Epilepsy & Behavior 88 (2018) 244–250

Contents lists available at ScienceDirect

Epilepsy & Behavior

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

Theory of Mind impairment in focal versus generalized epilepsy

Nicky Morou a, Vassilis Papaliagkas b, Eleni Markouli a, Maria Karagianni a, Elena Nazlidou a, Martha Spilioti b,
Theodora Afrantou c, Vassilis K. Kimiskidis b, Nicolas Foroglou d, Mary H. Kosmidis a,⁎
a
Lab of Cognitive Neuroscience, School of Psychology, Aristotle University of Thessaloniki, Thessaloniki, Greece
b
Lab of Clinical Neurophysiology, 2nd Department of Neurology, AHEPA General Hospital, Aristotle University of Thessaloniki, Thessaloniki, Greece
c
AHEPA General Hospital, Thessaloniki, Greece
d
Department of Neurosurgery, AHEPA General Hospital, Aristotle University of Thessaloniki, Thessaloniki, Greece

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

Article history: Theory of Mind (ToM) is a critical component of social cognition, and thus, its impairment may adversely affect
Received 8 July 2018 social functioning and quality of life. Recent evidence has suggested that it is impaired in epilepsy. What is not
Revised 22 August 2018 clear, however, is whether it is related to particular types of epilepsy or other factors. We undertook the present
Accepted 20 September 2018
study to explore ToM in patients with focal versus those with generalized epilepsy, the particular pattern of ToM
Available online xxxx
deficits, and the potential influence of antiepileptic medication load. Our sample included 149 adults: 79 patients
Keywords:
with epilepsy (34 with generalized epilepsy and 45 with focal epilepsy) and 70 healthy controls. Theory of Mind
Social cognition tasks included a) comprehension of hinting, b) comprehension of sarcasm and metaphor, c) comprehension of
Sarcasm false beliefs and deception, d) recognition of faux pas, and e) a visual ToM task in cartoon form. We found signif-
Metaphor icant ToM impairment in the group with focal epilepsy relative to the performance of both the healthy group and
Faux pas perception the group with generalized epilepsy on all tasks, with the exception of faux pas, on which the group with gener-
Adults alized epilepsy also performed more poorly than the healthy group. Additionally, early age at seizure onset, but
not antiepileptic drug (AED) load, was associated with ToM performance. Our findings suggest that focal tempo-
ral and frontal lobe, but not generalized, epilepsies were associated with impaired ToM. This may reflect the neu-
roanatomical abnormalities in the relevant neuronal networks and may have implications for differential
cognitive-behavioral interventions based on epilepsy type.
© 2018 Published by Elsevier Inc.

1. Introduction emotion recognition, as well as on intention attribution tasks [4]. In

Emerging evidence of Theory of Mind (ToM) impairment among
individuals with epilepsy has highlighted the importance of including
such an exploration in routine assessments of patients' cognitive and
social functioning. Theory of Mind refers to the ability to put oneself in
the perspective of another person and to infer his/her mental state as
potentially different from one's own [1,2]. It is an essential component
of social functioning and interpersonal interactions. If impaired, it may
have a negative impact on quality of life.
Social cognition in generalized epilepsy has received limited
research attention to date as compared with other disorders. Previous
reports [3,4] have shown moderate ToM abilities in patients with gener-
alized epilepsy relative to healthy participants or individuals with focal
(specifically temporal lobe) epilepsy. Patients with generalized epilepsy
have been found to have either normal faux pas (FP) recognition [3] or
performance between that of patients with focal epilepsy on FP and
⁎ Corresponding author at: Lab of Cognitive Neuroscience, School of Psychology,
Aristotle University of Thessaloniki, 54124 Thessaloniki, Greece.
E-mail address: kosmidis@psy.auth.gr (M.H. Kosmidis).
line with these findings, additional studies [5,6] have demonstrated
that ToM abilities in patients with generalized epilepsy were impaired
in anger and fear recognition only but displayed almost intact intention
and emotion attribution abilities with scores again between those
attained by the healthy group and group with temporal lobe epilepsy.
In contrast, Jiang and colleagues [7] observed ToM deficits in patients
with generalized epilepsy patients, specifically concerning eye emotion
recognition and cognitive empathy abilities with intact affective empathy.
Theory of Mind impairment in both temporal [8] and frontal lobe
epilepsies has received considerable empirical attention [9,10]. Patients
with mesial temporal lobe epilepsy have demonstrated severe deficits
in FP recognition relative to patients with generalized epilepsy and
extramesial temporal lobe epilepsy [3,4,11] despite normal perfor-
mance on face recognition [4]. Moreover, investigations of patients
with temporal lobe epilepsy have also reported deficits in emotional
intelligence, facial emotion recognition, and deceitful and sarcastic
speech comprehension in social exchanges [12,13]. Indeed, Wang and
colleagues [14] described extensive impairment in both basic and
advanced ToM functions (i.e., FP/false belief recognition, mental state
inference) in a group with temporal lobe epilepsy. Additionally, FP
recognition deficits were observed in patients with cryptogenic epilepsy

