2006 Golan -kont.wzrokowy, Autyzm(1)
[ Pobierz całość w formacie PDF ]
SOCIAL NEUROSCIENCE, 2006, 1 (2), 111 123
The ‘‘Reading the Mind in Films’’ Task: Complex emotion
recognition in adults with and without autism spectrum
conditions
Ofer Golan, Simon Baron-Cohen, Jacqueline J. Hill, and Yael Golan
Autism Research Centre, Psychiatry Dept, Cambridge University, UK
Background: Individuals with autism spectrum conditions (ASC) have difficulties recognizing mental
states in others. Most research has focused on recognition of basic emotions from faces and voices
separately. This study reports the results of a new task, assessing recognition of complex emotions and
mental states from social scenes taken from feature films. The film format arguably is more challenging
and ecologically closer to real social situations. Sample and method: A group of adults with ASC (
n
/
22)
were compared to a group of matched controls from the general population (
n
/
22). Participants were
tested individually. Results: Overall, individuals with ASC performed significantly lower than controls.
There was a positive correlation between verbal IQ and task scores. Using task scores, more than 90% of
the participants were correctly allocated to their group. Item analysis showed that the errors individuals
with ASC make when judging socioemotional information are subtle. Conclusions: This new test of
complex emotion and mental state recognition reveals that adults with ASC have residual difficulties in
this aspect of empathy. The use of language-based compensatory strategies for emotion recognition is
discussed.
Autism Spectrum Conditions (ASC) are neurode-
velopmental conditions, characterized by cogni-
tive and behavioral difficulties in communication
and social interaction (American Psychiatric As-
sociation, 1994; World Health Organization,
1994). The effects of ASC are lifelong, and
although learning occurs throughout develop-
ment, social and communication difficulties re-
main even among individuals diagnosed with
Asperger Syndrome (AS) or High Functioning
Autism (HFA; Attwood, 1998; Baron-Cohen,
Tager-Flusberg, & Cohen, 2000b; Frith, 1989;
Hobson, 1993).
One key cognitive theory of autism views the
social dysfunction in ASC as a result of a deficit in
what is variously referred to as ‘‘theory of mind’’
(ToM; Astington, Harris, & Olson, 1988), ‘‘mind-
reading’’ (Wellman, 1992), or ‘‘mentalizing’’
(Frith, 1989), that is, the ability to understand
other people’s minds, to decipher their intentions,
emotions and thoughts. Impaired in this ability,
individuals with ASC are bound to feel confused
by other people’s behavior, failing to understand
the motives that underlie human action, and
experiencing degrees of ‘‘mind-blindness’’
(Baron-Cohen, 1995) or degrees of deficit in
‘‘empathizing’’ (Baron-Cohen, Wheelwright,
Lawson, Griffin, & Hill, 2002).
The neural network that underlies these abil-
ities was first described by Brothers and Ring as
Correspondence should be addressed to: Ofer Golan, Autism Research Centre, Douglas House, 18b Trumpington Road,
Cambridge CB2 2AH, UK. E-mail: og211@cam.ac.uk
OG was supported by the Corob Charitable Trust, the Cambridge Overseas Trust, the National Alliance for Autism Research
(NAAR), and B’nai B’rith Scholarships. SBC and JH were supported by the Shirley Foundation, Medical Research Council (MRC),
and the Three Guineas Trust. We are grateful to Chris Ashwin, Sally Wheelwright, Sarah Johnson and Emma Chapman for their
support.
#
2006 Psychology Press, an imprint of the Taylor & Francis Group, an informa business
DOI:10.1080/17470910600980986
www.psypress.com/socialneuroscience
112
GOLAN ET AL.
Figure 1. The social brain network (from Baron-Cohen & Belmonte, 2005; reprinted with permission).
‘‘the social brain’’ (Brothers & Ring, 1992), and is
illustrated in Figure 1. The medial, inferior frontal
and superior temporal cortices, along with the
amygdala, form a network of brain regions that
implement computations relevant to social pro-
cesses. Perceptual inputs to these social computa-
tions may arise in part from regions in the
fusiform gyrus and from the adjacent inferior
occipital gyrus that activate in response to faces.
