Вы находитесь на странице: 1из 18

AMERICAN JOURNAL ON INTELLECTUAL AND DEVELOPMENTAL DISABILITIES ÓAAIDD

2017, Vol. 122, No. 2, 138–155 DOI: 10.1352/1944-7558-122.2.138

Emotion Recognition in Children With Down Syndrome:


Influence of Emotion Label and Expression Intensity
Katie R. Cebula, Jennifer G. Wishart, Diane S. Willis, and Tom K. Pitcairn

Abstract
Some children with Down syndrome may experience difficulties in recognizing facial
emotions, particularly fear, but it is not clear why, nor how such skills can best be
facilitated. Using a photo-matching task, emotion recognition was tested in children with
Down syndrome, children with nonspecific intellectual disability and cognitively
matched, typically developing children (all groups N ¼ 21) under four conditions:
veridical vs. exaggerated emotions and emotion-labelling vs. generic task instructions. In
all groups, exaggerating emotions facilitated recognition accuracy and speed, with emotion
labelling facilitating recognition accuracy. Overall accuracy and speed did not differ in the
children with Down syndrome, although recognition of fear was poorer than in the
typically developing children and unrelated to emotion label use. Implications for
interventions are considered.
Key Words: Down syndrome; emotion recognition; emotion labelling

Within intellectual disability research, emotion A number of studies conducted to date have
recognition is an area that has received much shown that children and adults with Down
attention. This is perhaps unsurprising, as it is a syndrome are significantly less proficient in the
core skill in social interaction, long recognised as understanding of emotions than would be expect-
underpinning many other areas of learning (e.g., ed on the basis of chronological age, and they
Rogoff, 1990). Much of this work has focused on have also indicated difficulties in comparison to
autism spectrum conditions, but in recent years typically developing individuals of similar devel-
emotion recognition research has been informed opmental levels (Hippolyte, Barisnikov, Van der
by a behavioral phenotype approach, exploring Linden, & Detraux, 2009; Kasari, Freeman, &
and contrasting patterns of strengths and difficul- Hughes, 2001; Pochon & Declercq, 2014; Virji-
Babul, Watt, Nathoo, & Johnson, 2012; Way &
ties across a range of specific aetiologies (e.g.,
Rojahn, 2012; Williams, Wishart, Pitcairn, &
Porter, Coltheart, & Langdon, 2007; Whittington
Willis, 2005; Wishart & Pitcairn, 2000, but see
& Holland, 2011). Such an approach, if combined
Celani, Battacchi, & Arcidiacono, 1999; Channell,
with a developmental perspective, has the poten- Conners, & Barth, 2014; Pochon & Declercq,
tial to inform the design of early and specifically 2013, for exceptions). A number of these studies
targeted interventions (Fidler & Nadel, 2007). Yet explored six basic emotions (happiness, sadness,
in order to maximise intervention utility it is anger, surprise, fear, disgust), and several reported
important to establish precisely which aspects of difficulties in relation to specific expressions, often
emotion recognition present difficulties for chil- that of fear (Kasari et al., 2001; Virji-Babul et al.,
dren with specific aetiologies, and which specific 2012; Williams et al., 2005; Wishart & Pitcairn,
techniques will best facilitate emotion recognition. 2000); error patterns with respect to fear recogni-
The present study explored these issues in relation tion were also found to be developmentally
to children with Down syndrome. atypical (Williams et al., 2005). Fear recognition

138 Emotion Recognition and Down Syndrome


AMERICAN JOURNAL ON INTELLECTUAL AND DEVELOPMENTAL DISABILITIES ÓAAIDD
2017, Vol. 122, No. 2, 138–155 DOI: 10.1352/1944-7558-122.2.138

