PAMELA J. REID
Emotional distress underlies many of the behavior disorders suffered by companion animals. While stress is a normal part of everyday life, stress becomes distress when the animal’s ability to cope compromises its well-being (Moberg, 2000). Distress-related behaviors can be challenging for pet owners, prompting them to seek professional help from certified behaviorists and veterinary behaviorists (Tiira et al., 2016; Cannas et al., 2018). They also rank high among reasons owners give for relinquishing their pets to shelters and rescues (Patronek et al., 1996; Segurson et al., 2005). The most common distress disorders reported by pet owners include fears and phobias, generalized or free-floating anxiety, separation anxiety, and defensive aggression.
Most problematic fears in companion animals are learned, although there is suggestion that fear of loud percussive noises may be innate in some dogs (Lindsay, 2005). Learned fears develop when an animal associates a previously innocuous stimulus or situation with the threat or the occurrence of an aversive event. The stimulus or situation comes to elicit fear and fear-related responses. This learning can take place in as few as one trial (Jarvik and Essman, 1960; Fanselow, 1990). As an example, it is not uncommon for a pet to learn to fear the sight of nail clippers after one painful experience of having a nail cut too short. A fear is labeled a phobia if the fearful response is far more intense than what is appropriate for the situation (Hothersall and Tuber, 1979) – for instance, a dog that repeatedly throws itself against confinement barriers in an effort to avoid contact with a person, causing itself injury.
Generalized anxiety disorder is characterized by distress that is not associated with any specific stimuli. Instead, the animal is in a chronic state of apprehension, agitation, heightened sensitivity, and hypervigilance (Lindsay, 2001; Ferdowsian et al., 2012), sharing many of the same symptoms as posttraumatic stress disorder in humans (Foa et al., 1992; see also Chapter 14, this volume). This type of disorder often develops from exposure to traumatic events but can also result from chronic experience with low level aversive stimuli.
While most dogs, and some cats, may experience some degree of discomfort when separated from their person (or people), an animal is considered to suffer from separation anxiety when it becomes so distressed that it engages in destruction, vocalization, elimination, and/or attempts to escape. These animals are often hyper-attached to their people as well, exhibiting anxiety even when separated by a door or gate and displaying effusive greeting behavior when reunited.
Defensive aggression develops when an animal learns that an effective means for fending off a frightening person or conspecific is to adopt the strategy of ‘the best defense is a good offense’. Most fearful animals will prefer to escape, but if escape is prevented, aggression is a viable defensive tactic. Unfortunately, defensive aggression often results in the threatening stimulus moving away and so the behavior is reinforced, with the result that this can become the go-to strategy for the frightened animal. Without knowing the history of the animal, this type of aggression can be misidentified as offensive.
Whether an animal is vulnerable to developing a particular distress disorder depends on factors such as temperament, early experience, prior exposure to stressors, and the interactions of these (Mineka and Oehlberg, 2008). There is solid evidence that there are genetic or epigenetic predispositions for fearfulness. In a classic demonstration, Murphree (1973) artificially selected for fearful behavior and produced a strain of ‘nervous’ pointer dogs that exhibited intense avoidance of humans, regardless of their socialization history. Deficits in socialization and environmental exposure during development can result in profoundly debilitating neophobia (Scott and Fuller, 1965). Dogs originating from substandard commercial breeding operations often suffer severe impoverishment in early life, resulting in a variety of maladaptive distress disorders that sometimes cause the dogs to be described as ‘autistic-like’ by their owners (McMillan et al., 2011, 2013; McMillan, 2017). Experiencing an acute stressor, such as a startling or traumatic event, can lead to rapid conditioned fear of any stimuli associated with the event (Haeson et al., 2017). However, repeated exposures to lower level aversive stimuli can also result in conditioned fear. For instance, Schilder and Van der Borg (2004) found that dogs trained with the use of low-intensity shock experienced distress during training sessions and came to associate the painful events with their owners, even outside of the training context. Finally, chronic experience with uncontrollable and/or unpredictable aversive events can, in some animals, lead to generalized anxiety disorder (Dess et al., 1983).
