The physical signs of PNEA are as variable as epileptic seizures themselves. They can range from mild tremulousness to full convulsions. Patients may remain aware, may be partially responsive, or may be unaware, unresponsive, and amnestic. Like epileptic seizures, PNEAs are difficult to diagnose just based on the history. Physicians rarely see them and typically have to rely on descriptions from untrained (and often distressed) witnesses.

Certain behaviors are more common in PNEA and less common in epileptic seizures (Reuber & Elger, 2003).These include undulating, rhythmic, asynchronous, purposeful, or goal-directed movements, pelvic thrusting, side-to-side head movements, crying, closed eyes and resistance to opening, verbal responses, lack of cyanosis during a prolonged “tonic” phase, and rapid reorientation postictally. These behaviors should raise suspicions of PNEA, but they sometimes can be misleading. For example, directed aggression during the ictus is almost always considered a sign of nonepileptic activity. However, in a recently reported case, a young man repeatedly made focused verbal threats to physicians that were accompanied by profanity and “shooting” them with his index finger. The events were associated with bifrontal epileptiform activity and hyperperfusion in the right lateral and orbitofrontal cortex on SPECT imaging (Shih, LeslieMazwi, Falcao, et al., 2009). One of our patients was partly responsive during his seizures and had a remote history of drug abuse. He was repeatedly barred from the local emergency room and eventually assaulted and incarcerated by police because staff assumed he was faking seizures to obtain drugs. He was later shown by video-EEG monitoring to have frontal lobe seizures. Major injuries during PNEA are rare but not unheard of. One of our PNEA patients had broken an arm during an attack. It later came out that she had intentionally and repeatedly broken her hand with a hammer or bat on several occasions as a child, in response to severe family dysfunction and abuse.

Because of the significant risk for misdiagnosis based on clinical signs, the “gold-standard” diagnostic method for PNEA is to show absence of epileptiform discharges during the patient’s typical event(s) which usually requires inpatient, video-EEG scalp monitoring. In rare cases, PNEAs diagnosed based on scalp monitoring later have been shown to be epileptic with invasive monitoring, but all these cases had simple partial or complex partial events and interictal EEG abnormalities or other evidence of neurologic involvement (Wyler, Hermann, Blumer, et al., 1993). Inpatient evaluation also allows a thorough medical and psychological evaluation to rule out nonepileptic, organic causes of seizures and to evaluate risk factors for PNEA. Unfortunately, definitive diagnosis is often delayed, by an average of 7 years in one study (Reuber, Fernandez, Bauer, et al., 2002). During this time, patients are exposed to side effects of antiepilepsy drugs (AEDs), social stigma, and activity restrictions, often resulting in an intractable, disabled lifestyle. In our experience, the longer the patient has carried an epilepsy diagnosis, the harder it is for them to understand and accept that the events are psychogenic. This is particularly important in light of data indicating that successful recovery from PNEA is directly correlated with the patient’s understanding and acceptance of the diagnosis (Carton, Thompson, & Duncan, 2003).

PNEAs are symptoms of one of a number of psychiatric diagnoses. Most commonly, PNEAs are manifestations of an underlying somatoform disorder. The most recent version of psychiatric nomenclature (DSM-V) refers to these as somatic symptom disorders (American Psychiatric Association, 2013). These patients have physical symptoms that suggest a general medical condition but cannot be explained on the basis of the physical findings. Patients show undue worry and concerns about their symptoms, even in the face of reassurance. In severe cases, symptoms become the focus of the patient’s life and may dominate their relationships with others.

PNEAs that occur as the lone unexplained symptom reflect a conversion disorder. When PNEAs are one of a number of medically unexplained symptoms, these patients should be diagnosed with somatic symptom disorder, with different specifiers, depending on the role of pain in the presentation, duration, and severity. Severity is based on the degree of associated behavioral disturbance and the number and severity of symptoms. PNEAs that appear as paroxysmal changes in mood, sensation, or cognition may be symptoms of other psychiatric diagnoses (e.g., panic disorder, PTSD, dissociative disorders). There are few studies where psychiatric diagnosis has been prospectively established by standard psychiatric interviews. The majority of diagnoses were conversion disorder (36–63 %) or other somatoform disorder (16–23 %), with other diagnoses including anxiety disorders (7–15 %) and dissociative disorders (4–7 %) (Galimberti, Ratti, Murelli, et al., 2003; Marchetti, Kurcgant, Neto, et al., 2008).

