Psychogenic:  Background Information

Psychogenic Menu

Psychological Mechanisms of Medically Unexplained Symptoms

Permission graciously given by the author to reproduce this paper

Psychological Bulletin © 2004 by the American Psychological Association 

September 2004 Vol. 130, No. 5, 793-812 For personal use only--not for distribution. Richard J. Brown, Academic Division of Clinical Psychology, University of Manchester.  Department of Clinical Epilepsy, Institute of Neurology 

Psychological Mechanisms of Medically Unexplained Symptoms
An Integrative Conceptual Model


Theories of medically unexplained illness based on the concepts of dissociation, conversion, and somatization are summarized. Evidence cited in support of these theories is described and the conceptual strengths and shortcomings of each approach are considered. It is argued that each of these approaches adds to the understanding of unexplained illness but that none is able to provide a comprehensive explanation of the phenomenon. An integrative conceptual model of unexplained illness based on cognitive psychological principles is then presented. This model attempts to combine existing theoretical approaches within a single explanatory framework, extending previous theory by explaining how compelling symptoms can exist in the absence of organic pathology. The clinical and empirical implications of the model are then considered.

Throughout history, medical practitioners have encountered individuals with symptoms of physical illness for which no adequate organic basis can be found. According to some estimates, between 25% and 60% of the symptoms investigated in family medical centers defy adequate physical explanation (Kirkwood et al., 1982). In some cases, symptoms represent the somatic component of a diagnosable psychiatric condition, such as anxiety or depression, but are not recognized as such by the patient or the physician; in others, normal bodily sensations or minor physical ailments may be perceived as evidence of serious underlying pathology by a hypochondriacal individual. In many cases, however, there is no obvious psychiatric explanation for the symptoms in question (Kroenke & Swindle, 2000). Although many such “medically unexplained” symptoms resolve quickly and spontaneously, symptoms often remain unresolved, causing distress and disability that persists over time. In addition to the suffering and uncertainty faced by patients in this context, the cost of repeated consultation and investigation for these individuals represents a considerable burden to the health care services. This is particularly true in the case of individuals who experience and seek help for multiple unexplained symptoms, for whom health care costs may be up to nine times the primary care average (G. R. Smith, Monson, & Ray, 1986). 

The psychological origin of medically unexplained symptoms has been widely recognized since the 18th century (Merskey, 1979), and a number of models attempting to explain these phenomena have been advanced since this time. This article provides a critical review of the different theories purporting to explain the pathogenesis of medically unexplained symptoms, considering both the merits and the limitations of current ideas in this area. On the basis of this review, I argue that (a) different theories are able to account for certain features of medically unexplained illness, but no one theory can accommodate all of the available evidence concerning these conditions, and that (b) no theory provides a satisfactory account of how apparently compelling symptoms can exist in the absence of significant organic pathology. I then describe a novel model that aims to address these shortcomings by reference to contemporary cognitive psychological research. Rather than superceding existing theories in this area, this model has been developed with a view to integrating current theoretical concepts within a single explanatory framework. This framework provides a useful tool to organize available research and theory concerning medically unexplained symptoms and allows for the generation of new hypotheses concerning these phenomena. 



Broadly speaking, the fourth edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV; American Psychiatric Association, 1994) categorizes medically unexplained complaints according to the nature, number, and duration of the symptoms in question. The somatoform disorders category encompasses a range of complaints characterized by the symptoms of somatic illness (e.g., pain, fatigue, general malaise) that cannot be accounted for by identifiable physical pathology, including the contemporary manifestation of polysymptomatic hysteria (“Briquet's syndrome”), so-called somatization disorder. The conversion disorders category, a subcategory of the somatoform disorders, encompasses unexplained symptoms suggestive of a neurological condition, such as gait disturbances, convulsions, sensory loss, and cognitive decline; symptoms specifically involving a disruption of memory, perception, and identity are classified as dissociative disorders, as are the phenomena of depersonalization and derealization. DSM-IV also identifies hypochondriasis and body dysmorphic disorder as somatoform phenomena; neither of these conditions has medically unexplained symptoms as a defining feature, and they are not considered in any detail here. 

In all cases, symptoms are classified as medically unexplained if adequate investigation has failed to identify a plausible physical cause for their occurrence or their associated level of impairment; they must also cause clinically significant functional disability or distress to meet diagnostic criteria. Moreover, the symptoms cannot be solely attributable to diagnosable anxiety, depression, hypochondriasis, or psychosis. 


Unexplained symptoms are ubiquitous, representing the commonest single category of complaints encountered in primary care (Kirmayer & Taillefer, 1997). Evidence suggests that more than a quarter of unselected primary care patients meet criteria for DSM-IV undifferentiated somatoform disorder, defined as the presence of one or more debilitating unexplained symptom of at least 6 months duration (Fink, Sørensen, Engberg, Holm, & Munk-Jørgensen, 1999). Patients with a history of multiple unexplained symptoms are also extremely common. Data reported by Escobar, Waitzkin, Cohen Silver, Gara, and Holman (1998), for example, indicate that 22% of primary care patients meet criteria for the so-called abridged somatization construct, defined as the lifetime occurrence of four unexplained symptoms in men, six in women. Many patients with a history of multiple unexplained symptoms report experiencing several symptoms simultaneously. Kroenke et al. (1997), for example, presented data indicating that more than 8% of unselected primary care patients meet criteria for so-called multisomatoform disorder, defined as the presence of three or more currently distressing unexplained symptoms alongside a history of somatoform illness. 

Research investigating the epidemiology of conversion disorder has typically used specialist populations, and consequently, little is known about its primary care prevalence. Mace and Trimble (1996) suggested that as many as 20% of specialist neurological patients present with symptoms that defy adequate organic explanation (for a review, see Akagi & House, 2001). 

Clinical Features

The most commonly observed unexplained symptoms in primary care are pain, fatigue, and general malaise (Kirkwood et al., 1982), although gastrointestinal and sexual/reproductive symptoms are frequently found. Within neurological settings, both “negative” (i.e., those characterized by a loss of normal functioning; e.g., sensory loss, amnesia, paralysis, nonconvulsive pseudoseizures) and “positive” symptoms (i.e., those characterized by the presence of additional symptoms that disrupt functioning; e.g., gait disturbance, tremor, pseudohallucinations, convulsive pseudoseizures) are common. Broadly speaking, the most common somatoform phenomena involve a significant subjective component (e.g., pain, fatigue, dizziness) or involve a disruption in the voluntary control of action (e.g., gait disturbance, urinary retention, dysphagia). Unexplained symptoms are significantly more common in women than in men, although the phenomenon is no longer considered an exclusively female problem. 


Distinguishing somatoform symptoms from their organic counterparts can be extremely difficult and typically requires extensive physical investigation before a firm diagnosis can be made (see, e.g., Brown & Ron, 2002; Brown & Trimble, 2000). In a small proportion of cases, symptoms are physiologically impossible (e.g., triple vision); in others, a somatoform or dissociative diagnosis may be suspected because of the variability or medical implausibility of the symptom in question. In cases of conversion aphonia, for example, the patient may have a preserved ability to cough (requiring intact vocal chord function) despite being unable to speak or whisper. Similarly, unexplained sensory loss is often inconsistent with actual patterns of sensory innervation (e.g., as in “glove” and “stocking” anesthesia in the absence of neuropathy). 

Although the validity of making a firm diagnosis of unexplained illness has been questioned over the years (e.g., Slater & Glitherco, 1965), recent research by Crimlisk et al. (1998) has shown that very few symptoms are actually misdiagnosed as unexplained if appropriate investigations have been carried out. Nevertheless, where doubt exists, diagnoses should remain tentative until careful follow-up has unequivocally ruled out a physical explanation (Brown & Ron, 2002). 


Although a large proportion of unexplained symptoms remit spontaneously, many continue to cause distress and disability over time (e.g., Stone, Sharpe, Rothwell, & Warlow, 2003). Several different treatment options are available, including short-term group therapy (e.g., Kashner, Rost, Cohen, Anderson, & Smith, 1995), cognitive-behavioral therapy (CBT; e.g., Chalder, 2001) and psychodynamic psychotherapy (e.g., Temple, 2001). Research suggests that each of these treatments can be used successfully with certain cases of somatoform illness (Blanchard & Scharff, 2002; Kashner et al., 1995; Kroenke & Swindle, 2000). Treating these conditions can be notoriously difficult, however, and many unexplained symptoms remain resistant to treatment. This is particularly true when symptoms are chronic or occur in the absence of obvious psychopathology (Brown & Ron, 2002). In such cases, somatoform illness is often managed by limiting patients' exposure to unnecessary investigations and interventions; encouraging them (and their physicians) to accept a psychological interpretation of their symptoms; and offering symptomatic treatments such as pain management, physiotherapy, and occupational therapy (e.g., Brown & Ron, 2002; G. R. Smith, Rost, & Kashner, 1995). It is noteworthy that the management approach in such cases is the same regardless of the assumed etiology of the symptoms. 

Theoretical Concepts

Recent attempts to explain the pathogenesis of medically unexplained symptoms can be grouped according to three explanatory themes that are based on the concepts ofdissociation, conversion, and somatization. 


Probably the earliest systematic account of medically unexplained illness—originally couched in the language of hysteria—was Janet's dissociation theory (e.g., Janet, 1889, 1907; for theoretical origins, see Charcot, 1889). According to this account, “hysterical” individuals experience a spontaneous narrowing of attention when exposed to traumatic events. This narrowing of attention leads to the development of unexplained symptoms via two basic mechanisms. On the one hand, attentional narrowing limits the number of sensory channels that can be attended to simultaneously. Over time, the individual may develop a habitual tendency to concentrate on some sensory channels at the expense of others, leading to the loss of deliberate attentional control over neglected channels. As a result, the individual is rendered anesthetic for any information in the previously unattended modality; nevertheless, the information in this channel is still processed outside conscious awareness (i.e., in a dissociated fashion). Janet (1907) cited this mechanism as an account of unexplained sensory loss, which at that time was regarded as a cardinal symptom of hysteria. 

Other symptoms reflect the activation of memories that have become dissociated from the main body of knowledge representing the patient's personality. By this view, attentional narrowing creates a paucity of information within consciousness that prevents new memories from being integrated with existing aspects of personal knowledge. Because of the lack of competing information, however, these memories may take on a compelling subjective quality that causes them to be misinterpreted as perceptions rather than recollections. Moreover, in the absence of integration, the patient has little or no control over the activation of these memories, which may be triggered by related events in the internal or external environment. According to Janet (1907), this process is the same as that operating when comparable phenomena are created using hypnotic suggestion. Indeed, Janet (1907) argued that both suggestion and hypnosis are essentially pathological phenomena, possible only in those individuals with the mental weakness characteristic of hysteria. 

Although a century old, the theoretical analysis offered by Janet (1889, 1907) continues to influence nosology, theory, research, and clinical practice concerning the somatoform and dissociative disorders. Ludwig (1972), for example, argued that the primary pathophysiological mechanism involved in the creation and maintenance of unexplained symptoms is an attentional dysfunction, resulting from an increase in the corticofugal inhibition of afferent stimulation (cf. Whitlock, 1967). This process of inhibition generates a dissociation between attention and sources of afferent stimulation, preventing the integration of sensory information within subjective awareness. In many respects, this view is a modern-day manifestation of Janet's (1907) argument that hysterical phenomena arise from functional dissociations related to a deficit in attention. 

Evidence for afferent inhibition in unexplained illness was obtained in an earlier electrophysiological study by Hernández-Peón, Chávez-Ibarra, and Aguilar-Figueroa (1963). In that study, somatosensory evoked responses were found to be lower in the affected leg of a patient with unexplained hemianesthesia when compared with those evoked from the patient's unaffected leg. Other studies have failed to find this effect, however (Alajouanine, Scherrer, Barbizet, Calvet, & Verley, 1958; Behrman & Levy, 1970; Halliday, 1968). Indeed, normal somatosensory evoked potentials are often regarded as positive support for a diagnosis of unexplained sensory loss. More recent electrophysiological studies appear to suggest that conversion disorder is associated with normal early evoked potentials but a disruption in the later P300 component (e.g., Fukudu et al., 1996; Lorenz, Kunze, & Bromm, 1998). These findings seem to indicate that conversion disorder involves the selective attentional gating of processed information at a late stage in the processing chain, prior to the generation of conscious awareness (Sierra & Berrios, 1999). This is consistent with Ludwig's (1972) assertion that the inhibitory process in unexplained illness occurs after cortical responses are evoked, at a level beyond the primary receptor cortex. 

Further evidence for Ludwig's (1972) reinterpretation of the dissociation model has been obtained in cognitive studies that seem to confirm the presence of an attentional deficit in individuals with unexplained illness. This deficit appears to extend across a range of attentional measures, with patients showing decrements on high-level tasks assessing vigilance, habituation, cognitive flexibility, set shifting, and mental transformation (Bendefeldt, Miller, & Ludwig, 1976; Hernández-Peón et al., 1963; Horvath, Friedman, & Meares, 1980). Currently, however, the precise nature of this attentional deficit remains unclear; further research using standardized measures of attention, as well as controls for motivation and anxiety, is required (Lader, 1982). 