24.1 Components of Treatment
Nonpharmacological procedures designed to reduce emotional distress and distress-related behaviors in companion animals consist of two components: management and treatment. The purpose of management is to decrease the opportunities for the expression of negative emotion-laden behaviors. Effective management serves to: (i) reduce the likelihood that the animal will continue to experience distress; (ii) halt the strengthening of any associations and patterns of responding that may result from the animal’s repeated exposure to distressing events; (iii) reduce the likelihood that the animal’s display of emotional behavior will result in damage to people, other animals, or property; and (iv) empower clients with a strategy they can use to reduce the risks associated with their pet’s undesired behaviors. Clients may be more willing to embark on a treatment program if they can be provided with at least some degree of respite from the anguish of experiencing their pet’s undesired, and sometimes dangerous, behavior.
The purpose of the treatment component is to diminish the animal’s distress and the associated emotional behaviors, and replace them with positive emotions and their associated emotional behaviors. Separating problem behavior into its affective and behavioral components is helpful in conceptualizing how to assess and approach treatment. Parceling out emotional state from emotional behavior forms the basis for much of the following discussion of treatment options.
24.2 Treatment Options for Distress-Related Behaviors
24.2.1 Flooding or implosion therapy
Flooding, also known as implosion therapy, involves exposing an animal to the full-intensity distress-eliciting stimulus while at the same time preventing the animal from escaping the situation. The goal is for escape and avoidance behaviors to extinguish, thereby enabling the animal to learn that its fear is unfounded. Blocking the animal’s ability to escape or avoid the feared stimulus, called response prevention, is a necessary component of flooding, as otherwise responding is very resistant to extinction. As long as the animal’s escape/avoidance response is blocked, this procedure leads to rapid extinction of fear responding (Lindsay, 2000). However, because the animal is forced to experience the full-intensity stimulus, it is extremely frightened during the initial stages of treatment (Fig. 24.1). Similarly, human patients find implosion therapy most unpleasant (Thyer et al., 1988).
Despite the effectiveness of flooding in experimental set-ups and in human therapy, few reports exist of its efficacy for treating companion animals. One notable exception is the successful treatment of stranger aggression in three dogs by Williams and Borchelt (2003). They developed a form of flooding and response blocking with their use of a ‘grain box’ to prevent aggressive responses in dogs undergoing repeated presentations of the triggering stimulus. As a form of response prevention, at the American Society for the Prevention of Cruelty to Animals (ASPCA) Behavioral Rehabilitation Center (BRC), dogs are initially provided an airline crate in which to hide from feared stimuli, but early in treatment the top of the crate is removed to prevent the dog from continued hiding. However, this can hardly be likened to flooding because dogs at the BRC are never exposed to full-intensity stimuli during treatment.
There are two reasons why this author does not recommend flooding/response blocking. First, flooding can be a tricky procedure to put into practice. In addition to needing a way to prevent the animal from escaping (or, in the case of defensive aggression, from attacking the stimulus), it is critical that sessions continue until the animal not only ceases attempting to escape but also relaxes in the situation. This can take an unreasonably lengthy amount of time. If the animal is still fearful at the end of the session, there is unlikely to be any beneficial effect and, in fact, the animal’s fear may intensify (Metja et al., 1974). Second, flooding can be severely traumatizing for both the animal and the client. Clients do not want to see their pets in distress, even if they are advised that the end will justify the means. And there is the very real risk that the procedure can backfire and the fear will worsen. This author feels that there are equally effective and more humane methods from which to choose.
