Executive Function Mechanisms of Theory of Mind

You might also like

Download as pdf or txt
Download as pdf or txt
You are on page 1of 12

J Autism Dev Disord (2011) 41:667–678

DOI 10.1007/s10803-010-1087-7

ORIGINAL PAPER

Executive Function Mechanisms of Theory of Mind


Fayeza S. Ahmed • L. Stephen Miller

Published online: 1 September 2010


Ó Springer Science+Business Media, LLC 2010

Abstract This study examined the relationship between determining the absolute presence of ToM does not explain
Executive Function (EF) and Theory of Mind (ToM) using its underlying mechanisms, examining its relationship with
the Delis-Kaplan Executive Function System (D-KEFS) EF may shed light on those processes and lead to a better
and three tests of ToM (Reading the Mind in the Eyes test understanding of it. Previous research has indicated that
(RMET), Strange Stories test, and Faux Pas test). Separate there may be multiple domains of ToM (Ahmed et al. 2007,
regression analyses were conducted, and EF predictors February; Brent et al. 2004), and thus there may be dif-
varied by ToM test. No EF domains accounted for signifi- fering EF mechanisms that drive each aspect of ToM.
cant variance in RMET scores; only estimated IQ scores
were significant predictors of RMET performance. Verbal
fluency and deductive reasoning were significant predictors Theory of Mind
of performance on the Strange Stories test, while verbal
fluency, problem solving, and gender accounted for a sig- The concept of ToM originated from Premack and
nificant variance in the Faux Pas test. Results suggest that Woodruff’s research on the ability of chimpanzees to take
the ToM tests each utilized differing cognitive mechanisms. the perspective of a person (Premack and Woodruff 1978).
Traditionally, in humans, ToM has been measured in false
Keywords Executive Function  Theory of Mind  belief paradigms. For example, the common Sally-and-
Neuropsychology Anne test (Baron-Cohen et al. 1985; Wimmer and Perner
1983) involves a scenario in which one character hides an
object while another character leaves the room. Partici-
The purpose of this study was to determine aspects of pants are then asked where they think the character who
Executive Function (EF) related to the construct of Theory has left the room will look for the object. Perspective-
of Mind (ToM). EF consists of higher-order cognitive taking capabilities are required in order for participants to
processes (Carlson et al. 2004) involved in goal-oriented recognize that the character who has left the room does not
behavior, such as planning and sequencing (Royall et al. know that the object was moved and therefore will look for
2002). ToM refers to the ability to understand the views the object in the original location (Brüne and Brüne-Cohrs
and beliefs of another person (Baron-Cohen 1988). This 2005). In addition to false belief paradigms, other measures
broad definition of ToM (Baron-Cohen 1988) illustrates the of ToM have included tasks assessing emotion recognition
complexity and abstractness of perspective-taking, and (Baron-Cohen et al. 2001), understanding of cartoon dia-
therefore it can benefit from further operationalism. While grams (Sarfati et al. 1997), and the ability to explain the
reason a character in a story behaved in a certain manner
(Gregory et al. 2002; Happé 1994). Though several tests of
ToM exist in the extant literature, studies will often cal-
F. S. Ahmed (&)  L. Stephen Miller
culate an aggregate ToM score (Saltzman et al. 2000; Fahie
Department of Psychology, University of Georgia, Athens,
Georgia 30602, USA and Symons 2003; Yirmiya et al. 1998). Given the variety
e-mail: fsag@uga.edu of ToM measures, consolidating the different types of ToM

123
668 J Autism Dev Disord (2011) 41:667–678

domains into one score may not be the best method in anxiety, and depression. Impairments have been found in
which to examine ToM. some clinical populations. For example, individuals with
ToM is considered a cognitive construct rather than an autism spectrum disorders, dementia, and bipolar disorder
affective one (Baron-Cohen 1988) and its development have demonstrated poor ToM skills (Baron-Cohen et al.
occurs across several stages of increasing complexity (for a 1985; Baron-Cohen et al. 1997; Bora et al. 2005; Brüne and
review, see Brüne and Brüne-Cohrs 2005). Development Brüne-Cohrs 2005; Cuerva et al. 2001; Gottlieb 2005;
begins at age one and culminates around ages 9 to 11 years Gregory et al. 2002; Happé 1994; Kaland et al. 2002; Kerr
(Brüne and Brüne-Cohrs 2005). Children learn to follow et al. 2003).
another’s gaze, engage in pretend play, understand that
another person can hold a different belief than oneself,
comprehend jokes and irony, and finally conceptualize the Executive Function
faux pas. In addition to the developmental trajectory of
ToM, there has been some research on sex differences in EF refers to higher-order cognitive processes required for
ToM ability. Specifically, females have been shown to individuals to complete goal driven tasks (Lezak et al. 2004
outperform males on measures of emotion recognition p. 35). Examples include the ability to inhibit behavior, shift
(Baron-Cohen et al. 1997) and understanding of faux pas cognitive sets, plan how to solve a problem, and generate
(Baron-Cohen et al. 1999. Additionally, females have been behaviors with categorical or phonemic constrictions
found to perform better on a measure of empathy, in which (Anderson 2002; Delis et al. 2001). The prefrontal cortex
intact ToM has been conceptualized as a necessary com- may control EF because of its ability to incorporate multiple
ponent (Baron-Cohen and Wheelwright 2004). Not all types of information needed for a task (Royall et al. 2002).
studies, however, have found a gender difference (Ozonoff Of note, there is still no agreement of what constitutes all of
et al. 1991). the components of EF, and many models exist to explain EF
From school, work, and peer and intimate relationships, domains (for a review, see Stuss and Knight 2002).
perspective-taking is paramount. Consider the importance Although there is a discrepancy on defining the exact
of successful social interaction in everyday functioning. components of EF, all models agree that EF constitutes the
For example, school requires a child to not only learn steps necessary to solve a complex task (Zelazo and Frye
course material, but to be able to communicate successfully 1998). One approach in particular, the Cognitive-Process
with classmates and teachers. A significant portion of approach, avoids separating out distinct domains of EF and
coursework is made up of group-based activities, such as instead focuses on the skills necessary to complete tradi-
group presentations. The ability to work as an effective tional EF tests (Homack et al. 2005; Delis et al. 2001).
member of a group extends beyond school. In the work-
place, one needs to be able to successfully work with
colleagues. A person needs to be able to listen to what Executive Function and Theory of Mind
other group members say and understand why they are
taking on a specific perspective, especially if it differs from A number of studies have examined the EF-ToM rela-
oneself, as is often the case. This requires ToM because the tionship, and results have been variable. For example, a
person needs to be able to place him/herself in the other case study of an adult who sustained amygdala damage and
person’s mental state to understand why they are making a was diagnosed with both Asperger’s Disorder and schizo-
specific suggestion. Additionally, ToM ability is necessary phrenia found EF deficits but normal ToM functioning
during job interviews and interaction with supervisors, as it (Fine et al. 2001). Another case study of an adult with
is very important to take the perspective of the interviewer impaired orbitofrontal capacity also reported no relation-
or supervisor as well as recognize a socially-inappropriate ship between EF and ToM (Bach et al. 2005). However,
response. Another area that requires successful ToM abil- most studies have found a relationship between EF and
ities is in friendships. Humans are social creatures and need ToM. In a comprehensive review, Hughes and Graham
social interaction. When establishing friendships, one (2002) concluded that in addition to EF and ToM deficits in
needs ToM abilities in order to understand others. This autism spectrum disorders, the two domains are related to
becomes much more important when searching for a sig- one another. In a study of children with autism, it was
nificant other. Because everyday living needs successful found that performance on ToM was significantly and
social interactions, any deficits in ToM abilities can sig- positively related to working memory and inhibition
nificantly impact one’s life. The inability to interact with (Joseph and Tager-Flusberg 2004). Fisher and Happé
others can impede success at school, work, and finding/ (2005) examined the clinical implications of the EF-ToM
maintaining friends. This may lead to a lack of positive relationship and trained two groups of children in either
social relationships, which can lead to feelings of isolation, ToM skills or EF skills. They found that better performance

