Neuropsychol Rev DOI 10.1007/s11065-015-9313-x
REVIEW
Theory of Mind in Patients with Epilepsy: a Systematic Review and Meta-analysis Elizabeth Stewart 1,2 & Cathy Catroppa 3 & Suncica Lah 1,2
Received: 2 August 2015 / Accepted: 16 December 2015 # Springer Science+Business Media New York 2016
Abstract The ability to understand our own thoughts, intentions, beliefs and emotions and those of others (Theory of Mind; ToM) is a high-order social cognitive skill that is vital for social interaction and which has been found to be impaired in patients with epilepsy. Studies examining ToM in patients with epilepsy, however, have yielded inconsistent findings. The main aim of this study is to determine whether the magnitude of ToM deficits varies as a function of the site of epilepsy focus and/or the type of ToM task used. Electronic databases searches included Psychinfo, Medline/PubMed and EMBASE. Studies were included if they examined a group of patients with epilepsy and a group of healthy controls, reported original research, were published in the English language in peer reviewed journals, and used one of five empirically validated measures of ToM: False Belief, Reading the Mind in the Eyes Task (RMET), Faux-pas, Strange Stories, Cartoon ToM vignettes. Twelve studies were identified, ten included adults and two included children with epilepsy. Findings revealed marked ToM deficits in adults with focal seizures emanating from core brain regions underpinning ToM: temporal and frontal lobes (frontal lobe epilepsy, FLE; temporal lobe epilepsy, TLE), but not in adults with focal seizures outside the temporal and frontal lobes (extra-TLE/FLE). ToM deficits were also observed in children with generalised
* Suncica Lah
[email protected]
1
School of Psychology, The University of Sydney, Brennan McCallum (A18), Sydney, NSW 2006, Australia
2
ARC Centre of Excellence in Cognition and its Disorders, Sydney, NSW, Australia
3
Murdoch Children’s Research Institute, Royal Children’s Hospital, Melbourne, Australia
seizures (idiopathic generalised epilepsy, IGE). ToM deficits were documented across ToM tasks. In conclusion, ToM deficits represent a robust finding in adults with frontal and temporal epilepsy, but are also found in children with generalised seizures. Further research into ToM is needed, especially in children with epilepsy as early ToM may have cumulative, negative effects on development of social skills that continues into adulthood. Keywords Theory of mind . Mentalising . Epilepsy . Seizure Social cognition is a broad, multifaceted concept that involves an ability to understand the internal mental states of others and oneself (Fiske and Taylor 1991). It encompasses a range of skills including emotion perception and decoding abilities (facial emotion recognition) and higher-level skills such as Theory of Mind (ToM; Ladegaard et al. 2014; Sosa et al. 2011). Theory of Mind (ToM) refers to the ability to understand the thoughts, intentions, beliefs and emotions of others and oneself (Sodian and Kristen 2010). It has been differentiated from more basic social perception abilities, such as facial emotion recognition, in that it requires detection of ambiguous or covert social cues in order to understand both cognitive and affective internal mental states (e.g., eye-gaze expression, irony, metaphors, implicit meanings in speech) (Kim et al. 2011; Ladegaard et al. 2014). ToM thus encompasses component processes, including cognitive perspective taking (cognitive ToM) and emotional understanding (affective ToM) (Fiske and Taylor 2013). These skills develop at different rates through childhood and adolescence and are measured via different tasks (Kalbe et al. 2010). An important step in the development of ToM takes place at around 4 to 6 years old, when children develop an understanding of false belief: that other people can have beliefs that
Neuropsychol Rev
are mistaken or different from their own (Wimmer and Perner 1983). This skill is most frequently assessed by false belief tasks, which require a child to distinguish their own beliefs from another person’s and to reason that the other person’s beliefs may be mistaken or different from their own (Perner and Lang 1999). At 4 years old, children can generally predict an mistaken belief of another person (first-order false belief) and at around 6 years old, they develop the capacity to infer beliefs about beliefs: that a person may hold a mistaken belief about a third person’s belief (second-order false belief) (Bauminger-Zviely 2013). Performance on first and secondorder false belief tasks can be used to distinguish 4 from 6 year-old children, however, they lack specificity in older children and adults, who have moved beyond this simple developmental capacities (Baron‐Cohen et al. 1997). Because of this, more advanced ToM tasks have been developed to assess later-developing skills. These tasks require understanding of social faux pas, jokes, metaphors, irony, sarcasm and implicit meanings in speech (Baron-Cohen et al. 1999). They require individuals to move beyond the literal meaning of actions or language to reason about cognitive and affective states (Baron-Cohen et al. 1999; Happé 1994). Three commonly used advanced tests of ToM, which assess both cognitive and affective ToM, are the faux pas task, strange stories and cartoon ToM vignettes. The Reading the Mind in the Eyes Task (RMET) is another advanced test of ToM that specifically assesses affective ToM (Baron‐Cohen et al. 1997). The RMET requires perception of emotional states based on subtle facial emotion cues (i.e., eye gaze expression) (Baron‐Cohen et al. 1997). The RMET is generally passed by children from 7 to 8 years-old, with successful completion coinciding with the emergence of an important emotional reasoning skill: the ability to understand hidden emotions–that people may feel differently on the inside to the way they are expressing their emotions on the outside (Pons et al. 2004). Traditionally, ToM has been measured via these behavioural tasks and as such, it has long been thought that ToM emerged at 4 to 5 years old, when children developed mastery of false belief tasks. However, more advanced paradigms that assess subtler social cues (e.g., observing eye-gaze to measure belief-based anticipatory looking) suggest that children develop an understanding of other people’s minds from as young as 15 months old (Onishi and Baillargeon 2005) and 18 months old (Song et al. 2008). These findings have led to suggestions that performance on traditional ToM tasks may relate to alterations in cognitive skills other than ToM, such as verbal ability and executive skills (Martín-Rodríguez and León-Carrión 2010). Executive Functions (EFs) refer to a range of high order cognitive abilities such as working memory, inhibition and attentional flexibility. Both EF and verbal ability has been found to relate to ToM performance in typically developing children (Carlson and Moses 2001; Hughes 1998) and adults
(Bull et al. 2008) on a range of ToM tasks, including false belief (Astington and Jenkins 1999; Mutter et al. 2006; Wellman et al. 2001), faux pas, strange stories and cartoon ToM (Bird et al. 2004; Happé 1994). The RMET differs slightly, in that it relies on verbal ability but imposes comparatively few demands EF relative to story and cartoon tasks (Henry et al. 2015). Importantly, although these skills contribute to performance on ToM tasks, studies have consistently shown that they are distinct and dissociable abilities and that deficits in ToM may remain even after controlling for verbal ability and executive skills (Hughes 1998; Cavallini et al. 2013). This suggests that while the tasks in part rely on other cognitive skills, they do also measure ToM. The rising interest in ToM among patients with epilepsy comes from the fact that they experience marked deficits in social skills, which have been associated with higher rates of loneliness and psychological distress among adults with epilepsy (Suurmeijer et al. 2001) and impaired social communication in children with epilepsy (Caplan et al. 2005). Given the importance of ToM for social skills performance in typically developing children (Astington and Jenkins 1995), it has been purported that impoverished social skills in adults (Giovagnoli 2014) and in children (Rantanen et al. 2012) with epilepsy may be in part underpinned by impaired ToM. In patients with epilepsy, deficits in ToM may be secondary to disruption of a distributed neural network that supports this complex function. Neuroimaging studies conducted with healthy participants have implicated the anterior temporal lobe, precuneus (Olson et al. 2007), posterior superior temporal sulci (pSTS; Frith 2007), temporoparietal junction (TPJ; Saxe and Wexler 2005; Saxe and Kanwisher 2003), medial prefrontal cortices (Amodio and Frith 2006), and anterior paracingulate cortex in the medial frontal lobes (Gallagher and Frith 2003). Patients with focal epilepsy provide a unique test bed to study functional implications of lesions/disruptions of the core parts of the ToM network, as seizures most commonly emanate from temporal and frontal lobes. Only a small number of studies, however, investigated ToM in adults with temporal lobe epilepsy (TLE) and frontal lobe epilepsy (FLE). Deficits in ToM have not been consistently found (Giovagnoli et al. 2009, 2011; Farrant et al. 2005). For example, in patients with TLE Li et al. (2013) found deficits on three advanced ToM tasks (faux pas, strange stories, cartoon ToM). In contrast, Giovagnoli et al. (2009) found no deficits on the faux-pas task in patients with TLE. There is some evidence that side and site of seizures/ pathology within the temporal lobes affect the severity of ToM deficits. For example, right-sided seizure focus and/or lesions in the mesial temporal lobes have been associated with more marked ToM deficits in some studies (Giovagnoli et al. 2011; Schacher et al. 2006). One study, however, has found no significant relations between side and site of seizures/pathology and ToM (Li et al. 2013). Patients with FLE were impaired on faux
Neuropsychol Rev
pas and cartoon ToM tasks, but not on the strange stories (Farrant et al. 2005), despite all three of these tasks purportedly measuring similar components of ToM. At present, it is unclear which subgroups of patients with epilepsy are impaired in ToM and on which measures their deficits are apparent. In patients with focal epilepsy, age of seizure onset and duration of epilepsy have also been found to relate to the gravity and type of ToM deficits. Patients with early-onset TLE (mean age of seizure onset = 7 years-old) were more markedly impaired in affective ToM compared to patients with later-onset TLE (mean age of seizure onset = 24 yearsold), whereas the cognitive ToM abilities of these two groups did not differ (Giovagnoli et al. 2011). This study measured cognitive and affective ToM via the faux pas task. The faux pas task contains a number of questions that assess the ability to detect social actions that are reprehensible/morally wrong (faux-pas detection), understand beliefs/intentions (cognitive Tom) and predict emotions of others (affective ToM). It is designed for children from 9 to 11 years old. Nevertheless, the individual questions assess skills that differ slightly in the age at which they are acquired. The cognitive ToM question (BWhy did he/she say that?^) requires participants to understand the beliefs/intentions of another person; a skill that is acquired at around 4 to 6 years old (Pons et al. 2004). In contrast, the affective ToM question (BHow did he/she feel?^) assesses the ability to predict the emotions of a second character based on an external event, which develops at around 8 years old (Pons et al. 2004). Thus, onset of seizures during a critical period of ToM development may have disrupted the development of affective ToM that is assessed on this task, but spared cognitive ToM, which was already developed by this age. These findings are consistent with developmental neuropsychological literature, which has found that skills are particularly vulnerable to disruption during periods of intense development, so called critical periods (Anderson et al. 2011; Anderson et al. 2009; Anderson and Moore 1995). In addition to age of onset, age at assessment is also important to consider, particularly when studying children and adolescents, as deficits in later developing ToM skills may not be apparent until those skills come online in typically developing children. To date little research has been conducted with patients with generalised epilepsy (i.e., Idiopathic Generalised Epilepsy; IGE) (Jiang et al. 2014; Lew et al. 2015). Although in these patients seizures do not emanate from either the frontal or temporal lobes, generalised seizures may interfere with functioning of these brain regions (Engel 2001). Indeed, it has been shown that IGE is associated with deficits in early social cognitive processes (i.e., facial emotion identification) (Gomez-Ibañez et al. 2014) and ToM, as measured by the strange stories tasks (Lew et al. 2015) and the RMET in one study (Jiang et al. 2014), but not another (Lew et al. 2015). These deficits in social cognition are unlikely to be secondary to low IQ, as IQ is neither impaired in patients with IGE
(Nolan et al. 2003) nor related to ToM in typically developing children (Rajkumar et al. 2008). IGE participants, however, present with selective cognitive impairments, such as impaired EF (e.g., working memory, inhibition, and attentional flexibility) (Gelžinienė et al. 2010), which are critical for ToM in healthy individuals (Carlson et al. 2002). Thus, deficits in ToM in children with IGE may be related to impaired EF. This may partly account for ToM deficits in patients with TLE and FLE, who also experience impairments in executive skills (Black et al. 2010; Culhane-Shelburne et al. 2002; Gelžinienė et al. 2010). At present, it is not known to what degree EF contributes to ToM impairments in patients with different types of epilepsy (TLE, FLE and IGE) and on different ToM tasks. Finally, treatment factors, such as and anti-epileptic drugs (AEDs) and surgery, are important to consider, as they may impact upon ToM. Long-term use of multiple AEDs has been associated with problems with inattention, working memory, and verbal skills (Kwan and Brodie 2001; Park and Kwon 2008), which are all critical for ToM (Hughes 1998). Epilepsy surgery, which commonly involves resection of temporal or frontal regions, may also be expected to carry a risk of a decline in ToM abilities. Contrary to this expectation, a review that examined the effects of resective epilepsy surgery of temporal and frontal lobes on aspects of social cognition that are closely related to ToM (e.g., ability to identify emotions from faces and voices, to remember details from personally experienced life events; autobiographical memory), found no evidence of a decline in social cognitive function pre to post surgery (Kirsch 2006). The authors suggest that although surgery produces anatomic lesions, it may lessen the frequency and severity of seizures and/or lead to a reduction in AEDs, which may improve overall social cognitive functions (Kirsch 2006). ToM, however, was not specifically examined.
