Neurological Sciences

https://doi.org/10.1007/s10072-019-04108-7

ORIGINAL ARTICLE

Structural brain abnormalities in patients

with psychogenic nonepileptic seizures
Ali A. Asadi-Pooya 1,2 & Maryam Homayoun 1

Received: 20 June 2019 / Accepted: 14 October 2019

# Fondazione Società Italiana di Neurologia 2019

Abstract
Objectives We assessed the relationship between the clinical features of patients with psychogenic nonepileptic seizures (PNES)
and referrals for brain imaging tests. We also hypothesized that some clinical factors might be associated with structural brain
imaging abnormalities in these patients.
Methods In this retrospective study, patients with PNES, who were investigated at Shiraz Comprehensive Epilepsy Center, Iran,
from 2008 until 2019, were studied.
Results One hundred thirty-two patients had a brain magnetic resonance imaging (MRI) study available. Forty-seven patients
(36%) had abnormal finding(s) in their MRI. Age (odds ratio = 1.040, p = 0.02), and comorbid epilepsy (odds ratio = 3.006, p =
0.005) were significantly associated with having an abnormal brain MRI. In a subanalysis, we excluded the patients with
comorbid epilepsy (46 patients). From the remaining 86 patients (with PNES only), 23 patients (26.7%) had abnormal findings
on their MRIs. Common epileptogenic structural brain abnormalities (e.g., tumors, mesial temporal sclerosis, encephalomalacia,
and developmental anomalies) were common in patients with comorbid PNES and epilepsy (in 19 out of 46 patients; 41%), but
not in those with PNES only (in 4 out of 86 patients; 5%) (p = 0.00001).
Conclusion While the evidence is convergent on the relatively high prevalence of structural brain abnormalities in patients with
PNES, the data so far is suboptimal. In order to investigate the significance of structural brain abnormalities in the development of
PNES, future well-designed multicenter studies, which include a large number of patients with a unified methodology of
imaging, are desirable.

Keywords Epilepsy . Imaging . PNES . Psychogenic . Seizure

Introduction most common and important differential diagnoses of epilep-

Psychogenic nonepileptic seizures (PNES) are relatively com-
mon occurrences in epilepsy centers [1]. Pathophysiology of
PNES is still poorly understood [2]. However, association
with head injury has been observed in many studies [3, 4].
Psychogenic nonepileptic seizures are considered to be the
* Ali A. Asadi-Pooya
aliasadipooya@yahoo.com
Maryam Homayoun
maryam.homayoun@gmail.com
1
Neuroscience Research Center, Shiraz Medical School, Shiraz
University of Medical Sciences, Shiraz, Iran
2
Jefferson Comprehensive Epilepsy Center, Department of
Neurology, Thomas Jefferson University, Philadelphia, PA, USA
sy; however, misdiagnosis and diagnostic delay are common
in daily practice. As a result, patients with PNES are at risk of
receiving unnecessary diagnostic procedures and treatments
[e.g., antiepileptic drugs (AEDs)] [5]. Prior to a definite diag-
nosis, patients with PNES are often investigated with
neurodiagnostic imaging(s) that contributes to the high cost
of their care [6]. In this study, we assessed the relationship
between the clinical features of patients with PNES and refer-
rals for brain imaging tests.
On the other hand, previous studies have investigated struc-
tural brain imaging abnormalities in patients with PNES in an
attempt to discover the pathophysiology of this condition
[7–9]. Structural brain imaging abnormalities in patients with
PNES appear to be more prevalent compared with that in the
healthy population [7]. However, no consistent structural im-
aging abnormalities have been reported in previous studies
[7]. Another purpose of this study was to improve on previous

