Positive Training using food treats.
“Dogs are not obedient to commands: they are obedient to the laws of learning.” Jean Donaldson
Anything that will increase the likelihood that an act will occur again is called a reinforcer. A positive reinforcer, in dog training, is something that the dog wants, such as food, a toy, petting or praise. Simply offering food treats as positive reinforcement to strengthen any behaviour is the most rudimentary part of dog training. And for a puppy, what can be better than a nice treat? Fortunately we do not have to teach a dog to eat. Behaviour that is already occurring, no matter how seldom, can always be intensified with positive reinforcement training.
There is an often stated misconception that using food is “bribery.” That by using a positive reinforcer such as food treats you will always have to carry treats with you and when the food stops the dog won’t listen anymore and the unwanted behaviour returns. This is not true because constant reinforcement is only needed in the learning stages. You do not still clap hands when your teenager uses the toilet or rides his bicycle.
Food can be used as a lure or prompt to capture a position or movement but also as a reward. When you let a dog sniff a treat in your hand and you then back away for the dog to follow you, you have lured his coming to you. If you then give him that same treat, you are using it as a reward. Problems arise when the dog learns that the reward was the same one used to lure him. So when food is not used as a lure anymore, he knows there is not going to be a reward.
People complain that their dogs only obey when they have food in their hands. What they don’t realise is that they have actively trained this over time by constantly giving a reward when it was visible and almost never when it was not visible. Dogs will do what works for them and soon learn that if no reward is visible means that there will not be a reward.
When a dog can see the food he desires, in his mind, he is doing things (working) to get the reward he is seeing. He is working for/ to get the treat = reactive dog behaviour (driven by the reward itself). But, by now using “Yes” when he makes eye-contact, i.e. asking “where is my reward?” – without him seeing it – he is working to get the handler to produce the treat reward that was hidden from his sight = active dog (operant dog).
In order to maintain an already-learned behaviour to become more reliable you do not have toreinforce it every time. It is actually vital that you do not reinforce on a regular basis but instead change to reward intermittently: on a random and unpredictable basis. A variable schedule is far more effective in maintaining behaviour than a constant, predictable schedule of reinforcement.
I recently replaced my lawnmower engine with a new one. It started well while very new. Soon, however, I had to pull many times before it started. It got so bad that I used to get into the habit of counting how many times I had to pull to get life into it. What had happened is that because I had a random starting response, I kept trying harder to get it going. This is what is happening when our treats are given in random or unpredictable fashion. The animal is kept hoping and trying harder for the treat he knows will come.
We know that we do not have to go on punishing bad behaviour that has stopped, but we do not see that it is not necessary or desirable to reward correct behaviour continuously.
A variable schedule of reinforcement means that sometimes you treat a correct behaviour and sometimes you don’t. However, puppies may have to be helped to tolerate the occasional skipped reinforcer after a period of predicted rewards. Let the dog do it twice – two paw lifts or bumps. Reward at the end of a sequence of moves. Sit down-reward. Sit down sit-reward. Sit down sit stand-reward etc. With a lure/treat in your hand get the dog to make a number of changes as before but do not reward at the end. Then, without the dog seeing, put it in your other hand which is behind your back. Show the dog your empty hand and ask him to do one of the positions like down. As soon as he does it, immediately give him the hidden reward from your other hand. Repeat by rewarding from different hands. The dog must learn to work even if you do not have food in your hand.
Intermittent means that the dog does not always get rewarded with food every time he responds correctly, but with praise or a hug or a release. Because the dog never knows when a treat is coming he will continue to respond in anticipation that food will appear again in future.
Timing is information. It tells the dog exactly what it is you like. When the dog is still learning it is the “Yes” or “Good” that becomes even more important than the food. Beginner trainers often struggle to get their timing right. They get the puppy to sit but by the time they say “Good dog” the dog is standing again. Whenever you have difficulties in a training situation, the first question to ask yourself is whether you are reinforcing too late. Giving treats for behaviour that hasn’t occurred yet does not improve what is happening at the time. This is a major problem with owners who treat without thinking.
At my club we have found that the beginner trainers are often confused as to how big a treat must be. Some arrive with cheese, chicken or such large tidbits that the dog has to stand chewing before swallowing and continuing with the exercise. Our guide is, “No larger than the nail on your pinkie.” In fact, the answer should be: as small as you can get away with. The smaller the treat, the faster the dog eats. It not only cuts down on wasting time, but it allows for more reinforces per session, before the animal is full. Early morning training sessions should contain about a quarter of a dog’s rations.
The jackpot is a reward that is much bigger, maybe five times bigger, than the normal reward. It is unexpected and may be used for a sudden breakthrough. The dog performs a difficult task correctly for the first time or his recall was very fast etc. It could be effective to improve fearful or object resistant behaviour.
Food and the brain
Food, we know, is very important to a dog but scientific studies have found that food literally alters an animal’s brain chemistry. It not only improves a dog’s ability to learn but also helps it to overcome fear and anxiety by raising the levels of dopamine in the brain and stimulates the desire to “seek” and move towards the food reward.
Dopamine is a neurotransmitter that plays a major role in reward-driven learning and helps regulate movement and emotional responses. If a dog is given food before he reaches a high stress level when facing something that scares him, then a positive emotional response occurs. (High stress levels lead to a fight or flight situation. Cortisol and Adrenalin are pumped into the brain and the dog goes into overdrive.)
There are circuits in the dog’s brain that encourage seeking or hunting behaviour and circuits that elicit the fear response. Presenting food to your dog turns on the seeker system and effectively turning off fear. At the club I regularly demonstrate this by taking two dogs that are aggressive towards each by letting them stand on either side of me. I offer both a food treat and turn on their thinking brain and deactivate the emotional brain. Repeating this action, I bring my hands closer to each other and in so doing bring their heads closer until these enemies virtually touch noses.
Because food is incompatible with fear, using food treats for teaching is incredibly valuable, especially when it comes to modifying a dog’s anxiety and fear. That is why the food used to modify a dog to learn must be of high value to him until he responds reliably.