Professional Documents
Culture Documents
Affective and Cognitive Theory of Mind Abilities in
Affective and Cognitive Theory of Mind Abilities in
PII: S0165-1781(16)30783-1
DOI: http://dx.doi.org/10.1016/j.psychres.2017.06.016
Reference: PSY10563
To appear in: Psychiatry Research
Received date: 4 May 2016
Revised date: 1 June 2017
Accepted date: 4 June 2017
Cite this article as: Sarah-Ann Tay, Carol A. Hulbert, Henry J. Jackson and
Andrew M. Chanen, Affective and cognitive theory of mind abilities in youth
with borderline personality disorder or major depressive disorder, Psychiatry
Research, http://dx.doi.org/10.1016/j.psychres.2017.06.016
This is a PDF file of an unedited manuscript that has been accepted for
publication. As a service to our customers we are providing this early version of
the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting galley proof before it is published in its final citable form.
Please note that during the production process errors may be discovered which
could affect the content, and all legal disclaimers that apply to the journal pertain.
Affective and cognitive theory of mind abilities in youth with borderline personality
disorder or major depressive disorder
Sarah-Ann Taya, Carol A Hulberta*, Henry J Jacksona, Andrew M Chanenb,c,d
a
Melbourne School of Psychological Sciences, The University of Melbourne, Melbourne,
Australia
b
Orygen, the National Centre of Excellence in Youth Mental Health, Melbourne, Australia
c
Centre for Youth Mental Health, The University of Melbourne, Melbourne, Australia
d
Orygen Youth Health, Northwestern Mental Health, Melbourne, Australia
*
Corresponding Author: Melbourne School of Psychological Sciences, The University of
Melbourne, Parkville 3010, VIC, Australia. cah@unimelb.edu.au
Abstract
Background:
Theory of mind (ToM) is an important social cognitive ability that has been investigated in
BPD, with inconsistent findings indicating impaired, comparable, and enhanced ToM in
BPD. This study aimed to clarify and extend previous findings by investigating affective and
cognitive ToM abilities in youth early in the course of BPD, by including a clinical
Methods:
Female participants aged 15 – 24 years diagnosed with BPD (n = 41) or MDD (n = 37)
completed the Reading the Mind in the Eyes Test (RMET) and Happé’s Cartoon Task,
Results:
The BPD group performed significantly worse than the MDD group on the affective ToM
task, even after controlling for age, intelligence and depressive symptoms. Results for
Conclusions:
youth with MDD early in the course of BPD suggest a developmental failure of
sociocognitive abilities needed for mentalising and which are theorised as giving rise to core
features of BPD. Future research should employ more naturalistic paradigms to study social
cognition and should assess individuals even earlier in the course of BPD.
1. Introduction
Borderline personality disorder (BPD) is a severe mental disorder that is characterised by a
relationships and self-image (Leichsenring et al., 2011). The interpersonal features of BPD
have received increasing emphasis in recent years (Herpertz, 2013), and dysfunction in social
cognition has been proposed to account for the severe interpersonal difficulties experienced
understand and to respond effectively to others’ thoughts and feelings (Herpertz, 2013). ToM
refers to an individual’s ability to attribute mental states (e.g., thoughts, feelings, beliefs,
intentions) to accurately predict and explain another person’s behaviour (Premack and
Woodruff, 1978). Mentalising is a closely related concept (Choi-Kain and Gunderson, 2008),
which is defined as the mental process by which an individual perceives and interprets the
actions of others and themselves in terms of intentional mental states (e.g., thoughts, feelings,
wishes, desires, beliefs and reasons; Allen, Fonagy, and Bateman, 2008; Fonagy and
Bateman, 2008). Despite several key distinctions between the two constructs (see Choi-Kain
and Gunderson, 2008), the terms ‘mentalising’ and ‘ToM’ are often used interchangeably
ToM can be separated into cognitive and affective components (Shamay-Tsoory et al., 2007).
Affective ToM refers to the capacity to understand others’ emotional states, and cognitive
ToM refers to the ability to reason about other people’s beliefs (Shamay-Tsoory et al., 2007).
False belief tasks are regarded as a prototypical assessment of cognitive ToM, as they require
individuals to infer that another person’s behaviour will be directed by their beliefs, even if
they are based on misinformation about the environment (Poletti et al., 2012; Wimmer and
Perner, 1983). On the other hand, the Reading the Mind in the Eyes Task (RMET; Baron-
Cohen et al., 2001) is considered a prototypical measure of affective ToM (Poletti et al.,
2012). The RMET involves making subtle discriminations of others’ complex affective
Two studies of cognitive ToM, each using different methodologies, yielded inconsistent
results. Relative to healthy controls, BPD patients were found to display comparable
performance (Ghiassi, Dimaggio, and Brüne, 2010) on a cartoon task (Brüne, 2005), and
enhanced performance (Arntz et al., 2009) on the advanced ToM test (Happé et al., 1996).
