In several significant ways, mood and anxiety disorders stand apart from the other psychiatric conditions we discuss in this book. Perhaps the most obvious is that their base rate occurrence in the population is substantially higher. For example, in Western countries, about one in five people experience clinically significant symptoms of anxiety or depression at some point during their lives, whereas most other disorders have much lower prevalence rates – of the order of one to three per cent. The symptoms that define these disorders are also much more continuous with the personality traits that give rise to them than we can claim for other Axis I conditions. Generalized Anxiety Disorder represents a particularly intuitive example of the continuity from normal personality to pathology since it describes, in large part, the extreme end of a dimension of trait anxiety, similar to the way that hypertension, as a medical condition, is simply the pathological elevation of normal blood pressure.
As we explained in Chapter 2, it is our intention, when framing the link between personality and disorder, to follow the diagnostic classification system imposed by the DSM. However, in the case of mood and anxiety disorders, we decided that it was more appropriate to integrate the two conditions in a single chapter. While several factors influenced our decision, perhaps the most compelling is their substantial symptom overlap. In the past, the term ‘neurosis’ was often used to reflect the great similarity between these disorders, but this terms has rather fallen out of favour and been replaced by ‘distress disorder’ to cover both conditions (Clark et al., 1994). There are also stable vulnerability factors that are common to both types of disorder. For example, in a large-scale study of female twin pairs, Kendler and colleagues (1992) found that the same genetic factors appear to influence both Major Depressive Disorder and Generalized Anxiety Disorder, suggesting that environmental experiences largely determine whether ‘at risk’ individuals develop anxiousness or other depressive symptoms. In addition, although anxious and depressive ‘cognitions’ have a certain unique content, they share many more similarities, such as an habitual attention to negativity and threat. The common component between the two seems to be a state of general distress with the factors that differentiate them developing from that general base (Beck & Perkins, 2001; Hertel, 2002).
Also, highlighting the strong interrelationship between the two types of disorder is that some have proposed a new diagnostic category for the DSM, which they call Mixed Anxiety-Depression (MAD), because many patients with severe psychological impairment do not meet the full diagnostic criteria for either the anxiety disorders or depression; presenting instead with subclinical levels of both anxious and depressive symptoms (Barlow & Campbell, 2000). Even from a developmental perspective, we often find that parents have difficulty differentiating between symptoms of anxiety and depression in their children.
Another important factor that prompted our decision to discuss anxiety and depression in the same chapter is that stress is seen as a common factor in the aetiology of both disorders – as well as a consequence of them – suggesting that the three phenomena are strongly interrelated. Furthermore, as we shall see in the next section, the animal paradigms that have been developed to model anxiety and depression have considerable overlap and are, in many cases, based on exposure to stress.
Blurring of the diagnostic categories is especially clear in the case of Major Depressive Disorder where the heterogeneity of this condition has plagued clinical research for decades. In fact, the idea that ‘depression’ may actually comprise a cluster of loosely connected, discrete disorders has existed in some form since the early twentieth century (see Just et al., 2001). One way this can be seen is by the diversity of response to psychological and pharmacological therapies in patients diagnosed with depression (Drevets, 1998). Furthermore, the different types of depression seem to have quite distinguishable and separate underlying neurophysiological mechanisms, as we shall see later in this chapter.
Although there are some common and recognizable features of all depressions, such as negativity, pessimism, and rumination, it is also clear that at least two, quite distinct, forms of the disorder exist. A pronounced deficit in the ability to find positive incentives pleasing and reinforcing – with associated symptoms of apathy, inactivity, and excessive sleeping – describes the severely depressed mood of some patients. Other depressed individuals display a more agitated form of the disorder, as seen by their difficulty in recovering from emotionally stressful events, and their obvious signs of anxiety and restlessness (Davidson et al., 2002). Indeed, according to the DSM, as many as 80 to 90 per cent of depressed patients have symptoms associated with the anxiety disorders, such as poor concentration, sleep disturbances, loss of energy, irritability, excessive health concern, and panic attacks. Conversely, a frank ‘depressed mood’ is an associated feature of all anxiety disorders (see Table 5.1 for a brief list and description of the disorders that are subsumed under the general heading of Anxiety Disorders).
Table 5.1 DSM-IV anxiety-related disorders.
| Panic disorder |
| The reoccurrence of panic attacks (that is, a constellation of symptoms, including increased heart rate, sweating, trembling, chest pain, choking sensation, feeling of losing control, and fear of dying) with anticipatory anxiety associated with the attack. |
| Agoraphobia |
| Its literal definition suggests a fear of 'open spaces', though it is not a strictly accurate one. Individuals with this condition fear any place where they develop panicky feelings, such as crowded supermarkets, elevators, or standing in line. Some become 'housebound' as a result of this anxiety. |
| Specific phobia |
| The excessive fear of some object (like snakes) or a situation (like being near the water) to the point that it is strenuously avoided, or endured with intense anxiety. |
| Social phobia |
| The excessive fear of social situations, especially things like talking or eating in public, that usually involves the intense anxiety about feeling humiliated and blushing. |
| Obsessive-compulsive disorder |
| See Chapter 6 for a detailed description. |
| Post traumatic stress disorder |
| After a traumatic situation has occurred (such as a car accident) there is a re-experience of the event in the form of nightmares or flashbacks which results in extensive avoidant behaviour (in this case, driving in cars) or a numbing of general responsiveness which interferes with social and occupation activities. |
| Generalized anxiety disorder |
| Chronic and excessive worry and anxiety associated with a variety of somatic symptoms, such as restlessness, impaired concentration, irritability, muscle tension, and insomnia. |
As with many of the disorders we discuss in this book, animal models have provided us with some valuable insights, especially concerning the neurobiology and pathophysiology of the condition. Furthermore, because of the practical difficulties associated with the prospective study of mental illnesses (for example, their relatively low base rate in the population), animals can serve as useful proxies in research of this kind. However, certain features of a good animal model are necessary if this purpose is to be served adequately. For example, the paradigm should model a core symptom of the disorder and respond to the treatment drug for the condition it is alleged to mimic (for example, antidepressants for an animal model of depression and anxiolytic agents for models of anxiety). It should also use inducing conditions that are ecologically valid (Willner, 1997). In this regard, some animal models have been rather better than others in fulfilling these important validating criteria.
In certain senses, animal models of anxiety are more straightforward, and their validity easier to achieve, than those for depression. For instance, the defensive or fearful behaviours that we can induce in animals are closely analogous to anxiety-related behaviours in humans. Both can be seen as the excessive activation of our innate defence mechanisms arising from an exaggerated assessment of, or reaction to, danger. Second, since both share a common biological substrate, the animal response clearly has significance for human anxiety disorders (Palanza, 2001).