difficulties have not always been found in all evidence of this: in comparison to younger MA-
emotional understanding tasks, nor have they matched typically developing children, Pochon
always been shown to be syndrome-specific (see and Declerq (2013, 2014) have reported difficul-
Cebula & Wishart, 2008 for review). Difficulties ties amongst children with Down syndrome in an
with other emotions have also been reported, emotion recognition task employing emotion
albeit less consistently (e.g., Kasari, et al., 2001; labels, but no significant difficulties in an emotion
Wishart & Pitcairn, 2000). Taken together though, vocalisation-to-face matching task which did not
the findings of these studies do suggest that use emotion labels. They concluded that children
emotion recognition in general, and fear recogni- with Down syndrome may have a specific
tion in particular, may be an area of concern with emotional lexicon deficit, rather than a difficulty
some children with Down syndrome. in recognizing emotional expressions per se,
A variety of methods have already been although it is unclear whether this was the case
developed to support children with developmen- across all of the emotions they studied. Such a
tal disabilities in their understanding of emotions difficulty would, though, fit with evidence noted
(e.g., Baron-Cohen, Golan, Wheelwright, & Hill, by Kasari et al. (2001) that children with Down
2004; Golan et al., 2010). However, children with syndrome are exposed to less conversation about
Down syndrome may show a different profile of emotional terminology than are typically devel-
emotional understanding skills from that found in oping children (Tingley, Gleason, & Hooshyar,
other developmental disabilities (e.g., Celani et al., 1994); in particular, Kasari et al. suggested that,
1999), and there may also be differences in the because young children with Down syndrome are
origin of any difficulties. To develop targeted often perceived to be of a friendly and happy
interventions, it is therefore important to establish disposition, caregivers may use fewer negative
in more detail which specific components of emotion words with them, providing children
emotion recognition are problematic for children with fewer opportunities to learn these particular
with Down syndrome. emotion labels.
One difficulty here is that emotion recogni- Rather than difficulties with linguistic emo-
tion tasks tend to conflate semantic knowledge of tion labels, distinguishing and decoding the facial
emotion labels with perceptual face processing emotion stimuli themselves may be problematic
ability (Lindquist & Gendron, 2013; Vicari, for individuals with Down syndrome. This
Reilly, Pasqualetti, Vizzotto, & Caltagirone, possibility is supported by the finding that
2000), often using emotion labels in verbal task emotion recognition difficulties have been report-
instructions (happy, sad, etc.), in combination ed even in tasks that have been designed to
with facial expressions of emotion (e.g., photo- minimize language demands (e.g., Williams et al.,
graphs, schematic drawings). As a result, it is 2005). It is also supported by findings of
sometimes unclear whether the verbal emotion differences in the emotional expressions experi-
labels, the facial stimuli, or both are problematic enced in interactions between parents and their
for study participants. children with Down syndrome from early in life
Language plays a central role in the under- (e.g., Carvajal & Iglesias, 1997).
standing of emotions (Lindquist & Gendron, One way in which the influence of the facial
2013; Reed & Steed, 2015), and in typical emotions themselves can be explored is to
development emotion labels are used from early manipulate the intensity of expressions in tasks,
childhood (Lindquist & Gendron, 2013; Ridge- and assess whether recognition improves with
way, Waters & Kuczaj, 1985). Some labels, such as increased intensity. Most of the studies to date
happy, sad, and angry are used accurately earlier that have explored the effects of emotion intensity
than others, such as fear and disgust (Székely et al., on facial expression recognition have been with
2011; Vicari et al., 2000; Widen & Russell, 2003), typically developing participants (Calder, Young,
but accurate use of emotion labels continues to Rowland, & Perrett, 1997; Gao & Maurer, 2009;
develop throughout the school years (Widen & Herba, Landau, Russell, Ecker, & Phillips., 2006).
Russell, 2008; Vicari et al., 2000). These have often used computer-manipulated
It is possible that the emotion recognition facial expressions of emotion and results generally
difficulty found in children with Down syndrome show that more intense or exaggerated facial
relates specifically to difficulties with the emotion expressions of emotion are recognized more
labels used in tasks. Indeed, there is some recent quickly, and with equivalent or increased accura-

K. R. Cebula et al. 139


AMERICAN JOURNAL ON INTELLECTUAL AND DEVELOPMENTAL DISABILITIES ÓAAIDD
2017, Vol. 122, No. 2, 138–155 DOI: 10.1352/1944-7558-122.2.138

cy, than less intense expressions. However, the size clear prediction on whether such a difficulty
of this effect can vary with emotion (Herba et al., would be predominantly perceptual or linguistic.
2006). Exaggerated emotions might result in more
accurate recognition because fine detail is more Method
easily detected (Calder et al., 1997; Guo & Maurer,
2009), and may attract greater attentional resources Participants
(Kumfor et al., 2011). Increased intensity of Eighty-one children participated in the study: 28
emotion has also been shown to lead to increased with Down syndrome (10 male, 18 female; 9;03–
neural response. For example, Morris et al. (1996, 18;09 years), 25 with nonspecific intellectual
1998) found that increased intensity of fearful disability (11 male, 14 female; 9;01–17;05 years)
expressions led to heightened amygdala response and 28 typically developing children (14 male, 14
in typically developing participants. Although female; 3;04–6;06 years). Children participated
there has been some work with atypical child from six mainstream schools, four special schools,
and one mainstream nursery. The majority of
populations using computer manipulated facial
children with intellectual disability were in special
emotion stimuli (e.g., Blair, Colledge, Murray, &
school placements, reflecting educational place-
Mitchell, 2001; Castelli, 2005), the effects of
ment patterns in the region in which the study was
intensity manipulation have not, to our knowl- carried out.
edge, previously been examined in children with Parents were sent letters about the study and
Down syndrome. consent forms. These went to families already on a
These two components of emotion recogni- database of voluntary study participants held by
tion—perceptual (facial expression recognition) the research team, and also via local mainstream
and linguistic (emotion label understanding)— and special schools where the head teachers
may be closely related, but follow very different identified pupils who met the inclusion criteria.
developmental pathways (Vicari et al., 2000). Response rates were high, but because researchers
Although it is unlikely to be possible to disentan- were not always informed of the number of letters
gle them completely, a better understanding of sent by the school, precise response rates cannot
how these two elements contribute towards any be reported.
difficulties experienced by children with Down Diagnosis of the children in the Down
syndrome might begin to pave the way for the syndrome and nonspecific intellectual disability
development of targeted interventions. groups was confirmed by cross-checking local
The present study explored emotion recognition educational and central health records. The
in children with Down syndrome, in comparison to nonspecific intellectual disability group was
children with nonspecific intellectual disability and composed of children for whom etiology of
typically developing children matched for develop- intellectual disabilities was unknown. Children
mental age. It aimed to establish: with identified genetic syndromes, a family
history of intellectual disability or with known
 the extent to which exaggeration of facial neurological insult were excluded. In addition,
expressions and the use of emotion labels children were excluded from all three participant
effect emotion recognition accuracy and speed groups if they had attention deficit hyperactivity
 whether any differences exist across groups in disorder, severe sensory or physical impairment,
ability to recognise specific emotions and, if severe behavioral difficulties, or were on the
so, whether these difficulties are predominant- autism spectrum. In the case of the two
ly perceptual or linguistic intellectual disability groups an additional exclu-
sion criterion was profound or complex intellec-
It was hypothesized that exaggerated emotions and tual disability as task demands were likely to be
emotion labels would improve the overall accura- beyond the capabilities of these children.
cy and speed of emotion recognition in all three
groups of children, but not necessarily equally. In Design and Procedure
line with previous research, it was hypothesized Assessment measures. Cognitive ability was
that children with Down syndrome would expe- assessed using a four-subtest short-form of the
rience particular difficulties in fear recognition, Wechsler Preschool and Primary Scales of Intelli-
although the literature to date does not allow for a gence-Revised (WPPSI-R: Wechsler, 1990). This