studies on brain imaging abnormalities in a large cohort of
patients with PNES. We hypothesized that some clinical fac-
tors (e.g., seizure severity) might be associated with structural
brain imaging abnormalities in these patients.
Patients and methods
In the current database retrospective study, patients with
PNES, who were investigated at Shiraz Comprehensive
Epilepsy Center at Shiraz University of Medical Sciences,
Iran, from 2008 until 2019, were studied. We always make a
diagnosis of PNES if a detailed clinical history is compatible
with the diagnosis; attacks (seizures) are witnessed by the
epileptologist, showing semiology typical of PNES while on
video-EEG monitoring, and finally, no epileptiform activity is
detected immediately before, during, or after the attack that
has been captured on the ictal video-EEG recording. We also
always obtain a detailed clinical history in order to investigate
the coexistence of epileptic seizures (e.g., presence of other
seizure types, different from what we have captured during
their video-EEG monitoring, if their description is compatible
with epileptic seizures). We review the recorded interictal
EEG carefully to search for any possible epileptiform
discharges.
The epileptologist examines all patients, and if they con-
sent to share their information in the database, it will be used.
We have kept all the data confidential. We routinely do not
perform brain magnetic resonance imaging (MRI) in patients
with PNES; however, in reality, many patients have already
done a brain MRI study before being referred to our center. If
we suspect comorbid epilepsy in a patient, we may perform a
brain imaging study as indicated. The radiologists, who carry
out the brain MR studies, are blind to the final diagnosis of
epilepsy vs. PNES; they know that the patient has seizures.
Age, gender, age at seizure onset, seizure characteristics
(semiology and frequency), factors potentially predisposing
to PNES [e.g., a history of sexual abuse, a history of physical
abuse (corporal punishment or any physical injury resulted
from aggressive behavior towards the patient), academic fail-
ure (school dropout or repeated grades), family function (i.e.,
divorce, single parent, significant family disputes, etc.), past
history of head injury, and a family history of seizures], and
video-EEG recordings of all patients were recorded routinely.
Demographic variables and pertinent clinical variables were
summarized descriptively to characterize the study popula-
tion. First, we performed univariate analyses using Pearson
chi-squared, Fisher’s exact, Mann-Whitney, Kolmogorov-
Smirnov, and t test, as appropriate. Then, variables that were
significant (p < 0.05) were assessed in a logistic regression
analysis. Odds ratio (OR) and 95% confidence interval (CI)
were calculated. p value less than 0.05 was considered as
significant. In the first attempt, we studied all patients and
Neurol Sci
compared the clinical features of patients who had a brain
MRI with those who did not, in order to discover any statisti-
cally and clinically significant variables that might have led
physicians to order an imaging study in these patients. In the
second part of the study, we investigated patients who had a
brain MRI available to see whether some clinical factors are
associated with structural brain imaging abnormalities in these
patients. We conducted this study with the approval by the
Shiraz University of Medical Sciences Review Board and
the Ethics Committee.
Results
During the study period, 330 patients were recorded in our
database; 132 patients (40%) had a brain MRI study, and
198 patients (60%) did not. Mean age (± standard deviation)
of the patients was 29 (± 10) years, and their mean age at the
onset was 24 (± 11) years. The sex ratio of the patients was
1.89 (females 216/males 114). Table 1 shows the clinical fea-
tures of the patients who had a brain MRI study in comparison
with those who did not. Patients, who had a brain MRI, had
higher seizure frequencies and more often had abnormal neu-
rological examinations and comorbid epilepsy in univariate
analyses. We then performed a binary logistic regression anal-
ysis considering these three variables. The model that was
generated by regression analysis was significant (p =
0.0001) and could predict the possibility of having a brain
MRI study in 71.2% of the patients. Comorbid epilepsy (OR
= 5.490, 95% CI 2.965–10.165; p = 0.0001) and seizure fre-
quency (OR = 1.004, 95% CI 1.000–1.008; p = 0.03) were
significantly associated with having a brain MRI study in
patients with PNES. Abnormal neurological examination lost
its significance within the model that was generated by regres-
sion analysis (p = 0.09).
Of the 132 patients with a brain MRI study, 47 patients
(36%) had abnormal finding(s) in their brain MRI studies
and 85 patients (64%) had a normal imaging study. Table 2
shows brain MRI findings in these patients. We investigated
the associations between demographic and clinical variables
and abnormal brain MRI study in univariate analyses
(Table 3). Age (p = 0.01) and comorbid epilepsy (p = 0.004)
were significantly associated with having abnormal brain MRI
findings in univariate analyses. Patients with PNES, who had
an abnormal brain MRI finding, were older and more often
had comorbid epilepsy. We then performed a binary logistic
regression analysis considering these two variables. The mod-
el that was generated by regression analysis was significant (p
= 0.001) and could predict the possibility of an abnormal brain
MRI finding in 70.5% of the patients. Both variables retained
their significance within the model that was generated by re-
gression analysis; age (OR = 1.040, 95% CI 1.005–1.077; p =
0.02) and comorbid epilepsy (OR = 3.006, 95% CI 1.394–
Neurol Sci

