Article Text


Functional (conversion) neurological symptoms: research since the millennium
  1. Alan J Carson1,2,
  2. Richard Brown3,
  3. Anthony S David4,
  4. Roderick Duncan5,
  5. Mark J Edwards6,
  6. Laura H Goldstein4,
  7. Richard Grunewald7,
  8. Stephaine Howlett8,
  9. Richard Kanaan4,
  10. John Mellers4,
  11. Timothy R Nicholson4,
  12. M Reuber7,
  13. Anette-Eleonore Schrag6,
  14. Jon Stone1,
  15. Valerie Voon9,10,
  16. on behalf of UK-FNS*
  1. 1Department of Clinical Neurosciences, University of Edinburgh, Edinburgh, UK
  2. 2Department of Rehabilitation Medicine, Edinburgh, UK
  3. 3School of Psychological Sciences, University of Manchester, Manchester, UK
  4. 4Institute of Psychiatry, King's College London, London, UK
  5. 5West of Scotland Regional Epilepsy Service, Department of Neurology, Southern General Hospital, Glasgow, UK
  6. 6UCL Institute of Neurology, University College London, London, UK
  7. 7Academic Neurology Unit, University of Sheffield, Sheffield, UK
  8. 8Department of Neurology, Sheffield Teaching Hospitals NHS Foundation Trust, Sheffield, UK
  9. 9Department of Psychiatry, University of Cambridge, Cambridge, UK
  10. 10Behavioural and Clinical Neurosciences Institute, University of Cambridge, Cambridge, UK
  1. Correspondence to Dr A J Carson, Department of Psychiatry, University of Edinburgh, Robert Fergusson Unit, Royal Edinburgh Hospital, Tipperlin Road, Edinburgh, EH10 5HF, UK; a.carson{at}


Functional neurological symptoms (FNS) are commonly encountered but have engendered remarkably little academic interest. ‘UK-Functional Neurological Symptoms (UK-FNS)’ was an informal inaugural meeting of UK based clinicians in March 2011 with a variety of research and clinical interests in the field. This narrative review reflects the content of the meeting, and our opinion of key findings in the field since the turn of the millennium.

Statistics from


UK neurologists have described patients with functional neurological symptoms (FNS) as the ‘most difficult to help’.1 FNS came bottom of a recent survey of US neurologists' ‘most likeable conditions’.2 This has been reflected in a dearth of research interest over the previous century. Yet it is increasingly recognised that FNS are one of the most common conditions a neurologist will encounter. This has been reflected by increased research activity and burgeoning interest over the past decade. The research has been multifaceted, spanning a range of techniques and disciplines, from electrophysiology through cognitive neuroscience to psychotherapy trial design. The diversity is intellectually attractive but can be daunting as it inevitably takes the clinician out of familiar fields.

‘UK-Functional Neurological Symptoms (UK-FNS)’ was an informal inaugural meeting of UK based clinicians in March 2011 with a variety of research and clinical interests in the field. This narrative review is based on the content of the meeting. The meeting focused on sensorimotor and seizure presentations. The third main domain of cognitive complaints was not discussed.3 The wider but related domains of functional somatic symptoms in general, such as irritable bowel syndrome, chronic fatigue syndrome, fibromyalgia, etc,4 were also not discussed. These omissions were not a reflection of lack of importance but simply the available time and the focus of the meeting. In this review, we have aimed to describe key contributions to the literature on functional sensorimotor and seizure symptoms since the millennium across a range of disparate investigatory techniques. We hope to provide an account of the most recent findings of relevance to diagnosis, underpinning mechanisms, risks factors and treatment.


Nomenclature in this field remains problematic with a host of different terms being recommended, each having their proponents and critics, and may be subject to further revision in the forthcoming editions of International Classification of Diseases (ICD) and Diagnostic and Statistical Manual of Mental Disorders (DSM) (see box 1). In the absence of a universally agreed taxonomy, discussion of the issue tends to reflect, often passionately held, beliefs and opinions5 ,6 but rarely scientific data. The current proposals for revision to DSM-V are that the criterion B (‘the positive identification of psychological features’) and criterion C (‘the condition should not be deliberately feigned’) should both be dropped. With regard to criterion B, this is a research question that although widely believed remains to be confirmed. Criterion C, while essential at a theoretical level, is practically unworkable in the evaluation of any individual patient. It has also been suggested that an additional criterion of inconsistent or incongruent physical signs should be added to the definition as this is the basis of clinical differentiation from other neurological disorders. Rather than attempt to resolve the issue here, the interested reader is referred to these reviews.7–9

