Neuropsychology

© 2018 American Psychological Association 2018, Vol. 32, No. 6, 754 –763
0894-4105/18/$12.00 http://dx.doi.org/10.1037/neu0000464

Cognitive and Affective Associations With an Ecologically Valid Test of

Theory of Mind Across the Lifespan

Emilie Johansson Nolaker Kim Murray and Francesca Happé

Goldsmiths University of London King’s College London

Rebecca A. Charlton
Goldsmiths University of London
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
This document is copyrighted by the American Psychological Association or one of its allied publishers.

Objectives: Many studies have demonstrated that theory of mind (ToM) ability declines with increasing
age. Research has found that ToM-age associations are often mediated by other cognitive abilities
particularly executive function. However, older adults rarely complain about real-world ToM difficulties.
It has been suggested that older adults may perform better in real-world situations compared with
experimental settings. Method: We examined performance on the Strange Stories Film Task (SSFT)
which has been designed to assess ToM using naturalistic, video scenarios. Sixty adults aged between 17-
and 95-years-old completed the SSFT, inhibitory control (Stroop) and working memory (letter-number
sequencing) measures, the basic empathy scale (cognitive and affective empathy), and the broad autism
phenotype questionnaire. Results: ToM performance correlated significantly with age, whereas perfor-
mance on a control task did not. Partial correlations and stepwise regression analyses demonstrated that
performance on the three SSFT ToM measures was explained by a combination of executive function and
empathy measures, with age explaining none of the variance. Conclusions: Using a naturalistic test of
ToM, performance was shown to decline with age for ToM but not control scenarios. Across the lifespan,
the variance in ToM performance was explained by cognitive abilities and empathy but not age. Age
alone may not influence ToM ability, but may be associated with age-related changes in cognition and
social-cognition.

General Scientific Summary

Older age is associated with poorer ability to understand the mental states of others, although older
adults rarely complain about the quality of their social interactions. Using a video task of real-world
scenarios, we found that although poorer theory of mind ability was associated with older age,
performance was explained by other cognitive abilities. Age-related cognitive difficulties rather than
age alone may explain changes in understanding others in later life.

Keywords: empathy, executive function, lifespan, social cognition, theory of mind

Theory of mind (ToM) is the ability to attribute mental states— passed around 4 years and implicit tasks (e.g., anticipatory looking
for example to recognize that others may have beliefs or desires based on attributed belief) in the second year of life. More complex
that differ from one’s own (Flavell, 1999). Basic ToM develops in understanding of others’ mental states (e.g., recognizing the dif-
childhood, with explicit tests such as false belief tasks being ference between irony and lying) continues to develop through
childhood (Dumontheil, Apperly, & Blakemore, 2010). ToM is
thought to be stable in adulthood but it has seldom been examined
This article was published Online First May 24, 2018. across the adult lifespan, although a greater number of studies have
Emilie Johansson Nolaker, Department of Psychology, Goldsmiths Uni- examined ToM in later life (see Henry, Phillips, Ruffman, &
versity of London; Kim Murray and Francesca Happé, Social, Genetic and Bailey, 2013 for a review). ToM underlies successful social inter-
Developmental Psychiatry Research Centre, Institute of Psychiatry, Psy- actions and decline in this ability may impact social functioning
chology and Neuroscience, King’s College London; Rebecca A. Charlton, giving rise to social isolation and loneliness. Therefore, ToM
Department of Psychology, Goldsmiths University of London. difficulties may be particularly important in clinical populations
We thank the individuals who volunteered their time to contribute to this
and among older adults where social isolation has been shown to
research.
Correspondence concerning this article should be addressed to
have a significant impact on health and well-being (Adams, Sand-
Rebecca A. Charlton, Department of Psychology, Goldsmiths Univer- ers, & Auth, 2004; Hawkley, Thisted, Masi, & Cacioppo, 2010).
sity of London, New Cross, London, United Kingdom SE14 6NW. The aspects of ToM that are measured, often differ across the
E-mail: [email protected] lifespan but have generally focused on understanding social inter-
754
 THEORY OF MIND ACROSS THE ADULT LIFESPAN 755

actions and specifically the ability to attribute mental states. In pendent (Bernstein et al., 2011). There are notable differences in
childhood ToM is often examined using (first-order) false belief the ages of the groups used in these studies (mean age difference
tasks (the understanding that others may hold a different belief to between young-middle groups (Duval et al., 2011 ⫽ 29 years;
oneself; Baron-Cohen, Leslie, & Frith, 1985). More complex at- Bernstein et al., 2011 ⫽ 37 years; mean age difference between
tribution of false beliefs is measured using second-order (“Jane middle-old groups, Duval et al., 2011 ⫽ 18 years; Bernstein et al.,
thinks John thinks X”) false belief tasks, often used with adoles- 2011 ⫽ 11 years) and many analyses used age-group rather than
cents or adults (Happé, 1995). Other more advanced tests of age as a continuous variable. Age differences and arbitrary age
mentalizing have been developed for use with adult populations. groupings may have influenced the different patterns of results by
These measures, including tasks such as the Strange Stories emphasizing group differences without considering the continuous
(Happé, 1994) or Frith-Happe Triangles (Castelli, Happé, Frith, & nature of age.
Frith, 2000), require interpretation of a protagonist’s intention More consistent results have been observed examining ToM
regarding another’s mental state, for example, irony, persuasion, performance in later life. Many studies have demonstrated poorer
deception, white lies, double-bluff, and so forth. Task success is ToM abilities in older compared with younger adults even when
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

generally measured by accurate attribution of characters’ inten- accounting for intelligence or other cognitive abilities (Charlton,
This document is copyrighted by the American Psychological Association or one of its allied publishers.