According to Janet (1889, 1907), the fragmentation of memory associated with unexplained symptoms is triggered by the occurrence of traumatic events, which were regarded as having a destabilizing influence on mental processing more generally. Consistent with this, there is good evidence to suggest that traumatized individuals have more dissociative experiences and more instances of medically unexplained symptoms than those who have not experienced trauma. Childhood sexual, physical, and emotional abuse have all been linked to unexplained illness (e.g., Badura, Reiter, Altmaier, Rhomberg, & Elas, 1997; Brown, Schrag, & Trimble, in press; Kinzl, Traweger, & Biebl, 1995; Nijenhuis, Spinhoven, van Dyck, van der Hart, & Vanderlinden, 1998; Pribor, Yutzi, Dean, & Wetzel, 1993; Reilly, Baker, Rhodes, & Salmon, 1999; Roelofs, Keijsers, Hoogduin, Näring, & Moene, 2002; Walker et al., 1988; for a review, see Brown, in press), and many somatoform symptoms develop following exposure to acute stressors (Binzer, Andersen, & Kullgren, 1997), such as recent loss (Van Ommeren et al., 2001), relationship difficulties (Craig, 2001), and exposure to dead bodies (Labbate, Cardeña, Dimitreva, Roy, & Engel, 1998). Nevertheless, a number of commentators have questioned the necessity of identifying psychosocial precipitants in making a firm somatoform or dissociative diagnosis (e.g., American Psychiatric Association, 1994; Merskey, 1979; Ron, 1994; Wessely, 2001; cf. World Health Organization, 1992). Although clear precipitants are often found, they are absent in many cases (Ron, 1994; Wessely, 2001). 

Although the similarities between Janet's (1889, 1907) original model and more recent accounts of medically unexplained symptoms (e.g., Kihlstrom, 1992; Ludwig, 1972; Whitlock, 1967) are considerable, there is one fundamental difference. According to Janet (1907), dissociation is an abnormal process provoked by stress in the constitutionally weak mind of the hysterical patient; as such, only those individuals with such a psychological weakness will experience dissociation and hence unexplained symptoms. In contrast, the majority of contemporary models adopt the view that dissociation is a normal psychological process, used by the organism as a defense mechanism in the face of trauma (see, e.g., Ludwig, 1972; Whitlock, 1967); unexplained symptoms arise when this normally adaptive process is overgeneralized to a point at which it becomes maladaptive. By this view, dissociation is a coping response available to all individuals, although the degree to which it is used varies from person to person. 

Hilgard's (1977) neodissociation theory, itself a partial reworking of Janet's (1889, 1907) original model, has been particularly influential in its support of the notion that dissociation is an adaptive mechanism that can become pathological if overused. Unlike other models, Hilgard's model is noteworthy in its assertion that dissociation is actually a fundamental aspect of cognitive processing in general, rather than an unusual response elicited by traumatic events. According to Hilgard, the majority of behavior is controlled by a set of functionally autonomous but interconnected cognitive control systems (schemata) specialized for the execution of particular behavioral acts. These control systems are hierarchically organized beneath an executive ego, which selects the cognitive control systems required at any given moment. Once selected, these systems are able to function without ongoing input from the executive ego. This “dissociation” between the executive ego and the systems responsible for action control allows well-learned behaviors to be performed effortlessly, concurrently, and outside awareness. 

Although originally developed as a general account of the mechanisms underlying behavioral control, neodissociation theory has been particularly influential as an account of hypnotic behavior. According to the theory, hypnosis represents one situation in which the communication between individual control systems and the executive ego is subject to change. Following suggestions aimed at inhibiting the executive, a vertical dissociation can be created within the ego forming two separate executive “streams,” one of which is concealed beneath an amnesic barrier. This unconscious part of the executive is able to communicate with targeted subsystems and therefore control behavior as normal, although this is not represented in conscious awareness. As the conscious part of the executive has no direct access to the systems controlling the suggested behavior, it is perceived as occurring involuntarily. More recently, Bowers (1992; see also Woody & Bowers, 1994) has suggested that hypnotic phenomena result from a horizontal dissociation between the executive and low-level control systems. According to this “dissociated control” theory, hypnosis serves to inhibit the executive altogether, allowing low-level control systems to be automatically activated by the words of the hypnotist. By this view, hypnotic involuntariness reflects the fact that hypnotic phenomena are generated without executive involvement (i.e., automatically). 

Echoing the earlier work of Janet (1889), Kihlstrom (1992) has applied neodissociation theory to the unexplained neurological symptoms characteristic of the dissociative and conversion disorders, citing the same kinds of functional dissociations as those outlined by Hilgard (1977) for hypnotic phenomena. (For other accounts that have endorsed the idea of common mechanisms for hypnotic and unexplained medical phenomena, see Bryant & McConkey, 1999; Ludwig, 1972; Oakley, 1999; Sackeim, Nordlie, & Gur, 1979; Whitlock, 1967.) The symptoms of unexplained amnesia, for example, would be linked to a dissociation that prevents the systems responsible for encoding and storing memorial information from communicating with the executive (see Brown, 2002, in press). Despite this lack of communication, memorial systems themselves remain intact and continue to function as normal. Compelling support for this hypothesis comes from an elegant study by Kuyk, Spinhoven, and van Dyck (1999). Kuyk et al. (1999) compared a group of patients reporting postictal amnesia following nonepileptic seizures and a group reporting similar memory loss following generalized epileptic events. All participants were hypnotized and given suggestions designed to facilitate the recovery of material encoded during the ictal period. Using a free-recall paradigm, 85% of the nonepileptic seizure patients recalled events occurring during the ictus that could be corroborated by independent observations. None of the epilepsy patients recalled ictal material. This study clearly demonstrates that the nonepileptic attack patients had encoded information during the ictal period but were unable to recall it because of a retrieval deficit; in this sense, the “forgotten” material could be regarded as dissociated from conscious awareness. The epilepsy patients, in contrast, experienced a disruption in encoding during the seizure and therefore had no ictal information in memory to retrieve. 

Further support for this account comes from studies demonstrating that individuals experiencing unexplained blindness exhibit intact visual perception in the absence of visual experience (e.g., Bryant & McConkey, 1989; Sackeim et al., 1979; for related examples, see Kihlstrom, 1992; cf. Kirsch & Lynn, 1998). A similar phenomenon is also observed in hypnosis, with participants displaying intact processing of information that is otherwise unavailable to awareness because of suggestion (so-called implicit processing; see, e.g., David, Brown, Pojoga, & David, 2000; Kihlstrom, 1987, 1992). Other findings also lend support to the notion that medically unexplained symptoms involve similar mechanisms to those underlying hypnotic phenomena. Many unexplained symptoms (e.g., amnesia, sensory loss, paralysis) have a hypnotic counterpart, and these conditions are often accompanied by high levels of hypnotic suggestibility (Bendefeldt et al., 1976; Janet, 1907; Ludwig, 1972; Roelofs, Hoogduin, et al., 2002). There is also good evidence to suggest that the occurrence of both unexplained symptoms and suggested phenomena are subject to similar moderating factors, such as belief, expectation, and attentional focus (e.g., Kirsch, 1985; McConkey, 2001; Salkovskis, 1989; Tellegen & Atkinson, 1974). Furthermore, recent functional neuroimaging studies provide some support for the idea that suggested and unexplained medical phenomena involve similar cerebral mechanisms. In a positron emission tomography (PET) study of unexplained paralysis, Marshall, Halligan, Fink, Wade, and Frackowiak (1997) found increased activation in the right orbitofrontal and anterior cingulate cortices, as well as an absence of activity in the right primary motor cortex, when an attempt to move the paralyzed left leg was made. A similar pattern of activation was found in a case of hypnotic left leg paralysis in a PET study conducted by Halligan, Athwal, Oakley, and Frackowiak (2000). These findings have been taken as evidence for the idea that unexplained paralysis (both hypnotic and nonhypnotic) involves the inhibition of primary motor activity by areas in the prefrontal cortex (Halligan et al., 2000; Marshall et al., 1997). 

Although its influence remains ubiquitous within contemporary theorizing, a number of recent commentators have raised doubts about the explanatory significance of the dissociation construct (e.g., Cardeña, 1994; Frankel, 1994). Frankel (1994), for example, has argued that the concept of dissociation has been reified and overextended within clinical practice to the point where it has only descriptive and not mechanistic significance. According to Frankel, any symptom that appears to involve some loss of control or experience has been described by clinicians as dissociative, despite there being no substantive grounds to assume that dissociative mechanisms (in the Janetian sense) are responsible for the clinical presentation in such cases (see also Brown, 2002; Cardeña, 1994). Depersonalization and derealization, for example, are typically regarded as dissociative phenomena (see, e.g., American Psychiatric Association, 1994), despite the fact that neither involves any profound alternation in perception, memory, or personal identity (World Health Organization, 1992). Indeed, recent theoretical developments indicate that these phenomena involve entirely different pathogenic mechanisms from those underlying medically unexplained symptoms (Holmes et al., 2004; Sierra & Berrios, 1998). This could explain why studies using measures such as the Dissociative Experiences Scale (Bernstein & Putnam, 1986) have often failed to find elevated levels of dissociativity in patients with unexplained symptoms (e.g., Alper et al., 1997; Pribor et al., 1993; Roelofs, Keijsers, et al., 2002; for a review, see Brown, in press). It is also possible to draw a meaningful link between somatoform symptoms and hypnotic phenomena without relying on the dissociation hypothesis (e.g., Oakley, 1999). 


The concept of conversion was introduced by Breuer and Freud (1893–1895/1991) as an extension to Janet's (1889, 1907) dissociative framework. According to Breuer and Freud, the brain often attempts to regulate the conscious experience of negative affect by unconsciously suppressing (or repressing) the conscious recall of memories associated with personal trauma. This process of repression can provide the basis for the creation of dissociated knowledge structures that are kept outside awareness by an amnesic barrier. Although the repression and dissociation of traumatic memories may protect the individual from potentially overwhelming negative affect, this process prevents the neural energy associated with that affect from being discharged in the usual fashion. As this must occur if the energetic balance of the brain is to be preserved, the negative affect is “converted” into a somatic symptom that either was present at the time of the original trauma or is some symbolic representation of it. In this way, the individual is able to express psychological distress without consciously acknowledging the psychic conflict responsible for it. As such, unexplained symptoms can be viewed as serving a defensive function, with the reduction of anxiety representing the primary gain from developing symptoms; in contrast, any additional advantage associated with being ill (e.g., sympathy, attention, avoidance of work) represents the secondary gain from symptoms. The appeal of the conversion concept is well illustrated in clinical case examples, such as that of the doctoral student who developed right-arm paralysis shortly before the deadline for his dissertation.1  According to the conversion model, the student's paralysis provides a means of avoiding his anxiety about completing, or failing to complete, his dissertation in a way that is both emotionally palatable and socially acceptable. Most clinicians with experience of unexplained illness can describe cases of this nature, which probably accounts for the continuing popularity of the conversion concept within clinical circles. 

This emphasis on the concept of psychological defense serves to differentiate the work of Breuer and Freud (1893–1895/1991) from the simple dissociation model described by Janet (1889, 1907). According to this view, unexplained symptoms reflect an unconscious emotional conflict that must be resolved if the patient is to experience relief from symptoms. To achieve this, the underlying conflict must be made conscious, and its accompanying affect expressed appropriately (Temple, 2001). In earlier versions of the theory (e.g., Breuer & Freud, 1893–1895/1991), the process of conversion was viewed as the product of early sexual abuse. This view was subsequently modified by Freud, who argued that patients' memories of abuse actually reflect an unresolved Oedipus complex driven by the repression of unacceptable sexual fantasies concerning the opposite-sex parent (see Temple, 2001). In contrast, later analytic theorists have placed greater explanatory emphasis on patients' internal representations of significant others and their influence in subsequent relationships (e.g., Temple, 2001). 

Although Breuer and Freud's (1893–1895/1991) views concerning psychic energies have long since been abandoned, the influence of these authors remains pervasive to this day. The notion that hysterical phenomena are created by the conversion of psychological distress into physical symptoms is widely endorsed, particularly within medical circles. Like Janet's (1889, 1907) model, such a view is supported by studies showing that many patients with unexplained illness have experienced extreme stress or trauma (often sexual) prior to the development of symptoms (see Brown, in press). In cases in which the patient reports no significant psychological problems or early trauma, it is assumed that the underlying conflict and its associated affect have been repressed out of consciousness. The apparent lack of concern displayed by many such patients, so-calledla belle indifférence, is also regarded as being consistent with the conversion concept (Iezzi & Adams, 1993). By this view, one would not anticipate significant psychopathology in cases of unexplained illness if the patient were successfully converting psychological distress into somatic symptoms. 

In addition, research concerning the alexithymia construct (Sifneos, 1973) has offered some support for the assertion that conversion occurs as a means of discharging unexpressed emotion. A number of studies have demonstrated that individuals who are less able to identify and report on their emotional states (i.e., alexithymia) are significantly more likely to exhibit unexplained medical phenomena than those who do not display this tendency (e.g., Shipko, 1982). Other concepts introduced by Breuer and Freud (1893–1895/1991) have been adopted by theorists working outside a psychoanalytic framework. In particular, the concepts of primary and secondary gain are consistent with a behavioral approach that emphasizes the role of reinforcement in the creation and maintenance of unexplained symptoms (Kellner, 1986). Some studies suggest that clear gains are associated with many cases of unexplained illness (e.g., Raskin, Talbott, & Myerson, 1966). 

Despite its continuing popularity, there are significant problems with the approach to unexplained illness described by Breuer and Freud (1893–1895/1991). In particular, there are doubts surrounding the assumption that symptoms are necessarily the product of an unresolved and unconscious psychological conflict (e.g., Ron, 1994; Wessely, 2001). Although obvious conflicts or psychosocial precipitants are present in a proportion of cases, they are absent in many others (for a review, see Brown, in press). According to the psychoanalytic approach, such patients are unable to describe the conflict underlying their symptoms because repression has rendered it unconscious; the absence of apparent conflict is taken as evidence for the operation of repression. Such circular reasoning clearly raises doubts concerning the testability of the conversion model. The problem is compounded by the assumption that unexplained symptoms are a symbol of the underlying conflict, rather than a direct expression of it. Taken together, these assumptions allow for an ad hoc approach to the formulation and treatment of medically unexplained illness that may be both misleading and potentially damaging. In particular, relentlessly probing for forgotten material, especially concerning sexual victimization, carries a risk of creating compelling but false memories of abuse (see, e.g., Lindsay & Read, 1994). 

