Article Text

Download PDFPDF

Research paper
Resting cortical PET metabolic changes in psychogenic non-epileptic seizures (PNES)
  1. M Arthuis1,2,
  2. J A Micoulaud-Franchi2,
  3. F Bartolomei1,3,4,
  4. Aileen McGonigal1,3,4,
  5. E Guedj5,6,7
  1. 1Service de Neurophysiologie Clinique, Hôpital de la Timone, Assistance Publique des Hôpitaux de Marseille, Marseille, France
  2. 2Pôle de Psychiatrie, Centre Hospitalier Universitaire de Sainte-Marguerite, Marseille, France
  3. 3Institut de Neurosciences des Systèmes, INSERM UMR 1106, Marseille, France
  4. 4Aix Marseille Université, Faculté de Médecine, Marseille, France
  5. 5Service Central de Biophysique et Médecine Nucléaire, Hôpital de la Timone, Assistance Publique des Hôpitaux de Marseille, Marseille, France
  6. 6Aix-Marseille Université, CERIMED, Marseille, France
  7. 7Aix-Marseille Université, CNRS, UMR7289, INT, Marseille, France
  1. Correspondence to Dr Aileen McGonigal, Service de Neurophysiologie Clinique, Hôpital de la Timone, Assistance Publique des Hôpitaux de Marseille, 264 Rue St Pierre Marseille 13005, France; aileen.mcgonigal{at}univ-amu.fr

Abstract

Background The pathophysiology of psychogenic non-epileptic seizures (PNES) is poorly understood. Functional neuroimaging data in various functional neurological disorders increasingly support specific neurobiological dysfunction. However, to date, no studies have been reported of positron emission tomography (PET) in patients presenting with PNES.

Methods Sixteen patients being evaluated in a specialist epilepsy centre underwent PET with 2-deoxy-2-[fluorine-18]fluoro-d-glucose (18FDG-PET) because of suspected intractable epileptic seizures. However, in all patients, the diagnosis was subsequently confirmed to be PNES with no coexisting epilepsy. 18FDG-PET was also performed in 16 healthy controls. A voxel by voxel intergroup analysis was performed to look for significant differences in interictal (resting state) cerebral metabolism. In addition, metabolic connectivity was studied using voxel-wise inter-regional correlation analysis.

Results In comparison to group analysis of healthy participants, the group analysis of patients with PNES exhibited significant PET hypometabolism within the right inferior parietal and central region, and within the bilateral anterior cingulate cortex. A significant increase in metabolic correlation was found in patients with PNES, in comparison to healthy participants, between the right inferior parietal/central region and the bilateral cerebellum, and between the bilateral anterior cingulate cortex and the left parahippocampal gyrus.

Conclusions To the best of our knowledge, this is the first study describing FDG-PET alterations in patients with PNES. Although we cannot exclude that our data reflect changes due to comorbidities, they may indicate a dysfunction of neural systems in patients with PNES. Hypometabolism regions might relate to two of the pathophysiological mechanisms that may be involved in PNES, that is, emotional dysregulation (anterior cingulate hypometabolism) and dysfunctional processes underlying the consciousness of the self and the environment (right parietal hypometabolism).

Trial registration number NCT00484523.

  • CEREBRAL METABOLISM
  • NEUROPSYCHIATRY
  • PET, FUNCTIONAL IMAGING
  • HYSTERIA

Statistics from Altmetric.com

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.

Introduction

Psychogenic non-epileptic seizures (PNES) are defined as abrupt paroxysmal changes in behaviour or in consciousness without electroencephalographic modifications.1 Video EEG (vEEG) with a recording of habitual events is considered as the gold standard for diagnosis.2 About 5–20% of patients presenting to an epilepsy unit for vEEG monitoring are diagnosed with PNES, and 20–30% of intractable seizures are finally diagnosed as PNES.3

Despite the fact that PNES are now well described, understanding their pathophysiology remains a challenge.1 In terms of psychiatric classification, PNES are currently included within the category of conversion disorder as defined by the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition, Text Revision (DSM-IV-TR) and DSM-5,4 or recently within the category of functional neurological symptom disorders as defined by DSM-5.4 While these disorders were conceptualised by Freud as the conversion of a psychic conflict to a somatic symptom,5 current research approaches are increasingly directed towards understanding the relation between functional neurological disorders (FND) and specific neurobiological impairment (SPN).1

There is increasing interest in functional neuroimaging studies in psychiatric disorders6 as well as in functional neurological disorders,7 in order to search for neural correlates, but so far relatively few studies dealing with PNES have been reported.8–10 However, an association of PNES with physical brain dysfunction appears to be likely based on the frequent existence of electroencephalographic and radiological abnormalities in these patients.11

In order to increase our knowledge about the possible neuroanatomical substrates of PNES, we report the results of a whole-brain voxel-based analysis of cerebral glucose metabolism and of metabolic connectivity, using interictal positron emission tomography with 2-deoxy-2-[fluorine-18]fluoro-d-glucose (18FDG-PET) in patients with PNES, in comparison to healthy participants of similar age and gender.