tions. However, research has shown that some individuals with Barrick, Markus, & Morris, 2009; Rakoczy, Harder-Kasten, &
real-life ToM difficulties (e.g., individuals with autism spectrum Sturm, 2012). Furthermore, a meta-analysis including a range of
disorders), often “pass” ToM test questions, suggesting that some different ToM measures, supported poorer performance in later-
lab-based measures are not sensitive to subtle real-world difficul- life and a moderate effect size across all measures (Henry et al.,
ties (Scheeren, de Rosnay, Koot, & Begeer, 2013). The reason for 2013). It is worth noting that several studies have found equivalent
this may be related to the speed and complexity of real world performance between older and younger adults (Keightley, Wino-
social interactions, where individuals must both interpret and react cur, Burianova, Hongwanishkul, & Grady, 2006; Saltzman,
to events on a moment-by-moment basis (Scheeren et al., 2013). Strauss, Hunter, & Archibald, 2000). Only one article has demon-
The amount of mental state speech present in verbal responses has strated superior performance among older versus young adults
also been measured as an indirect metric of ToM ability (Happé, (Happé, Winner, & Brownell, 1998); the authors note, however,
1994). Studies have demonstrated reduced use of mental state that the young and old groups were not IQ-matched, and it seems
terms (despite correct interpretation of intention) in individuals likely that selection effects explain this result. Many studies have
with ToM difficulties (Happé, 1994). More recently ToM tasks demonstrated that age-ToM associations are partly or fully medi-
have also begun to examine the ability to generate appropriate ated by other cognitive abilities, with some variability in results
behavior based on mental state attribution, for example, when that may be due to the specific cognitive ability (Charlton et al.,
presented with an ambiguous scenario “What would you do/say in 2009; Maylor, Moulson, Muncer, & Taylor, 2002; Rakoczy et al.,
that situation?” It is hoped that such tasks may more accurately 2012). Working memory and inhibitory control have been de-
reflect real world behaviors by measuring both understanding of scribed as core deficits in ageing (Charlton, Barrick, Lawes,
the scenario and prosocial behaviors. Studies have identified Markus, & Morris, 2010; Hasher & Zacks, 1988), and previous
poorer prosocial responses in individuals reporting more autistic studies have suggested that these abilities may influence ToM
traits, suggesting that this may reflect an additional aspect of ToM performance (Charlton et al., 2009; Henry et al., 2013; Maylor et
(Jameel, Vyas, Bellesi, Roberts, & Channon, 2014; Murray et al., al., 2002; Rakoczy et al., 2012). Furthermore, the task demands of
2017). many ToM tests require working memory abilities (holding infor-
When ToM has been examined across the adult lifespan, objec- mation in mind, evaluating it and then generating a response), and
tive measures show mixed results. Duval, Piolino, Bejanin, Eu- inhibitory control (inhibiting responses based on own knowledge
stache, and Desgranges (2011) found no age effects from young to in order to answer test questions about a character’s false belief).
middle adulthood, but significantly poorer performance in older In a typical ageing population, seven (out of eight) executive
adults on an attribution of intention task and second-order false function tests correlated significantly with ToM performance, in-
belief tasks. They further found that age had a direct effect on cluding domains of working memory and inhibitory control (Char-
second-order false belief performance, whereas the association lton et al., 2009). Robust correlations between ToM and executive
between age and first-order false belief performance was mediated function abilities have also been noted both in age-related disease
by executive function. In contrast Bernstein, Thornton, and Som- such as stroke (Pluta, Gawron, Sobańska, Wójcik, & Łojek, 2017).
merville (2011), using a continuous performance false belief task Research suggests an important role for executive function abili-
found greater false belief bias in middle-aged and older adults ties in ToM performance across the adult lifespan, but it is not
compared to young adults, as well as a general reduction in ToM clear whether ToM performance is affected by age directly or by
performance from young to old age. In a subsequent regression age-related changes in other cognitive domains. Understanding the
analysis, executive function, working memory, episodic memory mechanism for age-related changes in ToM abilities will allow us
and processing speed all failed to explain any additional variance to intervene to reduce social problems and social isolation that has
in false belief performance beyond age effects. Although these a significant negative impact in older age (Bailey, Henry, & Von
studies used different measures and analyses, the contrast in results Hippel, 2008). To date few studies have directly examined the
is still striking, with one study demonstrating middle-aged adults influence of other sociocognitive abilities such as emotional em-
perform like young adults (Duval et al., 2011) and the other pathy on ToM performance in ageing.
demonstrating they perform like older adults (Bernstein et al., In recent years the distinction has been made between “cogni-
2011); in one study age-ToM associations are affected by cogni- tive” and “affective” ToM (Duval et al., 2011; Mitchell & Phillips,
tive abilities (Duval et al., 2011) and in the other they are inde- 2015), that is between attribution of mental states and affective
 756 JOHANSSON NOLAKER, MURRAY, HAPPÉ, AND CHARLTON

empathy for others’ feelings. Cognitive ToM has been described as multiple sources of information that may be both subtle and
the cognitive understanding of the beliefs and intentions of others, fleeting. In this study we utilize a recently developed video ToM
whereas affective ToM has been described as requiring empathic task (the Strange Stories Film Task [SSFT]) designed to assess
understanding of others’ emotions (Mitchell & Phillips, 2015). ToM in a naturalistic manner in adults (Murray et al., 2017). The
These definitions often overlap with descriptions of cognitive and SSFT was developed based on the Strange Stories Task (Happé,
affective empathy. Empathy is often described as an emotional 1994), as a series of video scenarios showing the same two
response relating to someone else’s affective state, with cognitive characters and requiring interpretation of speaker’s intention and
empathy being the representation of others’ mental states (i.e., generation of an appropriate response (see Method section for a
ToM) and affective empathy being the recognition of and response more detailed description; full details of development can be found
to others’ emotions (Blair, 2005). Although different aspects of in Murray et al., 2017). The SSFT includes assessment of several
empathy are often strongly associated with each other, they may different aspects of ToM; (a) correct identification of the intention
also dissociate particularly in clinical populations including of a character, (b) measurement of use of mental state speech, and
autism spectrum disorders (Sucksmith, Allison, Baron-Cohen, (c) correct generation of a suitable social interaction. Understand-
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Chakrabarti, & Hoekstra, 2013). For example, although studies ing others’ mental states, such as intentions is the core feature of
This document is copyrighted by the American Psychological Association or one of its allied publishers.

have identified difficulty with cognitive empathy/ToM in individ- ToM according to most definitions (Mitchell & Phillips, 2015),
uals with autism spectrum disorders, many have found no diffi- and mental state speech has been used as an index of a partici-
culties in affective empathy (Rogers, Dziobek, Hassenstab, Wolf, pant’s ability to think about others’ mental states (Happé, 1994).
& Convit, 2007). Furthermore, ToM and empathy have been Less widely used are questions requiring a response describing an
shown to dissociate in other neurodevelopmental groups (e.g., interaction, where a participant must generate their “own” social
conduct disorder with callous/unemotional traits vs. ASD; Jones, response to each scenario. A previous study has shown that indi-
Happé, Gilbert, Burnett, & Viding, 2010). The presence of neuro- viduals reporting high autistic traits responded less prosocially to
developmental traits (e.g., the broad autism phenotype [BAP]) is scenarios (Jameel et al., 2014), thus this measure may be a useful
significantly associated with emotional empathy abilities in rela- reflection of real world behavior. By using a test that mimics
tives of individuals with autism spectrum disorders and the general real-world social situations, measuring multiple aspects of ToM
population (Grove, Baillie, Allison, Baron-Cohen, & Hoekstra, we hope to better understand the subtle changes in ToM across the
2014; Lamport & Turner, 2014). However, few studies have ex- lifespan.
amined relationships between ToM, affective empathy and traits Given that “cognitive empathy” has been shown to mediate
associated with the BAP in ageing (Bailey et al., 2008; Mitchell & age-social functioning associations (Bailey et al., 2008) and the
Phillips, 2015). To our knowledge one study has examined the known negative consequences of social isolation in ageing (Blozik
relationship between ToM and cognitive versus affective empathy. et al., 2009), understanding these associations may inform well-
Bailey et al. (2008) demonstrated that older adults performed being in older age. In this study we examined ToM performance
significantly worse than young adults on an affective ToM task cross-sectionally across the adult lifespan using a naturalistic mea-
(Reading the Mind in the Eyes Task; Baron-Cohen, Wheelwright, sure. We further examined associations between ToM performance
Hill, Raste, & Plumb, 2001), as well as reporting poorer cognitive with age, executive function abilities (specifically working mem-
empathy on a self-report questionnaire. However, no age differ- ory and inhibitory control), cognitive and affective empathy. We
ences were observed on self-reported affective empathy. In other hypothesize that ToM performance will decline with increasing
studies of cognitive versus affective abilities in old age, findings age, but that this will be explained by age-related declines in
are mixed. Bottiroli et al. (2016) demonstrated poorer performance domain-general cognitive abilities. We hypothesize that ToM per-
for older adults (compared with young adults) on a cognitive but formance will be associated with self-rated cognitive but not
not affective ToM measure (a Faux Pas Task designed to include affective empathy.
both cognitive and affective measures). However other studies
have also identified age-effects on affective ToM measures (Duval
et al., 2011; Mahy et al., 2014). Overall, these results suggest there Method
may be a different pattern of age-related decline in cognitive
versus affective abilities. Participants
An important consideration when discussing ToM across the
neurotypical lifespan, is that any differences or difficulties in Sixty adults aged 17–95 (M ⫽ 44, SD ⫽ 18.5) were recruited to
performance are likely to be small in comparison with clinical participate in the study. One participant (aged over 60) was ex-
disorders associated with ToM difficulties (i.e., autism spectrum cluded from the analyses, having had a stroke that significantly
disorders or schizophrenia). Indeed, even among clinical groups, influenced his cognitive ability. On the Stroop task (see below) he
individuals may perform well on lab-based tests despite difficulties performed more than two standard deviations below the mean of
in the real world (Murray et al., 2017). In order to detect the subtle his age group, and was excluded from the analysis. Demographic
differences in performance that are likely to occur across the data on the remaining 59 participants are presented in Table 1
lifespan we require measures that are both sensitive and ecologi- where the sample has been divided into three age groups for
cally valid. It has been suggested that video-based tasks may descriptive purposes, although age is used as a continuous variable
overcome some of the difficulties with traditional lab-based tasks in all analysis. All participants provided written informed consent.
(Heavey, Phillips, Baron-Cohen, & Rutter, 2000; Murray et al., After the study, all participants were debriefed both verbally and
2017; Sullivan & Ruffman, 2004). The advantage of videos is that via a written debrief form. The study was approved by Goldsmiths
they allow the dynamic presentation of complex scenes with University of London Ethics Committee.
 THEORY OF MIND ACROSS THE ADULT LIFESPAN 757