Findings using the RMET to assess affective ToM in BPD, also, have been inconsistent. Two
clinical studies found that adults with BPD performed significantly better than healthy
controls on the RMET (Fertuck et al., 2009; Frick et al., 2012), as did a study of a non-
clinical sample of university students with features of BPD (Scott et al., 2011). However, two
clinical studies (Preiler et al., 2010; Schilling et al., 2012) did not find differences in
performance on the RMET between adults with BPD and healthy controls. Taken together,
RMET studies seem to indicate intact or enhanced affective ToM ability in BPD, with
discrepancies reported also in terms of whether BPD individuals show sensitivity toward a
particular affective valence, and the potential influence of depression on RMET performance
inconsistent results (Harari et al., 2010; Preiler et al., 2010). The first used story vignettes in
the Understanding Faux Pas task (Baron-Cohen et al., 1997) and found evidence of impaired
cognitive ToM in BPD patients, compared with healthy controls (Harari et al., 2010).
However, there were no significant differences between the groups on affective ToM,
suggesting dissociation between cognitive and affective ToM in BPD. The second study
(Preiler et al., 2010) used the Movie for the Assessment of Social Cognition (MASC;
Dziobek et al., 2006), a computerised, video-based test, designed to simulate the social
reduced ToM. The authors found recognition impairments in all three aspects of feelings,
thoughts, and intentions in patients with BPD, indicating impairment in both cognitive and
affective ToM. Two studies used the MASC (Dziobek et al., 2006) in adolescents with BPD
(Sharp et al., 2011, 2013) and found a strong association between an “excessive ToM” or
Not all studies of ToM in BPD have distinguished between affective and cognitive
BPD patients displayed superior attribution of mental states to interaction partners, while
In summary, studies of ToM capacities in BPD have yielded seemingly inconsistent findings.
Some studies have documented overall impairments in ToM (Preiβler et al., 2010), and more
specifically in cognitive ToM (Harari et al., 2010). Other studies argue against mentalising
impairment in BPD, finding better ToM performance (Arntz et al., 2009; Franzen et al.,
2011), excessive ToM (Sharp et al., 2011, 2013), or comparable mentalising ability (Ghiassi
et al., 2010). RMET studies in particular, have indicated enhanced (Fertuck et al., 2009; Frick
et al., 2012; Scott et al., 2011) or intact (Preiβler et al., 2010; Schilling et al., 2012) mental
state decoding or affective ToM ability in BPD. In short, gross deficiencies in ToM do not
appear to be evident in BPD, relative to healthy controls (Richman and Unoka, 2015). Rather,
there appears to be a trend towards equivalent or enhanced ToM in BPD patients in relation
to affective ToM and mental state decoding (Fertuck et al., 2009; Harari et al., 2010) but
compromised cognitive aspects of ToM (Harari et al., 2010; Preiβler et al., 2010).
Three possible explanations for these discrepant findings include measurement issues, failure
to account for the role of personality pathology and severity and/or specific co-occurring
psychopathologies, such as mood disorder and post-traumatic stress disorder, and failure to
Some studies have used more traditional ‘offline’ cartoon- or story-based assessment
methods (e.g., Arntz et al., 2009; Ghiassi et al., 2010; Harari et al., 2010), whereas another
study used a simulated social interaction situation (Franzen et al., 2011). Even studies using
the MASC might not be directly comparable with one another, with different components of
the influence of emotional arousal and interpersonal context on social cognition in BPD
(Fonagy and Luyten, 2009), which might vary, depending upon the type of task utilised.
Social cognition research across various diagnostic groups indicates a complex array of
(e.g., Andreou et al., 2015; Lee et al., 2005; Semerai et al., 2014, 2015; Unoka et al, 2014;
Unoka et al., 2015; Vaskinn et al., 2015; Wang et al., 2008). For example, the results of a
recent meta-analysis indicated that relative to healthy controls, schizophrenia and major
depressive disorder (MDD), were associated with greater ToM impairment than BPD, with
differences identified also in the patterns of impairments found for each group (Richman and
Unoka, 2015). The authors noted, also, that participants with both BPD and MDD were more
accurate in RMET assessments than those with BPD or MDD alone. Given that MDD
commonly co-occurs with BPD, including in youth (Chanen et al., 2007), these findings have
relevance for the investigation of the contributions of mood and ToM to the enduring social
dysfunction found in BPD. For, example, adult MDD groups demonstrated poorer
performance in detecting positive valence on the RMET while those with BPD fared worse
with neutral stimuli (Richman and Unoka, 2015). Additionally, participants with BPD and
co-occurring cluster B or C personality disorders were more impaired than those with BPD
only. These findings, together with findings from Semerai and colleagues (2015) showing
(compared to healthy controls) and associations with personality disorder type and severity,
Lastly, in much of this research there has been a lack of clarity about the influence of age,
developmental stage, extent of prior treatment and other characteristics among BPD
participants. For example, studies of adults have used samples with full-syndrome BPD,
whereas those of adolescents have used inpatients with a mixture of sub-syndromal and full
syndrome BPD (Scott et al., 2011; Sharp et al., 2011; 2013). Moreover, the adolescent groups
employed by Sharp et al. were from a specialised inpatient unit where patients typically have
lengthy prior treatment histories, had high comorbidity with MDD, and were of above
average intelligence. Both the adult and adolescent samples in these studies might be
polypharmacy, other treatments, and cumulative adverse events and mental state disorders.