PNEAs as an expression of malingering or factitious disorder are rare, in our experience.² These diagnoses should be suspected in forensic and other settings where there is an external incentive (injury compensation, avoidance of major role responsibility, or culpability) or other evidence that suggests deliberate attempts to deceive (e.g., significant inconsistencies in the history and examination). Rarely, children have admitted to manufacturing attacks to draw attention to ongoing abuse or neglect. Certain results of psychometric assessment may be useful in supporting these diagnoses (see “Utility of Psychometric Assessment,” below).

Medically unexplained symptoms are the most common symptoms in primary care settings. Prevalence ranges from 8 to 25 %, depending of the severity and number of symptoms (Escobar, Waitzkin, Silver, et al., 1998; Kroenke, Spitzer, deGruy, et al., 1997). Prevalence of conversion in neurology clinics has been estimated at 20 % (Mace & Trimble, 1996). The exact incidence and prevalence of PNEA is not known. There are no community-based studies, no doubt, because of the expense of obtaining video-EEG monitoring as a gold-standard diagnosis in a community-based sample. PNEA patients represent ~10–20 % of patients referred to epilepsy centers and a somewhat higher percentage of monitored patients. This has led to an estimated prevalence of 2–33/100,000 (Benbadis & Hauser, 2000). PNEA usually onsets between 20 and 30 years of age, but there have been cases as young as 4 and as old as 70. Ten to 50 % of patients also have current or past epileptic seizures, making diagnosis and treatment even more difficult.

The characteristics that distinguish patients with PNEA from patients with epilepsy and healthy controls have been reviewed extensively (Bodde, Brooks, Baker, et al., 2009, Reuber & Elger, 2003). Since antiquity, PNEAs have been seen more often in women, such that they were often felt to be a uniquely female problem. Modern studies suggest that 75 % of patients are women (Lesser, 1996). PNEA patients are more likely than epilepsy patients to have other psychiatric diagnoses, particularly personality disorders with borderline features (Bowman & Markand, 1996; Reuber, Pukrop, Bauer, et al., 2004). High rates of clinically significant depressive symptoms (52 %), suicidal ideation (40 %), and prior suicide attempts (20 %) have also been reported (Ettinger, Devinsky, Weisbrot, et al., 1999). PNEA patients have high levels of psychopathology, as measured by the MMPI. The variety of profile types is felt to represent the diversity of psychopathology underlying the symptoms (Cragar, Berry, Schmitt, et al., 2005; Dodrill, Wilkus, & Batzel, 1993). PNEA patients are more likely to have difficulty describing their own emotional states or recognizing them in others (alexithymia) (Bewley, Murphy, Mallows, et al., 2005).

PNEA patients often have a past history of trauma and recent social stress, particularly involving family. They report a higher frequency of past traumatic experiences in general (Rosenberg, Rosenberg, Williamson, et al., 2000). Reported childhood sexual or physical abuse is more common in PNEA (33 %) than in epilepsy patients (9 %) (Alper, Devinsky, Perrine, et al., 1993). PNEA patients experienced more adverse or traumatic events in the 12 months prior to diagnosis than patients with epilepsy (Binzer, Stone, & Sharpe, 2004). Their family members have higher levels of concerns about their own health. They perceive their families as more hostile and critical than family members of patients with epilepsy (Wood, McDaniel, Burchfiel, et al., 1998).

Risk factors for the broader range of somatoform disorders are also of interest, since many PNEA patients carry these diagnoses. In a large, German population-based cohort, somatoform symptoms were associated with female gender, older age, lower education, lower income, and rural residence (Hiller, Rief, & Brahler, 2006). In a prospectively followed, population-based cohort of adolescents, the new onset of a somatoform diagnosis was predicted by female gender, lower social class, substance use, anxiety or affective disorder, and the experience of traumatic sexual or physical threat events (Lieb, Zimmermann, Friis, et al., 2002). Severe illness in a parent predicted later somatoform symptoms in a population-based birth cohort (Hotopf, 2002). Patients with somatoform disorder and histrionic personality disorder also are more likely to have a first-degree relative with antisocial personality disorder (Lilienfeld, Van, Larntz, et al., 1986). Somatoform patients also are more susceptible to hypnotic suggestion (Roelofs, Hoogduin, Keijsers, et al., 2002).