https://doi.org/10.1016/j.yebeh.2018.09.026
1525-5050/© 2018 Published by Elsevier Inc.
 N. Morou et al. / Epilepsy & Behavior 88 (2018) 244–250
[15], but the same research group found no ToM impairment in patients
with temporal lobe epilepsy in a subsequent study [11].
Given the role of the frontal lobes in aspects of social cognition, re-
searchers have also explored ToM in patients with frontal lobe epilepsy
[16]. In particular, Farrant and colleagues [17] described impairment in
the Reading the Mind in the Eyes task, humor appreciation, and emotion
recognition tasks but normal performance on FP and “Happe's strange
stories” in patients with frontal lobe epilepsy. Happe's strange stories
is a task that involves the comprehension of nonliteral statements in
hedged expressions, metaphors, irony, lies, sarcasm, and first and sec-
ond order deception. On the other hand, subsequent studies [11,18] re-
ported significant FP deficits in patients with frontal lobe epilepsy. Other
researchers [19] have also observed significant impairment in the Yoni
task (which assesses cognitive and affective ToM abilities based on ver-
bal cues, eye gaze, and facial expressions) and cognitive empathy de-
spite intact affective empathy.
Furthermore, additional explorations have revealed a relationship
between ToM impairment and clinical variables in patients with focal
epilepsy [16]. Seizure-related factors demonstrating the strongest asso-
ciations with poor ToM performance were early age of seizure onset,
disease duration, and the number of antiepileptic drugs (AEDs) taken.
Moreover, findings pointed to indications of neurodevelopmental injury
(e.g., dysembryoplastic neuroepithelial tumor and mesial temporal lobe
sclerosis) [11,16]. Thus, clinical features related to focal epilepsy have
been found to adversely affect ToM performance.
Given the role of the amygdala in ToM processing, some studies have
explored ToM in mesial temporal lobe epilepsy associated with amyg-
dala damage. These studies have yielded conflicting findings, perhaps
reflective of the acute or early manifestation of the damage. For exam-
ple, Shaw and colleagues [20] observed that patients who underwent
amygdalectomy showed no postoperative change in their ToM perfor-
mance but demonstrated improvement in the recognition of fear in fa-
cial expressions approaching normal levels. Indeed, the role of the
amygdala in higher level ToM has been supported in studies of patients
with congenital or early amygdala damage, whose ToM abilities were
impaired in detecting “socially incomplete” (FP task) or sarcastic
speech, recognition of second order false beliefs, understanding of
another's motivation, and the interpretation of nonverbal emotional ex-
pressions [21,22]. Thus, the amygdala appears to play a role in the devel-
opment of the neural circuitry mediating ToM.
Our goal in undertaking the current research study was to investi-
gate potential differences in ToM between patients with generalized ep-
ilepsy and those with focal epilepsy (specifically temporal and frontal
lobe epilepsies), expecting that damage specifically to temporal and
frontal regions would lead to greater and more extensive ToM impair-
ment than in cases with presumably generalized or diffuse damage.
We also sought to explore the potential influence of clinical features
namely age of seizure onset and illness duration, as well as AED load
on ToM performance, expecting that these variables would further com-
promise ToM.
2. Methods
2.1. Participants
A total of 79 patients with different epilepsy syndromes were
recruited for the study. The patients had been diagnosed or were
being treated over a 36-month period in the epilepsy monitoring units
of two neurology clinics at Papanikolaou and AHEPA Hospitals in
Thessaloniki, Greece. Clinical diagnosis for the selected cases (n = 79)
was supported by electroencephalography (EEG) recordings, followed
by magnetic resonance imaging (MRI) scans (n = 71). We studied
these clinical data in order to determine the localization and the etiol-
ogy of the epileptic seizures (lesion localization and type). Thirty-five
of the patients had a diagnosis of generalized epilepsy (group with gen-
eralized epilepsy: nine with juvenile myoclonic epilepsy, seven with
245
juvenile absence epilepsy, and 19 with other types of generalized epi-
lepsy), 30 with temporal lobe epilepsy (group with temporal lobe epi-
lepsy: 12 with right and 11 with left hemisphere loci), four with
frontal lobe epilepsy (group with frontal lobe epilepsy), seven with
focal lesions originating in areas other than the frontal and temporal
lobes (e.g., occipital regions), and three with an undefined type of epi-
lepsy. Magnetic resonance imaging scans indicated unilateral hippo-
campal sclerosis in 12 of the patients with temporal lobe epilepsy
(eight in the right hemisphere, three in the left hemisphere, and one bi-
lateral). Those patients with focal lesions outside the frontal and tempo-
ral lobes as well as those with an undefined type of epilepsy were
excluded from the statistical analyses. We also excluded patients with
a history of psychiatric illness (e.g., psychosis), developmental disorders
related to intellectual deficits, or other physical diseases which could
have caused mental impairment. We did not exclude patients with
medically treated anxiety or depression as these conditions are frequent
comorbid symptoms of epilepsy. Also, we did not exclude patients
based on medical comorbidity as most additional diagnoses did not in-
volve the functioning of the central nervous system and they were
equally dispersed between the two patient groups.
The healthy control group consisted of 70 healthy participants, who
had no history of a neurological or a psychiatric disorder and received
no treatment. The patients with generalized epilepsy served as a patient
control group; they were comparable with the group with focal epilepsy
with respect to AED load and had not had daily seizures for at least a six-
month period or their seizures were rare. The seizure frequency for each
patient was classified into one of three categories (I. daily seizures — one
or more seizures per day, II. persistent seizures — less than one per day
but at least one per six months, and III. rare or no seizures — less than
one per six months) based on Lüders' classification system [23]. For
the purposes of the present study, patients with frontal and temporal
lobe epilepsies were combined to form a single group with focal epi-
lepsy, comprising only focal epilepsy syndromes involving the neural
network related to ToM. The three groups (healthy control, generalized,
and focal epilepsy) were matched on level of education [F(2, 113) =
2.423, p = .093] but differed with respect to gender ratio [there were
more men than women in the healthy control group but the reverse
pattern in both groups with epilepsy; χ2(2) = 15,078, p = .001] and
age [F(2, 113) = 6.249, p = .001; patients with focal epilepsy were sig-
nificantly older than those in the healthy control group and group with
generalized epilepsy; the latter two groups did not differ in age]. Since
we found a significant correlation between age and ToM scores, we
employed age as a covariate in the statistical analyses. Gender was not
associated with ToM performance. Table 1 summarizes the demo-
graphic characteristics for all groups as well as the clinical data for the
patient groups. All participants gave written informed consent to partic-
ipate in the study and were treated in accordance with the Code of
Ethics of the World Medical Association (Declaration of Helsinki) for ex-
periments involving humans.
2.2. Procedure
Participants were informed about the nature of the study and agreed
to participate. They then completed a series of tasks examining verbal
and visual ToM, as well as emotion recognition. The tasks were admin-
istered individually to the participants and took about 1–1.5 h to com-
plete. The administration order of the tasks differed systematically for
each participant so as to eliminate possible order effects. Specifically,
we applied a pseudo-randomized administration order with two deliv-
ery options. For this purpose, tasks were divided into two parts: part A
consisted of the visual ToM task while part B comprised the verbal
ToM tasks in the following order: hinting task, sarcasm and metaphor
task, false belief, FP, and deception scenarios. We switched the delivery
order of the two parts to the participants alternately so that for each
participant, test category order was different as compared with the
previous and the next one. As a result, half of the participants were
246 N. Morou et al. / Epilepsy & Behavior 88 (2018) 244–250