Also, prior research has shown that being older (Astington and Jenkins, 1995), and having
higher intelligence are associated with better ToM performance (Astington and Jenkins,
1999). Likewise, the co-occurrence of BPD and MDD in adults was associated with greater
accuracy on the RMET, compared to responses of those with BPD alone (Richman & Unoka.
2015).
The research context for the present study includes a focus on clinical research with the
potential to guide early intervention with youth with BPD, including those with acute
presentations including high levels of risk related to self -harm and suicide and mood
disorder. Hence, the primary aims of the present study were: (1) to investigate the specificity
presenting early in the course of BPD, relative to a clinical comparison group of community-
treated youth with MDD screened for personality disorder; (2) to adjust for potentially
confounding variables, such as age, intelligence, and severity of depression; and (3) to use
measures that would allow comparisons with previous studies in BPD and other conditions.
The RMET was utilised as a measure of affective ToM. Cognitive ToM was assessed using
Happé’s Cartoon Task (Happé et al, 1999), a measure that demands mental state reasoning.
Noting the very limited ToM research comparing youth with MDD and BPD, the very mixed
findings from BPD studies utilising healthy controls, and findings from studies of adults with
MDD showing poorer ToM overall than those with BPD (e.g., Richman & Unoka, 2015) it
was predicted that, compared with MDD participants, individuals with BPD would show
higher accuracy on both the RMET, and Happé’s Cartoon Task. The study also explored the
effects of comorbid post-traumatic stress disorder within the BPD group on RMET and
2. Method
2.1 Participants
Two groups (N = 78) of female participants between the ages of 15 and 25 years were
recruited from two government-funded youth mental health services in western metropolitan
Melbourne, Australia. Orygen Youth Health (OYH) is funded by the Victorian State
Government and headspace Western Melbourne is funded by the Federal Government. Only
females were recruited in order to avoid potential sex effects on social cognition (Baron-
Cohen, 2003).
The BPD group comprised 41 youth recruited from Helping Young People Early (HYPE;
Chanen et al., 2009), a specialised early intervention program for BPD at OYH. Participants
met Structured Clinical Interview for DSM-IV (SCID) Axis II disorders (SCID-II; First et al.,
1997b) criteria for BPD. The BPD module of the SCID-II was completed as part of the
routine HYPE entry assessment by treating clinicians, who are specifically trained in BPD
assessment, using a rigorous standard that has been published elsewhere (Lawrence, Allen,
and Chanen, 2010). These clinicians also administered the Structured Clinical Interview for
DSM-IV Axis I disorders, patient version (SCID-I/P; First et al., 1997a) as part of this
assessment. Thirty five BPD participants (84.5%) met criteria for a mood disorder, while 16
(39%) met criteria for PTSD (see Table 1). Exclusion criteria were a schizophrenia-spectrum
English comprehension, or a severe disturbance that might interfere with their ability to give
informed consent or adhere to the study protocol.
The comparison group comprised 37 youth who were recruited from the Youth Mood Clinic
at Orygen Youth Health and from headspace. Participants met SCID-I/P criteria for MDD in
the past 12 months. An initial telephone screen was conducted for BPD and antisocial
features, prior to the invitation to participate in the study, using the SCID-II Personality
Questionnaire (SCID-II-PQ; First et al., 1997b) and clinical diagnoses were extracted from
participants’ medical files. In addition to the exclusion criteria applied to the BPD group,
participants were excluded if they had 3 or more SCID-II BPD features, any antisocial PD
features, or if they met full diagnostic criteria for any other DSM-IV personality disorder.
2.2 Measures
2.21 Screening measures for control participants
The SCID-II PQ (First et al., 1997b) uses a Yes/No response format. For the present study,
the 15 BPD items and the 15 items corresponding to Criterion A for antisocial PD were used.