Box 1

Current Diagnostic and Statistical Manual of Mental Disorders, fourth edition (DSM-IV) criteria for conversion disorder and proposed revisions for DSM-V (which may be subject to further revision)

DSM-IV diagnostic criteria for 300.11 conversion disorder

  1. One or more symptoms or deficits affecting voluntary motor or sensory function that suggest a neurological or other general medical condition.

  2. Psychological factors are judged to be associated with the symptom or deficit because the initiation or exacerbation of the symptom or deficit is preceded by conflicts or other stressors.

  3. The symptom or deficit is not intentionally produced or feigned (as in factitious disorder or malingering).

  4. The symptom or deficit cannot, after appropriate investigation, be fully explained by a general medical condition, or by the direct effects of a substance or as a culturally sanctioned behaviour or experience.

  5. The symptom or deficit causes clinically significant distress or impairment in social, occupational or other important areas of functioning, or warrants medical evaluation.

  6. The symptom or deficit is not limited to pain or sexual dysfunction, does not occur exclusively during the course of somatisation disorder and is not better accounted for by another mental disorder.

Specify type of symptom or deficit:

  • With motor symptom or deficit

  • With sensory symptom or deficit

  • With seizures or convulsions

  • With mixed presentation

Current proposed criteria for DSM-V (name under discussion at the time of writing)

Criteria A, B, C and D must all be fulfilled to make the diagnosis:

  1. One or more neurological symptoms such as altered voluntary motor, sensory function, cognition or seizure-like episodes.

  2. The symptom, after appropriate medical assessment, is found not to be due to a general medical condition, the direct effects of a substance or a culturally sanctioned behaviour.

  3. The physical signs or diagnostic findings are internally inconsistent or incongruent with recognised neurological disorder.

  4. The symptom causes clinically significant distress or impairment in social, occupational or other important areas of functioning or warrants medical evaluation.


There are two major barriers to epidemiological study in the field of FNS:

The first is case definition.7 ,9 One approach is to regard ‘cases’ as all patients who present with neurological symptoms that lack pathophysiological explanations.10 The alternate approach is to attempt to operationalise the psychogenic component with the hope of identifying a more tightly defined group,11 but this may leave many unclassified; for instance, Schrag et al12 concluded that only 11% of patients in their series with fixed dystonia, which is thought to be ‘psychogenic’ in the majority of cases, had clear evidence of a DSM-IV conversion disorder although many more had a prior history of other somatoform symptoms. The unclassified group may simply represent patients in whom the clinician has failed correctly to identify the psychogenic mechanism(s).13 A more challenging alternative explanation is that such mechanisms may not be the only mechanism of causation. The Movement Disorders Society surveyed 519 neurologists and found that only 18% regarded the presence of emotional disturbance as essential to the diagnosis.14 Additional problems with definition include how one classifies FNS that occur in the context of defined neurological conditions.15

The second barrier is case ascertainment. The requirement for a detailed neurological examination to confirm the diagnosis limits the likelihood of large community studies and, realistically, epidemiological studies are limited to patients having contact with neurological services.

Despite this, there is some consistency of results: Akagi and House16 concluded that, notwithstanding often weak methodologies and disparate geographical settings, the reported incidence rates of ‘conversion disorder’ were reasonably consistent between 4 and 12 per 100 000 population per year. A detailed study of incident cases of functional weakness alone in South East Scotland17 found an estimated minimum incidence of 3.9 per 100 000 per year. The lower estimates of community prevalence of all conversion disorder, extracted from population based case registers, are much higher at around 50/100 000 population.16

The Scottish Neurological Symptoms Study (SNSS), a multicentre prospective cohort study of 3781 new neurology outpatients, found at baseline that 30% had neurological symptoms that were either ‘not at all’ or only ‘somewhat’ explained by neurological disease; 5.6% of all outpatients were thought to have conversion disorders.18 The patients with poorly explained symptoms had equivalent physical disability and higher rates of emotional distress than those with neurological disease. Of patients with poorly explained symptoms 27% were not working for health reasons and in receipt of capacity related benefits (vs 19% of those with explained symptoms).19