An important limitation of the literature on PNEA risk factors is that most studies compared patients with PNEA to patients with epileptic seizures. Comparisons to psychiatric samples without seizure-like events are rare. Therefore, some of these risk factors may reflect risk for psychopathology in general and not specifically for PNEA. In one of the few studies to include other psychiatric patients, a history of childhood sexual abuse was more frequent in PNEA patients than in inpatients with other psychiatric diagnoses (Betts & Boden, 1992). In another study, PNEA patients and patients with other somatoform disorders had similar rates of comorbid psychiatric disorders and levels of anxiety, depression, and anger. These rates were higher than in healthy controls (Mokleby, Blomhoff, Malt, et al., 2002). However, PNEA patients were more likely than other somatoform patients and controls to report a history of minor head injuries, have more subtle EEG abnormalities, and express higher levels of hostility.

Pre-modern Theories: Many premodern writers recognized that PNEAs were more common in women and were associated with emotionality and psychosocial stressors, but their theories to explain these observations were naturally constrained by the knowledge of the time (Temkin, 1971). The Hippocratic writers saw hysterical seizures as a form of epilepsy. The Roman physician Celsus wrote that “a disease originating from the uterus sometimes made women so weak it prostrated them as in epilepsy.” Renaissance physicians explained all hysterical phenomena as the effect of vapors rising from the uterus to the brain. Recognizing the interplay of constitutional and environmental factors, Todd, a nineteenth-century physician, described a “highly excitable” hysteric who had been “subjected to moral, and perhaps physical influences also, well-calculated to keep up that state.” Briquet theorized that hystero-epilepsy embodied painful emotions that prompted reactions of the emotional centers of the brain. Women were more prone, due to an innate emotionality that allowed them to fulfill their “noble mission in life.” Breuer’s description of the hysterical pregnancy of Anna O., “a young woman of exceptional cultivation and talents,” played a key role in initially convincing Freud that repression of unpleasant memories and emotions (especially sexual ones) led to “conversion” into physical symptoms that symbolized the original trauma (Gay, 1988). Janet’s concept of dissociation was one of the first purely cognitive theories. It held that psychosomatic symptoms reflect a narrowing or fragmenting of attention in response to extreme emotional stress. This caused symptoms ranging from over-attending to subjective sensations to unresponsiveness and amnesia (Janet, 1907).

Despite continued heuristic value in lay and clinical circles, older conceptualizations of conversion and dissociation are limited as explanatory theories, from a strictly scientific point of view. Both hold that sensory and motor changes are caused by subjective, unobservable mental processes whose neurologic basis is not specified. The psychoanalytic concepts of repression and symbolic representation of past events in the unconscious are not falsifiable. Modern explanations of the phenomenon of PNEA rely on objective observations, are potentially falsifiable, and invoke mechanisms by which sensation and behavior might plausibly emerge from the central nervous system.

Psychological Theories: Attachment theory has been invoked to explain observed associations between over-reporting of physical symptoms and early traumatic experiences (Lamberty, 2008). Attachment theory holds that early experiences with caretakers have an enduring influence on how people perceive themselves and how they form relationships as adults (Bowlby, 1969). Validated measures allow researchers to classify various attachment “styles” (e.g., “secure,” “anxious/attached,” “anxious/avoidant,” “dismissing,” “preoccupied,” and “fearful”) that reflect interest in, and comfort with, engagement in intimate and autonomous relationships (Ainsworth, 1967; Bartholomew & Horowitz, 1991). Preoccupied and fearfully attached persons report more physical symptoms than securely attached persons (Ciechanowski, Walker, Katon, et al., 2002). Preoccupied persons also used more health care than fearful ones, suggesting that attachment style also influences strategies for managing symptoms. In a community-based sample, persons with a history of childhood trauma were more likely to report physical symptoms and to have an insecure attachment style (Waldinger, Schulz, Barsky, et al., 2006). Patients with insecure or preoccupied attachment styles may over-report physical symptoms because they are hypervigilant to a range of perceived threats. Over-reporting also may elicit concerns and support that strengthen their attachment in key relationships.