2.22 Depression
Depression severity was rated using the 20-item Centre for Epidemiologic Studies
Depression Scale – Revised (CESD-R; Eaton, Smith, Ybarra, Muntaner, and Tien, 2004). The
CESD-R has excellent internal consistency and good or excellent convergent and divergent
validity (Eaton et al., 2004; Van Dam and Earleywine, 2011). In the current study, internal
IQ was measured with the Wechsler Abbreviated Scale of Intelligence (WASI; Wechsler,
1999) two-subtest form. Vocabulary and Matrix Reasoning subtests were used to derive a T-
score for Verbal IQ and Performance IQ, respectively. The sum of T-scores on these two
subscales was used to obtain age-corrected Full Scale IQ. A Digit Span task was also
administered to assess participants’ level of attention. The WASI has good to excellent
Affective ToM
The RMET consists of 36 black and white photographs of the eye region of faces. For each
item, participants are presented with a pair of eyes and asked to choose from four complex
mental state descriptors (i.e., one correct word and three distracter words) the word that best
describes the eyes. The RMET complex mental state descriptors were selected on the basis
that they went beyond basic emotions ( fear, sadness) and allowed the identification of an
no time limit to this task and participants were provided with a glossary. Following Baron-
Cohen et al. (2001), each participant evaluated the same set of RMET stimuli in the same
sequence. Scores are calculated by adding up the total number of correct mental state
attributions for all 36 items. Additionally, to examine mental state decoding accuracy for
stimuli of particular emotional valence, the 36 RMET stimuli were classified into positive
(nine items), neutral (17 items), and negative (10 items) valence categories based on the
Cognitive ToM
Happé’s Cartoon Task (Happé et al., 1999) has been described as a more “traditional” ToM
task that probes cognitive ToM, as it demands “thinking about thinking” and deductive
reasoning skills (Russell et al., 2009, p. 182). The task consists of 12 single-frame cartoons.
Six cartoons make up the ToM or mental state inference condition (MS), requiring the
participant to utilise understanding of the character’s beliefs or intentions (e.g., false belief or
ignorance) to accurately explain the humour, and six physical state (PS) cartoons make up the
non-mental control condition, in which the humour only involves recognition of physical
anomalies or violation of a social norm. Four cartoons in the MS set and four in the PS set
displayed facial expressions. Six cartoons (three MS and three PS) included captions that
were read aloud to participants. Cartoons were presented in random order, one at a time, and
participants were asked to explain why each was funny. Answers were recorded and scored
according to a standard scoring scheme in which 3 is given for a full and explicit explanation,
2 for a partial/implicit explanation, and 1 for reference to relevant parts of the cartoon without
further explanation. Irrelevant, incorrect or ‘don’t know’ answers are scored 0. A higher score
2.3 Procedure
The study was approved by the Melbourne Health Research and Ethics Committee. After
complete description of the study procedures to participants, written informed consent was
obtained. Participants were first asked to complete the WASI and Digit Span. All other
measures were presented on a laptop computer, with the CESD-R completed first, followed
by the experimental measures. Permission was sought to audio record responses to Happé’s
Cartoon Task. The first author (ST) scored all the responses according to detailed scoring
criteria provided by Happé et al. (1999). Participants were reimbursed AU$30 and provided
3. Results
3.1 Demographic and clinical comparisons
Table 1 shows that there were no significant differences between the groups on age,
depressive symptoms, country of birth, living situation, and use of psychotropic medication.
Both groups faced similarly high levels of social disadvantage. Although there were no
differences between the groups on attention, the MDD group had a significantly higher full-
scale IQ score than the BPD group and a significantly higher proportion of students were in
the MDD group, whereas the BPD group had a higher proportion of individuals who were
employed.
Table 2 displays co-occurring diagnoses in both groups. In the BPD group, mood disorder
was the most commonly reported comorbidity, followed by an anxiety disorder. Almost 40%
of BPD participants also reported PTSD, compared with none in the MDD group; thus, post-
hoc analyses to explore the potential impact of PTSD diagnosis on ToM were conducted. A
majority of participants in each group were taking antidepressant medication (BPD: 73.2%,
MDD: 59.5%).
A mixed model ANCOVA was conducted, with valence as the within-participants factor,
group as the between-participants factor, and percent accuracy on the RMET as the
dependent measure. Participants’ age, intelligence, and depressive symptoms were retained as
Annaz et al., 2009), because the within-participants main effects of emotion are independent
was first conducted excluding the covariates in order to examine pure within-participants
main effects. There was a significant main effect of valence, F(2, 75) = 34.61, p < 0.001,
across groups, accuracy was higher for negative stimuli (M = 76.79%, SE = 1.97) than for
both neutral (M = 64.71%, SE = 1.83) and positive (M = 63.39%, SE = 2.15) stimuli, ps <
0.001, but accuracy for positive stimuli did not differ from accuracy for neutral stimuli, p =
0.463.
