A number of theories have been suggested and tested; for example, faulty patterns of thinking or perception have been blamed, and many possible biochemical or physiological markers have been proposed. Still, the exact etiology has not been found. It will be obvious from the following discussion that lack of effort is certainly not at the bottom of our present ignorance.
There are several theoretical models that aspire to explain the etiology of panic disorder. Despite the limitations of categorization, the most plausible theories may be sorted out into major models: cognitivo-behavioral and biomedical.
Clark's theory of catastrophic interpretations (1988) sees panic attacks as a result of maladaptive and faulty interpretation of bodily sensations. Physical sensations that would be processed as normal, not alarming, or not registered at all by healthy individuals, are perceived as more dangerous than they really are and interpreted as an imminent physical or mental catastrophe. For instance, palpitations would be interpreted as a heart attack. According to Clark, such catastrophic interpretations trigger the panic attack. This theory would provide a plausible explanation for panic attacks during relaxation, since bodily sensations are more apparent when the person is relaxed. However, it cannot account for panic attacks during the deep stages of sleep, even though some might argue that some kind of a cognitive filter remains active, screens out unimportant signals, and lets those of personal importance go through. Such mechanism can be exemplified by sleeping mother who wakes up when her baby cries while nothing else disturbs her. However, it is questionable whether the mother's reflex to her child's distress is really comparable to the processing required for cognitive interpretation of bodily sensations. Thus, this theory may provide us with a partial explanation for the panic attack trigger or else of the escalation of symptoms once started; however, we cannot assume that such trigger is the only stimulus responsible for panic attacks.
Beck et al. (1985) proposed a similar model that emphasizes the importance of both predisposing and precipitating factors. Heredity, certain physical conditions, ineffective coping skills, or trauma, for example, may serve as predisposing factors that make some individuals vulnerable to the effects of precipitating factors, i.e. immediate stressors such as loss of someone close, anniversaries, physical illness, use of drugs, exposure to toxic substances and so forth. According to Beck et al., the symptoms of anxiety follow, in a chain reaction manner, the initial impression of dying. Having experienced a series of panic attacks, the patient elaborates a set of automatic thoughts, i.e. cognitive shortcuts that require little processing and jump directly to faulty conclusions which focus mostly on impeding danger, madness, harm or death. Therefore, Beck et al. advance that agoraphobia is an association that easily forms between panic attacks and certain places or circumstances, especially those often feared by small children, such as tunnel, heights, dark or large crowds. The patients then fear being far from home, avoid certain places and often need a safe companion to provide them with some sense of security while away.
Ehlers et al. (1988) proposed a model that explains panic attacks as a result of selective focusing on panicogenic interoception. According to this model, PD patients pay excessive attention to their somatic sensations. They are thus more likely to notice, perceive and react to common interoceptive stimuli. Such sensations then trigger a panic attack. In a recent article, Ehlers (1993) suggests a modification to this theory of interoceptive phobia, arguing that, rather than allocating too much attentional resources to interoceptive stimuli, PD patients are more accurate and capable of detecting actual somatic changes than their healthy counterparts. These suggestions were challenged by the outcome of a study conducted by Rapee (1994) who assessed accuracy of detection of physiological changes following an inhalation of 5, 10, and 20 % CO2 or room air in PD patients and healthy controls. Their results show no significant differences between the two groups either in terms of approximating the CO2 content of the inhaled air, or in the number of physiological symptoms reported.
The model of learned alarm reaction conceived by Barlow (1968) is based on the postulated similarity between panic attack and the physiological fight or flight reaction. It asserts that panic attack is essentially a fight or flight reaction in the absence of real danger. Therefore the alarm of the entire organism, that is so useful in case of real danger, becomes false and thus maladaptive. Because of the strength of the experience, pairing rapidly occurs between the false alarm and the interoceptive physiological sensations experienced during a panic attack. After this conditioning, whenever there is a stimulus resembling the physiological sensations associated with the attack, a false alarm reaction and consequently a panic attack are triggered. This theory does not discard the neurobiological bases of PD; indeed, it is suggested that the false alarm itself presents an expression of some neurobiological malfunction.
Goldstein and Chambless (1978) propose a model of panic disorder with agoraphobia that distinguishes between simple and complex agoraphobia. Anxiety provoked by traumatism, drugs or physical illness falls into their category of simple agoraphobia. Patients with complex agoraphobia, on the other hand, would often lack assertiveness and independence, and would have low self-sufficiency appraisal. Panic attacks in these patients trigger anticipatory anxiety and a vicious fear-of-fear circle that ends in avoidance of specific situations in which distressing feelings have been experienced and conditioning has occurred between the fear and the circumstances.
All the cognitivo-behavioral theories provide plausible hypotheses of the etiology of PD, and certainly reflect the experience of people who suffer from it. After all, these models are based on clinical experience with PD patients. On the other hand, their truthfulness does not discredit the claim that these patients have also an abnormality in their neurobiology that is either the cause or a consequence of their life experiences.