Neuroanatomic correlates of psychogenic nonepileptic seizures: A cortical thickness and VBM study


  • Angelo Labate,

    1. From the Institute of Neurology, University Magna Græcia, Catanzaro, Italy
    2. Neuroimaging Research Unit, Institute of Neurological Sciences, National Research Council, Germaneto (CZ), Italy
    Search for more papers by this author
  • Antonio Cerasa,

    1. Neuroimaging Research Unit, Institute of Neurological Sciences, National Research Council, Germaneto (CZ), Italy
    Search for more papers by this author
  • Marco Mula,

    1. Department of Clinical and Experimental Medicine, Amedeo Avogadro University and Division of Neurology, Maggiore Hospital, Novara, Italy
    Search for more papers by this author
  • Laura Mumoli,

    1. From the Institute of Neurology, University Magna Græcia, Catanzaro, Italy
    Search for more papers by this author
  • Maria Cecilia Gioia,

    1. Institute of Neurological Sciences, National Research Council, Piano Lago, Mangone, Cosenza, Italy
    2. Evolutionary System Group, University of Calabria, Arcavacata di Rende, Cosenza, Italy
    Search for more papers by this author
  • Umberto Aguglia,

    1. From the Institute of Neurology, University Magna Græcia, Catanzaro, Italy
    Search for more papers by this author
  • Aldo Quattrone,

    1. From the Institute of Neurology, University Magna Græcia, Catanzaro, Italy
    2. Neuroimaging Research Unit, Institute of Neurological Sciences, National Research Council, Germaneto (CZ), Italy
    Search for more papers by this author
  • Antonio Gambardella

    1. From the Institute of Neurology, University Magna Græcia, Catanzaro, Italy
    2. Institute of Neurological Sciences, National Research Council, Piano Lago, Mangone, Cosenza, Italy
    Search for more papers by this author

Address correspondence to Dr Angelo Labate, Cattedra ed U.O. di Neurologia, Università degli Studi “Magna Graecia,” Campus Universitario Germaneto, Viale Europa, 88100 Catanzaro, Italy. E-mail:


Purpose:  Psychogenic nonepileptic seizures (PNES) are among the most common clinical manifestations of conversion disorder and consist of paroxysmal behavior that resembles epileptic seizures. Preliminary data from functional neuroimaging studies gave plausible evidence that limbic circuits and sensorimotor cortex might be engaged in conversion disorder. Nonetheless, no advanced magnetic resonance imaging (MRI) studies have focused on patients with PNES.

Methods:  We enrolled 20 consecutive patients in whom the diagnosis of PNES was based on ictal video–electroencephalography (EEG) of the habitual episodes and 40 healthy subjects matched for age and sex All patients underwent a formal neuropsychological investigation and a neuropsychiatric assessment. All of the patients also underwent two distinct morphologic whole-brain MR measurements, voxel-based morphometry (VBM), and cortical thickness analysis, in a multimethod approach.

Key Findings:  None of the patients had serious medical or neurologic illness, substance abuse, or psychotic disorder, or were taking antipsychotic drugs. VBM and cortical thickness analyses in the patients with PNES revealed abnormal cortical atrophy of the motor and premotor regions in the right hemisphere and the cerebellum bilaterally. We also observed a significant association between increasing depression scores and atrophy involving the premotor regions.

Significance:  The results of this study illustrate that motor and premotor regions in the right hemisphere and the cerebellum bilaterally play an important role in the pathogenesis of PNES and that these structures are correlated with depressive symptoms. Our findings suggest a multistep model in the pathogenesis of PNES, in which the phenomenology is driven by psychological factors interacting with specific biologic abnormalities.

Conversion disorder represents the unconscious production of neurologic symptoms that cannot be attributed to organic brain injury and are commonly associated with emotional stressors or conflicts (Voon et al., 2010a). Psychogenic nonepileptic seizures (PNES) are among the most common clinical manifestations of such a disorder and consist of paroxysmal behavior that resembles epileptic seizures (Lesser, 1996; Krumholz, 1999). PNES are often associated with chronic disability and an increased risk of morbidity and mortality (Lesser, 1996) and represent more than 20% of patients referred to epilepsy centers for having refractory recurrent seizures (Bowman & Markand, 1996; Krumholz, 1999; Benbadis, 2010).