123
J Autism Dev Disord (2011) 41:667–678 669

in EF skills was positively associated with better perfor- becomes more difficult to draw conclusions. Additionally,
mance in ToM, suggesting an association between EF and Apperly et al. (2009) emphasized that it is important to
ToM in the autism spectrum disorders (Fisher and Happé examine the role of EF in ToM in adults, as it can give
2005). insights about the importance of EF in the development
This EF-ToM association has also been found outside and/or maintenance of ToM in adulthood. The authors
this group. In children with attentional and behavioral stated that without studying ToM in adults, it is difficult to
problems, ToM and working memory were found to be pinpoint the role of EF. They noted that EF may be crucial
significantly and positively related (Fahie and Symons in the development of ToM but may not be necessary in
2003). In a similar sample, Hughes et al. (1998) found that adulthood, or, that EF may also be an important process in
EF and ToM were positively correlated. This relationship the fully-developed ToM of adults. This question cannot be
has been found in non-clinical populations as well. In a answered by studying children alone, but needs to be
study of preschoolers by Gordon and Olson (1998), results coupled with research on adults as well (Apperly et al.
indicated better ToM performance as a function of better 2009). This, therefore, is another reason that it is important
performance of EF. Cole and Mitchell (2000) found similar to examine the EF-ToM relationship in a typically-devel-
conclusions in three to five-year-old children. A longitu- oping adult population.
dinal study by Hughes (1998) demonstrated that EF pre-
dicted ToM performance over time, but ToM did not
predict EF. The EF-ToM relationship has also been dem- Aims
onstrated in a non-Western culture. In a study that exam-
ined preschoolers from China and the United States, To avoid the complication among the effects of cognitive
researchers found that planning, cognitive flexibility, and development when examining the ToM-EF relationship as
inhibition were positively correlated with ToM (Sabbagh well as further examine the role of EF in the fully-devel-
et al. 2006). This relationship is not specific to children, oped ToM, this study used a sample of typically developing
either. Saltzman et al. (2000) found that design fluency, adults because they should have fully developed their ToM
problem solving, and verbal fluency were positively cor- abilities (Brüne and Brüne-Cohrs 2005). By using a typi-
related with ToM in an elder population. cally-developing population, it can be easier to infer the
Of note, there has also been research examining whether relationship between EF and ToM. Based on the results of
specific EFs drive ToM or if they simply share a basic Hughes (1998), in which EF was found to predict ToM but
general processing domain. McKinnon and Moscovitch not the other way around, this study will examine the
(2007) found that on measures of social cognition (ToM predictive ability of EF on ToM using a comprehensive set
and deontic reasoning), performance relied on a general of EF domains. Furthermore, to our knowledge, there has
central processing component. This study was followed by not been a study that examines the relationship among a
Bull et al. (2008), in which a dual-task interference task wide range of EF components and ToM. Previous studies
was added to ToM measures. The interference tasks were only compared a limited number of EF domains, such as
differing types of general EF domains (inhibition, switch- inhibition and working memory (Carlson et al. 2002,
ing, updating). The dual-task conditions significantly 2004a, b). This study utilizes multiple EF measures and
affected ToM performance and the authors suggested that multiple levels of ToM tests to obtain a comprehensive
this occurred because both ToM and EFs share global assessment of the EF-ToM relationship.
attentional reserves (Bull et al. 2008). Thus far, there The specific aims of this study were (a) to evaluate the
appear to be only a few studies have attempted to identify relationship between ToM tests with one another; and (b)
the specific EF domains which share the most variance to identify which domains of EF best predict ToM, as
with ToM. Limited research examining this question have obtained from a non-clinical adult population. By con-
noted that both inhibition and working memory appear to ducting this study, we hope to provide further insight into
have the strongest association with ToM, and that this the cognitive mechanisms that drive the complex process
relationship emerges at 39 months of age (Carlson et al. of ToM. As aforementioned, there are multiple clinical
2002, 2004a, b). Further, most of these studies have only populations who have shown deficits in ToM and thus
examined children (Carlson et al. 2002, 2004a, b; Cole and suffer from unsuccessful social interaction (Baron-Cohen
Mitchell 2000; Fahie and Symons 2003; Fisher and Happé et al. 1985, 1997; Bora et al. 2005; Brüne and Brüne-Cohrs
2005; Joseph and Tager-Flusberg 2004; Perner and Lang 2005; Cuerva et al. 2001; Gottlieb 2005; Gregory et al.
1999; Sabbagh et al. 2006). The limitation of this is that 2002; Happé 1994; Kaland et al. 2002; Kerr et al. 2003).
children are still in the developing stages of ToM (Brüne Therefore, we hope to contribute to the ongoing ToM
and Brüne-Cohrs 2005), and, EF is not fully developed research on clinical populations by further describing the
until after childhood (Anderson et al. 2002). Therefore, it underlying skills necessary for successful ToM.

