Epilepsy & Behavior 60 (2016) 197–201

Contents lists available at ScienceDirect

Epilepsy & Behavior

journal homepage: www.elsevier.com/locate/yebeh

Validation of a novel classification model of psychogenic nonepileptic

seizures by video-EEG analysis and a machine learning approach
Adriana Magaudda a,⁎, Angela Laganà a, Alessandro Calamuneri a, Teresa Brizzi a, Cinzia Scalera a,
Massimiliano Beghi b, Cesare Maria Cornaggia c, Gabriella Di Rosa d
a
Epilepsy Center, Department of Clinical and Experimental Medicine, University of Messina, Italy
b
Psychiatric Clinic, G. Salvini Hospital, Garbagnate Milanese, Italy
c
Department of Surgery and Medicine, University of Milano-Bicocca, Italy
d
Department of Human Pathology of Adult and Child, Unit of Infantile Neuropsychiatry, University of Messina, Italy

a r t i c l e i n f o a b s t r a c t

Article history: The aim of this study was to validate a novel classification for the diagnosis of PNESs. Fifty-five PNES video-EEG
Received 11 November 2015 recordings were retrospectively analyzed by four epileptologists and one psychiatrist in a blind manner and clas-
Revised 19 March 2016 sified into four distinct groups: Hypermotor (H), Akinetic (A), Focal Motor (FM), and with Subjective Symptoms
Accepted 21 March 2016
(SS). Eleven signs and symptoms, which are frequently found in PNESs, were chosen for statistical validation of
Available online xxxx
our classification. An artificial neural network (ANN) analyzed PNES video recordings based on the signs and
Keywords:
symptoms mentioned above. By comparing results produced by the ANN with classifications given by examiners,
Psychogenic seizures we were able to understand whether such classification was objective and generalizable. Through accordance
Video-EEG metrics based on signs and symptoms (range: 0–100%), we found that most of the seizures belonging to class
Machine learning A showed a high degree of accordance (mean ± SD = 73% ± 5%); a similar pattern was found for class SS
Classification of PNESs (80% slightly lower accordance was reported for class H (58% ± 18%)), with a minimum of 30% in some cases.
Low agreement arose from the FM group. Seizures were univocally assigned to a given class in 83.6% of seizures.
The ANN classified PNESs in the same way as visual examination in 86.7%. Agreement between ANN classification
and visual classification reached 83.3% (SD = 17.8%) accordance for class H, 100% (SD = 22%) for class A, 83.3%
(SD = 21.2%) for class SS, and 50% (SD = 19.52%) for class FM. This is the first study in which the validity of a new
PNES classification was established and reached in two different ways. Video-EEG evaluation needs to be
performed by an experienced clinician, but later on, it may be fed into ANN analysis, whose feedback will provide
guidance for differential diagnosis. Our analysis, supported by the ML approach, showed that this model of
classification could be objectively performed by video-EEG examination.
© 2016 Elsevier Inc. All rights reserved.