Research has also raised doubts concerning the validity of la belle indifférence as a diagnostic indicator for the presence of medically unexplained illness. Many patients with unexplained symptoms are acutely upset about their problems (e.g., Lader & Sartorius, 1968), for example, and an apparent indifference to symptoms is also common in general medical illness (Brown, in press; Gould, Miller, Goldberg, & Benson, 1986; Raskin et al., 1966). Other studies have questioned the relationship between alexithymia and unexplained illness (e.g., Cohen, Auld, & Brooker, 1994), with evidence indicating that the measurement of alexithymia is confounded by psychopathology in general (e.g., Bach, Bach, Böehmer, & Nutzinger, 1994). Research also suggests that significant secondary gains are absent in many instances of medically unexplained illness and are no more common than in general medical illness (Brown, in press). 


Dissociation and conversion represent the main conceptual precursors to contemporary accounts of unexplained illness based on the concept of somatization. Lipowski (1968) has defined somatization as the tendency to experience or express psychological distress as the symptoms of physical illness. Somatization models place less explanatory emphasis on the specific mechanisms of unexplained illness, concentrating instead on the processes underlying normal somatic perception and the biopsychosocial context surrounding the experience of physical and mental illness (for reviews, see, e.g., Kellner, 1986, 1990; Kirmayer & Robbins, 1991a; Lipowski, 1988). Research inspired by the somatization concept has identified a number of factors that may be important in the generation and maintenance of medically unexplained symptoms. A recent model of the interaction between several of these factors has been described by Kirmayer and Taillefer (1997; see Figure 1). According to Kirmayer and Taillefer, illness, emotional arousal, or everyday physiological processes produce bodily sensations that capture attention to varying degrees. In some instances, these sensations may be interpreted as indicators of disease (i.e., symptoms), an attribution that may serve to generate illness worry, catastrophizing, and demoralization. As a result, individuals may adopt the “sick role” by pursuing assessment and treatment for their putative condition, thereby exposing themselves to social forces (e.g., the reactions of others, the media, etc.) that may reinforce their illness behavior and experience. Although such a process is a normal response to the signs of illness, in some cases it may reach disabling proportions, depending on the nature of the individual and the sociocultural context in which he or she is embedded. This process can be moderated by a number of contextual and dispositional factors, including previous illness experience; the response of significant others; illness worry; and individual differences in personality, attentional set, coping behavior, and autonomic reactivity (for evidence, see Kirmayer & Robbins, 1991a; Kirmayer & Taillefer, 1997; Robbins & Kirmayer, 1991; the empirical literature in this area is discussed below). 

The model provided by Kirmayer and Taillefer (1997) offers a useful scheme for understanding the relationship between the different factors underlying somatization. It is particularly appealing because it provides a model of normal illness behavior that can be generalized to more maladaptive circumstances. The model also provides a useful account of the evolution of unexplained symptoms through its emphasis on the temporally extended aspects of somatization. Moreover, it is consistent with the available empirical evidence concerning the risk factors for medically unexplained illness. 

Despite its obvious appeal, however, there are a number of problems with the somatization concept. In particular, it concentrates primarily on the factors that moderate the occurrence of medically unexplained symptoms and only speaks in very general terms about the psychological mechanisms involved in their mediation. Such a gap in the model is problematic because it obscures potentially important etiological differences between different types of somatization (Kirmayer & Robbins, 1991a, 1991b). According to Kirmayer and Robbins (1991a, 1991b), a distinction should be drawn among unexplained physical symptoms that are (a) the physiological components of anxiety or depressive psychopathology (presenting somatization), (b) normal bodily sensations or minor pathological events that are misinterpreted as signs of serious illness (hypochondriacal somatization), and (c) subjectively compelling symptoms that cannot be attributed to either physical or psychiatric illness or hypochondriasis (functional somatization). Although there is substantial comorbidity among these different forms of somatization, they can be differentiated both conceptually and empirically (Kirmayer & Robbins, 1991a, 1991b). Only the concept of functional somatization is of relevance in this context. 

Also problematic is the assumption that medically unexplained symptoms are the result of identifiable physiological processes that are exacerbated by social, cognitive, and perceptual factors. Although this may be appropriate in some cases, particularly those involving presenting and hypochondriacal somatization, it does not provide a compelling account of the processes involved in the generation of many medically unexplained complaints. It is, for example, difficult to conceive of unidentified physical pathology that could be responsible for the profound disruptions of function observed in unexplained neurological illness. 

A further problem is the assumption, intrinsic also to the concepts of dissociation and conversion, that medically unexplained symptoms must be the product of psychological distress. Although such an assertion may be appropriate in the case of presenting and hypochondriacal somatizers, this is not well established empirically in the case of individuals with somatoform (i.e., functional) symptoms (Ron, 1994; Wessely, 2001). 

Interim Summary

The dissociation, conversion, and somatization models have added much to the field's understanding of medically unexplained illness. Each offers a different perspective on the pathogenesis of these phenomena, and each is able to accommodate different aspects of the available empirical evidence. The dissociation model and its derivatives have helped illuminate the basic processing mechanisms involved in the generation of unexplained symptoms. This perspective is supported by evidence from a number of cognitive, electrophysiological, and neuroimaging studies. The concept of conversion has extended the dissociation model by providing a more detailed account of the way in which emotional and motivational factors contribute to unexplained illness. The conversion model is supported by research demonstrating that traumatic events and other emotional disturbances are often associated with the development of unexplained symptoms. More recently, the somatization concept has placed current understanding of unexplained illness in the wider biopsychosocial context, providing a framework that elucidates the risk factors associated with this phenomenon. Converging evidence from several different sources supports this approach. 

Nevertheless, it is clear that there are problems with the concepts of dissociation, conversion, and somatization. Dissociation models have been widely criticized for lacking precision (e.g., Frankel, 1994), something that has generated confusion about the nature of dissociation as a psychological phenomenon. Although concepts such as attentional gating and top-down inhibition have advanced the argument, these are essentially descriptive labels for the processes they purport to explain. In contrast, the conversion model can be criticized for its circularity, something that has rendered many aspects of this approach unfalsifiable. Moreover, the fact that emotional and motivational factors appear to play a significant role in only a proportion of cases of unexplained illness casts doubts on the generality of the model (Ron, 1994; Wessely, 2001). The somatization approach can also be criticized for its reliance on the concept of emotional precipitation, and its descriptive nature prevents the generation of precise hypotheses. 

An Integrative Conceptual Model of Medically Unexplained Symptoms

The remainder of this article focuses on a novel account of unexplained illness that builds on the strengths of existing models while attempting to address some of their shortcomings. The model is integrative in that it accommodates the basic elements of the dissociation, conversion, and somatization concepts within a common explanatory framework. As such, it should be viewed as an extension of, rather than an alternative to, these approaches. The model has two principal aims. First, it aims to provide a precise, mechanistic account of how it is possible to experience compelling symptoms in the absence of underlying pathology. Thus, the proposed model explicitly assumes that genuine somatoform symptoms are subjectively “real” (i.e., similar to explained physical symptoms) to the individual reporting them; the model therefore has little to say about malingering, which is regarded as a separate phenomenon entirely. Second, the model aims to clarify how the development and maintenance of unexplained symptoms are moderated by different risk factors associated with the phenomenon. In achieving these aims, the model attempts to explain unexplained illness in a way that does not rely on the concept of psychosocial precipitation but accommodates its undoubted importance in many cases. 

Explanatory Domain

The model is intended as an account of functional somatization (Kirmayer & Robbins, 1991a, 1991b), encompassing all of the DSM-IV somatoform and dissociative disorders, with the exception of hypochondriasis, body dysmorphic disorder, depersonalization disorder, and dissociative identity disorder. Depersonalization disorder is regarded here as distinct from other dissociative phenomena, involving mechanisms that are qualitatively different to those described below (Holmes et al., 2004). Dissociative identity disorder, in contrast, may have certain mechanisms in common with functional somatization, although a comprehensive understanding of this condition is beyond the scope of this model. Other unexplained symptom syndromes, such as chronic fatigue, fibromyalgia, and irritable bowel syndrome, are also outside the remit of this account, as are unexplained symptoms with an observable physical component (e.g., phantom pregnancy). Finally, the model does not attempt to explain why some individuals with unexplained symptoms display disproportionate levels of abnormal illness behavior (Pilowsky, 1978; e.g., “doctor-shopping,” refusal to comply with treatment, demands for repeated investigation). 

Inevitably, there will be variation in the specific processes underlying different medically unexplained symptoms. In line with most previous theories in this domain, however, the model attempts to provide a general account of medically unexplained phenomena that can accommodate all members of this category. Such an approach is necessitated by the fact that many individuals suffering from unexplained illness are polysymptomatic, often displaying a combination of neurological and nonneurological symptoms. This approach is not without its critics. Kihlstrom (1992), for example, has argued that the mechanisms of somatoform complaints, including somatization disorder, are fundamentally different from those involved in the generation of unexplained neurological symptoms (conversion disorder). However, apart from differences in symptom numbers (including the presence of unexplained symptoms in other systems) and illness chronicity, the available data indicate that there is very little that distinguishes between conversion and somatization disorders (Ron, 2001). Although such differences are important, they do not provide sufficient grounds to assume a priori that fundamentally different mechanisms are operating in these conditions. 

It has also been argued that, unlike conversion disorder, somatization disorder is simply an extreme form of abnormal illness behavior, whereby symptom reports are used as a means of emotional expression and social control (Kihlstrom & Canter Kihlstrom, 1999). By this view, distortions in somatic perception (i.e., the experience of unexplained symptoms) may not be part of somatization disorder at all. It is unclear, however, why the symptom reports of patients with conversion disorder should be considered any more reliable than those of patients with multiple unexplained symptoms. In the absence of data to the contrary, it seems appropriate to pursue an explanation of unexplained illness that can accommodate both neurological and nonneurological symptoms. 

Although the proposed model assumes that there are common mechanisms underlying different unexplained symptoms, it is clear that a range of factors moderate the occurrence of this phenomenon and that the relative contribution of these factors will vary from case to case. As such, exactly the same processes may not be involved in the creation and maintenance of otherwise apparently similar cases of unexplained illness. Moreover, if one regards the development of medically unexplained symptoms as a temporally extended process, it is possible that a variety of different moderating factors will be implicated over the course of individual cases also. In the next section, I describe the psychological character of unexplained symptoms and the processes involved in their generation. The factors moderating the creation and maintenance of these complaints are then addressed. 

Attention, Consciousness, and Control

The central tenet of the proposed model is that unexplained symptoms can be understood by reference to research and theory from mainstream cognitive psychology. In particular, the model is informed by research concerning the nature of different attentional mechanisms in the cognitive system and their role in (a) shaping the contents of consciousness and (b) controlling thought and action. 

At any given time, the cognitive system is inundated with information that has the potential to influence thought and behavior. The selection of relevant aspects of this information for further processing and the control of action is generally regarded as one of the most important tasks of the attentional system (Styles, 1997). Much of the research in this area has focused on the so-called locus of selection, that is, the point in the processing chain where attentional selection occurs. Evidence suggests that complex perceptual and semantic analyses are often performed prior to the selection process (although see Lavie, 1995) through a parallel spread of activation in associative networks triggered directly by the receipt of sense data (e.g., Kosslyn, 1996; Marcel, 1983a, 1983b; Neely, 1977; Posner & Snyder, 1975; Rumelhart, McClelland, & the PDP Research Group, 1986; for reviews, see Styles, 1997, and Velmans, 2000). This spread of activation can be regarded as an inferential process (Sloman, 1996; Marcel, 1983a) whereby existing knowledge assists in the identification of relevant material for priority processing and hence attentional selection. The selection process itself is thought to integrate the most active information in sensory, perceptual, and memorial systems, producing multimodal representations that allow for the control of routine behavior as well as further processing by attentional mechanisms at higher levels of the system (e.g., Allport, 1987; Coltheart, 1980; Crick & Koch, 1990; Logan, 1988; Marcel, 1983b; Neuman, 1987; Norman & Shallice, 1986; Velmans, 1991, 2000); the representations produced by this process are thought to relate closely to the contents of conscious awareness (Allport, 1988; Coltheart, 1980; Crick & Koch, 1990; Marcel, 1983a; Velmans, 1991, 2000). 

A summary model of the cognitive system based on this research is presented in Figure 2. In this model, the parallel spread of activation in perceptual and memorial systems generates a number of perceptual hypotheses, each representing a possible interpretation of the stimulus world on the basis of previous experience (Marcel, 1983a). The most active perceptual hypothesis is then selected by a primary attentional system (PAS) and used to organize relevant sensory information into integrated multimodal perceptual units, or primary representations, that provide a working account of the environment for the control of action; in this model, perceptual awareness corresponds broadly to the content of these representations (cf. Allport, 1988; Coltheart, 1980; Crick & Koch, 1990; Marcel, 1983b; Neuman, 1987). The PAS is influenced by several factors, including the nature of the available sense data, the activation of competing and complementary representations in memory, the selection threshold of the perceptual hypotheses in question, and top-down input from high-level attention. 

Behavior is controlled via two basic routes in this model, which follows from the work of Norman and Shallice (1986). By this view, routine behaviors are controlled by a hierarchical system of procedural representations (schemata) specifying the attentional, cognitive, and motoric processes involved in executing well-learned actions. In the proposed model, schemata are activated automatically (in the sense used by Logan, 1988) by the construction of primary representations and are triggered when a threshold level of activation has been reached (in accordance with a competitive scheduling mechanism; see Norman & Shallice, 1986). This automatic activation of schemata by the products of PAS selection provides the system with a means of controlling cognition and action that is rapid, highly efficient, and consumes relatively few processing resources. Behaviors controlled in this way are experienced as occurring without conscious effort, whereas perception and cognition at this level are associated with a sense of intuition and self-evident validity. In situations in which the system does not possess the appropriate schemata, an additional means of control is required. In this model, novel actions are controlled by a secondary attentional system (SAS; equivalent to the supervisory attentional system in the Norman & Shallice, 1986, model) that controls action indirectly by biasing the relative activation levels of schemata via the PAS, in accordance with general purpose algorithms (for recent theoretical developments, see Burgess & Shallice, 1996; Shallice, 1988; Shallice & Burgess, 1996). The SAS is the province of self-regulatory processing, moderating cognition and action in the pursuit of high-level system goals (cf. Wells & Matthews, 1994). Processing controlled by the SAS draws on a limited pool of resources, is perceived as mentally demanding, and is associated with a sense of conscious volition and self-awareness (cf. Norman & Shallice, 1986; Wells & Matthews, 1994). 