In a similar way, somatoform symptoms have been described as a “bodily idiom of distress,” a way to acceptably communicate emotions that might otherwise disrupt important relationships. In one case series, in-depth interviews revealed that 13/14 conversion patients had been experiencing intolerable family situations that they felt unable to address openly (Griffith, Polles, & Griffith, 1998). The authors described these as “unspeakable dilemmas” where “family, social, religious or political circumstances… have imposed forced choices… and consequent suffering must remain hidden from important persons involved in the situation.” Many of the patients were the least expressive family members during the family interviews. They later confided that they presently were feeling threatened by physical or sexual assault or were having difficulty coping with a psychiatrically disabled family member. Social communication theories may also explain some of the many outbreaks of “mass hysteria” that have occurred across cultures and over time (Evans & Bartholomew, 2009). For example, episodes of mass hysteria in Malaysia were virtually unheard of prior to the imposition of strict Islamic codes of female behavior in 1960. Subsequent episodes often involved female students or sweatshop workers and were accompanied by outspoken criticism of authorities. Interpreting them as signs of a physical problem provided a culturally acceptable explanation for an otherwise intolerable breach of the social order (Bartholomew, 1997).

Attachment and social communication theories have been combined to explain conversion symptoms presenting in “good” children in the absence of the usual risk factors (e.g., abuse, overt stressors, family dysfunction) (Kozlowska, 2001). These children often do very well in school, are admired by peers, and generally lack recognizable psychopathology. Their families tend to be “supernormal” and expect a high degree of academic and athletic achievement, social propriety, and family cohesion. Providers tend to find these parents mildly intimidating. According to Kozlowska, children become “compulsively compliant” in order to bolster insecure attachments in the face of mounting expectations. They anticipate parents’ needs and expectations. They focus on performing tasks instead of attending to relationships. In this setting, conversion symptoms are seen as passive resistance that secures attachment by avoiding conflict. Symptoms elicit functional and emotional support from the parents in a face-saving way that also provides relief from overwhelming demands. Adults also may use a compulsively compliant strategy to manage “unspeakable dilemmas” where they feel they have to comply with intolerable demands and suppress their own needs out of a fear of threatening important attachments (e.g., the abused spouse, the overworked single parent).

Neurobiological Theories: Attachment and social communication theories point to testable hypotheses about how a person’s history of relationships might predispose them to develop PNEA. However, they do not address a central question: how does the central nervous system produce such dramatic changes in sensation or behavior without apparent awareness or voluntary control? Watching some of these events, it is difficult to imagine that they are not done on purpose, yet patients’ protestations typically seem credible. There also are so many other less painful symptoms that would have a similar effect. Updating the concept of dissociation, Brown (2004) has proposed that medically unexplained symptoms reflect disruption of normal attention systems. These systems shape the contents of consciousness and control behavior via spreading activation of hierarchically organized, neural networks. According to one such model, primary attention systems (PAS) govern sensory and motor representations instantiated in neural networks that are activated rapidly and efficiently (e.g., sensory perception, skilled motor movements, implicit memory) (Norman & Shallice, 1986). Behaviors activated by PAS are typically experienced as occurring without conscious effort (e.g., walking, some aspects of driving), while perceptions and thoughts are experienced as intuitive, self-evident, or automatic (e.g., simple visual discrimination, “realizing” that you forgot to put out the trash). Secondary attention systems (SAS) activate sensory, cognitive, and motor representations that subserve higher-level, goal-directed behaviors. These are experienced as effortful and associated with feelings of conscious volition and self-awareness.