1. Introduction Among patients affected by paroxysmal nonepileptic events, PNESs

Psychogenic nonepileptic seizures (PNESs) are episodes of altered
movements, sensations, or experiences similar to epileptic seizures
(ES). However, they are not caused by abnormal electric discharges in
the brain but rather by a psychological process [1].
The annual incidence of PNESs has been estimated to be around 1.5
cases per 100,000 [2]. Psychogenic nonepileptic seizures occur in 5–20%
of adult populations with epilepsy and 15–40% of adult referrals to epi-
lepsy centers [1,3–5]. The prevalence of PNESs in children referred to
epilepsy centers with suspected epilepsy ranged from 1 to 9% [6–10].
Abbreviations: ML, machine learning; H, Hypermotor; A, Akinetic; FM, Focal Motor; SS,
Subjective Symptoms.
⁎ Corresponding author at: Department of Clinical and Experimental Medicine,
Policlinico Universitario G. Martino, Via Consolare Valeria, 98125 Contesse, Messina,
Italy. Tel.: +39 090 2212954, +39 3929413735; fax: +39 090 2212301.
E-mail address: amagaudda@unime.it (A. Magaudda).
account for 11% to 46% [11–13]. The misdiagnosis of PNESs may lead
to inappropriate treatment with antiepileptic drugs (AEDs) with a sig-
nificant risk of iatrogenic morbidity and elevated costs for the patients
and the health-care system [13–15]. Diagnosis of PNESs remains a chal-
lenging topic. Misdiagnosis is enhanced by the coexistence of PNESs and
ES in the same patient, with a percentage ranging from 3.6% to 58% in
different series [15] and by the presence of EEG abnormalities as docu-
mented in 22.3% of subjects with PNES in one series [16]. Video-
electroencephalography (video-EEG) represents the gold standard for
PNES diagnosis. However, it is not available worldwide and has high
costs, especially when patients undergo long-term monitoring. Alterna-
tively, although home video could be a useful tool to observe seizures,
its diagnostic sensitivity is lower than that of video-EEG. The ILAE com-
mitted a group of epileptologists, neuropsychiatrists, and neuropsychol-
ogists to assess clear guidelines for standards for PNES diagnosis [17].
Using a consensus review of the literature, this group evaluated key di-
agnostic approaches, including detailed history and seizure descriptions