Although there have been a plethora of studies to explain conversion disorders as well as PNES in the context of psychogenic stress or emotional disturbances, recent functional neuroimaging studies have started to shift our understanding of these deficits from a psychological and psychodynamic model to a neurobiologic model (Marshall et al., 1997; Vuilleumier et al., 2001; Vuilleumier, 2005; Montoya et al., 2006). So far, preliminary data from neuroimaging studies provide information on possible networks engaging limbic regions and sensorimotor cortices in conversion disorders, especially in the prefrontal cortex and anterior cingulate cortex (Marshall et al., 1997; Trimble, 2004; Vuilleumier, 2005). It has been hypothesized that conversion disorder affects the abnormal inhibition of motor systems by limbic regions or impairments of motor conceptualization (Marshall et al., 1997; Halligan et al., 2000; Spence et al., 2000; Ward et al., 2003; Burgmer et al., 2006; de Lange et al., 2007). Furthermore, a recent functional magnetic resonance imaging (MRI) study illustrated that patients with conversion disorder had greater functional connectivity between the right amygdala and the right supplementary motor area, providing direct support for a model of conversion disorder based on abnormal limbic–motor interactions (Voon et al., 2010a,b).

However, these latter studies are affected by a number of limitations such as the focus on psychogenic movement disorders, especially with negative symptoms (paralysis or anaesthesia), and the heterogeneity of the investigated sample (Vuilleumier et al., 2001; Vuilleumier, 2005). Moreover, despite their frequency and relevance, patients with PNES have received little attention to date and, as far as we are aware, no structural advanced neuroimaging studies have been focused on such a population.

Herein, we investigated structural brain correlates of patients with PNES. We hypothesised that the prefrontal and sensorimotor circuits and their connections may also be involved in PNES. For this reason, we combined two distinct morphologic whole-brain MR measurements: voxel-based morphometry (VBM) (Ashburner & Friston, 2000; Labate et al., 2008, 2010) and cortical thickness (Freesurfer) (Fischl & Dale, 2000) in a multimethod unbiased approach. We used both methods because they may provide complementary information that cannot be obtained by only using one of the methods alone. Indeed, whereas cortical thickness analysis captures the columnar architecture of the cortex, VBM provides a general measure of gray matter (GM) volume, which conflates the contributions of thickness and surface. Therefore, we wish to demonstrate cortical thinning with Freesurfer in the same regions as the volume losses seen with VBM and possible converging results would strengthen our observation in such a particular population.


Patients and controls

From September 2007 to August 2010, we prospectively identified 23 consecutive patients with PNES at two twin University Epilepsy Centers located in Catanzaro and Reggio Calabria, Calabria, Southern Italy. Three of the patients refused to participate; therefore, 20 were included in the study: 11 women and 9 men with a mean age (± standard deviation, SD) of 36.7 ± 13.5 years (range 17–58 years). The diagnosis of definite PNES was made when patients with indicative clinical history had spontaneous seizures recorded with videoelectroencephalography (EEG) or habitual attacks provoked by a suggestive colored patch applied over the forehead while under video-EEG surveillance (Lancman et al., 1994). All attacks in a context of negative ictal EEG were considered typical of habitual seizures by seizure witnesses. PNES were characterized by stereotyped motor phenomenon. All patients included in the study underwent long-term video-EEG monitoring to record at least two stereotyped spontaneous or provoked nonepileptic events. Furthermore, the entire group was studied based on a protocol routinely used for patients with epilepsy throughout awake and sleep-deprived EEG and MRI sequences extensively described elsewhere (Labate et al., 2010, 2011). Only patients with a diagnosis of definite PNES were included in the study. Patients with both epileptic seizures and PNES, and patients with no recorded spontaneous or provoked typical seizures, were not enrolled in the study. None of the patients had serious medical or neurologic illness, substance abuse, substance dependence, history of alcohol abuse, or psychotic disorder or were taking antipsychotic drugs. Furthermore, 40 consecutive right-handed healthy controls matched for age (mean age 36.2 ± 9.8 years) and sex (21 women) with no previous history of neurologic or psychiatric diseases were enrolled from the staff of our University in Catanzaro and the Neuroimaging Research Unit of the National Research Council in Catanzaro, Italy. Neurologic examinations were unremarkable in all patients. None of our patients had a history of epilepsy. A chart review and examination was conducted by a trained epileptologist (A.L.) on all participants to document the normal neurologic history and examination in each individual. All video-EEG recordings followed a standard protocol on an analog, 24-channel system according to the international 10–20 system, including usual activation methods and induction seizure procedures (after approved and signed informed consent) in patients without recorded spontaneous seizures. All EEG studies were reviewed in a double-blinded fashion by two trained EEG experts (A.G. & A.L.) at the time of the patients’ presentations. Neither patients nor healthy controls showed evidence of vascular brain lesions, brain tumor, and/or marked cortical and subcortical atrophies on MRI scan. The study was approved by the University ethics committee. All patients and healthy controls provided informed consent to participate in this study. For the controls, the consent form also included a paragraph on incidental findings and the option to refuse to be informed about any unexpected abnormality.