Probably the most frequently used rodent model of anxiety has been the ‘open-field’ test referred to in Chapter 3. This is a paradigm based on the assumption that novel environments elicit defensive reactions in rats and mice who tend, in their natural habitats, to avoid brightly lit and unfamiliar places. Other models of anxiety have assessed the animal’s reactions to painful physical stimuli, such as a foot shock or an aversive noise. In most cases, researchers have measured the anxiety response of an animal in models such as these, either by drawing inferences from its behaviour – for example, increased locomotion in the open-field apparatus reflects higher levels of anxiety – and/or by assessing an animal’s physiological response to a stressful event from biological markers, such as corticosterone (a stress hormone).
In recent years, however, there has been a significant paradigm shift in the area of animal models for mental health research since the emergence of technology to genetically engineer or modify the DNA of mice.
Transgenic mice
One of the most remarkable technical advances in biomedical research in recent years has been the development of transgenic mice. It has now become possible to genetically alter the mouse genome with nucleotide precision. These mouse models can be made by altering the animal’s DNA in two specific ways. In one process, new DNA is injected into the zygote pronucleus of a fertilized egg so that the foreign DNA integrates into the mouse genome producing an animal which will ‘over-express’ a certain gene product. In another process a targeted gene is mutated producing a functional ‘knock-out’ of the gene. These models are especially effective for understanding how individual genes and environmental factors interact to affect human physical and mental health. To use an example from cancer biology, one gene being studied is a ‘tumour-suppressor’ gene whose functions seems to be the maintenance of genomic stability. Mice who have had this gene mutated are much more susceptible to lung cancer after treatment with a chemical carcinogen than animals who contain this gene.
In particular, a number of ‘knock-out’ mice models have been developed to study symptoms of anxiety and depression. For example, the 5-HT1AR knock-out strain (that is, mice lacking the gene that controls the expression of the serotonin 1A receptor) show increased anxiety-like behaviours in a variety of conflict tasks. This model has also been particularly useful in highlighting the role of environmental and early postnatal factors in promoting the proper expression of the 5-HT1A receptor gene and how it influences normal emotional behaviour later in life (Gross et al., 2002).
Other transgenic mice have been developed to target aspects of the hypothalamic–pituitary–adrenal (HPA) axis since abnormal stress hormone regulation is believed to be one factor in the increased risk for depression (Holsboer, 2000). Although transgenic models are useful for the study of endogenous mechanisms underlying abnormal behaviour at the molecular level, so far – and this is unlikely to change – the deletion of a single gene has not provided any useful and comprehensive animal template for a particular psychological disorder. However, this is hardly surprising since complex human behaviours are far more likely to be caused by a concatenation of minor changes in so-called susceptibility genes, and even more likely by the interaction of these vulnerability genes with various environmental factors (Muller & Keck, 2002).
Animal models of depression have been somewhat more difficult to study because the primary symptom of this disorder is a change in mood. Researchers have therefore been forced to rely on observable responses of the animal that seem to parallel the behavioural parameters, such as withdrawal, problems in social functioning, and poor coping ability, that are seen in human depression. The ‘stress hypothesis’ of depression has led to a series of now classic putative models of depression which all have in common that the observed behaviour of the animal is triggered or caused by uncontrollable and aversive stimuli (Palanza, 2001). The most common of these are the learned helplessness and chronic mild stress models – the former being the most widely studied of all the animal models of depression.
In the mid 1960s, Seligman and colleagues discovered an interesting phenomenon while they were carrying out a Pavlovian classical conditioning experiment with dogs.
The term ‘Pavlovian Conditioning’ derives from a series of famous experiments begun in 1889 by the Russian physiologist Ivan Pavlov, in which he demonstrated the potential for conditioned and unconditioned reflexes in dogs – that is, that a contingency could exist between environmental events and physiological responses. In these experiments, Pavlov rang a bell as he fed meat powder to his dogs. Each time the dogs heard the bell they knew that a meal was coming, and as a consequence, they began to salivate in anticipation. After a number of pairings of the bell and the food, Pavlov rang the bell without bringing food, and noticed that the dogs still salivated. They had been ‘conditioned’ to salivate at the sound of a bell. Pavlov believed that humans react to paired stimuli in much the same way.
From these original experiments the terminology of classical conditioning was born and may be directly related to the dogs, the salivation and the bell. The ‘Unconditioned Stimulus’ (UCS) is the event that triggers an automatic action which has been called the ‘Unconditioned Response’ (UCR) – the meat powder and the salivation respectively. Upon repeated pairing of a neutral stimulus – the bell – and the UCS, the former, called the ‘Conditioned Stimulus’ (CS), is reliably able to elicit the same response as the latter – the salivation – and has been called the ‘Conditioned Response’ (CR).
However, in this case, Seligman and colleagues paired the bell with an electrical shock instead of food while restraining the animal in a harness. After the dog was ‘conditioned’ to the bell by the inescapable shock, the experimenters put it into a box, which was divided into two compartments and had a low wall between the two over which the dog could jump to escape from either compartment. The experimenters fully expected that when they rang the bell (the CS), the dog would jump over the wall to escape, in the way that non-conditioned dogs did when they were shocked. Instead, to their great surprise, the animal didn’t move! They concluded that the dog had ‘learned’ to be helpless. In other words, it had learned that attempts to escape were fruitless, and so it simply did not try. Employing a variety of learning tasks, this helplessness model was subsequently used to demonstrate that uncontrollable stress can produce significant performance deficits not found in normal animals.
Initially, the learned-helplessness model was adopted enthusiastically because it seemed to parallel the dynamics of human depression. However, its popularity soon began to wane when the experimental effects could not be reproduced in other animal species. People also began to question whether learned helplessness was really a behavioural process characterizing most depressed individuals (Palanza, 2001). For instance, many people fail to become depressed even when they experience traumatic life events, making it evident that the predisposition for depression is highly variable. In recent years, however, the learned helplessness paradigm has re-emerged as a good model of vulnerability. What was once its weakness – its unreliability – is now seen as one of its strengths in depression research. Because animals develop the helplessness syndrome to varying degrees, or not at all, researchers have used this natural variability in subsequent neurobiological studies to examine inherent risk factors for depression.
The chronic mild stress (CMS) paradigm is another popular animal model of depression whose validity is based on the rationale that it simulates anhedonia – that is, the loss of responsiveness to pleasurable or rewarding events – which is a core symptom of the melancholic form of depression. In a typical CMS experiment, animals are exposed sequentially to a variety of mild stressors, such as tail pinching, overnight illumination of their cage, periods of food deprivation, tilting of the cage, and change of cage mate. Over time, there is typically a decrease in the extent to which the animal will engage in previously rewarding behaviours, such as the consumption of, and/or preference for, palatable food and drink. Not only can these behavioural deficits be maintained for several months, but the animal’s normal behaviour can also be restored quite reliably by treatment with a variety of commonly used antidepressant drugs. Another parallel with human depression is that antidepressant drugs only alter the behaviour of animals exposed to CMS, not that of non-stressed animals, similar to the way that antidepressants have no mood-enhancing effects on non-depressed individuals. The CMS model is especially appealing because, in addition to its anhedonic effects, it also causes the appearance of other symptoms of depression, such as diminished sexual behaviour, decreased locomotion, and increased aggression, without triggering symptoms of anxiety (Willner, 1997).