http://dx.doi.org/10.1016/j.yebeh.2016.03.031
1525-5050/© 2016 Elsevier Inc. All rights reserved.
198 A. Magaudda et al. / Epilepsy & Behavior 60 (2016) 197–201
from witnesses, EEG, video-EEG monitoring, neurophysiology, neuro-
humoral monitoring, neuroimaging, neuropsychological testing, hypno-
sis, and conversation analysis [17]. Differential diagnosis cannot only
rely on clinical features of PNESs because the majority of signs and
symptoms overlap with ES. A longer duration, a gradual onset,
waxing-and-waning movements, asynchronous movements, pelvic
thrusting, closed eyes, lateralized head movements, and ictal crying en-
compass the complex clinical pattern of PNESs, yet none of these is
disease-specific. Closed eyes and lateralized head movements have
been reported to be the most specific signs of PNESs. Psychogenic
nonepileptic seizures display a polymorphic semiology, mimicking dif-
ferent types of ES, thus, a classification of PNESs would be useful to in-
crease diagnostic accuracy.
Several PNES clinical classifications have been proposed [18], but
their validation and robustness have not yet been reached. The aim
of this study was to propose a novel simplified clinical classification of
PNESs and verify its validity in terms of reproducibility, either at a
subjective level or in an automatic analysis framework, as provided by
a machine learning (ML) approach.
2. Patients and methods
Fifty-five patients with PNESs diagnosed by video-EEG recordings
performed at the Adult Epilepsy Center and Unit of Infantile Neuropsy-
chiatry of the University of Messina were included in the study. Patients
or their caregivers gave written informed consent. We recorded at least
one PNES for each patient, occurring spontaneously or, in most patients,
induced by suggestion maneuvers (verbal suggestion, intermittent light
stimulation, a patch placed on the forearm, or oral administration of
a placebo). We showed PNES videos to a caregiver, to ascertain if the
semiology of recorded PNESs was the same as those usually presented
by the patient.
Four epileptologists, belonging to two different tertiary care epilepsy
centers, and one psychiatrist visually analyzed PNES videos in a blind
manner and classified recorded PNESs into four distinct groups:
Hypermotor (H), Akinetic (A), Focal Motor (FM), and with Subjective
Symptoms (SS), according to a clinical classification proposed by one
(AM) of the five physicians, highly experienced with PNESs. We then in-
vestigated the concordance of the classification of different observers.
We considered as belonging to class H those seizures mainly character-
ized by either tonic, clonic, or dystonic generalized movements.
Seizures mainly characterized by the absence of movement were
assigned to class A. Seizures with focal motor movements were instead
assigned to FM, whereas SS seizures were those mainly characterized by
experiential phenomena reported by the patients. Video-EEG record-
ings showing typical patterns for each of the four PNES types are
found in the Supplementary Material (Videos 1–4). In Video 1
(Hypermotor), the attack started with a moaning, fine tremors in both
the hands, resistance to eye opening, generalized hypertonus, and dif-
fuse jerks. In Video 2 (Akinetic), the patient showed unresponsiveness
and absence of movements. In Video 3 (Subjective Symptoms), the pa-
tient referred to epigastric discomfort and that “I feel as the arms are
dry”. Finally, in Video 4 (Focal), the patient showed flapping of the
right arm on the bed surface and wide jerks to the lower limbs. For sta-
tistical validation of our new classification, we choose eleven signs and
symptoms more frequently found in PNESs than in ES. It is usually diffi-
cult for a single symptom to be clearly coupled to a specific PNES class;
this leads to problems when attempting to find statistically significant
differences between subclasses. Hence, while a single sign may be un-
able to accomplish this task, an analysis that attempts to consider all
signs simultaneously as a unique multivariate type of information may