Table 1 questions. The first required the detection of the FP (‘Did anyone say
Demographic and clinical characteristics of the three groups. something s/he shouldn't have said?’). The other three questions (‘Who
Healthy Group with Group with said this and why shouldn't this person have said it?’, ‘Why did s/he say
control group generalized epilepsy focal epilepsy that?’, ‘How did X feel when Y said that?’) concerned the comprehension
(n = 70) (n = 35) (n = 34) of the FP and required the participant to identify the character who
Gender (male/female) 43/27 8/27 13/21 made the FP, understand social conventions and the behavioral slip
Age (yrs)/M (SD) 32.6 (10.99) 29.9 (11.5) 40.08 (14.37) made, attribute intentions, and recognize the listener's emotions. De-
Education (yrs)/M (SD) 12.48 (2.29) 12.66 (3.25) 11.27 (3.17)
pending on the adequacy of the responses, the participant received 0, 1,
Age at seizure onset – 17.28 (7.59) 16.22 (12.13)
(yrs) M (SD) or 2 points for each question resulting in a maximum score of 32 points.
Monthly seizure The total percentage of correct answers was calculated only for the
frequency (n) items on which the participant had answered the first question correctly.
Type II – 9 19
Type III – 25 14
AED load/M (SD) – 1.64 (1.27) 2.92 (1.81)
2.3.1.4. Visual ToM tasks. The visual ToM task consisted of 22 line-draw-
Presence of psychiatric 0 12 12 ings representing short stories in cartoon form on the top part of each
disease (n) card. After observing each cartoon story with no time constraint, the
Comorbidity (n) 0 9 8 participant was asked to choose the appropriate image among two or
three alternatives presented at the bottom of the same card, which
would complete it in a reasonable way. The stories concerned four
administered part A followed by part B, while the reverse order (part B types of mentalizing and were divided as follows: four stories involved
first then part A) was applied to the other half. first order false belief (i.e., predict a character's reaction when s/he has a
false belief regarding the events in a situation), six stories involved attri-
2.3. ToM assessment bution of intention (i.e., predict how the character is thinking of reacting
in a particular situation), and six stories required attribution of desire (i.e.,
The tasks used to examine ToM were developed in Greek for a previ- predict a character's actions after s/he has perceived another person's de-
ous study and have been described in detail elsewhere [24–26]. None of sire). The remaining six stories did not involve ToM and were used as con-
the tasks were timed, and in all cases, the stimuli remained available to trols to explore the participant's ability to comprehend the story plot.
the participant for review, thus avoiding potential difficulties due to in-
creased working memory load. A brief description of each task follows. 2.4. Statistical analyses