Material and methods

Population

From our database of adult patients who underwent an inpatient vEEG recording in one of the two epilepsy units of Marseille, France (Henri Gastaut Hospital or La Timone hospital) for diagnostic evaluation from 2003 to 2014, 36 patients with severe PNES underwent an 18FDG-PET scan for the indication at that time of suspected pharmacoresistant epilepsy. All patient data were extensively reviewed. Of these 36 patients, 20 were excluded from the present study because of the coexisting epilepsy, presence of brain lesion on cerebral MRI and/or possible confounding mental disorders (ie, anxiety disorder, current major depressive episode, bipolar disorder or psychotic disorder). This study therefore retrospectively included 16 patients with PNES. All patients underwent long-term (>48 h) vEEG recordings with a recording of habitual events, in order to distinguish between epileptic seizures and non-epileptic seizures. No patient had clear epileptiform abnormalities on EEG.

The patients were examined both by a neurologist and a psychiatrist, according to the International League Against Epilepsy (ILAE) guidelines.2 Beside clinical examination and vEEG, all had standard blood tests and cerebral MRI. An experienced psychiatrist carried out psychiatric evaluations, with diagnosis being based on the DSM-IV-TR criteria. However, no standardised psychiatric interview was performed. The consulting psychiatrist judged whether a mental disorder was present. Therefore, the diagnostic standard for this study was expert opinion. Beside clinical examination and vEEG, all had standard blood tests and cerebral MRI.

Age, gender, age at onset, PNES frequency and delay to diagnosis were documented. After receiving a detailed description of the study, participants gave their written informed consent. This study was conducted in accordance with the Declaration of Helsinki and French Good Clinical Practices. The data collection was approved by the Commission nationale de l'informatique et des libertés (CNIL number: 1223715).

18FDG-PET data obtained from patients with PNES were compared with data from 16 healthy participants of similar age and gender (33.5 years±9.3, p=0.93; 11 women, p=0.69). No statistical difference was found between the patient group and the control group in terms of hemispheric dominance (p=0.626). They were free from neurological/psychiatric disease and cognitive symptoms, and had normal brain MRI. Informed consent was obtained with a protocol approved by the local ethics committee and conforming to the Declaration of Helsinki on human investigation (registration number of the clinical trial: NCT00484523).

Clinical analysis of PNES

All vEEG recordings were studied to characterise clinical features of PNES, including duration of the event. The presence of altered level of contact, in the form of unresponsiveness to verbal commands and to environmental cues, was assessed as follows: 0, no loss of contact; 1, partial alteration of contact; and 2, complete alteration of contact. Semiology of PNES was defined according to a recent classification,12 in which five classes are described as follows: class 1, dystonic attacks with primary gestural activity; class 2, paucikinetic attacks with preserved responsiveness; class 3, pseudosyncope; class 4, hyperkinetic prolonged attacks; and class 5, axial dystonic prolonged attacks.

Brain 18FDG-PET

The interictal brain metabolism was studied in all patients, under the same conditions as in healthy participants. A PET scan was performed using an integrated PET/CT camera (Discovery ST, GE Healthcare, Waukesha, Wisconsin, USA), with a 6.2 mm axial resolution, allowing 47 contiguous transverse sections of the brain of 3.27 mm thickness. One hundred and fifty MBq of 18FDG were injected intravenously in the awake and resting state, with eyes closed, in a quiet environment. Image acquisition started 30 min after injection and ended 15 min later. Images were reconstructed using the ordered subsets expectation maximisation algorithm, with 5 iterations and 32 subsets, and corrected for attenuation using the CT transmission scan.