Table 1
Means (and Standard Deviations) for Demographic Data Divided by Age-Group for Descriptive Purposes Only

17–34 years 35–59 years 60 –95 years

Demographic and cognitive measures (N ⫽ 20) (N ⫽ 20) (N ⫽ 19)

Age 21 (4.17) 44 (8.40) 70 (8.0)

Gender (women:men) 17:3 9:11 10:10
Educationa 3.35 (.58) 4.30 (.98) 2.84 (1.30)
Mental healthb 0 (0) .25 (.64) .37 (.68)
Medication frequencyc No medication use 2 people taking 1-2 8 people taking 1-2 medications; 1 person
medications taking three or more medications
BAPQd 99.29 (18.52) 95.50 (22.33) 88.05 (25.49)
MMSEe — — 28.53 (1.30) Range 26–30
Letter-number sequencing scaled scoref 10.45 (2.31) 11.40 (2.66) 10.95 (1.99)
Stroop difference in seconds (interference-color)g 45 (14) 67 (18) 109 (34)
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

a
Education level: 0 ⫽ Left school; 1 ⫽ Vocational qualification including apprenticeships; 2 ⫽ General Certificate of Secondary Education (GCSE, United
This document is copyrighted by the American Psychological Association or one of its allied publishers.

Kingdom exams at 16-years-old); 3 ⫽ Advanced-level (A-level, United Kingdom exams at 18-years-old); 4 ⫽ Undergraduate Degree; 5 ⫽ Postgraduate
Degree; 6 ⫽ PhD or higher. b Mental health problems: 0 ⫽ never; 1 ⫽ one occasion; 2 ⫽ two-three occasions; 3 ⫽ more than three occasions. c Current
medication: 0 ⫽ none; 1 ⫽ one/two; 2 ⫽ three or more. d BAPQ ⫽ Score on the Broad Autism Phenotype Questionnaire (Max: 216). e MMSE ⫽ Mini
Mental State Examination, only performed for adults aged over 60-years-old. f F ⫽ .825, p ⫽ .444. g F ⫽ 39.1, p ⬍ .001.

Younger participants (aged 17–34 years) were students from in the other character’s situation, what would you say next?), and
Goldsmiths University. Twenty students were from the Depart- a memory question to check for attention or gross memory diffi-
ment of Psychology and were assigned student research participa- culties (e.g., Where was the character?). Participants answered
tion credits. Middle-aged (n ⫽ 20; aged 35–59 years) participants verbally and their responses were transcribed by the researcher.
were recruited at random from a cafe in Southeast London. Par- Intention questions were scored between 0 and 2 depending on
ticipants were given a hot drink as a thank you for their partici- accuracy and appropriateness; these responses were also coded for
pation. Older participants (aged 60 –95 years) were recruited from presence of mental state and metacognitive speech (0 –2 points,
an elderly independent housing community in the London Bor- henceforth mental state speech). Interaction questions were scored
ough of Camden. The community offers housing opportunities for between 0 and 2 depending on appropriateness. The memory
older adults with a low income, where people live independently in question was scored 0 –1 for accuracy. The intention question was
an apartment within a block for older adults. Participants aged over designed to measure ToM, whereas the interaction and mental state
60-years-old were screened for risk for dementia using the Mini- speech responses were intended to measure social cognition more
Mental State Examination (Folstein, Folstein, & McHugh, 1975). broadly. All scoring followed guidelines in Murray et al. (2017)
All older participants scored above the cut-off of 24 ruling out and was performed by a single rater (EJN). To assess consistency
likely dementia (M ⫽ 28.53, SD ⫽ 1.3; Tombaugh & McIntyre, of scoring, interrater reliability was calculated with an experienced
1992). All participants were fluent in English, any non-native coder (RAC) and two other research assistants in the lab, using
speaker had lived in the United Kingdom for at least 10 years. two-way random model intraclass correlations (absolute values) in
None of the participants reported problems with hearing and all IBM SPSS Statistics, Version 22 (IBM Corp., 2013), all values
had corrected-to-normal or normal vision. Participants were tested were deemed good and were as follows: ToM-intention, r ⫽ .962;
in a quiet environment in their own homes, in a quiet room in the ToM-mental state speech, r ⫽ .871; ToM-interaction, r ⫽ .916;
café or in a laboratory at Goldsmiths University. All participants ToM-memory, r ⫽ .969; control-intention, r ⫽ .894; control-
were tested independently in the company of the experimenter.
mental state speech, r ⫽ .907; control-interaction, r ⫽ .724;
control-memory, no variance was observed between raters.
Materials The sensitivity of the SSFT has been reported previously in
Strange Stories Film Task. The Strange Stories Film task young adults and adults with ASD (Murray et al., 2017), but has
(SSFT; Murray et al., 2017) was used to assess ToM and social not been widely used or used across the lifespan. For reference the
cognition in a naturalistic manner. The SSFT was developed and reliability of the measure is reported here. For the ToM scenarios
tested on 40 adults (20 with autism spectrum disorders and 20 Cronbach’s alphas were as follows: intention (␣ ⫽ .57); mental
without). The SSFT comprised 15 short video clips acted by two state speech (␣ ⫽ .48); and interaction (␣ ⫽ .55). For memory
semiprofessional actors, using similar social scenarios as Happé’s questions there was very little variance in the data and Cronbach’s
Strange Stories (Happé, 1994). Twelve scenarios involved mental alphas were not deemed appropriate: For young and midlife adults
states such as telling a white lie, using persuasion, or double bluff most individuals scored at ceiling (maximum score ⫽ 12, achieved
(henceforth referred to as ToM scenarios), and three were non- by: young ⫽ 19/20; midlife ⫽ 18/20); for older adults results had
ToM scenarios (henceforth referred to as control scenarios). The only slightly more variance (eight individuals scored ⫽ 12; seven
actors spoke directly to the camera, to mimic real world conver- scored ⫽ 11; 3 scored ⫽ 10; one scored ⫽ 9). For the control
sations from a first-person perspective. After watching each video scenarios Cronbach’s alphas were robust for the interaction scores
clip participants were asked three questions examining intention (⫽ .77), but much weaker for mental state speech (␣ ⫽ ⫺.001) and
(i.e., Why did the character say that?), interaction (i.e., If you were intention (␣ ⫽ ⫺.030); note however, that these variables were not
 758 JOHANSSON NOLAKER, MURRAY, HAPPÉ, AND CHARLTON

