Treatment of Behavioral Problems
In production animals, treatment focuses on group management, environmental or housing modifications, and in some cases removing individual animals out of or to other groups. Specifics are covered in the relevant species discussions.
In companion animals, the treatment of behavior problems varies with diagnosis and prognosis. In general, the program begins with prevention and avoidance of problems, while the owner develops effective strategies to modify the pet’s behavior so that it might gradually be reintroduced to the problem situations while achieving a desirable outcome. Initially, prevention is necessary to avoid further compromising the pet’s welfare and to ensure safety in cases of aggression. Repetition of the behavior further aggravates the problem if the pet successfully accomplishes its intended goal (eg, escape or retreat from the stimulus), while each exposure in which the outcome is unpleasant can condition further anxiety. Improvement is generally a slow and gradual process; therefore, owners must have realistic expectations of what might be achieved. Modifications to the environment may be required, so that the pet can be kept away from the stimuli (or the sights or sounds of the stimuli) that incite the problem or from the areas in which the problem occurs. Modifying the pet’s behavior is accomplished by applying the principles of learning and behavior modification, primarily achieving and rewarding desirable outcomes along with use of products that improve safety, reduce anxiety, or help to achieve the desired response more effectively (eg, muzzles, head halters, no-pull harnesses, etc). Drugs and natural products may also be indicated for some pets and some problems.
The most commonly used behavioral techniques include habituation, extinction, desensitization, counterconditioning, response substitution, and shaping. Flooding is often talked about but seldom used because it is likely to make most animals worse. While punishment is frequently used with varying degrees of success, few people correctly use this technique, and there are both humane and safety issues with the use of positive punishment. For punishment to be successful, the aversive stimulus (eg, startling with a loud noise, spraying compressed air) must occur sufficiently close to the onset of the behavior that the probability of the behavior occurring in the future is lessened. Often, punishment is more about the owner’s anger than about changing the behavior. In addition, some dog owners have been ill advised by training advice that advocates confrontation, with the intent of asserting leadership (dominance). In fact, numerous studies have demonstrated that punishment-based training and confrontational techniques are more likely to lead to fear, avoidance, and increased aggression.
Most of the humane, passive, or positive techniques involved in behavior modification are not hard to learn and together with preventive strategies are very successful. In fact, dogs trained with rewards have fewer behavior problems, less fear, and less avoidance than dogs trained with punishment. The following is a short review of the basic principles involved in the techniques and their associated strategies.
The pairing of an unconditioned stimulus with a neutral stimulus results in a conditioned stimulus and a conditioned response. Classical conditioning can occur in both positive and negative ways. Examples of a positive conditioned emotional response are the pairing of a clicker with favored treats (for clicker training) or a doorbell signaling visitors (for pets enthused about meeting new people).
Problems arise when a fearful conditioned emotional response is established toward a previously neutral stimulus (visual, odor, auditory, animate, inanimate) by repeated pairing with a fear-producing stimulus. Once this occurs, the stimulus itself will elicit the fear response, eg, a doorbell paired with the arrival of unfamiliar people (for pets fearful of visitors), or a doorbell paired with verbal or physical discipline applied by the owner for barking or jumping up (pinning, leash corrections). Similarly when a pet lunges or barks when meeting new people on the street or at the front door, the use of positive punishment to inhibit the behavior (such as choke collars, prong collars, shock, pinning) may condition a new response in which unfamiliar people become a fear-evoking conditioned stimulus. A visit to the veterinary clinic that may begin as a neutral situation may quickly become fear evoking if it is associated with unpleasant outcomes or is further enhanced by owner anxiety. In addition, all of the stimuli associated with the event (sights, sounds, smells) also become conditioned stimuli for fear. In much the same way, rain, wind, darkening skies, and lightning can quickly become conditioned fear-evoking stimuli for pets fearful of thunder.
Counterconditioning involves the consistent and repeated pairing of a stimulus that evokes an unpleasant response with something that is emotionally positive until a positive association is made. To be successful, counterconditioning should be coupled with desensitization in which the stimulus is minimized or reduced to a level that does not evoke the fear response (eg, by reducing volume, increasing distance, changing the environment, or modifying the stimulus to something less threatening). Once a positive association is made, rewards can be paired with stimuli of gradually increasing intensity.