After controlling for the between-participants effects of age, F(1, 73) = 6.10, p = 0.016,
partial ŋ² = 0.08, intelligence, F(1, 73) = 26.00, p < 0.001, partial ŋ² = 0.26, and depressive
symptoms, F(1, 73) = 0.31, p = 0.583, the main effect of group on accuracy was significant,
F(1, 73) = 4.36, p = 0.04, partial ŋ² = 0.06, but the interaction term (Group x Valence) was
not (p >0.05). As shown in Figure 1, the profile of accuracy for different valences was the
same for BPD and MDD. However, a significant group effect stemmed from overall greater
A mixed model ANCOVA was conducted, with cartoon type as the within-participants factor,
group as the between-participants factor, accuracy on the cartoons as the dependent measure,
and age, intelligence, and depressive symptoms as covariates. When the analysis was
conducted excluding the covariates to examine pure within-participants main effects, there
was no main effect of cartoon type, F(1, 76) = 3.67, p > 0.05, indicating that, collapsed
After adjusting for the between-participants effects of age, F(1, 73) = 7.81, p = 0.007, partial
ŋ² = 0.10, intelligence, F(1, 73) = 26.70, p < 0.001, partial ŋ² = 0.27, and depressive
symptoms, F(1, 73) = 0.97, p = 0.329, partial ŋ² = 0.01, the main effect of group on accuracy
was not significant, F(1, 73) = 3.16, p > 0.05, partial ŋ² = 0.04, indicating that the BPD and
MDD groups did not differ in overall task accuracy when collapsed across cartoon type.
3.4 Effect of PTSD Diagnosis on ToM Findings
Given the high rate of PTSD among the BPD group (39%), post hoc independent samples t-
tests were conducted to explore the potential impact of PTSD diagnosis on ToM
performance. There were no significant differences between BPD participants with PTSD (n
= 16) and those without PTSD (n = 25) on affective ToM (RMET total accuracy: t(39) = -
0.55, p = 0.585), and cognitive ToM (MS cartoon accuracy: (t(39) = 0.29, p = 0.774).
Additionally, within the BPD group, no significant correlations were found between PTSD
4. Discussion
The major findings from this study of ToM in youth with first-presentation BPD are that the
BPD group demonstrated poorer performance for affective ToM only, compared with youth
with MDD, even after controlling for age, intelligence, and depressive symptoms.
The finding of reduced accuracy on the RMET in the BPD group is inconsistent with studies
comparing BPD samples with healthy controls that reported enhanced (Fertuck et al., 2009;
Frick et al., 2012) or comparable (Preiβler et al., 2010; Schilling et al., 2012) mental state
decoding, or a negative bias (Scott et al., 2011) in BPD. Similarly, the present finding
showing that responses of the BPD and MDD groups to MS cartoons were not significantly
different is in contrast to findings from Richman and Unoka (2015) indicating that overall,
adults with MDD less well than BPD participants on ToM tasks. In comparison to results
from studies using healthy control groups, the present findings are inconsistent with two
previous studies that found poorer performances on both cognitive and affective aspects of
ToM or empathy in adults with BPD (Dziobek et al., 2011; Preiβler et al., 2010). However,
the present findings are consistent with studies reporting intact (Ghiassi et al., 2010) or
superior cognitive ToM and mental state reasoning abilities in adults with BPD, compared to
healthy controls (Arntz et al., 2009; Franzen et al., 2011), and hypermentalising in
adolescents with BPD (Sharp et al., 2011, 2013). Also, the present study was unable to
replicate Harari et al.’s (2010) finding of dissociation between affective and cognitive ToM in
BPD.
This is the first study to examine ToM performance using the RMET and cartoons in a
sample of youth with recently diagnosed with either full-syndrome MDD or BPD with high
levels of co-occurring mental disorders (including mood disorder and PTSD), and minimal
sample differences in terms of age, severity of disorder and/or duration of illness, and level of
other psychopathology. Previous ToM studies have largely utilised adult samples, with few
exceptions (Scott et al., 2011; Sharp et al., 2011; 2013). Compared with adult participants in
previous RMET studies, the youth in our sample were almost 10 years younger. Further,
Scott et al.’s (2011) non-clinical sample of university students, while of almost identical age
to the present sample, had much lesser severity of BPD features. Although Sharp et al. (2011)
also used a sample of youth, this group comprised inpatients with a mixture of sub-syndromal
and full syndrome BPD, and high comorbidity with MDD. The potential confounding
influence of IQ and depression was not taken into account in that study. Youth in the current
study appear to be at the more severe end of the BPD severity spectrum in terms of number of
BPD features, intelligence, level of social disadvantage, and co-occuring mental disorder.
Some previous RMET investigations of adult individuals with BPD compared to healthy
controls, such as Fertuck et al.’s (2009) better performing sample, are also likely to have had
longer duration of illness. It is possible then that enhanced RMET in the latter sample might
the mental health system or to lower levels of co-occuring disorders. Hence, enhanced mental
state decoding might develop later in the course of the disorder and represent an adaptive
attempt to gain control over social environments compromised by severity of BPD features,
cumulative trauma, and recurrent mental state disorders (Chanen et al., 2008).