Neuropsychological and psychiatric assessment

All patients underwent a formal neuropsychological investigation and a psychiatric assessment (CG & MM). Premorbid IQ was evaluated with the Italian version of the National Adult Reading Test (NART) called a Brief Intelligence Test (BIT) (Sartori et al., 1995; Colombo et al., 2000), executive functions with the Modified Card Sorting Test (MCST) (Caffarra et al., 2004), and attention and mental flexibility with the Trail Making Test (TMT) Part A and Part B (Giovagnoli et al., 1996). The psychiatric assessment included the Dissociative Experiences Scale version II (DES-II) (Bernstein & Putnam, 1986), the Somatoform Dissociation Questionnaire-20 (SDQ-20) (Nijenhuis et al., 1996), the Beck Depression Inventory (BDI) (Beck et al., 1961), and the State-Trait Anxiety Inventory Y1 and Y2 (STAI Y1 and Y2) (Spielberger, 1983). Diagnosis of current and lifetime DSM-IV Axis I disorders was made using the Mini International Neuropsychiatric Interview (MINI) Plus version 5.0.0 (Sheehan et al., 1998). The DES-II is a lifetime 28-item, self-rating questionnaire developed specifically as a screening instrument to identify subjects that are likely to have dissociative symptoms. Factor analysis studies have identified three main subscales (Amnesia, Depersonalization/Derealization, and Absorption). A cutoff score of 25 proved useful in identifying severe dissociative disorders (Draijer & Boon, 1993). Since its introduction, the DES has been used in hundreds of dissociation studies; a sophisticated taxonometric analysis of the DES, known as the DES-Taxon score (Simeon et al., 2003), identified eight designated pathologic dissociation items (Waller et al., 1997). The SDQ-20 is a self-rating scale developed to investigate somatic component of dissociation. It follows the original distinction, suggested by Nijenhuis et al., (1996), between negative and positive dissociative symptoms, which are a core element of somatoform disorder. The SDQ-20 discriminates between dissociative and affective disorders (mood and anxiety disorders) and psychotic symptoms, but a cutoff score is not available (Nijenhuis et al., 1996). All questionnaires were administered in a standardized way and in the same sequence in all patients.

Image acquisition

Brain MRI was performed according to our routine protocol with a 1.5-T unit (Signa NV/I; GE Medical Systems, Milwaukee, WI, U.S.A.). Structural MRI data were acquired using a three-dimensional (3D) T1-weighted spoiled gradient echo (SPGR) sequence with the following parameters: repetition time = 15.2 ms; echo time = 6.7 ms; flip angle 15 degrees; matrix size 256 × 256; field of view (FOV) = 24 cm; slice thickness = 1.2 mm. Participants were positioned to lie comfortably in the scanner with a forehead-restraining strap and various foam pads to ensure head fixation. The image protocol was identical for all participants studied.