A voxel-by-voxel intergroup study was then performed using SPM8 (Wellcome Department of Cognitive Neurology, University College, London, UK), running on Matlab (Mathworks Inc, Sherborn, Massachusetts, USA). The PET images were spatially normalised onto the Montreal Neurological Institute (MNI) atlas by using a 12-parameter affine transformation, followed by non-linear transformations and a trilinear interpolation. The dimensions of the resulting voxel were 2×2×2 mm. The images were then smoothed with a Gaussian filter (8 mm FWHM) to blur individual variations in gyral anatomy and to increase the signal-to-noise ratio. The ‘proportional scaling’ routine was used to control for individual variation in the global brain metabolism.

Significant changes in metabolism (hypometabolism or hypermetabolism) were first searched at the group level, comparing patients to healthy participants, using age and gender as a nuisance variable. In addition, metabolic connectivity from significant cluster(s) previously identified in the between-group comparison was studied using voxel-wise inter-regional correlation analysis (IRCA), as previously described.13 Briefly, mean values of extracted metabolic cluster(s) were used as covariates to find regions showing significant voxel-wise positive/negative correlations across participants and between groups, and using the same nuisance variables as previously described in our first SPM analysis.

All SPM maps were thresholded at the voxel level using p<0.001, and with a cluster extent of at least eight voxels (twice the FWHM of the Gaussian filter). Anatomical localisation of the most significant voxels was then identified by Talairach Daemon (http://ric.uthscsa.edu/projects/talairachdaemon.html), and found cluster(s) extracted to search for correlation with clinical characteristics.

Correlation between clinical data and 18FDG-PET findings

Sociodemographical characteristics (age, age of onset and time from onset of seizure disorder) and PNES characteristics (seizure frequency and duration of PNES) were statistically correlated with 18FDG-PET data. We used SPSS software and matched all the clinical data described above to the PET metabolism of the found cluster(s). We calculated the regression coefficient using the Spearman correlation test with a statistical threshold p<0.05. In order to compare metabolic clusters to subtypes of seizures and to the level of contact as defined above, we used the Analysis of Variance (ANOVA) method.

Results

Clinical data

Sociodemographical and clinical data are shown in tables 1 and 2. The majority of the sample was female (75%) with an average age of 33.1 years. The mean age at onset was 25 years and the median delay to diagnosis from onset of illness was 9.94 years (±9.61 years). The average frequency was 8.23 PNES per month (±11.22). The median duration of PNES was 3.67 min (±4.99). The level of contact was completely altered during PNES in the majority of patients (62.5%). The most frequently observed PNES semiological subtypes were 1, 2 and 3.

Table 1

Sociodemographical and PNES characteristics of patients with PNES (n=16)

Table 2

Level of contact and Hubsch classification of psychogenic non-epileptic seizures in the study (N=16)

18FDG-PET findings

In comparison to the group analysis of healthy participants, the group analysis of patients with PNES exhibited a significant PET hypometabolism within the right inferior parietal and central region (Talairach coordinates in mm: 63, −19, 38; BA6, BA3, BA1; p-voxel <0.001, T-score=5.21; k=464; figure 1), and within the bilateral anterior cingulate cortex (Talairach coordinates in mm: −2, 30, 19; BA24; p-voxel <0.001, T-score=4.28; k=486; figure 2). No significant hypermetabolism was found in patients with PNES, in comparison to healthy participants.

Figure 1

Hypometabolism in the right inferior parietal and central region in patients with psychogenic non-epileptic seizures, in comparison to healthy participants (p-voxel <0.001).

Figure 2

Hypometabolism in the bilateral anterior cingulate cortex in patients with psychogenic non-epileptic seizures, in comparison to healthy participants (p-voxel <0.001).

The metabolic connectivity of these two clusters was further studied in patients with PNES, in comparison to healthy participants. A significant increase in metabolic correlation was found in patients with PNES, in comparison to healthy participants, between the right inferior parietal/central region and the bilateral cerebellum (Talairach coordinates in mm: 2, −82, −16; p-voxel <0.001, T-score=4.59; k=447; figure 3), and between the bilateral anterior cingulate cortex and the left parahippocampal gyrus (Talairach coordinates in mm: −26, −32, −12; BA36; p-voxel <0.001, T-score=4.59; k=153; figure 4). No significant decrease in metabolic correlation was found for these two clusters in patients with PNES, in comparison to healthy participants.

Figure 3

Increase in metabolic connectivity in patients with psychogenic non-epileptic seizures, in comparison to healthy participants, between the right inferior parietal/central region and the bilateral cerebellum (p-voxel <0.001).

Figure 4

Increase in metabolic connectivity in patients with psychogenic non-epileptic seizures, in comparison to healthy participants, between the bilateral anterior cingulate cortex and the left parahippocampal gyrus (p-voxel <0.001).