Medically Unexplained Symptoms as Alterations in Perception and Control

The cognitive research and theory outlined above has a number of important implications for understanding medically unexplained illness. First, it suggests that the management of behavior is governed by systems that operate without direct volitional control. This is important in that it allows for functional dissociations between the experience of volition and the control of thought and action. Second, it identifies subjective awareness as a construction or interpretation of the world rather than an unqualified account of objective reality (for a similar view related to normal somatic perception, see Leventhal, 1986). This underscores the fact that on some occasions there may be no direct mapping between sensory stimulation and the contents of subjective experience. Indeed, there are numerous examples of how experience is often more consistent with what people know, believe, and want to believe about the world than with sensory data itself (e.g., hallucinations, certain illusions, misperceptions, “perceptual set” phenomena, placebo effects, hypnotic phenomena, defense mechanisms; see, e.g., Bruner & Postman, 1949; Gregory, 1970; Hilgard, 1977; Vaillant, 1993; Wall, 1993). Phenomena of this sort are important because they demonstrate how the generation of perceptual experience can be overdetermined by prior information (e.g., memories) in the cognitive system (Gregory, 1970; Leventhal, 1986). In the proposed model, medically unexplained symptoms represent a compelling example of this phenomenon. 

By this view, unexplained symptoms arise when the chronic activation of stored representations in memory causes the PAS to select inappropriate information during attentional selection and the automatic control of processing. The result is a misinterpretation of the sensory world that is experienced as subjectively valid (i.e., “real” symptoms) because the individual does not have introspective access to the inferences made during the creation of experience and control of action. The nature of the resulting symptom reflects the kind of information involved in this process. Certain symptoms, such as those characterized by alterations in experience (e.g., pain, feelings of discomfort, nausea, pseudohallucinations), arise when inappropriate (i.e., inconsistent with information provided by the senses) perceptual hypotheses are selected during the creation of primary representations by the PAS. In contrast, symptoms characterized by an inability to control perception, cognition, or action arise when inappropriate cognitive schemata are automatically triggered by the process of PAS selection. 

In a case of unexplained sensory loss, for example, the relevant schema would be one instructing the PAS to inhibit attention either to a particular sensory modality (as in blindness or deafness), a part of the body (as in focal sensory loss), or an aspect of the external phenomenal world (as in tunnel vision). Similarly, dissociative amnesia would be generated by schemata directing the PAS to withdraw attention from certain types of memorial information or inhibiting the operation of certain memory retrieval routines. Schemata preventing the recall of all autobiographical information would lead to amnesia with an associated loss of personal identity, as observed in cases of dissociative fugue. In the case of certain unexplained motor symptoms, schemata specifying particular limb positions or body postures (as in fixed dystonia, gait disturbance, etc.) or the inhibition of movement altogether (as in paralysis) would be involved. In each case, this selection of inappropriate schemata is often driven by the creation of relevant primary representations by the PAS. In the case of unexplained paralysis, for example, the underlying schema may be triggered by primary representations consistent with the idea that limb movement is deficient in some way. 

Unlike other somatoform symptoms, unexplained seizures (so-called nonepileptic attacks) are intermittent time-limited phenomena that probably involve slightly different mechanisms to those outlined above. Whereas other symptoms involve the chronic activation and selection of perceptual or behavioral representations, unexplained seizures may involve the automatic activation of a schema as an acute response to cues from the internal or external environment (Brown, 2002). The same mechanism may be responsible for other time-limited behavioral symptoms, such as pseudoballismus (unexplained violent movements of the limbs, akin to those seen in chorea). In cases involving repetitive movements that occur in a time-limited fashion (e.g., tremors), schemata may be subject to repeated reactivation through the operation of the SAS. This notion accords well with the common clinical experience that such movements are often considerably reduced when the patient is distracted. 

The model proposed here is clearly consistent with contemporary dissociation theory, which asserts that unexplained symptoms are generated preconsciously by an attentional gating mechanism operating late in the processing chain (e.g., Ludwig, 1972; Sierra & Berrios, 1999). It also captures Janet's (1889) original idea that unexplained symptoms reflect a distortion in awareness resulting from information that has become “stuck” in the cognitive system (so-called fixed ideas). The proposed model also assumes that a horizontal dissociation between different levels of processing in the cognitive system is central to all somatoform symptoms (cf. Hilgard, 1977; Janet, 1907; Kihlstrom, 1992; Oakley, 1999). Indeed, it is this assumption that allows the model to resolve the apparent paradox presented by unexplained symptoms characterized by an inability to control perception, cognition, or action despite repeated and honest attempts to exert control by the sufferer. In this model, the experience of volition is associated with the operation of the SAS, but the ultimate locus of cognitive and behavioral control is with the PAS and the schemata underlying routine action. When the patient attempts to control cognition or action, he or she is engaging the SAS; the attempt is unsuccessful because the locus of the patient's deficit is the chronic activation and selection of representations by the PAS. One important implication of this approach is that symptoms are generated by psychological mechanisms but are not produced deliberately. 

Unlike Janet's (1889, 1907) original dissociation theory, however, the proposed model rejects the idea that unexplained symptoms are the product of a purely pathological process. In the proposed account, symptoms result from a subtle disruption in processes that are fundamental to the everyday control of behavior and experience (cf. Hilgard, 1977; Kihlstrom, 1992); in this sense, unexplained symptoms are essentially “normal” psychological phenomena (cf. Halligan & David, 1999). This is clearly consistent with epidemiological evidence demonstrating the ubiquity of unexplained illness (Escobar et al., 1998; Fink et al., 1999; Kroenke et al., 1997) and the fact that many individuals experience disabling symptoms in the absence of other psychopathology (Wessely, 2001). 

The Origins of “Rogue Representations”

According to this approach, unexplained symptoms constitute an alteration in the body image generated by information in the cognitive system, rather than disturbances in the neural hardware itself (cf. Leventhal, 1986; Ramachandran & Hirstein, 1998). The term rogue representation is used here as a generic label for the inappropriate information that is selected by the PAS during this process. Broadly speaking, any kind of information within the cognitive system concerning the nature of physical symptoms can provide a template for the development of an unexplained complaint. Rogue representations can therefore be acquired from many different sources, including direct exposure to physical states in the self, indirect exposure to physical states in others, the sociocultural transmission of information about health and illness, and direct verbal suggestion. 

Exposure to physical states in the self

The fact that somatoform patients often have a history of physical illness (e.g., Crimlisk et al., 1998; Merskey & Buhrich, 1975; Slater & Glitherco, 1965; cf. Schrag, Brown, & Trimble, 2004) suggests that many rogue representations arise out of memory traces acquired during episodes of organic pathology. Similar traces are also established when symptoms result from minor pathological events, such as the loss of sensation in a limb following transient ischemia (Breuer & Freud, 1893–1895/1991; see also Sharpe & Bass, 1992). The physical components of emotional states, such as those associated with anxiety (e.g., shaking, palpitations, nausea, muscle tension, chest pain, dizziness, dysphagia, light-headedness, visual disturbance, tinnitus), also leave representations in memory that could provide the basis for the later development of unexplained symptoms. Also included in this category are the sensorimotor components of certain defensive reactions that can occur in response to traumatic events, including numbing and analgesia (Nijenhuis, Vanderlinden, & Spinhoven, 1998), freezing (Nijenhuis, Vanderlinden, & Spinhoven, 1998), the sham-death reflex, and the violent motor reaction (Kretschmer, 1926; Ludwig, 1972). The storage of representations pertaining to these events could account for some of the most common unexplained neurological symptoms, including sensory loss, paralyses, and certain types of nonepileptic seizure (Kretschmer, 1926; Ludwig, 1972; Nijenhuis, Vanderlinden, & Spinhoven, 1998). Arguably, traumatic experiences such as physical or sexual abuse provide one of the richest sources of material for the development of rogue representations, often encompassing organic damage, intense emotional arousal, and the defensive use of physical reactions within the same experience. 

Exposure to physical states in others

There is good evidence to suggest that rogue representations can be acquired indirectly through exposure to physical symptoms in others (e.g., Charcot, 1889; Fallik & Sigal, 1971; Van Ommeren et al., 2001). Many somatoform patients have been exposed to abnormal levels of illness in the family environment (Hotopf, Mayou, Wadsworth, & Wessely, 1999), and unexplained complaints seem to be more common in individuals who are exposed to physical illness on a regular basis (e.g., medical students; Woods, Natterson, & Silverman, 1966). The effect is also illustrated by the occurrence of so-called mass hysteria, when unexplained symptoms can spread through a community like a contagious physical condition (e.g., Van Ommeren et al., 2001). To account for these findings, the model assumes that observing symptoms in others creates memory traces that are functionally similar to those generated when the same symptoms are experienced in the self. This is consistent with evidence demonstrating that observing actions in others creates representations that are comparable with those involved in the production of the same act (e.g., Jeannerod, 1994; Prinz, 1997; Sebanz, Knoblich, & Prinz, 2003). 

Sociocultural transmission

It is also well established that unexplained symptoms can be shaped by sociocultural conceptions of illness, with many somatoform phenomena conforming to a popular conception of what constitutes a “genuine” symptom (e.g., Reynolds, 1869). Thus, a higher proportion of nonepileptic attacks involve collapse or prominent motor activity when compared with seizures resulting from epilepsy (Meierkord, Will, Fish, & Shorvon, 1991). Similarly, the experience of nonpsychotic parasitosis (the sensation of insects crawling beneath the skin) and burning hands and/or feet, are considerably more common in African and Asian cultures than in North America (American Psychiatric Association, 1994), demonstrating the role of local concerns about illness in the development of unexplained symptoms (see Isaac et al., 1995; Van Ommeren et al., 2001). Indeed, all cultures have socially sanctioned models of illness (the so-called sick role; Parsons, 1964) that are likely to shape the representations responsible for the development of unexplained symptoms. These views are generated and transmitted by many different sources, including the family, the medical profession, the media, the Internet, and society more generally (Fallik & Sigal, 1971; Stewart, 1990). 

The proposed model assumes that the creation and alteration of rogue representations through sociocultural transmission involve a similar process to that operating when symptoms develop after exposure to illness in others. In other words, exposure to general information about physical illness creates memory representations that are functionally equivalent to those encoded during episodes of illness in the self. This is consistent with research showing that executing an action and thinking about that action involve similar neurocognitive processes (e.g., Jeannerod, 1994; Prinz, 1997). The model also assumes that the activation levels of rogue representations are partly determined by the activation of any associated representations in the memory network (Marcel, 1983a, 1983b; Neely, 1977), including those representing conceptual and semantic information. As a result, symptoms tend to become more consistent with other information in the system over time (see also Kihlstrom, 1992; Leventhal, 1986), which could help account for the influence of beliefs in the development of unexplained symptoms. 

Verbal suggestion

The processes identified here as central to the development of unexplained symptoms have also been linked to the creation of hypnotic phenomena (Brown & Oakley, 2004; cf. Janet, 1907; Kihlstrom, 1992; Ludwig, 1972; Bryant & McConkey, 1999; Oakley, 1999). Assuming that similar processes are involved in these phenomena, it seems likely that some unexplained symptoms originate from rogue representations developed in response to direct verbal suggestions, akin to those encountered in the hypnotic setting. These could either be in the form of autosuggestions (e.g., thinking “I still won't be able to see when they take the bandages off”) or heterosuggestions from relevant others (e.g., being told “This may hurt” during a physical examination). This process may also overlap with that involved in phenomena such as the placebo effect; in this case, however, the result is the absence of genuine symptoms rather than the presence of unexplained ones, reflecting the activation of different representations in each case. 

Self-Focused Attention and the Development of Symptom Chronicity

According to this account, all people possess material about symptoms that could provide the basis for rogue representations and somatoform conditions. Although some people have much more material in this respect, it is unlikely that this alone could account for why there are such marked individual differences in the tendency to develop unexplained symptoms. How does the model account for these differences? 

The answer to this question lies in understanding how the initial selection of a rogue representation can lead to the development of a chronic problem. Evidence from a number of sources suggests that self-focused attention is an important factor in this respect. Robbins and Kirmayer (1991), for instance, have identified a body-focused attentional bias in patients with unexplained symptoms, and a link between self-focused attention and subjective symptom reports has been obtained in several nonclinical studies (e.g., Pennebaker & Brittingham, 1982; Wegner & Guiliano, 1980). A number of investigators, for example, have found that introspective individuals are more likely to experience somatic symptoms (e.g., Hansell & Mechanic, 1986; Pennebaker, 1982; Robbins & Kirmayer, 1986), and manipulations aimed at directing attention toward the self tend to increase physical symptom reports (e.g., Pennebaker & Brittingham, 1982; Wegner & Guiliano, 1980). Evidence also suggests that elevated symptom reports are associated with boring environments rather than stimulating ones, apparently because of the decreased attentional load in the former (e.g., Fillingim & Fine, 1986; Pennebaker & Lightner, 1980). In addition, there is strong evidence that self-focus is a feature of depression and anxiety (e.g., Hope & Heimberg, 1985; T. W. Smith & Greenberg, 1981; Wells & Matthews, 1994), which often occur alongside somatoform illness (Brown & Ron, 2002). Attending to the body has also been identified as an important factor in hypochondriacal somatization (Salkovskis & Warwick, 1986). 