2.3.1. Verbal ToM tasks
2.3.1.1. Comprehension of hinting. The stimuli consisted of six short
stories presenting an interaction between two characters and ending
with a hint, dropped by one of the characters [25]. The participant was
required to explain what the character really meant by what s/he said
after reading each scenario on a single page. This question required par-
ticipants to infer the real intention hidden behind indirect speech utter-
ances. If the participant's response was appropriate and explicitly
correct, it was awarded two points, if partially correct, 1 point. The max-
imum score on this task was 12 points.
2.3.1.2. Comprehension of sarcasm and metaphor. With this task, we ex-
plored pragmatic comprehension skills such as sarcasm and metaphor.
Participants read 24 short stories concluding with a statement made
by one of the main characters. Then they were asked to explain what
the character meant by this statement. The 24 stories were subdivided
into four conditions of six stories, including sarcastic utterances, meta-
phors, literal statements, and irrelevant utterances. The maximum
score for each of the four conditions was 12 points.
2.3.1.3. Comprehension of ToM in verbal scenarios. This task comprised of
14 stories examining three types of ToM: eight scenarios involved first
and second order false belief, two involved first and second order de-
ception (bluff and double bluff), and four involved FP decoding [24].
The participants read each scenario and answered relevant questions.
Additionally, we included one ToM question which required making in-
ferences about the characters' mental states (beliefs and intentions). If a
participant failed to answer a reality question, then the ToM question
was not administered and the score for this item was not included in
the total score of the task. We then calculated the percentage of correct
responses without these items.
The FP stories involved a situation in which a character inadvertently
made an offensive, unpleasant, or ungracious statement about another
person. This test required participants to make both cognitive and affec-
tive state attributions in order to perceive and identify the social FP in
each scenario. After reading each story, participants answered four
Statistical analyses were carried out using SPSS version 24.0. After
calculating descriptive statistics for each group, we tested the effect of
group (healthy controls, generalized, and focal epilepsy) and age on
the scores achieved on the ToM tasks. We also conducted a series of
one-way and multivariate analyses of covariance (ANCOVAs and
MANCOVAs) to examine group effects (with age as a covariate) on the
total scores and subscores of the ToM tasks. Then, we conducted
Bonferroni post hoc comparisons. Finally, we explored the potential ef-
fects of three factors (age, age at seizure onset, and AED load) on total
ToM scores and subscores through multiple regression analysis only
for the patient groups (generalized and focal epilepsies). We set the sig-
nificance level at p b .05 (two-tailed) for all statistical analyses.
3. Results
3.1. ToM tasks
3.1.1. Group comparisons
We conducted MANCOVAs or ANCOVAs to compare the two sub-
groups with epilepsy (focal and generalized) and the healthy control
group with regard to ToM variables with age as a covariate in all of
the statistical tests. Table 2 summarizes the mean total scores and
subscores of the three groups on all tasks. These comparisons are de-
scribed in the sections that follow.
3.1.1.1. Visual ToM tasks. Multivariate analyses of covariance (based on
Pillai's trace) yielded a significant group effect favoring the healthy
control group and group with generalized epilepsy relative to the
group with focal epilepsy for the following scenarios: ToM [F(2, 135)
= 3.113, p = .048], intention [F(2, 135) = 22.57, p b .001], desire
[F(2, 135) = 5.283, p = .006], and for the total score [F(2, 135) =
12.149,p b .001]. There was a significant age effect on the total
score [F(1, 135) = 7.454, p = .007], the ToM [F(1, 135) = 6.507, p
= .01], and the desire scenarios [F(1, 135) = 6.313, p = .01]. Post
hoc tests revealed that the group with focal epilepsy performed sig-
nificantly more poorly than the healthy control group on the ToM
and desire scenarios and worse than both healthy control group
 N. Morou et al. / Epilepsy & Behavior 88 (2018) 244–250 247

Table 2
Mean (SD) scores on ToM tasks for healthy participants and patients with generalized epilepsy and focal epilepsy.

Healthy Group with Group with Bonferroni

control group generalized epilepsy focal epilepsy post hoc
n = 70 n = 35 n = 34 comparisons

Mean (SD) Mean (SD) Mean (SD)

Visual ToM scenarios

• 1st order false belief 3.23 0.82 2.94 1.05 2.59 1.35 FE⁎ b HC
• Intention 5.74 0.50 5.57 0.74 4.29 1.62 FE⁎ b HC, GE
• Desire 4.48 1.22 4.26 1.27 3.35 1.76 FE⁎ b HC
• Total score 19.04 2.21 18.17 2.95 14.91 5.58 FE⁎ b HC, GE

Verbal ToM scenarios

Hinting task 11.45 1.10 11.03 2.24 9.15 3.39 FE⁎ b HC, GE
Indirect speech comprehension:
• Irony 10.14 2.75 9.46 3.49 7.88 4.00 FE⁎ b HC
• Metaphor 10.40 1.70 7.66 2.57 6.70 2.94 FE⁎ b HC/GE⁎ b HC
• Literal speech 11.42 1.14 10.86 1.26 9.85 2.28 FE⁎ b HC
• Irrelevant speech 9.30 2.49 9.51 2.59 9.03 3.40 n/s
False belief scenarios:
• 1st order 3.90 0.35 3.77 0.56 3.15 1.23 FE⁎ b HC, GE
• 2nd order 3.07 0.81 2.60 1.17 2.18 1.51 FE⁎ b HC
Comprehension of deception:
• 1st order 0.96 0.21 0.86 0.35 0.65 0.49 FE⁎ b HC
• 2nd order 0.90 0.30 0.74 0.44 0.53 0.51 FE⁎ b HC
Faux pas task
• Detection FP question #1 7.09 2.07 7.43 1.50 6.23 2.77 n/s
• Question #2 (‘why shouldn't s/he X’) 6.25 2.07 6.34 2.28 5.26 2.95 n/s
• Question #3 (‘why did s/he X’) 6.07 2.32 3.81 3.01 2.94 2.85 FE⁎ b HC/GE⁎ b HC
• Question #4 emotion 7.03 2.15 7.20 1.57 5.79 2.85 FE⁎ b HC, GE