instead be able to individuate hidden relations between these symp-
toms in a way to clearly differentiate PNES classes. An artificial neural
network (ANN) analyzed PNES video recordings based on the signs
and symptoms mentioned above in this multivariate framework, giving
as output subclasses which were automatically individuated. By
comparing results produced by the ANN with classifications given by
examiners after visual inspection, we were in a position to understand
whether such classification was objective and generalizable.
2.1. Data analysis
2.1.1. Feature inspection
Binary features considered were the following: Minor Limb Tremors,
Vocalization, Eyelid Myoclonia, Rigor-like Movements, Responsiveness,
Waxing-and-Waning Patterns, Pedaling, Body Rocking, Pelvic Thrust-
ing, Head Lateralization Movements, and Whole Body Rigidity. First,
we correlated those variables by computing the phi coefficient; the sta-
tistically significant level was set to α = 0.05, corrected for multiple
comparisons using Sidak correction.
2.1.2. Objectiveness of PNES visual classification
We examined the generalizability of PNES classification by looking
at the degree of accordance between groupings, as performed by five
expert examiners who separately inspected videos and classified the
entire dataset.
2.2. Machine learning classification
Machine learning is a data modeling tool borrowed from scientific
computing, which is able to find complex patterns within data and dif-
ferentiate them into subclasses. To validate an ML model, datasets are
usually divided into a training set (TS) and a cross-validation set (CV).
The TS is used to estimate parameters of the model; the latter is then ap-
plied to the CV set in order to test how this model is able to correctly
classify new data. Data are randomly assigned to TS and CV several
times; different models obtained from these trials are tested to find
the one that, on average, produces the best results. For this study, at
each trial, 60% of PNES classifications were assigned to TS and the
remaining 40% to CV. Finally, average performances were reported. An
important aspect to be taken into account is the numerosity of each sub-
group to be included in the analysis. The sample included in this study is
not perfectly balanced, i.e., there are some subclasses less represented
than others. This situation could present problems when assigning sam-
ples to TS and CV: in the worst scenario, an entire subclass could be ei-
ther assigned only to TS or to CV, thus making the classification task
erroneous. We resolved this potential problem by performing a pseudo-
random subdivision, in a way to be sure, at each trial, that 60% of each
subclass was part of TS. Among the ML techniques available, we chose
the ANN because it is considered as having the best performance, due
to its ability to find nonlinear hidden interactions within data [19]. We
fed the ANN with data features previously reported and classification
provided by examiners: for a given PNES, more than one class might
be elected, thus, the classification used for ANN analysis was based on
a majority vote between individual subdivisions. Details of ANN models
are the following: one input layer with 11 nodes, one hidden layer with
10 nodes, and one output layer consisting of 4 nodes; 1000 rounds of
cross-validation were used to obtain results. Analyses were performed
using the MATLAB Software Package (www.mathworks.com).
3. Results
3.1. Feature inspection
The percentage of each selected feature in the different subclasses is
summarized in Table 1. Among feature correlations, only Pedaling and
Head Lateralized Movements were significantly positively correlated
(phi = 0.523, p b 1 ∗ 10−6), i.e., they were either both present or both
absent. In Fig. 1, the degree of accordance between PNESs is shown,
based on the concomitant presence or absence of a given sign: for a
given couple of seizures, a count of features that are in common was
performed, and the normalized sum was displayed with a colored
 A. Magaudda et al. / Epilepsy & Behavior 60 (2016) 197–201 199