Table 1 Clinical features of the

patients with psychogenic With MRI Without MRI p value
nonepileptic seizures, who had a (132 patients) (198 patients)
brain MRI study, in comparison
with those who did not, in Age (mean ± standard deviation) (years) 30 ± 11 28 ± 10 0.1
univariate analyses Age at onset (mean ± standard deviation) (years) 25 ± 12 23 ± 10 0.1
Duration of the illness (mean ± standard deviation) (years) 5.2 ± 7.2 5.4 ± 7.6 0.8
Sex ratio (female/male) 85: 47 131: 67 0.8
Seizure frequency (mean ± standard deviation) 44 ± 87 28 ± 49 0.03
Aura 81 (61%) 128 (65%) 0.6
Loss of responsiveness 106 (80%) 172 (87%) 0.09
Wax and wane motor movements 70 (53%) 109 (55%) 0.8
Generalized motor seizures 114 (86%) 173 (87%) 0.7
Akinetic seizures 14 (11%) 20 (10%) 1.0
Urinary incontinence 15 (11%) 21 (11%) 0.8
Ictal injury 43 (33%) 58 (29%) 0.4
A history of head injury 4 (3%) 11 (6%) 0.4
A history of sexual abuse 9 (7%) 19 (10%) 0.5
A history of physical abuse 15 (11%) 26 (13%) 0.7
A family history of seizures 45 (34%) 56 (28%) 0.2
Comorbid epilepsy 46 (35%) 18 (9%) 0.0001
Abnormal neurological examination 8 (6%) 3 (2%) 0.02

6.481; p = 0.005) were significantly associated with having an and epilepsy (in 19 out of 46 patients; 41%), but these abnor-
abnormal brain MRI finding in patients with PNES. malities were rarely observed in those with PNES only (in 4
In a subanalysis, we excluded the patients with comorbid out of 86 patients; 5%) (p = 0.00001).
epilepsy (46 patients). From the remaining 86 patients (with
PNES only), 63 patients (73.3%) had a normal brain MRI
study and 23 patients (26.7%) had abnormal findings on their Discussion
imaging study. Table 4 shows brain MRI findings in patients
with PNES only. No demographic or clinical variable was In this study, 40% of the patients in a large cohort of those with
significantly associated with an abnormal brain MRI in pa- PNES had a brain MRI study. Other than comorbid epilepsy,
tients with PNES only (all p values > 0.05). Common epilep- we could not find any significant clinical variable that was
togenic structural brain abnormalities (e.g., tumors, mesial associated with performing a brain MRI in patients with
temporal sclerosis, encephalomalacia, and developmental PNES; seizure frequency was just borderline significant (odds
anomalies) were common in patients with comorbid PNES ratio = 1.004). Clinically, this makes sense and is justifiable to
perform a brain MRI in most patients with epilepsy (with or
Table 2 Brain MRI findings in patients with psychogenic nonepileptic without PNES). However, in a previous study of 224 patients
seizures with PNES only, 60% of the patients had brain imaging stud-
Brain MRI finding Number Percent
ies [MRIs or computerized tomographic (CT) scans] [6].
Those authors observed that there was a significant associa-
Normal 85 64.4 tion between the frequency of having neurodiagnostic tests
White matter nonspecific changes 12 9 and the PNES phenotype in their patients [6]. We did not have
Brain tumor 8 6 such an observation. The methodological differences may ex-
Mesial temporal sclerosis 7 5 plain the observed contradiction between these two studies.
Encephalomalacia and gliosis 6 4.5 In the current study, 36% of the whole cohort of the patients
Cyst 5 3.8 with PNES, who had a brain MRI study, and 27% of those
Atrophy 4 3 with PNES only had abnormal brain MRI findings. This is
Developmental anomaly 2 1.5 consistent with most previous studies [6–10]. In a study of
Vascular lesion 2 1.5 87 patients with PNES, who had brain MRI available, from
Demyelinating lesion 1 0.8 the USA [7], 37% of the whole cohort and 34% of those with
Total 132 100 PNES only had brain MRI abnormalities. In another study of
169 patients from the UK [8], 44% of the whole cohort and
 Neurol Sci