140 Emotion Recognition and Down Syndrome


AMERICAN JOURNAL ON INTELLECTUAL AND DEVELOPMENTAL DISABILITIES ÓAAIDD
2017, Vol. 122, No. 2, 138–155 DOI: 10.1352/1944-7558-122.2.138

short-form consisted of two subtests from the series (FEEST: Young, Perrett, Calder, Sprengel-
original performance and verbal scales and has meyer, & Ekman, 2002). These were the same
shown high reliability coefficients and high photographs as those used for the veridical trials,
validity in terms of correlation with full test scores but digitally altered so that the emotional
(rtt ¼ .932, r ¼ .914: Sattler, 1992). Vocabulary expressions appear exaggerated in comparison to
comprehension was measured using the British the veridical Ekman and Friesen photographs.
Picture Vocabulary Scale (BPVS-II: Dunn, Dunn, This digital alteration enhances the emotional
Whetton, & Burley, 1997). Children were also intensity of the expression displayed (Calder et al.,
administered the Benton Facial Recognition Test 2000). Various levels of exaggeration are available,
(Benton, Sivan, Hamsher, Varney & Spreen, but as this study was the first to utilize these
1983). This is a standardized procedure for images with children with Down syndrome, the
assessing the ability to identify unfamiliar human largest degree of exaggeration (175%) was used to
faces. Participants are presented with a single black ensure that any effects were detectable. These
and white photograph of a face and then exaggerated FEEST photographs are only available
instructed to locate that face in a display of six for one male and one female model from the
photographs, where it appeared either one or three Ekman and Friesen series. Therefore both the
times. The 27-item short-form of this test was used veridical and the exaggerated trials used photo-
on which a score of 11 might be expected on the graphs of only these models.
basis of chance alone (Benton, Sivan, Hamsher, & Altogether there were 48 trials—presented as
Spreen, 1994). 24 within each of two test sessions. Each set of 24
Emotion-matching task. This experimental trials consisted of 12 presented as veridical
task required participants to match photographs of emotions and 12 exaggerated emotions. In each set
unfamiliar adults on the basis of facial emotion. It of 12 trials each of the six basic emotions was
used photographs from Ekman and Friesen’s presented as the target twice. Exactly the same 24
Facial Affect Slides (1976), a series of 11.5 3 trials were used in each of the two test sessions.
17.5 cm black and white photographs of men and However, in one test session generic task
women expressing the six basic emotions (happi- instructions were used (the children were shown
ness, sadness, anger, surprise, fear, and disgust) and the target photo and asked: ‘‘Who feels the same
a set of neutral poses. These photographs have as this person?’’). In the other test session
been used extensively in research with typical and emotion labelling task instructions were used
atypical children and adults, and have high inter- (the children were shown the target photo and
rater agreement regarding the emotion displayed told: ‘‘This person feels [e.g.] happy. Who feels
(Ekman & Friesen, 1976). Children were shown the same as this person?’’). These sessions were
one facial photograph, and were then asked to presented in counterbalanced order, and for each
point to a photograph of a different person child, trial order was identical in each session.
showing the same emotion from a choice of three There were therefore four conditions in total:
photographs (the correct emotion and two dis- veridical emotions with generic task instructions;
tracters) that were then uncovered underneath. exaggerated emotions with generic task instruc-
The photographs were all mounted on a single tions; veridical emotions with emotion labelling
piece of cardboard. The photographs selected task instructions; exaggerated emotions with
from the Ekman and Friesen series for this task emotion labelling task instructions.
consisted of one male and one female face In any trial, all the photographs (i.e., the
displaying each of the six basic emotions. Where initial photograph and the three choice photo-
the male model was shown in the initial graphs) were either veridical or exaggerated. The
photograph to be matched, the female model emotions used in the distracter photographs were
was shown in the three choice photographs and identical for each of the corresponding veridical
vice versa. Position of correct answer was counter- and exaggerated trials, as was the position of the
balanced across trials. correct answer: the trials therefore only differed in
Two intensity levels of emotion were used in one respect, the intensity of the emotions
this task: veridical and exaggerated. The veridical depicted. The 24 veridical and exaggerated trials
trials used the Ekman and Friesen faces. The were presented in a fixed random order, with the
exaggerated trials used photos taken from the constraints that: the first and last trials were the
Facial Expressions of Emotion—Stimuli and Tests veridical happy expressions (in the case of the first