Neuroimaging studies of ToM in ASC reveal
that individuals with autism show less activation
than controls from the general population in
these ‘‘social brain’’ areas, e.g., lower activation
of the left medial prefrontal cortex when reading
stories that require attribution of mental states
(Happe et al., 1996; Nieminen-von Wendt et al.,
2003), or lower activation in the medial prefrontal
cortex, the superior temporal sulcus and the
temporal poles when watching geometric shapes
moving in a way that can be interpreted as social
(Castelli, Frith, Happe, & Frith, 2002). In an MRI
study using a task of emotion and mental state
recognition from pictures of eyes, participants
with ASC had less extensive frontal activation
than controls, and no activation of the amygdala
(Baron-Cohen et al., 1999b). This latter finding
prompted the amygdala theory of autism (Baron-
Cohen, Ring, Bullmore, Wheelwright, Ashwin, &
Williams, 2000a). In addition, the control group
had a significantly stronger response in the left
amygdala, the right insula and the left inferior
frontal gyrus (Baron-Cohen et al., 1999b). A
related study using mental state vs. non-mental
state words in an auditory paradigm involving
SPECT found the orbito-frontal cortex was less
active in autism, compared to controls (Baron-
Cohen, Ring, Moriarty, Schmitz, Costa, & Ell,
1994). A recent MRI study scanning parents of
children with ASC during the same ‘‘Eyes’’ task
demonstrated that reduced activations in parts of
the social brain during this task are evident in
first-degree relatives, suggesting this may repre-
sent an ‘‘endophenotype’’ (Baron-Cohen et al.,
2006).
Another key cognitive theory of autism views
the socioemotional deficit in terms of ‘‘weak
central coherence’’ (WCC; Frith, 1989; Happe,
1999), suggesting that ASC are characterized by
an increased focus on detail and difficulties
integrating information into a coherent whole.
According to this theory, without the ability to
group details in context to derive meaning,
individuals with autism experience a fragmented
world. Social functioning, which requires fast
integration of context-dependent information in
real time, would be seriously hampered under
such conditions.
Models of brain function in autism suggest
that WCC is the cognitive manifestation of
altered connectivity between local ‘‘low-level’’
READING THE MIND IN FILMS
113
perceptual brain systems and frontal brain re-
gions in charge of integration or ‘‘coherence’’. If
brain regions that control integration are less
connected to their perceptual inputs, the resulting
cognitive-behavioral outcome would be enhanced
local, on account of reduced global, processing
(Baron-Cohen & Belmonte, 2005; Belmonte,
Allen, Beckel-Mitchener, Boulanger, Carper, &
Webb, 2004a). In a visual signal detection MRI
study adults with ASC showed heightened ventral
occipital brain activation and lowered pre-frontal,
parietal, and temporal activation, compared to
controls (Belmonte & Yurgelun-Todd, 2003). A
related MRI study using the Embedded Figures
Task with adults with ASC found a similar pattern
of differences (Ring et al., 1999). Although these
findings relate to low level visual information
processing, such under-connectivity between
brain regions could explain the social deficit in
ASC in terms of failure to utilize social cues to
understand socioemotional phenomena (Bel-
monte et al., 2004a, 2004b; Critchley et al., 2000).
The current study focuses on the recognition of
emotions and mental states in others, which is a
fundamental part of empathizing. In addition,
emotion recognition strongly depends on the
ability to integrate multimodal information in
context. Hence, both ToM and WCC theories
would predict difficulties in this domain in ASC.
Indeed, research shows that emotion and mental
state recognition are core difficulties in children
and adults with ASC (Baron-Cohen, 1995; Hob-
son, 1994). Such difficulties have been found
through cognitive, behavioral and neuroimaging
studies, and across different sensory modalities
(Frith & Hill, 2004). Neuroimaging studies reveal
that people with ASC show less activation in
brain regions not just related to emotion recogni-
tion but to face processing in general, such as the
fusiform gyrus (Critchley et al., 2000; Pierce,
Muller, Ambrose, Allen, & Courchesne, 2001;
Schultz et al., 2003). As mentioned above, there
is evidence of reduced activation in brain areas
that play a major role in processing of emotion,
such as the amygdala (Ashwin, Baron-Cohen,
Wheelwright, O’Riordan, & Bullmore, in press;
Baron-Cohen et al., 1999b; Critchley et al., 2000).
These studies suggest that the processing of
socioemotional input in ASC is qualitatively
different to that of controls from the general
population.
Most emotion recognition studies carried out
with individuals with ASC have focused on the
recognition of six emotions, considered ‘‘basic’’
(happiness, sadness, fear, anger, surprise and
disgust). These ‘‘basic emotions’’ are expressed
and recognized cross-culturally (Ekman, 1993;
Ekman & Friesen, 1971) and may be neurologi-
cally distinct (Adolphs, 2002; Griffiths, 1997).
Studies assessing the recognition of these emo-
tions report inconclusive findings with children
and adults with ASC. Some studies report diffi-
culties in recognition of basic emotions from
dynamic or static facial expressions, from voice
recordings, and from matching of stimuli from the
two modalities (Celani, Battacchi, & Arcidiacono,
1999; Deruelle, Rondan, Gepner, & Tardif, 2004;
Hobson, 1986a, 1986b; Loveland, Tunali Kotoski,
Chen, & Brelsford, 1995; Macdonald
et al., 1989; Yirmiya, Sigman, Kasari, & Mundy,
1992). Other studies have found no such difficul-
ties in recognition of the basic emotions in
children and adults with ASC (Adolphs, 2001;
Baron-Cohen, Spitz, & Cross, 1993; Boucher,
Lewis, & Collis, 2000; Grossman, Klin, Carter,
& Volkmar, 2000; Loveland et al., 1997).