Despite their enormous utility in many aspects of anxiety and depression research, it would be misleading to suggest that animal models are without problems. In depression, for example, studies have shown that when animals are tested in more than one paradigm, there is often low convergent validity across the methods. This suggests that different models may be mediated by separate biological processes, thereby tapping into different aspects of the condition. There is also the questionable ecological validity of many animal models. For instance, animal studies have mostly relied on physical stressors, such as electric shock, tail pinching, or restraint, to produce depressive or anxious responses, even though we know that the main sources of human stress are social in nature. Therefore, a more appropriate animal analogue of the human stress reaction would be an experimentally induced social stressor, such as overcrowding. Another highly problematic issue in animal research is response differences across species and between genders. One poignant illustration is that although prevalence rates of depression are consistently higher among women than men, in many rodent models of depression, the females are actually less susceptible to the syndrome induction than the males. We also see a pronounced discrepancy in the workable methods of inducing social stress in males and females – and regrettably, the majority of animal studies have only used male animals. For example, when female animals are housed alone, they tend to produce higher levels of corticosterone and display behavioural changes such as withdrawal and reduced exploration, which resemble symptoms of human depressive disorders. Male animals, on the other hand, do not show these changes when housed alone; instead they produce higher corticosterone levels when they are caged with other animals. In other words, crowding induces social stress in male animals, whereas solitude has that effect on females. These findings point to the ecological importance of developing models of social stress that take account of gender- and species-specific differences in the behavioural strategies each uses to cope with its social environment.
Given the considerable diagnostic overlap across the various anxiety and mood disorders, as well as the changes in how they have been defined in the various versions of the DSM, it seems more appropriate to discuss the primary symptoms that cut across these disorders rather than trying to fathom the unique aetiology – if there is one – of each disorder separately.
Like all the psychiatric conditions we address in this book, clues to their biology come from a variety of sources, including responses to drug treatment, markers of neurotransmitter status or activation, pharmacological and behavioural challenges, and neuroimaging procedures. Therefore, the greatest difficulty in trying to present a digestible account of their physical origins is the sheer number of neurochemical, neuroendocrine, neurophysiologic and neuroanatomical factors that have been investigated, and for which some evidence seems to exist. For that reason, we have had to be selective in our descriptions, and extend a caveat to our readers that we are only highlighting and touching upon themes for which there is the most consensus. Even when deciding which symptoms to address, we have had to be selective, eventually deciding only to include fearfulness, anxiousness and melancholy. However, by limiting our discussion to these three dimensions, we are aware that other common symptoms of mood and anxiety disorders, such as irritability, poor concentration, and abnormal fatigue, may have their own biological substrates.
Before proceeding, it is also important to touch briefly on our rationale for treating fearfulness and anxiousness as separate constructs. Historically, and even in some current writing, the terms fear and anxiety are more or less interchangeable and have frequently been conceptualized as a single dimension of personality. More recently, however, the evidence favours the view that they are distinctly different emotional systems. White and Depue (1999) have provided an excellent summary of the evidence that fearful responses are generated by an emotional system that is sensitive to both the unconditioned and conditioned stimuli of physical punishment or harm. Initially, these responses tend to inhibit behaviour, but if necessary will result in active escape behaviours. Anxiousness, on the other hand, is a state of high emotional arousal fostered by situations of uncertainty and social comparison, the possibility of the negative evaluation of one’s worth, and the threat of personal failure; in other words, it is fundamentally tied to social interactions. The responses to anxiety-provoking stimuli can range from worry and feelings of agitated distress to catastrophic ruminations.
However, some of the strongest evidence of the distinctness of the two emotional systems comes from psychometric studies. As we have seen in Chapter 3, all major trait theories, and every factor-analytic approach to personality, identifies a superfactor called variously neuroticism, negative emotionality, or negative affectivity, which taps into the anxiety domain. The correlations between these scales and specific fear scales are typically very close to zero. Other research has similarly found no correlation between neuroticism and questionnaire measures of constraint (which is viewed as a marker of inhibited behaviour).
Lastly, in an elegant series of experiments, White and Depue (1999) used pupil diameter as a non-invasive measure to test the association between fear and anxiety – a highly suitable paradigm since the pupil of the eye is rapidly enlarged during emotional activation (in order to maximize visual sensory input) and since two different processes affect dilation of the pupil. One is responsive to acute stimuli and is believed to relate to the response to fearful stimuli, whereas the second reflects a more tonic and stable function associated with individual differences in anxious reactivity. Their data led to the strong conclusion that the dissociation between the neural systems of fear and anxiety is, in fact, much greater than has traditionally been appreciated.
Fear is one of the most basic and normal of all human and animal reactions. Indeed, because of its fundamental role in promoting safety and security, it is fair to say that organisms would not survive for long without the ability to experience fear. However, in addition to our inherent capacity for feeling fearful – and our ability to respond appropriately – we are also able to recognize the environmental events that warn or predict of forthcoming threats. Psychologists have called this type of learning, fear conditioning because it seems to occur in the manner of classical Pavlovian associative conditioning – through the temporal pairing of relatively neutral stimuli with those that are aversive or dangerous. The study of fear conditioning has attracted great interest in the past decade because of its apparent role in the development of various fear-related disorders. For example, the ‘fear network’ in the brain is almost certainly involved in the aetiology of phobias, and in Panic Disorder.
From an evolutionary perspective, it is clearly adaptive for us to learn the appropriate associations between environmentally relevant stimuli and the threat of danger. However, many of the things that people now learn to fear are not necessarily biologically relevant to our survival, but are learned through experience with the stresses and strains of life – like the fear of flying in aeroplanes. How fears can develop to inherently neutral events is aptly demonstrated in Watson’s famous 1920 experiment with the baby, Albert, who became greatly frightened by the sight of a white rat after a loud aversive noise was sounded every time the animal was shown to the child. Before the experiment, Albert showed no fear, but after the conditioning trials he startled violently and began to cry whenever he saw the white rat. It is now generally agreed that fear conditioning involves more than the simple conditioned-stimulus responding seen in the experiment with Albert. In most cases, it seems to involve the learning of more complex hierarchical relationships between a threatening event and the various stimuli and contextual cues that predict it (Maren, 2001).
With the use of newer and better animal models and more advanced brain-imaging technology, the neurobiology of fear is now reasonably well-understood. Most authorities agree (for example, M. Davis, 1997; LeDoux, 1998; Maren, 2001) that the amygdala, which is a small walnut-sized structure deep inside the limbic brain area, is central to the regulation and conditioning of fear via its activation of various neurotransmitters and hormones. The primary function of the amygdala seems to be as a ‘protection device’ to evaluate environmental stimuli, and to disseminate information to a number of interconnected structures that co-ordinate autonomic and behavioural responses on the basis of this evaluation (Amaral, 2002).