Table 3 Associations between

demographic and clinical Normal MRI Abnormal MRI p value
variables and abnormal brain (85 patients) (47 patients)
MRI study in univariate analyses
Age (mean ± standard deviation) (years) 28 ± 10 33 ± 13 0.01
Age at onset (mean ± standard deviation) (years) 24 ± 11 26 ± 13 0.2
Duration of the illness (mean ± standard deviation) (years) 4.4 ± 6.1 6.6 ± 8.7 0.1
Sex ratio (female/male) 57: 28 28: 19 0.4
Seizure frequency (mean ± standard deviation) 39 ± 72 52 ± 108 0.4
Aura 56 (66%) 25 (53%) 0.2
Loss of responsiveness 68 (80%) 38 (81%) 1.0
Generalized motor seizures 71 (84%) 43 (91%) 0.4
Urinary incontinence 12 (14%) 3 (6%) 0.2
Ictal injury 27 (32%) 16 (34%) 0.8
A history of head injury 1 (1%) 3 (6%) 0.1
A history of sexual abuse 6 (7%) 3 (6%) 1.0
A history of physical abuse 10 (12%) 5 (11%) 1.0
A family history of seizures 32 (38%) 13 (28%) 0.3
Comorbid epilepsy 22 (26%) 24 (51%) 0.004