Table 1 we decided to assign such seizures to a particular class based on the

Percentage of each selected feature in the different PNES subclasses. higher average accordance with the PNES class into which it had been
Sign Class H Class A Class SS Class FM previously classified. In only one case was there high disagreement
Minor Limb Tremors 12/24 4/13 2/13 2/5
on the classification: that seizure was assigned either to SS (3/5) or
Vocalization 8/24 2/13 3/13 2/5 A (2/5). Thus, at the end of the visual classification step between exam-
Eyelid Myoclonia 4/24 1/13 4/13 0/5 iners, the following PNES subtype targeting was decided: 24 out of 55
Rigor-like Movements 13/24 2/13 1/13 1/5 PNESs were classified as belonging to class H, 13 were instead assigned
Responsiveness 4/24 2/13 11/13 3/5
to class A, and 13 were assigned to class SS, while the remaining 5 were
Waxing-and-Waning Patterns 8/24 2/13 1/13 3/5
Pedaling 4/24 0/13 0/13 0/5 assigned to class FM. Such PNES classification was the one which was in-
Body Rocking 4/24 0/13 0/13 0/5 vestigated by means of the ANN model.
Pelvic Thrusting 9/24 0/13 0/13 1/5
Head Lateralization Movements 5/24 1/13 0/13 1/5 3.2. Machine learning analysis
Whole Body Rigidity 16/24 2/13 2/13 0/5

In Fig. 2, the ML results are visualized: the ANN automatically cor-

scale ranging from 0% (no features in common) to 100% (same feature
distribution). Most of the seizures belonging to class A showed a high
degree of accordance (mean ± SD = 73% ± 5%); a similar pattern
was found for class SS (mean ± SD = 80% ± 2%). A slightly lower
accordance was reported for class H (mean ± SD = 58% ± 18%), with
a minimum of 30% in some cases. Low agreement arose from the FM
group. This means that A and SS PNESs show a similar semiology
among different patients. The higher variability observed for classes H
and FM means that there is not a defined subgroup of signs and symp-
toms that individuate the H and FM types of PNESs. For example, H
PNESs could be mainly characterized by stiffness in one patient and by
clonic or dystonic movements in another.
3.1.1. Interexaminer accordance of PNES visual classification
Seizures were unequivocally assigned to a given class in 83.6% (46/
55) of the seizures. In detail, the subdivision was the following: 22
were assigned to H, 10 to A, 12 to SS, and 3 to FM. For 4 out of 55
PNESs, only one of the examiners was not in agreement with the
other. In three of such situations, a seizure assigned to class H by the
others was instead considered by one examiner to belong to class FM;
in the other case, a seizure in class A was instead assigned by one exam-
iner to SS. For the above cases, it was not always the same examiner that
gave an alternate response. For 4 out of 55 PNESs, examiners were not
able to identify a single class because they displayed some characteris-
tics of two different classes. For instance, a given PNES may begin with
an akinetic phase, followed by hypermotor signs. However, all the ex-
aminers agreed for these boundary cases. In these cases, for ML analysis,
rectly classified 86.7% of CV sets, i.e., the ANN classified PNESs in the
same way as visual examination. In particular, agreement between
ANN classification and visual classification reached 83.3% (SD =
17.8%) accordance for PNESs of class H, 100% (SD = 22%) for class A,
83.3% (SD = 21.2%) for class SS, and 50% (SD = 19.52%) for class FM.
Low performances for FM are likely due to the low number of seizures
belonging to this class; in this respect, the results obtained for the FM
subtype may not be conclusive.
4. Discussion
This is the first study, to our knowledge, in which the validity of a
new PNES classification was established and reached in two different
ways. On the one hand, five different examiners independently classi-
fied PNESs on the basis of visual inspection of PNES video-EEG record-
ings. In 83.6% of the cases, all examiners classified PNESs in the same
manner. Reliability of this subdivision was furthermore tested by
means of a sophisticated mathematical tool, the ANN, with which we
performed classification based on eleven main signs and symptoms
that were individuated during video-EEG recordings. Features were
chosen on the basis of evidence already reported in the literature, and
they were those most frequently observed by the authors, having a
long experience on interpretation of PNES video-EEG recordings. More-
over, they correspond to signs and symptoms most commonly de-
scribed in PNESs [20–27]. Agreement between subjective and ANN
classification reached 86.7%; because of the relatively low numbers of
PNESs examined, this convincing result attributes a remarkable objec-
tiveness to our classification. Looking at single correlations between