Our first understanding of the importance of the amygdala came over a century ago, from observations of the effects of brain damage on emotional behaviour. Since that time, numerous other clinical and experimental studies have confirmed a notable loss of fear in people and animals with temporal lobe lesions. We have also seen that damage to this area has dramatic effects on one’s capacity to learn about stimuli that warn of danger. For example, rats lose their fear of cats, and people lose their ability to recognize that certain stimuli, such as the smell of smoke, signal an impending threat. In other words, not only does temporal lobe damage disrupt primary fears, but, under some conditions, it inhibits learned fears. More advanced research has revealed that the loss of fear accompanying temporal lobe lesions stems specifically from damage to the amygdala. Other studies have also shown that heightened fear may be induced by electrical stimulation to the amygdala.
Basically, our capacity to evaluate the risk or safety of stimuli in our environment is moderated by a complex and highly integrated circuit which is first activated by sensory information – the things we see, hear, smell or touch. All sensory neurons send projections to the amygdala, which in turn sends projections to various brain stem structures whose task is to ready us for action. For example, activation of the locus cereleus causes an increase in norepinephrine release which contributes to the increase in blood pressure, heart rate, and other behavioural responses that accompany fearfulness. This area also controls the level of attention to environmental events and the degree to which the individual is vigilant to signals of danger (Maren, 2001). Other amygdaloid projections to the hypothalamus stimulate the release of corticotrophin-releasing factor or hormone (CRF or CRH) which triggers the cascade of neurochemical reactions, known collectively as the ‘stress response’. The prime purpose of all these actions and reactions is to activate the sympathetic nervous system and to initiate the shipment of glucose to the brain and muscles to mobilize the organism in readiness for ‘flight or fight’.
Many now also believe that CRF is the master brain neurochemical regulator and integrator of the fear response because its cell bodies and receptors are found in abundance throughout the brain. For example, if CRF is infused in the brain of animals, they display a range of behaviours that is analogous to fear reactions in people, including elevations of heart rate and blood pressure. Moreover, when CRF is injected directly into the amygdaloid area, there is a prolongation of the fear reaction. Other studies have also shown that the density of CRF receptors is positively associated with the level of fear that experimental animals show to a novel situation.
Recent research has demonstrated the remarkable plasticity within the neural systems that mediate the fear response, and how environmental factors play a crucial role in determining whether or not we develop normal fear reactions. For instance, animal pups who have had good mothering (which has been defined in mice as frequent licking and grooming) show decreased behavioural fearfulness, and have increased benzodiazepine receptor densities in the amygdala and locus cereleus and decreased CRF receptor densities in the latter when they become adults (Caldji et al., 1998). These maternal influences are not simply due to genetics – that is, that fearful mothers simply pass on ‘fearful’ genes to their offspring – because mice pups from an inherently fearful strain showed less fear as adults when they were reared by an adoptive mother that groomed and licked them more frequently, compared to a control group of pups from the same strain.
In summary, there is now good evidence that CRF plays a central role in fear reactions, that the amygdala is a prominent structure in the brain’s fear circuit, and that one’s early environment is also crucial in regulating the normal development of these biological processes.
Efforts to understand the brain circuits and neural mechanisms of the fear system have also relied heavily on the study of fear conditioning. Some research has implicated the hippocampus, and its primary role in memory, in the conditioning to contextual cues, whereas other studies have shown that damage to the prefrontal cortex inhibits the extinction of conditioned responses. How these processes work is clearly relevant to our understanding of clinically significant fears such as we see in those with panic disorder and phobias. It is also interesting to note that stress impairs the function of the hippocampus and the prefrontal cortex, and that in many cases stressful life events (particularly in childhood) contribute to the onset of both panic and phobias (LeDoux, 1998).
Panic Attack Some hold the view that a panic attack is physiologically and behaviourally similar to a conditioned fear stimulus – albeit in an extreme and exaggerated form – and therefore that the same ‘fear network’ is activated. Elaborating on this view, Gorman and his colleagues (2000) proposed that a neurocognitive deficit in the cortical processing pathways results in a misinterpretation of bodily experiences, such as increased heart rate, via ‘upstream’ neural activation from the brain stem to the cortex, which then results in inappropriate ‘downstream’ activation of the fear network – in other words, misguided activation of the amygdala. Also, the fact that a whole host of agents/drugs with dissimilar biological properties are equally capable of inducing panic attacks in patients with Panic Disorder suggests that the onset of a panic attack may occur via several different triggers, but that whenever a trigger is present, the entire fear network is activated, not simply one pathway. Over time – or as a result of individual vulnerabilities – projections from the amygdala to the various brain stem sites, such as the locus cereleus, may become weaker or stronger, thereby accounting for the considerable variation in autonomic and neuroendocrine responses within an individual patient with panic disorder, or across the range of patients with this disorder.
Explained from a psychological viewpoint, several cognitive theorists have speculated on the factors that influence or initiate the vicious cycle of events that eventually culminates in a panic attack. The most popular model was proposed by Clark (1986) who argued that panic results primarily from the catastrophic misinterpretation of physical and mental sensations. In other words, some people tend to interpret the common symptoms of fear, like a racing heart beat, dizziness, or the shortness of breath, as much more dangerous than they really are. This, in turn, increases anxiety and apprehension producing even more pronounced bodily sensations which are then interpreted in an even more catastrophic manner and eventually the chain of synergistic events becomes a full-blown panic attack. Others have expanded on Clark’s (1986) model by proposing that the development of panic involves the fusion of a temperamental vulnerability and a congruent trigger (Moore & Zebb, 1999). The overarching vulnerability is a set of harm-avoidant and/or anxiety traits. However, more specific vulnerabilities, such as hypochondriacal health beliefs about certain bodily symptoms also play a role. Central to cognitive theories of panic is the phenomenon of selective attention since patients with this disorder seem especially hypervigilant to their own bodily sensations, and are highly prone to mentally scanning their bodies for signs of danger which then increases the likelihood that certain feared sensations are noticed and overinterpreted (Kroeze & van den Hout, 2000).
Phobias The study of fear is also fundamental to our understanding of how we come to acquire phobic reactions. Perhaps the most widespread and popular view – expressed in conditioning theories – is that phobias arise from, or are the consequence of, a previous traumatic experience, and that conditioning can occur in a variety of ways:
Although there is strong support for the notion that some phobias are acquired this way, for example, dental phobia (Poulton et al., 1997), this approach has failed to explain the development of all clinical phobias since only a minority of them can be attributed to Pavlovian conditioning events (see Poulton et al., 1998). One explanation why some people are apparently buffered whilst others are not has evoked the concept of latent inhibition and suggests that individuals who experience an aversive conditioning event later in life – without developing a phobic reaction – may have been ‘rescued’ because they had non-threatening exposures to the event (or something similar) early in life.