The present findings are consistent with some aspects of mentalisation theory and discrepant
with others. The demonstration of poorer performance on an affective ToM task in a sample
of youth is consistent with the developmental focus of mentalisation theory (Fonagy and
Luyten, 2009), which asserts that understanding others’ behaviour in terms of mental states is
associated with more developed social-cognitive processes, such as, response inhibition,
emotion regulation, the capacity to meta-monitor, and the capacity for abstract and
hypothetical thought (Nelson et al., 2005; Paus, 2005). As such, it is possible that a
The finding of reduced affective ToM in the RMET performance of the BPD participants
relative to the MDD group, however, is inconsistent with the proposition that social cognition
conscious, reflective mode of mentalising (Fonagy and Luyten, 2009; Sharp et al., 2013).
However, it is important to note that there are differing interpretations of what the RMET
claims to measure. While the RMET has been described as involving the unconscious, rapid,
and automatic decoding of mental states, without cognitive deduction of mental content
(Baron-Cohen et al., 2001), it has also been suggested that it taps explicit-controlled
mentalising, as it seems to call on the individual to read mental states from external cues, and
a decontextualised display of the eye region of the face (Sharp et al., 2013). It is possible that
the RMET might be more closely related to cognitive ToM, rather than affective ToM, as it
has been argued to place greater emphasis on the detection and attribution of beliefs and
4.2 Limitations
to the two clinical groups, as this may have enabled more informed interpretation of the
meaning of differences found in level and patterns of responses by the BPD and MDD
groups, and more direct comparisons with the majority of previous studies. However, we
argue that there are issues about what constitutes a healthy control group in the context of the
current study. Identifying putative healthy control (non-psychiatric) members who are
equivalent in age may mean be that they differ in terms of other characteristics, e.g.,
education level, occupation, and so on. They will differ also into how they are selected into
such a study. Attempting to match participants is always difficult and may lead to the
“matching fallacy” (Meehl, 1970; Snitz, MacDonald, and Carter, 2006; see Kremen et al.
1996 for an illustrative empirical example). Given the ubiquity of mood disorder in early as
well as later presentations of BPD and the value in clinical terms of differentiating the
contributions of the two disorders to TOM difficulties early in the course of BPD, we argue
that comparison of BPD and MDD using acutely unwell participants provides a compelling
test, as differences between healthy controls and BPD participants could be attributed to the
A second major limitation relates to the ecological validity of the ToM measures, which
involved passive observation and interpretation of others’ interactions through pictures (‘cold
cognition’). Although the present study demonstrated that deficits in ToM in youth with BPD
are not limited to contexts involving themes of potential emotional significance or underlying
schemas, it has been argued that disturbances in social cognition might become more
apparent when the individual is personally invested in an interaction, and when that
interaction is emotionally-charged (‘hot cognition’, Arntz and ten Haaf, 2012; Fonagy and
Luyten, 2009; Sharp et al., 2013; Andreou et al., 2015). Indeed, Sharp et al. (2013)
emphasised that tests like the MASC, which are more closely related to interpersonal
functioning rather than pure cognitive mentalising, might be more likely to trigger
Munholland, 2008), along with hyperactive emotional responsivity, giving rise to anomalous
mentalising in BPD patients . Available findings from studies using the MASC showing
evidence of hypermentalising provide some support for this assertion (e.g., Andreou et al.,
2015: Sharp et al., 2013; Vaskinn et al., 2015) .The ToM measures used in the present study
are relatively emotionally neutral compared with the MASC. Furthermore, they were not
designed to be able to detect hypermentalising; responses on the RMET and Cartoon Task
simply reflect either accurate or inaccurate ToM, and do not indicate whether incorrect
responses are due to over-interpretation. Therefore, it would be desirable for future studies to
assess social cognition using more naturalistic paradigms, in more interpersonal, BPD-
Another potential limitation relates to the lack of assessment of other potential confounding
factors, such as externalising problems and childhood maltreatment. Given the high
comorbidity between BPD and externalising problems, and an association between social-
cognitive deficits and externalising (Sharp, 2008), investigation of whether ToM impairments
are driven by externalising pathology more broadly is needed. Childhood maltreatment has
also been known to be a factor that potentially influences social perception (Preiler et al.,
2010; Scott et al., 2011). However, we note that the higher levels of PTSD in our BPD
Finally, the lack of inter-rater reliability on Happé’s Cartoon Task and the use of clinical
diagnoses for co-occurring syndromes in the MDD group also represent limitations. Other
sample-related limitations include only studying females. It remains unclear how present the
Despite these limitations, a significant strength of this study is its investigation of ToM
BPD, minimising duration of illness factors. Also, the use of a clinical control group of youth
with MDD drawn from the same public sector setting, along with measurement of depressive
BPD.We note that the majority of the extant research has investigated ToM in adults across a
range of diagnostic groups in comparison to healthy controls. For the most part this research
shows that, across age groups, mental and personality disorder diagnosis and severity of
pathology influence performance on social cognition tasks differentially (e.g., Andreou et al.,
2015; Lee et al., 2005; Semerai et al., 2014, 2015; Unoka et al, 2014; Wang et al 2008;
Vaskinn et al., 2015). We believe the present study investigating two clinical samples
comprised of at risk, symptomatic youth presenting early in the course of illness makes a
Importantly, the finding that, with depressive symptoms controlled for, youth early in the
course of BPD had less accurate affective ToM, underscores the importance of early
intervention. Specifically, it reinforces the notion that mentalising might be a valuable target
for treatment, which is consistent with existing treatment programs that focus on helping
In summary, the present study provides evidence very early in the course of BPD, of reduced
performance on measures of both affective and cognitive ToM, relative to youth with MDD.