intended to be meaningful for the control clips. Memory scores Results

were at ceiling, with all but three individuals (two midlife, one
older) achieving the maximum score. Scores from the current data
are reported in Table 2. Age Effects
Cognitive and empathy measures. Inhibitory control was For descriptive purposes, the means and standard deviations are
measured by the Stroop test (Golden, 1978); completion time on presented by age-group in Table 2. Mean scores for cognitive
the color-ink subscale minus completion time for the color-word measures are reported in Table 1, no age-group differences are
subscale was used in the analysis therefore lower scores indicate noted for the working memory (letter-number sequencing) age-
better performance. Working memory was measured by the letter- scaled scores but the non-age corrected difference between the
number sequencing task from the Wechsler’s Intelligence Test IV Stroop color and interference tasks demonstrate expected age
(Wechsler, 2008); the total raw score was used in the analysis with differences (slower performance with older age). Please note that
higher scores indicating better performance. Raw scores (rather age is used as a continuous variable in all analysis presented
than age-corrected scores) were used for both cognitive measures below.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

in order to allow examination of age-effects. The Basic Empathy

This document is copyrighted by the American Psychological Association or one of its allied publishers.

Scale (BES) was used to measure cognitive and affective empathy

(Jolliffe & Farrington, 2006), with questions such as “I’m not Gender Differences
usually aware of my friend’s feelings,” and “I tend to feel scared
No gender differences were noted in highest education level, but
when I’m with friends who are afraid,” respectively. The Broad
were apparent in age, see Table 3. For the SSFT, gender differ-
Autism Phenotype Questionnaire (BAPQ; Hurley, Losh, Parlier,
ences were observed on ToM intention and mental state speech,
Reznick, & Piven, 2007) was used to assess presence of autism
but no other measures. After correction for multiple comparisons,
traits. As traits are considered to vary across the population,
only ToM intention ability demonstrated a significant gender dif-
participants with high scores were not excluded. For reference, six
ference (see Table 3 for full details). Women scored more highly
individuals scored above the suggested BAPQ cut-off (⬎3.35 for
on affective empathy (women: M ⫽ 42.06, SD ⫽ 6.90; men: M ⫽
males and ⬎3.25 for females; Hurley et al., 2007), three men (aged
37.74, SD ⫽ 6.96; F ⫽ 5.46, p ⫽ .023); but not cognitive empathy
37, 65, 95) and three women (aged 19, 20, 46), although it is worth
(women: M ⫽ 38.06, SD ⫽ 3.45; men: M ⫽ 37.22, SD ⫽ 4.83;
noting that these cut-offs were developed in a sample of young and
F ⫽ .545, p ⫽ .464).
midlife adults.

Statistical Analysis Correlation Analyses

Statistical analysis was performed using IBM SPSS Statistics, Age. Significant negative correlations were observed between
Version 22 (IBM Corp., 2013). The relationship between age and age and each aspect of ToM (see Figure 1 and Table 4), indicating
ToM performance was investigated using Pearson’s product mo- declining performance as age increases. All associations remained
ment correlations. The associations between ToM performance, significant after controlling for multiple comparison correction.
age, executive function abilities, cognitive and affective empathy For control SSFT scores, age correlated significantly with mental
were examined using Pearson’s product moment correlations. state speech and interaction scores, but not with intention or
Stepwise linear regressions were used to examine which variables memory scores; results did not survive multiple comparison cor-
best explained the variance in each aspect of ToM ability measured rection.
by the SSFT. To assure that gender did not influence results, Age correlated significantly with performance on both inhibi-
gender differences on SSFT measures were calculated using inde- tory control (r ⫽ .729, p ⬍ .001) and working memory
pendent sample t test. Multiple comparison correction was per- (r ⫽ ⫺.373, p ⫽ .004). No significant correlations were observed
formed using the Holm-Bonferroni sequence correction. between age and either affective (r ⫽ ⫺.135, p ⫽ .310) or

Table 2
Means and Standard Deviations for the SSFT Scores, Illustrated in Three Age Groups

Control group means

(Mean age ⫽ 30.6,
17–34 years 35–59 years 60 –95 years Whole sample n ⫽ 20) from
SSFT scores (Young, n ⫽ 20) (Middle-aged, n ⫽ 20) (Older, n ⫽ 19) (n ⫽ 59) Murray et al., 2017

ToM
Intention (Max: 24) 19.70 (2.72) 18.80 (2.07) 17.16 (3.66) 18.58 (3.01) 18.80 (2.33)
Mental state talk (Max: 24) 12.75 (2.72) 10.95 (2.16) 10.00 (3.09) 11.25 (2.87) 13.75 (2.45)
Interaction (Max: 24) 17.35 (2.50) 17.55 (2.78) 14.47 (2.80) 16.49 (3.00) 16.95 (4.14)
Memory (Max: 12) 11.95 (.22) 11.90 (.31) 11.16 (.90) 11.68 (.66) 11.85 (.37)
Control
Intention (Max: 6) 4.15 (.81) 4.85 (.93) 4.53 (1.02) 4.51 (.95) 4.40 (.99)
Mental state talk (Max: 6) 1.20 (1.01) 1.15 (1.04) .74 (.65) 1.03 (.93) 1.25 (1.16)
Interaction (Max: 6) 5.05 (1.39) 5.35 (.99) 4.21 (1.27) 4.88 (1.30) 5.70 (.65)
Memory (Max: 3) 3.00 (0) 2.90 (.31) 2.95 (.23) 2.95 (.22) 3.00 (0)
 THEORY OF MIND ACROSS THE ADULT LIFESPAN 759

Table 3
Means and Standard Deviations by Gender and Group Differences

Demographic and SSFT

scores Males Females Group differences

Age 53.48 (SD ⫽ 17.7) 39.06 (SD ⫽ 21.1) t ⫽ ⫺2.72, p ⫽ .009ⴱⴱⴱ

Highest education level 3.52 (1.34) 3.50 (1.03) t ⫽ ⫺.070, p ⫽ .944
ToM SSFT scores
Intention 17.09 (SD ⫽ 2.9) 19.53 (SD ⫽ 2.7) t ⴝ 3.28, p ⴝ .002ⴱⴱⴱ
Mental state talk 10.17 (SD ⫽ 2.4) 11.94 (SD ⫽ 2.9) t ⫽ 2.41, p ⫽ .019ⴱ
Interaction 15.83 (SD ⫽ 3.4) 16.92 (SD ⫽ 2.7) t ⫽ 1.37, p ⫽ .175
Memory 11.48 (SD ⫽ .85) 11.81 (SD ⫽ .47) t ⫽ 1.70, p ⫽ .100
Control SSFT scores
Intention 4.65 (SD ⫽ .98) 4.42 (SD ⫽ .94) t ⫽ ⫺.914, p ⫽ .365
Mental state talk 1.0 (SD ⫽ .85) 1.06 (SD ⫽ .98) t ⫽ .222, p ⫽ .825
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Interaction 4.65 (SD ⫽ 1.4) 5.03 (SD ⫽ 1.2) t ⫽ 1.08, p ⫽ .283

This document is copyrighted by the American Psychological Association or one of its allied publishers.