Rationale for the Present Review There is increasing evidence that patients with epilepsy are impaired in ToM. At present, however, the literature neither agrees on the severity of ToM impairment in this patient population, nor on the best tools to be used for its assessment. While a recent review, conducted by Giovagnoli (2014), outlined findings of studies examining ToM in patients with epilepsy, this review was limited in a number of respects. First, the review of Giovagnoli (2014) was restricted to studies involving adults with epilepsy and had little engagement with developmental literature, in which ToM originated. Omission of studies involving children with epilepsy and lack of discussion of the developmental trajectory of ToM (as evidenced in ToM tasks) is an important shortcoming, as presence of ToM deficits at an early age may provide evidence of a trait (a biomarker) of certain types of epilepsy, which have been found to present with marked ToM deficits in adulthood (TLE and FLE). Moreover, early ToM deficits may
Neuropsychol Rev
have detrimental impact on development of social skills, which are known to be reduced in adults with epilepsy. If ToM deficits are found in children with epilepsy, this could have significant clinical implications, as early diagnosis and treatment may reduce or minimise deficits in social skills. Further engagement with the developmental literature on ToM can provide important insight into cognitive mechanisms that underpin ToM and inform suitable interventions. Second, Giovagnoli (2014) did not conduct a systematic search of the literature, and as such several key papers were missed. In addition, a number of key papers have subsequently been published (i.e., Jiang et al. 2014; Lew et al. 2015). Third, Giovagnoli (2014) provided a narrative review only. Meta-analysis was not conducted. Meta-analysis enables more precise characterisation of the magnitude of deficits in ToM in patients with epilepsy and allows examination of relations between epilepsy and/or cognitive variables and ToM. Metaanalysis could also help identify the tools/tasks that are most sensitive to detecting ToM deficits in patients with epilepsy. As such, findings of meta-analysis could improve diagnosis of ToM in patients with epilepsy, by establishing which epilepsy related factors place patients at risk of ToM deficits and by guiding selection of tests to be included in neuropsychological clinical assessments. Fourth, the review of Giovagnoli (2014) did not undertake a quality assessment of the published papers. The purpose of the current study is to further this earlier review by addressing its shortcomings: employing a developmental conceptual stance, including both child and adult studies, using a systematic approach to literature search, utilising meta-analysis to establish the magnitude of deficits in ToM in patients with epilepsy, identifying factors that are associated with risk a ToM deficits in patients with epilepsy and tasks that are most likely to detect these deficits. The primary aims of this review are to determine whether the magnitude of ToM deficits varies as a function of the site of epilepsy focus and/or the type of task used to assess ToM. The secondary aims are to establish whether ToM performance is moderated by epilepsy variables (age of seizure onset, duration of epilepsy, seizure frequency), treatment factors (no. AEDs, surgery) and/or demographic factors (age of participants at testing) and to establish whether ToM is related to early social cognitive abilities (facial emotion recognition) and/or relevant cognitive skills (EF, verbal ability) in both children and adults with epilepsy.
on 1st March 2015. The following terms were used [((Theory of Mind) OR (Theory AND of AND mind) OR (social cognition) OR (social AND cognition) OR (social perception) OR (social AND perception) OR (social behaviour) OR (social AND behaviour) OR (perspective taking) OR (perspective AND taking) OR (mentalising) OR (mentalising) OR (mind reading) OR (mind AND reading) OR (empathy)) AND ((epileps*) OR (seizure*))]. No date limits were placed on any of the database searches. All Medical Subject Heading (MeSH) terms were exploded to broaden the search for relevant studies. In addition, the reference lists of relevant reviews and empirical studies were searched to identify further studies, as per the ancestry method. Study Selection Criteria Inclusion Criteria Eligible studies included (a) Patients with a diagnosis of epilepsy. (b) A healthy control group. (c) Used one of five empirically validated and psychometrically sound behavioural tasks to measure ToM: False Belief, Reading the Mind in the Eyes Task (RMET), Fauxpas, Strange Stories, Cartoon ToM vignettes (Henry et al. 2015). See Appendix 1 for details of these tasks. (d) Reported original research. (e) Were published in the English language and involved human participants. (f) Were published in peer reviewed journals. Excluded studies included a) Single case studies. b) Studies that grouped patients with different site of seizure focus together (e.g., FLE, TLE, other focal origin) and presented one ToM score for this group. c) Non-clinical outcome studies (i.e., reviews or studies validating ToM measures). d) Studies that used interviews, behavioural tasks or questionnaires to examine ToM other than those specified above. e) Studies that did not report adequate data to calculate a mean, weighted effect.
Method Quality Ratings of Selected Papers Search Strategy The search was conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) (Moher et al. 2009). PsycINFO, PubMed/ MEDLINE, and EMBASE were searched via OVID by ES
The Downs and Black checklist (1998) was used to evaluate the overall methodological quality of the included studies (Downs and Black 1998). This checklist measures quality of both randomized clinical trials and nonrandomized studies using several items distributed across the following subscales: reporting,
Neuropsychol Rev
Identification
4377 records identified through database searching (PsychINFO, PubMed/MEDLINE, and EMBASE) on 1st March 2015
6 additional articles identified through other sources (ancestry method)
PsychINFO MEDLINE 987 1691 EMABASE 1699
Screening
4056 records after duplicates removed
4056 records screened (titles and abstracts) 3633 records excluded
21 full text articles assessed for eligibility
Eligibility
9 articles excluded 2 studies did not measure ToM 7 studies measured ToM but did not meet other inclusion criteria: o Behavioural task other than one of the five tasks specified in selection criteria (n = 2) o Case studies (n = 2) o Case study with no control group (n = 1) o Data reported not sufficient to calculate an effect (n = 2)
Included
external validity, internal validity (bias and confounding), and power. The checklist demonstrates good test-retest reliability (r = .88), inter-rater reliability (r = .75), and internal consistency (Kruder-Richardson formula 20 = .89) (Downs and Black 1998). The original checklist contains 27 items, however, ten items relating to interventional trials studies were excluded as no interventional studies were identified (See Table 5 full details of items). Items are grouped into those assessing quality of reporting (items 1–8), external validity (item 9) and internal validity: statistical and methodological bias (items 10–12) or selection bias (items 13–17). Each item scored 0 (no or cannot tell) or 1 (yes), except for item 4, which was scored 0 (no), 1 (partially) or 2 (yes). Two of these items (items 7 and 16) relating to loss of patients at follow-up were only applicable to longitudinal studies. Each article could therefore achieve a score ranging from 0 to 18 points (longitudinal studies) or 0 to 16 points (cross-sectional studies). Papers were categorized as: high chance of bias (0 to 5 points), average chance of bias (6 to 11 points) and low chance of bias (12 to 18 points). The Downs and Black checklist is a comprehensive, albeit general checklist that does not consider various clinical variables that are known to impact cognition in patients with epilepsy (e.g., surgical status, age of onset/duration of epilepsy), which is important for our study. For this reason, a modified version of a second quality checklist, designed specifically for patients with epilepsy, was employed (Hrabok et al. 2013). This checklist included four items relating to the reporting of specific clinical and demographic information, inclusion/ exclusion criteria and recruitment methods. The original checklist did not include a specific item to differentiate patients with respect to surgical prospect or status, as it was designed specifically for surgical patients. Thus a fifth item that assesses whether patients had undergone surgery or were candidates for surgery was added to the scale used in our study. Items 1 and 2 were scored as 0 (no), 1 (partially) or 2 (yes). Item’s 3 and 4 were scored as 0 (no) or 1 (yes). Item 5 was a categorical item (surgical candidates, pre-surgery, postsurgery, non-surgical). Two reviewers (ES, SL) applied these checklists to all studies. Occasional discrepancies in ratings were resolved via discussion. Inter-rater reliability ranged from 90 to 100 % on each article reviewed.
See Appendix 1, Table 6 for details of the 7 studies. 12 studies included in metaanalysis
Fig. 1 Flow diagram of identification and selection of studies
only if the article clearly met inclusion criteria, but also if the reviewers could not determine with certainty whether inclusion criteria were met. Six additional papers were identified by hand searching the reference lists of these papers as well as 2 relevant reviews identified in the search. The manuscripts of these 21 studies were reviewed by two independent raters (ES and CR), and discrepancies were resolved via discussion and consensus. Nine articles were excluded after this, 2 of which did not measure ToM and 7 of which measured ToM but did met other exclusion criteria: used a behavioural task other than those specified in selection criteria (n = 2), case studies (n = 3), no control group (n = 1), and/or did not report sufficient data to answer the research questions (n = 3). See Table 7 in Appendix 1 for details of these 6 studies.
Methods of Review Data Extraction Figure 1 displays the flow diagram describing the process of study selection for the review. The initial search retrieved 4377 articles. After duplicates were removed, 4056 articles remained. All titles and abstracts were screened by one reviewer (ES). A second independent reviewer (CR) screened a random selection of 20 % of these articles to ensure inter-rater reliability, which was 95 %. Full manuscripts of 15 studies that seemed to meet inclusion criteria were obtained. A conservative approach was taken, such that articles were obtained not
Prior to conducting the meta-analysis, the following was extracted from each paper: 1. Name of first author and year of publication. 2. Number of participants in epilepsy and control groups. 3. Means of demographic variables (age, years of education), epilepsy variables (type, age of onset, duration, seizure frequency), treatment factors (surgery, number of AEDs),
Neuropsychol Rev
basic social cognition (Facial Emotional Recognition; FER) and relevant cognitive variables (EF; verbal ability). 4. Means and SDs of each group on TOM tasks. Where means and SDs were not available, t-values were obtained, followed by p-values. 5. A mean effect size (Hedges’ g) was calculated for each individual task. 6. For studies that measured more than one ToM task, a mean weighted effect size was calculated for the primary analysis. If papers did not report necessary statistical or demographic information, authors were contacted and those who did not respond within 4 weeks of an email being sent were excluded from the analysis (n = 3).
Data Analysis Data was analysed using the Comprehensive Meta Analyses Program, Version 3 (CMA; Borenstein et al. 2005). Hedges’ g was used as a measure of the standardized difference between epilepsy and control groups, as it corrects for biases that can lead to overestimation of standardized mean difference in small samples (Borenstein et al. 2009). This effect size is interpreted just like Cohen’s d: 0.2 represents a small effect, 0.5 is a medium effect, and 0.8 is a large effect (Cohen 1988). Whenever patients with epilepsy performed poorer than controls, we reported between-group differences as positive effect sizes. A significance level of p < 0.05 was used for all analyses, including main effects, meta-regressions, and publication bias. All tests were 2-tailed. The primary analysis examined ToM performance as a function of epilepsy group: TLE, FLE, extra-TLE/FLE (focal seizures outside the temporal and frontal lobes), IGE. For studies that used more than one ToM task (n = 5) a mean weighted effect size was calculated for each epilepsy group, as this prevents handpicking of tasks and eliminates bias in extracting data (Higgins and Green 2008). In studies that included more than one group of epilepsy patients (n = 4), each epilepsy group was compared to the control group separately. All studies that included more than one group of patients with epilepsy examined patients with TLE and either FLE or extra-TLE/FLE. To compare effect sizes between groups, TLE patients were examined in a separate analysis to other epilepsy groups, as this avoided a unit-of-analysis error that arises with multiple correlated comparisons (Higgins and Green 2008). This allowed for performance of patients with FLE, extra-TLE/FLE and IGE to be compared to one another, as there was no overlap of control groups within this analysis. An unavoidable limitation was that the effect size for TLE patients could not be directly compared to other epilepsy groups.
The secondary analysis examined whether effect sizes varied as a function of the task type. For studies that included more than one epilepsy group (n = 4), scores obtained on each ToM task by patients with epilepsy were collapsed and an overall effect size was calculated for each task within each study (Higgins and Green 2008). In both analyses homogeneity of the resulting mean weighted effect sizes were tested with Q test. A random effects model was used for the meta-analysis. Moderator analyses for continuous variables employed meta-regressions to assess whether age of seizure onset, duration of illness, seizure frequency, no AEDs. AEDs, and/or age of participants at testing were related to ToM in each epilepsy group. Meta-regression could not be run for non-continuous moderator variables (side and site of seizures/pathology, surgery), as there was an overlap of control groups within studies. Instead, studies reporting on non-continuous moderator variables were reviewed in systematic manner. Publication bias was tested using funnel plots and Egger’s regression test (Egger et al. 1997). To reduce the risk of false positive results and to further investigate the source of funnel plot asymmetry, tasks with a significant asymmetry (Egger’s test, p < 0.05) were further analysed. Individual characteristics of the studies were investigated and a Fail Safe number (number of negative studies necessary to make the group difference insignificant) was calculated.