K. R. Cebula et al. 141


AMERICAN JOURNAL ON INTELLECTUAL AND DEVELOPMENTAL DISABILITIES ÓAAIDD
2017, Vol. 122, No. 2, 138–155 DOI: 10.1352/1944-7558-122.2.138

trial, to ensure that the child could easily highest average level of viewer agreement on the
understand the task, and in the case of the last intended emotions, with agreement level for any
trial to ensure that they ended the session with a single emotion not less than 88%.
positive rather than negative emotion); the same Matching procedures. Following previous
emotion was never presented twice in a row; no studies in this field, groups were matched on the
more than two trials in a row depicted the same basis of MA. From the original participant pool of
level of intensity (100 or 175%) or the same gender 81, it was possible to MA-match 21 children from
in the target photograph. The order of trials was each participant group on the basis of their
therefore fixed, but half of the children in each Wechsler scores. Each child from the nonspecific
group were presented with them in reverse order in intellectual disability and the typically developing
case fatigue caused less accurate responding in groups was individually pair-wise matched as
later trials. closely as possible to a child from the Down
At the beginning of each test session children syndrome group (with no MA difference between
were asked to place their hands on a horizontal red individual children greater than 10 months and
line on the table, and reminded to return their across all individually paired children a mean
hands to this position at the beginning of each difference in MA of 4.3 months). Table 1 gives a
trial (this was so that response time could be summary of characteristics in the three groups.
measured as accurately as possible). To avoid Mean MA—a composite of performance mental
confounding expressive language and memory age (PMA) and verbal mental age (VMA)—did not
difficulties with any emotion-specific deficits, significantly differ across groups (around 4 years in
children were required to indicate their answer all groups) and ranges were also similar. There were
only by pointing, and all four photographs also no significant group differences in PMA,
remained on the table to ensure that emotion VMA, vocabulary comprehension or gender
recognition rather than recognition memory was balance. There were significant differences in
being tested. Children were presented with the Benton scores and chronological age. Although
task by one experimenter and responses recorded there were no differences between the two
by another experimenter who sat unobtrusively in intellectual disability groups in terms of Benton
the corner of the room. As recommended by scores, the nonspecific intellectual disability group
Herba and Phillips (2004) and De Sonneville et al. did obtain significantly higher scores than the
(2002) response speed data were collected in typically developing group (Posthoc Scheffe p ,
addition to accuracy data, and all sessions were .05). In addition, the Down syndrome and
filmed for later analysis of response times. nonspecific intellectual disability group did not
Emotion label production. This task was differ in chronological age but, as would be
presented around one week after the final expected, both groups were significantly older
emotion-matching task session, and it again used than the developmentally matched typically
photographs from the Ekman and Friesen series. developing group (Posthoc Scheffe p , .001).
Children were shown a photograph from this The age range in the typically developing group
series of a woman showing a neutral expression was also much smaller than in the other two
and told ‘‘This is Mary.’’ They were then shown a groups.
photograph of the same woman with a happy Procedure. Ethical consent was obtained
expression and asked ‘‘And now she feels. . .?’’ The from the lead author’s University ethics commit-
children were then shown photographs of the tee, and participants were only included once
same woman displaying sadness, anger, surprise, school and parental permission had been granted.
fear, and disgust, in a random order, and for each Parents were provided with an information sheet
one, the same question repeated. There was one and consent form, and were also asked to explain
stimuli photo per emotion. This was a free-choice the study to their child, using child-friendly
response task: no list of possible responses was information and consent sheets where they judged
provided. If children did not provide a response, these to be appropriate. Child assent was recon-
the prompt ‘‘Mary feels. . .?’’ was provided. Child- firmed at each testing session, either verbally or
ren’s responses were noted, but incorrect responses using pictorial methods. Children were tested in a
were not corrected. The model chosen from the quiet room in their school, over a number of
Ekman series was the one which Ekman and sessions. Length and number of sessions varied,
Friesen’s data showed to have produced the depending on the child’s availability and levels of

142 Emotion Recognition and Down Syndrome


K. R. Cebula et al.
2017, Vol. 122, No. 2, 138–155

Table 1
Characteristics of Mental Age-Matched Groups
Group (N ¼ 21 per group)
Down syndrome Nonspecific intellectual disability Typically developing ANOVA values
Variable Mean (SD) Range Mean (SD) Range Mean (SD) Range F (2, 60) p value
CA 14.76 (25.3) 10.92–18.75 13.35 (35.3) 9.08–17.42 4.51 (9.4) 3.33–5.75 141.72 .000***
MA 4.27 (6.1) 3.75–5.75 4.56 (9.8) 3.08–5.92 4.68 (7.4) 3.58–5.50 2.15 .125
PMA 4.48 (7.0) 3.25–5.75 4.67 (10.2) 3.17–6.25 4.98 (9.04) 3.92–6.50 2.44 .096
VMA 4.07 (7.2) 3.42–5.75 4.48 (11.2) 3.00–6.00 4.41 (8.32) 3.00–5.75 1.74 .185
AMERICAN JOURNAL ON INTELLECTUAL AND DEVELOPMENTAL DISABILITIES

Vocabulary comprehension 5.26 (21.3) 2.75–8.00 5.83 (19.7) 2.83–8.17 5.42 (18.5) 2.00 –7.58 0.66 .523
Benton Facial 14.0 (2.7) 9–19 16.1 (3.3) 1122 13.7 (2.0) 11–17 4.74 .012*
Recognition Test
N male/female 8/13 9/12 12/9 v2(2) ¼ 1.66
Note. With the exception of the Benton Facial Recognition Test all test scores are reported as age equivalents in years and SDs in months. CA ¼ chronological age; MA ¼
mental age; PMA ¼ performance mental age; VMA ¼ verbal mental age.
*p , .05; ***p , 0.001.
DOI: 10.1352/1944-7558-122.2.138
ÓAAIDD

