Can treats reinforce a dog’s fear or aggression?
A person that posted a comment on our blog asked us such a good question in response to our article, 5 Harsh Realities of Treating Dog Aggression that we thought it was worth a blog post. The question came after a statement that treats can’t reinforce (strengthen or cause it to repeat) fear. The poster asked if we would mind explaining why/how an emotion cannot be reinforced/rewarded.
A common idea of reinforcement works is like this: rewarding a dog after a dog has performed a behavior will cause that behavior to thereby reinforcing it. So what happens if you give a fearful dog a treat? Does it not cause the fear to be strengthened? It is a legitimate question, but it’s not that straight forward.
What we learn has an adaptive logic
The concept of rewarding a dog for a behavior will cause it to repeat is a nice tidy idea, but dogs are not simple vending machines where you put something in and something comes out. There is more at work; we make associations with situations that are most likely to help us survive given the context of the situation we are in. In other words, what we learn has an adaptive logic.
There is a term known as biological preparedness. Biological preparedness is a theory that suggests that we are more ready to make certain associations than others and it plays a role in some forms of training and learning.
For example, in a study of chaffinches (a kind of bird), it was discovered that recorded chaffinch songs are an effective reinforcement to get them to perch in a certain spot, but not for training them to key-peck. Yet, food was an effective reinforcer in training them to key peck, but not for training them to perch in a certain spot. Rats are more likely to learn to press a lever to get food pellets, but not to avoid shock. However, they do readily learn to freeze or run to avoid shock. Pigeons will learn to fly from one perch to another to avoid shock, but won’t learn to key-peck to avoid shock.
So we have to ask ourselves, if food is a reward, what are dogs biologically prepared to do to obtain food? To what degree is there a biological basis to obtain food by being fearful in nature? How motivated is a dog to use fear to obtain food if fear makes them lose their appetite? More on that in a moment.
This question of what we are willing to do brings us to another complication: fear is an emotion. It’s not something we do. In other words it’s a reaction – primarily an involuntary one. However it is definitely possible to strengthen – reinforce – an emotion through practice. The more a dog becomes fearful, the more likely those neural pathways will become more efficient at processing it.
But we are not talking about practicing; we are talking about making the link between fear and food. So let’s have a look at a couple of different types of associative learning starting with classical conditioning. Is classical conditioning a possible model for reinforcing fear since classical conditioning deal with reflexes?
Can classical conditioning using food reinforce a dog’s fear?
Classical conditioning is the learning where a reflex occurs in response to something that does not ordinarily produce the reflex. This happens by learning associations and predicting the occurrence of events as a result of those associations. Pavlov’s dog salivating at the sound of the bell when the dog learns that that food is the follows the sound of the bell, is the classic example). It is largely involuntary learning.
But assuming that following fear with a treat will cause the dog to be more fearful by using this example ignores an important gap. Dogs do not tend to react to food with fear. They react to food by salivating. This is an involuntary reflex. They don’t choose to, it just happens. When they hear the bell, they start to think about food. The thought of food makes them salivate. The thought of food does not make them fearful. So from this perspective we know from classical conditioning experiments is that the association depends on the dog’s natural responses. That means food must be the neutral stimulus and come before the fear.
The way to make food be a predictor of something scary is to pair it with an aversive experience. The food in this case needs to be a predictor of the thing the dog is afraid of, in which case the food is a neutral stimulus and not a reward. For this to happen, food needs to consistently produce the shock (or the fearful thing) immediately after so the dog makes the prediction, or the association becomes weakened. In this case food might be a reinforcer of fear, but it is not considered a reward so this is the wrong model to consider.
But even in this scenario, food as a reinforcer is problematic. There needs to be a clear and consistent link between the two. What happens when we feed the dog other times and it doesn’t immediately produce the fearful thing? The food stops being a predictor of the thing that causes the fear and the association weakens.
Food is has been associated with pleasure since the dog was born. The association is very strong in the brain on many levels. Given that strong neural networks compete against and bully out weaker neural networks, it would be very difficult to replace the associations dogs have with food with other associations. It’s not impossible, but chances are the dog would be much more likely to associate the specifics of the situation as a predictor of fear than the food itself.
Can operant conditioning reinforce a dog’s fear?
Rewards are used in Operant conditioning so what about operant conditioning then? This kind of learning holds that if reward follows a behavior, the dog is more likely to repeat that behavior in order to obtain the reward. For example: every time a dog sits and he is given a treat, he learns to predict that a treat will follow if he sits. Once he makes the association, he actively – voluntarily – tests the theory.
This is not a passive response. The dog is using his attention and problem solving abilities to think about how to get the food. Once he discovers it works, he continues to do it. If the treats stop completely, he eventually determines it doesn’t work and gives up. So we see there is a rational cognitive process here.
But what we have learned from research on the brain is that fear causes activation in the part of the brain known as the amygdala. In times of stress and strong emotions, the amygdala will hijack our rational cognitive processes. This is one of the cautions we need to be concerned with when treating dogs that have had a history of aggression. Learning and memory can be affected by stress, in some cases by enhancing memory, and in other times, impairing memory. It is thought that stress enhances memory when the stress is related to what is being learned. That is to say, context is important.
The body and brain appear to be set up to make stress a high priority, meaning that whatever is stressing us out; whatever we have determined is threatening us: is going to be the top priority for the allocation of our brain’s resources. So what is the dog concentrating on in times of high stress? Does food represent threat or safety? Or is the dog even making the link at all between the thing he is afraid of and the food?
Again we have to ask how clearly is the fear and food linked together and is the dog in the right mind space to rationally make the connection to such a degree that he voluntarily chooses to do whatever causes him or her to be fearful. Because most creatures do not enjoy being fearful, how willing are they to be fearful to get food?
Another consideration is appetite. When a dog is experiencing fear the sympathetic nervous system kicks into gear (casually known as the “fight or flight” system), and the dog prepare to freeze, flee or fight. Part of this involuntary response is an increase in vigilance and the suppression of appetite. In other words, the dog is not in a state to desire food. This is one reason why dogs won’t take food under stress or if they do they don’t appear to enjoy it. Without that desire, the food no longer serves as a reward.
In the case of food rewarding aggression, the same issues of biological preparedness, the activation of the sympathetic nervous system and cognitive process are still a factor in whether or not food can actually serve as reward for aggression as well.
So in summary, while emotions can be reinforced, and treats can be used to reinforce certain behaviors, food does not tend to reward or strengthen fear. However, if aggression drives the fearful thing away, then aggression can be reinforced. The reward? The cessation of fear.