123
670 J Autism Dev Disord (2011) 41:667–678

Method RMET, Strange Stories test, and Faux Pas test. The Strange
Stories test included six randomized sets that were cycled
Participants through every six participants. Tests were administered and
scored according to standard protocol (Baron-Cohen et al.
One hundred and thirty-five research participants from the 2001; Delis et al. 2001; Gregory et al. 2002; Happé 1994;
University of Georgia Research Participant Pool were Stone et al. 1998).
recruited to participate in this study. Participants were After the completion of testing, participants received a
excluded from the study based on the presence of ongoing written and verbal debriefing of the study’s aims.
psychosis, depression, current psychiatric medication, or
poor effort. To screen for psychosis, the Psychotic Screen Measures
from the Structured Clinical Interview for DSM-IV Axis I
Disorders (First et al. 1996) was used as a measure of gross Delis-Kaplan Executive Function System
psychological impairment. Endorsement of any of these
items (e.g., visual or auditory hallucinations, persecutory or The Delis-Kaplan Executive Function System (D-KEFS) is
grandiose delusions, etc.) excluded the participant’s data a battery of nine subtests that measure components of EF.
from this study. To screen for depression, participants were This battery is based on several traditional neuropsycho-
administered the Beck Depression Inventory (BDI; The logical tests that have been updated and normed on the
Psychological Corporation 1996), which indicates current same sample. This is the first battery of EF tests that has
depressive symptoms. Data from participants who obtained been normed on such a large sample (1,750 adults and
a BDI score of C20 (i.e., moderate depression) or more children) and can be administered to individuals between 8
were excluded from the study. The Medical Symptom and 89 years. There are nine subtests, including (a) Word
Validity Test (MSVT), an empirically-supported, quantita- Context Test, (b) Sorting Test, (c) Twenty Questions Test,
tive measure of effort, was administered to screen for poor (d) Tower Test, (e) Color-Word Interference, (f) Verbal
effort (Richman et al. 2006; Merten et al. 2005). Five par- Fluency Test, (g) Design Fluency Test, (h) Trail Making
ticipants were excluded from this study due to endorsing Test, and (i) Proverb Test. They each yield multiple scores;
symptoms of current psychosis. Three participants were however one EF domain score can be pulled from each test,
excluded due to self report of significant depressive symp- which are (a) cognitive flexibility, (b) verbal fluency, (c)
toms, and current use of psychiatric medication excluded design fluency, (d) inhibition, (e) problem solving, (f)
four participants. No participants were excluded due to poor categorical processing, (g) deductive reasoning, (h) spatial
performance on the MSVT. The final number of participants planning, and (i) verbal abstraction (Delis et al. 2001).
included in this study was 123. Participants ranged in age Cognitive flexibility is postulated as the capacity to adapt
from 18 to 27 years (M = 19.04, SD = 1.30). to new rules. Verbal and design fluency constitute the
ability to quickly make unique words or designs. Inhibition
Procedure requires one to resist a prepotent response. Problem solving
refers to the capacity to successfully solve a problem. In
Each participant was tested in a single session in the categorical processing, one has to be able to systematically
Neuropsychology and Memory Assessment Laboratory at organize information. In order to have adequate deductive
the University of Georgia. Informed consent was obtained reasoning skills, one has to be able to use clues in order to
by a written form approved by the University’s Institu- solve a puzzle. Verbal abstraction refers to the capacity to
tional Review Board. Participants also received verbal not literally comprehend all statements (Delis et al. 2001).
explanation of their rights as well as course credit. Because of the large age range of the D-KEFS, it has low
Demographic information was obtained and participants floors and high ceilings. Furthermore, there is moderate
were administered the Wechsler Test of Adult Reading reliability and validity. Test–retest reliability across subtest
(WTAR; The Psychological Corporation 2001) in order to scores is between .06 and .90. The internal consistency
obtain an estimated Full Scale IQ (FSIQ) score. This was across subtests is between .33 and .90 (Delis et al. 2001).
calculated to ensure IQ equivalence among participants. The D-KEFS tests were administered and scored
Next, participants were administered the D-KEFS and according to standard protocol (Delis et al. 2001).
three well-known ToM tests (Reading the Mind in the Eyes
test (RMET; Baron-Cohen et al. 2001), Strange Stories Reading the Mind in the Eyes Test
Test (Brent et al. 2004; Happé 1994), and Faux Pas Test
(Gregory et al. 2002; Stone et al. 1998). The order of these The Reading the Mind in the Eyes Test (RMET) has been
tests (EF battery first or ToM battery first) were counter- developed as a subtle measure of ToM assessing a funda-
balanced. The order of the ToM tests were as follows: mental skill needed for ToM: emotion recognition. The

123
J Autism Dev Disord (2011) 41:667–678 671

RMET has been able to to identify ToM deficits in both Administration of the Strange Stories test adhered to the
healthy controls and individuals with high-functioning standard protocol which consisted of reading the vignettes
Autism and Asperger’s Disorder (Baron-Cohen et al. aloud to the participants while the participants were pro-
1997). This test is made up of 36 photos of actors’ eyes and vided with a written version of the vignettes as well (Happé
requires the participant to identify the emotion that the 1994) except for nine participants who read the stories
actor is portraying (Baron-Cohen et al. 2001). themselves. An independent samples t-test indicated no
This study used a revised version of the test that is differences between those who had the stories read to them
designed for adults. It has been shown to have better psy- and followed along versus those who only read the stories
chometric qualities than the original version, primarily in themselves (t (121) = .367, p [ .05). The majority of the
reducing normal performance to below ceiling compared to Strange Stories tests were scored by two raters, and inter-
the original (Baron-Cohen et al. 2001). Validity has been rater reliability was established (q = .87, p \ .01).
good. Scores on the Autism Quotient and the RMET were Therefore, an average variable was calculated for the
inversely correlated (r = -.53, p \ .01), with increasing majority of participants. The remainder of the tests were
severity of Autism associated with poorer scores on the scored by the first author alone, and interrater reliability
RMET (Baron-Cohen et al. 2001). As previously noted, the was established among the first author and the two raters
amount of literature regarding variance among ToM mea- from a subset of the data (q = .89, q = .89, p \ .01). This
sures is limited. The RMET was administered and scored yielded a test variable that consisted of either the average
according to standard protocol (Baron-Cohen et al. 2001). score of the two raters or the sole score rated by the first
author.
Strange Stories Test
Faux Pas Test
The Strange Stories test is another subtle test of ToM
(Happé 1994; Jolliffe and Baron-Cohen 1999). It is sensi- The Faux Pas test is another subtle measure of ToM. A
tive in detecting ToM deficits in individuals who generally faux pas is defined as an interaction in which one person
pass other simpler measures (Happé 1994). This test inadvertently makes an inappropriate comment that results
assesses more advanced concepts such as double bluff, in negative feelings from the other person (Gregory et al.
white lies, and persuasion. Participants are presented with a 2002; Stone et al. 1998).
series of vignettes and are asked to explain why the main The Faux Pas test has been shown to have good inter-
character acted in a particular manner. rater reliability (r = .98; Gregory et al. 2002). Research
A revised shorter version has also been found to be a also indicates that it is a valid measure of understated ToM
sensitive measure of ToM in adults (Happé 1994; Jolliffe impairment, as children with Asperger’s Disorder were
and Baron-Cohen 1999; Happé et al. 1998; Maylor et al. unable to correctly detect a faux pas but were able to pass
2002; Sullivan and Ruffman 2004). A different revised simple false-belief tasks (Baron-Cohen et al. 1999).
version was used in this study (with permission from the Additionally, research on adults found that those with
test developer) which contains 16 items (8 control and 8 damage to the orbitofrontal cortex were unable to detect a
experimental) and developed specifically for younger and faux pas but were able to pass false belief tasks (Stone et al.
older participants. To the best of our knowledge, this par- 1998). To the knowledge of the authors, concordance
ticular subset of stories has not been used though the var- between the Faux Pas test and other ToM tests has not
ious individual stories have all been used (Brent et al. generally been reported in studies using this test (Shaw
2004; Happé 1994). Therefore, specific data of the reli- et al. 2004).
ability and validity of the exact set of stories used for this Administration of the Faux Pas test adhered to the
study are unknown. standard protocol which consisted of reading the vignettes
The Strange Stories test scores has been found to have aloud to the participants while the participants were pro-
strong interrater reliability (87%) (Happé et al. 1998). vided with a written version of the vignettes as well
Validity has also been good. Specifically, it has consis- (Gregory et al. 2002; Stone et al. 1998). As with the
tently shown poorer performance in participants with Strange Stories, the same nine participants read the stories
Autism Spectrum Disorders compared to typically devel- to themselves and analyses similarly found no differences
oping controls, thus demonstrating a deficient ToM ability between the groups (t (121) = -1.56, p [ .05). Interrater
in those with Autism Spectrum Disorders. Additionally, reliability was similarly established as with the Strange
the Strange Stories test has demonstrated subtle differences Stories Test (q = .91, p \ .01) and an average variable
in test performance between severely autistic participants calculated for those with two raters. Remaining scores were
and higher-functioning autistic participants (Happé 1994; singularly scored by the first author following adequate
Kaland et al. 2002; Gottlieb 2005). interrater reliability of a subset of the data (q = .96,