These inconclusive findings may be due to
ceiling effects in basic emotion recognition tasks
that focus on faces or voices alone. In other
words, though they are delayed in recognition of
basic emotions compared to typically developing
controls, many adults with ASC may compensate
for their early deficit in this area and thus pass
these tasks to a similar level seen in controls, if
the tasks are not too challenging (Adolphs, Sears,
& Piven, 2001; Grandin, 1995).
This possible ceiling effect in basic emotion
recognition tasks raises two questions: First, can
adults with ASC recognize emotions and mental
states that are more complex? These could be
cognitive, belief-based rather than situation-based
emotions (Harris, 1989), e.g.,
troubled
, or
resigned
. They could also be social emotions,
e.g.,
caring
or
embarrassed
(Baron-Cohen, Jol-
liffe, Mortimore, & Robertson, 1997a; Kasari,
Chamberlain, & Bauminger, 2001).
The second question deals with the ability of
individuals with ASC to recognize emotions from
more ecological, life-like situations. These usually
require integration of facial expressions, body
language, intonation, verbal content and the
contextual cues into a coherent picture, rather
than relying on discrete emotional stimuli like
faces or voices, or specific features of these.
Typically developing children and adults integrate
all these features automatically and unconsciously
when processing social situations, but can indivi-
duals with ASC do this?
114
GOLAN ET AL.
The first question has been addressed in
several studies conducted with children and
adults with ASC in which they were asked to
recognize complex emotions and mental states
from visual emotional stimuli. Tasks included
pictures of eyes (Baron-Cohen, Wheelwright,
Hill, Raste, & Plumb, 2001a; Baron-Cohen,
Wheelwright, & Jolliffe, 1997b; Baron-Cohen,
Wheelwright, Spong, Scahill, & Lawson, 2001c),
and static and dynamic facial expressions (Baron-
Cohen et al., 1997b; Golan, Baron-Cohen, & Hill,
2006b). Auditory tasks have used short utterances
(Golan, Baron-Cohen, Hill, & Rutherford, in
press; Kleinman, Marciano, & Ault, 2001; Ruther-
ford, Baron-Cohen, & Wheelwright, 2002). Other
tasks have assessed the ability to detect mental
states from contextual cues (Baron-Cohen,
O’Riordan, Stone, Jones, & Plaisted, 1999a;
Fein, Lucci, Braverman, & Waterhouse, 1992;
Happe, 1994). In all of these studies, individuals
with ASC performed at a significantly lower level
compared to controls, matched for age and IQ.
These results suggest that recognition of basic
emotions might be relatively preserved (or com-
pensated for) in individuals with ASC, but that
they show difficulties recognizing more complex
emotional and mental states.
The second question has received less empiri-
cal attention. As far as we are aware, only two
studies have assessed the ability of adults with
ASC to recognize emotions and mental states
from more ecological stimuli. ‘‘The Awkward
Moments Test’’ (Heavey, Phillips, Baron-Cohen,
& Rutter, 2000) presented participants with
seven short social situations, all taken from
television advertisements. They were asked to
judge the protagonist’s mental state at the end of
each scene. Participants with ASC performed at
a significantly lower level compared to general
population controls. In another study (Klin,
Jones, Schultz, Volkmar, & Cohen, 2002a,
2002b), a single social scene from a feature film
was presented to adults with ASC, while tracking
their gaze. Compared to typically developed
controls, participants with ASC looked less at
the eyes of characters, and more at characters’
mouths and surrounding objects, thus missing
socioemotional information pertinent to the
understanding of the social situation. Both of
these studies suggest high-functioning adults with
ASC are impaired on more ecological tests of
social understanding. However, both studies in-
volved tasks with a relatively small number of
scenes (seven in one, and one in the other),
which limits their validity and power and risks
floor or ceiling effects. ‘‘The Awkward Moments
Test’’ also used adverts, which are by definition
exaggerated. A more subtle collection of social
situations may better represent the complexity of
socioemotional interactions that we typically
encounter, and provide a more valid measure
of the ability of adults with ASC to interpret
them.
In this study, we report the ‘‘Reading the Mind
in Films’’ (RMF) task, which assesses complex
emotion and mental state recognition. We com-
pare the ability of adults with Asperger Syndrome
(AS) or High Functioning Autism (HFA) with
that of adults from the general population. The
task uses 22 short scenes, taken from feature films
(with permission). The scenes include visual input
(facial expressions, body language, action), audi-
tory input (prosody, verbal content) and context,
thus making the task more ecological than pre-
vious tasks testing each modality separately. The
emotions and mental states selected vary in
valence, intensity and complexity. They were
chosen for their relevance to everyday social
interaction. Hence, any recognition deficit could
impede one’s functioning in social situations. For
example, if the speaker fails to recognize the
listener’s embarrassment in a conversation, she or
he may not attempt to change the subject, and
may, without realizing it, offend or cause distress
in the listener (Grandin, 1995).