Memory 2.96 (SD ⫽ .219) 2.94 (SD ⫽ .23) t ⫽ ⫺.202, p ⫽ .840

Note. Bold ⫽ remains significant after multiple comparison testing.
ⴱ
p ⬍ .05. ⴱⴱⴱ p ⬍ .001.

cognitive (r ⫽ ⫺.039, p ⫽ .770) empathy, or BAP traits Regression Analysis

(r ⫽ ⫺.176, p ⫽ .181).
BAP traits. Self-reported BAP traits did not correlate signif- Forward stepwise regression analyses were performed with each
icantly with any measure of ToM. However, partial correlations aspect of ToM performance (intention, mental state speech, inter-
between BAP traits and ToM performance controlling for age, action) in turn as the dependent variable. Included as independent
reveal a significant association between BAP traits and ToM variables were age, education, gender, working memory, inhibi-
Intention ability (see Table 4). Results did not survive multiple tory control, cognitive and affective empathy, and BAP traits. Full
comparison correction. results are presented in Table 6.
Cognitive function. All aspects of ToM performance (inten- For ToM intention, working memory, affective empathy, and
tion, mental state speech, interaction) correlated significantly with gender contributed to explaining 49.9% of the variance in perfor-
both inhibitory control and working memory measures, see Table mance. Age did not contribute to the model. For mental state
5. Results indicate better ToM performance was associated with speech, working memory, gender, and cognitive empathy contrib-
better cognitive abilities. If age was controlled for in a partial uted, explaining 41.7% of the variance in performance. Variance in
correlational analysis the association between ToM performance ToM interaction ability (37.9%) was explained by performance in
and working memory remained significant. inhibitory control and education level. Memory performance on
Empathy. ToM intention ability correlated significantly with the ToM SSFT task (37.4%) was explained by age and working
both cognitive and affective empathy (see Table 5). Mental state memory ability (see Table 5 for full details and statistics).
speech correlated significantly with cognitive empathy but not
Discussion
affective empathy. No significant association was observed be-
tween ToM interaction ability and either cognitive or affective The aim of this study was to examine ToM abilities across the
empathy scores. The pattern of results remained the same if a lifespan using a naturalistic video-based test, and to examine
partial correlation was performed controlling for age (see Table 5). associations with executive function, cognitive and affective em-
Also see Table 5 for details of multiple comparison correction. pathy. In keeping with previous studies in ageing (Bottiroli, Cav-
Note, the correlation between cognitive and affective empathy was allini, Ceccato, Vecchi, & Lecce, 2016; Henry et al., 2013), all
not significant (r ⫽ .215, p ⫽ .102), suggesting that these sub- three ToM measures derived from the SSFT (and the memory
scales measure different constructs. question) correlated negatively with age. Despite using a measure

Figure 1. Scatterplots showing correlations between age and SSFT intention, mental state speech, and
interaction scores. See the online article for the color version of this figure.
 760 JOHANSSON NOLAKER, MURRAY, HAPPÉ, AND CHARLTON

Table 4
Correlations Between SSFT Scores and Age and BAPQ Scores

Correlations with
Correlations with BAPQ controlling for
SSFT scores Correlations with age BAPQ age

ToM SSFT scores

Intention r ⴝ ⴚ.441, p < .001ⴱⴱⴱ r ⫽ ⫺.210, p ⫽ .111 r ⫽ ⫺.325, p ⫽ .013
Mental state talk r ⴝ ⴚ.452, p < .001ⴱⴱⴱ r ⫽ ⫺.117, p ⫽ .377 r ⫽ ⫺.224, p ⫽ .090
Interaction r ⴝ ⴚ.464, p < .001ⴱⴱⴱ r ⫽ ⫺.072, p ⫽ .588 r ⫽ ⫺.176, p ⫽ .186
Memory r ⴝ ⴚ.556, p < .001ⴱⴱⴱ r ⫽ ⫺.043, p ⫽ .748 r ⫽ ⫺.172, p ⫽ .196
Control SSFT scores
Intention r ⫽ .033, p ⫽ .802 r ⫽ ⫺.066, p ⫽ .619 r ⫽ ⫺.061, p ⫽ .648
Mental state talk r ⫽ ⫺.259, p ⫽ .048ⴱ r ⫽ ⫺.042, p ⫽ .751 r ⫽ ⫺.092, p ⫽ .490
Interaction r ⫽ ⫺.318, p ⫽ .014ⴱ r ⫽ .069, p ⫽ .603 r ⫽ .014, p ⫽ .917
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Memory r ⫽ ⫺.056, p ⫽ .674 r ⫽ ⫺.048, p ⫽ .717 r ⫽ ⫺.059, p ⫽ .659

This document is copyrighted by the American Psychological Association or one of its allied publishers.

Note. Bold ⫽ remains significant after multiple comparison testing.

ⴱ
p ⬍ .05. ⴱⴱⴱ p ⬍ .001.

designed to optimize ecologically validity, age-related declines in mance) but not “subjective” self-report measures (BES). In the
performance were still observed. In previous studies examining current study, on the objective measure, ToM intention and mental
ToM across the lifespan, analyses have largely explored group state speech did correlate significantly with the subjective self-
differences between young, middle-aged, and older adults, and report measure of cognitive empathy, but only modestly (r⬇.3)
have found age effects either before (young ⬎ midlife ⫽ old) or and no correlation was observed between self-report cognitive
after (young ⫽ midlife ⬎ old) middle-age; Bernstein et al., 2011; empathy and ToM interaction ability. Furthermore, subjectively
Duval et al., 2011). In contrast, this study considered age as a reported affective empathy significantly correlated with objec-
continuous variable and significant correlations were noted be- tively measured ToM Intention ability, but not the other two ToM
tween the three aspects of ToM measured and age (see Figure 1). measures (mental state speech or interaction). Similar results to
No correlations with age were observed for either cognitive or those found here were identified in a previous study, where no
significant correlations were noted between objective and subjec-
affective empathy, suggesting that self-reported empathy is not tive ToM measures (Duval et al., 2011). These results may suggest
associated with age. This finding is in keeping with a previous that perhaps cognitive empathy and ToM do not measure the same
study which examined cognitive and affective ToM, which found underlying construct, although they may share key components
no age-associations on self-reported cognitive or affective mea- (Mitchell & Phillips, 2015). Alternatively, results could reflect
sures (Duval et al., 2011). However, another study described older discrepancies between objective tasks performance versus subjec-
adults as self-reporting lower scores on cognitive empathy com- tive self-report measures (usually questionnaires), where individ-
uals may underestimate their difficulties, or perhaps objective
pared with young adults, but no age-group differences on self- lab-based measures are more strongly influenced by cognitive
report affective empathy (Bailey et al., 2008). demands than everyday life social skills, the latter reflected in
Some authors consider cognitive empathy and ToM to be syn- better self-reported abilities.
onymous (Mitchell & Phillips, 2015) but if they are then there are Interestingly, this study did not identify significant correlations
discrepancies in the age effects observed in this study, with age between the three aspects of ToM measured and BAP traits,
effects observed on “objective” task performance (SSFT perfor- although once age-effects were accounted for ToM intention cor-