Note: HC = healthy control group, GE = group with generalized epilepsy, FE = group with focal epilepsy.
⁎ p b .05.

and group with generalized epilepsy on intention scenarios and the
total score. We found no significant difference between the perfor-
mance of the group with generalized epilepsy and healthy control
group on any of the scenarios or the total score on the visual ToM
task.
3.1.1.2. Hinting task. Analyses of covariance revealed a significant group
effect on the comprehension of hinting [F(2, 134) = 9.661, p b .001]
(when controlling for age, which had a significant effect [p = .03]). Spe-
cifically, patients with focal epilepsy scored significantly lower than
both the group with generalized epilepsy and healthy control group;
the latter did not differ for each other.
3.1.1.3. Comprehension of sarcasm and metaphor. A MANCOVA revealed a
significant group effect on sarcasm [F(2, 134) = 3.534, p = .03], meta-
phor [F(2, 134) = 33.136, p b .001] and literal speech scenarios [F(2,
134) = 10.293, p b .001] but not irrelevant speech scenarios [F(2,
134) = 0.182, p = .83]. There was a significant age effect only on the
sarcasm scenarios [F(1, 134) = 5.028, p = .027]. Post hoc tests showed
significant ToM impairment in the group with focal epilepsy compared
with the healthy control group on three scenario types (sarcasm, meta-
phor, and literal, but not irrelevant, speech), while patients with gener-
alized epilepsy displayed lower scores than the healthy control group
only on the metaphor scenarios.
3.1.1.4. ToM task with verbal scenarios. A MANCOVA revealed a signifi-
cant group effect on first and second order false belief [F(2, 134) =
10.084, p b .001 and F(2, 134) = 5.787, p = .004, respectively] and de-
ception scores [F(2, 134) = 7.387, p = .001 and F(2, 134) = 7.982, p
= .001, respectively]. The age effect was significant only for the second
order false belief score [F(1, 134) = 8.034, p = .005]. The group with
focal epilepsy performed significantly worse than the healthy control
group on both false belief and deception scenarios (first and second
order) and also scored lower than the group with generalized epilepsy
on all first order scenarios. On the other hand, there was no difference
between participants with generalized epilepsy and healthy control
participants on false belief and deception variables.
In addition, a significant effect of group was observed through
MANCOVA in relation to the FP scenarios for the question ‘why did s/
he say that (FP)?’ [F(2, 134) = 17.873, p b .001] and the recognition of
the other character's emotions [F(2, 134) = 4.118, p = .018] but not
for the initial detection of the FP [F(2, 134) = 2.704, p = .07] and the
question ‘why shouldn't s/he have said that (FP)?’ [F(2, 134) = 2.337,
p = .101]. There was no age effect [F(4, 131) = 0.143, p = .96]. Groups
with focal epilepsy and generalized epilepsy were less accurate than the
healthy control group on the question ‘why did s/he say that?’, but did
not differ from each other. The group with focal epilepsy also scored sig-
nificantly lower than both the group with generalized epilepsy and the
healthy control group on the emotion recognition questions; the latter
two groups did not differ from each other.
3.1.1.5. The effects of age, age at seizure onset, and AED load on ToM perfor-
mance for the groups with epilepsy. Hierarchical regression analyses were
conducted in order to examine the effect of three independent variables
(age, age at seizure onset, and AED load) on the quantitative ToM vari-
ables only for the patients with generalized epilepsy and focal epilepsy
considered as a single group with epilepsy. This statistical method was
applied on the total scores and subscores of the visual ToM scenarios,
the sarcasm and metaphor comprehension task, the hinting task, the
false belief task, and the FP scenarios. The remaining two ToM scores
(first and second order deception vignettes) were binary, so we con-
ducted logistic regression analyses for these variables. Table 3 presents
the R2, R2 change, F change, beta values, and their significance for the
best fitting model of each of the hierarchical regression analyses. Re-
garding the order of the independent variables, age was entered into
the first block, age at seizure onset into the second block, and AED
load into the third block.
The analyses revealed significant effects of age and age at seizure
onset on the total score of the visual ToM task, as well as the intention
and desire vignettes, decoding of metaphors, the first order false belief
task, FP detection, the FP question “why shouldn't s/he have said
that?”, and the FP question regarding emotion recognition. In addition,
there was a significant age effect only on the hinting task and the second
order false belief task. Beta values indicated that ToM abilities in
248 N. Morou et al. / Epilepsy & Behavior 88 (2018) 244–250

Table 3
Hierarchical regression analyses of the effects of age, age at seizure onset, and AED load on ToM performance for the groups with epilepsy combined.