Judging complex emotions from ecological
social stimuli requires integration of multimodal
information into a coherent holistic picture as
well as attribution of mental states to others.
Therefore, based on the ToM or WCC models of
autism, we predicted that participants with ASC
would score significantly lower than controls on
the RMF task. In addition, we hypothesized that
the more autistic traits one possesses, the lower
one’s score on the task. In order to rule out that
difficulties in responding to the task are simply
due to working memory problems, we tested for
a correlation between task performance and
item length, with the number of characters
appearing in the item. We also tested for a
correlation between task score and age or IQ.
We hypothesized that a positive association
between task scores and verbal IQ would be
found for both groups, given that the task
involves matching a mental state word to a
character’s action. Lastly, we tested for correla-
tions between the RMF task score with other
READING THE MIND IN FILMS
115
complex emotion recognition tasks (from faces
and voices separately), to validate our new task.
Instruments
Reading the Mind in Films (RMF): Task
development
METHOD
The film clips are available at www.autism
researchcentre.com/tests. Thirty short scenes
(530 seconds long,
M
/
14.8,
SD
/
9.2), were
sampled from three feature films and one mini
series by two of the authors. The films were
picked for their dramatic value and frequent
occurrence of emotional scenes. The selected
scenes involved emotional interaction between
14 characters, and the expression of complex
emotions and mental states (e.g.,
smug
,
awkward
,
concerned
). In each scene, a protagonist was
identified and their emotion or mental state at
the end of the scene was labeled. Three foils were
selected for each item. In order to match the foils
and the target word for verbal difficulty, an
emotion taxonomy was used (Baron-Cohen, Go-
lan, Wheelwright, & Hill, 2004). The taxonomy
comprises 412 emotions and mental states, in six
developmental levels. Selected foils were either in
the same or easier levels than the target. Foils
were selected so that they matched some of the
emotional information in the scene but not all of
it, e.g., matching the content of the language but
not the intonation or the context. The labels and
foils were then reviewed by two independent
judges. A handout with definitions for all the
target and foil words in the items included was
prepared for participants’ use before and during
the task.
The items were then played to 15 adults (7 men
and 8 women) randomly selected from the gen-
eral population. They were played to them on a
laptop computer, using DMDX experimental
software (Forster & Forster, 2003). Two examples
of items from the task are shown in Figure 2.
Participants
The AS/HFA group comprised 22 adults (17
males and 5 females), aged 1752 (
M
/
29.0,
SD
/
9.8). Participants had all been diagnosed
with AS/HFA in specialist centers using estab-
lished criteria (American Psychiatric Association,
1994; World Health Organization, 1994). They
were recruited from a local clinic for adults with
ASC, and from a research volunteer database.
The control group comprised 22 adults (18 males
and 4 females) from the general population, aged
1851 (
M
/
25.4,
SD
/
9.5). They were recruited
from a local employment agency. All participants
were given the Wechsler Abbreviated Scale of
Intelligence (WASI), comprising the vocabulary,
similarities, block design and matrix reasoning
tests. The WASI produces verbal, performance
and full scale IQ scores, with correlations of .88,
.84 and .92 with the full Wechsler scales (Wechs-
ler, 1999). All participants scored above 85 on
both verbal and performance scales. Their IQ
scores are shown in Table 1. To screen for autism
spectrum conditions, participants also filled in the
Autism Spectrum Quotient (AQ; Baron-Cohen,
Wheelwright, Skinner, Martin, & Clubley, 2001b).
Replicating Baron-Cohen et al.’s (2001b) finding,
86.4% of the AS/HFA group and none of
the control group scored above the cut-off score
of 31. The two groups were matched on sex
(x
2
(1)
/
0.14,
ns
), age, verbal IQ, and perfor-
mance IQ. The groups’ background data appears
in Table 1.
TABLE 1
Means, standard deviations and ranges of AQ, chronological age and WASI scores for the AS/HFA group and the control group
AS/HFA group (n
/
22)
Control group (n
/
22)
Mean
SD
Range
Mean
SD
Range
t(42)
AQ
38.5
7.8
16 49
14.0
5.4
6 26
12.11**
Age
29.0
9.8
17.4 52.0
25.4
9.6
17.6 51.2
1.25
Verbal IQ
110.6
10.8
87 129
116.4
14.5
86 138
1.49
Performance IQ
114.8
12.5
97 140
113.1
8.4
92 129
0.51
Full scale IQ
114.2
10.8
91 130
116.5
11.0
97 138
0.71
Note
:**
p
B
/
.001. For all the other measures
p
/
.1.