Table 5
Correlations (and Age Corrected Partial Correlations) Between ToM SSFT Scores and Cognitive
and Empathy Measures

Cognitive and
empathy measures ToM intention ToM mental state speech ToM interaction
ⴱⴱⴱ ⴱⴱ
Executive control r ⴝ ⴚ.438, p ⴝ .001 r ⴝ ⴚ.389, p ⴝ .002 r ⴝ ⴚ.472, p < .001ⴱⴱⴱ
(r ⫽ ⫺.189, p ⫽ .156) (r ⫽ ⫺.098, p ⫽ .466) (r ⫽ ⫺.222, p ⫽ .094)

Working memory r ⴝ .495, p < .001ⴱⴱⴱ r ⴝ .538, p < .001ⴱⴱⴱ r ⴝ .392, p ⴝ .002ⴱⴱ
(r ⴝ .396, p ⴝ .002ⴱⴱ) (r ⴝ .446, p < .001ⴱⴱⴱ) (r ⫽ .266, p ⫽ .044ⴱ)

Cognitive empathy r ⫽ .312, p ⫽ .016 r ⫽ .288, p ⫽ .027ⴱ r ⫽ .137, p ⫽ .301

(r ⫽ .329, p ⫽ .012ⴱ) (r ⫽ .303, p ⫽ .021ⴱ) (r ⫽ .134, p ⫽ .315)

Affective empathy r ⴝ .394, p ⴝ .002 r ⫽ .154, p ⫽ .243 r ⫽ .219, p ⫽ .096

(r ⴝ .376, p ⴝ .004ⴱⴱ) (r ⫽ .106, p ⫽ .429) (r ⫽ .178, p ⫽ .181)
Note. Age-corrected correlations in parenthesis. Bold ⫽ remains significant after multiple comparison testing.
ⴱ
p ⬍ .05. ⴱⴱ p ⬍ .01. ⴱⴱⴱ p ⬍ .001.
 THEORY OF MIND ACROSS THE ADULT LIFESPAN 761

Table 6
Results for Forward Stepwise Regression Analyses for Each SSFT ToM Score

Standardized
coefficients– beta % Variance
SSFT ToM scores weights explained ANOVA

Intention Total ⫽ 49.9% F ⫽ 18.29, p ⬍ .001ⴱⴱⴱ

Working memory ␤ ⫽ .509, p ⬍ .001 24.5%
Affective empathy ␤ ⫽ .348, p ⫽ .001 18.5%
Gender ␤ ⫽ ⫺.277, p ⫽ .008 7%
Mental state speech Total ⫽ 41.7% F ⫽ 13.13, p ⬍ .001ⴱⴱⴱ
Working memory ␤ ⫽ .510, p ⬍ .001 28.9%
Gender ␤ ⫽ ⫺.264, p ⫽ .014 7.2%
Cognitive empathy ␤ ⫽ .219, p ⫽ .039 4.5%
Interaction Total ⫽ 37.9% F ⫽ 17.11, p ⬍ .001ⴱⴱⴱ
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Executive control ␤ ⫽ ⫺.465, p ⬍ .001 28.5%

This document is copyrighted by the American Psychological Association or one of its allied publishers.

Education ␤ ⫽ .315, p ⫽ .005 9.4%

Memory Total ⫽ 37.4% F ⫽ 16.71, p ⬍ .001ⴱⴱⴱ
Age ␤ ⫽ ⫺.454, p ⬍ .001 30.9%
Working memory ␤ ⫽ .273, p ⫽ .020 6.4%
Note. Independent variables entered into models were: age, education, gender, working memory, inhibitory
control, cognitive and affective empathy, and BAP traits.
ⴱⴱⴱ
p ⬍ .001.

related significantly with BAP traits. This finding is very similar to correlations controlling for age suggest that these abilities are
that of Murray et al. (2017) who found that ToM Intention per- important not just in ageing but across the adult lifespan.
formance correlated significantly with autism quotient scores in Results of this study should be considered keeping in mind some
typical young adults, although the association did not reach the limitations. Recruitment occurred in different settings depending
high significance level imposed (p ⬍ .01). These findings may on age (young adults were predominantly university students,
reflect the fact that although ToM difficulties reflect the core social whereas middle-aged adults were recruited at a café, and older
communication feature in ASD, they are not the only important adults at an independent residential association), and this may lead
diagnostic characteristic. to sampling biases. As in many ageing studies, the older individ-
The results of both the correlation and regression analyses uals who participated were relatively healthy, with eight taking one
suggest that executive function and empathy explain a proportion to two medications and only one taking three or more medications.
of the variance in ToM performance, whereas age does not con- This compares with United Kingdom national statistics showing
tribute statistically to the regression models. Results suggests that 88% of those aged over 65 take at least one medication, and 68%
working memory contributes to ToM performance across the take at least three medications (Scholes, Faulding, & Mindell,
lifespan, not just in ageing. However, the association between 2014). This suggests that this community housing sample are not
ToM performance and inhibitory control no longer reached signif-
unusually unwell for their age. Age differences were also noted
icance after controlling for age. Although age correlated signifi-
between men and women. We focused on two aspects of executive
cantly with ToM abilities, it was not the most important metric for
function (inhibitory control and working memory), but future
explaining performance. ToM intention was explained by working
studies may benefit from examining the contribution of other
memory, affective empathy and gender; and ToM interactions by
aspects of executive function and a formal measure of episodic
inhibitory control and education. Variance in mental state speech
memory. In addition, this study used a recently developed, novel
was explained by working memory, gender, and cognitive empa-
ToM task. Although our study demonstrates a pattern of results
thy. Age did not contribute to explaining the variance on any of the
measure of ToM. Only for the memory component of ToM did age similar to studies using more established measures and a previous
explain a large proportion of the variance in performance. Al- study has demonstrated convergent validity for ToM SSFT inten-
though some previous studies have found that age-ToM associa- tion scores compared with the widely used Strange Stories Task
tions remain after taking into account executive function (Bern- (Murray et al., 2017), we did not include a comparison ToM
stein et al., 2011; Bottiroli et al., 2016; Maylor et al., 2002), others measure in this study. In future studies it would be useful to
have found that executive function mediates age-ToM associations compare the SSFT to established ToM measures. Another potential
(Bailey & Henry, 2008; Charlton et al., 2009; Duval et al., 2011; limitation of the SSFT task across the lifespan is the use of young
Rakoczy et al., 2012). Results from this study support the impor- actors in the scenarios. Previous research has demonstrated own-
tance of executive function abilities for ToM performance, not age biases in recognition memory tasks, but the effect of own-age
only in ageing but across the adult lifespan. Few studies have biases in attention is less clear (Harrison & Hole, 2009). Future
explicitly assessed the impact of both executive function and task development could expand the scenarios to include actors
empathy on ToM across the lifespan (Duval et al., 2011). The across a wider age-range, to allow the examination of possible
results of this study suggest that cognition and empathy both own-age biases. Furthermore, as the SSFT was designed as an
contribute to ToM performance, and the results of the partial ecologically valid measure, future studies could assess real-life
 762 JOHANSSON NOLAKER, MURRAY, HAPPÉ, AND CHARLTON

