Trouble Shooting Counter Conditioning

15.11 in conditioning, counter, Dog, shooting, trouble

This video is from Emily Larlham.  Her site www.dogmantics.com has many more great videos. I have chosen this one today as it trouble shoots problems that people have when using Counter Conditioning. When people tell me that their dog isnt responding to Counter Conditioning (the process of pairing something nice with the appearance of something they previously or currently find unpleasant or scary) I often have to find out why this is. This short video explains many of pitfalls that people will have. Including when the dog hasnt associated that the apperance of the "scary thing" triggers the rewards from the handler. When the dog is over aroused by the rewards, when the dog has incorrectly paired the rewards with the clicks or the marker work "yes".  If you are interested in seeing more of Emilys videos search "Kikopup" on You Tube. 

Happy Watching, Miranda xxx

Read more

Counter conditioning and desensitization

19.01 in and, conditioning, counter, desensitization, Dog

These two techniques are often used to change unwanted behaviour in dogs and cats. Just as the term implies, counter conditioning means conditioning (training) an animal to display a behaviour that is counter to (mutually exclusive of) an unacceptable behaviour in response to a particular stimulus. For example, a dog cannot be trying to bite the post lady and at the same time greeting them in a friendly, excited manner.

Desensitization involves gradually exposing a pet to the situation, without provoking the unwanted reaction. If an animal is highly motivated to perform an undesirable behaviour, and if that behaviour is easily and quickly displayed, competing behaviours may be difficult to elicit. That’s where the desensitization part of the process comes into play.

Desensitization is the process of exposing an animal to a stimulus beginning at a very low intensity. So low that it does not result in the undesired behaviour. For example, if a dog becomes fearful and barks at visitors, then the first step would be to find a distance at which the dog does not bark, growl, attempt to flee, or show other signs of fear. The stimulus intensity is then increased gradually (bringing the dog closer to people in the example), without eliciting the unwanted behaviour.

Performed simultaneously, these techniques provide a way in which an animal can be gradually taught to show acceptable behaviour in the face of a stimulus that used to elicit reactive behaviour.

Counter conditioning and desensitization must be implemented very systematically. If the incremental increases are too large, or occur too quickly, the techniques will either not be effective, or may even make the problem worse. Implementing a counter conditioning and desensitization program requires some thought and planning.

Define the starting point. Ideally, a behaviour modification program of this sort should be designed and carried out in such small steps that the problem behaviour never occurs. This means all the stimuli that elicit the behaviour must be identified and ways found to lower their intensity until your pet doesn’t react to them. For example, if a dog becomes afraid if someone approaches closer than 6 feet, then the starting point would be a distance significantly greater than 6 feet. In order for these techniques to be most successful, your pet should not be put in any situation that triggers the problem behaviour.

Define the dimensions or characteristics of the stimulus that influence your pet’s response. For example, if we are working with a dog who is afraid of people, we need to know which aspects of that process influence the dog’s fear:

Is she more afraid of adults than children?
Is she more afraid of men than women?
Is she more afraid of a family member or someone she doesn’t know?
Is she more afraid when someone moves fast or slow?
Is she more afraid in a particular room?
Is she more afraid if the person speaks to her or is silent?
Is she more afraid if someone is sitting or standing?

Some common factors to consider include location, loudness, distance, speed of movement, length of time near the other animal or person, response of the other animal or person or body postures of an animal who induces fear or aggression.

Arrange these characteristics in order from least to most likely to produce a negative response. A counter conditioning and desensitization program needs to begin by using combinations of stimuli that are least likely to cause a fearful reaction. In our dog example above, perhaps the dog is least afraid of being handled by a familiar adult female who approaches slowly and speaks softly to her, while she’s lying on the bed in the bedroom. She is most afraid of a nephew who runs up to her yelling while she’s in the kitchen.

Always begin with the characteristics or dimensions that are least likely to elicit the problem behaviour. We would begin with the easiest combination of characteristics of the situation, and gradually work up to the most difficult. If we find that this cat will be less afraid of a male child approaching slowly than an adult female approaching fast, then we know speed of approach is more critical than type of person. Thus, the working order on these two characteristics, from easiest to hardest, would be:

Adult female, slow approach

Male child, slow approach

Adult female, fast approach

Male child, fast approach

If necessary, devise ways to make each dimension less intense. If a dog is afraid of the sound of the hair dryer, the sound must be presented to the dog at a low (sub-threshold) intensity, one that does not provoke the fearful behaviour. This could be done by turning the dryer on and off quickly before the dog showed fear, turning the hair dryer on in another room, covering the dryer with towels, etc.