[ Pobierz całość w formacie PDF ]
zanotowane.pl doc.pisz.pl pdf.pisz.pl charloteee.keep.pl
SOCIAL NEUROSCIENCE, 2006, 1 (2), 111 123
The ‘‘Reading the Mind in Films’’ Task: Complex emotion
recognition in adults with and without autism spectrum
conditions
Ofer Golan, Simon Baron-Cohen, Jacqueline J. Hill, and Yael Golan
Autism Research Centre, Psychiatry Dept, Cambridge University, UK
Background: Individuals with autism spectrum conditions (ASC) have difficulties recognizing mental
states in others. Most research has focused on recognition of basic emotions from faces and voices
separately. This study reports the results of a new task, assessing recognition of complex emotions and
mental states from social scenes taken from feature films. The film format arguably is more challenging
and ecologically closer to real social situations. Sample and method: A group of adults with ASC (
n
/
22)
were compared to a group of matched controls from the general population (
n
/
22). Participants were
tested individually. Results: Overall, individuals with ASC performed significantly lower than controls.
There was a positive correlation between verbal IQ and task scores. Using task scores, more than 90% of
the participants were correctly allocated to their group. Item analysis showed that the errors individuals
with ASC make when judging socioemotional information are subtle. Conclusions: This new test of
complex emotion and mental state recognition reveals that adults with ASC have residual difficulties in
this aspect of empathy. The use of language-based compensatory strategies for emotion recognition is
discussed.
Autism Spectrum Conditions (ASC) are neurode-
velopmental conditions, characterized by cogni-
tive and behavioral difficulties in communication
and social interaction (American Psychiatric As-
sociation, 1994; World Health Organization,
1994). The effects of ASC are lifelong, and
although learning occurs throughout develop-
ment, social and communication difficulties re-
main even among individuals diagnosed with
Asperger Syndrome (AS) or High Functioning
Autism (HFA; Attwood, 1998; Baron-Cohen,
Tager-Flusberg, & Cohen, 2000b; Frith, 1989;
Hobson, 1993).
One key cognitive theory of autism views the
social dysfunction in ASC as a result of a deficit in
what is variously referred to as ‘‘theory of mind’’
(ToM; Astington, Harris, & Olson, 1988), ‘‘mind-
reading’’ (Wellman, 1992), or ‘‘mentalizing’’
(Frith, 1989), that is, the ability to understand
other people’s minds, to decipher their intentions,
emotions and thoughts. Impaired in this ability,
individuals with ASC are bound to feel confused
by other people’s behavior, failing to understand
the motives that underlie human action, and
experiencing degrees of ‘‘mind-blindness’’
(Baron-Cohen, 1995) or degrees of deficit in
‘‘empathizing’’ (Baron-Cohen, Wheelwright,
Lawson, Griffin, & Hill, 2002).
The neural network that underlies these abil-
ities was first described by Brothers and Ring as
Correspondence should be addressed to: Ofer Golan, Autism Research Centre, Douglas House, 18b Trumpington Road,
Cambridge CB2 2AH, UK. E-mail: og211@cam.ac.uk
OG was supported by the Corob Charitable Trust, the Cambridge Overseas Trust, the National Alliance for Autism Research
(NAAR), and B’nai B’rith Scholarships. SBC and JH were supported by the Shirley Foundation, Medical Research Council (MRC),
and the Three Guineas Trust. We are grateful to Chris Ashwin, Sally Wheelwright, Sarah Johnson and Emma Chapman for their
support.
#
2006 Psychology Press, an imprint of the Taylor & Francis Group, an informa business
DOI:10.1080/17470910600980986
www.psypress.com/socialneuroscience
112
GOLAN ET AL.
Figure 1. The social brain network (from Baron-Cohen & Belmonte, 2005; reprinted with permission).
‘‘the social brain’’ (Brothers & Ring, 1992), and is
illustrated in Figure 1. The medial, inferior frontal
and superior temporal cortices, along with the
amygdala, form a network of brain regions that
implement computations relevant to social pro-
cesses. Perceptual inputs to these social computa-
tions may arise in part from regions in the
fusiform gyrus and from the adjacent inferior
occipital gyrus that activate in response to faces.