24.2.2 Desensitization and counterconditioning
According to Wolpe and Lazarus (1966), desensitization and counterconditioning (DSCC) involves the ‘breaking down of neurotic anxiety-response habits, employing a physiological state incompatible with anxiety to inhibit the anxiety response to the stimulus that evokes it weakly, repeating the exposure until the stimulus loses completely its anxiety-provoking ability’. To say that is a complicated description would be an understatement! First, DSCC works to change an animal’s response to a stimulus by repeatedly presenting the stimulus at such a low level that the animal’s arousal is kept to a minimum, thereby setting the animal up to habituate (or ‘desensitize’) to the stimulus. At the same time, however, the stimulus is paired with the presentation of another stimulus that elicits responses motivationally and/or physically incompatible with the distress response. The distress response originally exhibited by the animal to the stimulus is ‘countered’ by this new and very different association.
The first report of the clinical use of DSCC, also known as graduated exposure therapy (Antony and Barlow, 1997), described its use in eliminating fear in children (Jones, 1924). Children temporarily housed in an institution were assessed for their reactions to a variety of stimuli, including a snake, a rat, a rabbit, a frog, loud noises, and scary faces. Jones (1924) subjected children, who responded with extreme distress to one or more of the stimuli, to procedures designed to eradicate their fear. The most successful intervention consisted of the therapist bringing the child to the cafeteria, then placing the feared stimulus sufficiently far away that it did not interfere with the child’s desire to eat, and feeding the child a favorite food, such as ice cream. While the child ate, the stimulus was slowly brought nearer the table, then placed on the table and, finally, brought close enough for the child to touch. In one instance, the child ate ice cream while a rabbit, which initially had terrified the child, sat in the child’s lap. Jones also found that increasing the children’s hunger enhanced the effectiveness of the procedure (Dickinson and Pearce, 1977 also confirm that, at least when it comes to food, deprivation leads to more effective counterconditioning).
Wolpe (1958) documented the eradication of conditioned fear in cats using DSCC. He established ‘experimental neurosis’ in hungry cats by associating the act of feeding with the delivery of electric shock. After the initial conditioning, the cats refused to eat in the experimental room, despite 1 to 2 days of food deprivation. This inhibition of feeding generalized to rooms that were similar in appearance to the experimental room. Wolpe identified a room that was sufficiently unlike the experimental room in which the cats were able to eat despite still being visibly anxious. Successive feedings in the new room eliminated all signs of distress. The cats were then moved to a room slightly closer in appearance to the experimental room and offered food. The same routine was repeated in a series of rooms of increasing similarity to the experimental room, remaining in each room until the cats no longer showed signs of distress before progressing to the next. With some cats, Wolpe paired an auditory stimulus with the delivery of shock and found that the same procedure successfully eliminated the cats’ fear of the sound. The cat was initially moved far enough away from the source of the sound that it was able to eat. Much the same as in Jones’s study with children, the cat was gradually moved closer to the sound until the sound no longer elicited fear nor inhibited feeding.
A classic example of DSCC in clinical animal behavior can be provided through the example of a dog that fears the sound of thunder (Fig. 24.2). First there must be a way of presenting the noise at a volume so low that it fails to evoke a distress response. For instance, some dogs will react to an audio recording of thunder. While the dog hears the recording at a very quiet level, it is fed especially tasty food. The dog comes to learn a new association between thunder and food. Anticipation of the food elicits responses that are motivationally incompatible with fear and these responses eventually come to replace the distress responses. In other words, the dog comes to expect food when it hears thunder at a low volume. Gradually, the volume of the recording is increased while the dog’s arousal is maintained at a low level. Distress responses continue to decline while appetitive behaviors predominate. Eventually, the dog can tolerate the sound of thunder at realistic volumes without becoming afraid. Like Pavlov’s dogs, a dog successfully undergoing DSCC for thunder phobia presumably salivates upon hearing thunder.