Correlations between 18FDG-PET scan data and clinical features

For both hypometabolic clusters (the right inferior parietal/central region and the bilateral anterior cingulate cortex) and connectivity clusters (the cerebellum and the left parahippocampal gyrus), there was no significant correlation between PET metabolism and sociodemographical characteristics (age, age of onset and time from onset of seizure disorder) and PNES characteristics (seizure frequency and duration of PNES; table 1). No significant difference was found between hypometabolic clusters and subtypes of PNES (F=1.303, p=0.328 for the right inferior parietal/central region and F=0.391, p=0.811 for the bilateral anterior cingulate cortex); or the level of contact (F=0.601, p=0.724 for the right inferior parietal/central region and F=0.609, p=0.719 for the bilateral anterior cingulate cortex).

Discussion

We report the results of interictal (resting state) 18FDG-PET metabolic alteration in a sample of 16 patients with PNES using whole-brain voxel-based analysis of cerebral metabolic of glucose and metabolic connectivity. Group analysis indicated two significant areas of hypometabolism: one in the right inferior parietal/central region, and the other in the bilateral anterior cingulate cortex. Moreover, increased metabolic correlation was found between these regions and the cerebellum and the left parahippocampal gyrus, respectively. These metabolic changes obtained in comparison with healthy controls may reflect interictal disturbances in brain networks underlying PNES.

The limitations of our study include its retrospective nature, heterogeneous patient sample and relatively small group size. Patients were selected because of intractable seizures (originally considered to be of epileptic origin but revised to a diagnosis of PNES) requiring inpatient assessment in a specialist epilepsy unit, potentially a source of bias towards the more severe end of the PNES spectrum. Parameters relating to dissociative traits, emotional processing and the presence of psychiatric comorbidities (anxiety, depression or presence of post-traumatic stress disorder (PTSD)) were not formally measured in all patients and the degree of heterogeneity in this regard is thus not documented; the presence of these would potentially constitute an important confounding factor in this patient group.14 Moreover, the sample size is too small to permit a meaningful subgroup analysis according to the semiological features12 or psychological profile (eg, dissociative scale9 ,10). However, with respect to other variables, for instance, sex ratio, mean age, duration, seizure type and altered level of contact,15 our sample was not inconsistent with the wider PNES population.

To the best of our knowledge, these results are original since this is the first PET study of PNES. In terms of research using other imaging modalities in PNES, a recent MRI study of 20 participants using both cortical thickness analysis and voxel-based morphometry found neuroanatomic correlates in the form of altered grey matter volume reduction (cortical atrophy) in motor and premotor regions in the right hemisphere including the anterior cingulate gyrus, and in the bilateral cerebellum as well as precuneus.16 Comorbid depression was found to correlate with the degree of atrophy of premotor regions. The authors proposed that alterations in sensorimotor circuits might underlie the propensity to develop PNES, perhaps representing a shared neural substrate with paroxysmal movement disorders. The fact that the right hemisphere was preferentially affected, as observed in this study, was also discussed in terms of hemispheric specialisation for processing of negative emotional experience in relation to the body's state.17 A functional MRI (fMRI) study9 including 11 patients with PNES, all of whom had significantly elevated dissociation scores, showed stronger connectivity in a global network involving the insula, inferior frontal gyrus, parietal cortex and precentral sulcus. Authors found a left-sided predominance of motor cortex connectivity changes. Connectivity values in this study were correlated to dissociation scores, thus leading the authors to hypothesise that this abnormal connectivity could be the neural correlate of dissociation. However, since in our more heterogeneous series no measure was made of dissociative tendency, it is difficult to directly compare these results of connectivity obtained via fMRI and the PET resting state. Finally, another recent study described fMRI alteration in functional and structural connectivity in 20 patients with PNES,18 with widespread topological alteration including the frontal cortex, sensorimotor cortex, cingulate gyrus, insula and occipital cortex.19 Such widespread abnormalities argue in favour of multiple and multilevel cognitive processes, including those governing attention, sensorimotor control and emotion processing, as being involved in the pathophysiology of PNES. Why this study should show a somewhat different pattern of altered brain structures remains to be elucidated; further larger, well-controlled studies may help to answer this question.