Voxel-based morphometry analysis

VBM analysis was performed with an optimized protocol (Good et al., 2001) using the Statistical Parametric Mapping2 (SPM2) software ( Briefly, a customised gray matter (GM) template was generated and subsequently used to normalize all of the structural images in native space into the stereotaxic Montreal Neurological Institute (MNI) space. To create the customized GM template, all images were first spatially normalised (16-parameter affine) using the standard MNI template in SPM2. Each normalized image was then segmented and subsequently modulated and smoothed with a 10-mm full-width at half-maximum (FWHM) Gaussian kernel. The normalized, segmented, modulated, and smoothed GM volume maps were analyzed statistically using the general linear model (GLM). Statistical analysis consisted of an analysis of covariance (ANCOVA) with total intracranial volume (ICV) and with age and sex as the covariates of no interest. We applied a whole brain statistical threshold correction using the false-discovery rate (FDR, p < 0.05, 20 contiguous voxels). To evaluate any covariation between GM volume changes and clinical data, we performed a correlation analysis using the multiple regression function of SPM2. All clinical and neuropsychiatric scores were treated as covariates of interest, with total intracranial voulme, age, and sex as confounding covariates. Correlation analyses were performed inside and outside specific regions of interest (ROIs). We identified ROIs as regions that showed the most significant GM changes in patients with PNES compared to controls. To limit the analysis to these target areas, the resulting statistical map was transformed into a binary mask using the marsbar tool (Brett, 1999), which was applied explicitly to compute regression analysis. Within each of these ROIs, corrected p-values for multiple comparisons were used (FDR, p < 0.05). Furthermore, for exploratory purposes, we subsequently performed correlation analyses at a whole-brain level outside the ROIs (FDR, p < 0.05).

Cortical thickness analysis

MRI-based quantification of cortical thickness was performed using Freesurfer (v. 4.05) software package ( This method has been described in detail previously (Dale et al., 1999; Fischl & Dale, 2000). Cortical thickness measurements were obtained by reconstructing representations of the gray–white matter boundary and the cortical surface. The distance between these two surfaces was calculated individually at each point across the cortical mantle. The surface representing the gray–white border was “inflated” (Fischl et al., 1999), differences among individuals in the depth of gyri–sulci were normalized, and each participant’s reconstructed brain was then morphed and registered to an average spherical surface representation that optimally aligned sulcal and gyral features across subjects (Fischl et al., 1999). This spherical morphing procedure was used to map the thickness measurements at each vertex on each participant’s cortical surface into a common spherical coordinate system. Finally, cortical maps were smoothed with a 10-mm FWHM Gaussian kernel.

For each hemisphere, differences in cortical thickness between groups were tested by computing a GLM of the effects of “group” on cortical thickness at each vertex, controlling for the effects of age and sex. An FDR of ≤0.05 was applied to correct for multiple comparisons. Furthermore, a vertex by vertex multiple linear regression analysis was carried out to investigate the relationship between regional cortical thickness and critical clinical symptoms. This analysis was performed within ROIs, selected in the areas where the PNES group also had significantly thinner cortices. These ROIs were mapped back to each individual participant using spherical morphing to find homologous regions across subjects. For exploratory purpose, we subsequently performed correlation analyses also at a whole-brain level outside the ROIs (FDR, p < 0.05).

Statistical analysis

Categorical variables were analysed using the chi-square analysis or Fisher’s exact test, whereas continuous variables were analyzed using the Mann-Whitney test or the Kruskal-Wallis test. Correlations were tested using bivariate two-tailed nonparametric correlation procedures (Spearman’s coefficient). The alpha error was set at 0.05. All statistical analyses were two-sided and conducted using the Statistical Package for Social Sciences (Version 12 for Windows; SPSS Inc., Chicago, IL, U.S.A.).


Demographic features

The mean age at onset of PNES was 21.6 years (standard deviation, SD 9.0). PNES semiology was highly stereotyped in each patient, and mainly with convulsive components, such as tonic, clonic, or bizarre motor manifestations usually involving upper or lower limbs bilaterally. None of our patients manifested nonmotor events such as paralysis, sensory feelings, or unresponsiveness. Four (20%) of 20 patients had a positive family history of psychogenic events. Neurologic examination was unremarkable in all patients. None of our patients had interictal or ictal EEG changes during video-EEG. At least two attacks were on average recorded for each patient. Six patients were excluded: four because had co-occurrence of epileptic seizures and PNES, whereas two patients had no spontaneous or provoked typical seizures recorded. In all patients, the brain MRI was visually analyzed by an epileptologically experienced neuroradiologist, who did not detect any mass lesion such as tumor, cortical dysgenesis, mesiotemporal sclerosis, vascular lesion, malformation, or posttraumatic scars.

