Dog Training/dog aggressiom progressively getting worse
I just adopted her a few months ago. She is 4 years old and was in the shelter for 4 years, or so they told me. She is a boxer mix. She loves people but not somuch with dogs.
my dog has an aggression problem towards ALL new dogs she sees. She gets so aggressive itsnjot barking first then attacking, she goes straight to attack mode even with nice calm dogs! I have been strict & training her vigorously but she seems to be getting worse! Im at my wits end, i have no idea what to do!
There are two methods to try and help her with this and I have listed them below.
I would suggest finding a Canine behaviourist in your area for help with these protocols.
Behavioural Animal Training - BAT
BAT uses functional and bonus rewards, combined with clicker training and systematic desensitization to help dogs make better choices in an error-free way.
The things that you find obnoxious or scary (barking & lunging, for example) serve a purpose for the dog. That is the idea behind what behaviour researchers call Functional Analysis. Imagine you hated your high-stress job but loved the pay. It leaves you exhausted and panting, but you enjoy the reward at the end of it. Given the chance to earn your same great pay check by doing something easier that you liked more, wouldn’t you do it?
BAT is similar to a method used with humans, called Functional Communication Training. Students learn to communicate their needs instead of using aggression or other problem behaviour. Sound familiar?
Functional Communication Training used with children with developmental disabilities, is one of the inspirations for BAT, along with the best parts of clicker training, systematic desensitization, Constructional Aggression Treatment, the Two-Reward method, Treat & Retreat, click and retreat by Alexandra Kurland, and more. Functional rewards can fall into the quadrant of positive reinforcement (like ‘real life rewards’) and negative reinforcement (like relief from social contact). The upcoming book on BAT goes into that in more detail.
Functional rewards are not just for aggression or fear, but eliminating distress that leads to reactivity (including frustration, anger, and fear) is the main thrust of BAT. BAT uses Differential Reinforcement of Alternate behaviours (DRA) with environmental rewards: the ones the dog is already working to get. BAT also uses ‘Bonus Rewards’ when necessary for motivation – food, toys, etc. The key to BAT working quickly is low stress, never leaving the dog in the ‘deep end of the pool’ or punishing them for incorrect choices.
First, here’s why I’m calling it BAT. The goal is a well-adjusted dog and behaviour is all we’re really trying to adjust. When you’re done, you have the same trigger as before, and the same consequence, but a new behaviour. The core idea that I love from Functional Analysis is the idea of using the reward that the dog is working to earn to reinforce a new, more ‘appropriate’ replacement behaviour. So if the dog is already getting paid for barking at the postman by having him leave, the new set-up is to pay the dog for calm behaviour by having a pseudo-postman leave as a consequence for good behaviour.
With the Antecedent -> Behaviour -> Consequence model, you have something like: Antecedent -> Behaviour 1 -> Consequence being replaced by Antecedent -> Behaviour 2 -> Consequence (Think of ‘antecedent’ as a cue or something that signals the behaviour is about to be paid for.) So Postman arrives -> Barking -> Postman leaves is replaced by Postman arrives -> dog turns away -> Postman leaves I love that part. It means that the environment itself, which created the problem, will now start rewarding your dog.
You will need to have several friends practice being the postman before it works for real. The bigger picture of BAT is that it includes real-world positive reinforcement, too: Mom comes home -> jumping -> petting/attention is replaced by Mom comes home -> sitting -> petting/attention Not that that’s new, but I love that the BAT model fits in with what we already do! The behaviour is what we adjust, within the environment that provides the antecedent and consequence. That’s why I’ve called it Behaviour Adjustment Training. Now to get that behaviour change, we have to make sure that either Behaviour 1 -> Consequence is a chain that gets broken, as we do when we ignore jumping, or we work on more errorless learning. What I mean by that is that we start with an antecedent that is different enough from the usual antecedent that the dog doesn’t do the naughty Behaviour 1.
It gradually morphs over time until the picture becomes Antecedent -> Behaviour 2 -> Consequence, as we wanted. Postman really, really far away -> dog turns away -> Postman leaves … Postman really far away -> dog turns away -> Postman leaves …and so on, until we have Postman arrives -> dog turns away -> Postman leaves
In that example, the dog is still getting rid of the postman. With reactivity to strangers or other dogs, the idea is that the dog learns that it’s safe to approach. Curiosity can blossom, which then leads to trust and sociability. For example, my dog who was afraid of people now leans in for petting from strangers, because he learned that he could just walk away when he felt a little nervous.