143
AMERICAN JOURNAL ON INTELLECTUAL AND DEVELOPMENTAL DISABILITIES ÓAAIDD
2017, Vol. 122, No. 2, 138–155 DOI: 10.1352/1944-7558-122.2.138

fatigue, but children received some standardised Within each participant group total accuracy
and some experimental tasks in each session. The scores and mean response times did not differ as
two photo-matching task sessions were always a function of either participant gender or order of
given at least one week apart. presentation of conditions. Data were therefore
Analysis. Accuracy data was manually record- collapsed in these respects.
ed for each of the experimental tasks. In addition, Accuracy data. Table 2 shows the total
response time for each emotion-matching trial was accuracy scores for the three MA-matched
measured from the film clips of test sessions using groups, as well as these scores broken down
the Observer Video-Pro, a software package which according to condition.
allows frame-by-frame analysis of events, and Scores were analyzed using a 3 (participant
hence a high level of accuracy in data analysis group) x 2 (intensity condition) x 2 (labelling
(Noldus, Trienes, Hendriksen, Jansen, & Jansen, condition) mixed model ANOVA. This showed
2000). Coding of response time was conducted by significant main effects of group, F(2, 60) ¼ 3.26, p
the first author. For each trial, timing of response , .05, g2p ¼ .10, of intensity, F(1, 60), ¼ 9.48, p
began when the card covering the three choice ,.01, g2p ¼ .14, and of labelling, F(1, 60) ¼ 18.79,
photographs was removed and stopped when the
p , .001, g2p ¼ .24, but no significant interaction
child’s finger touched a photograph to indicate
effects amongst group, intensity and labelling.
their answer. To assess inter-rater agreement of
Post-hoc analysis of the group effect showed
response time coding, 12 sessions from the full
that the typically developing children were more
participant pool were re-coded by the third
author. The mean kappa value was 0.95. The accurate overall on this task than the two
emotion production task was scored as follows: a intellectual disability groups, but that this differ-
score of one point was given for each emotion ence was only significant in relation to the
correctly identified (i.e., a maximum possible score children with nonspecific intellectual disability
of 6). The following emotion terms were classed as (Scheffe test p , .05). In relation to emotion
correct: happy; sad; angry or cross; surprised; fearful intensity, children were significantly more accurate
or scared; disgusted or yuck. overall in emotion recognition when the stimuli
were of greater intensity. In relation to emotion
labelling, children were significantly more accurate
Results overall in emotion recognition when the emotions
Emotion-Matching Task were labeled.
This task required participants to match photo- Response time data. Table 3 shows the total
graphs of unfamiliar adults on the basis of facial mean response time data for the three MA-
expressions. Emotions were either veridical or matched groups, as well as these scores broken
exaggerated, and either labelled or nonlabelled. down according to condition.

Table 2
Emotion-Matching Task Accuracy Scores
Group (N ¼ 21 per group)
Down syndrome Nonspecific intellectual disability Typically developing
Trials Mean (SD) Mean (SD) Mean (SD)
veridical emotion 16.86 (4.17) 15.10 (5.00) 18.43 (4.31)
exaggerated emotion 17.48 (3.70) 15.81 (5.88) 19.43 (3.70)
nonlabelled emotion 15.81 (4.47) 15.10 (5.20) 17.52 (4.85)
labelled emotion 18.52 (3.75) 15.81 (6.17) 20.33 (3.99)
Total score† 34.33 (7.59) 30.90a (10.70) 37.86a (7.84)
Note. Maximum score ¼ 24 for each condition and 48 for total score. Score achievable by chance alone ¼ 8 for each
condition and 16 for total score.
a
significant difference across groups, p , .05. †Total scores are not the sum of the four conditions, because the four
conditions were not separately presented. Rather, the emotion presented in each trial was either veridical or exaggerated and
either labelled or nonlabelled.

144 Emotion Recognition and Down Syndrome


AMERICAN JOURNAL ON INTELLECTUAL AND DEVELOPMENTAL DISABILITIES ÓAAIDD
2017, Vol. 122, No. 2, 138–155 DOI: 10.1352/1944-7558-122.2.138

Table 3
Emotion-Matching Task Average Response Times (Seconds)—All Trials (Correct and Incorrect)
Group (N ¼ 20 per group†)
Down syndrome Nonspecific intellectual disability Typically developing
Trials Mean (SD) Mean (SD) Mean (SD)
veridical emotion 2.93 (0.92) 3.18 (2.02) 3.95 (2.00)
exaggerated emotion 2.59 (0.78) 2.78 (1.82) 3.45 (1.68)
nonlabelled emotion 2.69 (1.18) 3.05 (2.14) 3.50 (1.56)
labelled emotion 2.83 (0.81) 2.91 (1.99) 3.90 (2.64)
Mean (SD) response time 2.76 (0.83) 2.98 (1.90) 3.70 (1.82)

Response times for three participants (one from each group) were omitted from analysis, as they were . 3 SDs above the
group mean. Groups did not differ significantly on mental age or vocabulary comprehension levels once these participants
had been removed.

Scores were analyzed using a 3 (participant responded correctly. The pattern of significant
group) x 2 (intensity condition) x 2 (labelling and non-significant results was unchanged.
condition) mixed model ANOVA. This showed Individual emotion accuracy data. Table 4
no significant main effect of group, F(2, 57) ¼ shows accuracy scores for the six individual
1.90, ns, nor of labelling, F(1, 57) ¼ 0.36, ns, but a emotions (conditions collapsed). To establish
significant effect of intensity, F(1, 57), ¼ 33.91, p whether groups differed in their ability to
, .001, g2p ¼ .37. There were no significant recognize the six individual emotions, scores were
interaction effects. The groups did not therefore analyzed using a 3 (participant group) x 6
(emotion) ANOVA. This showed significant main
significantly differ from each other in how quickly
effects of group, F (2,60) ¼ 3.26, p , .05, g2p ¼ .10
they gave an answer. Nor did the use of emotion
and emotion, F (5, 300) ¼ 44.85, p , .001, g2p ¼
labelling facilitate or impede speed of perfor- .43 and a significant interaction between group
mance. However, on the whole the children were and emotion, F (10, 300) ¼ 2.89, p , .01, g2p ¼
significantly faster at recognizing the exaggerated .09. When individual emotions were analyzed it
emotions compared to the veridical ones. To was found that the Down syndrome group was
control for the possibility that some children may significantly less accurate than the typically
have been sacrificing accuracy for speed, the developing group in the recognition of fear, F (2,
analysis was rerun, including only the response 60) ¼ 4.02; p , .05, g2p ¼ .12; post-hoc Scheffe p
times for trials in which the children had , .05. In addition, the nonspecific intellectual