Neuropsychological and neuropsychiatric assessment

All scores are summarized in Table 1. DES scores ranged between 0.7 and 95.7, SDQ-20 between 20 and 100, and BDI between 2 and 43. The number of completed categories on the MCST correlated significantly with the SDQ score (ρ = −0.580 p = 0.009), DES total score (ρ = −0.509 p = 0.031), DES-amnesia subscale score (ρ = −0.483 p = 0.042), and DES-absorption subscale score (ρ = −0.531 p = 0.023). The number of perseverative responses on the MCST significantly correlated with SDQ score (ρ = −0.510, p = 0.026) and DES-absorption subscale score (ρ = −0.519 p = 0.027).

Table 1.   Clinical and demographic features of patients with PNES
 N = 20
  1. Neuropsychological and neuropsychiatric assessments.

  2. BDI, Beck Depression Inventory; DES, Dissociative Experiences Scale version II; MCST, Modified Card Sorting Test; STAI Y1 and Y2, State-Trait Anxiety Inventory Y1 and Y2; SDQ, Somatoform Dissociation Questionnaire-20; TMT, Trail Making Test.

  3. Total is more than 100% because of the presence of comorbidities.

Age36.7 ± 13.5
Gender M/F9/11
Age at onset PNES23.4 ± 7.6
Premorbid IQ tot103.9 ± 10.9
Premorbid IQ verb103.3 ± 12.1
Premorbid IQ perf100.0 ± 10.6
TMT A41.8 ± 18.9
TMT B135.6 ± 62.5
TMT B-A97.6 ± 48.1
MCST Cat4.7 ± 1.4
MCST Pers7.2 ± 5.5
MINI diagnosis 
 Any mood disorder (current/lifetime)15 (75%)
 Any anxiety disorder (current/lifetime)19 (95%)
 Conversion disorder7 (35%)
BDI18.2 ± 12
STAI Y130.3 ± 19.4
STAI Y234.9 ± 16.6
SDQ9.0 ± 3.9
DES Tot19.3 ± 20.5
 DES-amnesia16.4 ± 20.1
 DES-Dp/Dr19.2 ± 21.6
 DES-absorption25.7 ± 26.2
 DES taxon13.5 ± 21.9

VBM results

VBM analyses revealed a selective patterns of atrophy reaching the p = 0.05 significance threshold as determined by FDR. In particular, patients with PNES showed (Fig. 1a) a significant loss of GM volume in the bilateral cerebellum, right precentral gyrus (encompassing both the primary motor and premotor cortices), right middle frontal gyrus, anterior cingulate cortex, and supplementary motor area (see Table S1). When the threshold for significance was lowered to p ≤ 0.001 without correction for multiple comparisons, PNES showed additional loci of atrophy in the bilateral postcentral gyrus and left precentral gyrus. No significant findings were detected for the inverse contrast. In addition, we performed a correlation analysis to delineate possible links between GM abnormalities and clinical and neuropsychiatric data. Regression analysis within ROIs showed a single significant result concerning depression scores of patients with PNES. In fact, we found a negative correlation between depression scores and the GM volume of the right dorsal premotor cortex (t-value = 6.21, cluster (k) = 182, PFDR-corr = 0.005; local maxima: x: 30, y: −10, z: 51) (Fig. 2a). No significant correlation was detected when considering other ROIs. Moreover, at a significant FDR correction (p < 0.05), no significant correlation was detected considering voxels outside the ROIs.

Figure 1.

Neurodegenerative patterns in PNES. (a) 3D rendering of GM volume reduction in the cerebral and cerebellar cortices displayed by the PNES group with respect to controls as detected by VBM analysis. The colored bars represent the range of t scores. For display purpose, significances were displayed at an uncorrected threshold of p < 0.001. (b) Whole-brain vertex-wise analysis of cortical thickness. Mean difference maps were generated by aligning and averaging brain MRIs across participants in spherical space to demonstrate the main cortical thickness differences between the two groups at each point on the cortex. Maps are presented on the pial cortical surface that shows regions of significant thinning. Red (t-value = 3.00) and yellow (t-value = 5.00) represent areas where patients with PNES had significantly thinner cortices than healthy controls, with peak differences of almost 0.3 mm. FDR-determined p < 0.05 significance threshold was at t-value = 4.5.

Figure 2.