An alternative view is that the predisposition to develop most common phobias, such as the fear of heights, blood, water, and reptiles, is innate, mostly universal, and that what we actually learn as we go through life is how to overcome those fears (Rachman, 1978). In other words, non-associative or Darwinian models of fear acquisition suggest that most phobias occur without the involvement of learning or conditioning, and that the liability to phobias has arisen from evolutionary selection (Kendler et al., 2002). One interesting study found that falls resulting in head injury in childhood (before the age of nine years) had occurred more frequently in those without a fear of height at age 18 – a finding opposite to the predictions arising from conditioning theories of phobia, but consistent with the innate or non-associative theories of fear acquisition (Poulton et al., 1998).
On the other hand, non-associative theories may also be criticized since relatively few people suffer from clinically significant specific phobias, and because a large number of people have no reaction at all to these putatively innate fears. One explanation is that the ability to habituate to natural fear-arousing stimuli in the environment is influenced by individual differences in the tendency to arouse quickly and habituate slowly – ascribable to ‘neuroticism’, ‘stress reactivity’, or ‘negative affectivity’ in the prominent theories of personality we described in Chapter 3. Poulton and colleagues (2001) found that limited exposure to height stimuli in childhood (defined as time spent playing on swings and bars, and climbing trees and fences) – and therefore presumably less opportunity for habituation – was associated with a higher incidence of fear of heights during adolescence. Also, levels of neuroticism in childhood were higher among those who were height phobic in late adolescence compared to those who were dental phobic or had no fears. The authors concluded that some people do not overcome their innate fear of heights either because of limited opportunities to do so and/or because they are at the higher end of a continuum of emotional reactivity and therefore fail to habituate.
The Stress-Diathesis model of phobias extends the interaction model of vulnerability by proposing that a combination of environmental adversity and individual vulnerability gives rise to the condition, but importantly, that an inverse relationship exists between the two. In other words, individuals whose disorder was associated with severe environmental stress should, on average, have lower levels of personal vulnerability than those whose disorder was not associated with environmental stress. However, in a recent test of this model, Kendler et al. (2002) found no support for traditional associative theories, such as the Stress-Diathesis model, which assume environmental conditioning. Instead, their results were more compatible with Darwinian theories and individual vulnerability models. First of all, while neuroticism strongly predicted the risk of all phobias, it was not higher in phobias with no memory of an environmental trauma, nor was it lower in those who associated the onset of their phobia with a personal trauma. In all aspects, Kendler’s study is in accord with the growing body of research suggesting that the development of many phobias is non-associative and that personal vulnerability is the key predictor.
Perhaps the most parsimonious and intuitive way to integrate associative and non-associative approaches to phobia is to take the position that conditioning processes are more salient in the development of fears and phobias that are ‘evolutionary-neutral’, such as dental fear, that inherent biological processes are the primary determinants for ‘evolutionary-relevant’ fears, such as the fear of height, and that personality factors contribute to the onset of both types of phobia (Poulton et al., 2001).
Over the years, and with each new version of the DSM, the number of categories subsumed under the heading of anxiety disorders has increased – from three in the first version to 12 in the current edition. Some have expressed concern that efforts at diagnostic precision may have come at the expense of a better understanding of the shared and overlapping features of these disorders. Perhaps the set of symptoms that define Generalized Anxiety Disorder – a state of chronic worry, apprehension, and hypervigilance which is usually accompanied by somatic symptoms, such as muscle tension – best captures what most of us understand by the term anxiousness and how it differs from, though is obviously related to, the concept of fearfulness. Generalized Anxiety Disorder is probably the most basic emotional disorder because it comprises features that are present in varying degrees in all other mood and anxiety disorders, and because it reflects the principal personality dimensions of risk for these syndromes; factors such as neuroticism and negative affect. In fact, there is a debate about whether Generalized Anxiety Disorder is better conceptualized as the extreme of a trait that confers non-specific risk to a number of disorders, rather than as a distinct Axis I construct that is influenced, along with other anxiety and mood disorders (albeit to a greater extent), by the higher-order personality factors referred to above (Brown et al., 1998). That neuroticism is a risk in the development of anxiety disorders is hardly surprising because, by definition, it is a trait that predicts people’s susceptibility to anxious states. However, both Eysenck’s and Gray’s theories also predict that anxiousness is inversely correlated with the extraversion dimension. Although much research has focused on the additive effects of these two trait dimensions, some studies have found an interactive and synergistic effect of neuroticism and introversion on the risk for anxiety (see Jorm et al., 2000, for a review). In other words, in these studies neuroticism was only associated with heightened anxiety in the presence of introversion.
There is also considerable evidence that cognitive factors play a role in the development and maintenance of anxiety disorders. For example, theorists such as Beck (1986) and Clark (1986) have emphasized the increased risk amongst those whose information-processing style tends to overinterpret the threat or danger associated with physical and psychosocial events in their environment. Other research has focused on attentional biases. Either because of biological factors or simply habit, some individuals selectively attend to certain stimuli in their environment; then again, because of habit or self-control, their attention is either sustained or it shifts to other events. We have learnt that people with anxious orientations tend to shift their attention frequently, being finely tuned to, and often scanning for, sources of threat in their environment. They also have a heightened tendency to perceive threat in ambiguous events. For example, a strange and unfamiliar noise is more likely to be viewed as anxiety-provoking than as something unimportant and worth ignoring. Another key characteristic of anxiousness is that it is quintessentially about ‘worry’ and therefore it is future-oriented. In other words, prospective judgements tend to have a more negative tone in anxious states, compared to melancholic states which are focused on past events, as we shall see in the next section (Hertel, 2002).
Consistent with cognitive theories is the concept of anxiety sensitivity which is viewed as a relatively stable cognitive individual difference variable that basically describes the ‘fear of fear’. Those high in this trait tend to fear the overt bodily sensations, such as trembling, a racing heart rate, dizziness, and blushing, that are related to high physiological arousal. Usually this fear is generated by a belief that their anxiety symptoms are a portent of some forthcoming catastrophic physical event like a heart attack, that they signal a loss of cognitive control (as seen in some serious mental disturbance), or because of social concerns, such as the embarrassment and humiliation they might cause (Stewart & Kushner, 2001).
There has also been some seemingly unnecessary debate about the degree to which anxiety sensitivity is related to, or distinct from, trait anxiety. We say ‘unnecessary’ because it would be wholly surprising and counterintuitive if the two constructs were not associated. Indeed, in most studies they are moderately correlated, in the order of 0.4–0.5. The most obvious view is that anxiety sensitivity is a lower-order factor of trait anxiety which can accelerate anxiousness in those with high trait anxiety, or even foster it among those with low levels of anxiety. On the other hand, a measure of their distinctiveness is that there is not much more than about 20 per cent shared variance between their respective measures.
