This piece first appeared in the June 25, 2024 issue of our newsletter.
A cue signals the learner that if they perform the behavior NOW, they will receive reinforcement.
In our ABCs (antecedent, behavior, consequence) the cue is the antecedent to the behavior we want to reinforce. In more technical terms, the cue is a “discriminative stimulus” that we have paired with the operant behavior and predicts reinforcement (Cooper, Heron, and Heward, 2020) and becomes salient when we always provide the reinforcement.
Most people only consider two types of cues when training animals: verbal cues and hand signals. There are so many more options than just those two! For example, we also have environmental cues, tactile cues, whole body cues, compound cues, and olfactory cues. If it can be perceived by the learner, it can be a cue. There are also behavior chains where the behavior before cues the behavior after (such as a service dog opening the fridge, getting a bottle of water, and carrying it to the handler). In such a chain, each behavior does not need its own verbal cue, it is cued by the behavior before which originated with the handler’s cue. Service dogs represent the gold standard of understanding different types of cues. Many are taught to respond to olfactory cues (scenting high or low blood sugar for diabetes, or scenting an impending seizure) as well as whole body cues (handler seizing or fainting means go find help), and environmental cues (eyesight dogs guiding around objects in their path to lead their handler safely).
Now you’re wondering “Why is SPA talking about service dogs”? Cues are easy to apply to the antecedent category, but in our most ideal world, our shelter animals are not present at the shelter long enough to learn the cues we want to teach them in controlled learning sessions. Most won’t learn that “sit” means “place butt on floor,” and that doing so will unlock a tasty treat, to the level of fluency we desire. Very few will ever learn complex behavior chains that we intentionally teach them. They might learn behavior chains we don’t intend to teach, though (leash at front of kennel cues rushing forward, rebounding off kennel door, biting at the leash, whole body flailing, barking, getting leashed and lunging out of the kennel, dragging handler to yard which is reinforced with leash being removed and dog running full speed zoomies). Sound familiar? These scenarios are why we encourage and teach low arousal behaviors from kennel exit to kennel return. We can change the response to the “leash being presented” cue from high arousal to low arousal behaviors just by adding reinforcement to the ground from the very beginning. “Leash presented” becomes the antecedent for the behavior of “head down eating treat scatter or canned food smear.” Remember that “eating” is a behavior. (If you want to get nerdy, check out Kathy Sdao).
Some of our shelter animals are so stressed in the shelter environment that we might need to work on eating at the kennel front prior to taking them out. We might need to find the food that is high-value enough for them to be able to eat despite the stress (see below to learn more about preference testing). Once the animal has acclimated to the environment, learned to eat at the kennel front, and is under less general stress, the type of food may be of less significance.
The most important part of a cue is that it must be perceived by your learner. Verbal cues are great until you have a deaf dog who physically cannot hear you saying “sit” (learn about training deaf dogs from Terrie Hayward below). A cue may also not be perceived if the animal is distracted or under a large amount of duress. It is very hard to solve a calculus problem if you’re running from a lion. If a dog is uncomfortable with strangers, it may be impossible for the dog to perceive they are asking her to sit, especially if her brain is actively screaming “danger!” at her. In that scenario, it would be much more realistic to give a cue that helps build distance from the scary thing (such as tossing a treat away for her to chase and eat away from the scary thing). Yes, tossing food can be the cue to go eat that food.
When first learning a cue it can take time to sort through which behavior is actually being asked for. This is why it’s important to give your learner time to respond instead of repeating yourself immediately. We are essentially teaching them a foreign language. Repeating yourself multiple times does not make the cue more clear. It only turns the cue into “sit sit sit”. When giving a cue say it once, then wait for the animal to respond. If the animal doesn’t respond, you can either draw the conclusion that the animal does not understand the cue or that the animal was unable to perceive the cue (too loud to hear the verbal, not looking at the hand signal, too stressed to respond).
In an effort toward errorless learning, if you give a cue and the animal does not respond (or performs a different behavior) the ideal response is to cue a different behavior, which can be as easy as tossing food on the ground as a cue to go eat it.
When training by shaping, luring, and capturing behaviors, adding the cue is the last step to the process. We generally want a behavior fully formed to the final product before placing it on cue. Once the cue has been added, no more editing should be done. Most shelter animals aren’t in care long enough to achieve the number of successful repetitions needed for adding a cue. If they are in care long enough, it is likely that they’ve been too stressed to be in a learning frame of mind. We also need to take into consideration that most animals are being taught similar skills by a multitude of handlers using slightly different mechanics and will go home to a handler who has no or limited knowledge of what they learned. It might be pointless to apply an English verbal cue to a behavior when the dog goes home to a Mandarin or Spanish speaker.
Instead, focusing on the process of shaping, capturing, and luring new skills can be most beneficial as the animal learns how to learn. They learn that their behavior affects change and can produce reinforcement from people. They are getting classically conditioned that people mean good things happen and operantly conditioned on how to engage with humans. The only skills we prefer to have on cue are Look At That which is an environmental cue of ‘see a trigger and I’ll feed you’ (see more on this by visiting Leslie McDevitt’s ControlUnleashed page).
Other environmental cues are skills we may be constantly shaping, such as calm exits from the kennel, eating from enrichment stations, automatic check-ins, etc. Remember that many behaviors learned in the shelter setting don’t generalize to life outside the shelter. Even if a dog has a sit behavior on a verbal cue in the shelter environment, she will likely be “retrained” by her new owner once they’re home. Maybe her new owner’s preferred sit cue is “sit sit sit”.
References:
Cooper, J. O., Heron, T. E., & Heward, W. L. (2020). Applied Behavior Analysis (3rd Edition). Hoboken, NJ: Pearson Education.
Hayward, T. (2015). A Deaf Dog Joins the Family: Training, Education, and Communication for a Smooth Transition. Amazon publishing.
McDevitt, L. (2007). Control Unleashed: Creating a Focused and Confident Dog. Clean Run Productions.
Comments