Desensitization and counterconditioning are extremely time consuming. The exercises must be constantly repeated, so that the response is altered to a positive one. All stimuli that evoke fear (sights, sounds, odors, tactile) must be considered. Clients often want both quicker fixes and less work. However, moving too quickly provokes anxiety and sabotages any behavior modification program.
Operant conditioning is a method based on making an association between a behavior and consequences of that behavior. The results either increase or decrease the likelihood of future responses. There are four types of behavior-consequence relations: positive and negative punishment and positive and negative reinforcement. Reinforcement increases the likelihood a behavior will be repeated, and punishment leads to a reduction in behavior. Negative refers to the removal of a stimulus, and positive refers to the application of a stimulus.
Positive reinforcement training occurs if behavior is increased by something applied (generally something pleasant or appealing); negative reinforcement occurs if behavior is increased by something removed (generally something unpleasant). In positive reinforcement training, a reward should be given immediately and consistently until the behavior is reliably repeated. If the behavior is to be trained on command or cue, a word or hand signal should then be added before the behavior-reward sequence. Once learned, behavior can be reinforced on a variable schedule, so that the period of time or number of responses before the reward is given is varied. Rewards are used for positive reinforcement, but a reward is not synonymous with positive reinforcement. A reward is anything desirable to the pet, from an activity such as petting, walking, or play, to an item such as a toy, food, chew, or treat. However, unless there is a clear relationship between the behavior and the reward (timing, consistency, contiguity), then the reward does not achieve the goal of positively reinforcing behavior.
Negative reinforcement must not be confused with punishment, because punishment decreases behaviors and reinforcement increases behaviors. One example of negative reinforcement is avoidance or escape behavior. For example, if an animal anticipates an unpleasant outcome (eg, meeting another dog, veterinary visit), then the aversive outcome will not occur if the animal retreats. Similarly, if the owner puts pressure on a head halter until the desired behavior is achieved (eg, sit, back up), the release of tension is negative reinforcement. One potential consequence of negative reinforcement is that if a pet’s threats or aggression lead to removal of a stimulus (eg, dog, delivery person, owner), the behavior is reinforced by the retreat of the stimulus.
Positive punishment occurs when a behavior decreases when something is applied (generally something unpleasant), and negative punishment occurs when a behavior is decreased when something is removed (generally something pleasant or appealing). In positive punishment, if behavior does not decrease after the first few applications, then the punishment is not being appropriately timed or the behavior is too strongly motivated to be deterred by punishment. Positive punishment applied by a person (owner, trainer) is intended to cause the pet to become fearful of repeating the behavior. However, a potential consequence is that the pet becomes fearful or defensive to the punisher or to an approaching hand. Relationships with people should always remain positive! Also, if an unpleasant consequence occurs only when the owner is present, the behavior may continue in the owner’s absence. Another problem with positive punishment is that punishment paired with exposure to a stimulus (barking at cars, meeting other dogs on walks) can result in a conditioned fear of the stimulus (see above).
Punishment cannot be used to achieve desirable behaviors, only to stop what is undesirable. If the goal is to make the pet fearful of repeating a behavior (eg, garbage raiding, taking things from counters, chewing plants) or to keep the pet away from an area (room, couch, bed), then environmental punishment or pet-activated punishment (eg, motion detector alarms or sprays, upside-down carpet runners, aversive tastes, double-sided tape, or bark-activated sprays) or remote punishment (eg, spraying water while out of sight, remote-activated alarm or spray) might be most appropriate. However, before focusing on how to stop what is undesirable, the owner should first focus on providing a desirable alternative (eg, where to sleep, where to climb, what to chew).
Negative punishment is the reduction of a behavior by the removal of something pleasant. For example, if the pet is receiving affection or play when an undesirable behavior begins (eg, play biting, mouthing, mounting), the immediate removal of the play or affection will "negatively" punish the pet. However, unless the pet can determine what behavior leads to the removal of play, the behavior may actually intensify because of frustration at not receiving its reward.
Signals that can be used at a distance to convey that the reward is coming are second-order reinforcers. Commonly used second-order reinforcers are words (eg, “Good dog!”), clickers, or whistles. By repeatedly and continuously pairing these with a primary reward such as a toy or treat, second-order reinforcers can elicit the same response that the reward would, as long as the pairing is repeatedly maintained. Clicker training requires frequent practice and excellent timing, but once achieved the animal can be reinforced each time the desired behavior is observed. Clicker training is an excellent way to immediately “mark” desirable responses, gradually shape new or more desirable behaviors (eg, longer, more relaxed), or associate a positive emotional response with the stimulus. (A useful resource is www.clickertraining.com.)