Results Study Characteristics and Patient Demographics Please see Table 1 for details of studies included in this manuscript. Overall, the studies included 55 patients with FLE (3 studies), 437 patients with TLE (8 studies), 62 patients with IGE (2 studies), and 41 patients with extra-TLE (2 studies). The mean age of epilepsy onset ranged from 11.8 to 26.1 years old (FLE), 13.3 to 21.4 years old (TLE), 15.6 to 18.6 years old (Extra-TLE/FLE) and 7.6 to 12.0 years old (IGE). The mean duration of epilepsy was 8.3 to 22.6 years (FLE), 13.3 to 22.7 years (TLE), 14.8 to 20.3 years (Extra-TLE/FLE) and 3.0 to 3.4 years (IGE) . The mean age of participants at testing ranged from 34.4 to 37.1 (FLE), 32.2 to 42.3 (TLE), 33.4 to 35.9 years (Extra-TLE/FLE) and 11.6 to 16.5 years (IGE) All studies included a group of healthy controls with no history of seizures or epilepsy and/or neurological or psychiatric disability. Control groups were matched with epilepsy groups on age, gender, years of education and IQ. Meta-Analysis by Epilepsy Group Meta analyses demonstrated that patients with FLE (n = 3, g = 1.025, 95 % CI 0.717–1.334, Z = 6.541, p < 0.0001), TLE
37.8
37.17
TLE (n = 54) HC (n = 42)
FLE (n = 12) HC (n = 42)
IGE (n = 42) HC (n = 47)
IGE (n = 20) HC (n = 57)
Jiang et al. (2014)
Lew et al. (2015)
11.6
16.5
26.07
35.77
FLE (n = 29; 4 left, 3 right, 22 bilateral) HC (n = 69)
Giovagnoli et al. (2013)
21.39
37.01
TLE (n = 109; 62 left, 47 right) HC (n = 69)
Giovagnoli et al. (2011)
7.60
12.00
25.33
18.70
17.43
39.76
TLE (n = 21) HC (n = 21)
Giovagnoli et al. (2009)
11.80
18.57
20.21
34.36
33.36
34.40
FLE (n = 14; 8 left, 5 right) HC (n = 14)
HC (n = 29)
Extra-TLE/FLE (n = 14)
HC (n = 29)
TLE (n = 28; 17 left, 11 right)
Farrant et al. (2005)
Broicher et al. (2012)
HC (n = 20)
17.50
3.40
3.00
11.83
18.89
8.32
15.58
22.67
22.56
14.79
14.25
18.0
Faux-pas
Faux-pas
–
–
–
2.09
–
–
1.31
–
14.73
9.33
2.09
1.91
95.10 (FSIQ)
RMET Strange stories
RMET
Faux-pas
–
8.91
2.09
Faux-pas
9.17
–
–
1.95
–
Strange stories RMET Cartoon ToM Faux-pas
95.57 (VIQ)
RMET
Faux-pas
RMET
Faux-pas
Faux-pas
105.0 (VIQ)
101.3 (VIQ)
Faux-pas
96.57 (FSIQ)
–
–
–
97.0 (FSIQ)
–
–
–
–
3.25
Task
0.204 (0.258) 0.810 (0.265)
0.717 (0.217)
1.162 (0.341)
0.880 (0.214)
1.261 (0.237)
0.665 (0.157)
0.092 (0.303)
0.979 (0.390) 0.901 (0.386)
0.431 (0.371)
0.921 (0.387)
0.296 (0.321)
0.279 (0.329)
0.392 (0.264)
0.948 (0.279)
0.690 (0.255)
Hedges' g (SE)
0.498 (0.185)
0.717 (0.217)
1.162 (0.341)
0.880 (0.214)
1.261 (0.237)
0.665 (0.157)
0.092 (0.303)
0.799 (0.192)
0.288 (0.229)
0.655 (0.192)
0.690 (0.255)
Mean weighted effect size
35.50
FSIQ/ VIQ
No. AED
TLE (n = 74; 37 left, 37 right)
Seizures p/month
Amlerova et al. (2014)
Duration (years)
Mean age (years)
Groups
Study
Mean age at onset (years)
Characteristics of studies included in the meta-analysis and mean weighted effect sizes for Theory of Mind (ToM) tasks
Table 1
Side. NA Site. NA Surgery. NA
Surgery. NR Side. NR Site. Reported but not compared (33 MTS, 21 non-MTS) Surgery. NR Side. NR Site. NR Surgery. NR Side. NA Site. NA Surgery. NA
Site. NA
Side. Reported but not compared Site. MTS < non-MTS, p = 0.011 Surgery. NR Side. Reported but not compared
Side. NR Site. NR Surgery. NR
Surgery. All patients pre-surgery
Site. NR Surgery. NR Side. Reported but not compared Site. NR
Site. All patients w MTS Surgery. NR Side. NA
Side. No sig. diff between left and right-TLE, p = 0.27. Both groups < HC, p < 0.05 Site. NR Surgery. No sig. change pre to post temporal lobe resection, p > 0.05, (n = 30, same patients) Side. Reported but not compared
Within epilepsy group comparisons (side, site, surgery)
NM
IGE < HC, p < 0.05. Correlation with ToM, n.s.
NM
NM
NM
NM
NM
Not examined in relation to ToM
Correlation between FER and ToM, n.s. Extra-TLE/ FLE < HC, p < 0.05 Correlation with ToM, n.s FLE < HC, p < 0.05.
TLE < HC, p < 0.05
Not examined in relation to ToM
TLE < HC, p < 0.05
Facial emotion recognition (FER)
IGE < HC, p < 0.05 (Stroop task). Non-sig. correlation for EF and ToM NM
NM
NM
Non-sig. correlation for EF and ToM
FLE < HC, p < 0.05. (Tower of London)
TLE < HC, p < 0.05 (Tower of London) Not examined in relation to ToM TLE < HC, p < 0.05 (Tower of London) sig correlation for EF and ToM, p < 0.05
Extra-TLE/FLE < HC, p = 0.03 (Iowa Gambling Task). Not examined in relation to ToM FLE < HC, p < 0.05 (Hayling Test) No. sig correlation for EF and ToM.
TLE < HC, p = 0.02 (Iowa Gambling Task) Not examined in relation to ToM
NM
Executive function (EF)
Neuropsychol Rev
HC (n = 30)
TLE (n = 67)
HC (n = 38)
32.19
33.50
18.51
14.5
15.60
35.90
ExtraTLE/FLE (n = 27) HC (n = 12)
TLE (n = 26)
13.30
36.50
24.58
42.33
TLE (n = 27; 13 left-TLE, 14right-TLE) HC (n = 12)
Mean age at onset (years)
Mean age (years)
13.72
19.00
20.30
22.20
19.31
Duration (years)
–
–
2.61
–
–
3.2
–
2.09
No. AED
–
1.32
Seizures p/month
91.00 (VIQ)
93.10 (FSIQ)
97.00 (VIQ)
100.00 (Performance IQ)
109.5 (FSIQ)
107.2 (FSIQ)
103.12 (VIQ)
99.10 (FSIQ)
FSIQ/ VIQ
1.674 (0.249)
Strange stories Cartoon ToM
1.596 (0.246)
1.094 (0.232)
1.156 (0.233)
1.525 (0.285)
0.443 (0.344)
1.215 (0.367)
1.029 (0.285)
1.6599 (0.311) 0.961 (0.283)
0.763 (0.278)
Hedges' g (SE)
Faux-pas
Strange stories False belief
Faux-pas
False belief
Faux-pas
Strange stories Cartoon ToM Faux-pas
Faux-pas
False belief
Task
1.364 (0.119)
1.525 (0.285)
0.443 (0.344)
1.215 (0.367)
1.075 (0.144)
Mean weighted effect size
Site. Reported but not compared (36 MTS, 31 non-MTS) Surgery. All patients pre-surgery
Site. NR Surgery. NR Side. No sig. diff between left-TLE and right-TLE p = 0.98, both groups < HC, p < 0.05 Site. No sig. diff between amygdala damage or no amygdala damage Surgery. All patients post-surgery Side. NR
Side. Right-TLE < leftTLE, p < 0.05. Both groups < HC, p < 0.05 Site. All had MTS Surgery. No sig diff. between pre-surgery and (n = 16) and post-surgery (n = 11) groups. Temporal lobe resection (separate groups), p > 0.05 Side. NR
Surgery. NR
Side. No sig. diff between left, right, or bilateral, p > 0.05 Further pairwise comparisons not performed. All groups < HC, ps < 0.05. Site. NA
Within epilepsy group comparisons (side, site, surgery)
NM
NM
NM
NM
TLE = HC, p > 0.05 (Modified Card Sorting Task)
NM
TLE = HC, p = 0.06–0.21 (Modified card sorting task) Not examined in relation to ToM
Not examined in relation to ToM
TLE < HC, p < 0.05 (Hayling and Brixton tests).
NM
NM
Not examined in relation to ToM
Executive function (EF)
Facial emotion recognition (FER)
AED Anti-Epileptic Drugs, FSIQ Full Scale IQ from the Weschler Adult Intelligence Scale – Third Edition or Fourth Edition (WAIS-III or WAIS-IV), VIQ Verbal IQ from the WAIS-III or WAIS-IV, FLE Frontal lobe epilepsy, TLE Temporal lobe epilepsy, IGE Idiopathic generalised epilepsy, Extra-TLE/FLE Patients with focal seizures outside thetemporal and frontal lobes, HC Healthy controls, RMET Reading the Mind in the Eyes Task, ToM Theory of Mind, MTS mesial temporal sclerosis, NR Not reported, NM not measured
Wang et al. (2015)
Shaw et al. (2004)
Schacher et al. (2006)
TLE (n = 31; 11 left-TLE, 13 right-TLE, 7 bilateral-TLE)
Li et al. (2013)
HC (n = 24)
Groups
Study
Table 1 (continued)
Neuropsychol Rev
Neuropsychol Rev
(n = 9, g = 0.915, 95 % CI 0.652–1.178, Z = 6.820, p < 0.0001), and IGE (n = 2, g = 0.590, 95 % CI 0.314–0.866, Z = 4.191, p < 0.001) performed significantly below controls on ToM tasks (Table 2, Fig. 2). In contrast, scores of patients with extra-TLE/FLE (focal seizures originating outside the temporal and frontal lobes) did not differ significantly from scores obtained by controls on ToM tasks (n = 2, g = 0.336, 95 % CI −0.038–0.710, Z = 1.759, p = 0.079). A between group comparison was conducted to compare performance of patients with FLE, IGE and extra-TLE/FLE. Analysis with Q-test revealed significant heterogeneity between groups (Q = 8.470, df = 2, p = 0.014). Further analyses showed significantly greater deficits in patients with FLE relative to patients with IGE (p = 0.039) and extra-TLE/FLE (p = 0.005). Patients with IGE did not differ significantly from those with extra-TLE/FLE (p = 0.284). The TLE group could not be compared to patients with FLE, IGE or extra-TLE due to overlap of control groups within studies.