Epidemiological studies have highlighted a range of risk factors for the occurrence of FNS. The risk factors are generally similar to those described for emotional disorders such as anxiety and depression. Some other findings deserve highlighting. People with functional weakness17 and psychogenic non-epileptic seizures (PNES)20 have been found to have similar illness beliefs to their corresponding disease counterparts except that paradoxically they tend to be less likely to attribute their symptoms to stress than patients with disease. There is good evidence that aversive childhood experience, including but not exclusively childhood sexual abuse, is more common in patients with FNS than disease or community controls, although not necessarily psychiatric controls. Nonetheless this only applies to a minority of patients. The frequency of recent life events, around the time of onset, is also somewhat increased in most but not all studies.21


Potential misdiagnosis is often a concern leading to reluctance in making this diagnosis. However, systematic review of 27 studies of conversion symptoms (n=1466) found that the reported frequency of misdiagnosis has been approximately 4% since 1970; the rate was unaffected by the widespread introduction of clinical imaging.22 In the more recent SNSS, there were only four clearcut misdiagnoses in 1114 patients after 18 months of follow-up.23

Studies commonly report lack of recovery for between half and two-thirds of patients.24 Chronicity appears as a predictor of poorer outcome.25 Similarly, in SNSS, patient rated follow-up data were available for 63% of 1144 patients at 12 months, and 67% reported being the same or worse. In a multivariate analysis, poor outcome was predicted by expectation of non-recovery, non-attribution of symptoms to psychological factors and the receipt of health related benefits at the time of the initial consultation.26

Motor symptoms

Progress in refining clinical symptomatology of FNS (weakness/movement disorders) has been slow. Hoover's sign of hip extension weakness17 and a hip abductor version of Hoovers sign27 gained some empirical credibility for distinguishing functional/psychogenic paralysis (and ‘la belle indifference’ lost some28). The association of convergence spasm with psychogenic movement disorders was described29 but more blinded studies of simple clinical signs are sorely needed.

Clinical studies to distinguish ‘organic’ from ‘psychogenic’ tremor have demonstrated the utility of manoeuvres such as the ‘entrainment’ test30 ,31 (copying a 3 Hz tapping movement from the examiner with one hand either results in a significant shift in tremor frequency or an inexplicable inability to perform the tapping task) and these have been complemented by quantitative work using electrophysiological tremor recordings (see below).

A large study of fixed dystonia, typically a clenched fist or inverted ankle, highlighted the large overlap with what is labelled complex regional pain syndrome and DSM-IV somatisation disorder.12 Sixty-three per cent of the patients in this study had a physical injury at onset. A high frequency of physical injury (37%) at onset was also found in a systematic review of 132 studies of functional motor symptoms (n=869)32; subsequent studies have also highlighted this relationship.33 Panic and dissociation are also common features at onset in those patients whose symptoms develop acutely.34

The advent of thrombolysis as an acute stroke treatment has highlighted the fact that functional paralysis is an important stroke mimic. Helpfully, a study of thrombolysis in stroke mimics found no excess haemorrhage in this group.35

Probably the most important development of the decade in terms of changing global attitudes was the promotion of psychogenic movement disorders as a serious topic through two international multidisciplinary conferences and books36 ,37 (one in press), coordinated by a group of respected North American movement disorder neurologists.

Psychogenic non-epileptic seizures

The millennium began with the best epidemiological estimate so far made for the incidence of PNES,38 suggesting 4.6/100 000/year. This included video EEG confirmed diagnoses only and is likely to be an underestimate. In a large series of patients, diagnostic delay to diagnosis of 7 years was documented39 although a recent study demonstrated that this could be shortened to 1 year or less with diagnosis at a first seizure clinic and rapid access to video EEG.40 A potential contributory aetiological role of organic neural deficits was identified, with indices of brain dysfunction occurring in a higher than expected proportion of PNES patients.41 Two identifiable subgroups, distinct from the generality of patients, were characterised. First, patients with PNES and learning disability42 had a high rate of comorbid epilepsy and a high likelihood of having situational triggers for attacks. Second, patients with late onset PNES43 who had high rates of life threatening medical illness and associated health related anxiety which were presumed to be causal factors for PNES in their case.