Fig. 1. On the left, symmetric correlation matrix between subjects that participated in the study. Blocks indicate class subdivision according to majority voting. Subjects are ordered
according to the class they have been assigned to: from top to bottom (left to right): Hypermotor, Akinetic, Subjective Symptoms, and Focal Motor. Colors represent a normalized
count of features in common between each pair of subjects. Class subdivision is highlighted by means of black blocks, showing comparisons within each class (main diagonal) and
comparisons between subjects belonging to different classes. On the right, correlation matrix scheme. (For interpretation of the references to color in this figure legend, the reader is
referred to the web version of this article.)
200 A. Magaudda et al. / Epilepsy & Behavior 60 (2016) 197–201
Fig. 2. Fitting of classification performances during training (on the left in blue), CV (in the middle in red) processes, and overall mean results (on the right in green). Fitting is computed to
the posterior probability of a sample to belong to a given class: the closer the colored line is to the black dotted one, the higher is the probability of subjects to belong to the class they were
assigned to. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
symptoms used for ANN automatic analysis, only Pedaling and Head
Lateralization Movements were significantly positively related. This
demonstrates that the symptoms that we chose (Minor Limb Tremors,
Vocalization, Eyelid Myoclonia, Rigor-like Movements, Responsiveness,
Waxing-and-Waning Patterns, Pedaling, Body Rocking, Pelvic Thrust-
ing, Head Lateralization Movements, and Whole Body Rigidity) are rel-
atively independent: this aspect is important from a mathematical point
of view for the quality of input space used for the ANN technique. Fur-
thermore, these symptoms are greatly significant in terms of clinical
classification of PNESs. Recently, ML approaches have been applied to
epilepsy and ictal video decoding. A complete automatic procedure
was performed based on image processing of video recordings with
the goal of training a robotic system to understand arousal of epileptic
seizures [28]. Such an approach was not oriented to PNESs and could
not be implemented in clinical settings. Several classification schemes
were proposed based on cluster analysis of the clinical signs of the
PNESs collected by video-EEG recordings [17] or on detailed seizure ob-
servation and tabulation [28–30]. Groppel et al. analyzed the clinical se-
miology of seizures in 27 patients with PNESs collected by video-EEG
recording [28]. Cluster analysis of recurrent clinical features allowed
them to identify three groups of PNESs: “psychogenic motor seizures”
with predominantly hypertonic and clonic movements of the head
and limbs, pelvic thrusting, and tonic posturing; “psychogenic minor
motor or trembling seizures” with parcellar motor manifestations or
fine tremors; and “psychogenic atonic seizures” with a sudden loss of
postural tone and consciousness. Hubsch et al. proposed a more symp-
tom-specific cluster analysis for PNES classification that distinguished
PNESs into the following: dystonic attacks, paucikinetic attacks with
preserved responsiveness, pseudosyncope, hyperkinetic prolonged at-
tacks with hyperventilation and aura, and axial dystonic prolonged at-
tacks [29]. Seneviratne et al. reported a proposal of classification of
330 PNESs recorded from 61 patients, visually analyzed by just one ex-
pert epileptologist [31]. In this case, PNESs were grouped into six sub-
types: rhythmic motor (with minor motor phenomena or rigor-like
movements), hyperkinetic, complex motor, dialeptic, nonepileptic
auras, and mixed. Furthermore, Selwa et al. proposed a symptom-
based model of PNES classification, grouping seizures into the follow-
ing: catatonic, thrusting, automatism, tremor, intermittent, and subjec-
tive, among which the catatonic subtype showed a better outcome than
the thrashing ones [32]. Further simplified models of PNES classifica-
tion, based on the predominance of positive or negative motor and/or
sensory manifestations, have been proposed [33–35]. Bhatia and Sapra
reported that PNESs that mimicked generalized tonic–clonic ES were
the most common (60%), followed by the focal–clonic type (10%) and
the akinetic type (10%) [36]. None of these classifications was validated
by interobserver reliability combined with statistical analysis as in our
work.
We proposed four classes corresponding to the most frequently
described ones in clinical practice. Their definition has been dictated,
within this context, as well as in previous articles by other groups, by
the need of a differential diagnosis with respect to ES. In our experience,