Meta-Analysis by Task Type Seven studies used a single behavioural task to measure ToM and 5 studies used multiple tasks. The tasks used included false belief tasks (3 studies), RMET (4 studies), faux pas (9 studies), strange stories (5 studies), and cartoon ToM vignettes (3 studies). A task level analysis was conducted to determine whether effect sizes differed as a function of tasks (see Table 3). All tasks detected significant differences between epilepsy and control groups and there were no significant differences in effect sizes between tasks (Q = 2.334, df = 4, p = 0.675) (Table 3). Examination of mean weighted effect sizes revealed that the largest to smallest effects were observed for cartoon ToM (g = 1.231), strange stories (g = 1.004), and false belief (g = 0.988), faux pas (g = 0.833), and RMET (g = 0.517) (Table X). An inspection of mean weighted effect sizes revealed medium to very large (0.618 to 1.660) effects sizes across studies and across tasks, except for one study on one of the tasks (see Table 4). In the study by Giovagnoli et al. (2009) the effect size of the Faux pas task was small (g = 0.092, p = 0.760). Further inspection of the studies indicated that while Giovagnoli et al. (2009) used ‘Faux-pas hits’, which measures Table 2 Mean weighted effect sizes as a function of epilepsy group
Epilepsy group
FLE TLE IGE Extra-TLE/FLE
Number of studies
3 9 2 2
the ability to correctly identify faux-pas, all other studies employed the cumulative score obtained on the Faux-pas task, which is a composite score that includes questions assessing both cognitive and affective ToM. We re-ran the analyses without this outlier study to establish the effect size for the Faux-pas task and examine between task differences. The magnitude of the effect size for the Fauxpas tasks changed from medium (n = 9, g = 0.833 95 % CI 0.553–1.041, Z = 6.394, p < 0.0001) to large (n = 8, g = 0.894, 95 % CI 0.706–1.082, Z = 9.310, p < 0.0001). Heterogeneity of mean weighted effect sizes between tasks remained insignificant (Q = 2.625, df = 4, p = 0.622). Meta-Regression of Moderator Variables Meta-regression only included studies that involved patients with TLE or FLE. Meta-regression was not conducted for patients with IGE or extra-TLE/FLE, as the number of studies that examine impact of moderators on TOM in these patient populations was below the minimum (n ≥ 3) required for meta-regression (Borenstein et al. 2009). Meta-regression revealed a significant negative association between effect sizes and one moderator variable: age at testing in studies including patients with TLE (R = − 0.04, 95 % CI −0.072–−0.004, Z = −2.194, p = 0.028), but not in studies including patients with FLE. The magnitude of ToM deficits increased as participants’ age decreased. The effect sizes of ToM deficits were not associated with other epilepsy variables: age of seizure onset, disease duration, seizure frequency and/or number of AEDs in either patients with TLE or FLE (all ps > 0.05). Systematic Review of Other Moderator variables Side of Seizures/Pathology Four studies compared ToM performance in patients with left-TLE and right-TLE. Of these, one study found significantly worse performance in patients with right-TLE compared to left-TLE (p < 0.05; Schacher et al. 2006) and three studies found no differences between right-TLE, left-TLE and/or bilateral-TLE (ps =0.21–0.98; Amlerova et al. 2014; Li et al. 2013; Shaw et al. 2004). Further pairwise comparisons were not performed in these studies. In
Number of participants Epilepsy
Control
55 437 62 41
125 285 104 41
Hedges' g
95 % CI
Z
P
1.025 0.915 0.590 0.336
0.717–1.334 0.593–1.086 0.314–0.866 −0.038–0.710
6.541 6.671 4.191 1.759
p < 0.001 p < 0.001 p < 0.001 p = 0.079
CI Confidence interval, FLE frontal lobe epilepsy, TLE temporal lobe epilepsy, IGE idiopathic generalised epilepsy, Extra-TLE/FLE focal seizures outside the temporal and frontal lobes
Neuropsychol Rev Fig. 2 Forest plot of individual and mean weighted effect sizes for patients with FLE, TLE, IGE and extra-TLE/FLE
all studies, however, patients performed significantly more poorly than healthy controls, regardless of the side of seizure/ pathology (ps < 0.05). Three other studies reported numbers of patients with left or right-sided seizures/pathology (TLE or FLE), but did not statistically compare groups (Broicher et al. 2012; Farrant et al. 2005; Giovagnoli et al. 2011).
Site of Seizure Focus/Pathology in TLE Two studies examined whether site of seizure focus is related to ToM in patients with TLE. One study compared ToM scores obtained by patients with TLE and mesial temporal sclerosis (TLE/MTS+) to patients with TLE without mesial temporal sclerosis (TLE/ MTS-) (Giovagnoli et al. 2011). Patients with TLE/MTS+ Table 3 Mean weighted effect sizes as a function of Theory of Mind (ToM) task
Task
False Belief RMET Faux pas Strange stories Cartoon ToM
Number of studies
2 4 9 4 3
performed significantly more poorly than patients with TLE/ MTS- (p < 0.05; Giovagnoli et al. 2011). Another study examined the impact of amygdala lesion and age at which this lesion was obtained on ToM in patients with TLE (Shaw et al. 2004). Patients with early amygdala damage obtained significantly lower ToM scores relative to patients with late amygdala damage and extra-amygdala damage. While patients with early amygdala damage were also impaired on ToM task compared to controls, patients with late amygdala damage and extra-amygdala damage were not.
Surgery Two studies investigated impact of temporal lobe resection on ToM (Amlerova et al. 2014; Schacher et al.
Number of participants Epilepsy
Control
98 117 479 132 112
54 147 261 125 68
Hedges' g
95 % CI
Z
P
0.988 0.517 0.833 1.004 1.231
0.607–1.370 0.212–0.822 0.608–1.057 0.491–1.518 0.792–1.671
5.076 3.319 7.281 3.835 5.488
p < 0.0001 p = 0.001 p < 0.0001 p < 0.0001 p < 0.0001
CI Confidence interval, RMET Reading the Mind in the Eyes Task, ToM Theory of Mind
Neuropsychol Rev Table 4
Mean weighted effect sizes for individual ToM tasks with epilepsy patients grouped together within studies
Studies
Group(s)
Pooled N
Effect Size Hedges' g (SE) by Task and Study False belief
RMET
Faux pas 0.690 (0.255) 0.618 (0.247)
Amlerova et al. (2014) Broicher et al. (2012)
TLE TLE, Extra-TLE/FLE
74 41
0.352 (0.241)
Farrant et al. (2005)
FLE
14
0.979 (0.390)
Giovagnoli et al. (2009) Giovagnoli et al. (2011) Giovagnoli et al. (2013)
TLE TLE, FLE TLE, FLE
21 138 66
Jiang et al. (2014) Lew et al. (2015)
IGE IGE
42 20
Li et al. (2013)
TLE
31
Schacher et al. (2006) Shaw et al. (2004)
TLE, Extra-TLE/FLE TLE
27 26
Wang et al. (2015) No of studies
TLE
67
Hedges' g
0.921 (0.387) 0.092# (0.303)
Strange stories
Cartoon ToM
0.431 (0.371)
0.901 (0.386)
0.811 (0.152) 0.862 (0.205) 0.717 (0.217) 0.204 (0.258) 0.763 (0.278)
0.810 (0.265) 1.660 (0.311)
0.961 (0.283)
1.029 (0.285)
0.804 (0.251) 1.156 (0.233) 2
4
1.094 (0.232) 9
1.674 (0.249) 4
1.596 (0.246) 3
0.988 (0.195)
0.517 (0.156)
0.833 (0.114)
1.004 (0.262)
1.231 (0.224)
FLE Frontal lobe epilepsy, TLE Temporal lobe epilepsy, IGE Idiopathic generalised epilepsy, Extra-TLE/FLE Patients with focal seizures outside the temporal and frontal lobes, RMET Reading the Mind in the Eyes Task, ToM Theory of Mind # p > 0.05
2006). Both studies compared performance of a group of patients who were candidates for temporal lobe surgery to a group of patient who underwent surgery on ToM task. Neither of these studies found any difference between the two groups on ToM task. One of these studies also included a longitudinal arm in which 30 of the surgical candidates who subsequently underwent surgery were reviewed 12 to 24 months postsurgery (Amlerova et al. 2014). This longitudinal arm of the study revealed no change in patients’ ToM scores from pre- to post- surgery (Amlerova et al. 2014). We also identified three additional cross-sectional studies that included either only pre-surgery patients (Farrant et al. 2005; Wang et al. 2015) or only post-surgery patients (Shaw et al. 2004). Large effect sizes of comparable magnitude were observed in the pre-surgical (g = 1.364) and post-surgical patients (g = 1.525). Facial Emotion Recognition (FER) Four studies measured FER. All studies found marked FER impairments among patients with TLE (2 studies; Amlerova et al. 2014; Broicher et al. 2012), FLE (1 study; Farrant et al. 2005), Extra-TLE/FLE (1 study; Broicher et al. 2012), and IGE (1 study; Jiang et al. 2014). There were no significant associations, however, between FER and ToM (all ps > 0.05; Broicher et al. 2012; Jiang et al. 2014). Executive Functions (EF) Seven studies included measures of EF. Tasks included the Modified Card Sorting Task (2 studies Li et al. 2013; Wang et al. 2015), Iowa Gambling Task
(1 study; Broicher et al. 2012), Hayling and Brixton Tests (2 studies; Farrant et al. 2005; Shaw et al. 2004), Tower of London (2 studies; Giovagnoli et al. 2009; 2011), Stroop Task (1 study; Jiang et al. 2014) and Digit Span backward (1 study; Jiang et al. 2014). These tasks measure various components of EF, including inhibition/control, attention flexibility/setshifting, and/or working memory. All studies found marked EF impairments in patients with epilepsy, however, only three of the seven studies reported correlations between EF and ToM. One study found a significant correlation between EF (composite score of derived from tests of abstraction, working memory and planning abilities) and ToM among patients with TLE (Giovagnoli et al. 2011). The two other studies found no significant association between EF (inhibition) and ToM in patients with FLE (Farrant et al. 2005) or IGE (Jiang et al. 2014). Intellectual ability. Six studies reported Full Scale Intellectual Quotient (FSIQ) from the Weschler Scales (Weschler Adult Intelligence Scale–Third or Fourth editions, WAIS-III or WAIS-IV; Weschler Children Intelligence Scale–Fourth Edition, WISC-IV). One study found a significant correlation between FSIQ and ToM in patients with TLE (Amlerova et al. 2014), three studies found no significant association between FSIQ and ToM (Farrant et al. 2005; Lew et al. 2015; Shaw et al. 2004) and two studies did not examine this relationship (Li et al. 2013; Schacher et al. 2006). Three additional studies used the Raven’s Coloured Progressive Matrices to measure FSIQ, but none examined the relationship between FSIQ and ToM (Giovagnoli et al. 2009, 2011, 2013). Five studies
Neuropsychol Rev
measured verbal intellectual ability by the Weschler Scales (WAIS-III, WAIS-IV, WISC-IV). None of these studies found a significant association between Verbal IQ (VIQ) and ToM. Language skills. Three studies measured measured certain aspects of language skills, namely word finding and naming using the Word Fluency and the Boston Naming Tests, but none of these studies examined the relationship between language and ToM (Giovagnoli et al. 2009, 2011, 2013). Publication Bias Egger’s regression test indicated no significant publication bias (ps = 0.689–0.971, 2-tailed). The classic failsafe N, which represents the number of missing studies needed to nullify the effect, was estimated at 134 studies, which far exceeds Rosenthal’s criteria of K + 10 studies (where k = no. of studies included in the meta-analysis). We also inspected funnel plots visually and found no evidence of publication bias.
Quality Ratings For methodological quality ratings on the Downs and Black checklist please refer to Table 5. The ratings ranged from 12 to 16 out of 16 in cross-sectional studies (n = 11) and 13 out of 18 in the single longitudinal study (n = 1). Inspection of quality assessment criteria revealed that 10 of the 12 studies obtained perfect scores on items assessing quality of reporting. The other two studies obtained near perfect scores. With regard to statistical and methodological bias (internal validity), all studies obtained perfect scores. In contrast, there was somewhat lower quality in selection bias (internal validity), with only 3 studies reporting the time-period over which recruitment took place and 7 studies reporting whether participants in different groups were recruited from the same population. Adequate external validity was reported in 7 out of 12 studies, whereas the remaining 5 studies did not meet the single external validity criteria. Quality ratings on the epilepsy specific checklist (please see Table 6) ranged from 3 to 6 out of 6. All studies completely or partially reported demographic and clinical information. Nine studies reported inclusion/exclusion criteria, while three studies did not (Giovagnoli et al. 2009; Schacher et al. 2006; Shaw et al. 2004). All but one study (Giovagnoli et al. 2009) reported how patients were selected/identified (e.g., consecutive patients). With regard to surgery status, two studies included pre- and post-surgical patients (Amlerova et al. 2014; Schacher et al. 2006), one study included post-surgery patients only (Shaw et al. 2004) and two studies included patients who were considered as candidates for epilepsy surgery (Farrant et al. 2005; Wang et al. 2015). The remaining seven studies included non-surgical patients (i.e., those not being specifically assessed for epilepsy surgery).