Neuroimaging studies of ToM in ASC reveal
that individuals with autism show less activation
than controls from the general population in
these ‘‘social brain’’ areas, e.g., lower activation
of the left medial prefrontal cortex when reading
stories that require attribution of mental states
(Happe et al., 1996; Nieminen-von Wendt et al.,
2003), or lower activation in the medial prefrontal
cortex, the superior temporal sulcus and the
temporal poles when watching geometric shapes
moving in a way that can be interpreted as social
(Castelli, Frith, Happe, & Frith, 2002). In an MRI
study using a task of emotion and mental state
recognition from pictures of eyes, participants
with ASC had less extensive frontal activation
than controls, and no activation of the amygdala
(Baron-Cohen et al., 1999b). This latter finding
prompted the amygdala theory of autism (Baron-
Cohen, Ring, Bullmore, Wheelwright, Ashwin, &
Williams, 2000a). In addition, the control group
had a significantly stronger response in the left
amygdala, the right insula and the left inferior
frontal gyrus (Baron-Cohen et al., 1999b). A
related study using mental state vs. non-mental
state words in an auditory paradigm involving
SPECT found the orbito-frontal cortex was less
active in autism, compared to controls (Baron-
Cohen, Ring, Moriarty, Schmitz, Costa, & Ell,
1994). A recent MRI study scanning parents of
children with ASC during the same ‘‘Eyes’’ task
demonstrated that reduced activations in parts of
the social brain during this task are evident in
first-degree relatives, suggesting this may repre-
sent an ‘‘endophenotype’’ (Baron-Cohen et al.,
2006).
Another key cognitive theory of autism views
the socioemotional deficit in terms of ‘‘weak
central coherence’’ (WCC; Frith, 1989; Happe,
1999), suggesting that ASC are characterized by
an increased focus on detail and difficulties
integrating information into a coherent whole.
According to this theory, without the ability to
group details in context to derive meaning,
individuals with autism experience a fragmented
world. Social functioning, which requires fast
integration of context-dependent information in
real time, would be seriously hampered under
such conditions.
Models of brain function in autism suggest
that WCC is the cognitive manifestation of
altered connectivity between local ‘‘low-level’’
READING THE MIND IN FILMS
113
perceptual brain systems and frontal brain re-
gions in charge of integration or ‘‘coherence’’. If
brain regions that control integration are less
connected to their perceptual inputs, the resulting
cognitive-behavioral outcome would be enhanced
local, on account of reduced global, processing
(Baron-Cohen & Belmonte, 2005; Belmonte,
Allen, Beckel-Mitchener, Boulanger, Carper, &
Webb, 2004a). In a visual signal detection MRI
study adults with ASC showed heightened ventral
occipital brain activation and lowered pre-frontal,
parietal, and temporal activation, compared to
controls (Belmonte & Yurgelun-Todd, 2003). A
related MRI study using the Embedded Figures
Task with adults with ASC found a similar pattern
of differences (Ring et al., 1999). Although these
findings relate to low level visual information
processing, such under-connectivity between
brain regions could explain the social deficit in
ASC in terms of failure to utilize social cues to
understand socioemotional phenomena (Bel-
monte et al., 2004a, 2004b; Critchley et al., 2000).
The current study focuses on the recognition of
emotions and mental states in others, which is a
fundamental part of empathizing. In addition,
emotion recognition strongly depends on the
ability to integrate multimodal information in
context. Hence, both ToM and WCC theories
would predict difficulties in this domain in ASC.
Indeed, research shows that emotion and mental
state recognition are core difficulties in children
and adults with ASC (Baron-Cohen, 1995; Hob-
son, 1994). Such difficulties have been found
through cognitive, behavioral and neuroimaging
studies, and across different sensory modalities
(Frith & Hill, 2004). Neuroimaging studies reveal
that people with ASC show less activation in
brain regions not just related to emotion recogni-
tion but to face processing in general, such as the
fusiform gyrus (Critchley et al., 2000; Pierce,
Muller, Ambrose, Allen, & Courchesne, 2001;
Schultz et al., 2003). As mentioned above, there
is evidence of reduced activation in brain areas
that play a major role in processing of emotion,
such as the amygdala (Ashwin, Baron-Cohen,
Wheelwright, O’Riordan, & Bullmore, in press;
Baron-Cohen et al., 1999b; Critchley et al., 2000).
These studies suggest that the processing of
socioemotional input in ASC is qualitatively
different to that of controls from the general
population.
Most emotion recognition studies carried out
with individuals with ASC have focused on the
recognition of six emotions, considered ‘‘basic’’
(happiness, sadness, fear, anger, surprise and
disgust). These ‘‘basic emotions’’ are expressed
and recognized cross-culturally (Ekman, 1993;
Ekman & Friesen, 1971) and may be neurologi-
cally distinct (Adolphs, 2002; Griffiths, 1997).
Studies assessing the recognition of these emo-
tions report inconclusive findings with children
and adults with ASC. Some studies report diffi-
culties in recognition of basic emotions from
dynamic or static facial expressions, from voice
recordings, and from matching of stimuli from the
two modalities (Celani, Battacchi, & Arcidiacono,
1999; Deruelle, Rondan, Gepner, & Tardif, 2004;
Hobson, 1986a, 1986b; Loveland, Tunali Kotoski,
Chen, & Brelsford, 1995; Macdonald
et al., 1989; Yirmiya, Sigman, Kasari, & Mundy,
1992). Other studies have found no such difficul-
ties in recognition of the basic emotions in
children and adults with ASC (Adolphs, 2001;
Baron-Cohen, Spitz, & Cross, 1993; Boucher,
Lewis, & Collis, 2000; Grossman, Klin, Carter,
& Volkmar, 2000; Loveland et al., 1997).