A cat that becomes distressed when a new kitten is introduced into the home can be treated in much the same way as Jones treated the children. For example, imagine a cat that flattens its ears and hisses at the sight of the kitten. If the kitten ventures too close, the cat will spit and swat at the kitten before running away. To employ DSCC in this case, the cat is positioned, on a harness and leash or in a wire crate, at one end of a room or hallway, while the kitten is held far enough away that the cat is willing to eat its favorite treats. The cat is fed continuously while the kitten is in view. The session can consist of one lengthy period of exposure, as Jones did, or of a series of discrete exposures (trials) in which the kitten is periodically removed from sight and the treats are discontinued. This further draws the cat’s attention to the positive association between the kitten and the treats, such that the kitten becomes a discriminative stimulus for food. As with any type of conditioning, the time between stimulus presentations (the intertrial interval) should be longer than the duration of exposure to the stimulus itself (Mackintosh, 1974). For instance, if the kitten is in view for 10 seconds, the time the kitten is removed from the cat’s view should be noticeably longer than 10 seconds.
In practical applications, the components of desensitization (DS) and counterconditioning (CC) can be employed together or separately. Analyses of CC applications further reveal that there are two similar, but theoretically and procedurally distinct, methods in use. Next, I describe how DS and CC procedures each work separately and in concert and outline their benefits and limitations.
The use of DS alone relies on the processes of habituation to the feared stimulus and/or extinction of an association between the feared stimulus and a second frightening stimulus that historically were linked. In a typical study of habituation, a stimulus is presented repeatedly. Over presentations, there is a relatively permanent decline in responding to the stimulus, presumably reflecting a general reduction in the animal’s attention to the stimulus. In a typical extinction study, the animal first undergoes a series of conditioning trials to establish an association between a stimulus and an outcome, and then that link is subsequently abolished. The underlying learning process is the same as flooding, in that fear reduction occurs as a result of the response extinguishing first and then the underlying emotion following suit.
In a clinical application of DS alone, the animal experiences repeated presentations of the feared stimulus, but at such a low intensity that it does not elicit distress responses. No attempt is made to countercondition by pairing the stimulus with a reinforcer such as food. For instance, a dog that fears the sound of thunder could be continuously exposed to a recording of thunder played at a very low volume. The dog would likely initially alert to the sounds but eventually would learn to ignore it. Gradually, probably over the course of many days, the volume may be increased, with the goal that the dog ceases alerting to the sound or showing any signs of distress.
Aside from the fact that it is logistically simpler to implement, there are few, if any, situations where it makes sense to use DS alone to treat distress disorders. The addition of CC invariably makes behavior change more probable.
CC involves an explicit attempt to counter the animal’s distress by associating the feared stimulus with a second stimulus that elicits an incompatible motivational state and/or physical response. When CC is implemented alone, the feared stimulus is presented at full intensity. This is comparable to flooding, but with the addition of reinforcement to ‘counter’ the original association. Sometimes graded exposure (desensitization) is simply not possible, as there is no way to lower the intensity of the stressor (e.g., fear of flying in an airplane).
For an example of CC alone, consider a dog that has learned to fear children because of a history of punishment whenever children were present. If a child is nearby, the dog attempts to hide. The dog could be counterconditioned by associating the presence of a child with a game of fetch. The motivational state elicited by ball play is incompatible with the distress elicited by the child. If the dog’s desire to play ball is stronger than the motivation to hide, anticipation of play will come to override the original anticipation of punishment. If it is not, the fear association will remain intact. In classic opponent-process style, the desire to play and the motivation to hide are assumed to exert antagonistic influences on each other until arousal in one system inhibits arousal in the opposing system and interferes with its motivating, reinforcing, and response-producing capabilities (Stein, 1964). Although it should be acknowledged that there are problems with this conceptualization (Capaldi et al., 1983), delving into the exact nature of the mechanisms underlying CC is unnecessary for this discussion.
The scientific literature on CC is confusing because the term is used to describe two theoretically and procedurally distinct CC methods: classical and operant. The following describes each individually.