Discussion The primary aims of this review were to establish the impact of the site of epilepsy focus on ToM and determine whether ToM tasks have differential sensitivity to ToM deficits in this patient population. The secondary aims were to explore whether epilepsy variables (side of seizure focus, age of seizure onset, duration of epilepsy, seizure frequency), treatment factors (number of AEDs, surgery) and/or demographic factors (age of participants at testing) moderated ToM in patients with epilepsy. Twelve studies were identified; ten included adults with focal seizures and two included children with generalised seizures. Our meta-analyses revealed that ToM was related to the site of epilepsy focus. While adult patients with seizures emanating from frontal or temporal lobes as well as children with generalised seizures had deficits in ToM, adult patients with seizures emanating from posterior brain regions did not. The meta-analyses also showed that deficits in ToM were documented across ToM tasks. Younger age at testing was associated with worse ToM in patients with TLE. In contrast, age of seizure onset, epilepsy duration, seizure frequency, number of AEDs and surgery were not related to ToM. As expected from our review of neuropsychological and neuroimaging literature, our meta-analyses revealed deficits in ToM in adults with FLE and TLE, but not extra-TLE/FLE. The effect sizes of ToM deficits differed significantly between the groups, suggesting that the site of epilepsy focus is likely to play a pivotal role in ToM in patients with epilepsy. Conversely, inspection of the magnitude of effect sizes shows (i) large effect sizes in patients with FLE (g = 1.025) and TLE (g = 0.915), and (ii) small effect size in patients with focal seizures emanating outside frontal and temporal regions (g = 0.336). While neuroimaging studies suggest that several different frontal and temporal brain regions are active during ToM task, the impact of site and side of epilepsy focus/ pathology on ToM has not been examined in patients within frontal lobes. Little evidence has been found that the side of epilepsy focus differently impact ToM in patients with TLE, although this was examined in 4 studies. The impact of site of epilepsy focus within temporal lobes requires further investigations, as it was only examined in two studies and findings were mixed. Importantly, we found that patients with generalised epilepsy (IGE) are also at risk of ToM deficits. The effect size was medium (g = 0.590), suggesting that deficits in ToM among patients with IGE were not subtle. The underpinnings of ToM deficits in patients with IGE are less clear than in patients with temporal and frontal seizure foci. In patients with IGE, ToM deficits cannot be attributed to underlying lesions and/or pathology, as IGE is diagnosed on the basis of a lack of structural abnormalities or lesions identified on brain scans (Sullivan and Dlugos 2004). It also does not appear to be the presence of seizures themselves that is causing ToM difficulties in
Neuropsychol Rev Table 5
Quality assessment of studies included in the review from general quality rating checklist
Studies
Quality assessment criteria from adapted Downs and Black Checklist Quality of reporting
Internal validity Statistical and methodological bias
Selection bias
1
2
3
4
5
6
7
8
External validity 9
10
11
12
13
14
15
16
17
Sum
Amlerova et al. (2014) Broicher et al. (2012)
1 1
1 1
1 1
1 2
1 1
1 1
0 NA
1 1
0 1
1 1
1 1
1 1
1 1
1 0
0 1
0 NA
1 1
13/18 15/16
Farrant et al. (2005)
1
1
1
2
1
1
NA
1
0
1
1
1
0
0
1
NA
1
13/16
Giovagnoli et al. (2009) Giovagnoli et al. (2011)
1 1
1 1
1 1
1 2
1 1
1 1
NA NA
1 1
0 1
1 1
1 1
1 1
0 1
0 1
1 1
NA NA
1 1
12/16 16/16
Giovagnoli et al. (2013) Jiang et al. (2014)
1 1
1 1
1 1
2 2
1 1
1 1
NA NA
1 1
1 1
1 1
1 1
1 1
1 0
0 0
1 1
NA NA
1 1
15/16 14/16
Lew et al. (2015) Li et al. (2013)
1 1
1 1
1 1
2 2
1 1
1 1
NA NA
1 1
1 1
1 1
1 1
1 1
1 0
1 0
1 1
NA NA
1 1
16/16 14/16
Schacher et al. (2006) Shaw et al. (2004)
1 1
1 1
1 1
2 2
1 1
1 1
NA NA
1 1
1 0
1 1
1 1
1 1
1 1
0 0
1 1
NA NA
1 1
15/16 14/16
Wang et al. (2015)
1
1
1
2
1
1
NA
1
0
1
1
1
0
0
1
NA
1
13/16
Items: (1) Is the hypothesis/aim/objective of the study clearly described? (2) Are the main outcomes to be measured clearly described in the Introduction or Methods section? (3) Are the characteristics of the patients included in the study described clearly? (4) Are the distributions of principal confounders in each group of subjects to be compared described clearly? (5) Are the main findings of the study described clearly? (6) Does the study provide estimates of the random variability in the data for the main outcomes? (7) Have the characteristics of patients lost to follow-up been described? (8) Have actual probability values been reported (for example, 0.035 rather than < 0.05) for the main outcomes except where the probability value is less than 0.001? (9) Were the subjects asked to participate in the study representative of the entire population from which they were recruited? (10) If any of the results of the study were based on ‘data dredging’, was this made clear? (11) Were the statistical tests used to assess the main outcomes appropriate? (12) Were the main outcome measures used accurate (valid and reliable)? (13) Were the patients in different groups recruited from the same population? (14) Were study subjects recruited over the same period of time? (15) Was there adequate adjustment for confounding in the analyses from which the main findings were drawn? (16) Were losses of patients to follow-up taken into account? (17) Did the study have sufficient power to detect a clinically important effect where the probability value for a difference being due to chance is less than 5 %? All items given 0 or 1 point, except for item 4 which was given 0, 1 or 2 points. Items 7 and 16 were not applicable to cross-sectional studies and were marked as NA (Not Applicable). Total possible score for longitudinal studies was 0 to 18 (1 study) and for cross sectional studies was 0 to 16 (11 studies) Items 1–8 (quality of reporting), item 9 (external validity), items 10–12 (internal validity: statistical and methodological bias), items 13–17 (internal validity: selection bias)
patients with IGE, as we found no association between seizure frequency and ToM (Jiang et al. 2014; Lew et al. 2015). Similarly, it is of interest to observe that seizure frequency also did not relate to ToM deficits in patients with either TLE or FLE (Amlerova et al. 2014; Giovagnoli et al. 2009). Finally, patients with extra-TLE/FLE, who experienced seizures that were of comparable frequency to other epilepsy groups, were not impaired in ToM. Together, these findings suggest that seizures alone play a minimal role in whether patients experience ToM difficulties or not. We now turn our attention to examination of epilepsy treatment factors that could impact ToM, especially surgery. The most common type of epilepsy surgery is temporal lobectomy, which typically involves resection of anterior temporal lobe (in combination with mesial temporal structures) that is activated during ToM tasks in healthy individuals (Olson et al. 2012; Olson et al. 2007). Thus temporal lobectomy could place patients at a risk of a drop in ToM. Only two studies
included in this review examined the impact of epilepsy surgery on ToM in patients who underwent temporal lobectomy for intractable epilepsy (Amlerova et al. 2014; Schacher et al. 2006). Both studies employed cross-sectional designs and compared a group of patients who were candidates for temporal lobectomy with another group of patients who have already undergone temporal lobectomy for epilepsy (Amlerova et al. 2014; Schacher et al. 2006). One of these studies also had an additional, longitudinal arm that followed-up the same group of patients from pre- to post-surgery (Amlerova et al. 2014). In all instance there was no evidence of a temporal lobectomy having a significant impact on ToM. While lack of significant findings is surprising, it is consistent with recent review that examined the effects of epilepsy surgery on another aspect of social cognition: facial emotion identification. This review found no decline in facial emotion identification pre- to post- either temporal or frontal lobe epilepsy surgery (Kirsch 2006). It was purported that although surgery produces
Neuropsychol Rev Table 6 Quality of studies included in the review from epilepsy specific checklist
Studies
Quality assessment criteria from an adapted version of the Hrabok et al. (2013) checklist
Scoring
0 = no, 1 = partially, or 2 = yes 1 2
3
4
Total
5
Amlerova et al. (2014) Broicher et al. (2012)
2 2
1 1
1 1
1 1
5/6 5/6
Pre- and post-surgery Non-surgical
0 = no or unable to determine, 1 = yes
Farrant et al. (2005)
2
1
1
1
5/6
Surgical candidates
Giovagnoli et al. (2009) Giovagnoli et al. (2011)
2 2
1 1
0 1
0 1
3/6 5/6
Non-surgical Non-surgical
Giovagnoli et al. (2013) Jiang et al. (2014)
2 2
1 1
1 1
1 1
5/6 5/6
Non-surgical Non-surgical
Lew et al. (2015)
2
1
1
1
5/6
Non-surgical
Li et al. (2013) Schacher et al. (2006)
2 1
2 1
1 0
1 1
6/6 3/6
Non-surgical Pre- and post-surgery
Shaw et al. (2004) Wang et al. (2015)
1 2
1 2
0 1
1 1
3/6 6/6
Post-surgery Surgical candidates
(1) Were patients described socio-demographically (i.e., age, sex, education/schooling)? (2) Were patients described clinically (i.e., age at onset, duration of epilepsy, seizure frequency, AED use, IQ, psychiatric comorbidity)? (3) Were the inclusion/exclusion criteria specified? (4) Was patient selection specified (e.g., consecutive patients)? (5) What was the patient’s surgical status (i.e., surgical candidates, pre-surgical, post-surgical, nonsurgical)?
anatomic lesions that could negatively impact social cognition, the surgery may render patients seizure and medication free, which could improve cognitive functions that are important for social skills, such as executive skills that are adversely affected by AEDs (Eddy et al. 2011) and seizures (Black et al. 2010). Our meta-regression of moderator variables, however, did not find association between ToM and either seizure frequency or number of AEDs. We note that our review revealed little evidence that the side of epilepsy focus plays a significant role in the magnitude of ToM deficits. It is possible that either of the hemispheres support ToM processes, which may also explain a lack of a drop in ToM following temporal lobectomy. We did not find any published studies to date that have examined impact of frontal lobe surgery on ToM in patients with epilepsy. Group findings can mask changes that could occur in individual patients. The only study that had a longitudinal arm, and found no change from pre- to post- surgery at a group level, also examined changes in ToM in individual patients (Amlerova et al. 2014). A significant improvement and deterioration were found in ToM scores in 9/30 and 8/ 30 patients post-temporal lobectomy, respectively. These individual changes were not related to any clinical or demographic variables, such as age at assessment, age at epilepsy onset, duration of epilepsy, intelligence or seizure frequency. It was proposed that further studies should examine whether changes in ToM performance might be explained by changes in other cognitive abilities that are critical for ToM, such as executive skills.
Our review found that patients with epilepsy also had marked deficits in one of the low level, early maturing social-cognitive skills: facial emotion identification, however, deficits in facial emotion recognition were not related to impairments in ToM (Broicher et al. 2012; Jiang et al. 2014). The lack of association between facial emotion identification and ToM has also been found in other clinical populations, such as adults with schizophrenia (Brune 2005) and traumatic brain injury (TBI sustained between 20 and 61 years old) (Henry et al. 2006). These findings are consistent with the neurocognitive model of social cognition proposed by Frith and Frith (1999), developed on studies of patients with brain lesions, in which emotion recognition and ToM are said to rely on distinct neural networks. Thus, skills are dissociable and may be independently impaired. From a treatment perspective, these findings suggest that interventions targeting lower level abilities (facial emotion recognition) are unlikely to have flow-on effects to higher-order processes, such as ToM. Instead, separate interventions need to be developed for these two types of social-cognitive deficits. We pointed out that several other factors, namely age at seizure onset, duration of epilepsy and age at testing are also likely to impact ToM. Contrary to our expectations, our metaanalysis showed no significant relationship between age of seizure onset and disease duration with ToM. The lack of significant findings may be due to studies lacking statistical power to find significant relations between age of onset/ disease duration and ToM for patients with TLE (Li et al. 2013; Schacher et al. 2006), FLE (Farrant et al. 2005), extra-
Neuropsychol Rev
TLE (Broicher et al. 2012) and IGE (Jiang et al. 2014). Sample sizes in these studies ranged from 14 to 42 participants. In contrast, the single study that found significant associations between age of onset/disease duration and ToM included 66 participants (rs = 0.24–0.33) (Giovagnoli et al. 2013). Using G*Power 3.1 we found that 63 to 77 participants would be required to replicate a similar, medium-sized effect (alpha = 0.05, power = 0.8), which clearly exceeds the number of patients included in studies that failed to find significant relations between age of seizure onset/disease duration and ToM. Two other studies reported significant associations between age of onset and ToM by dichotomising patients into early vs. late onset seizure disorder (Amlerova et al. 2014; Giovagnoli et al. 2011). Amlerova et al. (2014) found that patients with early onset TLE (<6 years old) obtained a significantly lower overall faux-pas score (which incorporates cognitive and affective ToM) than those with late onset TLE (>6 years). Giovagnoli et al. (2011) found that patients with early onset TLE (<12 years) were significantly more impaired than those with late onset TLE (>12 years), but only in their affective ToM score from the faux-pas task. This finding is consistent with development accounts of ToM, which posit that affective ToM skills continue to develop until age 12 (Baron‐Cohen et al. 1997). Nevertheless, a shortcoming of dichotomising patients is that it was unclear whether any patients in the early onset group had seizures beginning prior to 6 years-old (Giovagnoli et al. 2011). These patients may have also had deficits in cognitive ToM. Alternatively, the lack of relation between age of epilepsy onset and ToM may reflect the fact that ToM is a trait (a biomarker) of TLE and FLE, rather than ToM development being disrupted by the onset of epilepsy. While our meta-analysis did not find that the age of onset was associated with ToM, the meta-analysis did find that the age at testing was associated with ToM, with participants tested at a younger age having greater deficits in ToM. Notwithstanding this, only two studies involved children with epilepsy and both examined children with IGE (Jiang et al. 2014; Lew et al. 2015). One study also included a group of children with mixed focal epilepsies (TLE, FLE, extra-TLE/FLE). Because adults with focal epilepsies (TLE, FLE, Extra-TLE/FLE) had very divergent ToM abilities, we excluded this mixed focal epilepsy group from analyses, as we anticipated it may confound results. With regard to children with IGE, findings were mixed. Jiang et al. (2014) found that adolescents with IGE (mean age = 16.5 years) were impaired on the RMET, whereas Lew et al. (2015) found that children with IGE (mean age = 11.6 years) were not impaired on this task. It is possible that these seemingly inconsistent findings are related to between study differences in participants’ age. As the skills measured by the RMET are still emerging at 11 years old, difficulties on this task may not be expressed or adequately assessed until an older age (Dumontheil et al. 2010). Failure to find differences in 11 year old children may reflect the fact that typically
developing controls had not yet acquired the skills assessed by this task (Lew et al. 2015). These findings highlight the importance of selecting age-appropriate tasks in future studies of children with epilepsy. Finally, the fact that we found an association between age at testing and ToM, but not age of seizure onset and ToM, may reflect the approach we used to analyse moderator variables. Meta-regression requires a spread of scores to detect a significant effect, with a larger range of scores increasing the chance of statistically significant findings. The range of scores for age of onset was 7.6 years old to 26.1 years old, whereas the age range for age at testing was 11.6 years to 42.3 years, which is much larger and thus, provides a much greater opportunity to detect significant results. Theory of Mind tasks ToM difficulties were observed across tasks, including tasks designed to assess early developing cognitive perspective taking (e.g., false belief), affective understanding (e.g., RMET), and later-developing cognitive and affective ToM skills (e.g., strange stories, cartoon ToM vignettes, faux-pas task). Deficits appeared to be task independent, in that effect sizes did not differ significantly between tasks. Despite this, we found that effect sizes tended to be larger on tasks requiring executive skills: false belief, faux-pas, strange stories and cartoon ToM. Performance on the RMET, which imposes comparatively few demands on executive skills (working memory and inhibition), was less impaired (Henry et al. 2015). Thus, it is possible that impairments on ToM tasks reflect alterations in cognitive skills other than ToM (Martín-Rodríguez and León-Carrión 2010). For instance, false belief, faux pas and strange stories all rely on verbal ability and executive skills, as individuals must understand and respond rapidly to questions about the mental and emotional states of characters in stories (Astington and Jenkins 1999; Bird et al. 2004; Happé 1994; Mutter et al. 2006; Wellman et al. 2001). Similarly, cartoon ToM tasks require verbal ability and EF, as participants must infer mental states of characters in cartoons and verbally articulate these. Only three of the twelve studies included in our review examined associations between executive skills and ToM (Farrant et al. 2005; Giovagnoli et al. 2011; Jiang et al. 2014). Giovagnoli et al. (2011) found a significant association between a composite score of executive functions (derived from tests of abstraction, working memory and planning abilities) and performance on the faux-pas task in patients with TLE. In contrast, Jiang et al. (2014) found no association between inhibition (as measured by the Stroop Task) and performance on the RMET in patients with IGE. Similarly, Farrant et al. (2005) found no significant association between inhibition (as measured by the Hayling Test) and performance on the RMET, faux pas, strange stories or cartoon ToM among patients with FLE.