One of the most practical clinical developments was the demonstration of the potential for making the diagnosis using suggestion in combination with short outpatient video EEG.44 ,45 One of the most interesting ancillary diagnostic aids has been the application of conversation analysis and other linguistic techniques.46 ,47 This technique identified clear and clinically discriminating differences in the way patients with PNES and epilepsy spoke about their attacks.

Aetiological work has focused on predisposing factors and psychological models, finding high rates of personality disorder,48 anxiety and dissociation49 ,50 Of note, alexithymia (a state of deficiency in understanding, processing or describing emotions) could not discriminate individuals with NES from those with organic manifestations.51

Two recent outcome papers have been published, one looking at outcome over a 1–10 year period, finding that only 29% of patients were free of attacks and that socioeconomic outcome was also poor.52 The second study was of short term (6–12 months) outcomes only, and found that 38% of patients became attack free with the only intervention being taken off antiepileptic drugs and being told the diagnosis.53 This study also demonstrated striking and immediate reduction in emergency healthcare utilisation, reflecting increasing awareness of outcomes other than attack frequency. This has led to increased attention on how the diagnosis should be communicated.40 ,54 The main factor predicting outcome has been employment status.40 ,53 ,55 Significantly, chronicity is not a predictor of outcome in PNES.52 ,53


There have been two main strands to electrophysiological research in the field: (1) development of positive diagnostic markers separating functional from organic neurological disorders; and (2) research exploring the pathophysiology of functional disorders.

Diagnostic markers

For a long time, electrophysiological help with the diagnosis of functional disorders was largely restricted to the use of video EEG in patients with suspected PNES. However, there is now reasonable evidence that patients with functional tremor can be distinguished from organic tremor using tremor recordings; a shift in tremor frequency towards the frequency of tapping movements made with another limb and a pause in the tremor (usually of <1 s) when the patient makes a sudden ballistic movement to a target with the other limb are among a small group of measures that appear to have reasonable sensitivity and specificity, recently confirmed to be the case even for distinguishing functional tremor from more unusual organic tremor syndromes such as dystonic tremor.56–61 Pre-movement potentials, EEG potentials (‘Bereitschaftspotentials’) occurring in the second or so before voluntary movement, can be detected before movements in patients with functional myoclonus. Using these recordings, two groups have independently shown that the vast majority of patients classified as having idiopathic propriospinal myoclonus or spinal segmental myoclonus have pre-movement potentials before jerks, and are therefore best classified as functional.62 ,63 Lastly, a simple blink reflex test has been shown in a small group to have high sensitivity and specificity in distinguishing patients with functional from those with organic blepharospasm.64 Such tests (and others that may be developed) are an adjunct to and not a replacement for clinical observation, and are only of use in the small minority of patients where there is diagnostic difficulty despite careful examination. They require a laboratory set up and a clinician able to interpret the results that may not be available in all centres.


Liepert et al explored motor system excitability in patients with functional paralysis using transcranial magnetic stimulation. Excitability appears to be normal at rest, but when imagining movement with the affected limb, it appears to reduce in contrast with healthy participants where it increases.65 ,66 However, observation of movement in another person causes an increase in motor excitability recorded from the homologous limb in patients, similar to healthy participants.66 There will always be difficulties in interpreting the results of studies using interventions that are hard to quantify, such as motor imagery, but these findings do suggest an interesting potential dissociation between movement control and mirror neurons (which mediate effects of action observation). In the area of functional movement disorders, interesting results have emerged from studies of patients with fixed abnormal postures (usually called fixed dystonia). They appear to have some electrophysiological abnormalities that are typically associated with organic dystonia.67 ,68 However, they lack the fundamental abnormality of enhanced response to plasticity protocols seen in primary dystonia.69 The electrophysiological parameters that are abnormal are open to confounds from muscle activity and from psychological state (attention,70 anxiety71). Therefore, interpretation is difficult, and this uncertainty has done little to help the rather polarised debate about how to classify this particular group of patients.

Functional imaging

Using functional imaging to explore the mechanisms of functional symptoms has a logic that may seem inescapable: if symptoms have no structural basis, then there must be a functional disturbance that causes them. Two pioneering case studies in the 1990s72 ,73 established the paradigms—passive stimulation in functional sensory loss and attempted movement in functional paralysis. In each case, the sensory deactivation or failure of motor activity was accompanied by right frontal activation, suggesting inhibition of sensorimotor action by ‘higher’ areas: a good start.