Like so many other symptoms and disorders we discuss in this book, the neurobiology of anxiousness has been related to potential abnormalities in a range of neurochemical, neuroendocrine, and neuroanatomical processes. Perhaps the strongest evidence – in part, because it has been the most studied – has implicated overactivity of brain 5-HT. This was a serendipitous finding first brought to light in the 1980s when it was discovered that a drug developed to treat psychotic patients was also useful for treating anxiety disorders (Snyder, 2002). Since then we have learnt that when post-synaptic 5-HT2receptors are activated in the limbic region there is an increase in anxiety and avoidance behaviour (Connor & Davidson, 1998). It has also been confirmed that 5-HT activation has a dual role in the regulation of defence. Relevant to generalized anxiousness is the fact that 5-HT enhances learnt responses to potential threats in the environment through its action in the forebrain (Graeff, 2002). It is also known that specific agonists of the 5-HT1A autoreceptors in the forebrain have anxiolytic properties in both humans and animals, and that increased anxious behaviours are seen in 5-HT1Areceptor ‘knock-out’ mice (Ramboz et al., 1998). Some very recent research has also demonstrated the crucial role of the early post-natal period in the subsequent development of normal anxious reactions in adulthood. Specifically, it appears that appropriate stimulation of the 5-HT1Areceptor during this period is necessary to set in motion the long-lasting changes in the brain that are essential for normal anxiety behaviours (Gross et al., 2002).
Social Anxiety In the remainder of this section, we shall look at two other relatively common disorders whose primary feature is the extreme expression of anxiousness. Social Anxiety Disorder (sometimes called ‘social phobia’) is a relatively common condition – indeed, the most common of the anxiety disorders, and the third most common psychiatric disorder following major depression and alcohol dependence – with a lifetime prevalence of about 15 per cent. Yet, many feel that it is still relatively underdiagnosed, undertreated, and one of the least understood of all psychiatric illnesses (Brunello et al., 2000; Li et al., 2001). The quintessential feature of this disorder is the extreme anxiety that is experienced in social situations, and most acutely when the individual has to perform in public or be exposed to the scrutiny of others, such as during a job interview or a class presentation. Typically, sufferers fear that while they are being watched by others they may be evaluated negatively, and/or that the signs of their anxiousness, such as blushing, shaking and sweating, will cause them great embarrassment. The non-generalized or stereotypical version of social anxiety is mostly confined to a single, or no more than a few, performance situations, such as speaking in front of an audience, whereas the generalized form involves the excessive fear of a broader array of social situations (Stein, 1998). Although social anxiety generally starts in early childhood or adolescence, it seldom occurs in isolation – being more frequently comorbid with a range of other disorders, including depression, other anxiety disorders, substance abuse, and eating disorders. What is interesting, however, is that social anxiety is usually primary to this ‘cascade of comorbidity’, suggesting that it may serve a causal role in these related forms of psychopathology (Brunello et al., 2000).
The DSM classification of social anxiety has caused a certain amount of debate, and some cynicism, because its diagnosis seems to be largely a quantitative issue and a matter of clinical judgement based on the level of associated impairment. Indeed, some have strongly criticized the very existence of social anxiety as a ‘disorder’ claiming it to be nothing more than the ‘medicalization’ of shyness (for example, Talbot, 2000). However, this sort of criticism may aptly be applied to many other mental and physical illnesses whose symptoms, although they cause considerable distress to the individual, are simply the extreme end of a continuum of normal behaviour. Since we are, by nature, a species that typically finds pleasure in social interaction, there is little doubt that some level of pathology exists when one’s social experience becomes a source of distress instead of comfort (Insel, 2002). An important question that arises is how social anxiousness is different from other forms of anxiousness? One answer lies in the fact that social information seems to be processed in our brains in a different way from other sorts of information. In other words, specific brain pathways seem to have evolved for the sole purpose of processing social stimuli (Young, 2002).
Most aetiological models of social anxiety propose an interaction of psychobiological risk factors that are exacerbated by negative thoughts and avoidance behaviours (Kashdan & Herbert, 2001). An easily identifiable temperamental style, which Kagan (1989) called behavioural inhibition, and which is already apparent in infancy, is a strong predictor of the later onset of social anxiety (and other anxiety disorders). Children with this disposition withdraw from novel social settings, are irritable and sleepless as babies, and become shy and socially inhibited as they get older. It is easy to see how certain life experiences, laid over this temperamental foundation, could contribute to excessive social anxiety. For example, some research has implicated certain kinds of parent–child interactions, such as harsh maternal criticism, overprotection and overcontrol, and the modelling of anxiety (Coupland, 2001).
Early biological research on social anxiety focused primarily on the noradrenaline system because the symptoms of this disorder, such as blushing, sweating, and palpitations, are all characteristic of autonomic overactivity. However, more recently there has been a focus on the serotonergic and dopaminergic pathways. Although there is little evidence to suggest that current 5-HT models of anxiety are directly informative about social anxiety, we do know that the traits that correlate with this condition, such as submissiveness and social interaction, are regulated by 5-HT function, as are other characteristics of more generalized anxiousness (Li et al., 2001).
A more informative line of neurobiological research has linked social anxiety with diminished dopamine transmission in the mesolimbic reward pathway, as seen by lower D2 receptor binding potential in patients with this disorder, and by the observation that neuroleptic drugs (dopamine blockers) tend to increase social anxiety. There are interesting parallels in primates of subordinate social status, and in a strain of particularly timid mice, both of whom show lower striatal dopamine (see Schneier et al., 2000; Stein et al., 2002 for reviews). Low striatal dopamine has also been implicated in other aspects of personality, such as avoidant, schizoid, and detachment traits (Li et al., 2001). And, social anxiety is increased in Parkinson’s disease in which there is a pronounced deficit of striatal dopamine. Stein (1998) presents a particularly interesting integration of the two primary neurotransmitter pathways discussed above by proposing an imbalance between risk assessment (probably modulated at the level of the amygdala) and the capacity for experiencing reward, so that the ‘risk’ of engaging in social interaction far outweighs any ‘reward’ that it could offer.
Post Traumatic Stress Disorder Post Traumatic Stress Disorder (PTSD) is alleged to afflict between 10 per cent and 15 per cent of the population sometime in their life, with its prevalence twice as frequent in women as in men. Although this condition is, by definition, linked to a traumatic, fearful event, patients with PTSD typically show quite normal fear reactions, although their anxiety reactions are elevated (M. Davis, 1998). Almost all individuals who experience a greatly frightening or traumatic event develop some psychological symptoms of anxiety in its aftermath. However, in most individuals, and over the space of time, these symptoms begin to abate. In only a subset of trauma survivors do clinically significant PTSD symptoms become chronic. In recent years, the factors that influence why some are buffered while others succumb to the distress of lingering PTSD symptoms have generated considerable research interest. One of the most salient predictors of PTSD seems to be the nature of the traumatic event, with human violence, such as torture or victimization, producing a substantially higher prevalence than natural disasters, such as earthquakes and floods (Yehuda et al., 1998).