Neuropsychol Rev
There are several possible explanations for these inconsistent findings. First, it may be that only specific EFs are associated with ToM in patients with epilepsy. For instance, working memory has been found to be a stronger predictor of ToM than response inhibition among typically developing children (as measured by the false belief task) (Mutter et al. 2006). Thus, working memory, but not inhibition, may relate to ToM in patients with epilepsy. Second, it may be that certain EFs are implicated in some ToM tasks, but not others. For example, Bull et al. (2008) found that inhibition, working memory and attentional flexibility were all implicated in the strange stories task, whereas only inhibition was required for the RMET (Bull et al. 2008). Nevertheless, we found no evidence of an association between inhibition and performance on the RMET or strange stories in patients with epilepsy. This raises the possibility that the relationship between EF and ToM is different in patients with epilepsy compared to healthy controls. This is consistent with findings from patients with other neurological/psychiatric illnesses (e.g., schizophrenia) in that impairments in EF are independent of impairments in ToM (Pickup 2008). At present, it is not clear how specific executive skills contribute to ToM performance in patients with different types of epilepsy (TLE, FLE and IGE) or on different ToM tasks. This is a limitation of the literature of patients with other neurological disorders (e.g., acquired brain injury) in that no consensus has been reached on the dissociation between ToM and executive skills or the role of specific executive skills on ToM tasks (Martín-Rodríguez and LeónCarrión 2010). The fact that patients with epilepsy were more impaired on the false belief than the faux pas task is suggests that mastery over ToM tasks may follow a different trajectory in patients with epilepsy compared to healthy controls. From an assessment and careening perspective, researchers must take special care in selecting instruments to measure ToM, especially when working with children, as tasks that are suitable for healthy controls may be beyond the developmental capacity of epilepsy groups. Another cognitive skill that has been implicated in ToM among healthy individuals are language skills (Astington et al. 2002; Milligan et al. 2007). A previous meta-analysis that included studies undertaken by healthy control participants found a strong association between various aspects of verbal/language skills (e.g., general vocabulary, semantics, receptive vocabulary, syntax) and ToM (as measured by the false belief task; Milligan et al. 2007). Three studies (Giovagnoli et al. 2009, 2011, 2013) included in our review measured word finding (the Word Fluency and the Boston Naming tests). None of these studies, however, examined the relationship between word findings and ToM, which should be addressed in future studies. Five studies measured verbal intelligence and examined how it related to ToM in patients with epilepsy. None of these
studies found a significant association between verbal intelligence and ToM, which is consistent with findings for other patients groups, such as those with schizophrenia, as well as healthy controls (Mo et al. 2008). One study worth mentioning is Farrant et al.’s (2005) study, which used three different tasks to test ToM in patients with FLE. While all three tasks are all designed to assess cognitive and affective ToM, impairments were not found on all three tasks. Instead, patients with FLE scored significantly below healthy controls on faux pas and cartoon tasks, but not on the strange stories task. It is possible that subtle differences in the social content and/or the types of language skills required to complete these tasks explain discrepant results. The strange stories task involves understanding of figurative (non-literal) language, such as metaphors, irony, sarcasm, white lies, metaphorical expressions and indirect requests, which may be intentionally used to adhere to certain social rules (Kaland et al. 2005). For instance, a person receiving a birthday present might say: BThankyou, it’s lovely^ either because they really do find it lovely or because they intend to be polite to spare the other person’s feelings. In contrast, the faux pas task requires an understanding of literal language and utterances that unintentionally violate social rules. For instance, a person who is at a friend’s house for dinner may proclaim that they Bhate apple pie^, not knowing that the host has baked apple pie for dessert. The cartoon ToM task is similar to faux pas in that humour/jokes depends on unintended social actions because of something someone mistakenly thinks or does not know. Differences in the types of language skills required to complete tasks (literal vs. non-literal language) and/or degree of intentionality/motivation behind the speech may underpin differential performance on these tasks. In addition, the brain networks that underpin performance on these ToM tasks may not be entirely overlapping. For example, Gallagher and colleagues (2000) used functional Magnetic Resonance Imaging (fMRI) to compare the brain regions activated during strange stories and cartoon ToM tasks in healthy participants. While both tasks activated the medial prefrontal cortex, the strange stories task activated a broader network, which involved the anterior and inferior medial prefrontal lobes. As somewhat larger prefrontal activation was implicated in the strange stories task, performance of this task may be less sensitive to disruption as there may be more scope for reorganisation of function. Studies are needed to determine whether location of pathology/seizure focus in the frontal lobes differentially impacts performance on different ToM tasks. Finally, we turn out attention to the RMET, which merits special consideration, as the effect size was somewhat, albeit not significantly, smaller than the effect sizes noticed on four other tasks. Aside from the RMET relying less on executive skills, RMET differs from other ToM tasks in several ways. First, while on story and cartoon tasks, participants must spontaneously generate responses with minimal prompts on the
Neuropsychol Rev
RMET participants are provided with four semantic categories and asked to pick the one that is the best match for the emotions expressed by the eyes. Thus, on the RMET participant has a 1 in 4 chance of responding correctly by guessing, which is not possible on the story and cartoon tasks. Second, although the RMET is considered an ‘advanced’ test of ToM, the authors who developed the task acknowledge that the RMET requires less complex ToM reasoning than faux-pas, strange stories and cartoon ToM. The RMET requires only the first stage of ToM attribution (i.e., naming the relevant mental state), but not the second stage of ToM attribution (i.e., inferring the content of that mental state; Baron‐Cohen et al. 2001) required by advanced tasks. Third, the RMET is the only task that exclusively assesses affective ToM. It is also possible that patients with epilepsy are less impaired in affective ToM than in cognitive ToM. In summary, while all tasks included in this review detect deficits in ToM in patients with epilepsy, they differ in the scope of assessment. The RMET assesses only affective ToM, while the faux pas, strange stories and cartoon ToM tasks assess cognitive and affective ToM. In addition the scenarios depicted in the faux pas, strange stories and cartoon ToM tasks mimic everyday social interactions and may highlight the types of social situations that patients might have difficulty understanding when interacting with family and friends (e.g., understanding social faux-pas, jokes, metaphors, irony, sarcasm, white lies and/or other implicit meanings in speech). Interestingly, on close inspection of the studies employing the faux-pas task, we noticed that patients were unimpaired in their ability to detect faux-pas (Farrant et al. 2005; Giovagnoli et al. 2009), but had marked difficulties explaining the thoughts and feelings of characters following faux-pas. This suggests that while patients with epilepsy are able to detect social actions that are inappropriate or rude, their understanding of the intentions, thoughts, beliefs and emotions of others (i.e., ToM) is impaired. Limitations of the Review As research into ToM in patients with epilepsy is an emerging field, our review included a modest number of studies (n = 12), which may have limited our ability to detect significant effects in meta-regression. Moreover, the approach we utilised to analyse data in our meta-regressions may have been limited, as we used means of variables reported in papers (e.g., mean age of seizure onset, mean disease duration, mean AEDs). Utilising group means reduced variability of scores between studies, which is required to detect significant effects in regression. Moreover, it did not take into account individual variability of participants within studies, which may have masked the effects of very early onset seizures on ToM. In addition, overlap in control groups precluded meta-regression being conducted for certain moderator variables (e.g., side of
seizure focus). Instead, these variables were examined in the systematic review included in this paper. Finally, we limited our review to papers that utilised one of five empirically validated and psychometrically sound behavioural measures of ToM (Henry et al. 2015). While such an approach is likely to have increase the reliability and validity of our study, it might have excluded some studies that used novel tasks to examine ToM. One final point worth noting is that the two IGE studies examined children/adolescents. As earlier age at testing was associated with more severe ToM impairments across studies, the effect size obtained for IGE patients (children/adolescents) might not hold in the adult sample, which may have smaller ToM deficits, which needs to be investigated in future studies. Overall, quality of papers included in this review, as determined by the Downs and Black (1998) checklist, was good. Nevertheless, a detailed inspection of ratings made on the Downs and Black (1998) checklist identified specific areas of weakness across studies, specifically: incomplete reporting of the source and timing of participant recruitment. This raises the possibility of selection bias, as subjects included in studies may not be representative of the population of patients with epilepsy. Another way to determine whether patients are representative of an epilepsy population is to examine whether they have been considered for epilepsy surgery and/or had already undergone surgery. These patients are likely to have severe epilepsy and experience seizures that are difficult to control with AEDs. Inspection of papers revealed that two studies included patients who were considered as candidates for epilepsy surgery (Farrant et al. 2005; Wang et al. 2015) and another three papers examined patients pre and/or postsurgery (Amlerova et al. 2014; Schacher et al. 2006; Shaw et al. 2004). The majority of studies (7/12) recruited consecutive patients through specialised epilepsy monitoring units. No study recruited patients with epilepsy from the community/general population. Thus, the findings are likely to be representative for patients who require specialised epilepsy management, most likely due to epilepsy being complex and/or difficult to control with medication. This is important to keep in mind when interpreting the findings.
Conclusions To our knowledge, this is the first comprehensive metaanalyses examining ToM in patients with epilepsy. The meta-analysis revealed large deficits in ToM in adult patients with frontal and temporal epilepsy (TLE, FLE), as well as in children with generalised seizures (IGE), irrespective of task type. In patients with TLE, a younger age at testing was associated with significantly greater deficits in ToM, suggesting that children with TLE may have a particularly high risk of deficits in ToM. To date, no study has examined ToM performance in children with TLE or in children with FLE as
Neuropsychol Rev
independent groups. It is difficult to draw firm conclusions about how epilepsy variables (age of onset, duration of epilepsy, seizure frequency, AEDs, side and site of seizure focus) moderate ToM due to the small number of studies included. Finally, further research is needed to determine how language skills and executive skills (i.e., working memory, inhibition, attentional flexibility) relate to ToM in both children and adults with epilepsy. Compliance with Ethical Standards Conflict of Interest The authors declare that they have no conflict of interest.