Table 4
Total Accuracy Scores for Individual Emotions (Conditions Collapsed)
Group (N ¼ 21 per group)
Down syndrome Nonspecific intellectual disability Typically developing
Mean (SD) Mean (SD) Mean (SD)
Happiness 7.29 (1.06) 6.38 (1.80) 7.24 (1.18)
Sadness 6.43 (1.69) 5.67 (2.69) 6.24 (1.67)
Anger 3.67 (2.33) 2.76a (1.97) 5.14a (2.56)
Surprise 6.62 (1.53) 6.00 (2.15) 6.67 (1.65)
Fear 3.71b (2.31) 4.52 (2.34) 5.62b (1.88)
Disgust 6.62 (1.69) 5.57 (2.42) 6.95 (1.63)
Note. Maximum possible score ¼ 8; score achievable by chance ¼ 2.67.
a
significant nonspecific intellectual disability-typically developing group difference in anger recognition, p , .01;
b
significant Down syndrome-typically developing group difference in fear recognition, p , .05.

K. R. Cebula et al. 145


AMERICAN JOURNAL ON INTELLECTUAL AND DEVELOPMENTAL DISABILITIES ÓAAIDD
2017, Vol. 122, No. 2, 138–155 DOI: 10.1352/1944-7558-122.2.138

disability group was significantly less accurate than 2.33 (1.46) for the nonspecific intellectual disability
the typically developing group in the recognition and typically developing groups respectively, t(40)
of anger, F (2, 60) ¼ 5.74 p , .01; post-hoc Scheffe ¼ 2.80, p , .01. The nonspecific intellectual
p , .01. disability group was less accurate than the typically
To explore the effect of increased intensity developing group when the exaggerated images
and emotion labelling on recognition of these two were used: M(SD) ¼ 1.62 (1.16) and 2.62 (1.43) for
emotions they were analyzed individually. As the the nonspecific intellectual disability and typically
differences in fear recognition ability lay between developing groups respectively, t(40) ¼ 2.49, p ,
the Down syndrome and typically developing .05. They also performed less accurately when the
groups further analysis focused only on these two emotion was veridical: M(SD) ¼ 1.14 (1.11) and
groups. This showed that the Down syndrome 2.52 (1.29) for the nonspecific intellectual disability
group was less accurate in fear recognition than the and typically developing groups respectively, t(40)
typically developing group when emotion label- ¼ 3.72, p , .01.
ling was used: M(SD) ¼ 2.05 (1.36) and 3.05 (1.16) Again, anger recognition scores in the non-
for the Down syndrome and typically developing specific intellectual disability group were not
groups respectively, t(40) ¼ 2.56, p , .05. They significantly different from what would be expect-
were also less accurate when emotion labelling was ed on the basis of chance responding alone, t(20) ¼
not used: M(SD) ¼ 1.67 (1.20) and 2.57 (1.29) for 0.21, ns. In terms of facilitation of anger
the Down syndrome and typically developing recognition, paired t-tests showed that the non-
groups respectively, t(40) ¼ 2.36, p , .05. The specific intellectual disability group showed no
Down syndrome group was less accurate in fear significant improvement in anger recognition
recognition than the typically developing group performance if labels were used, t(20) ¼ 1.40,
when the exaggerated images were used: M(SD) ¼ ns. Similarly, they showed no significant improve-
1.71 (1.23) and 2.86 (1.11) for the Down ment in performance when exaggerated emotions
syndrome and typically developing groups respec- were used, t(20) ¼ 1.94, ns.
tively, t(40) ¼ 3.16, p , .01. They were also less
accurate with the veridical emotions, although the Emotion Label Production Task
difference was non-significant: M(SD) ¼ 2.00
In this task children were asked to label each of the
(1.38) and 2.76 (1.09) for the Down syndrome
six emotions from photographs. The mean
and typically developing groups respectively, t(40)
number of emotions correctly labeled were 3.45
¼ 1.99, p ¼ .054.
(1.10), 2.95 (1.12) and 3.62 (1.07) for the Down
Fear recognition scores in the Down syn-
syndrome, nonspecific intellectual disability and
drome group were in fact not significantly
typically developing groups respectively. Scores
different from what would be expected on the
from this task were analyzed using a 3 (participant
basis of chance responding alone, t(20) ¼ 2.08, p ¼
group) x 6 (emotion) mixed model ANOVA. This
0.051. In terms of facilitation of fear recognition,
showed no significant main effect of participant
paired t-tests showed that the Down syndrome
group, F (2,60) ¼ 2.09, ns, a significant effect of
group showed no significant improvement in fear
recognition performance if labels were used, t(20) emotion, F (5, 300) ¼ 63.12, p , .001 g2p ¼ .51
¼ 1.56, ns. Similarly, they showed no significant and no significant interaction, F (10, 300) ¼ 0.32,
improvement in performance when exaggerated ns. The groups did not therefore differ in their
emotions were used, t(20) ¼ 1.06, ns. overall ability to label the emotions, although
The difference in anger recognition between some emotions were significantly more likely to be
nonspecific intellectual disability and typically labelled correctly (happy was easier than all other
developing groups was then explored. It was found emotions; sad was easier than all other emotions
that the nonspecific intellectual disability group except for happy and angry; angry was easier than
was less accurate than the typically developing surprise, fear and disgust: p  .001 in all cases).
group in anger recognition when emotion labelling
was used: M(SD) ¼ 1.57 (1.17) and 2.81 (1.47) for Relationships Between Assessment
the nonspecific intellectual disability and typically Measures and Experimental Task
developing groups respectively, t(40) ¼ 3.02, p , Performance
.01. They were also less accurate when emotion A series of correlations were carried out within
labelling was not used: M(SD) ¼ 1.19 (1.17) and each of the MA-matched groups using all available