These findings are consistent with a developmental failure of socio-cognitive abilities needed
for mentalisation, which might give rise to core features of BPD (Fonagy and Luyten, 2009;
cognition in BPD using more naturalistic paradigms that also allow investigation of the
Acknowledgements
We would like to thank Drs Jennifer Betts and Christopher Davey for their support in the
recruitment of participants and Drs Jean Cheng and Kate Caldwell for their assistance with
data collection. We are indebted to Professor Simon Baron-Cohen and Professor Francesca
Happé for providing the Reading the Mind in the Eyes Test and the Cartoon Task
respectively.
Declaration of Interest
The authors declare that there is no conflict of interest. This research received no specific
grant from any funding agency in the public, commercial, or not-for-profit sectors.
References
Allen, J., Fonagy, P., Bateman, A., 2008. Mentalizing in Clinical Practice. American
Annaz, D., Karmiloff-Smith, A., Johnson, M.H., Thomas, M.S.C., 2009. A cross-syndrome
study of the development of holistic face recognition in children with autism, Down
Andreou, C., Kelm, L., Bierbrodt, J., Braum, V., Lipp, M., Yassari, AH, Moritz, S. 2015.
Arntz, A., Bernstein, D., Oorschot, M., Schobre, P., 2009. Theory of mind in Borderline and
Cluster-C personality disorder. Journal of Nervous and Mental Disease 197, 801–
807.
Arntz, A., ten Haaf, J., 2012. Social cognition in borderline personality disorder: Evidence for
Astington, J.W., Jenkins, J.M., 1995. Theory of mind development and social understanding.
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
Baron-Cohen, S., Wheelwright, S., Hill, J., Raste, Y., Plumb, I., 2001. The “Reading the
Mind in the Eyes” Test Revised Version: A study with normal adults, and adults with
Baron-Cohen, S., 2003. The Essential Difference: The Truth About the Male and Female
relationships: A construct revisited, in: Cassidy, J., Shaver, P.R. (Eds.) Handbook of
Attachment: Theory, Research and Clinical Applications. Guilford Press, New York,
pp. 89 – 111.
Brüne, M., 2005. Emotion recognition, ‘‘theory of mind’’ and social behaviour in
Chanen, A.M., Jovev, M., Jackson, H.J., 2007. Adaptive functioning and psychiatric
Chanen, A.M., Jovev, M., McCutcheon, L.K., Jackson, H.J., McGorry, P.D., 2008.
Borderline personality disorder in young people and the prospects for prevention and
2009. The HYPE Clinic: An early intervention service for borderline personality
Chanen, A.M., McCutcheon, L., 2013. Prevention and early intervention for borderline
personality disorder: current status and recent evidence. The British Journal of
Chanen, A.M., Thompson, K., 2014. Preventive strategies for borderline personality disorder
Dziobek, I., Fleck, S., Kalbe, E., Rogers, K., Hassenstab, J., Brand, M., Kessler, J., Woike,
J.K., Wolf, O.T., Convit, A. 2006. Introducing MASC: A movie for the assessment of
social cognition. Journal of Autism and Developmental Disorders 36(5), 623 – 636.
Dziobek, I., Preiler, S., Grozdanovic, Z., Heuser, I., Heekeren, H.R., Roepke, S., 2011.
Eaton, W.W., Smith, C., Ybarra, M., Muntaner, C., Tien, A., 2004. Center for Epidemiologic
Studies Depression Scale: Review and revision (CESD and CESD-R), in: Maruish,
M.E. (Ed.) The Use of Psychological Testing for Treatment Planning and Outcomes
Assessment (3rd ed.). Lawrence Erlbaum, Mahwah, NJ, pp. 363 – 377.
Fertuck, E.A., Jekal, A., Song, I., Wyman, B., Morris, M.C., Wilson, S.T., Brodsky, B.S.,
Stanley, B., 2009. Enhanced 'Reading the Mind in the Eyes' in borderline personality
First, M.B., Spitzer, R.L., Gibbon, M., Williams, J.B., 1997b. Structured Clinical Interview
Washington DC.