In the proposed model, the recurrent redirection of attention onto symptoms by the SAS is the primary pathogenic factor in the development of symptom chronicity (see Figure 3). Allocating high-level attention to a symptom (e.g., checking to see if is still present) serves to augment the activation of the rogue representation and lowers the amount of activation required for it to be selected in future (i.e., the selection threshold). This is a relatively self-perpetuating process, facilitated by the automatic allocation of SAS-level attention onto environmental events that are consistent with currently selected representations (a form of confirmatory bias). If symptom-focused attention is maintained for a sufficient period, the activation levels of rogue representations may become high enough to ensure their continued selection over time. In such cases, the generation of an unexplained symptom is complete. 

According to this account, anything that serves to increase symptom-focused attention will contribute to the development and maintenance of medically unexplained symptoms. Included in this category are the misattribution of symptoms to physical illness, negative affect, illness worry and rumination, illness behavior, and certain personality factors. To this extent, the proposed model overlaps considerably with the model of somatization described by Kirmayer and Taillefer (1997; see also Chalder, 2001). The relationship between these different factors is shown in Figure 4. 

Misattribution of symptoms

It is well established that patients' experiences of symptoms, and their behavioral responses to them, are strongly influenced by how they interpret their symptomatology (e.g., Cioffi, 1991; Leventhal, 1986; Pennebaker, 1982; Skelton & Croyle, 1991). A number of studies have shown, for example, that symptoms perceived as a sign of serious physical pathology are more likely to attract attention than those that are attributed to a benign cause (e.g., Cioffi, 1991). For this reason, it is likely that misattributing somatoform symptoms to a serious medical cause—akin to the catastrophic misinterpretation of normal sensations found in hypochondriacal somatization—will contribute to their maintenance in many cases. 

There are two basic routes to symptom misattribution in the proposed model, reflecting the hierarchical nature of the cognitive system. On the one hand, misattribution may arise automatically during the creation of primary representations by PAS selection. Misinterpretations of this sort may be in the form of “negative automatic thoughts” associated with a subjective sense of spontaneity and inevitability (cf. Beck, 1976). On the other hand, misinterpretations may be the product of ruminative activity associated with the SAS (see Illness worry and rumination section; also Wells & Matthews, 1994). In both cases, however, the product is an escalation of negative affect and perceived disability, increased worry and rumination about illness, and potentially inappropriate illness behavior. In turn, these factors increase the likelihood of symptoms being misattributed to a physical illness, setting up a vicious cycle (see Figure 4). 

It is likely that several different factors can contribute to the misattribution of symptoms to a physical cause. Unusual, painful, or debilitating symptoms are likely to be associated with a greater degree of negative affect and are therefore more likely to be appraised as a threat warranting attention and self-regulatory action (Cioffi, 1991). Maladaptive beliefs concerning the nature of health and the meaning of symptoms are also likely to be important, particularly the beliefs that health is a state devoid of symptoms and that symptoms always indicate pathology, both of which are commonly found in patients with unexplained symptoms (Rief, Hiller, & Margraf, 1998). “Irrational” beliefs of this sort are likely to originate in the familial context, particularly where an individual has been exposed to high levels of illness (Hotopf et al., 1999) or disproportionate parental concern over benign physical symptoms (Benjamin & Eminson, 1992). In itself, catastrophic misinterpretation is not necessarily irrational, however. Indeed, attributing subjectively compelling symptoms to a serious physical cause could be regarded as a normal part of the sick role, particularly where certain symptoms are concerned (e.g., unexplained neurological symptoms). It may even be positively sanctioned by contact with health care professionals (Lipowski, 1988), who often collude with patients' attributions about their conditions (Salmon, Peters, & Stanley, 1999), refer for unnecessary investigations and physical treatments (paradoxically “to put the patient's mind at rest”), and even misdiagnose somatoform symptoms as physical problems (Kouyanou, Pither, Rabe-Hesketh, & Wessely, 1998). 

Negative emotional states

Negative affect is not only a product of somatoform symptoms and their attribution to physical illness but also a potential contributor to the development and maintenance of these conditions. As well as increasing the likelihood of catastrophic misinterpretation, ongoing negative emotional states (such as those associated with anxiety and depression) serve to increase symptom-focused attention (see Wells & Matthews, 1994) and may trigger or perpetuate both rumination and illness behavior. It is also possible that negative affect directly influences the actual encoding and storage of rogue representations. Such a concept could help explain the spontaneous development of unexplained symptoms in the face of acute stressors. It may be that extreme fear or anxiety leads to the chronic selection of rogue representations, rather like the acquisition of conditioned aversion to noxious stimuli. A second, and perhaps related, possibility might be that symptoms are acquired as the result of a spontaneous narrowing of attention occurring as a biological response to extreme anxiety or fear (Cardeña & Spiegel, 1993; Janet, 1907). It may be that this narrowing of attention serves to increase the activation level of rogue representations, thereby contributing to the development of symptom chronicity. 

Illness worry and rumination

Both negative affect and the process of symptom misattribution trigger illness worry and rumination, which may play a particularly important role in the development of symptom chronicity. Following Wells (2000), illness worry and rumination are defined as repetitive forms of self-regulatory activity by the SAS aimed at reducing the discrepancy between current and desired states related to the presence of physical symptoms. Subjectively, they are experienced as streams of conscious, typically negative mental activity concerning the nature, meaning, and implications of symptoms and appropriate courses of action to be taken in their presence. They are also associated with a state of self-focused attention and constant monitoring of the body for symptom-relevant information. In general, these forms of cognition contribute to the maintenance of unexplained symptoms by perpetuating the allocation of high-level attention to rogue representations (and other perceptual and memorial information that supports their activation) and by increasing negative affect, symptom misinterpretation and illness behavior. The degree of rumination and worry depends on a number of factors, including the level of associated affect, the nature of any symptom misinterpretation, and the information available to individuals about their problems. Certain beliefs may also perpetuate illness worry and rumination, particularly “positive” metabeliefs, which suggest that these processes are useful for ensuring appropriate vigilance, mental preparation, and illness behavior (Wells, 2000). By maintaining rumination and worry, however, such beliefs can have the paradoxical effect of perpetuating symptoms. 

Once an individual has begun to worry and ruminate about their symptoms, terminating these processes may prove difficult unless symptoms improve. Over time, this may lead to the development of a “cognitive-attentional syndrome” characterized by chronic rumination, worry, and symptom focus; maladaptive coping; and a reduction of central processing resources (Wells, 2000). This could account for the attentional deficit seen in some patients with unexplained symptoms (Bendefeldt et al., 1976; Horvath et al., 1980; Janet, 1907). Chronic rumination may also lead to the development of ruminative schemata that can be triggered by PAS selection alone, producing the same cognitive effects in the absence of SAS-level activity (cf. Wells, 2000). In this way, a previously held element of control over the processes involved in the generation and maintenance of symptoms is lost. 

Illness behavior

Depending on the way symptoms are interpreted, the individual may engage in several different types of illness behavior, including “checking,” medical consulting, information seeking, reassurance seeking, avoidance, and doctor shopping, each of which directs attention toward the body and symptoms. Continued attempts to check whether symptoms are present (e.g., by mentally and physically scanning the body) is likely to be particularly important in this respect (cf. Salkovskis, 1989; Salkovskis & Warwick, 1986), especially when behavior has become automatized over time and introspective access to the process has been reduced. Repeated help seeking may also be central in maintaining attention to symptoms, especially when casual instructions to keep an eye on symptoms explicitly endorse this process. In addition, attempts to seek reassurance through the acquisition of illness-related information (e.g., from the Internet) may serve to increase anxiety and provides ambiguous material that could serve as a basis for rumination, catastrophic misinterpretation, and the development of new rogue representations. Similarly, doctor shopping may lead to patients receiving conflicting advice and diagnoses, increasing the likelihood of rumination and misinterpretation and reducing patients' faith in the reliability of the health care system. In contrast, avoidance of illness-related material serves to prevent the disconfirmation of any negative beliefs about symptoms, as does avoidance of activities that are perceived as having the potential to exacerbate symptoms (Chalder, 2001); the latter may be particularly problematic in that it can lead to physical deconditioning and health problems in its own right. 

Personality factors

The proposed model is consistent with evidence demonstrating that individuals with certain personality features are particularly vulnerable to developing unexplained symptoms. Trait measures of negative affectivity (Watson & Clark, 1984), for example, consistently correlate with subjective symptom reports but show little or no correlation with objective health markers. Studies show that high negative affectivity is associated with introspection, self-focus, worry, symptom misinterpretation, and illness behavior (Watson & Clark, 1984; Watson & Pennebaker, 1989), each of which has been linked to the development of unexplained symptom chronicity in this account. In addition, individuals high in hypnotic susceptibility may be particularly sensitive to the automatic activation of cognitive and behavioral schemata by internal and external events (Brown & Oakley, 2004), which could increase their vulnerability to developing somatoform symptoms. 

Body-Focused Attention as Psychological Defense: Conversion Reconsidered

The processes outlined thus far provide an explanation of how unexplained symptoms can develop in the absence of trauma. Why is a history of trauma so common in patients with these conditions? The model accounts for this by assuming that traumatic events such as physical, sexual, and emotional abuse often lead to the use of body-focused attention as a means of avoiding the affect and cognitive activity associated with experiences of this sort. 

Under normal circumstances, the individual is able to manage negative affect via self-regulatory processing and goal-oriented action (Wells, 2000; Wells & Matthews, 1994). The high level of negative affect associated with trauma, however, increases the difficulty of effective self-regulation, particularly when previous assumptions about the self, others, and the world have been violated by the nature of the traumatic event (cf. Horowitz, 1986; Janoff-Bulman, 1992). Self-regulation may also be impeded by appraisals indicating that externally oriented actions (e.g., resisting, attempting to escape) could prove futile or even lead to retaliatory revictimization. Under these circumstances, the system is forced to rely almost exclusively on internal processes for the regulation of traumatic affect. According to the proposed model, one way of reducing this potentially overwhelming affect is to divert SAS resources from self-regulatory processing and onto the body. Such a strategy could be used both during an episode of victimization and/or afterward, when a motivation to avoid trauma-related material and its associated affect remains; this is particularly likely in cases in which revictimization is perceived as inevitable (e.g., in a consistently abusive family environment). These defensive actions serve to protect the individual from overwhelming affect, although the resulting body-focused attention facilitates the creation and maintenance of unexplained symptoms via the processes outlined above. 

Peritraumatic body-focused attention would also increase the salience of the physiological aspects of trauma-related emotion (including those related to emotional behavior) while suppressing its affective and cognitive components. This could go some way toward explaining the development of symptoms that in the traditional conversion model would be regarded as symbolic of their underlying psychodynamics. Thus, the rogue representation underlying a fixed dystonic hand could result from a clenched fist associated with angry feelings experienced at the time of trauma; the individual may not recall these feelings, however, due to the inhibition of high-level cognitive and affective processing during the event. Focusing attention on the body during trauma would also augment the salience of other somatosensory or sensorimotor experiences associated with the event, such as pain (cf. Badura et al., 1997), sexual feelings and movements (cf. Betts & Boden, 1992), and different physical defense mechanisms (Nijenhuis, Spinhoven, et al., 1998; Nijenhuis, Vanderlinden, & Spinhoven, 1998), which could lead to the development of symptoms that recapitulate other aspects of the initial trauma. Where other effective self-regulatory strategies are unavailable, focusing attention on the body may become a habitual mode of dealing with negative affect and stressful life events, producing a traitlike vulnerability to unexplained symptoms such as that seen in somatization disorder. 

The development of symptoms also provides a way of expressing negative affect without acknowledging its psychosocial source, a concept that is central to the original conversion model. In addition to the primary gain of reducing negative affect, symptoms may also confer additional advantages to the individual, or secondary gains, with both positive (e.g., greater emotional support and affection from others) and negative reinforcement (e.g., prevention of revictimization) contributing to their maintenance (e.g., Breuer & Freud, 1893–1895/1991; Kellner, 1986; Kretschmer, 1926; Ludwig, 1972). 



The proposed model has obvious implications for the treatment of patients with medically unexplained symptoms. Given the cognitive foundations of the model, it is particularly amenable to CBT, although other treatments such as psychodynamic psychotherapy, physiotherapy, occupational therapy, and pharmacotherapy (e.g., antidepressants) are also consistent with this approach. Once the extent of any physical pathology has been established and an organic explanation for symptoms has been ruled out, the treatment of choice depends on a detailed assessment of the patient's symptoms and the factors associated with their onset and maintenance, including the number, nature, and history of both current and previous symptoms; the presence of any obvious emotional or relational difficulties associated with the onset of symptoms; the nature and extent of emotional disturbance more generally; the degree of disability associated with symptoms; the patient's illness beliefs and attributions about the causes of his or her symptoms (particularly when anxious) and the patient's tendency for catastrophic misinterpretation; the extent of any worry and rumination about symptoms; the nature and extent of any illness behavior, including body checking, help seeking and consulting behavior, reassurance seeking, information seeking, avoidance, and doctor shopping; any history of traumatic events, particularly in childhood; the presence of possible secondary gains (e.g., litigation claims); and relevant personality factors (e.g., negative affectivity, dispositional self-focus). The likely impact of these factors on the individual's tendency to allocate attention to symptoms should be considered throughout. Establishing the possible origins of the rogue representations underlying symptoms is not essential, although doing so could assist in socializing the patient to a psychological account of his or her problems. 