Test variables Model predictors R2 R2 change F change p (F) Beta p (beta)

Visual ToM scenarios

• 1st order false belief a&b 0.198 0.133 10.959 .002 −0.032 .002
0.046 .002
• Intention a&b 0.243 0.132 11.503 .001 −0.044 .000
0.053 .001
• Desire a&b 0.259 0.219 19.480 .000 −0.039 .003
0.077 .000
• Total score a&b 0.304 0.199 18.862 .000 −0.154 .000
2.18 .000

Verbal ToM scenarios

Hinting task a 0.083 0.083 6.090 .016 −0.062 .016
Indirect speech comprehension:
• Sarcasm a&b 0.067 0.052 3.662 .600 −0.052 .134 (n/s)
n/s 0.090 .060 (n/s)
• Metaphor a&b 0.095 0.063 4.569 .036 −0.051 .042
0.072 .036
• Literal speech a, b, & c 0.099 0.044 3.175 .079 −0.018 .306 (n/s)
n/s 0.024 .324 (n/s)
−0.253 .079 (n/s)
• Irrelevant speech a, b, & c 0.102 0.053 3.819 .055 −0.031 .268 (n/s)
n/s 0.012 .750 (n/s)
−0.440 .055 (n/s)
False belief scenarios:
• 1st order a&b 0.116 0.067 5.001 .029 −0.021 .016
0.027 .029
• 2nd order a 0.108 0.018 8.148 .006 −0.032 .006
Faux pas task
• Detection FP question #1 a&b 0.128 0.089 6.702 .012 −0.047 .020
0.071 .012
• Question #2 (‘why shouldn't s/he X’) a&b 0.146 0.111 8.571 .005 −0.055 .019
0.092 .005
• Question #3 (‘why did s/he X’) a, b, & c 0.053 0.047 3.236 .077 0.000 .994 (n/s)
n/s −0.017 .652 (n/s)
−0.407 .077 (n/s)
• Question #4 emotion a&b 0.126 0.081 6.135 .016 −0.051 .016
0.070 .016

Note: R2 concerns the most significant model for each dependent variable. Predictors: a = age, b = age at seizure onset, and c = AED load.

patients were significantly more impaired in the aforementioned tasks 4. Discussion