CLASSICAL COUNTERCONDITIONING. In the examples presented thus far, the feared stimulus (the conditioned stimulus [CS]) is linked with a pleasant unconditioned stimulus (UCS), such as food, in an attempt to replace distress responses with conditioned appetitive responses (CRs or conditioned emotional responses [CERs]). This is more accurately termed classical counterconditioning (CCC) because the two stimuli are presented contiguously, with no explicit conditioning of a target behavior. In other words, the UCS (food) is presented in conjunction with the CS (the feared stimulus), noncontingently, regardless of the animal’s behavior. The objective is, of course, to change behavior but this is expected to occur as a result of a change in the animal’s emotional or motivational state. For instance, in the example presented in the previous section, the dog is exposed to a child and then the dog is invited to play. The play is offered regardless of the dog’s behavior toward the child; the dog might initially be trembling, panting, pacing, growling, barking, or any number of behaviors designed to increase distance between itself and the child. Provided play is a sufficiently powerful motivator, the dog will come to associate play, rather than punishment, with the presence of children, and so the responses elicited by the child will come to reflect this new association – responses that are elements of the play behavior system, such as tail wagging, play bowing, etc. In a nutshell, CCC focuses on altering the emotional state of the animal with the assumption that the form of the behaviors elicited by the emotions, the CERs, will also emerge.
CCC is an extremely powerful agent for behavior change. Laboratory studies of CCC are often referred to as cross-motivational transfer experiments. In most demonstrations, an initially neutral stimulus is first paired with one UCS such as shock, and then, at a later time, the stimulus is paired with a motivationally disparate UCS such as food. In other demonstrations, though, two biologically relevant but incompatible UCSs are paired together. For instance, Erofeeva (1921) used a strong electric shock to signal the delivery of food to hungry dogs. Erofeeva reported that the dogs initially responded with defensive behaviors such as struggling and yelping. Yet, as conditioning progressed, the dogs began to show typical appetitive responses toward the shock, including lip licking and salivating. Even more surprising, Dearing and Dickinson (1979) found that after counterconditioning an aversive stimulus to signal an appetitive one, the aversive stimulus was incapable of functioning as a punishing stimulus in an instrumental paradigm. This supports the interpretation that CCC can actually produce a change in the motivational and reinforcing properties of a UCS.
OPERANT COUNTERCONDITIONING. The second CC approach is to explicitly condition a volitional behavior that is physically incompatible with the undesirable distress behavior, with the goal that the new behavior increases in frequency while the distress behavior fades away. Tarpy (1982) defines CC as a form of training in which ‘a new behavior, counter to the original response, is reinforced, while at the same time the original response is not rewarded’. Defined as such, this CC is the result of instrumental contingencies and should correctly be referred to as operant counterconditioning (OCC). The focus is on replacing behaviors rather than changing emotions.
OCC, although identified simply as counterconditioning, is the method often described in the applied animal behavior literature. For instance, Overall (1997) writes that ‘in counterconditioning …. The dog is taught to engage in a behavior that competitively inhibits the performance of the undesirable behavior.’ Overall provides an example of OCC in the treatment for submissive urination: ‘as soon as the dog’s rump touches the ground without any leakage, the treat is released. Clearly, if the dog rolls, grovels, or leaks, it does not get the food.’ In this example, delivery of the UCS (food) is contingent on the dog displaying or inhibiting a specific response. Contrast that with a CCC approach to submissive urination. Suppose the CS (the stressor) is identified as direct eye contact from a person. The dog experiences direct eye contact followed immediately by the delivery of a treat, regardless of whether the dog urinates or not. With sufficient pairings, the objective is for the dog to associate eye contact with treats and, if the dog finds the treats pleasurable enough, then urinating should be replaced with solicitous behaviors such as approaching, lip licking, and tail wagging.