Here is the BAT protocol for aggression or fear in its simplest form.
1. Expose: Start sub-threshold and remain below the dog’s threshold as you increase stimulus intensity a little (e.g. student dog moves closer to stooge (a.k.a. helper/decoy/actor), stooge moves closer, or both). If dog is getting worse instead of better, abort.
2. Wait for or manufacture acceptable alternative behaviour in a non-aversive way. Be sure to take very small behaviours, like blinks or head turns. Waiting is preferred, whenever possible.
3. Mark using a verbal Yes or a clicker.
4. Functional Reward: Decrease stimulus intensity (but not to zero). ex. Student gets to walk 20 feet away, but stooge remains in view.
5. (optional) Give a treat or toy as a Bonus Reward (best to use on Walks, not usually necessary or desired for set-ups).
This need not be a big, long set-up. It can just be a single repetition, in the middle of doing something else, like TTouch ground work or passing by a gardener in front of her home. Whenever the dog is in a slightly stressful situation and they do a nice appropriate signal (instead of aggression) you could mark with “Yes” and retreat with them as the reward. What’s very cool about BAT for reactivity is that the dog starts to actually become friendly to the decoy, even though all they wanted, at the beginning, was more distance. It’s as if they now have the locus of control and the world makes sense again. Their peaceful choices are controlling their environment, and having an internal Locus of Control feels good!
Unlike counter-conditioning, which assumes that emotions drive behaviour, BAT acknowledges the theory that behaviour and its consequences can also drive emotions. I think both are probably true, so I use both CC/DS and BAT, in combination or in different circumstances. Key points to remember when doing BAT for aggression or fear:
• Above all, stay sub-threshold. Take frequent breaks, especially when the dog asks for them.
• If the dog becomes agitated (breathing heavier, looks like he might have an outburst), interrupt the dog or counter-command (give the cue for an easy incompatible behaviour) and then decrease the stimulus intensity (for example, walk away from the decoy or have them go away, whichever is least reinforcing). If the dog has an outburst and you just stand there, you’re wasting time, because he’s more sensitive to the stimulus for your next trials, and the rate of reinforcement goes way down. Barking/lunging is also hard on everybody else’s stress level, including the stooge and the student dog’s owner.
• Shape for a nice big set of clear cut-off signals. Don’t go for variety too soon. The student dog (or horse or goat or person or whatever) should experience rapid success. You can repeat this later to look for variety. Replacement behaviours include: blinking, jaw loosening (being able to pant again), looking away, turning away, ground sniffing, air sniffing, tail carriage getting looser.
• Reinforce calmly gathering information with allowing the dog to continue doing so, or with quiet praise.
• I really like having the student dog able to move, rather than being tethered, and able to walk away from the situation (safely leashed or behind a barrier, but not tethered to a wall). If the student requests a break by checking out, she gets one. Dogs should have control of their exposure to the trigger during the session (your job is often just to keep them from getting so close that they go over threshold).
• For most dogs, especially fearful dogs, I prefer for them to retreat from the Scary Monster as their reward when they’ve done an acceptable alternative behaviour, versus only having the monster leave. Fear is the emotion of ‘get me out of here!’ so it makes more sense to me that the student dog gets to leave versus chasing off the bad guy. You do need to have trials where the stooge approaches and retreats, but I think the bigger reward for a fearful dog is being able to leave. So even if the stooge approaches the student dog, the student dog still walks away as the reward, simultaneously or just after the other dog leaves. Even dogs who are offensively aggressive (angry, territorial, whatever you want to call it) benefit from learning how to just walk away. Lateral/tangential retreats work well for them, at first.
• If the student dog barks on retreat, continue to retreat, but remember that next time you’ll need to tone down the stimulus (a bit farther away, less motion or whatever).
• It is totally reasonable to pre-train some of these behaviours via clicker training without the decoy there.
• What about regular walks between BAT sessions? Any time you encounter triggers that will hold still (dogs behind fences, people out gardening, etc.), you can do BAT by approaching and retreating based on the dog’s behaviour.
• On walks where you don’t control the triggers, follow the steps on BAT Handout PDF Document by Grisha Stewart. The core of BAT is simply that you use environmental rewards when you can, to change A -> B1 -> C to A -> B2 ->C. When you can’t use environmental rewards, at least give the dog a chance to earn reward somehow, even by just looking at the trigger!
• You don’t need to do a marathon session to complete switchover (where you can tell trust has built and the dog is seeking social contact), although you can.