the older they were or the earlier their seizure disorder onset was.
Regarding the logistic regression analyses, the main values (beta and
their significance, Wald statistic, and Odds ratio) are displayed in Table
4. Age, age at seizure onset, and AED load were entered as independent
variables initially, and the produced models were compared in order to
determine which one was confirmed. For both dependent variables
(first order and second order deception), only the model testing the
age effect was significant, while adding the other two variables or
their interaction did not make a significant difference. Therefore, the
data presented here concern the confirmed model for each dependent
variable. Both models had a good fit, as indicated by Hosmer and
Lemeshow's goodness of fit test (1st order deception: χ2(8) = 5.638,
p = .688; 2nd order deception: χ2(8) = 9.056, p = .338). Based on
Wald values for beta coefficients, increased age contributed to a signifi-
cantly impaired performance of the group with epilepsy on both 1st and
2nd order deception tasks.
Table 4
Logistic regression analyses coefficient values of the effect of age on the two binary ToM
variables for the two groups with epilepsy.
Dependent variable b Wald p Odds ratio
Deception stories
• 1st order −0.058 7.322 .007
• 2nd order −0.037 3.963 .047
In the present study, we assessed components of ToM ability using
verbal and visual material in a patient group with generalized epilepsy,
a second patient group with focal temporal or frontal lobe epilepsy, and
a healthy control group. Our findings confirmed our hypothesis regard-
ing ToM impairment primarily among patients with focal epilepsy but
with more circumscribed difficulties among patients with generalized
epilepsy. Specifically, we found extensive deficits in the patients with
focal epilepsy relative to the healthy participants on most of the ToM
tasks. Patients with focal epilepsy had difficulty with the attribution of
intention and desire, first order false belief, as well as overall ToM on a
visual scenario task. This difficulty inferring another person's intention
among the group with focal epilepsy was also evident on their poor in-
terpretation of hinting. Regarding indirect speech comprehension, the
group with focal epilepsy had difficulty perceiving sarcasm, metaphor,
and literal speech, as well as comprehending first and second order
false belief and deception. This group also demonstrated difficulty
with attribution of intention on FP vignettes (‘why did s/he say that?’)
and on emotion recognition but had no difficulty relative to the healthy
control group on FP detection, identification of the social slip, and the
comprehension of irrelevant speech.
Regarding the group with generalized epilepsy, the present findings
showed intact performance on most of the ToM tasks and mild impair-
ment only in two ToM measures. Specifically, patients with generalized
epilepsy did more poorly than the healthy control group only on the
0.944 comprehension of metaphor and gave less accurate answers on the FP
0.963
question related to the attribution of intention (‘why did s/he say
 N. Morou et al. / Epilepsy & Behavior 88 (2018) 244–250
that?’) but performed similarly to controls on the visual ToM scenarios,
hinting, comprehension of sarcasm, literal and irrelevant speech, com-
prehension of false belief and deception scenarios, and three of the
four questions on the FP recognition task. When compared with the
group with focal epilepsy, patients with generalized epilepsy made bet-
ter attribution of intentions, scored higher overall on the visual ToM
task, and performed significantly higher on the hinting task, first order
false belief and deception scenarios, as well as on the emotion recogni-
tion question of the FP task. With regard to the attribution of desire and
first order false belief assessed through the visual ToM task, the compre-
hension of sarcasm, literal and irrelevant speech, second order false be-
lief and deception, and detection of a social slip on FP scenarios, the
group with generalized epilepsy did not differ from either of the other
two groups. Concerning these tasks, the scores of the patients with gen-
eralized epilepsy were between those of the healthy control group and
group with focal epilepsy.
The present findings are consistent with the majority of previous
studies reporting notable ToM impairment in patients with focal frontal,
temporal, temporal cryptogenic, or mesial temporal epilepsy on FP rec-
ognition tasks [3,4,11,15,18]. Our findings are also consonant with those
of studies (employing either verbal or cartoon scenarios) that suggested
deficits of patients with temporal epilepsy or early amygdala damage in
the comprehension of hinting, first order false belief and sarcasm, and
the interpretation of nonverbal expressions [21,22,27]. Other reports,
however, are contradictory to our findings, indicating intact FP and fa-
cial emotion recognition in patients with mesial or lateral temporal ep-
ilepsy [4,11], while yet other studies have shown no decrease in ToM
ability in patients after left or right amygdalectomy [20]. Moreover,
there is also a previous report [17] on patients with frontal lobe epilepsy
suggesting no deficits in FP recognition and a series of social cognition
abilities concerning the attribution of intention and the comprehension
of sarcasm, humor, and first and second order deception, which were
assessed with the use of a cartoon story task.
Furthermore, we explored the relationship between three clinically
relevant variables and performance on the ToM tasks. Age influenced
all participant groups on all measures related to the visual ToM task
(total score, ToM, and desire vignettes) and three verbal ToM tasks
(hinting task, sarcasm, and second order false belief scenarios). A signif-
icant and broader contribution of age also emerged for ToM abilities of
the patients with epilepsy only as estimated by the visual ToM task
(total score, intention, and desire vignettes) and most verbal ToM
tasks (hinting task, metaphor subtest, first and second order false belief
and deception scenarios, three questions of FP task). Early age at seizure
onset also contributed to patients' lower performance on the visual ToM
task (total score, intention, and desire vignettes) but not all the verbal
tasks (metaphor subscore, first order false belief task, and three ques-
tions from the FP scenarios). These findings clearly suggest that ToM de-
clines with aging. They also indicated a relationship between social
cognition and early seizure onset. Previous studies, however, have re-
ported conflicting findings regarding clinical variables such as age at
seizure onset [3], seizure frequency [11], illness duration [3], and seizure
control [20] with respect to their predictive value for ToM performance.
These discrepancies could be attributed to heterogeneity in the types of