123
672 J Autism Dev Disord (2011) 41:667–678

q = .89, p \ .01). This yielded a test variable that con- D-KEFS EF domains yielded a single scaled score from
sisted of either the average score of the two raters or the one representative task as a predictor variable. Condition 4
sole score rated by the first author. of the Trail-Making Test referred to cognitive flexibility.
Of note, none of the three ToM measures used were The Letter Fluency condition of the Verbal Fluency Test
false belief tasks. These three tasks were utilized in this indicated verbal fluency, and design fluency was repre-
study because they are considered higher-order tasks that sented by Condition 1 of the Design Fluency Test. Con-
are designed for people who do not generally fail false dition 3 of the Color-Word Interference Test designated
belief tasks (Baron-Cohen et al. 1999). According to Happé inhibition, problem solving was indicated by the confirmed
(1994), it is common for individuals with higher-func- correct sorts of the Free Sort condition of the Sorting Test,
tioning status on the autism spectrum to pass first-order and the total weighted achievement of the Twenty Ques-
false belief tasks. Since this study examined typically- tions Test represented categorical processing. Deductive
developing adults, it appeared appropriate to use these reasoning was indicated by the total number consecutively
measures. correct of the Word Context Test, total achievement of the
Tower Test represented spatial planning, and the cumula-
Variables tive percentile rank from the Multiple Choice condition of
the Proverb Test indicated verbal abstraction (Delis et al.
Though various ToM tests are commonly aggregated into 2001).
one score (Saltzman et al. 2000; Fahie and Symons 2003;
Yirmiya et al. 1998), earlier pilot data suggested a potential
lack of correlation (Ahmed et al. 2007). Thus we first Results
examined whether the ToM scores were correlated before
considering aggregating them. For the ToM tests, variables Demographic information for the 123 participants is sum-
were derived from the total number correct from the marized in Table 1.
RMET, total score of the Strange Stories test, and the total Descriptive data on the D-KEFS and ToM scores are
items correct from the Faux Pas test. Each of the nine summarized in Table 2.

Table 1 Demographic data


Frequency Percent Minimum Maximum Mean Standard
deviation

Agea 18 27 19.04 1.30


Gender
Male 53 42.7
Female 70 56.5
Ethnicity
White/non hispanic 96 78.0
African American/Black 6 4.9
Hispanic/Latino 5 4.1
Asian/Pacific Islander 14 11.4
Other 2 1.6
Family income
Below $10,000 1 .8
$10,001–$30,000 7 5.7
$30,001–$60,000 23 18.7
$60,001–$90,000 37 30.1
$90,001 and above 55 44.7
Geographical region raised
South 106 86.2
North Central 1 .8
N = 123 West 3 2.4
a
In years Northeast 6 4.9
b
WTAR-predicted Other 6 4.9
FSIQ = Wechsler test of adult b
WTAR-predicted FSIQ 86 122 108.91 6.90
reading-predicted full scale IQ

123
J Autism Dev Disord (2011) 41:667–678 673

Table 2 Descriptive statistics of D-KEFSa and ToMb variables Correlational analyses were used to evaluate the rela-
Minimum Maximum Mean Standard
tionship among ToM tests. As all three ToM tests were in
deviation violation of normality based on the Kolmogorov–Smirnov
statistic (all p’s \ .01), rank-order correlations were con-
D-KEFS: cognitive 4 14 10.78 1.87
ducted. None of the three ToM tests were significantly
flexibility
correlated with each other (see Table 3).
D-KEFS: verbal fluency 5 19 12.19 3.48
A correlation matrix was also calculated for the D-KEFS
D-KEFS: design fluency 4 19 11.98 3.05
and ToM variables (see Table 3).
D-KEFS: inhibition 6 19 11.51 2.25
Since the ToM tests were not correlated with one
D-KEFS: problem solving 1 15 9.13 2.62
another, three separate regression analyses were conducted
D-KEFS: categorical 5 15 11.54 2.34
processing
with each of the three variables from the ToM tests as
single dependent variables. D-KEFS variables that were
D-KEFS: deductive 6 16 11.71 2.11
reasoning correlated with each ToM measure were entered as inde-
D-KEFS: spatial planning 4 17 11.17 2.18 pendent variables (see Table 3). Hierarchical regression
D-KEFS: verbal 5 100 87.49 28.63 analyses were employed by entering any significantly
abstraction correlated demographic variables in the first step of the
RMETc 8 34 27.28 3.75 model and then adding D-KEFS variables into the second
Strange stories 22 32 27.41 2.23 step of the model. Outlier analyses were also calculated
Faux Pas 76 121 112.94 8.58 separately for each regression. All results from regression
analyses were in a positive direction.
N = 123
a
D-KEFS Delis-Kaplan Executive Function System. D-KEFS scores
based on M = 10, SD = 3. D-KEFS verbal abstraction based on Executive Function and Performance on the RMET
percent correct
b
ToM Theory of Mind. ToM scores based on total correct (RMET To determine whether any of the demographic variables
ranges from 0 to 36; Strange Stories Test ranges from 0 to 32; Faux (i.e., age, gender, ethnicity, family income, geographical
Pas Test ranges from 0 to 121) region raised, and WTAR-predicted FSIQ) significantly
c
RMET Reading the Mind in the Eyes Test accounted for variance in the RMET scores, they were

Table 3 Correlation Matrix of D-KEFSa and ToMb variables


CF VF DF I PS CP DR SP VA RMET Strange stories Faux Pas

CF 1.00 .25** .26** .28** .28** .07 .15 .09 .05 .03 -.01 .07
VF 1.00 .29** .44** .22* .18* .29** .02 .11 .23* .22* .22*
DF 1.00 .26** .10 .12 .15 .04 -.020 .24** .19* .10
I 1.00 .26** .02 .28** .04 .16 .22* -.00 .06
PS 1.00 .19* .23* .07 .21* -.09 -.05 .21*
CP 1.00 .04 .06 .10 .06 .01 .19*
DR 1.00 .13 .17 .21* .18* .14
SP 1.00 -.04 -.03 .18* .04
VA 1.00 .120 .08 .13
RMETc 1.000 .14 .13
Strange stories 1.00 .11
Faux Pas 1.00
N = 123. Spearman’s Rank-Order Correlations were used for Correlations among ToM tests. All other correlations used Pearson’s correlations
* p \ .05; ** p \ .01
a
D-KEFS Delis-Kaplan Executive Function System, D-KEFS variables: CF cognitive flexibility, VF verbal fluency, DF design fluency,
I inhibition, PS problem solving, CP categorical processing, DR deductive reasoning, SP spatial planning, VA verbal abstraction
b
ToM Theory of Mind
c
RMET Reading the Mind in the Eyes Test