Two studies at the beginning of the millennium appeared to give confident direction to the field, using stimulation74 and attempted movement paradigms.75 The study of Vuilleumier et al using single photon emission computed tomography studied seven patients with unilateral sensorimotor symptoms using bilateral vibration, during both symptomatic and recovery phases. They found no sensory changes but demonstrated reduced contralateral basal ganglia, and particularly caudate, activity during the symptomatic phase but not on recovery. This supported a model of functional disruption of cortico-striato-thalamo-cortical loops underlying symptoms. Subsequent functional MRI studies were smaller, but yielded conflicting findings, either because they found sensory deactivation,76 ,77 or because the basal ganglia was found to be more activated. Ghaffar and colleagues78 suggested a solution to the varying sensory deactivation: stimulation of the affected limb did not activate the contralateral primary somatosensory (S1) region in three patients whereas bilateral limb stimulation did. On the other hand, all of the studies found activations in the anterior cingulate or orbitofrontal cortex during stimulation although Vuilleumier found no difference in activity in these regions between the symptomatic and recovered state, suggesting this may be a feature of the stimulus (or the patients) rather than the symptoms.

When a patient with functional paralysis fails to move (whether in the scanner or the clinic), the question is often asked as to whether they really tried or indeed whether they merely pretended to be paralysed. Spence et al addressed this in a small series of three patients with unilateral motor weakness moving a joystick while undergoing PET compared with four controls feigning weakness in the same task.75 They found decreased left dorsolateral prefrontal cortical activity in patients when attempting to move their affected limb, regardless of which limb was affected. This suggested that conversion disorder was demonstrably distinct from feigning. The distinction appeared to be with impairments in motor intention (planning); an executive component. Subsequent functional MRI studies using similar paradigms also found differences from feigning but unfortunately not a consistent pattern of activation.79 ,80

Cojan and colleagues81 examined the idea of excessive executive inhibition in a task that involved making a decision to either make a specific movement or inhibit it in response to a cue (a Go–NoGo paradigm). They found that the motor cortical regions during movement preparation and right inferior frontal regions during inhibition were unaffected which led them to suggest that both motor intention and inhibition were intact. By contrast, they found functional paralysis to be associated with ventrolateral prefrontal cortex activity, intriguingly a region with substantive input from areas involved in emotional processing.

Recently, Voon et al used functional MRI (figure 1) to examine the relationship of emotion and symptom production obliquely—asking patients with conversion disorder to perform an emotional rather than motor based task.82 ,83 They found abnormal correlations in activation between the amygdala (where it would be expected given the paradigm) and the supplementary motor area (where it would not)—hinting at a model, not only for symptom generation but also for why these might arise in the first place—a question addressed directly by those looking at potentially triggering life events.84

Figure 1

(A) Conversion disorder patients had lower activity of the right temporoparietal junction (TPJ) (top) when comparing involuntary conversion tremor versus their own voluntary mimicked tremor. Patients also had lower functional connectivity between the right TPJ and sensorimotor cortices and cerebellum (bottom). The TPJ is implicated in the feed forward model of motor control and suggested to act as a comparator of intended (sensory prediction) and the actual sensory outcome for online monitoring of movement. As sensory outcome is intact, the authors suggest that decreased TPJ comparator activity may be related to impaired generation of the intended sensory prediction during motor preparation.75 (B) Conversion disorder patients had greater right amygdala activity to positive and negative affective stimuli compared with healthy controls (top) along with greater limbic–motor functional connectivity between the amygdala and a region involved in motor preparation, the supplementary motor area (bottom). Arousing stimuli might aberrantly influence motor preparation, providing a potential explanation for the influence of stress on motor function76 (with permission from Neurology82 and Brain83).

More recent studies have suggested potential abnormalities in top down prefrontal control.85–87 Although the imaging studies are small and use different tasks and populations, studies are starting to point towards abnormalities in neural networks88–91 rather than specific areas. It is too early to claim that a comprehensive theory is emerging but one possibility is that there are overly sensitive amygdala fear responses (ie, abnormal response to stimuli (even objectively neutral stimuli), possibly conditioned by previous learning experiences) that drive changes in networks mediating perceptual experiences and/or movement plans. These changes, in the presence of abnormal self-directed attention92 (cf prefrontal activations in functional imaging studies), are capable of producing movements or perceptual experiences which are not associated with a normal sense of self-agency and are therefore interpreted by patients as involuntary symptoms of an underlying disease.