• The first session of BAT may take a long time to switchover (the point where the dog wants to be near the stooge versus only leaving), the second will take less time, and so on. Eventually it’s like the real world, with no time to switchover. That’s how you know the dog has been rehabilitated (assuming it’s a generalized response). If you don’t get to switchover in the first session, that’s fine. Just try to end each session with some down-time, hanging out at whatever distance feels comfortable to the dog. You can use the same stooge again and work up to socializing with them, unless that’s not safe.
• Remember that we’re teaching new skills to negotiate with the Scary Monster, so the first several times that you do this will take more time than, say, counter-conditioning, or simply parallel walking. But as I said before, the time-to-switchover shrinks rapidly. And now the dog has a new set of skills.
• This involves some negative reinforcement, because ‘relief’ is the treat. While I would never apply pain in order to remove it, this kind of negative reinforcement is very natural, and done right, need not be too aversive, and we need not apply anything aversive to the dog in order to take it away as a reward. In many cases, their own species is the unwanted stimulus and in most cases, they are walking themselves up to the approach line. They will be exposed to other dogs all the time anyway. They are already learning behaviours anyway. We are just teaching them different behaviours, using their natural environment.
• The marker signal of “yes” becomes a classically conditioned safety cue for the dog, because it has signalled the end of a trial and a retreat.
Counter conditioning and desensitization are the cornerstones to treatment of fears and anxieties. Yet the treatment of these conditions can be difficult. First, it is important to accurately identify the fearful or anxiety producing stimulus. Then, it is important to break the stimulus down into discrete units. Progress with treatment may be slow and animals may not generalize from the treatment sessions to real experiences. In order to be successful in modifying behaviour in dogs and cats a good understanding of learning, reinforcers and punishers is necessary.
What is learning? Learning is the relationship between behaviour and consequence, between making a response and the outcome of that behaviour. The results obtained affect later behaviours by either increasing or decreasing the likelihood of future similar responses. What is important is to realize that behaviour is something that occurs all the time. Therefore, when you remove a behaviour from an animal's repertoire, it will be replaced with something else. The goal of behaviour therapy is to structure that replacement behaviour in the correct way.
Since behaviour is the result of behaviour-consequence relationships understanding how they function is helpful. Generally speaking there are four types of behaviour-consequence relations. First, behaviour can result in positive consequences. This is positive reinforcement and will produce an increase in the behaviour that resulted in the positive event. Second, behaviour can result in negative consequences. This is punishment, and should result in a decrease in the behaviour that caused it. Third, the behaviour can result in the removal of something unpleasant. This is negative reinforcement or escape and will increase the likelihood that the preceding behaviour will occur again. Lastly, there are behaviours that result in the elimination of something pleasant, and this is called omission. 1 Learning occurs best when there is a clear relationship between the timing of the two events (behaviour and consequence) and the predictiveness of the consequences.
Learning principles and behaviour modification techniques
The previous discussion showed that behaviour is controlled by its consequences-either pleasant ones or unpleasant ones.
Reinforcement is a positive relationship between behaviour and outcome. The more you do, the more you get, and what you get is good. In other words, behaviour will be repeated. There can be positive reinforcement, often called a reward. Positive reinforcement and rewards are often used synonymously, but that is not always true. It is easy to see a food treat or a pat on the head as a reward for the dog. However, it is often more difficult to determine why some behaviours still exist because the reward is not clear. For example, what it the reward for a dog when it barks at the mailperson? The reinforcement comes from the mailperson leaving, and therefore the dog will continue to bark. The dog has erroneously assumed that the barking behaviour made the person leave, and therefore will engage in the behaviour again.
Negative reinforcement is the removal of something unpleasant that increases the likelihood that a behaviour will be repeated. In this situation there is a negative relationship between behaviour and outcome. One way to look at this relationship is to realize the more an animal engages in the behaviour, the less negative outcome is obtained. The easiest example of negative reinforcement to understand is escape behaviour. If an animal is anticipating an aversive outcome, perhaps a reprimand from the owner, if the animal can escape, or not be caught, then the aversive outcome will not occur.
Punishment is a situation where there is a positive relationship between behaviour and outcome, but the outcome is negative. In other words, the more of a behaviour an animal does, the more of a negative outcome is obtained. This is a situation that should make behaviour decrease. This is an extremely important component of punishment. When punishment is used to change behaviour, there should be a decrease in the target behaviour in very few applications of the aversive event. If not, then either the punishment is not being appropriately applied, or applied to the incorrect behaviour. Punishment also has the possibility of causing anxiety, fear and aggression and is not the recommended means of changing behaviour.