146 Emotion Recognition and Down Syndrome


AMERICAN JOURNAL ON INTELLECTUAL AND DEVELOPMENTAL DISABILITIES ÓAAIDD
2017, Vol. 122, No. 2, 138–155 DOI: 10.1352/1944-7558-122.2.138

child data (Table 5). This showed that emotion- and accuracy of emotion matching overall.
matching accuracy scores correlated significantly However, in relation to specific emotions, they
with mental age, verbal mental age and vocabulary were significantly less accurate than the typically
comprehension levels in all three participant developing children in the recognition of fear.
groups. A correlation between accuracy scores Accuracy scores of children with Down syndrome
and chronological age was unique to the typically for fear were not significantly different from
developing group. Accuracy scores correlated with chance. These fear recognition difficulties were
performance mental age and with Benton test not obviously linguistic in nature, as they were still
scores in the nonspecific intellectual disability and evident when emotion labels were absent from
Down syndrome groups. Emotion-matching accu- task instructions. Recognition of fear was facilitat-
racy scores did not correlate with emotion label ed neither by the use of emotion labels nor by the
production in any of the three groups. use of exaggerated expressions.
Children with nonspecific intellectual dis-
Discussion ability were significantly less accurate than the
typically developing children in emotion recog-
The present study explored the effect of exag- nition in general, though in contrast, significantly
gerated facial emotions and of emotion labelling
more accurate in the Benton face recognition
on the speed and accuracy of emotion recogni-
task. This latter finding perhaps points to the role
tion in children with Down syndrome, children
of experience in face recognition (though for
with nonspecific intellectual disability, and
recent discussion see McKone, Crookes, Jeffery,
typically developing children matched on cog-
nitive ability. Across the participant groups as a & Dilks, 2012). Specifically, within emotion
whole, the children were more accurate in the recognition they were significantly less accurate
recognition of emotions when they were more in the recognition of anger, again with recogni-
intense and when they were labeled, supporting tion abilities not significantly different from
previous findings with typically developing chance. Indeed, participants across all groups
children (e.g., Herba et al., 2006; Russell & found anger difficult to recognize. This may have
Widen, 2002). Children overall also recognized been because the target and response anger
exaggerated emotional expressions more quickly, stimuli were more dissimilar than in the case of
although the addition of emotion labels did not the other emotions (a closed and an open-mouth
increase their speed of recognition. expression of anger). In all three groups emotion
The children with Down syndrome did not recognition accuracy was correlated with mental
differ significantly from the other groups in their age and vocabulary comprehension, but not with
ability to label basic emotions, or in their speed emotion labelling ability.

Table 5
Correlations Between Emotion-Matching Score and Assessment Measures, and Pretask Emotion Label
Production
Emotion-matching task score
Group (N ¼ 21 per group)
Variable Down syndrome Nonspecific intellectual disability Typically developing
CA 0.11 -0.36 0.47*
MA 0.54* 0.78** 0.53*
PMA 0.46* 0.75** 0.41
VMA 0.45* 0.67** 0.46*
Benton 0.44* 0.73** 0.17
Vocabulary comprehension 0.75** 0.60** 0.64**
Emotion label production 0.38 0.42 0.24
Note. CA ¼ chronological age; MA ¼ mental age; PMA ¼ performance mental age; VMA ¼ verbal mental age.
*p , .05; **p , .01.

K. R. Cebula et al. 147


AMERICAN JOURNAL ON INTELLECTUAL AND DEVELOPMENTAL DISABILITIES ÓAAIDD
2017, Vol. 122, No. 2, 138–155 DOI: 10.1352/1944-7558-122.2.138