PNES subclasses in clinical practice are similar to different subtypes
belonging to the domain of proper epileptic seizures. That is, the H
group could be placed in correspondence with the so-called convulsive
ES; similarly, the A group would correspond to atonic or absence ES.
Then, FM subtypes could be considered in correspondence with epilep-
tic partial motor seizure; eventually, the SS subgroup could be placed in
correspondence with epileptic partial seizure with experiential phe-
nomena. We did not place a mixed group. Looking at previous literature,
Groppel et al. and Hubsch et al. [28,29] did not report a mixed group. A
mixed PNES group (combination of features characterizing the other
subgroups of the classification) was reported with very different
percentages: 5.2% in Seneviratne's paper [31] and 32.9% in Dhiman's re-
port [37]. We have classified PNESs by considering the most prominent
seizure characteristics. For example, as shown in Table 1, “minor limb
tremors”, “rigor-like movements”, and “whole body rigidity”, which
are really frequent in class H, can be found, to a lower extent, in class
A as well. However, we classified as belonging to the A subtype those
PNESs in which the main and most evident symptom throughout a sei-
zure was akinesia.
Moreover, it is our advice that the creation of a classification that
would consider too many subclasses, at this stage, could cause confusion
in the absence of a classification accepted by the medical community. It
is worth mentioning that categorization does not relate to the treatment
choice, which depends on the comorbid psychiatric disorder.
Our analysis, supported by the ML approach, showed that this model
of PNES classification could be objectively performed by video-EEG
examination. Different examiners agreed for the most part on the subdi-
vision into classes, and only in very few cases was there discordance in
at least one of the examiner's assignments. Usually, ML algorithms are
applied to huge datasets; in our study, only 55 subjects were involved
in the analysis. It might be argued that our optimistic results need the
model to be applied to a larger dataset and tested again for its accuracy.
It is worth noticing that, as we already mentioned, assignment of
each PNES to a subclass was decided via a majority vote between exam-
iners; such decision was then fed to the ML algorithm to provide results.
 A. Magaudda et al. / Epilepsy & Behavior 60 (2016) 197–201
Because in some cases there was discordance regarding classification of
PNES subtype, a probabilistic labeling could have been used as input to
the ANN. Such choice might permit taking into account uncertainty in
category assignment. For the present work, we preferred to uniquely as-
sign each PNES to a different subtype in the attempt to follow clinical
practice, in which a unique decision has to be taken even when clini-
cians disagree in diagnosis. Exploring the impact of such probabilistic
assignment to the ANN may thus be of interest for future work.
Moreover, to the best of our knowledge, this is the first time in which
such a sophisticated ML model was used to provide a blind classification
of PNESs. We know that different ML models may be chosen, such as
decision trees. Future studies may be dedicated to explore results and
robustness coming from those alternative approaches.
In conclusion, a semiologically focused classification of PNESs may
address proper diagnosis and standardization across different studies.
Ultimately, etiologic understanding and management of this important
public health issue would improve.
Supplementary data to this article can be found online at http://dx.
doi.org/10.1016/j.yebeh.2016.03.031.
Conflict of interest
The authors have nothing to disclose.
References
[1] Lesser RP. Psychogenic seizures. Neurology 1996;46:1499–507.
[2] Sigurdadottir KR, Olafsson E. Incidence of psychogenic seizures in adults: a
population-based study in Iceland. Epilepsia 1998;39:749–52.
[3] Sirven JI, Glosser DS. Psychogenic nonepileptic seizures: theoretic and clinical
considerations. Neuropsychiatry Neuropsychol Behav Neurol 1998;11:225–35.
[4] Gumnit RJ, Walczak TS. Guidelines for essential services, personnel, and facilities in
specialized epilepsy centers in the United States. National Association of Epilepsy
Centers. Epilepsia 2001;42:804–:814.
[5] Reuber M, Micoulaud-Franchi J-A, Gülich E, Bartolomei F, McGonigal A. What the
patient's history tells us about their nonepileptic seizures. Neurophysiol Clin 2014;
44:375–88.
[6] Smith PE, Myson V, Gibbon F. A teenager epilepsy clinic: observational study. Eur J
Neurol 2002;9:373–6.
[7] Pakalnis A, Paolicchi J. Frequency of secondary conversion symptoms in children
with psychogenic nonepileptic seizures. Epilepsy Behav 2003;4:753–6.