Cognitive models

More than a century since the earliest psychological accounts of hysterical phenomena there remains a dearth of well studied and widely supported psychological theories of FNS. Perhaps one of the most significant recent developments is the increasing emphasis on research and theory from mainstream cognitive psychology. Although recent models have recapitulated many of the ideas outlined in historical accounts,93–95 a closer look at the cognitive processes underlying symptoms has suggested new hypotheses and allowed novel empirical approaches to be developed.

Possibly the most important insight in recent models96 ,97 is that the cognitive system is organised hierarchically, with most perceptual and behavioural processing occurring automatically and outside awareness. Symptoms could arise as a result of psychological influences at these lower levels of processing, without involving conscious systems at all. Consistent with this, Bakvis et al found evidence of altered preconscious threat processing in patients with PNES.98 In this experiment, patients who had been sexually abused were more likely than non-abused patients to have their attention drawn to angry faces presented outside of their awareness.

FNS have been described96 as being produced when someone's ‘executive’ control systems (thought to be located in the frontal cortex) attempt to solve their seemingly insolvable emotional problems by feeding false perceptions into their self-awareness; a cognitive neuroscience model for Freudian ideas of conversion.

Alternately,97 FNS may simply arise from distortions in self-awareness that can develop in the absence of emotional difficulties. By this view, FNS result from the over activation of ‘rogue’ mental representations, with anything that fuels this activation contributing to symptom development and maintenance (see figure 2).

Figure 2

Proposed mechanism of functional neurological symptoms in one recent cognitive model.83 Exposure to symptom relevant information leads to the creation of corresponding representations in memory. Functional disorders arise when the activation level of such symptom representations increases to the point where they override sensory input and intrude into experience and/or action. In this example, encountering epileptic seizures in oneself, others or the media generates a mental representation of ‘what it is like’ to have a seizure. This ‘seizure’ programme (or rogue representation) may then be activated by the individual worrying excessively about the possibility of having a seizure and looking out for possible evidence that one might be about to occur; such as symptoms of increased arousal from anxiety or symptoms of orthostatic dizziness. Psychogenic non-epileptic seizures are said to arise when the activation of the ‘seizure’ programme reaches a threshold level, leading to seizure-like distortions in awareness and behaviour.

As in historical theories of hysteria, attentional processes are afforded a central role in contemporary accounts of FNS. Various paradigms informed by basic cognitive research have been used to study these processes. Attentional dysfunction has been observed in one small study of conversion paresis99 while impairments in working memory have been found in patients with PNES.100 Evidence of alteration in attentional processes, including hypervigilance, delayed disengagement of attention and attentional avoidance has also been found in non-clinical groups with high symptom reports.101 ,102 Another recent line of research has sought to measure individual differences in the tendency to experience distortions in awareness and the relationship between this, other cognitive processes103 and physical symptom reporting.104

Although the increasing focus on cognition in FNS is a welcome theoretical development, research in this area remains limited. Relatively few studies have been carried out and all are restricted by small sample sizes, the use of non-clinical populations and/or uncertainty over what is measured by the cognitive tasks in question. While the available data are promising, much larger studies involving clinical groups and better cognitive paradigms are urgently required. On a theoretical level, more work needs to be done to integrate cognitive approaches to FNS with other areas of research and theory, such as those on attachment, emotion regulation and interpersonal dysfunction.105 Efforts to integrate cognitive and neurobiological approaches may be particularly productive.21 ,106