123
674 J Autism Dev Disord (2011) 41:667–678

Table 4 Multiple regression analysis for prediction of RMETa by R = .42, R2 = .17 Adj. R2 = .14, R2 change = .05, Standard Error
D-KEFSb testsc,d Step 1: R = .36, R2 = .13, Adj. R2 = .12, Standard of Estimate = 3.07, F (4,115) = 4.84, p \ .01
Error of Estimate = 3.11, F (1,119) = 17.22, p \ .01 Step 2:
Step Beta t Sig. (p) Correlations
Zero-order Partial Part

1 WTAR-predicted .17 4.15 \.01 .36 .36 .36


FSIQ
2 WTAR-predicted .13 2.82 \.01 .36 .25 .24
FSIQ
Verbal fluency .04 .36 [.05 .23 .03 .03
Design fluency .14 1.39 [.05 .22 .13 .12
Inhibition .13 .83 [.05 .26 .08 .07
Deductive reasoning .15 1.10 [.05 .22 .10 .09

N = 121. The dependent variable is the RMET total raw score (range = 0–36)
a
RMET Reading the Mind in the Eyes Test
b
D-KEFS Delis-Kaplan Executive Function System
c
Only D-KEFS variables that were significantly correlated with the RMET were entered in the model
d
D-KEFS-RMET significant correlations: a verbal fluency (r = .23, p \ .05), b design fluency (r = .24, p \ .01), c inhibition (r = .22,
p \ .05), and d deductive reasoning (r = .21, p \ .05)

regressed on the RMET variable. The only significant Verbal fluency (r = .22, p \ .05), design fluency
variable was WTAR-predicted FSIQ scores (b = .39, (r = .19, p \ .05), deductive reasoning (r = .18, p \ .05),
t (115) = 4.49, p \ .01), so it was entered into the first step and spatial planning (r = .18, p \ .05) were all signifi-
of the model. cantly correlated with Strange Stories performance (see
The D-KEFS variables that correlated with RMET per- Table 3) so they were simultaneously entered into the
formance were entered into the second step of the model regression model, with the Strange Stories test as the
(see Table 3). Verbal fluency (r = .23, p \ .05), design dependent variable. A Cook’s D analysis (value of 4/123
fluency (r = .24, p \ .01), inhibition (r = .22, p \ .05), [.0325]) identified seven outliers. After removing these
and deductive reasoning (r = .21, p \ .05) were the only outliers from the data set, the regression analysis was
significantly correlated variables. Next, a Cook’s D anal- re-run (see Table 5).
ysis (value of 4/123 [.0325]) identified two outliers. After
removing these two participants from the data set, the
Table 5 Multiple regression analysis for prediction of strange stories
regression analysis was re-run (see Table 4).
testa by D-KEFSa testsc,d
The first step of the model significantly predicted vari-
ance in RMET (R2 = .126, F (1, 119 = 17.22, p \ .01) Beta t Sig. (p) Correlations
and WTAR-predicted FSIQ scores significantly predicted Zero-order Partial Part
performance on the RMET (b = .17, t (119) = 4.15,
Verbal fluency .14 2.44 \.05 .32 .23 .21
p \ .01). The overall model also accounted for a signifi-
Design fluency .10 1.58 [.05 .25 .15 .13
cant amount of variance in RMET scores (R2 = .17,
Deductive reasoning .21 2.33 \.05 .31 .22 .20
R2 = .05, F (5, 115) = 4.84, p \ .01). However, this
Spatial planning .17 1.97 [.05 .20 .18 .17
continued to be driven by WTAR-predicted FSIQ
2 2
(b = .13, t (115) = 2.82, p \ .05) and none of the R = .45, R = .21, Adjusted R = .18 Standard Error of Esti-
D-KEFS variables were significant. mate = 1.88, F (4, 111) = 7.14; p \ .01
N = 116
a
Executive Function and Performance on the Strange The dependent variable is the Strange Stories Test total raw score
Stories Test (range = 0–32)
b
D-KEFS Delis-Kaplan Executive Function System
c
As above, demographic variables were initially entered in a Only D-KEFS variables that were significantly correlated with the
regression analysis to determine whether they accounted Strange Stories Test were entered in the model
d
for a significant amount of variance in Strange Stories test D-KEFS-Strange Stories Test significant correlations: a verbal
fluency (r = .22, p \ .05), b design fluency (r = .19, p \ .05),
scores. None were significant, so they were left out of c deductive reasoning (r = .18, p \ .05), and d spatial planning
further analysis. (r = .18, p \ .05)

123
J Autism Dev Disord (2011) 41:667–678 675

The model accounted for a significant proportion of continued to account for a significant portion of variance
variance in Strange Stories performance (R2 = .21, F (4, (b = 4.99, t (111) = 5.04, p \ .01). Of the D-KEFS vari-
111) = 7.139, p \ .01; see Table 5). Verbal fluency ables, verbal fluency (b = .32, t (111) = 2.22, p \ .05) and
(b = .14, t (111) = 2.44, p \ .05) and deductive reasoning problem solving (b = .50, t (111) = 2.45, p \ .05)
(b = .21, t (111) = 2.33, p \ .05) predicted a significant explained a statistically significant amount of variance in
amount of variance in Strange Stories scores. Design flu- Faux Past test performance. Categorical processing was
ency and spatial planning were nonsignificant. nonsignificant. In order to examine the gender effect, an
independent-samples t-test was used. Because Levene’s test
Executive Function and Performance on the Faux Pas for equality of variance was significant, a Brown-Forsyth
Test test was used (Brown-Forsyth test statistic = 16.442,
p \ .01), which indicated that females performed modestly,
The amount of variance accounted for by demographic though significantly, better than males (mean score
variables was again calculated. Gender significantly pre- females = 116, mean scores males = 112).
dicted performance on the Faux Pas test (b = .34,
t (116) = 3.90, p \ .01) so it was entered into the first step
of the regression analysis. Verbal fluency (r = .22, Discussion
p \ .05), problem solving (r = .21, p \ .05), and cate-
gorical processing (r = .19, p \ .05) were significantly Our first aim was to evaluate the relationship between the
correlated with the Faux Pas test (see Table 3) and were three ToM measures. The lack of correlations among the
entered into the second step of the regression model. three ToM tests suggest that although all were designed to
A Cook’s D analysis (value of 4/123 [.0325]) identified measure ToM (Baron-Cohen et al. 2001, 1997; Gottlieb
seven outliers. After removing these participants from the 2005; Gregory et al. 2002; Happé 1994; Happé et al. 1998;
data set, the regression analysis was re-run (see Table 6). Jolliffe and Baron-Cohen 1999; Kaland et al. 2002; Maylor
The first step of the model accounted for a significant et al. 2002; Stone et al. 1998; Sullivan and Ruffman 2004),
portion of variance (R2 = .14, F (1, 114) = 18.75, p \ .01) they may be each tapping into very different domains. This
and gender was significant (b = 4.47, t (114) = 4.33, supports and expands preliminary research from this lab-
p \ .01). The overall model was also significant (R2 = .26, oratory that also suggested each of these ToM tasks may
R2 change = .11, F (4, 111) = 9.48, p \ .01). Gender utilize different cognitive processes (Ahmed et al. 2007).
Research by Brent et al. (2004) also suggested that various
Table 6 Multiple regression analysis for prediction of Faux Pas Testa higher-order ToM tasks may be related to differing infor-
by D-KEFSa Testsc,d mation-processing components. Furthermore, different EF
Step Beta t Sig. Correlations
patterns were found for each ToM test, which suggests that
(p) there may be differing cognitive processes that are asso-
Zero- Partial Part ciated with the apparent different domains of ToM mea-
order
sured by each test. It is not rare for ToM tests to be
1 Gender 4.47 4.33 \.01 .38 .38 .38 aggregated into a composite score (Salztman et al. 2000;
2 Gender 4.99 5.04 \.01 .38 .43 .41 Fahie and Symons 2003; Yirmiya et al. 1998). These
Verbal fluency .32 2.22 \.05 .24 .21 .18 results, however, suggest that this practice may not be the
Problem solving .50 2.45 \.05 .19 .23 .20 best manner in which to analyze one’s data.
Categorical .20 .92 [.05 .15 .09 .08 The RMET requires a participant to study the expres-
processing sions of actors’ eyes and determine which emotion they are
Step 1: R = .38, R2 = .14, Adjusted R2 = .13, Standard Error of conveying from four possible choices (Baron-Cohen et al.
Estimate = 5.49, F (1,114) = 18.75, p \ .01 Step 2: R = .51, 2001). None of the EF variables accounted for a significant
R2 = .26, Adjusted R2 = .23, R2 change = .11, Standard Error of proportion of variance in RMET performance. A measure
Estimate = 5.18 F (4,111) = 9.48, p \ .01
of word-reading ability, however, was predictive of RMET
N = 116 performance. The RMET has been conceptualized as a
a
The dependent variable is the Faux Pas Test total raw score ToM task due to the concept that emotion recognition is the
(range = 0–121)
b foundation of ToM abilities (Baron-Cohen et al. 2001). The
D-KEFS Delis-Kaplan Executive Function System
c
results found in this study suggest that the EF domains
Only D-KEFS variables that were significantly correlated with the
Faux Pas Test were entered in the model
measured in this study did not drive this particular task of
d
D-KEFS-Faux Pas Test significant correlations: a verbal fluency
emotion recognition. It may be that the aspect of ToM
(r = .22, p \ .05), b problem solving (r = .21, p \ .05), and c cate- measured in the RMET, emotion recognition, is so differ-
gorical processing (r = .19, p \ .05) ent from the areas of ToM measured by the Strange Stories