ToM abilities through self and other-report, to examine whether in normal aging. Cortex, 46, 474 – 489. http://dx.doi.org/10.1016/j.cortex
the SSFT is providing information valid to the real-world. .2009.07.005
In summary, in this lifespan study of ToM we found that even Charlton, R. A., Barrick, T. R., Markus, H. S., & Morris, R. G. (2009).
using a test designed to have ecological validity, performance on Theory of mind associations with other cognitive functions and brain
imaging in normal aging. Psychology and Aging, 24, 338 –348. http://
ToM measures was negatively associated with age. However, ToM
dx.doi.org/10.1037/a0015225
performance was largely explained by executive function and
Dumontheil, I., Apperly, I. A., & Blakemore, S. J. (2010). Online usage of
empathy, rather than age. Results suggest that age-related ToM theory of mind continues to develop in late adolescence. Developmental
changes may be due to other cognitive and sociocognitive changes Science, 13, 331–338. http://dx.doi.org/10.1111/j.1467-7687.2009
rather than age alone. This suggests that declines in ToM perfor- .00888.x
mance, associated difficulties in social interaction and possible Duval, C., Piolino, P., Bejanin, A., Eustache, F., & Desgranges, B. (2011).
sequelae of social isolation, may not be inevitable in ageing. By Age effects on different components of theory of mind. Consciousness
supporting cognitive functions it may be possible to reduce the and Cognition, 20, 627– 642. http://dx.doi.org/10.1016/j.concog.2010.10
impact of ToM changes in later life. Further studies are required to .025
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

fully understand the associations between sociocognitive and so- Flavell, J. H. (1999). Cognitive development: Children’s knowledge about
This document is copyrighted by the American Psychological Association or one of its allied publishers.

the mind. Annual Review of Psychology, 50, 21– 45. http://dx.doi.org/

cial functions across the lifespan, in order to mediate the poten-
10.1146/annurev.psych.50.1.21
tially negative consequences on well-being.
Folstein, M. F., Folstein, S. E., & McHugh, P. R. (1975). “Mini-mental
state.” A practical method for grading the cognitive state of patients for
the clinician. Journal of Psychiatric Research, 12, 189 –198. http://dx
References
.doi.org/10.1016/0022-3956(75)90026-6
Adams, K. B., Sanders, S., & Auth, E. A. (2004). Loneliness and depres- Golden, C. J. (1978). Stroop color and word test: A manual for clinical and
sion in independent living retirement communities: Risk and resilience experimental uses. Chicago, IL: Skoelting.
factors. Aging & Mental Health, 8, 475– 485. http://dx.doi.org/10.1080/ Grove, R., Baillie, A., Allison, C., Baron-Cohen, S., & Hoekstra, R. A.
13607860410001725054 (2014). The latent structure of cognitive and emotional empathy in
Bailey, P. E., & Henry, J. D. (2008). Growing less empathic with age: individuals with autism, first-degree relatives and typical individuals.
Disinhibition of the self-perspective. The Journals of Gerontology: Molecular Autism, 5, 42. http://dx.doi.org/10.1186/2040-2392-5-42
Series B, Psychological Sciences and Social Sciences, 63, 219–P226. Happé, F. G. E. (1994). An advanced test of theory of mind: Understanding
http://dx.doi.org/10.1093/geronb/63.4.P219 of story characters’ thoughts and feelings by able autistic, mentally
Bailey, P. E., Henry, J. D., & Von Hippel, W. (2008). Empathy and social handicapped, and normal children and adults. Journal of Autism and
functioning in late adulthood. Aging & Mental Health, 12, 499 –503. Developmental Disorders, 24, 129 –154. http://dx.doi.org/10.1007/
http://dx.doi.org/10.1080/13607860802224243 BF02172093
Baron-Cohen, S., Leslie, A. M., & Frith, U. (1985). Does the autistic child Happé, F. G. E. (1995). The role of age and verbal ability in the theory of
have a “theory of mind”? Cognition, 21, 37– 46. http://dx.doi.org/10 mind task performance of subjects with autism. Child Development, 66,
.1016/0010-0277(85)90022-8 843– 855. http://dx.doi.org/10.2307/1131954
Baron-Cohen, S., Wheelwright, S., Hill, J., Raste, Y., & Plumb, I. (2001). Happé, F. G. E., Winner, E., & Brownell, H. (1998). The getting of
The “Reading the Mind in the Eyes” test revised version: A study with wisdom: Theory of mind in old age. Developmental Psychology, 34,
normal adults, and adults with Asperger syndrome or high-functioning 358 –362. http://dx.doi.org/10.1037/0012-1649.34.2.358
autism. Journal of Child Psychology and Psychiatry, and Allied Disci- Harrison, V., & Hole, G. J. (2009). Evidence for a contact-based explana-
plines, 42, 241–251. http://dx.doi.org/10.1111/1469-7610.00715 tion of the own-age bias in face recognition. Psychonomic Bulletin &
Bernstein, D. M., Thornton, W. L., & Sommerville, J. A. (2011). Theory of Review, 16, 264 –269.
mind through the ages: Older and middle-aged adults exhibit more errors Hasher, L., & Zacks, R. T. (1988). Working memory, comprehension, and
than do younger adults on a continuous false belief task. Experimental aging: A review and a new view. In G. H. Bower (Ed.), The psychology
Aging Research, 37, 481–502. http://dx.doi.org/10.1080/0361073X of learning and motivation: Advances in research and theory (Vol. 22
.2011.619466 ed., pp. 193–225). San Diego, CA: Academic Press, Inc. http://dx.doi
Blair, R. J. R. (2005). Responding to the emotions of others: Dissociating .org/10.1016/S0079-7421(08)60041-9
forms of empathy through the study of typical and psychiatric popula- Hawkley, L. C., Thisted, R. A., Masi, C. M., & Cacioppo, J. T. (2010).
tions. Consciousness and Cognition, 14, 698 –718. http://dx.doi.org/10 Loneliness predicts increased blood pressure: 5-year cross-lagged anal-
.1016/j.concog.2005.06.004 yses in middle-aged and older adults. Psychology and Aging, 25, 132–
Blozik, E., Wagner, J. T., Gillmann, G., Iliffe, S., von Renteln-Kruse, W., 141. http://dx.doi.org/10.1037/a0017805
Lubben, J., . . . Clough-Gorr, K. M. (2009). Social network assessment Heavey, L., Phillips, W., Baron-Cohen, S., & Rutter, M. (2000). The
in community-dwelling older persons: Results from a study of three awkward moments test: A naturalistic measure of social understanding
European populations. Aging Clinical and Experimental Research, 21, in autism. Journal of Autism and Developmental Disorders, 30, 225–
150 –157. http://dx.doi.org/10.1007/BF03325223 236. http://dx.doi.org/10.1023/A:1005544518785
Bottiroli, S., Cavallini, E., Ceccato, I., Vecchi, T., & Lecce, S. (2016). Henry, J. D., Phillips, L. H., Ruffman, T., & Bailey, P. E. (2013). A
Theory of Mind in aging: Comparing cognitive and affective compo- meta-analytic review of age differences in theory of mind. Psychology
nents in the faux pas test. Archives of Gerontology and Geriatrics, 62, and Aging, 28, 826 – 839. http://dx.doi.org/10.1037/a0030677
152–162. http://dx.doi.org/10.1016/j.archger.2015.09.009 Hurley, R., Losh, M., Parlier, M., Reznick, J. S., & Piven, J. (2007). The
Castelli, F., Happé, F., Frith, U., & Frith, C. (2000). Movement and mind: Broad Autism Phenotype Questionnaire. Journal of Autism and Devel-
A functional imaging study of perception and interpretation of complex opmental Disorders, 37, 1679 –1690.
intentional movement patterns. NeuroImage, 12, 314 –325. http://dx.doi IBM Corp. (2013). IBM SPSS statistics for windows (version 22.0) [Com-
.org/10.1006/nimg.2000.0612 puter software]. Armonk, NY: IBM Corp.
Charlton, R. A., Barrick, T. R., Lawes, I. N. C., Markus, H. S., & Morris, Jameel, L., Vyas, K., Bellesi, G., Roberts, V., & Channon, S. (2014). Going
R. G. (2010). White matter pathways associated with working memory ‘above and beyond’: Are those high in autistic traits less pro-social?
 THEORY OF MIND ACROSS THE ADULT LIFESPAN 763