Although more than half of those who develop PTSD meet the diagnostic criteria for other mood and anxiety disorders – indicating once again the substantial overlap in the symptoms of all these disorders – PTSD seems to have a neurobiology rather distinct from other mood and anxiety disorders. The evidence points to a process whereby the initial traumatic experience becomes transduced into an anxiety reaction because of a disruption in the normal cascade of biological events that form the fear response and, more particularly, its resolution (Yehuda et al., 1998). For example, different from what is typically observed in cases of depression, patients with PTSD are characterized by decreased levels of circulating cortisol and an increased sensitivity of the HPA negative feedback inhibition. In other words, it appears that the HPA axis is hypersensitive to inhibition feedback because of an increase in the number and the sensitivity of glucocorticoid receptors. An animal model of PTSD has also shown that although the basal cortisol levels in these animals were no different from control animals, after the stress challenge they consistently exhibited HPA hyporesponsivity to subsequent major stressors (King et al., 2001). Based on available data, Yehuda and colleagues proposed that due to certain – as yet not fully understood – risk factors some individuals, in response to a traumatic event, may not release sufficient levels of cortisol to shut down the sympathetic nervous system (SNS) response. In turn, a prolonged and heightened SNS response tends to disrupt the normal processing of memory and sustains distress by causing repeated re-experiencing and reprocessing of the event.
It also appears that pre-traumatic vulnerability factors may contribute to the cause of this disrupted process. For instance, studies have found that psychological disturbance identified before the trauma, and a disruption of social support networks, consistently emerged as strong predictors of psychopathology following exposure to trauma (Pine & Cohen, 2002). Not surprisingly, neuroticism has also been positively associated with PTSD both in retrospective and prospective studies (see Holeva & Tarrier, 2001). Perhaps the risk is conferred because those with high levels of this trait show a propensity to become aroused and conditioned more easily. However, since neuroticism correlates with almost all aspects of psychopathology, this sort of information is not particularly helpful. Nevertheless, it is consistent with the accumulating evidence that the response to trauma and the likelihood of developing PTSD have a genetic basis (King et al., 2001)
At the beginning of this chapter we emphasized that, despite some obvious common ground, depression is a very heterogeneous clinical condition with at least two, quite distinct, subtypes, the first having symptoms that overlap greatly with those of the anxiety disorders. Therefore, in this section, we shall concentrate our discussion on the core symptom of the melancholic type of depression, characteriseed by a profound anhedonia and an almost total inability to experience pleasure, which has been explained in biological terms by an under-activation of the brain circuits that regulate reward and reinforcement.
Research over the past several decades has improved our understanding of two higher-order motivational systems. One is related to approach, and often called Positive Activation (or Affectivity), and the other to withdrawal and called Negative Activation (or Affectivity). These two dimensions represent the subjective or experiential components of more general biobehavioural systems that mediate activities directed towards the acquisition of positive goals, and those that are designed to avoid negative events. Although these systems have been given a variety of names, behavioural inhibition and behavioural activation are perhaps the most commonly used (as we have seen in Chapter 3). Not surprisingly, they tend to correlate strongly and systematically with measures of neuroticism and extraversion, respectively (Watson et al., 1999). As we have seen in the previous section, high Negative Activation relates strongly to symptoms of fear and anxiety. On the other hand, low Positive Activation is the defining feature of melancholy.
The longest standing, and most persistent, biological theory of depression derives from the monoamine hypothesis which proposes that a central nervous system depletion of serotonin, norepinephrine and/or dopamine underlies the melancholic symptoms of this disorder. By far the most studied of these neurotransmitters is 5-HT. However, the supporting evidence is largely based on working backwards from the known mechanisms of drugs which either induce depression, or can successfully alleviate its symptoms. A key player in the formulation of this hypothesis was the discovery, half a century ago, that depression was a common side-effect of one of the first effective hypertensive medications (reserpine) – the link being made because it was also known that reserpine depleted brain 5-HT stores (Hirshfeld, 2000). Since then, the neurobiological study of depression has been inextricably linked to the mechanism of action of effective antidepressants, the newer of which (like the serotonin reuptake inhibitors (SRIs)) elevate brain levels of 5-HT. A brief description of the SRIs is given below.
Serotonin reuptake inhibitor drugs (SRIs)
SRIs, such as fluoxetine (the most familiar of these is marketed as Prozac), have a well-established antidepressant and antiobsessional effect, making them a highly popular treatment option for many patients with OCD and depressive disorders. The pharmacological mechanism of these drugs is to inhibit the removal of 5-HT from the synaptic cleft, back into the releasing neuron, by blocking the transporter or uptake pump that executes this function. In theory, the result of this blocking action is that more 5-HT is available in the synapse, and therefore able to bind to postsynaptic receptors. However, an important question, that has puzzled researchers since the introduction of these drugs, is why there is a delayed reaction of several weeks in their effectiveness for symptom improvement.
One convincing explanation is that although the initial use of these drugs serves to flood the synapse with 5-HT when the uptake pump is blocked, this action causes rapid sensitization of the 5-HT inhibitory autoreceptors located on the cell body and terminals of the serotonergic neurons. This serves to slow down the firing rate of the neuron. The net result is a small, but insignificant, enhancement of 5-HT. However, it is believed that over time, chronic use of SRI drugs helps to desensitize the overreactive 5-HT autoreceptors. In other words, with long-term use, 5-HT availability is eventually increased because of enhanced release from the neuron and diminished reuptake of 5-HT.
Despite the appeal of pharmacokinetics explanations, the action of these drugs is still only partially understood. And because their precise mechanism continues to defy a clear explanation, neurochemical experts often resort to a rather evasive explanation by saying that the action of the SRIs is not simply to increase or decrease 5-HT, but rather to make this neurotransmitter system work more precisely in modulating mood and impulse control. It has also been suggested that the SRI drugs cause a shift in the balance of tone in the indirect versus the direct orbitofrontal-subcortical pathways, decreasing activity in the overall circuit (Saxena et al., 1998).
Over the years, a wealth of indirect evidence has also supported the role of 5-HT in melancholy. For example, concentrations of a certain 5-HT metabolite are lower in patients with depression, and low concentrations in the central spinal fluid and cortex have been found in those who commit violent suicide. However, the difficulty with this sort of research is that no specific biochemical lesion has been successfully and consistently implicated, although certain candidates appear more likely than others. For instance, depleted 5-HT could be the result of defective synthesis in the cell body, reduced release from the cell, enhanced activity of the inhibitory autoreceptors, and impaired post-synaptic function (Leonard, 2000).
Recently, a more direct test of the monoamine hypothesis has been achieved with 5-HT depletion studies – the underlying rationale being that if melancholy is caused by a deficiency of this neurotransmitter, an experimentally induced depletion in normal subjects should foster the depressive symptoms seen in patients with this disorder. However, results from such studies have not been supportive since depletion of 5-HT and norepinephrine does not usually cause depressive symptoms in healthy volunteers. Nor does it worsen symptoms in unmedicated depressed patients (Delgado, 2000).