The assessment provides the basis for an idiosyncratic formulation identifying the factors involved in the creation of symptoms and the processes maintaining symptom-focused attention, allowing avenues for intervention to be identified. When symptoms are thought to result from interpersonal difficulties and/or unresolved emotional conflicts, such as those pertaining to early abuse or other childhood traumas, psychodynamically oriented therapy may be an appropriate treatment option. In cases in which interpersonal or emotional conflicts do not appear to be relevant, or a psychodynamic approach is precluded for other reasons (e.g., client preference, unsuitability), the use of CBT is indicated. The present model extends current cognitive-behavioral work in this area by (a) providing a detailed rationale for the use of CBT with unexplained symptoms, (b) identifying the mechanisms involved in clinical change, and (c) suggesting certain additions to the therapeutic arsenal. 

From the outset, efforts should be directed toward helping patients accept a psychological interpretation of their problems. This could be facilitated by presenting a rationale for therapy based on the model proposed here. The advantage of this approach is that it provides a normalizing interpretation of somatoform illness that does not rely on the concept of unconscious conflict, an idea that is often explicitly rejected by patients. Possible socialization exercises might include the use of analogies demonstrating how experience often misrepresents reality, such as the placebo effect, battlefield analgesia, and hypnotic phenomena. The creation or removal of symptoms using suggestion or hypnosis may provide particularly compelling real-time examples of this process (Oakley, 2001). Once patients are suitably socialized to the model, interventions aimed at reducing symptom-focused attention can be considered. Maladaptive beliefs, catastrophic misinterpretations, and illness behaviors can all be addressed using conventional verbal and behavioral reattribution techniques, and the use of controlled worry periods and “detached mindfulness” may assist in minimizing patients' worry and rumination about symptoms (Wells, 1997, 2000). Symptom-focused attention could also be addressed more directly by the use of techniques such as attention training (Wells, 1990), which aims to foster increased control over attentional processing. Unlike traditional CBT, the current approach also provides a rationale for the use of techniques such as suggestion (with or without hypnosis), to deactivate rogue representations more directly (Oakley, 2001), as well as treatments that could help replace them with more adaptive representations (e.g., occupational therapy, physiotherapy). Finally, the role of possible secondary gains should be considered and interventions applied where feasible.

Testing the Model

At the heart of the present approach is the assumption that all somatoform conditions (with the exception of those involving observable physical phenomena) are governed by the same basic mechanism, namely, the repetitive reallocation of high-level attention onto symptoms. This basic assumption allows for the generation of several novel hypotheses by which the model might be falsified. 

Hypothesis 1: The severity and duration of unexplained symptoms will be increased or decreased by manipulations directing attention toward or away from symptoms, respectively.   In addition to simple laboratory studies, this prediction could be assessed clinically by investigating the impact of attention training (Wells, 1990, 2000), a procedure designed to improve attentional control and the ability to focus externally, on patients with unexplained symptoms. 

Hypotheses 2a and 2b: Patients with unexplained symptoms will show a symptom-focused attentional bias when compared with nonsomatoform controls; individuals with a traitlike tendency to develop unexplained symptoms will show a body-focused bias even when asymptomatic.

These hypotheses may be particularly amenable to investigation using electrophysiological (e.g., event-related potentials), neuroimaging (e.g., functional magnetic resonance imaging) and cognitive (e.g., modified emotional Stroop) paradigms. 

Hypothesis 3: Individuals with a tendency to focus attention on the body will be disproportionately vulnerable to developing unexplained symptoms.

This prediction could be addressed in studies investigating the comorbidity between unexplained illness and other conditions associated with body-focused attention, such as body dysmorphic disorder (e.g., Veale, 2001) and the eating disorders (e.g., Fairburn, Shafran, & Cooper, 1999; for preliminary data, see Krishnamoorthy, Brown, & Trimble, 2001). Further research investigating the occurrence of unexplained symptoms in individuals with personality traits such as high private self-consciousness (Fenigstein, Scheier, & Buss, 1975) is also warranted. 

Hypothesis 4: The relationship between traumatic experience and unexplained symptoms is mediated by the use of body-focused attention as a coping strategy

The use of mediator analysis (e.g., Baron & Kenny, 1986) may be particularly useful in this context. 

Hypothesis 5: Modalities affected by unexplained symptoms will be associated with deficits in high-level postattentive processing but not low-level preattentive processing.

A number of different cognitive methodologies could be used to assess this hypothesis (see, e.g., Roelofs et al., 2001; Roelofs, van Galen, Eling, Keijsers, & Hoogduin, 2003). 

Hypothesis 6: The catastrophic misinterpretation of symptoms, illness worry and rumination, illness behavior, and negative affect are all linked to the occurrence and severity of unexplained symptoms by their effect on symptom-focused attention and vice versa.

Multivariate research using appropriate analytic techniques (e.g., mediator analysis, path analysis, structural equation modeling) may be a useful option in this case. Such research could also help identify which, if any, of these factors (or combination of factors) is the most important in the development of unexplained symptoms. 

Further research should also be conducted to establish the validity of the link between suggestion and unexplained illness. In particular, research should address the possibility of using suggested phenomena as laboratory analogues of medically unexplained symptoms. The development of such analogues would facilitate controlled investigation of the factors that mediate and moderate the creation and maintenance of unexplained symptoms. Other studies should further investigate the relationship between unexplained illness and suggestibility, using well-defined patient groups and validated instruments that assess both objective and subjective responses to suggestions. It would also be instructive to address levels of suggestibility in both hypnotic and nonhypnotic contexts, with a view to assessing the relative contribution of hypnosis and suggestibility per se to the creation of unexplained symptoms (cf. Brown & Oakley, 2004; Kirsch, 1997). 


This review has attempted to describe current theoretical models of medically unexplained symptoms and the research cited in their support. It is clear that research motivated by the concepts of dissociation, conversion, and somatization has added much to the understanding of unexplained symptoms. Nevertheless, it is clear that there are certain shortcomings—both empirical and conceptual—with existing approaches in this area. In particular, none of the available models is able to provide an adequate account of how subjectively compelling symptoms can persist in the absence of organic pathology. In this article, an integrative conceptual framework that attempts to address these shortcomings, and that explicitly endorses the idea that unexplained symptoms are subjectively real to the sufferer, has been described. 

The proposed model is consistent with Janet's (1889, 1907) dissociation theory in its assertion that symptoms are caused by stored information in the cognitive system that disrupts the interaction between conscious and preconscious aspects of information processing. In line with conversion theory, the model suggests that this process is often driven by a defensive reaction that operates to reduce the individual's exposure to traumatic affect. Like somatization theory, the model identifies symptom-focused attention as central to the creation and maintenance of unexplained symptoms, and highlights the importance of catastrophic misinterpretation, illness beliefs, rumination and worry, illness behavior, negative affect, and personality features in this process. Although consistent with the concepts of dissociation, conversion, and somatization, the model extends existing theories in this area by reformulating them within a common explanatory framework that is based on cognitive psychological principles. In this way, the model attempts to place the current understanding of unexplained symptoms within the remit of everyday psychology, allowing for a more normalizing interpretation of this phenomenon. It is hoped that the model provides a useful scheme for organizing existing research and theory in this area, as well as an impetus for further research into unexplained symptoms and the development of more effective strategies for their treatment. 



Akagi, H.,  & House, A. (2001). The epidemiology of hysterical conversion   In P. W. Halligan, C. Bass, & J. Marshall (Eds.), Contemporary approaches to the study of hysteria: Clinical and theoretical perspectives (pp. 73–87). Oxford, England: Oxford University Press.

Alajouanine, T., Scherrer, J., Barbizet, J., Calvet, J.,  & Verley, R. (1958). Potentiels évoqués corticaux chez des sujets atteints de troubles somesthésiques   [Cortical evoked potentials in patients with sensory loss].  Revue Neurologique, 98, 757. 

Allport, A. (1987). Selection for action: Some behavioural and neurophysiological considerations of attention and action.   In H. Heuer & A. F. Sanders (Eds.), Perspectives on selection and action (pp. 395–419). Hillsdale, NJ: Erlbaum.

Allport, A. (1988). What concept of consciousness?   In A. J. Marcel & E. Bisiach (Eds.), Consciousness in contemporary science (pp. 159–182). Oxford, England: Oxford University Press.

Alper, K., Devinsky, O., Perrine, K., Luciano, D., Vazquez, B., Pacia, S.,  & Rhee, E. (1997). Dissociation in epilepsy and conversion nonepileptic seizures.   Epilepsia, 38, 991-997. 

American Psychiatric Association (1994).  Diagnostic and statistical manual of mental disorders    (4th ed.). Washington, DC: Author.

Bach, M., Bach, D., Böehmer, F.,  & Nutzinger, D. O. (1994). Alexithymia and somatization: Relationship to DSM-III diagnoses.   Journal of Psychosomatic Research, 38, 529-538. 

Badura, A. S., Reiter, R. C., Altmaier, E. M., Rhomberg, A.,  & Elas, D. (1997). Dissociation, somatization, substance abuse, and coping in women with chronic pelvic pain.   Obstetrics & Gynecology, 90, 405-410. 

Baron, R. M.,  & Kenny, D. A. (1986). The moderator-mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations.   Journal of Personality and Social Psychology, 51, 1173-1182. 

Beck, A. T. (1976).  Cognitive therapy and the emotional disorders.   New York: International Universities Press.

Behrman, J.,  & Levy, R. (1970). Neurophysiological studies on patients with hysterical disturbances of vision.   Journal of Psychosomatic Research, 14, 187-194. 

Bendefeldt, F., Miller, L. L.,  & Ludwig, A. M. (1976). Cognitive performance in conversion hysteria.   Archives of Psychiatry, 33, 1250-1254. 

Benjamin, S.,  & Eminson, D. M. (1992). Abnormal illness behavior: Childhood experiences and long-term consequences.   International Review of Psychiatry, 4, 55-70. 

Bernstein, E. M.,  & Putnam, F. W. (1986). Development, reliability, and validity of a dissociation scale.   Journal of Nervous and Mental Disease, 174, 727-735. 

Betts, T.,  & Boden, S. (1992). Diagnosis, management and prognosis of a group of 128 patients with non-epileptic attack disorder. Part II. Previous childhood sexual abuse in the aetiology of these disorders.   Seizure, 1, 27-32. 

Binzer, M., Andersen, P. M.,  & Kullgren, G. (1997). Clinical characteristics of patients with motor disability due to conversion disorder: A prospective control group study.   Journal of Neurology, Neurosurgery, and Psychiatry, 63, 63-68. 

Blanchard, E. B.,  & Scharff, L. (2002). Psychosocial aspects of assessment and treatment of irritable bowel syndrome in adults and recurrent abdominal pain in children.   Journal of Consulting and Clinical Psychology, 70, 725-738. 

Bowers, K. S. (1992). Dissociated control and the limits of hypnotic responsiveness.   Consciousness & Cognition, 1, 32-39. 

Breuer, J.,  & Freud, S. (1991). Studies on hysteria.   In J. Strachey & A. Strachey (Eds. & Trans.), The Penguin Freud library (Vol. 3). London: Penguin. (Original work published 1893–1895)

Brown, R. J. (2002). Epilepsy, dissociation and nonepileptic seizures.   In M. R. Trimble & B. Schmitz (Eds.), The neuropsychiatry of epilepsy (pp. 189–209). Cambridge, England: Cambridge University Press.

Brown, R. J.  (in press) Dissociation and conversion in psychogenic illness.   In M. Hallett, S. Fahn, J. Jankovic, A. E. Lang, C. R. Cloninger, & S. C. Yudofsky (Eds.), Psychogenic movement disorders: Psychobiology and treatment of a functional disorder. Philadelphia: Lippincott, Williams & Wilkins.

Brown, R. J.,  & Oakley, D. A. (2004). An integrative cognitive model of hypnosis and high hypnotisability.   In M. Heap, R. J. Brown, & D. A. Oakley (Eds.), The highly hypnotizable person: Theoretical, experimental and clinical issues (pp. 152–186). London: Brunner-Routledge.

Brown, R. J.,  & Ron, M. A. (2002). Conversion and somatoform disorders.   In V. S. Ramachandran (Ed.), Encyclopedia of the human brain (Vol. 2, pp. 37–49). San Diego, CA: Academic Press.

Brown, R. J., Schrag, A.,  & Trimble, M. R.  (in press) Dissociation, childhood interpersonal trauma, and family functioning in somatization disorder.   American Journal of Psychiatry, . 

Brown, R. J.,  & Trimble, M. R. (2000). Dissociative psychopathology, non-epileptic seizures, and neurology.   Journal of Neurology, Neurosurgery, and Psychiatry, 69, 285-291. 

Bruner, J. S.,  & Postman, L. (1949). On the perception of incongruity: A paradigm.   Journal of Personality, 18, 206-223. 

Bryant, R. A.,  & McConkey, K. M. (1989). Visual conversion disorder: A case analysis of the influence of visual information.   Journal of Abnormal Psychology, 98, 326-329. 

Bryant, R. A.,  & McConkey, K. M. (1999). Functional blindness: A construction of cognitive and social influences.   Cognitive Neuropsychiatry, 4, 227-241. 

Burgess, P. W.,  & Shallice, T. (1996). Confabulation and the control of recollection.   Memory, 4, 359-412. 

Cardeña, E. (1994). The domain of dissociation.   In S. J. Lynn & J. W. Rhue (Eds.), Dissociation: Clinical and theoretical perspectives (pp. 15–31). New York: Guilford Press.

Cardeña, E.,  & Spiegel, D. (1993). Dissociative reactions to the Bay Area earthquake.   American Journal of Psychiatry, 150, 474-478. 

Chalder, T. (2001). Cognitive behavioural therapy as a treatment for conversion disorders.   In P. W. Halligan, C. Bass, & J. Marshall (Eds.), Contemporary approaches to the study of hysteria: Clinical and theoretical perspectives (pp. 298–311). Oxford, England: Oxford University Press.

Charcot, J. M. (1889).  Clinical lectures on diseases of the nervous system    (Vol. 3). London: New Sydenham Society.