Journal of Autism and Developmental Disorders, 44, 1846 –1858. http:// Rakoczy, H., Harder-Kasten, A., & Sturm, L. (2012). The decline of theory
dx.doi.org/10.1007/s10803-014-2056-3 of mind in old age is (partly) mediated by developmental changes in
Jolliffe, D., & Farrington, D. P. (2006). Development and validation of the domain-general abilities. British Journal of Psychology, 103, 58 –72.
Basic Empathy Scale. Journal of Adolescence, 29, 589 – 611. http://dx http://dx.doi.org/10.1111/j.2044-8295.2011.02040.x
.doi.org/10.1016/j.adolescence.2005.08.010 Rogers, K., Dziobek, I., Hassenstab, J., Wolf, O. T., & Convit, A. (2007).
Jones, A. P., Happé, F. G. E., Gilbert, F., Burnett, S., & Viding, E. (2010). Who cares? Revisiting empathy in Asperger syndrome. Journal of
Feeling, caring, knowing: Different types of empathy deficit in boys Autism and Developmental Disorders, 37, 709 –715. http://dx.doi.org/10
with psychopathic tendencies and autism spectrum disorder. Journal of .1007/s10803-006-0197-8
Child Psychology and Psychiatry, 51, 1188 –1197. http://dx.doi.org/10 Saltzman, J., Strauss, E., Hunter, M., & Archibald, S. (2000). Theory of
.1111/j.1469-7610.2010.02280.x mind and executive functions in normal human aging and Parkinson’s
Keightley, M. L., Winocur, G., Burianova, H., Hongwanishkul, D., & disease. Journal of the International Neuropsychological Society, 6,
Grady, C. L. (2006). Age effects on social cognition: Faces tell a 781–788. http://dx.doi.org/10.1017/S1355617700677056
different story. Psychology and Aging, 21, 558 –572. http://dx.doi.org/ Scheeren, A. M., de Rosnay, M., Koot, H. M., & Begeer, S. (2013).
10.1037/0882-7974.21.3.558 Rethinking theory of mind in high-functioning autism spectrum disor-
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Lamport, D., & Turner, L. A. (2014). Romantic attachment, empathy, and der. Journal of Child Psychology and Psychiatry, 54, 628 – 635. http://
the broader autism phenotype among college students. The Journal of
This document is copyrighted by the American Psychological Association or one of its allied publishers.

dx.doi.org/10.1111/jcpp.12007
Genetic Psychology, 175, 202–213. http://dx.doi.org/10.1080/00221325
Scholes, S., Faulding, S., & Mindell, J. (2014). Prescribed medicines. In R.
.2013.856838
Craig & J. Mindell (Eds.), Health survey for England, 2013 (pp. 1–27).
Mahy, C. E. V., Vetter, N., Kühn-Popp, N., Löcher, C., Krautschuk, S., &
Leeds, UK: Health and Social Care Information Centre. Retrieved from
Kliegel, M. (2014). The influence of inhibitory processes on affective
http://digital.nhs.uk/catalogue/PUB16076
theory of mind in young and old adults. Neuropsychology, Development,
Sucksmith, E., Allison, C., Baron-Cohen, S., Chakrabarti, B., & Hoekstra,
and Cognition Section B, Aging, Neuropsychology and Cognition, 21,
R. A. (2013). Empathy and emotion recognition in people with autism,
129 –145. http://dx.doi.org/10.1080/13825585.2013.789096
Maylor, E. A., Moulson, J. M., Muncer, A. M., & Taylor, L. A. (2002). first-degree relatives, and controls. Neuropsychologia, 51, 98 –105.
Does performance on theory of mind tasks decline in old age? British http://dx.doi.org/10.1016/j.neuropsychologia.2012.11.013
Journal of Psychology, 93, 465– 485. http://dx.doi.org/10.1348/ Sullivan, S., & Ruffman, T. (2004). Social understanding: How does it fare
000712602761381358 with advancing years? British Journal of Psychology, 95, 1–18. http://
Mitchell, R. L. C., & Phillips, L. H. (2015). The overlapping relationship dx.doi.org/10.1348/000712604322779424
between emotion perception and theory of mind. Neuropsychologia, 70, Tombaugh, T. N., & McIntyre, N. J. (1992). The mini-mental state exam-
1–10. http://dx.doi.org/10.1016/j.neuropsychologia.2015.02.018 ination: A comprehensive review. Journal of the American Geriatrics
Murray, K., Johnston, K., Cunnane, H., Kerr, C., Spain, D., Gillan, N., . . . Society, 40, 922–935. http://dx.doi.org/10.1111/j.1532-5415.1992
Happé, F. (2017). A new test of advanced theory of mind: The “Strange .tb01992.x
Stories Film Task” captures social processing differences in adults with Wechsler, D. (2008). Wechsler Adult Intelligence Scale - IV (4th ed.). San
autism spectrum disorders. Autism Research, 10, 1120 –1132. http://dx Antonio, TX: Psychological Corporation.
.doi.org/10.1002/aur.1744
Pluta, A., Gawron, N., Sobańska, M., Wójcik, A. D., & Łojek, E. (2017).
The nature of the relationship between neurocognition and theory of Received August 14, 2017
mind impairments in stroke patients. Neuropsychology, 31, 666 – 681. Revision received January 9, 2018
http://dx.doi.org/10.1037/neu0000379 Accepted January 15, 2018 䡲