Correlation analyses. (a) Scatter plot showing the distribution of the mean % signal change (y axis) of all voxels within the right dorsal premotor cortex and the depression scores (x axis) in the patients with PNES as revealed by VBM. The multiple regression analysis of the GM volume extracted from this region shows a significant negative correlation. (b) Illustration of ROIs and correlational plots between depression scores and cortical thickness measurements in the right superior frontal gyrus (A) and right paracentral gyrus (B).

Cortical thickness results

We performed a whole-brain cortical thickness analysis between groups. We found significant thinning (FDR < 0.05) in the right precentral gyrus, right superior frontal gyrus, right precuneus, and right paracentral gyrus (Fig. 1b). When the threshold for significance was lowered to p ≤ 0.001 without correction for multiple comparisons, PNES showed additional loci of pronounced thinning (see Table S2). There were no suprathreshold peaks indicating greater thickness in the PNES group. Moreover, we performed regression analyses in order to investigate whether clinical factors might be associated with the detected abnormal cortical thickness measurements. Results of ROIs analyses are shown in Fig. 2b. Results revealed a significant association between increasing depression scores and the thinning of the right superior frontal gyrus and the right paracentral gyrus (see Table S3). No significant correlation considering other ROIs was detected. Moreover, considering vertices outside ROIs, a significant negative correlation surviving whole-brain FDR correction was detected between: (i) Beck scores and right orbitofrontal sulcus; and (ii) SDQ scores and the left inferior frontal gyrus (pars opercularis) together with the left central sulcus.


Charcot and Briquet were strongly convinced that the so-called “hysteroepilepsy” was a structural brain disorder (Goetz, 1987); nonetheless, the neural mechanisms to explain the seizure manifestation of the conversion disorder have remained unknown, although psychological etiologic factors have been established. Using a multimethod MRI approach, we have demonstrated that, despite a normal MRI standard investigation, patients with PNES show abnormal cortical thinning of the motor and premotor regions in the right hemisphere and the cerebellum bilaterally. We also observed a significant association between increasing depression scores and premotor region atrophy, as almost all patients had current diagnoses of depressive disorders, and depressive and anxiety scores together with other neuropsychological variables were used as covariates in the imaging analysis. Overall, these findings are important, since a better understanding of the underlying dynamic neuroanatomic networks will probably improve both diagnosis and therapy of PNES.

Previous neuroimaging studies revealed similar abnormalities of the motor network, especially of the prefrontal cortex and anterior cingulate cortex, in patients with psychogenic movement disorders. We also found abnormalities in the cerebellum bilaterally, highlighting that the cerebellum is a node in the neural network that underlies the subjective experience of emotion and is also involved in cognitive and emotional functions (Sacchetti et al., 2009). In fact, a recent study in patients with conversion tremor shows a lower functional connectivity between the sensorimotor circuits that include sensorimotor cortices and the cerebellum, and limbic regions (ventral anterior cingulate and right ventral striatum), leading to a potential abnormality of integration of the internal sensory prediction with the actual sensory state (Voon et al., 2010a,b). Therefore, the current findings together with previous functional studies (Vuilleumier et al., 2001; Voon et al., 2010a,b) strengthen the view that the sensorimotor circuit and its connections play an important role in the pathogenesis of conversion disorders regardless of their semiologic features of pseudoepileptic seizures or paroxysmal movement disorders. Moreover, the current findings, especially the observation that more pronounced structural abnormalities were associated with higher depression, are in favor of an adaptive cortical–subcortical plasticity within these regions, giving rise to PNES. It is reasonable, therefore, to hypothesize that there is a multistep model where PNES phenomenology is driven by the psychological factor interacting with specific biologic abnormalities.

It is also interesting to note that none of our identified brain regions correlated with neuropsychological parameters, but only with psychopathological variables, suggesting that cognitive dysfunction is probably secondary to psychopathology rather than being primarily affected. In fact, executive functions were inversely correlated with dissociation scores only (DES total and subscale scores, SDQ) and not with structural MRI measurements. Our findings clearly point out the importance, in PNES, of associated psychopathology that probably represents the fertile ground on which dissociation may lead to a dysfunctional cognitive representation and integration of the endogenous body state. There is a specific vulnerability factor that may play a role in the development of PNES, possibly as an “organic” extra factor. In our opinion, our study brings further evidence that structural brain abnormalities are present in patients with PNES, and suggests that such abnormalities are related to the associated psychopathology that concurs, together with psychogenic causation (e.g., traumatic experiences), in the development of PNES.