These inconclusive findings may be due to
ceiling effects in basic emotion recognition tasks
that focus on faces or voices alone. In other
words, though they are delayed in recognition of
basic emotions compared to typically developing
controls, many adults with ASC may compensate
for their early deficit in this area and thus pass
these tasks to a similar level seen in controls, if
the tasks are not too challenging (Adolphs, Sears,
& Piven, 2001; Grandin, 1995).
This possible ceiling effect in basic emotion
recognition tasks raises two questions: First, can
adults with ASC recognize emotions and mental
states that are more complex? These could be
cognitive, belief-based rather than situation-based
emotions (Harris, 1989), e.g.,
troubled
, or
resigned
. They could also be social emotions,
e.g.,
caring
or
embarrassed
(Baron-Cohen, Jol-
liffe, Mortimore, & Robertson, 1997a; Kasari,
Chamberlain, & Bauminger, 2001).
The second question deals with the ability of
individuals with ASC to recognize emotions from
more ecological, life-like situations. These usually
require integration of facial expressions, body
language, intonation, verbal content and the
contextual cues into a coherent picture, rather
than relying on discrete emotional stimuli like
faces or voices, or specific features of these.
Typically developing children and adults integrate
all these features automatically and unconsciously
when processing social situations, but can indivi-
duals with ASC do this?
114
GOLAN ET AL.
The first question has been addressed in
several studies conducted with children and
adults with ASC in which they were asked to
recognize complex emotions and mental states
from visual emotional stimuli. Tasks included
pictures of eyes (Baron-Cohen, Wheelwright,
Hill, Raste, & Plumb, 2001a; Baron-Cohen,
Wheelwright, & Jolliffe, 1997b; Baron-Cohen,
Wheelwright, Spong, Scahill, & Lawson, 2001c),
and static and dynamic facial expressions (Baron-
Cohen et al., 1997b; Golan, Baron-Cohen, & Hill,
2006b). Auditory tasks have used short utterances
(Golan, Baron-Cohen, Hill, & Rutherford, in
press; Kleinman, Marciano, & Ault, 2001; Ruther-
ford, Baron-Cohen, & Wheelwright, 2002). Other
tasks have assessed the ability to detect mental
states from contextual cues (Baron-Cohen,
O’Riordan, Stone, Jones, & Plaisted, 1999a;
Fein, Lucci, Braverman, & Waterhouse, 1992;
Happe, 1994). In all of these studies, individuals
with ASC performed at a significantly lower level
compared to controls, matched for age and IQ.
These results suggest that recognition of basic
emotions might be relatively preserved (or com-
pensated for) in individuals with ASC, but that
they show difficulties recognizing more complex
emotional and mental states.
The second question has received less empiri-
cal attention. As far as we are aware, only two
studies have assessed the ability of adults with
ASC to recognize emotions and mental states
from more ecological stimuli. ‘‘The Awkward
Moments Test’’ (Heavey, Phillips, Baron-Cohen,
& Rutter, 2000) presented participants with
seven short social situations, all taken from
television advertisements. They were asked to
judge the protagonist’s mental state at the end of
each scene. Participants with ASC performed at
a significantly lower level compared to general
population controls. In another study (Klin,
Jones, Schultz, Volkmar, & Cohen, 2002a,
2002b), a single social scene from a feature film
was presented to adults with ASC, while tracking
their gaze. Compared to typically developed
controls, participants with ASC looked less at
the eyes of characters, and more at characters’
mouths and surrounding objects, thus missing
socioemotional information pertinent to the
understanding of the social situation. Both of
these studies suggest high-functioning adults with
ASC are impaired on more ecological tests of
social understanding. However, both studies in-
volved tasks with a relatively small number of
scenes (seven in one, and one in the other),
which limits their validity and power and risks
floor or ceiling effects. ‘‘The Awkward Moments
Test’’ also used adverts, which are by definition
exaggerated. A more subtle collection of social
situations may better represent the complexity of
socioemotional interactions that we typically
encounter, and provide a more valid measure
of the ability of adults with ASC to interpret
them.
In this study, we report the ‘‘Reading the Mind
in Films’’ (RMF) task, which assesses complex
emotion and mental state recognition. We com-
pare the ability of adults with Asperger Syndrome
(AS) or High Functioning Autism (HFA) with
that of adults from the general population. The
task uses 22 short scenes, taken from feature films
(with permission). The scenes include visual input
(facial expressions, body language, action), audi-
tory input (prosody, verbal content) and context,
thus making the task more ecological than pre-
vious tasks testing each modality separately. The
emotions and mental states selected vary in
valence, intensity and complexity. They were
chosen for their relevance to everyday social
interaction. Hence, any recognition deficit could
impede one’s functioning in social situations. For
example, if the speaker fails to recognize the
listener’s embarrassment in a conversation, she or
he may not attempt to change the subject, and
may, without realizing it, offend or cause distress
in the listener (Grandin, 1995).