Cioffi, D. (1991). Beyond attentional strategies: A cognitive-perceptual model of somatic interpretation.   Psychological Bulletin, 109, 25-41. 

Cohen, K., Auld, F.,  & Brooker, H. (1994). Is alexithymia related to psychosomatic disorder and somatizing?   Journal of Psychosomatic Research, 38, 119-127. 

Coltheart, M. (1980). Iconic memory and visible persistence.   Perception & Psychophysics, 27, 183-228. 

Craig, T. K. J. (2001). Life events: Meanings and precursors.   In P. W. Halligan, C. Bass, & J. Marshall (Eds.), Contemporary approaches to the study of hysteria: Clinical and theoretical perspectives (pp. 88–101). Oxford, England: Oxford University Press.

Crick, F.,  & Koch, C. (1990). Towards a neurobiological theory of consciousness.   Seminars in the Neurosciences, 2, 263-275. 

Crimlisk, H. L., Bhatia, K., Cope, H., David, A., Marsden, C. D.,  & Ron, M. A. (1998). Slater revisited: 6 year follow up of patients with medically unexplained motor symptoms.   British Medical Journal, 316, 582-586. 

David, D., Brown, R. J., Pojoga, C.,  & David, A. (2000). The impact of posthypnotic amnesia and directed forgetting on implicit and explicit memory: New insights from a modified process dissociation procedure.   International Journal of Clinical and Experimental Hypnosis, 48, 267-289. 

Escobar, J. I., Waitzkin, H., Cohen Silver, R., Gara, M.,  & Holman, A. (1998). Abridged somatization: A study in primary care.   Psychosomatic Medicine, 60, 466-472. 

Fairburn, C., Shafran, R.,  & Cooper, Z. (1999). A cognitive behavioural theory of anorexia nervosa.   Behaviour Research and Therapy, 37, 1-13. 

Fallik, A.,  & Sigal, M. (1971). Hysteria: The choice of symptom site. A review of 40 cases of conversion hysteria.   Psychotherapy and Psychosomatics, 19, 310-318. 

Fenigstein, A., Scheier, M. F.,  & Buss, A. H. (1975). Public and private self-consciousness: Assessment and theory.   Journal of Consulting and Clinical Psychology, 43, 522-527. 

Fillingim, R. B.,  & Fine, M. A. (1986). The effects of internal versus external information processing on symptom perception in an exercise setting.   Health Psychology, 5, 115-123. 

Fink, P., Sørensen, L., Engberg, M., Holm, M.,  & Munk-Jørgensen, P. (1999). Somatization in primary care. Prevalence, health care utilization and general practitioner recognition.   Psychosomatics, 40, 330-338. 

Frankel, F. H. (1994). Dissociation in hysteria and hypnosis: A concept aggrandized.   In S. J. Lynn & J. W. Rhue (Eds.), Dissociation: Clinical and theoretical perspectives (pp. 80–93). New York: Guilford Press.

Fukudu, M., Hata, A., Niwa, S., Hiramatsu, K., Yokokoji, M., Hayashida, S.  et al. (1996). Event-related potential correlates of functional hearing loss: Reduced P3 amplitude preserved N1 and N2 components in a unilateral case.   Neuropsychiatry and Clinical Neurosciences, 50, 85-87. 

Gould, R., Miller, B., Goldberg, M.,  & Benson, D. (1986). The validity of hysterical signs and symptoms.   Journal of Nervous and Mental Disease, 174, 593-597. 

Gregory, R. L. (1970).  The intelligent eye.   London: Weidenfeld & Nicolson.

Halliday, A. M. (1968). Computing techniques in neurological diagnosis.   British Medical Bulletin, 24, 253. 

Halligan, P. W., Athwal, B. S., Oakley, D. A.,  & Frackowiak, R. S. J. (2000, March 18). Imaging hypnotic paralysis: Implications for conversion hysteria.   Lancet, 355, 986-987. 

Halligan, P. W.,  & David, A. S. (1999). Conversion hysteria: Towards a cognitive neuropsychological account.   Cognitive Neuropsychiatry, 4, 161-163. 

Hansell, S.,  & Mechanic, D. (1986). The socialization of introspection and illness behavior.   In S. McHugh & T. M. Vallis (Eds.), Illness behavior: A multidisciplinary model (pp. 253–260). New York: Plenum Press.

Hernández-Peón, R., Chávez-Ibarra, G.,  & Aguilar-Figueroa, E. (1963). Somatic evoked potentials in one case of hysterical anesthesia.   Electroencephalography and Clinical Neurophysiology, 15, 889-892. 

Hilgard, E. R. (1977).  Divided consciousness: Multiple controls in human thought and action.   New York: Wiley.

Holmes, E., Brown, R. J., Mansell, W., Fearon, R. P., Hunter, E., Frasquilho, F.,  & Oakley, D. A. (2004).  Are there two qualitatively distinct forms of dissociation? A review and some clinical implications.   Manuscript submitted for publication.

Hope, D. A.,  & Heimberg, R. G. (1985). Public and private self-consciousness and social phobia.   Journal of Personality Assessment, 52, 626-639. 

Horowitz, M. J. (1986).  Stress response syndromes    (2nd ed.). Northvale, NJ: Aronson.

Horvath, T., Friedman, J.,  & Meares, R. (1980). Attention in hysteria: A study of Janet's hypothesis by means of habituation and arousal measures.   American Journal of Psychiatry, 137, 217-220. 

Hotopf, M., Mayou, R., Wadsworth, M.,  & Wessely, S. (1999). Childhood risk factors for adults with medically unexplained symptoms: Results from a national birth cohort study.   American Journal of Psychiatry, 156, 1796-1800. 

Iezzi, A.,  & Adams, H. E. (1993). Somatoform and factitious disorders.   In P. B. Sutker & H. E. Adams (Eds.), Comprehensive handbook of psychopathology (pp. 167–201). New York: Plenum Press.

Isaac, M., Janca, A., Burke, K. C., Costa e Silva, J. A., Acuda, S. W., Altanura, A. C.  et al. (1995). Medically unexplained symptoms somatic in different cultures: A preliminary report from Phase I of the World Health Organization international study of somatoform disorders.   Psychotherapy and Psychosomatics, 64, 88-93. 

Janet, P. (1889).  L'automatisme psychologique.   [Psychological automatisms]. Paris: Alcan.

Janet, P. (1907).  The major symptoms of hysteria.   New York: Macmillan.

Janoff-Bulman, R. (1992).  Shattered assumptions: Towards a new psychology of trauma.   New York: Free Press.

Jeannerod, M. (1994). The representing brain—Neural correlates of motor intention and imagery.   Behavioral and Brain Sciences, 17, 187-202. 

Kashner, T. M., Rost, K., Cohen, B., Anderson, M.,  & Smith, G. R. (1995). Enhancing the health of somatization disorder patients. Effectiveness of short-term group therapy.   Psychosomatics, 36, 462-470. 

Kellner, R. (1986).  Somatization and hypochondriasis.   New York: Praeger-Greenwood.

Kellner, R. (1990). Somatization: Theories and research.   Journal of Nervous and Mental Disease, 178, 150-160. 

Kihlstrom, J. F. (1987, September 18). The cognitive unconscious.   Science, 237, 1445-1451. 

Kihlstrom, J. F. (1992). Dissociative and conversion disorders.   In D. J. Stein & J. Young (Eds.), Cognitive science and clinical disorders (pp. 247–270). San Diego, CA: Academic Press.

Kihlstrom, J. F.,  & Canter Kihlstrom, L. (1999). Self, sickness, somatization, and systems of care.   In R. J. Contrada & R. D. Ashmore (Eds.), Rutgers series on self and social identity: Vol. 2. Self, social identity, and physical health: Interdisciplinary explorations (pp. 23–42). New York: Oxford University Press.

Kinzl, J. F., Traweger, C.,  & Biebl, W. (1995). Family background and sexual abuse associated with somatization.   Psychotherapy and Psychosomatics, 64, 82-87. 

Kirkwood, C. R., Clure, H. R., Brodsky, R., Gould, G. H., Knaak, R., Metcalf, M.,  & Romeo, S. (1982). The diagnostic content of family practice: 50 most common diagnoses recorded in the WAMI community practices.   Journal of Family Practice, 15, 485-492. 

Kirmayer, L. J.,  & Robbins, J. M. (Eds.) (1991a).  Current concepts of somatization: Research and clinical perspectives.   Washington, DC: American Psychiatric Press.

Kirmayer, L. J.,  & Robbins, J. M. (1991b). Three forms of somatization in primary care: Prevalence, co-occurrence, and sociodemographic characteristics.   Journal of Nervous and Mental Disease, 179, 647-655. 

Kirmayer, L. J.,  & Taillefer, S. (1997). Somatoform disorders.   In S. M. Turner & M. Hersen (Eds.), Adult psychopathology and diagnosis (3rd ed., pp. 333–383). New York: Wiley.

Kirsch, I. (1985). Response expectancy as a determinant of experience and behavior.   American Psychologist, 40, 1189-1202.

Kirsch, I. (1997). Suggestibility or hypnosis: What do our scales really measure?   International Journal of Clinical and Experimental Hypnosis, 45, 212-225. 

Kirsch, I.,  & Lynn, S. J. (1998). Dissociation theories of hypnosis.   Psychological Bulletin, 123, 100-115. 

Kosslyn, S. M. (1996).  Image and brain.   Cambridge, MA: MIT Press.

Kouyanou, K., Pither, C. E., Rabe-Hesketh, S.,  & Wessely, S. (1998). A comparative study of iatrogenesis, medication abuse, and psychiatric morbidity in chronic pain patients with and without medically explained symptoms.   Pain, 76, 417-426. 

Kretschmer, E. (1926).  Hysteria.   New York: Nervous & Mental Disease.

Krishnamoorthy, E. S., Brown, R. J.,  & Trimble, M. R. (2001). Personality and psychopathology in non-epileptic attack disorder (NEAD): A prospective study.   Epilepsy & Behavior, 2, 418-422. 

Kroenke, K., Spitzer, R. L., deGruy, F. V., Hahn, S. R., Linzer, M., Williams, J. B. W.  et al. (1997). Multisomatoform disorder. An alternative to undifferentiated somatoform disorder for the somatizing patient in primary care.   Archives of General Psychiatry, 54, 352-358. 

Kroenke, K.,  & Swindle, R. (2000). Cognitive-behavioral therapy for somatization and symptom syndromes: A critical review of controlled clinical trials.   Psychotherapy and Psychosomatics, 69, 205-215. 

Kuyk, J., Spinhoven, P.,  & van Dyck, R. (1999). Hypnotic recall: A positive criterion in the differential diagnosis between epileptic and pseudoepileptic seizures.   Epilepsia, 40, 485-491. 

Labbate, L. A., Cardeña, E., Dimitreva, J., Roy, M.,  & Engel, C. C. (1998). Psychiatric syndromes in Persian Gulf War veterans: An association of handling dead bodies with somatoform disorders.   Psychotherapy and Psychosomatics, 67, 275-279. 

Lader, M. (1982). The psychophysiology of hysteria.   In A. Roy (Ed.), Hysteria (pp. 81–87). New York: Wiley.

Lader, M.,  & Sartorius, N. (1968). Anxiety in patients with hysterical conversion symptoms.   Journal of Neurology, Neurosurgery, and Psychiatry, 31, 490-495. 

Lavie, N. (1995). Perceptual load as a necessary condition for selective attention.   Journal of Experimental Psychology: Human Perception and Performance, 21, 451-468. 

Leventhal, H. (1986). Symptom reporting: A focus on process.   In S. McHugh & T. M. Vallis (Eds.), Illness behavior: A multidisciplinary model (pp. 219–237). New York: Plenum Press.

Lindsay, S. D.,  & Read, J. D. (1994). Psychotherapy and memories of childhood sexual abuse: A cognitive perspective.   Applied Cognitive Psychology, 8, 281-338. 

Lipowski, Z. J. (1968). Review of consultation psychiatry and psychosomatic medicine. III. Theoretical issues.   Psychosomatic Medicine, 30, 395-422. 

Lipowski, Z. J. (1988). Somatization: The concept and its clinical application.   American Journal of Psychiatry, 145, 1358-1368. 

Logan, G. D. (1988). Toward an instance theory of automatization.   Psychological Review, 95, 492-527. 

Lorenz, J., Kunze, K.,  & Bromm, B. (1998). Differentiation of conversive sensory loss and malingering by P300 in a modified oddball task.   NeuroReport, 9, 187-191. 

Ludwig, A. M. (1972). Hysteria: A neurobiological theory.   Archives of General Psychiatry, 27, 771-777. 

Mace, C. J.,  & Trimble, M. R. (1996). Ten-year prognosis of conversion disorder.   British Journal of Psychiatry, 169, 282-288. 

Marcel, A. J. (1983a). Conscious and unconscious perception: An approach to the relations between phenomenal experience and perceptual processes.   Cognitive Psychology, 15, 238-300. 

Marcel, A. J. (1983b). Conscious and unconscious perception: Experiments on visual masking and word recognition.   Cognitive Psychology, 15, 197-237. 

Marshall, J. C., Halligan, P. W., Fink, G. R., Wade, D. T.,  & Frackowiak, R. S. (1997). The functional anatomy of a hysterical paralysis.   Cognition, 64, B1-B8. 

McConkey, K. M. (2001). Hysteria and hypnosis: cognitive and social influences.   In P. W. Halligan, C. Bass, & J. Marshall (Eds.), Contemporary approaches to the study of hysteria: Clinical and theoretical perspectives (pp. 203–215). Oxford, England: Oxford University Press.

Meierkord, H., Will, B., Fish, D.,  & Shorvon, S. (1991). The clinical features and prognosis of pseudoseizures diagnosed using video-EEG telemetry.   Neurology, 41, 1643-1646. 

Merskey, H. (1979).  The analysis of hysteria.   London: Balliere, Tindall & Cassell.