Within this framework, medically unexplained symptoms reflect inappropriate PAS activation of chronically overactive “rogue representations.” These have been acquired during prior experiences (e.g., trauma) or strengthened by chronic over-attention to bodily sensations (e.g., in somatically focused persons). In the sensory realm, inappropriate activation of “rogue” representations amplifies normal sensory inputs (e.g., psychogenic pain or discomfort) or suppresses them (e.g., psychogenic anesthesia or amnesia). These experiences feel “real” to the person but are inconsistent with sensory inputs (i.e., medically unexplained). Activation of “rogue” motor representations by PAS produces behaviors that are inappropriate to the goals subserved by representations activated by SAS. Behaviors are experienced as “real” but involuntary and outside awareness (e.g., PNEA).

Brown’s theory extends more modern theories of dissociative behavior (such as occurs during hypnosis) that posit an “amnesic barrier” between executive systems and lower-level perceptual and motor systems (Hilgard, 1977). Evidence that an amnesic barrier can be affected by suggestion in PNEA comes from a study of patients with ictal amnesia. Under hypnotic suggestion designed to facilitate recall, 85 % of PNEA patients, but no epilepsy patients, were able to recall what had occurred during a prior event (Kuyk, Spinhoven, & van Dyck, 1999). The idea that PNEA patients have aberrant, unconscious attention mechanisms is supported by a recent study where subjects were asked to view and rapidly name colors under interference conditions, similar to the Stroop task (Bakvis, Roelofs, Kuyk, et al., 2009). The colors were preceded by a picture of a face that was flashed for 14 ms and backward masked in order to prevent conscious processing. PNEA patients were slower than healthy controls to name colors when the colors were preceded by angry faces, but not by happy ones. This suggests that PNEA patients have a preconscious attentional bias to threatening stimuli that places demands on information processing and affects subsequent behavior. Interestingly, PNEA patients with a greater attentional bias were more likely to report a history of sexual trauma and to have higher basal cortisol levels, indicating HPA axis dysfunction (Bakvis, Spinhoven, & Roelofs, 2009). Whether these findings are specific to PNEA or are seen with other psychiatric disorders remains to be determined. In functional imaging studies, patients with conversion paralysis show decreased activity in primary motor areas and increased activity in functionally related, inhibitory areas that also are involved in attention networks (Halligan, Athwal, Oakley, et al., 2000; Marshall, Halligan, Fink, et al., 1997). In studies of conversion anesthesia, peripheral stimulation that was not perceived resulted in deactivation of contralateral, primary sensory cortex and corresponding activation of limbic structures (Mailis-Gagnon, Giannoylis, Downar, et al., 2003; Vuilleumier, 2005). Brown’s attentional theory and these neuroimaging findings are far from definitive, but they do provide a neurologically plausible framework for testing hypotheses about the neural correlates of abnormal perceptual processing in these patients.

Multi-factorial Models: Recently, two groups have proposed similar multifactorial models to explain the wide range of risk factors for PNEA (Bodde et al., 2009; Reuber, 2009). Each model suggests that PNEAs arise from a complex interplay of predisposing, precipitating, and perpetuating factors that varies from patient to patient. Predisposing factors include constitutional factors (e.g., genetics, temperament, learning disabilities, subtle structural brain abnormalities) and early experiences (e.g., abuse, neglect, other sources of insecure attachment). Later exposure to stressful experiences (e.g., actual or recalled trauma, actual or perceived threats, unspeakable dilemmas, or physical illness) precipitates the onset of events. Events are perpetuated by distress over social stigma, isolation, dependency and diagnostic uncertainty, and by factors that reinforce the “sick role” (e.g., increased family concern or support, avoidance of conflicts or burdensome responsibilities, disability benefits). The main advantage of these models is that they account for (1) the heterogeneity in the history and presentation of PNEA patients, (2) no single risk factor being either necessary or sufficient to cause PNEA, and (3) PNEA risk factors being risk factors for many other psychiatric conditions. They provide a clinically useful framework for thinking about how the many facets of an individual patient’s constitution and experience may interact to produce PNEA. They remain somewhat limited from a scientific standpoint. Predisposing, precipitating, and perpetuating factors are quite common and how some factors (and not others) interact to produce PNEA is not highly specified. Therefore, almost any set of factors can be retrospectively invoked to explain PNEA in a given patient. They also do not propose mechanisms that explain how PNEAs are experienced as “real” by the patient yet involuntary and outside awareness.