Judging complex emotions from ecological
social stimuli requires integration of multimodal
information into a coherent holistic picture as
well as attribution of mental states to others.
Therefore, based on the ToM or WCC models of
autism, we predicted that participants with ASC
would score significantly lower than controls on
the RMF task. In addition, we hypothesized that
the more autistic traits one possesses, the lower
one’s score on the task. In order to rule out that
difficulties in responding to the task are simply
due to working memory problems, we tested for
a correlation between task performance and
item length, with the number of characters
appearing in the item. We also tested for a
correlation between task score and age or IQ.
We hypothesized that a positive association
between task scores and verbal IQ would be
found for both groups, given that the task
involves matching a mental state word to a
character’s action. Lastly, we tested for correla-
tions between the RMF task score with other
READING THE MIND IN FILMS
115
complex emotion recognition tasks (from faces
and voices separately), to validate our new task.
Instruments
Reading the Mind in Films (RMF): Task
development
METHOD
The film clips are available at www.autism
researchcentre.com/tests. Thirty short scenes
(530 seconds long,
M
/
14.8,
SD
/
9.2), were
sampled from three feature films and one mini
series by two of the authors. The films were
picked for their dramatic value and frequent
occurrence of emotional scenes. The selected
scenes involved emotional interaction between
14 characters, and the expression of complex
emotions and mental states (e.g.,
smug
,
awkward
,
concerned
). In each scene, a protagonist was
identified and their emotion or mental state at
the end of the scene was labeled. Three foils were
selected for each item. In order to match the foils
and the target word for verbal difficulty, an
emotion taxonomy was used (Baron-Cohen, Go-
lan, Wheelwright, & Hill, 2004). The taxonomy
comprises 412 emotions and mental states, in six
developmental levels. Selected foils were either in
the same or easier levels than the target. Foils
were selected so that they matched some of the
emotional information in the scene but not all of
it, e.g., matching the content of the language but
not the intonation or the context. The labels and
foils were then reviewed by two independent
judges. A handout with definitions for all the
target and foil words in the items included was
prepared for participants’ use before and during
the task.
The items were then played to 15 adults (7 men
and 8 women) randomly selected from the gen-
eral population. They were played to them on a
laptop computer, using DMDX experimental
software (Forster & Forster, 2003). Two examples
of items from the task are shown in Figure 2.
Participants
The AS/HFA group comprised 22 adults (17
males and 5 females), aged 1752 (
M
/
29.0,
SD
/
9.8). Participants had all been diagnosed
with AS/HFA in specialist centers using estab-
lished criteria (American Psychiatric Association,
1994; World Health Organization, 1994). They
were recruited from a local clinic for adults with
ASC, and from a research volunteer database.
The control group comprised 22 adults (18 males
and 4 females) from the general population, aged
1851 (
M
/
25.4,
SD
/
9.5). They were recruited
from a local employment agency. All participants
were given the Wechsler Abbreviated Scale of
Intelligence (WASI), comprising the vocabulary,
similarities, block design and matrix reasoning
tests. The WASI produces verbal, performance
and full scale IQ scores, with correlations of .88,
.84 and .92 with the full Wechsler scales (Wechs-
ler, 1999). All participants scored above 85 on
both verbal and performance scales. Their IQ
scores are shown in Table 1. To screen for autism
spectrum conditions, participants also filled in the
Autism Spectrum Quotient (AQ; Baron-Cohen,
Wheelwright, Skinner, Martin, & Clubley, 2001b).
Replicating Baron-Cohen et al.’s (2001b) finding,
86.4% of the AS/HFA group and none of
the control group scored above the cut-off score
of 31. The two groups were matched on sex
(x
2
(1)
/
0.14,
ns
), age, verbal IQ, and perfor-
mance IQ. The groups’ background data appears
in Table 1.
TABLE 1
Means, standard deviations and ranges of AQ, chronological age and WASI scores for the AS/HFA group and the control group
AS/HFA group (n
/
22)
Control group (n
/
22)
Mean
SD
Range
Mean
SD
Range
t(42)
AQ
38.5
7.8
16 49
14.0
5.4
6 26
12.11**
Age
29.0
9.8
17.4 52.0
25.4
9.6
17.6 51.2
1.25
Verbal IQ
110.6
10.8
87 129
116.4
14.5
86 138
1.49
Performance IQ
114.8
12.5
97 140
113.1
8.4
92 129
0.51
Full scale IQ
114.2
10.8
91 130
116.5
11.0
97 138
0.71
Note
:**
p
B
/
.001. For all the other measures
p
/
.1.
[ Pobierz całość w formacie PDF ]