Fonagy, P., Bateman, A., 2008. The development of borderline personality disorder – A
Franzen, N., Hagenhoff, M., Baer, N., Schmidt, A., Mier, D., Sammer, G., Gallhofer, B.,
Kirsch, P., Lis, S., 2011. Superior ‘theory of mind’ in borderline personality disorder:
Frick, C., Lang, S., Kotchoubey, B., Sieswerda, S., Dinu-Biringer, R., Berger, M., Veser, S.,
Ghiassi, V., Dimaggio, G., Brüne, M., 2010. Dysfunctions in understanding other minds in
Happé, F., Brownell, H., Winner, E., 1999. Acquired ‘theory of mind’ impairments following
Frackowiak, R., Frith, C., 1996. “Theory of mind” in the brain: Evidence from a PET
characters’ thoughts and feelings by able autistic, mentally handicapped, and normal
children and adults. Journal of Autism and Developmental Disorders 24(2), 129 –
154.
Harari, H., Shamay-Tsoory, S.G., Ravid, M., Levkovitz, Y., 2010. Double dissociation
Herpertz, S.C., 2013. The social-cognitive basis of personality disorders: Commentary on the
Jennings, T.C., Hulbert, C.A., Jackson, H.J., 2012. Social perspective coordination in youth
140.
Lawrence, K.A., Allen, J.S., Chanen, A.M., 2010. Impulsivity in borderline personality
Kremem, W.S., Seidman, L.J., Faraone, S.W., Pepple, J.R., Lyons, M.J., Tsuang, M.T., 1996.
Lee, L., Harkness, K.L., Sabbagh, M.A., Jacobson, J.A., 2005. Mental state decoding abilities
Leichsenring, F., Leibing, E., Kruse, J., New, A.S., Leweke, F., 2011. Borderline personality
research, in: Radner, M., Winkour, S. (Eds.), Minesota Studies in the Philosophy of
Science: Vol. IV. Analyses of the Theories and Methods of Physics and Psychology.
Nelson, E.E., Leibenluft, E., McClure, E., Pine, D.S., 2005. The social re-orientation of
Paus, T., 2005. Mapping brain maturation and cognitive development during adolescence.
Poletti, M., Enrici, I., Adenzato, M., 2012. Cognitive and affective theory of mind in
Preiler, S., Dziobek, I., Ritter, K., Heekeren, H.R., Roepke, S., 2010. Social cognition in
Premack, D.G., Woodruff, G., 1978. Does the chimpanzee have a theory of mind? Behavioral
Richman, M.J., Unoka, Z., 2015. Mental state decoding impairment in major depression and
borderline personality disorder meta analysis. The British Journal of Psychiatry 207,
483-489.
Russell, T.A., Schmidt, U., Doherty, L., Young, V., Tchanturia, K., 2009. Aspects of social
Ryle, A., 1997. Cognitive Analytic Therapy and Borderline Personality Disorder: The Model
Schilling, L., Wingenfeld, K., Löwe, B., Moritz, S., Terfehr, K., Köther, U., Spitzer, C., 2012.
personality disorder patients. Psychiatry and Clinical Neurosciences 66(4), 322 – 327.
Scott, L.N., Levy, K.N., Adams, R.B.J., Stevenson, M.T., 2011. Mental state decoding
Semerai, A., Colle, L., Pellecchia, G., Carcione, A., Conti, L., Fiore, D., Moroni, F., Nicelo,
G., Procacci, M., Pedone, M. 2015. Personality disorders and mindreading: Specific
Semerai, A., Colle, L., Pellecchia, G., Buccione, I., Carcione, A., Dimaggio, G., Procacci, M,
Shamay-Tsoory, S., Shur, S., Barcai-Goodman, L., Medlovich, S., Harari, H., Levkovitz, Y.,
Sharp, C., Fonagy, P., 2008. Social cognition and attachment-related disorders, in: Sharp, C.,
Sharp, C., Pane, H., Ha, C., Venta, A., Patel, A.B., Fonagy, P., 2011. Theory of mind and
Snitz, B.E., MacDonald, A.W., Carter, C.S. (2006). Cognitive deficits in unaffected first-
Van Dam, N.T., Earleywine, M., 2011. Validation of the Center for Epidemiologic Studies
Vaskinn, A., Antonsen, B.T., Fretland, R.A., Dziobek, I., Sundel, K., Wilberg, T., 2015.
Wang, Y., Wang, Y., Chen, S., Zhu, C., Wang, K., 2008. Theory of mind disability in major
Wimmer, H., Perner, J., 1983. Beliefs about beliefs: Representation and constraining function
103 – 128.
Figure 1: Mean percent task accuracy in BPD (n = 41) compared to MDD groups (n = 37)
for RMET. Error bars indicate standard error of the mean.
Notes: BPD = Borderline Personality Disorder; MDD = Major Depressive Disorder; PTSD =
Post-Traumatic Stress Disorder
Highlights
Youth with BPD or depression completed cognitive and affective ToM tasks
BPD group performed significantly worse than the MDD group on both ToM tasks
Findings were maintained even after controlling for age, intelligence and depressive
symptoms