Although there is a growing literature describing abnormal functional imaging data in functional neurological disorders, notably conversion paresis,7 ,20–22 comparing these data with the present results is somewhat problematic since, by definition, patients with PNES present behavioural disturbance and/or alteration of consciousness that is paroxysmal rather than chronic, and are usually asymptomatic at the time when scanning is performed, unlike the majority of patients with conversion paresis. Paroxysmal movement disorders, including tremor,23 might arguably be a more reasonable group for comparison with PNES. In PNES as in other FNDs, interictal cognitive and psychiatric disturbances are described,1 a possible confounding factor when considering the significance of functional imaging abnormalities detected in the resting state period. Indeed, the resting state analysis of brain function in general is noted for its various pitfalls.24

While several studies of conversion paresis have highlighted hyperactivation of the anterior cingulate cortex during activation tasks including movement execution, on the other hand hypoactivation of the anterior cingulate cortex has been described as a robust finding in various psychiatric conditions including heightened anxiety25 and PTSD.26 Psychiatric comorbidities in PNES such as depression, anxiety, personality disorder and PTSD are well described.27 ,28 In addition, dysfunctional emotion regulation may be one of the key factors in PNES pathophysiology.29 ,30 Thus, the hypometabolism of the anterior cingulate region found in our patients might be related to psychiatric comorbidity rather than a specific effect related to PNES.

Impairment of consciousness and feelings of loss of control are key symptoms in PNES, but the nature of disruption of consciousness in this context is a complex and poorly understood area.31 A bidimensional model incorporating both the level and content of consciousness has been proposed for epileptic seizures32 and applied to PNES.33 The dynamic and variable nature of altered consciousness in PNES, and its differences from that occurring in epileptic seizures, have been highlighted.33 In recent years, various frameworks of the neural basis of consciousness have been proposed, notably the global workspace theory of consciousness describing the frontoparietal network as an essential basis for conscious access.34 Alteration of frontoparietal networks is thought to be specifically associated with conditions affecting awareness without affecting vigilance such as sleepwalking, epileptic seizures or vegetative states.35 The global workspace theory has been investigated in epileptic seizures using in-depth EEG studies36 ,37 but has not so far been considered in the context of functional neurological disorders. Given the predominance in our PNES series of an altered level of contact as a main semiological feature, and given the known role of the parietal cortex in the neural basis of consciousness, it is of interest in this perspective to reflect on the possible significance of the right inferior parietal cortex hypometabolic region found in this study. Interestingly, recent works using fMRI showed right temporoparietal hypoactivation in conversion tremor as compared with voluntary mimicked tremor.23 ,38 This was interpreted as support for the hypothesis of dysfunctional processing of motor intention and loss of self-agency underlying the development of psychogenic movement disorder. In addition, the inferior parietal region has been proposed as a key region for integration of bodily self-consciousness.39 ,40 The data presented here suggest dysfunction of a brain structure (parietal lobe) that has been implicated in both self-agency and mechanisms of consciousness, and this raises the possibility of a role for parietal regions in the alteration of consciousness in PNES. Whether the resting state parietal dysfunction is a true feature in patients with PNES requires to be evaluated in larger studies, but this hypothesis could offer a framework for future investigation. We speculate that, while methodological limitations require a cautious interpretation of results, the specific zones of hypometabolism identified might relate to two of the pathophysiological mechanisms that may be involved in PNES, that is, emotional dysregulation (anterior cingulate hypometabolism) and dysfunctional processes underlying the consciousness of self and the environment (right parietal hypometabolism).

Conclusion

This is the first study describing FDG-PET cerebral resting state metabolic alterations in patients with PNES. Although we cannot exclude that our data reflect changes due to comorbidities, they may indicate dysfunction of specific neural systems in patients with PNES, which merits further investigation. In particular, the prospective methodology and sufficient numbers of patients, thus allowing the assessment of effects of potential comorbidities and confounders, will be essential aspects of future studies.

References

Footnotes

  • Contributors FB and EG conceived and formulated the project. MA analysed all clinical data and wrote the first draft. JAM performed psychiatric evaluations on all patients and contributed to the first draft. AM analysed video-EEG data and wrote the final draft of the paper. EG analysed all PET data including statistical analysis. All authors contributed to the discussion and reviewed the final manuscript.

  • Funding This work was supported by INSERM (Centre d'Investigation Clinique, CIC, Hôpital de la Conception, Marseille), and AP-HM (PHRC 2007/09).

  • Competing interests None.

  • Ethics approval Commission nationale de l'informatique et des libertés (CNIL number: 1223715).

  • Provenance and peer review Not commissioned; externally peer reviewed.