an attempt to increase the distance between himself and the other dog (or person)?
- This is the step in which you should establish your baseline so you know what typical behavior looks like and when it happens.
He showed a few examples. In one case, the dog was only aggressive in certain locations. In another, the dog was aggressive when a certain family member was present, but not when she was absent.
In this step, Jesús introduced a dog that Mary has been working with using CAT. Drill Bit is being trained as a service dog but he had started to behave in undesirable ways when he was approached by strangers or by other dogs.
Step 2: Imagine the perfect dog
What do you want the dog to do? It’s important to choose a behavior that is already in the dog’s repertoire. In most cases, you want to choose a behavior that is calm and one that the dog would choose to do when it is relaxed.
He showed a video of Mary and Drill Bit waiting on the sidewalk while people walked around them. In this situation, a calm sit might be appropriate. In later videos, they showed the same dog relaxing in the back yard with Mary. In this case, the desirable behavior(s) were for him to continue doing whatever he was doing (lying down) before the person (trigger) appeared. This was a good example of choosing a set of related behaviors – ones that a dog might do while just lounging around (lying down, casually looking around, itching a spot, looking at Mary, etc.).
Step 3: Manage the current environment
CAT may take multiple sessions
- Minimize exposure to triggers outside of training
- You may need to modify your routines
- Brainstorm possible options with the owner
Step 4: Get ready for training
- Select a safe environment with minimal distractions where the training can take place.
- Select the first “decoy.” The decoy is your human helper (or dog helper and the dog’s handler) who will approach or be approached by the dog during the training.
- The human decoy or handler must be able to follow instructions
- The canine decoy must stay calm, level-headed, focused on handler, even if the training dog begins barking, lunging, etc.
- Select the appropriate gear
- well fitting collar or harness
- leash that allows the owner good control of the dog. Do not use a flexi-leash.
- Use the dog’s normal equipment unless the owner uses pinch, shock, choke collars or head halters. This will make it easier when you get to step 8: Training for generalization.
Step 5: Determine the starting point
Capture the alert.
- Have the decoy begin from a far distance, preferably out of sight and walk toward the dog.
- When the dog notices the decoy
- Stop at this point for just a second and then leave before the dog begins engaging in unwanted behavior.
- The decoy walks far enough away so that the dog can return to normal – that is, the dog returns to activities similar to those he was doing before the decoy appeared.
Note that the goal here is to have the decoy STOP before the dog can do any unwanted behavior. You are just establishing the point at which the dog notices the decoy.
Step 6: Shaping new behavior
There are two types of trials in CAT:
- Type 1: Reinforce the alert
- Whenever criteria is raised (closer distance, new decoy behavior)
- Type 2: Increase duration/reinforce appropriate behavior
- If the dog is relaxed, leave then the duration runs out. Let’s say you were going for 5 seconds of duration. If the dog remains relaxed, you complete the duration and then have the decoy leave.
- If the dog freezes or reacts at any time, then you end the trial (before the 5 seconds is done) and lower the criteria for the next trial.
- If the dog displays friendly or calm behavior, leave – even if it’s before the 5 seconds is up. You want to reinforce those changes in behavior and they are more important than duration.
- My understanding is that you are bouncing between reinforcing alerts, reinforcing calm/friendly behaviors and increasing duration – the goal being to slowly build duration of calm behavior as the decoy gets closer.
During each trial:
- Owner (or trainer) signals for the decoy to approach
- Decoy approaches in a normal manner and goes to the current point
In between trials:
- Decoy walks a far distance away and/or out of sight
- Waits at least a minute
- Starts the next trial when the dog has returned to normal
Step 7: Interaction
You are ready for interaction once the decoy is at a distance just out of range of the dog’s leash.
Stay at this point until:
- The dog remains calm/relaxed for at least one minute
- Look for “friendly” behaviors or approximations
- friendly behaviors include sniffing or calmly looking at/orienting toward the decoy, taking a step forward with relaxed body language, tail wagging with happy body language.
- Walk away when any of these behaviors occur or when the duration is up.
- The dog gets to approach
- Wait until the dog is dog is curious or interested
- Start by letting the dog sniff
- Then, allow interaction or reset
- Don’t use food during the interactions
Step 8: Training for generalization
Train sufficient exemplars
- New decoys in original locations
- Old decoys in new locations
- New decoys in old locations
- Increasing variations (hats, beards, sounds, etc.)
A dog example with Drill Bit and Mary
For each of the steps listed above, Jesús had some video showing how they worked through the process with Drill Bit. They started in a quiet backyard with Mary sitting in a chair and Drill Bit lying on the ground next to her. When he was quiet, Maasa came in the back gate (facing backward) and waited until Mary told her to leave. Mary did so as soon as Drill Bit noticed her. Maasa entered facing backward because Drill Bit was more likely to respond when he saw the person’s face.
Over the next trials, Maasa turned around and then came closer. Each time she left when Mary asked. Mary kept track of duration, directed criteria changes and observed Drill Bit for signs of relaxation. Drill Bit showed some changes in behavior but they were very mild. For the most part, he stayed relaxed throughout the entire process. After these initial sessions, Mary did some in other environments so that Drill Bit could practice with other people and in other locations. Again, he stayed calm.
This new version of CAT is easier for people to implement and has been used successfully with animals of many species. In the next talk, Barbara Heidenreich shared how she uses with fearful or aggressive animals in zoos.
Just for fun…
An alternative interpretation of the phrase “CAT rules.”
(Sorry – I couldn’t resist…)
Katie’s notes: I first learned about CAT when Jesús and Kelly were doing the original work with aggressive dogs. At the time, I was still fairly new to using using positive reinforcement and it seemed odd to me that someone would want to use negative reinforcement to address fear and aggression.
Even considering that, I was intrigued because my mare was aggressive toward people and while I had made progress, I felt like it was due to a combination of positive reinforcement and good management on my part, more than to a change in her attitude toward people. I gave her alternative behaviors to do so that she was safe, but I hadn’t changed how she felt about people. CAT seemed interesting. However, I found it was difficult to implement because she and I had worked out a system for handling visitors and using CAT meant setting aside my current strategy and starting all over again. Honestly, I just couldn’t do it.
While I didn’t use CAT for that particular challenge, I still found the idea intriguing and over the years I have experimented with using a CAT type approach when my horses show fear about approaching or being approached by strange objects or venturing into new locations. Distance can sometimes be a valuable reinforcer if a horse won’t take food or accept other types of positive reinforcement.
I have used CAT (as closely as I could) with several stray cats that have arrived at our farm.
This is my cat Indy. I found him in the road when he was about 4 months old. He was feral and I used what I knew about CAT to help him overcome his fear of people.
When CAT came out it was fairly controversial. I read a lot of criticism about how it was implemented and I also read a lot of discussions about how it could be done better. If nothing else, it generated a lot of discussions about alternate approaches to fear and aggression. Previously most of the approaches had involved a combination of counterconditioning and desensitization so CAT’s operant approach generated some fresh thinking. Some trainers took parts of CAT and added other pieces to come up with their own variations, which have been quite succesful.
I hadn’t heard much about CAT recently so I was glad to hear this presentation and learn that Jesús and his students have continued to work on improving the procedure. It sounds like they are having a lot more success.
I’ve been blogging about the ASAT conference for several years now. You can find a list of them in the articles page on this blog, but some older ones are on my website, www.equineclickertraining.com.
If you are interested in learning more about how to clicker train your horse, check out my book, Teaching Horses with Positive Reinforcement, available on Amazon.