Another important line of neurobiological research has linked hemispheric asymmetry in the prefrontal cortex to melancholy. In a review of the data that has accumulated in this area, Tomarken and Keener (1998) concluded that resting levels of left prefrontal activation primarily reflect individual differences in the motivation to approach rewarding stimuli and the degree of responsiveness to pleasurable environmental events. Conversely, resting levels of right prefrontal activation reflect individual differences in withdrawal from stimuli that are potentially threatening (describing a system that is clearly relevant to our earlier discussions of fearfulness and anxiousness). Depression studies have shown that decreased activation in the left prefrontal cortex relative to the right (as measured, for example, by electroencephalographic (EEG) activity), is consistently correlated with measures of anhedonia, whereas happy and cheerful people tend to have higher resting activation in the left prefrontal cortex (Henriques & Davidson, 2000). Studies have also examined frontal asymmetry in the infants and children of depressed mothers and found left frontal hypoactivation compared to children of non-depressed mothers, further suggesting that frontal asymmetry may be a genetically based marker of risk for depression. However, asymmetry could also be the result of socio-environmental influences since we know that patterns of stimulation at critical periods of a child’s development can produce changes in cortical connectivity. Therefore, it is possible that depressed mothers may show decreased signs of pleasure to their infant offspring, which may then retard the child’s development of essential cortical connections (Tomarken & Keener, 1998). Most probable, however, is that transmitted risk is a function both of genetic and of environmental factors.
Patterns complementary to the cortical activity described above are also seen at the subcortical level. The brain’s mesolimbic dopamine reward pathway is fundamentally implicated in melancholy since the level of functioning of this circuitry underlies individual differences in the capacity to experience pleasure, and the motivation to approach pleasurable stimuli. An extensive body of research – much of it conducted in the context of addiction research (see Chapter 7) – has confirmed the role of this neurotransmitter pathway in approach-related behaviours. Depue and colleagues have also shown positive correlations between dopamine availability and human variability in personality traits characterized by positive affect, such as extraversion (Depue & Collins, 1999). Relevant to the research on cortical asymmetry is the fact that projections from the mesolimbic region to the cortex tend to be concentrated in the left hemisphere (Watson et al., 1999).
The high comorbidity between Major Depressive Disorder and drug abuse has also drawn attention to similarities in the neurobiology of the two conditions, with suggestions that the two disorders are different symptomatic expressions of the same pre-existing biological abnormalities, or that the effects of chronic drug exposure lead to biochemical changes which have some common elements with the biological events that mediate depression. We also know that stressful environmental factors mimic the brain’s reaction to drug use (Markou et al., 1998).
Perhaps the best demonstration of the role of mesolimbic dopamine in both melancholy (and addiction) is seen in the chronic mild stress (CMS) model where it is clearly established that the anhedonic effects of CMS are mediated by a downregulation of certain dopamine receptors in the nucleus accumbens, and therefore, a net decrease in dopamine availability (Willner, 1997). However, in the context of this model, it is important to emphasize that reward-seeking behaviour (something more closely linked to the concept of impulsivity) is conceptually distinct from the response to a reward when it is presented. The latter is often evaluated by preference ratings in studies with human subjects. Although it seems reasonable to think the two would go together – viz. wanting something and liking it – the distinction between them has a ‘long and distinguished intellectual pedigree’ and forms the basis for many theories of addiction, as we explain in Chapter 7 (Willner et al., 1998).
A variety of studies have highlighted the importance of the dopamine reward pathway to behavioural activation. When compared to non-depressed subjects, both depressed and subclinically depressed subjects showed a decreased responsiveness to reward, but not punishment, in a verbal memory task that was carried out under three monetary reward conditions (neutral, reward, and punishment) (Henriques & Davidson, 2000). Another study monitored subjective responses to a mild dose of amphetamine and found that the degree of positive mood induced by the drug was strongly correlated with the severity of the subject’s depression (Tremblay et al., 2002). Depressed patients with severe symptoms experienced greater reward from the drug compared to control subjects, whereas the mood ratings of those with moderate depression were no different from control subjects. The investigators concluded that the hypersensitive brain reward-system response in the depressed patients is likely to reflect a downregulated or hypofunctional state in these patients.
While biological research has provided some insights into the causes of depression, cognitive explanations have also improved our understanding of depressive processes. Most cognitive models, including that of Beck and the later hopelessness/helplessness models, are based on a ‘cognitive vulnerability hypothesis’ and describe the way in which dysfunctional thinking patterns foster the development of depressive episodes (see Just et al., 2001 for a review). In other words, negative thinking is not seen simply as a symptom of depression; it is also a causal antecedent stemming from the maladaptive way depressed individuals process information from their social world. For example, they may selectively attend to information that is unflattering to them, to interpret information negatively when there is some ambiguity, and endorse largely pessimistic beliefs – information that ultimately results in a negative self-concept. As a result, they may pursue social goals that result in a lack of positive reinforcement or they may behave in a socially undesirable way because they lack the requisite social skills (Street et al., 1999).
Many experimental studies conducted with pleasant and unpleasant word recall, or with autobiographical recollection, have shown that depressed people tend to recall more negatively toned past events than non-depressed individuals, and they do this more frequently than they recall positive events. In other words, they tend to remember and focus their attention on unhappy and unflattering information from the past. Some have suggested that this memory bias may be a maintenance mechanism in depression because it inhibits the individual from engaging in effective ‘mood-repairing’ activities (Watkins, 2002). For example, when depressed individuals are invited to a social event they are more likely to remember something unpleasant that occurred on a previous occasion they attended, which then leads to the view that parties do not improve their mood, and so the event would probably be avoided. In this way the depression is maintained, and even exacerbated, by reinforced negative memories, by increased social isolation, and by missed opportunities to develop more adaptive coping skills.
One hallmark of depressive thinking is rumination which, by definition, entails an attentional focus on past events. Although the intensity of one’s thoughts during rumination suggests a high degree of cognitive control, in fact just the opposite is the case. There is a rather forceful lack of control in the ability to switch attention to other matters since negative thoughts seem to intrude automatically into the mind of the depressive, inviting more and more frequent gloomy thoughts. The work of Wegner and colleagues (1987), on the ironic effects of thought suppression, has had considerable impact on our understanding of ruminative thoughts in depression (see Purdon, 1999 for a review). In Wegner’s original studies, subjects were asked either to suppress or express thoughts about a white bear whilst thinking aloud their stream of consciousness. Later in the study, the suppression/expression instructions were reversed, and it was found that subjects who had ‘suppressed’ first had more thought occurrences of the bear in the second phase. In other words, the very cognitive processes that are used to help us suppress thoughts paradoxically also seem to work to foster the elicitation of those same thoughts.
Thought suppression has been implicated in various psychological disorders, such as depression, that are characterized by the intrusive distress of unwanted thoughts. One explanation for why suppression of pessimistic thoughts tends to result in their more frequent occurrence is because, at least among the depressed, there is an absence of positive distracting thoughts and a hyperaccessibility of other negative thoughts. However, not all studies have found these effects when applied to the suppression of neutral thoughts, highlighting once again the important role of individual vulnerability in all psychological phenomena.
The above mention of rumination and thought suppression draws attention to another psychological disorder where those phenomena play a major role in the mechanisms of symptom formation and maintenance. We are referring to Obsessive–Compulsive Disorder, a condition which we consider in detail in the following chapter.