Schedule of reinforcement
The way the both reinforcement and punishment are used can greatly influence their effectiveness. This is often termed the "schedule of reinforcement". How behaviours are rewarded can be powerful determinants of future behaviour. Schedules of reinforcement can be based either on time (intervals) or amount of work, or number of responses (ratio). Schedules can be fixed, meaning that after so much time or a set number of responses reinforcement is given. Alternatively, behaviour can be reinforced on a variable schedule, meaning that a period of time or a number of responses must take place, but that period varies from reinforcer to reinforcer, with an average time or number of responses. Reinforcement given on a variable schedule results in strong acquisition of the response. This means that the rewards are given intermittently, and the animal is not exactly sure when the reinforcement or the punishment will occur. The outcome also must be closely associated in time with the behaviour for best results.
Extinction is a procedure used to end a behaviour. Extinction occurs when behaviour is no longer reinforced. When this occurs, eventually the behaviour will stop. However, it is very common, especially if the behaviour has been maintained on a variable ratio of reinforcement, for the behaviour to temporarily increase and this is called an extinction burst. When trying to get a behaviour to extinguish it is extremely important to identify ALL reinforcers and eliminate them.
Habituation is a process by which a stimulus no longer evokes a response. Usually this occurs with repeated presentation of a stimulus and the animal learning that it does not signal anything important.
Flooding is used to treat fears of harmless stimuli by forcing the animal to stay in the presence of the stimuli until the fear is extinguished.
Classical conditioning is the pairing of an unconditioned stimulus, with a neutral stimulus that results in a conditioned stimulus and a conditioned response. Classical conditioning can occur in both positive and negative ways. The timing of the presentation of the stimulus, the saliency of the stimulus and the predictability of the stimulus and the reinforcement influence the conditioning process.
Conditioned emotional response refers to establishing fears through a classical conditioning paradigm. This entails the association of a fear-producing stimulus with a previously neutral object. This type of learning can be very powerful and hard to extinguish.
Operant conditioning is learning how ones actions result in consequences; i.e. the individual causes the results. This is a stimulus-response/response-consequence relationship. In other words, what the animal does is critical to what happens next and those results dictate if the behaviour will occur again. Behaviour becomes more likely if it is reinforced, less likely if it is punished.
Counter conditioning is teaching a behaviour that is incompatible with the previous response. An example is to teach a dog to sit and stay instead of lunging. What is wanted is that the response be behaviourally and physiologically different from the previous response. Therefore, facial expressions, body postures, respiratory rate etc. are all-important components in the response. The goal is to change the association with the stimulus. Classical counter conditioning occurs when you pair a previous stimulus with some unconditioned response such as food.
Systematic Desensitization is gradually exposing an animal to stimuli at a low level so as not to evoke an undesirable response and conditioning relaxation responses instead. Paired with counter conditioning, this allows animals to learn to behaviour properly to stimuli that caused fear, aggression or other problem behaviours. The stimuli must be presented on a gradient from low to high without evoking the inappropriate or unwanted response. Therefore, the arrangement of the stimuli becomes very important.
Obstacles to treatment success in counter conditioning and desensitization (CCDS)
Generally there are five obstacles to treatment success. First is stimulus discrimination, the ability of the animal to distinguish the stimulus. In addition, the stimuli presented must be relevant and control the behaviour or the animal does not learn the appropriate response. Then the animal must learn what to do in the presence of the stimulus. Second, transfer of learning must take place for CCDS to work. The animal must learn to pay attention to the relevant stimulus and ignore irrelevant stimuli. Third, the animal must learn to generalize from the learning situation to the real world. This requires the behaviourist to know what stimuli are controlling the response. Fourth, inappropriate rewards may allow the animal to discriminate improperly and learn a different stimulus-response relationship than what was intended. Finally, the animal can be come more sensitive rather than less sensitive to the stimulus.
To overcome these obstacles, accurate history taking, good observational skills and appropriately set up treatment plans are important. The behaviourist and the owner must be willing to proceed slowly and set up the animal to succeed. Finally, the animal must be exposed to a variety of stimuli once the behaviour is learned.
Changing behaviour takes a good history, a realistic treatment plan and good supervision and cooperation between behaviourist and pet owner as well as a complete understanding of learning and behaviour modification techniques.