Other psychological factors

Models of predisposing, precipitating and perpetuating factors in the development and maintenance of FNS have been an important contribution to the formulation and treatment of these disorders and have reflected their multifactorial nature.107–109 Conceptually, it remains unclear whether it may be better to derive separate models for motor FNS and PNES patients, given that clinical characteristics may differ between these patient groups, or whether co-occurrence of these phenomena in many patients speaks to the possibility of a unified underlying mechanism.20 Depending on the specific model, research emphasis has been placed on different aetiological variables, with this focus changing as the knowledge base develops. While initially it has, for example, been informative to explore the interrelationship between physical and sexual abuse in childhood and adult life,110 or the interrelationship between more distant and recent life events preceding the onset of the disorder,111 it is increasingly recognised that such explanations do not apply to the majority of patients and a broader conceptualisation using a biopsychosocial perspective21 ,107–109 has increasingly been viewed as essential. This permits consideration of individual differences such as genetic vulnerability that has been of relevance in other conditions when considering the link between childhood sexual abuse and dissociation.112 Studies of biomarkers in PNES patients,98 for example, suggest that psychological markers of PNES may be best conceptualised, or at least investigated, as part of an integrated biopsychosocial model of PNES and dissociative disorders more generally. Consideration of these broadening conceptualisations may, with sufficiently large samples, begin to elucidate why some people exposed, for example, to childhood abuse develop FNS, others depressive illness, while others remain healthy. They might also provide biomarkers for treatment outcomes.

Of the specific psychological mechanisms underlying FNS, dissociation (a somewhat ambiguous collective term describing a range of psychopathological processes altering a person's level of awareness and/or the integration of sensorimotor function, emotions, thoughts, memories and identity which may be subjectively perceived as a sense of disconnection) has attracted considerable interest. Dissociative disorders per se are recognisably difficult to study and dissociation has been conceptualised in different ways.113 In particular, distinctions have been made between psychoform and somatoform dissociation. The majority of studies have investigated psychoform dissociation in patients with FNS. Although not always the case,114 a number of studies have shown that PNES patients demonstrate evidence of a heightened dissociative tendency,49 ,50 ,115–117 as measured using scales of cognitive or psychoform dissociation, such as the Dissociative Experiences Scale118 (DES) (or adaptations) and other measures of pathological dissociation.49 ,116 However, DES scores in patients with PNES are typically less markedly increased than in other dissociative disorders.49 Additionally, different dissociative disorders may give rise to different DES profiles119 and different abuse histories may potentially lead to different relationships between mood and dissociation.120 Somatoform dissociation, the lack of integration of somatic experiences, functions and responses, has been much less studied.

Alexithymia does not appear to be a particularly helpful marker, particularly once measures of affective state are taken into account.51 ,121 This may be because alexithymia is conceptualised as broadly and perhaps measurement of specific subtypes may be of more benefit. What may be more relevant to treatment formulations may be findings suggesting increased vigilance to social threat stimuli,98 in a manner also seen in post-traumatic stress disorder, and therefore informative for psychological treatment approaches.

Findings of agoraphobic avoidance in PNES patients49 and of heightened fear sensitivity to a range of stimuli122 lend support to a fear avoidance model.123 ,124 Other potential clinical models of PNES might include one where the seizure related symptoms reported by patients relate to paroxysmal episodes of arousal which act as the underlying mechanism behind their seizures.49 Arousal symptoms may also be relevant in motor conversion disorder.83 With regard to potential illness perpetuating factors, there has been only limited application of Leventhal's self-regulation theory125 to PNES via the study of illness representations.126 ,127 The self-regulation model is the dominant model for considering health beliefs and suggests that an individual's response to health advice will depend on the extent that the advice fits into their own illness model based hierarchically around three constructs: (i) ‘representation’ of the illness which guides (ii) ‘coping’ responses and (iii) ‘appraisal’ of the success or failure of the coping responses. A major aim of cognitive behavioural treatment is to change any maladaptive illness representations.109

Treatment of functional neurological symptoms

There have been early attempts to look critically at the consultation as a form of treatment using a range of methodologies (see PNES above). The use of randomised trials in this context is particularly exciting, answering basic questions such as whether to withdraw anticonvulsant medication in PNES patients?128 The results were reassuring: no adverse events, slight improvement in attack frequency and a more convincing reduction of healthcare resource utilisation.

A Cochrane review found that sending clinic letters from secondary care (often following joint review with a psychiatrist) recommending structured primary care appointments with regular review, ‘hands-on’ examination of physical complaints and the avoidance of tests or referrals reduced healthcare spending in patients with the most chronic somatisation disorders. However, the effect of this approach on patients' symptoms or functioning remains uncertain.129

A number of self help resources have been developed to aid communication, including web resources available at and, and a self-help manual Overcoming functional neurological symptoms: a five areas approach (Williams et al, Hodder Arnold, London).