27% of those with PNES only had abnormal brain MRI scans. structural neuroimaging changes in white and grey matter
In the healthy populations, incidental structural brain abnor- structures have been observed in major depressive disorders
malities have been reported in 18% of individuals [11]. These [12–14].
observations suggest that structural brain abnormalities may Therefore, at this time, it is premature to draw a concrete
play a role in the development of PNES [8, 9]. However, our conclusion, but at the same time, reproducibility of this obser-
study and many previous studies were limited in being retro- vation of common occurrence of structural brain abnormalities
spective (hampering collection of several clinical variables in patients with PNES highlights the necessity of designing
such as the occurrence of auras or minor episodes of loss of future studies to investigate this plausible theory appropriate-
awareness reliably) and not including healthy control groups ly. It is noteworthy to mention that common epileptogenic
[7–9]. In addition, it is not clear if all patients had an imaging brain abnormalities (e.g., tumors, mesial temporal sclerosis,
study with an appropriate epilepsy protocol. Finally, we did encephalomalacia, and developmental anomalies) were not
not investigate coexisting psychopathologies in our patients, frequently observed in patients with PNES only in our study
and considering the high frequency of psychiatric comorbid- (just in 5%); these findings were commonly observed in those
ities in patients with PNES, it is impossible to infer that these with comorbid PNES and epilepsy (in 41%). This is some-
brain abnormalities are specifically associated with PNES and what different from the findings of a previous study, in which
not with other coexisting psychopathologies [9]. For example, abnormal brain imaging findings were more or less similar
between patients with PNES only and those with comorbid
Table 4 Brain MRI findings in patients with psychogenic nonepileptic PNES and epilepsy [7]. This important contradiction should
seizures (PNES) only be investigated in future studies. Finally, it is important to keep
in mind that while 27% of patients with PNES only did have
Brain MRI finding Number Percent
an abnormal brain MRI, 73% of them had normal MRIs. This
Normal 63 73.3 would suggest that factors other than structural brain abnor-
White matter nonspecific changes 10 11.6 malities are more importantly involved in the development
Cyst 5 5.8 and maintenance of PNES [15, 16].
Brain tumor 3 3.5 We also observed that age and comorbid epilepsy were
Atrophy 3 3.5 significantly associated with having abnormal brain MRI find-
Encephalomalacia and gliosis 1 1.2 ings in patients with PNES. These observations are intuitive.
Vascular lesion 1 1.2 As we discussed above patients with comorbid epilepsy more
Mesial temporal sclerosis 0 0 often had abnormal brain abnormalities, which is consistent
Developmental anomaly 0 0 with previous studies and is very much expected [7]. Also, it is
Demyelinating lesion 0 0 expected for older patients to have more frequent brain imag-
Total 86 100 ing abnormalities (e.g., white matter nonspecific changes,
which were frequently observed in our study). Finally, we
Neurol Sci
did not observe any associations between clinical characteris-
tics of the patients and existence of brain MRI abnormalities in
this study.
We have to mention that we did not study the laterality or
the exact location of the brain imaging abnormalities in the
current study. Some previous studies have indicated that non-
dominant hemisphere lesions may play a role in the develop-
ment of PNES [17], but other studies have negated this hy-
pothesis [9]. A recent systematic meta-review of neuroimag-
ing studies in patients with PNES concluded that “the lack of
clear hemispheric dominance or lobar preponderance emanat-
ing from these studies supports the notion of a heterogeneous
etiology and phenomenology of PNES” [9].
In conclusion, while the evidence is convergent on the
relatively high prevalence of structural brain abnormalities in
patients with PNES, the data so far is suboptimal. In order to
investigate the significance of structural brain abnormalities in
the development of PNES, future well-designed multicenter
studies, which include a large number of patients with a uni-
fied methodology of imaging, are desirable. In designing such
studies, important confounding variables including any co-
morbid psychopathologies should be considered.
Acknowledgments We thank the Neuroscience Research Center, Shiraz
University of Medical Sciences, for supporting this study.
Author contribution The authors conducted the statistical analyses.
Compliance with ethical standards
Conflict of interest Ali A. Asadi-Pooya, M.D.: Honoraria from Cobel
Daruo; Royalty: Oxford University Press (Book publication). Maryam
Homayoun, M.D.: none.
References
1. Asadi-Pooya AA, Sperling MR (2015) Epidemiology of psycho-
genic non-epileptic seizures. Epilepsy Behav 46:60–65
2. Asadi-Pooya AA (2015) Neurobiological origin of psychogenic
nonepileptic seizures: a review of imaging studies. Epilepsy
Behav 52:256–259
3. Asadi-Pooya AA, Emami Y, Emami M (2014) Psychogenic non-
epileptic seizures in Iran. Seizure 23(3):175–177
4. Asadi-Pooya AA, Tinker J, Fletman EW (2017) How variable are
psychogenic nonepileptic seizures? A retrospective semiological
study. J Neurol Sci 377:85–87
5. Asadi-Pooya AA, Tinker J (2017) Delay in diagnosis of psycho-
genic nonepileptic seizures in adults: a post hoc study. Epilepsy
Behav 75:143–145
6. Wei D, Garlinghouse M, Li W, Swingle N, Samson KK,
Taraschenko O (2018) Utilization of brain imaging in evaluating
patients with psychogenic nonepileptic spells. Epilepsy Behav 85:
177–182
7. Bolen RD, Koontz EH, Pritchard PB 3rd. (2016) Prevalence and
distribution of MRI abnormalities in patients with psychogenic
nonepileptic events. Epilepsy Behav 59:73–76
8. Reuber M, Fernandez G, Helmstaedter C, Qurishi A, Elger CE
(2002) Evidence of brain abnormality in patients with psychogenic
nonepileptic seizures. Epilepsy Behav 3:249–254
9. Mcsweeney M, Reuber M, Levita L (2017) Neuroimaging studies
in patients with psychogenic non-epileptic seizures: a systematic
meta-review. Neuroimage Clin 16:210–221
10. Jones SG, O’Brien TJ, Adams SJ et al (2010) Clinical characteris-
tics and outcome in patients with psychogenic nonepileptic sei-
zures. Psychosom Med 72:487–497
11. Katzman GL, Dagher AP, Patronas NJ (1999) Incidental findings
on brain magnetic resonance imaging from 1000 asymptomatic
volunteers. JAMA 282:36–39
12. Kanner AM (2004) Structural MRI changes of the brain in depres-
sion. Clin EEG Neurosci 35:46–52
13. Coloigner J, Batail JM, Commowick O et al (2019) White matter
abnormalities in depression: a categorical and phenotypic diffusion
MRI study. Neuroimage Clin 22:101710
14. Andreescu C, Butters MA, Begley A, Rajji T, Wu M, Meltzer CC,
Reynolds CF 3rd, Aizenstein H (2008) Gray matter changes in late
life depression—a structural MRI analysis.
Neuropsychopharmacology 33:2566–2572
15. Asadi-Pooya AA (2016) Biological underpinnings of psychogenic
nonepileptic seizures: directions for future research. Neurol Sci 37:
1033–1038
16. Asadi-Pooya AA (2017) Psychogenic nonepileptic seizures: a con-
cise review. Neurol Sci 38:935–940
17. Devinsky O, Mesad S, Alper K (2001) Nondominant hemisphere
lesions and conversion nonepileptic seizures. J Neuropsychiatr Clin
Neurosci 13:367–373
Publisher’s note Springer Nature remains neutral with regard to jurisdic-
tional claims in published maps and institutional affiliations.