epilepsy represented in different studies. They may also reflect second-
ary seizure-related factors affecting ToM functioning, yielding a
subsignificant relationship. In contrast, we found no association be-
tween AED load and performance of patients with epilepsy on any
ToM task. This suggests that antiepileptic medication neither amelio-
rates nor aggravates ToM impairment in patients with generalized or
focal epilepsy.
The present findings indicate different patterns of ToM abilities in
patients with focal and generalized epilepsy syndromes. Specifically,
focal epilepsies (frontal or temporal) appear to be linked with extensive
deficits across the full range of ToM processes. In fact, previous studies
[16] have also suggested intragroup differentiated patterns pointing
that patients with mesial temporal epilepsy displayed more extensive
249
impairment in ToM abilities as compared with patients with extramesial
or lateral temporal damage. Furthermore, ToM abilities in participants
with frontal lobe epilepsy were impaired in FP and emotion recognition.
The evidence is limited, however, since this type of focal epilepsy is quite
rare. Unfortunately, in the present study, we were unable to compare
subgroups with focal epilepsy (frontal vs temporal vs medial temporal)
on ToM performance because of the discrepancy in our sample sizes.
Therefore, we cannot speculate regarding intragroup ToM patterns
based on our data. The discrepancy between the current and previous
findings as mentioned above may reflect the particular cerebral regions
with damage (i.e., mesial, lateral temporal, or amygdala damage), vari-
ance in syndrome severity, or differences in task modality and difficulty
level. Additionally, it may reflect the existence of a distinct neural net-
work that serves ToM functions and includes brain areas located in the
frontal and temporal lobes.
Further research with larger and, to the extent possible, more ho-
mogenous samples than the present could help to elucidate the patterns
of ToM difficulties based on type, location, and AED load in patients with
epilepsy. Moreover, this information could be associated with measures
of daily functioning or quality of life so as to explore the impact of im-
paired ToM and to inform intervention programs to improve the func-
tional outcome of patients with epilepsy.
Conflict of interest
The authors have no conflict of interest to report.
References
[1] Premack D, Woodruff G. Does the chimpanzee have a theory of mind? Behav Brain
Sci 1978;4:515–26.
[2] Baron-Cohen S. Theory of mind and autism — a review. Int Rev Res Ment Retard
2000;23:169–84.
[3] Schacher M, Winkler R, Grunwald T, Kraemer G, Kurthen M, Reed V, et al. Mesial
temporal lobe epilepsy impairs advanced social cognition. Epilepsia 2006;47:
2141–6.
[4] Broicher SD, Kuchukhidze G, Grunwald T, Krämer G, Kurthen M, Jokeit H. “Tell me
how do I feel” — emotion recognition and theory of mind in symptomatic mesial
temporal lobe epilepsy. Neuropsychologia 2012;50:118–28.
[5] Realmuto S, Zummo L, Cerami C, Agrò L, Dodich A, Canessa N, et al. Social cognition
dysfunctions in patients with epilepsy: evidence from patients with temporal lobe
and idiopathic generalized epilepsies. Epilepsy Behav 2015;47:98–103.
[6] Reynders HJ, Broks P, Dickson JM, Lee CE, Turpin G. Investigation of social and emo-
tion information processing in temporal lobe epilepsy with ictal fear. Epilepsy Behav
2005;7:419–29.
[7] Jiang Y, Hu Y, Wang Y, Zhou N, Zhu L, Wang K. Empathy and emotion recognition in
patients with idiopathic generalized epilepsy. Epilepsy Behav 2014;37:139–44.
[8] Steiger BK, Jokeit H. Why epilepsy challenges social life. Seizure 2017;44:194–8.
[9] Szemere E, Jokeit H. Quality of life is social — towards an improvement of social abil-
ities in patients with epilepsy. Seizure 2015;26:12–21.
[10] Kirsch HE. Social cognition and epilepsy surgery. Epilepsy Behav 2006;8:71–80.
[11] Giovagnoli AR, Franceschetti S, Reati F, Parente A, MacCagnano C, Villani F, et al. The-
ory of mind in frontal and temporal lobe epilepsy: cognitive and neural aspects.
Epilepsia 2011;52:1995–2002.
[12] Walpole P, Isaac CL, Reynders HJ. A comparison of emotional and cognitive intelli-
gences in people with and without temporal lobe epilepsy. Epilepsia 2008;49:
1470–4.
[13] Cohn M, St-Laurent M, Barnett A, McAndrews MP. Social inference deficits in tempo-
ral lobe epilepsy and lobectomy: risk factors and neural substrates. Soc Cogn Affect
Neurosci 2015;10:636–44.
[14] Wang W-H, Shih Y-H, Yu H-Y, Yen D-J, Lin Y-Y, Kwan S-Y, et al. Theory of mind and
social functioning in patients with temporal lobe epilepsy. Epilepsia 2015;56:
1117–23.
[15] Giovagnoli AR, Canafoglia L, Reati F, Raviglione F, Franceschetti S. The neuropsycho-
logical pattern of Unverricht-Lundborg disease. Epilepsy Res 2009;57(6):920–30.
[16] Giovagnoli AR. The importance of theory of mind in epilepsy. Epilepsy Behav 2015;
39:145–53.
[17] Farrant A, Morris RG, Russell T, Elwes R, Akanuma N, Alarcón G, et al. Social cognition
in frontal lobe epilepsy. Epilepsy Behav 2005;7:506–16.
[18] Giovagnoli AR, Parente A, Villani F, Franceschetti S, Spreafico R. Theory of mind and
epilepsy: what clinical implications? Epilepsia 2013;54:1639–46.
[19] Hu Y, Jiang Y, Hu P, Ma H, Wang K. Impaired social cognition in patients with
interictal epileptiform discharges in the frontal lobe. Epilepsy Behav 2016;57:
46–54.
[20] Shaw P, Lawrence E, Bramham J, Brierley B, Radbourne C, David AS. A prospective
study of the effects of anterior temporal lobectomy on emotion recognition and the-
ory of mind. Neuropsychologia 2007;45:2783–90.
250 N. Morou et al. / Epilepsy & Behavior 88 (2018) 244–250
[21] Fine C, Lumsden J, Blair RJR. Dissociation between ‘theory of mind’ and executive
functions in a patient with early left amygdala damage. Brain 2001;124:287–98.
[22] Shaw P, Lawrence EJ, Radbourne C, Bramham J, Polkey CE, David AS. The impact of
early and late damage to the human amygdala on “theory of mind” reasoning.
Brain 2004;127:1535–48.
[23] Lüders H. Classification of epileptic seizures and epilepsies. In: Lüders H, editor. Text-
book of epilepsy surgery. 1st ed. London: CRC Press; 2008. p. 245–51.
[24] Bozikas VP, Giannakou M, Kosmidis MH, Kargopoulos P, Kioseoglou G, Liolios D, et al.
Insights into theory of mind in schizophrenia: the impact of cognitive impairment.
Schizophr Res 2011;130:130–6.
[25] Giannakou M. Thesis: theory of mind in schizophrenia. Aristotle Univ Thessaloniki;
2008http://thesis.ekt.gr/thesisBookReader/id/18938#page/1/mode/2up, Accessed
date: 19 May 2018.
[26] Kosmidis MH, Giannakou M, Garyfallos G, Kiosseoglou G, Bozikas VP. The impact of
impaired “theory of mind” on social interactions in schizophrenia. J Int
Neuropsychol Soc 2011;17:511–21.
[27] Li Y-H, Chiu M-J, Yeh Z-T, Liou H-H, Cheng T-W, Hua M-S. Theory of mind in patients
with temporal lobe epilepsy. J Int Neuropsychol Soc 2013;19:594–600.