123
676 J Autism Dev Disord (2011) 41:667–678

and Faux Pas tests that it does not include a significant EF gender differences finding females performing better than
mechanism. It should be noted that the Strange Stories and males (Baron-Cohen and Wheelwright 2004; Baron-Cohen
Faux Pas tests both require one to be able to clearly express et al. 1997, 1999) in general as well as specific to the Faux
their responses in a verbal format, whereas the RMET Pas test (Baron-Cohen et al. 1999).
requires one to choose an appropriate word from a choice A strength of this study was the use of a typically-
of four options (Baron-Cohen et al. 2001; Gregory et al. developing population in regards to expected ToM func-
2002; Happé 1994; Stone et al. 1998). Successful perfor- tioning. As previously mentioned, much of the ToM
mance on the RMET may only require a basic under- research has been conducted on populations in which there
standing of emotion words. Emotion recognition has been are deficiencies in ToM, by either studying children or
referred to as the first step in ToM without the more clinical populations with difficulties in ToM (Brüne and
complicated task of deciphering the beliefs of the charac- Brüne-Cohrs 2005; Carlson et al. 2002, 2004a, b; Cole and
ters (Baron-Cohen et al. 2001). In fact, it has been argued Mitchell 2000; Fahie and Symons 2003; Fisher and Happé
that the ability to recognize an emotional state develops as 2005; Joseph and Tager-Flusberg 2004; Perner and Lang
early as infancy (Nelson 2001). 1999; Sabbagh et al. 2006). It makes it more difficult to
Of domains of EF variables entered into the regression draw conclusion about an EF-ToM relationship when ToM
model for predicting Strange Stories performance, verbal is not intact. Therefore, this study attempted to control for
fluency and deductive reasoning were significant. The this problem by assessing a large sample of adults with
Strange Stories test uses a series of vignettes in which typically-developed ToM. Additionally, this study utilized
participants have to identify the intentions of the stories’ a broad range of EF variables, which to our knowledge has
characters (Happé 1994). Verbal and design fluency require not been done before. However, a related limitation of this
a person to quickly form words according to a rule (Delis study was the rather homogenous sample. There was a
et al. 2001). This can be conceptualized as the ability to limited range of age, predicted IQ, and ethnicity. This may
flexibly initiate responses. A person does not memorize one have restricted the amount of variance in the test results
type of social interaction. Instead, one needs to generalize and one should be cautious about generalizing these results
basic concepts of social interaction to a variety of unique beyond this limited sample group. A further limitation was
situations, thus utilizing ToM through successful applica- that, while gross psychological screening was conducted,
tion of flexible initiation of behavior. Deductive reasoning we did not specifically exclude persons with a history of
relies on the ability to solve a puzzle from clues (Delis autism spectrum disorder or attention deficit disorder.
et al. 2001). Since the Strange Stories test requires a person Although several studies have found links between
to figure out the intentions of the characters (Happé 1994), working memory and inhibition on ToM (Carlson et al.
one has to deduce why a character behaved the way he or 2002; 2004a; b; Mckinnon and Moscovitch 2007), results
she did. from this study did not find significant links between them.
For the Faux Pas test, verbal fluency and problem In young adults, working memory interfered with perfor-
solving significantly accounted for performance on that mance of tests of social reasoning (Mckinnon and Mos-
test. The problem-solving variable came from the D-KEFS covitch 2007). This study did not specifically test for
Sorting Test’s confirmed correct sorts, in which partici- working memory, as the D-KEFS measures do not give a
pants have to sort a set of cards in multiple ways (Delis working memory score (Delis et al. 2001). Therefore, the
et al. 2001). This appears to be a more complicated task results of the current study neither support nor deny the
than the measure of deductive reasoning, which requires impact of working memory on ToM. It should be noted,
one to solve a nonsense word using contextual cues (Delis however, that the nature of the D-KEFS measures requires
et al. 2001). The Faux Pas test measures a person’s ability one to be able to hold and manipulate information men-
to simultaneously understand the beliefs of two characters tally. Participants need to hold different instructional sets
(Gregory et al. 2002; Stone et al. 1998), whereas the as they carry out the tasks of each test, such as the changing
Strange Stories test measures a person’s ability to under- instructions of the Trail Making Test (Delis et al. 2001).
stand the beliefs of one character (Happé 1994). Perhaps Regarding inhibition, the results of this project did not find
this increased complication in demands on the Faux Pas inhibition to significantly predict ToM performance,
test explains why the more complex D-KEFS measure of although it was significantly correlated with the RMET
problem-solving accounted for a significant amount of (r = .22, p \ .05). As previously mentioned, working
variance than the somewhat simpler measure of deductive memory and inhibition were found to be significant pre-
reasoning. dictors of ToM in children (Carlson et al. 2002, 2004a, b).
Compared to the EF variables, however, gender Perhaps inhibition plays a more crucial role in the devel-
accounted for much more variance in Faux Pas perfor- opment of ToM but plays a more minor role in the
mance. Our findings supported previous literature of developed ToM. According to Apperly et al. (2009), it is