[8] Kutluay E, Selwa L, Minecan D, Edwards J, Beydoun A. Nonepileptic paroxysmal
events in a pediatric population. Epilepsy Behav 2010;17:272–5.
[9] Asadi-Pooya AA, Emami M. Reasons for uncontrolled seizures in children: the impact
of pseudointractability. Epilepsy Behav 2012;25:341–: 344.
[10] Szabo L, Siegler Z, Zubek L, Liptai Z, Körhegyi I, Bánsági B. A detailed semiologic
analysis of childhood psychogenic nonepileptic seizures. Epilepsia 2012;53:565–70.
[11] Kotagal P, Costa M, Wyllie E, Wolgamuth B. Paroxysmal nonepileptic events in
children and adolescents. Pediatrics 2002;11:46.
[12] Kim SH, Kim H, Lim BC, Chae JH, Kim KJ, Hwang YS, et al. Paroxysmal nonepileptic
events in pediatric patients confirmed by long-term video-EEG monitoring — single
tertiary center review of 143 patients. Epilepsy Behav 2012;24:336–40.
201
[13] Reuber M, Baker GA, Gill R, Smith DF, Chadwick DW. Failure to recognize psychogenic
nonepileptic seizures may cause death. Neurology 2004;62:834–5.
[14] LaFrance Jr WC, Benbadis SR. Avoiding the costs of unrecognized psychological
nonepileptic seizures. Neurology 2006;66:1620–1.
[15] Reuber M, Elger CE. Psychogenic nonepileptic seizures: review and update. Epilepsy
Behav 2003;4:205–16.
[16] Reuber M. Psychogenic nonepileptic seizures: answers and questions. Epilepsy
Behav 2008;12:622–35.
[17] LaFrance Jr WC, Baker GA, Duncan R, Goldstein LH, Reuber M. Minimum require-
ments for the diagnosis of psychogenic nonepileptic seizures: a staged approach. A
report from the International League Against Epilepsy Nonepileptic Seizures Task
Force. Epilepsia 2013;54:2005–18.
[18] Cianci V, Ferlazzo E, Condino, Mauvais HS, Farnarier G, Labate A, et al. Rating scale
for psychogenic nonepileptic seizures: scale development and clinimetric testing.
Epilepsy Behav 2011;21:128–31.
[19] Langley P. Elements of machine learning. Morgan Kaufmann; 1996.
[20] Li J, Zhen X, Liu X, Ouyang G. Classifying normal and abnormal status based on video
recordings of epileptic patients. Scientific World Journal 2014;2014:459636.
[21] Leis AA, Ross MA, Summers AK. Psychogenic seizures: ictal characteristics and
diagnostic pitfalls. Neurology 1992;42:9.5–9.
[22] Dreifuss F. Classification of epileptic seizures. In: Engel JJ, Pedley TA, editors. Epilepsy: a
comprehensive textbook. Philadelphia: Lippincott-Raven; 1997. p. 525–32.
[23] Luders H, Acharya J, Banmgartner C, Benbadis S, Bleasel A, Burgess R, et al. Semiological
seizure classification. Epilepsia 1998;39:1006–13.
[24] Riggio S. Psychogenic seizures. Emerg Med Clin North Am 1994;12:1001–12.
[25] Gulick TA, Spinks IP, King DW. Pseudoseizures: ictal phenomena. Neurology 1982;
32:24–30.
[26] Gates JR, Ramani V, Whalen S, Loewenson R. Ictal characteristics of pseudoseizures.
Arch Neurol 1985;42:1183–7.
[27] Chee MWL, Kotagal P, van Ness PC, Gragg L, Murphy D, Luders HO. Lateralizing signs
in intractable partial epilepsy: blinded multiple-observer analysis. Neurology 1993;
43:19–25.
[28] Groppel G, Kapitany T, Baumgartner C. Cluster analysis of clinical semiology of
psychogenic nonepileptic seizures. Epilepsia 2000;415:610–4.
[29] Hubsch C, Baumann C, Hingray C, Gospodaru N, Vignal JP, Vespignani H, et al. Clinical
classification of psychogenic non-epileptic seizures based on video-EEG analysis and
automatic clustering. J Neurol Neurosurg Psychiatry 2011;82:955–60.
[30] Griffith NM, Szaflarski JP, Schefft BK, Isaradisaikul D, Meckler JM, McNally KA, et al. Re-
lationship between semiology of psychogenic nonepileptic seizures and Minnesota
Multiphasic Personality Inventory profile. Epilepsy Behav 2007;11:105–11.
[31] Seneviratne U, Reutens D, D'Souza W. Stereotypy of psychogenic nonepileptic
seizures: insights from video-EEG monitoring. Epilepsia 2010;51:1159–68.
[32] Selwa LM, Geyer J, Nikakhtar N, Brown MB, Schuh LA, Drury I. Non-epileptic seizure
outcome varies by type of spell and duration of illness. Epilepsia 2000;41:1330–4.
[33] Meierkord H, Will B, Fish D, Shorvon S. The clinical features and prognosis of
pseudoseizures diagnosed using video-EEG telemetry. Neurology 1991;41:1643–6.
[34] Reuber M, Pukrop R, Bauer J, Helmstaedter C, Tessendorf N, Elger CE. Outcome in
psychogenic non-epileptic seizures: 1 to 10-year follow-up in 164 patients. Ann
Neurol 2003;53:305–11.
[35] van Merode T, de Krom MC, Knottnerus JA. Gender-related differences in non-
epileptic attacks: a study of patient's cases in the literature. Seizure 1997;6:311–6.
[36] Bhatia MS, Sapra S. Pseudoseizures in children: a profile of 50 cases. Clin Pediatr
(Phila) 2005;44:617–21.
[37] Dhiman V, Sinha S, Rawat VS, Harish T, Chaturvedi SK, Satishchandra P. Semiological
characteristics of adults with psychogenic nonepileptic seizures (PNESs): an attempt
towards a new classification. Epilepsy Behav 2013;27:427–32.