While there is evidence that pharmacological therapy (especially with antidepressant drugs) can alleviate functional symptoms in general, few trials have specifically examined neurological symptoms.130–132 A small randomised controlled trial (RCT) of antidepressant medication in PNES131 showed a non-significant trend towards a beneficial effect. While the trial demonstrated that RCT methodology could be applied in PNES, it also demonstrated some of the associated difficulties.

Psychological therapies

Many consider psychotherapy the mainstay of treatment but despite such backing the number of appropriately powered studies in this area is small.132 ,133 Recently, a number of randomised controlled studies have provided class 2–3 evidence of the effectiveness of such interventions for patients with FNS and PNES, at least in the short-term.124 ,134

Treatments have targeted a range of potentially modifiable factors, including illness perceptions,107 ,124 ,135–137 previous trauma and emotional processing,138 or health utilisation behaviours.129 A range of different psychotherapeutic approaches have been described. Most studies focusing on functional somatic symptoms (including FNS and PNES) often in primary care settings have used variants of reattribution, typically with modest success.136 ,139 An RCT of a guided four session self-help programme based on a cognitive behavioural approach using a ‘functional’ model of explanation in neurological outpatients improved health related quality of life in the short term and symptoms in the longer term.140 The absolute improvement in patients ‘better’ or ‘much better’ in this trial was modest at 13% (number needed to treat of 8) but, for a brief intervention, was encouraging for future studies.

In terms of more formal outpatient psychotherapy, the evidence for cognitive behavioural therapy approaches is best,124 but it remains uncertain whether early cognitive behavioural therapy associated improvements in terms of symptoms and quality of life can be maintained in the longer term. Other approaches, for instance the use of approaches based on psychodynamic interpersonal therapy, have also been described and tested in uncontrolled studies.55 ,138 ,141

Other treatment modalities

A small series of studies in functional paralysis142 ,143 (plus one case report in aphonia144) report dramatic benefit from the application of repetitive transcranial magnetic stimulation to the motor cortex. None of these studies has a placebo control, and the immediate dramatic benefit is in marked contrast with that seen in repetitive transcranial magnetic stimulation applied to organic movement disorders. It is possible that the experience of the paralysed limb moving when transcranial magnetic stimulation is applied at sufficient intensity over the motor cortex is an important driver of benefit rather than any specific neuromodulatory effect, but further studies with adequate placebo control are needed. Even as a placebo intervention, the apparent dramatic effectiveness, if confirmed, may justify its use in selected patients.

Inpatient treatment programmes have also been described although controlled studies are only available for certain aspects of inpatient treatments.145 ,146

Disappointingly, there is a paucity of evidence examining rehabilitative approaches that seem intuitively sensible, such as physiotherapy,147 and this would seem a priority area for future trials.


There are encouraging signs of a renaissance of interest in functional symptoms. A more integrative approach to neuroscience that avoids outdated dualistic assumptions while not denying the potential importance of social and cultural factors, coupled with new technological developments for investigating brain function, is offering new and exciting methodologies for exploring mechanisms underlying symptom production. Although in its infancy as an approach to these disorders, they appear to offer real potential. In parallel, the field has moved forward using proper methodologies to explore basic clinical questions of interpretation of clinical signs, how to communicate with patients and how to investigate them. There remain however many under answered questions, not least how to treat such patients. It is nevertheless encouraging that clinicians are starting to speak about these issues, and not just in the UK, and that seems a necessary first step to getting the questions answered.


View Abstract


  • * UK-FNS attendees: Richard J Brown, Alan Carson, Trudie Chalder, Anthony David, Roderick Duncan, Mark Edwards, Laura Goldstein, Richard Grunewald, Stephanie Howlett, Eileen Joyce, Richard Kanaan, John Mellers, Tim Nicholson, Maria Otto, Wendy Philips, Markus Reuber, Hugh Rickards, Anette-Eleonore Schrag, Michael Sharpe, Sharon Smith, Jon Stone, Valerie Voon, Charles Warlow.

  • Funding The UK FNS meeting was supported by an unrestricted educational grant from UCB Pharma.

  • Competing interests VV is a Wellcome Trust Clinical Fellow.

  • Provenance and peer review Commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.