123
J Autism Dev Disord (2011) 41:667–678 677

possible that EFs, or certain EFs, are more necessary in the frontal cortex in acquired brain injury. Cognitive Neuropsychi-
development of ToM but not for its maintenance. However, atry, 5, 175–192.
Baron-Cohen, S. (1988). Social and pragmatic deficits in autism:
it is still too early to make this claim, as it is possible that Cognitive or affective? Journal of Autism and Developmental
inhibition was simply not a necessary EF for the ToM tests Disorders, 18, 379–402.
used in this study. Baron-Cohen, S., Jolliffe, T., Mortimore, C., & Robertson, M. (1997).
This study supported previous preliminary research that Another advanced test of theory of mind: Evidence from very
high functioning adults with autism or asperger syndrome.
suggested that the RMET, Strange Stories test, and Faux Journal of Child Psychology and Psychiatry, 38, 813–822.
Pas tests utilized differing cognitive mechanisms. Except Baron-Cohen, S., Leslie, A. M., & Frith, U. (1985). Does the autistic
for the RMET, some EF domains significantly accounted child have a ‘‘theory of mind’’? Cognition, 21, 37–46.
for variance in ToM performance. It is possible that inci- Baron-Cohen, S., O’Riordan, M., Stone, V., Jones, R., & Plaisted, K.
(1999). Recognition of faux pas by normally developing children
dental task demands of ToM tests are similar to EF mea- and children with asperger syndrome or high-functioning autism.
sures. Specifically, these tests may be sharing a general Journal of Autism and Developmental Disorders, 29, 407–
attentional capacity (Bull et al. 2008; McKinnon and 418.
Moscovitch 2007). A follow-up to this study may be able to Baron-Cohen, S., & Wheelwright, S. (2004). The empathy quotient:
An investigation of adults with asperger syndrome or high-
address this question by partitioning the skills necessary to functioning autism, and normal sex differences. Journal of
successfully complete the EF measures that significantly Autism and Developmental Disorders, 34, 163–175.
predicted ToM performance (i.e., verbal fluency, deductive Baron-Cohen, S., Wheelwright, S., Hill, J., Raste, Y., & Plumb, I.
reasoning, and problem solving). Finally, these results were (2001). The ‘‘reading the mind in the eyes’’ test revised version:
A study with normal adults, and adults with asperger syndrome
observed in a non-clinical sample. It may be useful to or high-functioning autism. Journal of Child Psychology and
replicate this study using a population that does not have Psychiatry, 42, 241–251.
fully-developed ToM abilities. It is unknown whether these Bora, E., Vahip, S., Gonul, A. S., Akdeniz, F., Alkan, M., Ogut, M.,
same results would reappear, as clinical populations would et al. (2005). Evidence for theory of mind deficits in euthymic
patients with bipolar disorder. Acta Psychiatrica Scandinavica,
likely have a larger range of demographic variables and, 112, 110–116.
perhaps, test scores. With this increased variability, the Brent, E., Rios, P., Happé, F., & Charman, T. (2004). Performance of
results may be different. To be able to establish an EF-ToM children with autism spectrum disorder on advanced theory of
relationship, there has to be further research in both clinical mind tasks. Autism, 8, 283–299.
Brüne, M., & Brüne-Cohrs, U. (2005). Theory of mind-evolution,
and nonclinical populations using the most comprehensive ontogeny, brain mechanisms and psychopathology. Neurosci-
batteries possible. ence and Behavioral Reviews, xx, 1–19.
Bull, R., Phillips, L. H., & Conway, C. A. (2008). The role of control
Acknowledgments We would like to thank Francesca Happé and functions in mentalizing dual-task conditions of theory of mind
Anita Marsden for their permission and provision of the Strange and executive function. Cognition, 107, 663–672.
Stories test as well as Simon Baron-Cohen and the Autism Research Carlson, S. M., Mandell, D. J., & Williams, L. (2004a). Executive
Centre for allowing the use of the Reading the Mind in the Eyes and function and theory of mind: Stability and prediction from ages 2
the Faux Pas tests. Finally, we thank LaShonda Abner, Cami Godsey, to 3. Developmental Psychology, 40, 1105–1122.
Amy Kasian, and Gemma Patel for their help in the data collection Carlson, S. M., Moses, L. J., & Breton, C. (2002). How specific is the
process. relation between executive function and theory of mind?
Contributions of inhibitory control and working memory. Infant
and Child Development, 11, 73–92.
Carlson, S. M., Moses, L. J., & Claxton, L. J. (2004b). Individual
References differences in executive functioning and theory of mind: An
investigation of inhibitory control and planning ability. Journal
Ahmed, F., Miller, L. S., & Abner, L. (2007). Assessment of theory of of Experimental Child Psychology, 87, 299–319.
mind: Variance among measures. Presented at the thirty-fifth Cole, K., & Mitchell, P. (2000). Siblings in the development of
annual meeting of the international neuropsychology society, executive control and a theory of mind. British Journal of
Portland, OR. Abstract: (published online by Cambridge Univer- Developmental Psychology, 18, 279–295.
sity Press, 3/8/2007), Journal of the International Neuropsycho- Cuerva, A. G., Sabe, L., Kuzis, G., Tiberti, C., Dorrego, F., &
logical Society, 13, S1, 30. doi:10.1017/S1355617707079970. Starkstein, S. E. (2001). Theory of mind and pragmatic abilities
Anderson, P. (2002). Assessment and development of executive in dementia. Neuropsychiatry, Neuropsychology, & Behavioral
function (EF) during childhood. Child Neuropsychology, 8, Neurology, 14, 153–158.
71–82. Delis, D. C., Kaplan, E., & Kramer, J. H. (2001). Delis Kaplan
Anderson, V., Levin, H. S., & Jocobs, R. (2002). Executive functions executive function system examiner’s manual. San Antonio, TX:
after frontal lobe injury: A developmental perspective. In D. The Psychological Corporation.
T. Stuss & R. T. Knight (Eds.), Principles of frontal lobe Fahie, C. M., & Symons, D. K. (2003). Executive functioning and
function (pp. 1–7). theory of mind in children clinically referred for attention and
Apperly, I. A., Samson, D., & Humphreys, G. W. (2009). Studies of behavior problems. Applied Developmental Psychology, 24,
adults can inform accounts of theory of mind development. 51–73.
Developmental Psychology, 45, 190–201. Fine, C., Lumsden, J., & Blair, R. J. R. (2001). Dissociation between
Bach, L. J., Happé, F., Fleminger, S., & Powell, J. (2005). Theory of ‘theory of mind’ and executive functions in a patient with early
mind: Independence of executive function and the role of the left amygdala damage. Brain, 124, 287–298.