Merskey, H.,  & Buhrich, N. A. (1975). Hysteria and organic brain disease.   British Journal of Medical Psychology, 48, 359-366. 

Neely, J. H. (1977). Semantic priming and retrieval from lexical memory: Roles of inhibitionless spreading activation and limited capacity attention.   Journal of Experimental Psychology: General, 106, 226-254. 

Neuman, O. (1987). Beyond capacity: A functional view of attention.   In H. Heuer & A. F. Sanders (Eds.), Perspectives on selection and action (pp. 361–394). Hillsdale, NJ: Erlbaum.

Nijenhuis, E. R. S., Spinhoven, P., van Dyck, R., van der Hart, O.,  & Vanderlinden, J. (1998). Degree of somatoform dissociation and psychological dissociation in dissociative disorder is correlated with reported trauma.   Journal of Traumatic Stress, 11, 711-730. 

Nijenhuis, E. R. S., Vanderlinden, J.,  & Spinhoven, P. (1998). Animal defensive reactions as a model for trauma-induced dissociative reactions.   Journal of Traumatic Stress, 11, 243-260. 

Norman, D. A.,  & Shallice, T. (1986). Attention to action: Willed and automatic control of behavior.   In R. J. Davidson, G. E. Schwartz, & D. Shapiro (Eds.), Consciousness and self-regulation: Vol. 4. Advances in research and theory (pp. 1–18). New York: Plenum Press.

Oakley, D. A. (1999). Hypnosis and conversion hysteria: A unifying model.   Cognitive Neuropsychiatry, 4, 243-265. 

Oakley, D. A. (2001). Hypnosis and suggestion in the treatment of hysteria.   In P. W. Halligan, C. Bass, & J. Marshall (Eds.), Contemporary approaches to the study of hysteria: Clinical and theoretical perspectives (pp. 312–329). Oxford, England: Oxford University Press.

Parsons, T. (1964).  Social structure and personality.   London: Collier MacMillan.

Pennebaker, J. W. (1982).  The psychology of physical symptoms.   New York: Springer-Verlag.

Pennebaker, J. W.,  & Brittingham, G. L. (1982). Environmental and sensory cues affecting the perception of physical symptoms.   In A. Baum & J. Singer (Eds.), Advances in environmental psychology (Vol. 4, pp. 115–136). Hillsdale, NJ: Erlbaum.

Pennebaker, J. W.,  & Lightner, J. M. (1980). Competition of internal and external information in an exercise setting.   Journal of Personality and Social Psychology, 39, 165-174. 

Pilowsky, I. (1978). A general classification of abnormal illness behaviors.   British Journal of Medical Psychology, 51, 131-137. 

Posner, M. I.,  & Snyder, C. R. R. (1975). Facilitation and inhibition in the processing of signals.   In P. M. A. Rabbitt & S. Dornick (Eds.), Attention and Performance V (pp. 669–682). New York: Academic Press.

Pribor, E. E., Yutzi, S. H., Dean, T. J.,  & Wetzel, R. D. (1993). Briquet's syndrome, dissociation, and abuse.   American Journal of Psychiatry, 150, 1507-1511. 

Prinz, W. (1997). Perception and action planning.   European Journal of Cognitive Psychology, 9, 129-154. 

Ramachandran, V. S.,  & Hirstein, W. (1998). The perception of phantom limbs: The D. O. Hebb lecture.   Brain, 121, 1603-1630. 

Raskin, M., Talbott, J.,  & Myerson, A. (1966). Diagnosis of conversion reactions: Predictive value of psychiatric criteria.   Journal of the American Medical Association, 197, 530-534. 

Reilly, J., Baker, G. A., Rhodes, J.,  & Salmon, P. (1999). The association of sexual and physical abuse with somatization: Characteristics of patients presenting with irritable bowel syndrome and non-epileptic attack disorder.   Psychological Medicine, 29, 399-406. 

Reynolds, J. R. (1869). Paralysis and other disorders of motion and sensation dependent on idea.   British Medical Journal, 1, 483-485. 

Rief, W., Hiller, W.,  & Margraf, J. (1998). Cognitive aspects of hypochondriasis and somatization syndrome.   Journal of Abnormal Psychology, 107, 587-595. 

Robbins, J. M.,  & Kirmayer, L. J. (1986). Illness cognition, symptom reporting, and somatization in primary care.   In S. McHugh & T. M. Vallis (Eds.), Illness behavior: A multidisciplinary model (pp. 283–302). New York: Plenum Press.

Robbins, J. M.,  & Kirmayer, L. J. (1991). Cognitive and social factors in somatization.   In L. J. Kirmayer & J. M. Robbins (Eds.), Current concepts of somatization: Research and clinical perspectives (pp. 107–141). Washington, DC: American Psychiatric Press.

Roelofs, K., Hoogduin, C. A. L., Keijsers, G. P. J., Näring, G. W. B., Moene, F. C.,  & Sandijck, P. (2002). Hypnotic susceptibility in patients with conversion disorder.   Journal of Abnormal Psychology, 111, 390-395. 

Roelofs, K., Keijsers, G. P. J., Hoogduin, K. A. L., Näring, G. W. B.,  & Moene, F. C. (2002). Childhood abuse in patients with conversion disorder.   American Journal of Psychiatry, 159, 1908-1913. 

Roelofs, K., Näring, G. W. B., Keijsers, G. P. J., Hoogduin, C. A. L., van Galen, G. P.,  & Maris, E. (2001). Motor imagery in conversion paralysis.   Cognitive Neuropsychiatry, 6, 21-40. 

Roelofs, K., van Galen, G. P., Eling, P., Keijsers, G. P. J,  & Hoogduin, C. A. L. (2003). Endogenous and exogenous attention in patients with conversion paresis.   Cognitive Neuropsychology, 20, 733-745. 

Ron, M. A. (1994). Somatisation in neurological practice.   Journal of Neurology, Neurosurgery, and Psychiatry, 57, 1161-1164. 

Ron, M. A. (2001). The prognosis of hysteria/somatization disorder.   In P. W. Halligan, C. Bass, & J. Marshall (Eds.), Contemporary approaches to the study of hysteria: Clinical and theoretical perspectives (pp. 271–282). Oxford, England: Oxford University Press.

Rumelhart, D. E.,  & McClelland, J. L.the PDP Research Group (1986).  Parallel distributed processing.   Explorations in the microstructure of cognition. Volume 1: Foundations. Cambridge, MA: MIT Press.

Sackeim, H. A., Nordlie, J. W.,  & Gur, R. C. (1979). A model of hysterical and hypnotic blindness: Cognition, motivation, and awareness.   Journal of Abnormal Psychology, 88, 474-489. 

Salkovskis, P. M. (1989). Somatic problems.   In K. Hawton, P. M. Salkovskis, J. Kirk, & D. M. Clark (Eds.), Cognitive behaviour therapy for psychiatric problems: A practical guide (pp. 235–276). Oxford, England: Oxford University Press.

Salkovskis, P. M.,  & Warwick, H. M. C. (1986). Morbid preoccupations, health anxiety and reassurance: A cognitive behavioural approach to hypochondriasis.   Behaviour Research and Therapy, 24, 597-602. 

Salmon, P., Peters, S.,  & Stanley, I. (1999). Patients' perceptions of medical explanations for somatisation disorders: Qualitative analysis.   British Medical Journal, 318, 372-376. 

Schrag, A., Brown, R. J.,  & Trimble, M. R. (2004). The reliability of self-reported diagnoses in patients with somatoform disorder.   Journal of Neurology, Neurosurgery, and Psychiatry, 75, 608-611. 

Sebanz, N., Knoblich, G.,  & Prinz, W. (2003). Representing others' actions: Just like one's own?   Cognition, 88, B11-B21. 

Shallice, T. (1988).  From neuropsychology to mental structure.   Cambridge, England: Cambridge University Press.

Shallice, T.,  & Burgess, P. W. (1996). The domain of supervisory processes and temporal organization of behavior.   Philosophical Transactions of the Royal Society of London, Series B, 351, 1405-1412. 

Sharpe, M.,  & Bass, C. (1992). Pathophysiological mechanisms in somatization.   International Review of Psychiatry, 4, 81-97. 

Shipko, S. (1982). Alexithymia and somatization.   Psychotherapy and Psychosomatics, 37, 193-201. 

Sierra, M.,  & Berrios, G. E. (1998). Depersonalization: Neurobiological perspectives.   Biological Psychiatry, 44, 898-908. 

Sierra, M.,  & Berrios, G. E. (1999). Towards a neuropsychiatry of conversion hysteria.   Cognitive Neuropsychiatry, 4, 267-287. 

Sifneos, P. E. (1973). The prevalence of “alexithymic” characteristics in psychosomatic patients.   Psychotherapy and Psychosomatics, 22, 255-262. 

Skelton, J. A.,  & Croyle, R. T. (1991).  Mental representations in health and illness.   New York: Springer-Verlag.

Slater, E. T. O.,  & Glitherco, E. (1965). A follow-up of patients diagnosed as suffering from “hysteria.”   Journal of Psychosomatic Research, 9, 9-13. 

Sloman, S. A. (1996). The empirical case for two systems of reasoning.   Psychological Bulletin, 119, 3-22. 

Smith, G. R., Monson, R. A.,  & Ray, D. C. (1986). Patients with multiple unexplained symptoms: Their characteristics, functional health, and health care utilization.   Archives of Internal Medicine, 146, 69-72. 

Smith, G. R., Rost, K.,  & Kashner, M. (1995). A trial of the effect of a standardized psychiatric consultation on health outcomes and costs in somatizing patients.   Archives of General Psychiatry, 52, 238-243. 

Smith, T. W.,  & Greenberg, J. (1981). Self-focused attention and depression.   Motivation and Emotion, 5, 323-333. 

Stewart, D. E. (1990). The changing faces of somatization.   Psychosomatics, 31, 153-158. 

Stone, J., Sharpe, M., Rothwell, P. M.,  & Warlow, C. P. (2003). The 12 year prognosis of unilateral functional weakness and sensory disturbance.   Journal of Neurology, Neurosurgery, and Psychiatry, 74, 591-596. 

Styles, E. A. (1997).  The psychology of attention.   Hove, England: Psychology Press.

Tellegen, A.,  & Atkinson, G. (1974). Openness to absorbing and self-altering experiences (“absorption”), a trait related to hypnotic susceptibility.   Journal of Abnormal Psychology, 83, 268-277. 

Temple, N. (2001). Psychodynamic psychotherapy in the treatment of conversion hysteria.   In P. W. Halligan, C. Bass, & J. Marshall (Eds.), Contemporary approaches to the study of hysteria: Clinical and theoretical perspectives (pp. 283–297). Oxford, England: Oxford University Press.

Vaillant, G. E. (1993).  The wisdom of the ego.   Cambridge, MA: Harvard University Press.

Van Ommeren, M. V., Sharma, G. K., Komproe, I., Sharma, G. K., Cardeña, E., de Jong, J. T.  et al. (2001). Trauma and loss as determinants of medically unexplained epidemic illness in a Bhutanese refugee camp.   Psychological Medicine, 31, 1259-1267. 

Veale, D. (2001). Cognitive-behavioural therapy for body dysmorphic disorder.   Advances in Psychiatric Treatment, 7, 125-132. 

Velmans, M. (1991). Is human information processing conscious?   Behavioral and Brain Sciences, 14, 651-726. 

Velmans, M. (2000).  Understanding consciousness.   London: Routledge.

Walker, E., Katon, W., Harrop-Griffiths, J., Holm, L., Russo, J.,  & Hickok, L. R. (1988). Relationship of chronic pelvic pain to psychiatric diagnoses and childhood sexual abuse.   American Journal of Psychiatry, 145, 75-80. 

Wall, P. D. (1993). Pain and the placebo response.   In J. Marsh (Ed.), Ciba Foundation Symposium: Vol. 174. Experimental and theoretical studies of consciousness (pp. 187–216). Chichester, England: Wiley.

Watson, D.,  & Clark, L. A. (1984). Negative affectivity: The disposition to experience aversive emotional states.   Psychological Bulletin, 96, 465-490. 

Watson, D.,  & Pennebaker, J. W. (1989). Health complaints, stress and distress: Exploring the role of negative affectivity.   Psychological Review, 96, 234-254. 

Wegner, D. M.,  & Guiliano, T. (1980). Arousal-induced attention to the self.   Journal of Personality and Social Psychology, 38, 719-726. 

Wells, A. (1990). Panic disorder in association with relaxation-induced anxiety: An attentional training approach to treatment.   Behavior Therapy, 21, 273-280. 

Wells, A. (1997).  Cognitive therapy of anxiety disorders.   Chichester, England: Wiley.

Wells, A. (2000).  Emotional disorders and metacognition.   Chichester, England: Wiley.

Wells, A.,  & Matthews, G. (1994).  Attention and emotion: A clinical perspective.   Hove, England: Erlbaum.

Wessely, S. M. (2001). Discrepancies between diagnostic criteria and clinical practice.   In P. W. Halligan, C. Bass, & J. Marshall (Eds.), Contemporary approaches to the study of hysteria: Clinical and theoretical perspectives (pp. 63–72). Oxford, England: Oxford University Press.

Whitlock, F. A. (1967). The aetiology of hysteria.   Acta Psychiatrica Scandinavica, 43, 144-162. 

Woods, S. M., Natterson, J.,  & Silverman, J. (1966). Medical students' disease: Hypochondriasis in medical education.   Journal of Medical Education, 41, 785-790. 

Woody, E. Z.,  & Bowers, K. S. (1994). A frontal assault on dissociated control.   In S. J. Lynn & J. W. Rhue (Eds.), Dissociation: Clinical and theoretical perspectives (pp. 52–79). New York: Guilford Press.

World Health Organization (1992).  The ICD-10 classification of mental and behavioural disorders: Clinical descriptions and diagnostic guidelines.   Geneva, Switzerland: Author.