DIFFERENTIATING OPERANT AND CLASSICAL COUNTERCONDITIONING. Why is the distinction between CCC and OCC important? To start, a significant difference exists in how these procedures play out in an applied setting. Take, for instance, the example of a dog that experiences fear at the sound of the vacuum cleaner. This negative emotion prompts the dog to flee under the bed, where it pants and trembles. In a CCC paradigm, the behaviorist’s task is to identify a UCS, such as food or play, that is emotionally incompatible with the emotion of fear, and pair the sound of the vacuum with the new UCS. This pairing can be accomplished every time: present the vacuum, present the food. If the food is sufficiently appealing, the expectation of food will come to replace the fear associated with the sound of the vacuum. And as a result, approach behavior will take the place of avoidance behavior. Contrast this with an OCC paradigm. Imagine that the dog is required to sit–stay before receiving the UCS. The dog hears the vacuum and is then cued to sit and stay. If it does, it gets a treat. If it does not, it does not get a treat (no UCS).
With CCC, a correlation of 1.0 exists between CS presentation (vacuum) and UCS presentation (food). With OCC, the correlation between CS and UCS may be <1.0, because it is up to the animal whether or not to perform the requisite behavior. If the animal does not perform the behavior, the UCS is not presented in conjunction with the feared stimulus. Conditioning is shown to be stronger when there is a greater contingency between CS and UCS (Mackintosh, 1974).
The second primary distinction between CCC and OCC is the matter of what changes first, emotions or behavior? In CCC, the goal is to change emotion, with the assumption that a change in behavior will follow. In OCC, the goal is to change behavior, with the assumption that a change in the underlying emotional state is inevitable. Because emotion ‘drives’ most serious distress disorders such as aggression, fear, and separation anxiety, it is reasonable to conclude that conditioning procedures which focus on changing the quality and intensity of the animal’s emotional state first will ultimately lead to conditioned emotional responses that are reflective of the altered emotion. A convincing illustration is the guard dog that barks aggressively and lunges at the fence when the next-door neighbor enters the adjacent yard. If the neighbor tosses a delicious steak over the fence every time she comes out, it will not be long before the guard dog comes running happily toward the fence the instant he detects her door opening.
Instrumental contingencies may change the animal’s behavior but they are not guaranteed to produce a shift in the underlying affective state. In other words, while the animal’s behavior may be altered, the feelings of distress remain, and the original problematic behavior is likely to reoccur. Barlia (1988) provides an illustration of this in his report of a dog that behaved aggressively toward unfamiliar people. The client trained the dog to adopt specific postures (sit, down) with such high reliability that the dog would remain in position even when approached directly by a stranger. However, Barlia notes that the dog remained stiff and tense during interactions, despite the extensive training and exposure to people. Although the dog’s response toward the feared stimulus had been transformed through OCC, the underlying fear remained intact. Similarly, it is not uncommon for human subjects to report that, after therapy, although they are able to function more effectively in the presence of a feared stimulus, such as sitting through an airplane flight, they still experience extreme anxiety and fear (Hersen, 1973).
VOLUNTARY AND INVOLUNTARY BEHAVIOR. Another factor that comes into play is the nature of the behavior or behaviors linked to the animal’s distress. Specific behaviors can be thought of as ranging on a continuum of voluntary–involuntary, although, in reality, things may not be as clear cut as this (De Havas et al., 2016) (Fig. 24.3). Certain behaviors, such as sit and down, are under good volitional control and amenable to operant conditioning, whereas behaviors such as trembling or freezing are closely linked to underlying emotional states and are much more difficult, if not impossible, for an animal to control. Involuntary responses are easily classically conditioned but far less sensitive to instrumental contingencies (Skinner, 1957; Thompson, 1958). Other behaviors are more likely to fall somewhere in the middle. For instance, a dog’s bark can be highly emotive, yet still the dog may exercise some control over whether to bark. Although it is recognized that emotionally charged behaviors are sometimes placed under operant control (Salzinger and Waller, 1962), the typography of the behavior is altered and the response becomes emancipated from the emotion. An example of this is the dog that has been taught to bark on cue. The sound of the bark is often noticeably different from the bark produced by the dog, for instance, in response to an intruder. Likewise, a dog taught to growl for the reward of a tidbit can do so, but the growl sounds quite different from a growl the same dog produces when it is experiencing the associated affect.