Appendix 1 Details of Quality Rating Items The following 17 items from the Downs and Black checklist (1998) were employed to assess quality of included studies: 1. Is the hypothesis/aim/objective of the study clearly described? 2. Are the main outcomes to be measured clearly described in the Introduction or Methods section? 3. Are the characteristics of the patients included in the study described clearly? 5. Are the distributions of principal confounders in each group of subjects to be compared described clearly? 6. Are the main findings of the study described clearly? 7. Does the study provide estimates of the random variability in the data for the main outcomes? 9. Have the characteristics of patients lost to follow-up been described? 10. Have actual probability values been reported (for example, 0.035 rather than < 0.05) for the main outcomes except where the probability value is less than 0.001? 11. Were the subjects asked to participate in the study representative of the entire population from which they were recruited? 16. If any of the results of the study were based on ‘data dredging’, was this made clear? 18. Were the statistical tests used to assess the main outcomes appropriate? 20. Were the main outcome measures used accurate (valid and reliable)? 21. Were the patients in different groups recruited from the same population? 22. Were study subjects recruited over the same period of time? 25. Was there adequate adjustment for confounding in the analyses from which the main findings were drawn? 26. Were losses of patients to follow-up taken into account? 27: Did the study have sufficient power to detect a clinically important effect where the probability value for a difference being due to chance is less than 5 %? Details of Theory of Mind Tasks Five types of ToM tasks were included in this review. These tasks were selected based on a recent review by Henry et al. (2015), which identified the empirically validated and psychometrically sound behavioural measures of ToM.
False Belief Tasks Participants are read aloud short stories about a character who has a mistaken belief (i.e., a belief that is inconsistent with their own beliefs and/or or out of line with reality). The participant must correctly infer the mental state of the character by using their cognitive ToM skills. In the simplest version of these tasks, participants must make inferences about the mistaken beliefs of a character with regard to real events (first-order false belief). In more complicated versions of the task, participants must make meta-cognitive inferences: attributing the false belief of one person based on the thoughts of another (second-order false belief). Firstorder and second-order false belief are commonly grouped together to measure false belief (i.e., ToM) performance. Numerous versions of this task have been developed, which are mostly based on the initial task done by (Wimmer and Perner 1983). This task assesses early cognitive ToM (Henry et al. 2015). Reading the Mind in the Eyes Test (RMET) Participants are presented with a series of photographs of the eye regions of faces and are asked to select the emotion label that best depicts how a character is feeling. The Reading in the Mind in the Eyes Task (RMET), designed by Baron-Cohen, is the most commonly used of these measures and contains 36 photographs, presented sequentially with four emotion labels for each (Baron‐Cohen et al. 2001). Adaptations of this task have been created for participants from different cultural groups (Jiang et al. 2014). The RMET assesses early affective ToM (Henry et al. 2015). Faux-Pas Tasks Participants are read aloud stories that contain a social faux-pas (i.e., scenario in which a character says or does something unintended that may impact upon another characters feelings). Following each story, participants are asked a series of questions that assess their ability to detect the faux-pas (i.e., Did someone say something they shouldn’t have said or something awkward? Who said something they shouldn’t have said or something awkward?); their cognitive ToM (Why do you think he or she said it?); and their affective ToM (How do you think X felt when Y said that?). Each participant obtains a total ToM score by summing responses across questions on the faux-pas stories. They also obtain a score representing their ability to detect the faux pas (Faux-pas hits; calculated by summing responses to the question 1 on each story). Participants are also read
Neuropsychol Rev
control stories that do not contain faux-pas’ and are asked the first comprehension question to assess their ability to rule out non-existent faux-pas and obtain a score representing their ability to rule out non-existant faux-pas (Faux-pas correct rejections; calculated by summing responses to the first question asked on each of the non faux-pas stories). This task assessed advanced ToM (cognitive and affective components) (Henry et al. 2015). Strange Stories Participants are read aloud stories involving an interaction between two people, in which characters say or do something that they do not literally mean. The stories contain jokes, metaphors, double bluff, mistakes, persuasion, white lies and/or deception. Participants are asked a series of questions assessing their ability to comprehend the implicit meaning of the story (What it true, what X said?) and then to make inferences about characters’ thoughts, beliefs, intentions and emotional states (Why did X say it?). The original version of this task
Table 7
contains 24 short stories, however shortened versions have been used in recent studies (Happé 1994). This task assesses advanced ToM (cognitive and affective components) (Henry et al. 2015). Cartoon ToM Tasks Participants are shown a series of single frame cartoons in which humour depends on what a character mistakenly thought or did not know. Participants are typically asked two questions. The first question is open-ended (implicit form), asking participants why the picture is funny; the second question is presented in a more explicit manner (explicit form), asking what the motives of the character in each picture are. The total score for each type of question ranges from 0 to 20. Some studies have only used the implicit form of the question, however, to receive a correct response, participants must accurately infer what is funny in the story with reference to the thoughts, beliefs, intentions and feelings of characters. This task assesses advanced ToM (cognitive and affective component).
Studies excluded from meta-analysis
Study
Group(s)
Task
Main Findings
Reasons for exclusion from meta-analysis
Fournier et al. (2008)
2 patients (J.H. and S.M.) with generalized epilepsy, tested 30 years after hemispherdectomy for intractable epilepsy
RMET
Patient (J.H.) with seizures in the left hemisphere performed comparably to controls on all measures. Patient (S.M.) with seizures in the right hemisphere performed more poorly than controls on RMET and ability to understand another person’s intentions and feelings.
Data reported not sufficient to calculate an effect
Task that required children to use eye gaze direction to infer thoughts and feelings. Separated into first-order and second-order ToM (cognitive and affective) RMET
Children with Rolandic Epilepsy were impaired in second-order affective ToM, but not second-order cognitive ToM. They were unimpaired in first-order cognitive and affective ToM Average performance on MET (patient score = 25/36 or 69.5 % correct; no control group)
Behavioural task other than one of five specified in selection criteria
Faux-Pas
Deficits on Faux-pas task for both patients tested on the Faux-pas task (patient scores = 4/10 and 6/10; controls = 10/10)
Genizi et al. (2012)
15 children with Rolandic Epilepsy
Hynes and Mar (2009)
1 patient (F.S.) with drug resistant TLE. Surgery at 20 and 24 months-old removed the right anterior temporal lobe 6 patients who all had Corpus Collosum resection for drug resistant epilepsy. Specific type of epilepsy or site of seizures not reported
Miller et al. (2010)
Social Inference Test – participants’ answered questions about another person’s intentions, thoughts and feelings.
Moral Reasoning Task
Case study with no control group
Data reported not sufficient to calculate an effect. Behavioural task other than one of five specified in selection criteria (moral reasoning task assessed participants’ abilities to detect Bforbidden^ or Bpermissible^ actions, but did not require inferences
Neuropsychol Rev Table 7 (continued) Study
Group(s)
Task
Main Findings
Reasons for exclusion from meta-analysis about mental or emotional states)
Richard-Mornas et al. (2014)
Shaw et al. (2007)
Stone et al. (2003)
1 patient with TLE onset at 2 years old. Patient had right mesial temporal lobe lobectomy including amygdala and hippocampal region for drug resistant epilepsy at 24 years old. 19 patients with TLE. All patients had en bloc Temporal Lobe Resection and were tested 3 months before and after surgery.
1 patient (D.R.) with drug resistant epilepsy. Surgical resection of the left and right amygdala was conducted.
False-Belief task RMET
Faux-pas Strange Stories
Faux-pas RMET
References Amlerova, J., Cavanna, A., Bradac, O., Javurkova, A., Raudenska, J., & Marusic, P. (2014). Emotion recognition and social cognition in temporal lobe epilepsy and the effect of epilepsy surgery. Epilepsy & Behavior, 36, 86–89. Amodio, D., & Frith, C. (2006). Meeting of minds: the medial frontal cortex and social cognition. Nature Reviews Neuroscience, 7(4), 268–277. Anderson, V., Brown, S., Newitt, H., & Hoile, H. (2011). Long-term outcome from childhood traumatic brain injury: intellectual ability, personality, and quality of life. Neuropsychology, 25(2), 176. Anderson, V., Catroppa, C., Morse, S., Haritou, F., & Rosenfeld, J. V. (2009). Intellectual outcome from preschool traumatic brain injury: a 5-year prospective, longitudinal study. Pediatrics, 124(6), e1064– e1071. Anderson, V., & Moore, C. (1995). Age at injury as a predictor of outcome following pediatric head injury: a longitudinal perspective. Child Neuropsychology, 1(3), 187–202. Astington, J. W., & Jenkins, J. M. (1995). Theory of mind development and social understanding. Cognition & Emotion, 9(2–3), 151–165. Astington, J. W., & Jenkins, J. M. (1999). A longitudinal study of the relation between language and theory of mind development. Developmental Psychology, 35, 1311–1320. Astington, J. W., Pelletier, J., & Homer, B. (2002). Theory of mind and epistemological development: the relation between children’s second-order false-belief understanding and their ability to reason about evidence. New Ideas in Psychology, 20(2), 131–144. Baron-Cohen, S., O’Riordan, M., Jones, R., Stone, V., & Plaisted, K. (1999). A new test of social sensitivity: detection of faux pas in normal children and children with Asperger syndrome. Journal of Autism and Developmental Disorders, 29, 407–418.
Performance on False Belief task comparable to controls Patient performed better than controls on RMET (patient score = 26/36 or 72 % correct, controls = 21/36 or 58 %)
Case study
No differences between patients and controls either pre-operatively to post-operatively. No significant differences in change scores between patients and controls. Significant deficits on faux-pas task Poorer performance than controls on MET (patient score = 17/25 or 68 % correct, control score = 19.8/25 or 79 %)
Data reported not sufficient to calculate an effect
Case study Note. Two patients were tested in this study, however only one participant had epilepsy, the other had amygdala damage due to herpes encephalitis.
Baron‐Cohen, S., Jolliffe, T., Mortimore, C., & Robertson, M. (1997). Another advanced test of theory of mind: evidence from very high functioning adults with autism or Asperger syndrome. Journal of Child Psychology and Psychiatry, 38(7), 813–822. Baron‐Cohen, S., Wheelwright, S., Hill, J., Raste, Y., & Plumb, I. (2001). The BReading the Mind in the Eyes^ test revised version: a study with normal adults, and adults with Asperger syndrome or high‐ functioning autism. Journal of Child Psychology and Psychiatry, 42(2), 241–251. Bauminger-Zviely, N. (2013). False-belief task Encyclopedia of autism spectrum disorders (pp. 1249–1249): Springer. Black, L., Schefft, B., Howe, S. R., Szaflarski, J. P., Yeh, H. S., & Privitera, M. D. (2010). The effect of seizures on working memory and executive functioning performance. Epilepsy & Behavior, 17(3), 412–419. Bird, C. M., Castelli, F., Malik, O., Frith, U., & Husain, M. (2004). The impact of extensive medial frontal lobe damage on ‘Theory of Mind’and cognition. Brain, 127(4), 914–928. Borenstein, M., Hedges, L., Higgins, J., & Rothstein, H. (2005). Comprehensive meta-analysis version 2 (p. 104). Englewood: Biostat. Borenstein, M., Hedges, L., Higgins, J., & Rothstein, H. (2009). Introduction to meta-analysis. Chichester: Wiley. Broicher, S. D., Kuchukhidze, G., Grunwald, T., Kramer, G., Kurthen, M., & Jokeit, H. (2012). BTell me how do I feel^–Emotion recognition and theory of mind in symptomatic mesial temporal lobe epilepsy. Neuropsychologia, 50(1), 118–128. Brune, M. (2005). Emotion recognition, ‘theory of mind’,and social behavior in schizophrenia. Psychiatry Research, 133(2), 135–147. Bull, R., Phillips, L. H., & Conway, C. A. (2008). The role of control functions in mentalizing: dual-task studies of theory of mind and executive function. Cognition, 107(2), 663–672.