Pair each level of each characteristic with a positive consequence, as long as the problem behaviour is not displayed. At these sub-threshold intensities, the stimulus must be paired with something positive for your pet. In this way, the animal comes to associate good things happening in the situation rather than bad things. Alternative behaviours such as calmness and friendliness are then reinforced instead of fear, aggression, etc. being elicited. The reinforcement must be powerful. Good choices are food, especially favoured treats, toys, or social reinforcements such as petting, attention, or praise. If food is used (and it’s almost invariably helpful), it should be in very small pieces and be highly desired by your pet (cheese, hot dogs, or canned tuna often work well). You may need to experiment a little to see what food is the best motivator for your pet.

Do not progress to the next level until your pet is clearly anticipating the reinforcement. People commonly want to know how long they need to repeat each intensity level. This will depend entirely on your pet, who should be demonstrating that he is indeed expecting good things to happen. Perhaps he looks to you for a tidbit, or looks around for his toy. This should be in contrast to his previous reactions such as trembling, tensing up, or other fearful or aggressive responses.

Don’t make all dimensions more intense at the same time. In our dog example above, notice that we increased the intensity with type of person, and subsequently increased the speed of approach. We did not try to do both at once by shifting from a slowly moving adult to a fast child.

Progress slowly. It is common for people to tell me, “I tried some of those ideas and they didn’t work.” Usually what has happened is that they tried to progress too fast, didn’t take small incremental steps, or didn’t use highly motivating rewards. Counter conditioning and desensitization take time and must be done very gradually. You will need to think through the steps you need to take. Rather than expecting progress in terms of leaps and bounds, look for small, incremental change. It can be very helpful to keep a log or record of your results, since day to day changes will not be very big.

An example of a somewhat simplified counter conditioning and desensitization program for a dog who displays fear-motivated aggression toward men could be:

• Identify the point at which fearful or aggressive behaviour is first elicited (e.g., when the man is 6 feet away and is approaching to pet the dog).

• Begin the program with a situation the dog will tolerate without becoming aggressive or fearful (e.g., the man walks by at a distance of 7 feet, with non-threatening body postures, paying no attention to the dog).    

• Encourage the dog to assume a confident posture on a leash such as standing, walking, or sitting.

• Offer a small food reward and/or toy to generate expectant, excited, non-fearful behaviour.

• Respond in an up-beat manner, petting, praising, and talking “happy talk” to the dog.

• Conduct several sessions of a few minutes before making it any more difficult for the dog.

• Instruct the man to stand 7 feet away and make a small arm motion as though he was beginning to reach out and pet the dog.

• Give the dog food and praise for non-aggressive, non-fearful behavioural this happens, also conducting a set of short sessions.

• Practice this scenario until the dog is anticipating the food reward or the toy.

• Slowly decrease the distance between the man and the dog, adding to the arm motion on such a gradual basis that fear or aggression is never elicited.

• Many repetitions with more than one individual may be required. A new step should not be taken until the dog is clearly anticipating the reward.

You may need to supplement the behaviour modification program with other approaches, such as avoiding situations that provoke the problem, using a head collar like the Gentle Leader collar, or treating your pet with anti-anxiety medication. Ask your Sussex County Dog Training instructor or Vet for futher information on this.

Read more

Operant and Classical Conditioning

14.44 in and, classical, conditioning, Dog, operant

Operant and Classical Conditioning

Operant conditioning is a defined set of principles, developed in the 1930’s by a behavioural scientist called B.F. Skinner. The theory says we can make things happen, either good or bad, by choosing behaviours that are either rewarded or punished. We use the 4 methods Positive Reinforcement, Positive Punishment, Negative Reinforcement and Negative Punishment and where the dog acts upon the environment. Dogs will not generalise well, here, and need to be trained in all different environments in order to fully understand what is required. Once they have learned the behaviour, in the front room, we then need to train that behaviour in the garden, then in the park, and so on, until the behaviour has become generalised. E.g. the dog learns that they can make good things happen by offering the sitting behaviour. Next the dog needs to learn that those good things still happen if the sitting behaviour is performed elsewhere. We are using the Positive reinforcement quadrant here to teach the dog that sit is the behaviour we want, and we want it, to cue, in all environments.

With Classical Conditioning, we are not necessarily looking to train a behaviour, yet looking for an emotional response from our dog.  Often our dogs have learnt these responses from a single incident, and thus these become generalised very quickly. Behaviours that involve a strong emotional response where the environment acts upon the dog are classically conditioned. i.e. fear of thunder and lightning or the fear of children after an incident involving a child pulling his/her ears. Next time he/she see’s a child, they can generalise very quickly and have an emotional response of fear towards children. On the other hand, as with Pavlov’s dogs, the emotional response to the bell ringing was to salivate at the prospect of food, and dogs who have had pleasant experiences with children, soon generalise that all children are nice, and their emotional response is one of happy thoughts and not fear.