123
678 J Autism Dev Disord (2011) 41:667–678

First, M. B., Spitzer, R. L., Gibbon, M., & Williams, J. B. W. (1996). Merten, T., Green, P., Henry, M., Blaskewitz, N., & Brockhaus, R.
Structured clinical interview for Axis I DSM-IV disorders— (2005). Analog validation of german-language symptom validity
Patient Edition (with psychotic screening) SCID-I/P (W/psy- tests and the influence of coaching. Archives of Clinical
chotic screen) (version 2). New York, New York: Biometrics Neuropsychology, 20, 1–8.
Research Department, New York State Psychiatric Institute. Nelson, C. A. (2001). The development and neural bases of face
Fisher, N., & Happé, F. (2005). A training study of theory of mind recognition. Infant and Child Development, 10, 3–18.
and executive function in children with autistic spectrum Ozonoff, S., Pennington, B. F., & Rogers, S. J. (1991). Executive
disorders. Journal of Autism and Developmental Disorders, 35, function deficits in high-functioning autistic individuals: Rela-
757–771. tionship to theory of mind. Journal of Child Psychology and
Gordon, A. C. L., & Olson, D. R. (1998). The relation between Psychiatry, 32, 1081–1105.
acquisition of a theory of mind and the capacity to hold in mind. Perner, J., & Lang, B. (1999). Development of theory of mind and
Journal of Experimental Child Psychology, 68, 70–83. executive control. Trends in Cognitive Sciences, 3, 337–344.
Gottlieb, D. (2005). The strange stories test: A replication of children Premack, D., & Woodruff, G. (1978). Does the chimpanzee have a
and adolescents with asperger syndrome. European Child theory of mind? The Behavioral and Brain Sciences, 4, 515–526.
Adolescent Psychiatry, 14, 73–82. Richman, J., Green, P., Gervais, R., Flaro, L., Merten, T., Brockhaus,
Gregory, C., Lough, S., Stone, V., Erzinclioglu, S., Martin, L., Baron- T., et al. (2006). Objective tests of symptom exaggeration in
Cohen, S., et al. (2002). Theory of mind in patients with frontal independent medical examinations. Journal of Occupational
variant frontotemporal dementia and alzheimer’s disease: The- Environmental Medicine, 48, 303–311.
oretical and practical implications. Brain, 125, 752–764. Royall, D. R., Lauterbach, E. C., Cummings, J. L., Reeve, A.,
Happé, F. G. (1994). An advanced test of theory of mind: Rummans, T. A., Kaufer, D. I., et al. (2002). Executive control
Understanding of story characters’ thoughts and feelings by function: A review of its promise and challenges for clinical
able autistic, mentally handicapped, and normal children and research. Journal of Neuropsychiatry and Clinical Neurosci-
adults. Journal of Autism and Developmental Disorders, 24, ences, 14, 377–405.
129–154. Sabbagh, M. A., Xu, F., Carlson, S. M., Moses, L. J., & Lee, K.
Happé, F. G. E., Winner, E., & Brownell, H. (1998). The getting of (2006). The development of executive functioning and theory of
wisdom: Theory of mind in old age. Developmental Psychology, mind: A comparison of chinese and U.S. preschoolers. Psycho-
34, 358–362. logical Science, 17, 74–81.
Homack, S., Lee, D., & Riccio, C. A. (2005). Test review: Delis- Saltzman, J., Strauss, E., Hunter, M., & Archibald, S. (2000). Theory
Kaplan Executive Function System. Journal of Clinical and of mind and executive functions in normal aging and parkinson’s
Experimental Neuropsychology, 27, 599–609. disease. Journal of the International Neuropsychological Soci-
Hughes, C. (1998). Finding your marbles: Does preschoolers’ ety, 6, 781–788.
strategic behavior predict later understanding of mind? Devel- Sarfati, Y., Hardy-Baylé, M. C., Besche, C., & Widlöcher, D. (1997).
opmental Psychology, 34, 1326–1339. Attribution of intentions to others in people with schizophrenia:
Hughes, C., Dunn, J., & White, A. (1998). Trick or treat?: Uneven A non-verbal exploration with comic strips. Schizophrenia
understanding of mind and emotion and executive dysfunction in Research, 25, 199–209.
‘‘hard to manage’’ preschoolers. Journal of Child Psychology Shaw, P., Lawrence, E. J., Radbourne, C., Bramham, J., Polkey, C. E.,
and Psychiatry, 39, 981–994. & David, A. S. (2004). The impact of early and late damage to
Hughes, C., & Graham, A. (2002). Measuring executive functions in the human amygdale on ‘theory of mind’ reasoning. Brain, 127,
childhood: Problems and solutions? Child and Adolescent 1535–1548.
Mental Health, 7, 131–142. Stone, V. E., Baron-Cohen, S., & Knight, R. T. (1998). Frontal lobe
Jolliffe, T., & Baron-Cohen, S. (1999). The strange stories test: A contributions to theory of mind. Journal of Cognitive Neurosci-
replication with high-functioning adults with autism or asperger ence, 10, 640–656.
syndrome. Journal of Autism and Developmental Disorders, 29, Stuss, D. T., & Knight, R. T. (2002). Introduction. In D. T. Stuss & R.
395–404. T. Knight (Eds.), Principles of frontal lobe function (pp. 1–7).
Joseph, R. M., & Tager-Flusberg, H. (2004). The relationship of New York: Oxford University Press.
theory of mind and executive functions to symptom type and Sullivan, S., & Ruffman, T. (2004). Social understanding: How does
severity in children with autism. Development and Psychopa- it fare with advancing years? British Journal of Psychology, 95,
thology, 16, 137–155. 1–18.
Kaland, N., Moller-Nielsen, A., Callesen, K., Mortensen, E. L., The Psychological Corporation. (1996). Beck depression inventory-
Gottlieb, D., & Smith, L. (2002). A new ‘advanced’ test of second edition manual. San Antonio, TX: Harcourt Brace &
theory of mind: Evidence from children and adolescents with Company.
asperger syndrome. Journal of Child Psychology and Psychiatry, The Psychological Corporation. (2001). Wechsler test of adult
43, 517–528. reading. San Antonio, TX: Harcourt Brace & Company.
Kerr, N., Dunbar, R. I. M., & Bentall, R. P. (2003). Theory of mind Wimmer, H., & Perner, J. (1983). Beliefs about beliefs: Represen-
deficits in bipolar affective disorder. Journal of Affective tation and constraining function of wrong beliefs in young
Disorders, 73, 253–259. children’s understanding of deception. Cognition, 13, 41–68.
Lezak, M. D., Howieson, D. B., & Loring, D. W. (2004). Neuropsy- Yirmiya, N., Erel, O., Shaked, M., & Solomonica-Levi, D. (1998).
chological assessment (4th ed.). New York, NY: Oxford Meta-analyses comparing theory of mind abilities of individuals
University Press. with autism, individuals with mental retardation, and normally
Maylor, E., Moulson, J. M., Muncer, A., & Taylor, L. A. (2002). Does developing individuals. Psychological Bulletin, 124, 283–307.
performance of theory of mind tasks decline in old age? British Zelazo, P. D., & Frye, D. (1998). Cognitive complexity and control ii.
Journal of Psychology, 93, 465–485. The development of executive function in childhood. American
McKinnon, M. C., & Moscovitch, M. (2007). Domain-general Psychological Society, 7, 121–125.
contributions to social reasoning: Theory of mind and deontic
reasoning re-explored. Cognition, 102, 179–218.

123

You might also like