Neuropsychol Rev Caplan, R., Sagun, J., Siddarth, P., Gurbani, S., Koh, S., Gowrinathan, R., & Sankar, R. (2005). Social competence in pediatric epilepsy: insights into underlying mechanisms. Epilepsy & Behavior, 6(2), 218– 228. Carlson, S. M., & Moses, L. J. (2001). Individual differences in inhibitory control and children’s theory of mind. Child Development, 72(4), 1032–1053. Carlson, S. M., Moses, L. J., & Breton, C. (2002). How specific is the relation between executive function and theory of mind? Contributions of inhibitory control and working memory. Infant and Child Development, 11(2), 73–92. Cavallini, E., Lecce, S., Bottiroli, S., Palladino, P., & Pagnin, A. (2013). Beyond false belief: theory of mind in young, young-old, and oldold adults. The International Journal of Aging and Human Development, 76(3), 181–198. Cohen, J. D. (1988). Statistical power analysis for the behavioral sciences. Hillsdale: Lawrence Earlbaum Associates. Culhane-Shelburne, K., Chapieski, L., Hiscock, M., & Glaze, D. (2002). Executive functions in children with frontal and temporal lobe epilepsy. Journal of the International Neuropsychological Society, 8(05), 623–632. Downs, S. H., & Black, N. (1998). The feasibility of creating a checklist for the assessment of the methodological quality both of randomised and non-randomised studies of health care interventions. Journal of Epidemiology and Community Health, 52(6), 377–384. Dumontheil, I., Apperly, I. A., & Blakemore, S. J. (2010). Online usage of theory of mind continues to develop in late adolescence. Developmental Science, 13(2), 331–338. Eddy, C. M., Rickards, H. E., & Cavanna, A. E. (2011). The cognitive impact of antiepileptic drugs. Therapeutic Advances in Neurological Disorders, 4(6), 385–407. Egger, M., Smith, G. D., Schneider, M., & Minder, C. (1997). Bias in meta-analysis detected by a simple, graphical test. BMJ, 315(7109), 629–634. Engel, J. (2001). A proposed diagnostic scheme for people with epileptic seizures and with epilepsy: report of the ILAE Task Force on Classification and Terminology. Epilepsia, 42(6), 796–803. Farrant, A., Morris, R. G., Russell, T., Elwes, R., Akanuma, N., Alarcón, G., & Koutroumanidis, M. (2005). Social cognition in frontal lobe epilepsy. Epilepsy & Behavior, 7(3), 506–516. Fiske, S., & Taylor, S. (1991). Social cognition (2nd ed., pp. 16–15). NY: McGraw-Hill. Fiske, S. T., & Taylor, S. E. (2013). Social cognition: From brains to culture: Sage. Fournier, N., Calverley, K., Wagner, J., Poock, J., & Crossley, M. (2008). Impaired social cognition 30 years after hemispherectomy for intractable epilepsy: the importance of the right hemisphere in complex social functioning. Epilepsy & Behavior, 12(3), 460–471. Frith, C. D. (2007). The social brain? Philosophical Transactions of the Royal Society, B: Biological Sciences, 362(1480), 671–678. Frith, C. D., & Frith, U. (1999). Interacting minds-a biological basis. Science, 286(5445), 1692–1695. Gallagher, H. L., & Frith, C. D. (2003). Functional imaging of ‘theory of mind’. Trends in Cognitive Sciences, 7(2), 77–83. Gallagher, H. L., Happé, F., Brunswick, N., Fletcher, P. C., Frith, U., & Frith, C. D. (2000). Reading the mind in cartoons and stories: an fMRI study of ‘theory of mind’ in verbal and nonverbal tasks. Neuropsychologia, 38(1), 11–21. Gelžinienė, G., Jurkevičienė, G., Marmienė, V., Adomaitienė, V., & Endzinienė, M. (2010). Executive functions in adolescents with idiopathic generalized epilepsy. Medicina (Kaunas, Lithuania), 47(6), 313–319. Genizi, J., Shamay-Tsoory, S. G., Shahar, E., Yaniv, S., & Aharon-Perez, J. (2012). Impaired social behavior in children with benign childhood epilepsy with centrotemporal spikes. Journal of Child Neurology, 27(2), 156–161.
Giovagnoli, A. R. (2014). The importance of theory of mind in epilepsy. Epilepsy & Behavior, 39, 145–153. Giovagnoli, A. R., Canafoglia, L., Reati, F., Raviglione, F., & Franceschetti, S. (2009). The neuropsychological pattern of Unverricht–Lundborg disease. Epilepsy Research, 84(2), 217–223. Giovagnoli, A. R., Franceschetti, S., Reati, F., Parente, A., Maccagnano, C., Villani, F., & Spreafico, R. (2011). Theory of mind in frontal and temporal lobe epilepsy: cognitive and neural aspects. Epilepsia, 52(11), 1995–2002. Giovagnoli, A. R., Parente, A., Villani, F., Franceschetti, S., & Spreafico, R. (2013). Theory of mind and epilepsy: what clinical implications? Epilepsia, 54(9), 1639–1646. Gomez-Ibañez, A., Urrestarazu, E., & Viteri, C. (2014). Recognition of facial emotions and identity in patients with mesial temporal lobe and idiopathic generalized epilepsy: an eye-tracking study. Seizure, 23(10), 892–898. Happé, F. (1994). An advanced test of theory of mind: understanding of story characters’ thoughts and feelings by able autistic, mentally handicapped, and normal children and adults. Journal of Autism and Developmental Disorders, 24(2), 129–154. Henry, J., Cowan, D., Lee, T., & Sachdev, P. (2015). Recent trends in testing social cognition. Current Opinion in Psychiatry, 28(2), 133– 140. Henry, J. D., Phillips, L. H., Crawford, J. R., Ietswaart, M., & Summers, F. (2006). Theory of mind following traumatic brain injury: the role of emotion recognition and executive dysfunction. Neuropsychologia, 44(10), 1623–1628. Higgins, J., & Green, S. (2008). Cochrane handbook for systematic reviews of interventions. Version 5.0.1.: The Cochrane Collaboration. Hrabok, M., Dykeman, J., Sherman, E. M. S., & Wiebe, S. (2013). An evidence-based checklist to assess neuropsychological outcomes of epilepsy surgery: How good is the evidence? Epilepsy & Behavior, 29(3), 443–448. Hughes, C. (1998). Executive function in preschoolers: links with theory of mind and verbal ability. British Journal of Developmental Psychology, 16(2), 233–253. Hynes, C. A., & Mar, R. A. (2009). A case study of long-term cognitive and social functioning following a right temporal lobectomy in infancy. Neurocase, 15(1), 37–46. Jiang, Y., Hu, Y., Wang, Y., Zhou, N., Zhu, L., & Wang, K. (2014). Empathy and emotion recognition in patients with idiopathic generalized epilepsy. Epilepsy & Behavior, 37, 139–144. Kalbe, E., Schlegel, M., Sack, A. T., Nowak, D. A., Dafotakis, M., Bangard, C., & Kessler, J. (2010). Dissociating cognitive from affective theory of mind: a TMS study. Cortex, 46(6), 769–780. Kaland, N., Møller-Nielsen, A., Smith, L., Mortensen, E. L., Callesen, K., & Gottlieb, D. (2005). The strange stories test. European Child & Adolescent Psychiatry, 14(2), 73–82. Kim, Y. T., Kwon, D. H., & Chang, Y. (2011). Impairments of facial emotion recognition and theory of mind in methamphetamine abusers. Psychiatry Research, 186(1), 80–84. Kirsch, H. E. (2006). Social cognition and epilepsy surgery. Epilepsy & Behavior, 8(1), 71–80. Kwan, P., & Brodie, M. J. (2001). Neuropsychological effects of epilepsy and antiepileptic drugs. The Lancet, 357(9251), 216–222. Ladegaard, N., Larsen, E. R., Videbech, P., & Lysaker, P. H. (2014). Higher-order social cognition in first-episode major depression. Psychiatry Research, 216(1), 37–43. Lew, A. R., Lewis, C., Lunn, J., Tomlin, P., Basu, H., Roach, J., & Martland, T. (2015). Social cognition in children with epilepsy in mainstream education. Developmental Medicine & Child Neurology, 57(1), 53–59. Li, Y., Chiu, M., Yeh, Z., Liou, H., Cheng, T., & Hua, M. (2013). Theory of mind in patients with temporal lobe epilepsy. Journal of the International Neuropsychological Society, 19(05), 594–600.
Neuropsychol Rev Martín-Rodríguez, J. F., & León-Carrión, J. (2010). Theory of mind deficits in patients with acquired brain injury: a quantitative review. Neuropsychologia, 48(5), 1181–1191. Miller, M. B., Sinnott-Armstrong, W., Young, L., King, D., Paggi, A., Fabri, M., & Gazzaniga, M. S. (2010). Abnormal moral reasoning in complete and partial callosotomy patients. Neuropsychologia, 48(7), 2215–2220. Milligan, K., Astington, J. W., & Dack, L. A. (2007). Language and theory of mind: meta‐analysis of the relation between language ability and false‐belief understanding. Child Development, 78(2), 622– 646. Mo, S., Su, Y., Chan, R. C., & Liu, J. (2008). Comprehension of metaphor and irony in schizophrenia during remission: the role of theory of mind and IQ. Psychiatry Research, 157(1), 21–29. Moher, D., Liberati, A., Tetzlaff, J., & Altman, D. G. (2009). Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. Annals of Internal Medicine, 151(4), 264–269. Mutter, B., Alcorn, M. B., & Welsh, M. (2006). Theory of mind and executive function: working-memory capacity and inhibitory control as predictors of false-belief task performance. Perceptual and Motor Skills, 102(3), 819–835. Nolan, M. A., Redoblado, M. A., Lah, S., Sabaz, M., Lawson, J. A., Cunningham, A. M., & Bye, A. M. (2003). Intelligence in childhood epilepsy syndromes. Epilepsy Research, 53(1), 139–150. Olson, I. R., McCoy, D., Klobusicky, E., & Ross, L. A. (2012). Social cognition and the anterior temporal lobes: a review and theoretical framework. Social cognitive and affective neuroscience, 8(2), 123– 133. Olson, I. R., Plotzker, A., & Ezzyat, Y. (2007). The enigmatic temporal pole: a review of findings on social and emotional processing. Brain, 130(7), 1718–1731. Onishi, K. H., & Baillargeon, R. (2005). Do 15-month-old infants understand false beliefs? Science, 308(5719), 255–258. Park, S., & Kwon, S. (2008). Cognitive effects of antiepileptic drugs. Journal of Clinical Neurology, 4(3), 99–106. Perner, J., & Lang, B. (1999). Development of theory of mind and executive control. Trends in Cognitive Sciences, 3(9), 337–344. Pickup, G. J. (2008). Relationship between theory of mind and executive function in schizophrenia: a systematic review. Psychopathology, 41(4), 206–213. Pons, F., Harris, P. L., & de Rosnay, M. (2004). Emotion comprehension between 3 and 11 years: developmental periods and hierarchical organization. European Journal of Developmental Psychology, 1(2), 127–152. Rantanen, K., Eriksson, K., & Nieminen, P. (2012). Social competence in children with epilepsy—a review. Epilepsy & Behavior, 24(3), 295– 303. Rajkumar, A. P., Yovan, S., Raveendran, A. L., & Russell, P. S. (2008). Can only intelligent children do mind reading: the relationship
between intelligence and theory of mind in 8 to 11 years old. Behavioral and Brain Functions, 4(51), 1–7. Richard-Mornas, A., Mazzietti, A., Koenig, O., Borg, C., Convers, P., & Thomas-Antérion, C. (2014). Emergence of hyper empathy after right amygdalohippocampectomy. Neurocase, 20(6), 666–670. Saxe, R., & Kanwisher, N. (2003). People thinking about thinking people: the role of the temporo-parietal junction in Btheory of mind^. NeuroImage, 19(4), 1835–1842. Saxe, R., & Wexler, A. (2005). Making sense of another mind: the role of the right temporo-parietal junction. Neuropsychologia, 43(10), 1391–1399. Schacher, M., Winkler, R., Grunwald, T., Kraemer, G., Kurthen, M., Reed, V., & Jokeit, H. (2006). Mesial temporal lobe epilepsy impairs advanced social cognition. Epilepsia, 47(12), 2141–2146. Shaw, P., Lawrence, E., Bramham, J., Brierley, B., Radbourne, C., & David, A. (2007). A prospective study of the effects of anterior temporal lobectomy on emotion recognition and theory of mind. Neuropsychologia, 45(12), 2783–2790. Shaw, P., Lawrence, E., Radbourne, C., Bramham, J., Polkey, C., & David, A. (2004). The impact of early and late damage to the human amygdala on ‘theory of mind’ reasoning. Brain, 127(7), 1535–1548. Sodian, B., & Kristen, S. (2010). Theory of mind Towards a theory of thinking (pp. 189–201): Springer. Song, H. J., Onishi, K. H., Baillargeon, R., & Fisher, C. (2008). Can an agent’s false belief be corrected by an appropriate communication? Psychological reasoning in 18-month-old infants. Cognition, 109(3), 295–315. Sosa, J. T. R., Ojeda, M. A., & del Rosario, L. R. (2011). Theory of mind, facial recognition and emotional processing in schizophrenia. Revista de Psiquiatría y Salud Mental (English Edition), 4(1), 28– 37. Stone, V., Baron-Cohen, S., Calder, A., Keane, J., & Young, A. (2003). Acquired theory of mind impairments in individuals with bilateral amygdala lesions. Neuropsychologia, 41(2), 209–220. Sullivan, J. E., & Dlugos, D. J. (2004). Idiopathic generalized epilepsy. Current Treatment Options in Neurology, 6(3), 231–242. Suurmeijer, T. P., Reuvekamp, M. F., & Aldenkamp, B. P. (2001). Social functioning, psychological functioning, and quality of life in epilepsy. Epilepsia, 42(9), 1160–1168. Wang, W. H., Shih, Y. H., Yu, H. Y., Yen, D. J., Lin, Y. Y., Kwan, S. Y., Hua, M. S. (2015). Theory of mind and social functioning in patients with temporal lobe epilepsy. Epilepsia, 56(7), 1117–1123. Wellman, H. M., Cross, D., & Watson, J. (2001). Meta-analysis of theoryof-mind development: The truth about false belief. Child development, 72(3), 655–684. Wimmer, H., & Perner, J. (1983). Beliefs about beliefs: representation and constraining function of wrong beliefs in young children’s understanding of deception. Cognition, 13(1), 103–128.