When looking to apply operant conditioning or classical conditioning when dealing with a dog’s problem, first we must define the problem. Problems are often defined as problems for the owners and not for the dogs. A dog that loves to greet us by jumping up, may have been rewarded for this by a cuddle here and there, and yet now it has become a problem, especially as the dog has grown bigger and stronger. Here we would teach the dog a new behaviour (operant conditioning) as jumping up is not necessarily an emotional response (classically conditioned) but a learnt one, probably through reward.

We would in this case use positive reinforcement, and maybe, negative punishment to change the unwanted behaviour of jumping up to the desired one of sitting when you greet me.

When the dog jumps up we can turn away (negative punishment) and when the dog puts all 4 feet on the floor, we reward that behaviour (positive reinforcement) As we continue to do this exercise, the unwanted behaviour will extinguish, and the desired behaviour increase.

Alternatively, if we do not want to use negative punishment here, we can ask the dog for an incompatible behaviour to jumping up (desensitisation of incompatible behaviours), a sit, and reward this desired behaviour. Again the unwanted behaviour will eventually extinguish and the desired behaviour increase.

If we were to encounter a dog who would bark at another dog, due to an emotional response of fear (classically conditioned), we could look to change the dogs emotional response using classical conditioning, or a technique called counter conditioning.

For dogs to learn, we must ensure that they are in the right state of mind, and a dog that is so fearful may not be able to learn, as its emotions are such that it is not cognitive to learning. This happens if we subject the dog to too much of the stimuli that causes the fear in the first place. Some trainers will flood the dog here in an attempt to prove to the dog that there is nothing to fear and subsequently change the dogs emotional response and learn to be happier in the presence of other dogs. This method, in my opinion, can also cause a dog to shut down and not be able to learn that other dogs are nothing to be fearful of, or worse still lead to some aggression towards the other dogs. If we ensure that the dog is in a clam state before we introduce the stimuli, then this helps towards learning. We need to change the dogs emotional response from fear by adding something nice in the presence of the stimuli. We can use high quality treats to feed the dog, ensuring that our dog is not overly stressed at each stage. We will keep the stimuli(other dog) at a distance, where our dog is happily eating the treats, thus now pairing the presence of another dog with something nice. We can move the stimuli closer over a period of time or over different sessions, as our dog becomes more at ease, subsequently classically conditioning the dog to enjoy the presence of other dogs.

 

Jeff Sasse (2010)

Read more

4 Quadrants of Operant Conditioning

16.47 in 4, conditioning, Dog, of, operant, quadrants

Positive Reinforcement – We increase the chance of a desired behaviour happening again by rewarding it. The dog soon learns that in order to earn his reward, he just has to repeat that desired behaviour.

E.g. when the dog sits in front of us, we say “yes” in a happy tone and reward that with a treat. Marking the behaviour with a “yes” and rewarding the dog with a treat will increase the chances of that behaviour happening again. This is a strong way to enhance the bond/relationship with your dog whilst training. Rewarding all the desired behaviours is a great way to help you and your dog feel great about training, and increases the desire to train and learn new things. When we use positive reinforcement dogs learn to think more and look to offer us behaviours in order to be rewarded. This gives the dog confidence in the learning program.

Positive Punishment – We decrease the chance of an undesirable behaviour happening again by adding a punishment that the dog dislikes (an aversive). The dog learns that if he repeats this behaviour, it may be paired with an aversive.

E.g. the dog Jumps up at us, so we smack it on the end of the nose!! The dog learns that if he jumps up at us, this may be paired with a smack on the nose, and so decides against jumping up.

This method of training, will eventually break down the bond/relationship between dog and handler and if used regularly will not only stop the dog enjoying training, but may, if the dog becomes fearful lead to aggression. Dogs may become tolerant of this method, making it ineffective in changing the behaviour, yet still affecting your relationship. Unlike positive reinforcement, dogs here learn not to trust their handler, and look to suppress behaviours in case a punishment was to follow.

Negative Reinforcement – The dog’s behaviour makes something undesirable go away.

E.g. in order to increase the chances of a sit, we lift the lead, choking the dog, and remove when sitting. The dog will sit because he doesn’t like the choking when lifting the lead. Like the punishment method dogs just want bad stuff to go away. They will however pair you with the “bad Stuff” and this will break down the bond/relationship you have with them.

Negative Punishment – We decrease the chances of an undesirable behaviour happening again by removing a motivator (something the dog wants)

E.g. the dog jumps up to get our attention (something the dog wants), so we turn our back (removing ourselves) in order to reduce the behaviour happening again. This type of punishment is a non-force method and can include removal of a treat, a toy, a person or even a privilege. This is a non-aversive method of providing a consequence for an unwanted behaviour. This type of punishment is best paired with positive reinforcement, which quickly teaches the dog rights and wrongs. In the example above, if we rewarded the sit on the floor with a treat or our attention (positive reinforcement) then we would not only look to extinguish the unwanted jumping up, but also increase the chances of the sit on the floor.

 
Jeff Sasse(2010)

Read more