In our patients with PNES, the major abnormalities seen in the supplementary motor cortex are consistent with observations indicating that the function of the supplementary motor cortex is implicated in self-initiated action as well as in unconscious motor inhibition. Of interest, epileptic seizures originating in the supplementary motor cortex share many clinical similarities with PNES with difficult differential diagnosis; in either circumstance, seizures are usually explosive at onset, with bizarre bilaterally posturing of the arms or legs, frequent emotional or sexual content, and preserved consciousness (Williamson et al., 1985).

In our patients, the asymmetrical major atrophy observed over the right hemisphere is not fully surprising because different authors have suggested a specific role for the two hemispheres with the left (dominant) hemisphere implicated in defective processing of endogenous somatic percept (Flor Henry et al., 1981), whereas the impairment of the right (nondominant) hemisphere is reflected by a negative bias to emotional experience or a defective mapping of the body state (Ross et al., 1994). Our findings are in keeping with such an integrative approach. In fact, we identified brain regions in the left hemisphere (i.e., left precuneus, left cingulate sulcus, left entorhinal cortex, and left middle frontal gyrus) involved in consciousness and awareness of endogenous somatic percept, whereas the right hemisphere salience (i.e., right superior parietal gyrus, right cingulate sulcus, right superior frontal gyrus, right calcarine sulcus, and right superior frontal sulcus) is mainly related to the representation of the patient’s physical and emotional state.

One important strength of our study involves the employment of two distinct but matching advanced MRI techniques to provide a wider picture of neurodegenerative processes underlying PNES. However, some minor differences have been detected, such as atrophy in the precuneus revealed only by cortical thickness analysis and atrophy of the anterior cingulate cortex revealed only by VBM analysis. That VBM and cortical thickness analyses detected different neurodegenerative patterns is not surprising, since previous works highlighted the different features of these two advanced MRI techniques (Cerasa et al., 2010; Winkler et al., 2010). The advantage of combining these two techniques lies in the complementary nature of the two methods. In fact, VBM provides a mixed measure of cortical GM, including cortical surface area and/or cortical folding as well as cortical thickness; in contrast, cortical thickness analysis has the advantage of providing a quantitative value that represents a physical property of the cortical mantle. Therefore, demonstrating the presence of thinner cortex in the same regions as the volume losses might suggest that the abnormalities seen with VBM are mostly driven by cortical thinning (neuronal loss, gliosis, or neuronal degeneration) and less by an abnormal folding pattern. Nonetheless, the involvement of the precuneus and cingulate cortex in the pathophysiologic pathway underlying PNES deserves further investigations, since these regions are strongly involved in several neuropsychiatric disorders (Cavanna, 2007).

Our results need also to be considered in the context of possible limitations. First, although the published studies involve smaller samples, our sample size may represent a possible limitation because we analyzed only 20 subjects; however, the homogeneity of the investigated sample and the careful clinical and neuropsychiatric evaluations represent considerable strengths of our work. Second, our study could be interpreted in a context of antipsychotic medications or major depressive symptoms that actually contribute to a number of brain structural changes observed in psychiatric patients; however, this does not represent our case because our patients were not taking antipsychotic medications nor had they manifested major depression. Moreover, since previous studies found similar abnormalities in the prefrontal cortex and sensorimotor cortex in patients with depressive symptoms (de Lange et al., 2010; Li et al., 2010), further studies are needed to clarify whether these structural abnormalities are related to conversion disorders per se rather than PNES.


We confirm that we have read the Journal’s position on issues involved in ethical publication and affirm that this report is consistent with those guidelines. None of the authors have any conflict of interest to disclose.


Angelo Labate conceived, wrote, and reviewed the manuscript; Antonio Cerasa conducted the VBM and cortical thickness analyses; Marco Mula conducted the neuropsychiatric tests; Laura Mumoli collected all clinical, electrophysiological, and MRI data; Cecilia Gioia conducted the neuropsychological tests; Umberto Aguglia recruited patients and reviewed the manuscript; Aldo Quattrone reviewed the manuscript; and Antonio Gambardella conceived and reviewed the manuscript.