When a more desirable behavior is made contingent on a less desirable behavior, the less desirable behavior is more likely to be repeated. Thus, the more desirable behavior serves as the reinforcer. For example, if a pet wants to go out or cross the street for its walk, the owner can train a sit-stay before each of these behaviors. A horse or dog that wants to walk ahead can be taught that walking on a slack rein or leash will result in this behavior.
Overlearning is the repeated evocation and expression of an already learned response. It is a phenomenon frequently used in training for specific events but may be underused in preventing fearful responses in dogs. Overlearning accomplishes three things: it delays forgetting, it increases the resistance to extinction, and it increases the probability that the response will become a “knee-jerk” one, or response of first choice, when the circumstances are similar.
Shaping works through gradual approximations and allows the animal to be rewarded initially for any behavior that resembles the desired behavior. For instance, when teaching a puppy to sit, providing a food reward for a slight squat will increase the probability that squatting will be repeated. This squatting behavior is then rewarded only when it more closely resembles a sit, and finally, when it becomes a true sit. Shaping can also be used to reward an increase in duration of or progressively more relaxed behaviors.
The ending of a behavior once all reinforcement is removed is termed extinction. For example, if people pet a dog that jumps up on them for attention, the behavior continues; if they stop, the dog will eventually extinguish its response because the reward is no longer there. However, any form of intermittent reinforcement—even occasional petting of the dog in response to its jumping—will prolong the performance of the response. Valuable rewards, a long history of performance, and intermittent reinforcement all increase resistance to extinction. Owners also must be prepared for the intensity of the behavior to initially increase before it is extinguished. Giving in will make extinction even more difficult as the animal learns that higher intensity behaviors achieve the desired outcome.
Habituation is a gradual lessening of a response to a stimulus. Usually this occurs with repeated presentation of a stimulus whereby the animal learns that it does not signal anything important. For example, horses placed in a pasture bordering a road may at first run away when traffic passes but eventually learn to ignore it. Stimuli associated with potentially adverse consequences are more difficult to extinguish with habituation than other stimuli. In prey species, responses to sounds associated with predators would be difficult to habituate, because they have been selected for and generally are adaptive. If the fear response is too intense, instead of habituation the animal may become increasingly more fearful of the stimulus. This is termed sensitization.
If an extended interval has occurred since the time an animal last experienced a stimulus to which it had habituated, the animal may again react when reexposed to the stimulus. This is termed spontaneous recovery.
This 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. This procedure is seldom effective and has welfare implications in dogs, because it initially enhances fear and cannot be stopped until all physiologic and emotional signs of fear are gone. If done improperly, flooding may therefore increase problem behaviors. In practice, a controlled level of flooding is quite often used as a component of behavior modification, in which the stimulus is presented at a level that is low enough to cause mild fear and the pet is not removed until it habituates. This can then be combined with reinforcement, ie, the pet is positively reinforced or the stimulus removed (negative reinforcement) when the fear response subsides or abates.
This involves the replacement of an undesirable response with a desirable one. For example, high-value rewards can be used to train desirable target behaviors that are alternatives to the undesirable behavior. However, if the behavior is part of the pet’s natural repertoire (eg, greeting, barking), it can be particularly difficult to train alternative behaviors. Specific examples of response substitution include training a dog to sit or lie down as an alternative to jumping up, mounting, or play biting; or to sit, walk on loose leash, or back up for dogs forging ahead or running out the door. Training should begin in a variety of environments where success can be most readily achieved. The desired endpoint for the new response is for the animal to be quiet and calm. Therefore, the owner must learn to read the look in the eyes, body posture, facial expressions, and breathing to be able to gradually shape the desired behavior. Training could then move to environments with increasing distractions and locations where the problem is most likely to arise. Alternatively, the pet might be enticed to engage in a behavior that is incompatible with the undesirable behavior, eg, teaching the dog to fetch a toy when visitors arrive instead of jumping up.
To replace the undesirable behavior with one that is desirable, response substitution can be coupled with desensitization by beginning training with stimuli of low enough intensity while training the target behaviors (eg, relaxation) with high-value rewards. However, for pets that are fearful or anxious, the focus should be on desensitization and counterconditioning to change the pet’s emotional state rather than the behavioral response.