Other research studies have found greater (Kasari et al., 2001; Williams et al., 2005; Wishart
evidence of difficulties in emotional understand- & Pitcairn, 2000). However, it is worth noting that
ing in Down syndrome (e.g., Kasari et al., 2001; a difficulty in fear recognition was found only in
Williams et al., 2005; Wishart et al., 2007). The comparison to the typically developing group, not
finding here that emotion recognition in general in comparison to the nonspecific intellectual
was not an area of difficulty beyond what would disability group, suggesting that this difficulty
be expected given their mental age, is very may not be strongly syndrome-specific (see also
encouraging. Indeed, it was notable that the Turk & Cornish, 1998; Wishart et al., 2007). A
performance of the Down syndrome and the larger scale study would be helpful in confirming
typically developing groups in this study were, this.
overall, marked more by similarity than difference. The fear recognition difficulty did not seem to
This pattern of findings is in concordance with stem from any linguistic problems in understand-
findings from some previous studies (e.g., Chan- ing and using the emotion label scared: participant
nell et al., 2014; Pochon & Declerq, 2013). The groups did not differ significantly on verbal
reasons for the discrepancy in findings across comprehension ability or in their ability to label
different studies are unclear, but may relate to a photograph of a fearful face. The children with
differences in participants, stimuli or tasks (Chan- Down syndrome moreover remained significantly
nell et al., 2014). For example, participants in the less accurate than the typically developing chil-
present study had higher mean language ability dren even in the condition in which instructions
than those in previous studies with contrasting included no emotion label. This pattern of
findings (though different language measures findings appears to contradict those of Pochon
across studies makes comparison difficult), and and Declercq (2013, 2014), who reported that the
the present study used different tasks and stimuli emotion recognition difficulties found in children
than in some previous studies (e.g., Kasari et al., with Down syndrome stem primarily from prob-
2001 used a puppet paradigm). lems with emotion labels, rather than from a
In relation to the current study, it must be difficulty in recognizing facial expressions per se.
borne in mind that the three participant groups However, emotion labels are more important in
were matched on cognitive ability. Given that the recognition of some emotions than others
overall emotion matching scores were not yet at (Widen & Russell, 2004) and it is possible that this
ceiling in the much younger typically developing apparent contradiction in findings stems from a
children, differences in emotion recognition difference between our own study and that of
abilities might have been found had the children Pochon and Declerq in the extent to which
with Down syndrome been compared with peers analysis was conducted at the level of individual
of similar chronological age. As many children emotions. One further complication is that even
with Down syndrome now attend mainstream in the absence of its use by the experimenter,
school, additional support in emotion recognition children may use emotion labelling internally to
may therefore be required to facilitate social assist them with emotion recognition tasks. This
interaction with peers in their own age group. too would be worth investigating more fully in the
The results of the present study suggest that future, given that self-produced labelling has been
labelling emotions and using exaggerated emo- shown to enhance emotion perception in typical
tional expressions may be helpful here, at least for development (Lindquist et al., 2006).
some emotions. The continuing difficulty in fear recognition,
Although emotion recognition in general was regardless of the presence or absence of a linguistic
not found to be any more difficult for the children label, suggests that some children with Down
with Down syndrome than the two comparison syndrome have a perceptual difficulty in interpret-
groups, they were significantly less accurate than ing this specific facial expression. This may in turn
the typically developing children in the recogni- relate to more basic face processing difficulties
tion of fearful expressions. This significant differ- amongst this group. Vicari et al. (2000), for
ence was found in three of the four conditions example, have noted that recognition of emotions
(labelled; nonlabelled; exaggerated) and it ap- such as fear and anger require the processing of
proached significance in the fourth (veridical). both the upper and the lower parts of the face,
This specific weakness in fear recognition ties in while research by Carvajal, Ferández-Alcaraz,
with findings from a number of previous studies Rueda, & Sarrión (2012) has shown that adults

148 Emotion Recognition and Down Syndrome


AMERICAN JOURNAL ON INTELLECTUAL AND DEVELOPMENTAL DISABILITIES ÓAAIDD
2017, Vol. 122, No. 2, 138–155 DOI: 10.1352/1944-7558-122.2.138

with Down syndrome may focus on the lower Although findings from the specific paradigm
more than the upper half of the face when used here must be generalized with caution, they
processing emotional expressions. However, the do suggest that in the early stages of any
extent to which such findings, in the absence of intervention to support the development of
significant parallel difficulties with anger, can emotion recognition skills, it might be worth
explain fear recognition difficulties in Down exploring whether individual children with Down
syndrome is unclear. syndrome do require particular support with
In typical development, a developmental recognition of fearful expressions. It may also be
exploration of fear recognition at a neurological worth considering the use of exaggerated expres-
level has been shown to be informative (e.g., sions (either in depicted or real life contexts) to
Guyer et al., 2008), with fear recognition linked to help underpin identification of more subtle and
the amygdala (for review and a contrasting neural complex emotions at a later stage in development.
model of emotion recognition see Lindquist, Although there is little direct evidence in the
Wager, Kober, Bliss-Moreau, & Barrett, 2012). literature to support the use of exaggerated over
Although this merits closer consideration in veridical emotions in the learning of emotional
Down syndrome, it is worth noting that although expressions, there is some indirect support for this
some studies have reported reduced amygdala in studies of how caregivers interact with their
volume in Down syndrome, this does not always infants in the early stages of learning about
remain after adjustment for overall brain volume, emotions. Caregivers typically exaggerate their
and in adults is complicated by the possible facial expressions when interacting with infants
presence of Alzheimer’s dementia (Aylward et al., (e.g., Stern, 1974), and in relation to positive
1999; Constable et al., 2010; Krasuki et al., 2002; emotions such as smiling, more intense expres-
White et al., 2003—for a recent developmental sions better capture the attention of infants
(Kuchuk, Vibbert, & Bornstein, 1986). The use
neurological overview see Karmiloff-Smith et al.,
of cartoon/schematic emotions (which often
2016). Further exploration of environmental
accentuate the key features of expressions) is also
influences on the development of fear recognition
already established in emotion recognition inter-
skills in Down syndrome should also be explored,
ventions for children with developmental disabil-
for example investigating the role of the children’s
ities (e.g., Silver and Oakes, 2001), suggesting that
own experiences of displaying and observing this
such approaches might merit evaluation with
emotion. Bringing these various ‘‘layers’’ of children with Down syndrome, including atten-
developmental pathways together in a causal tion to how progress is made in recognizing more
model of socio-cognitive development might well subtly expressed emotion in everyday social
prove informative in the future (see Cebula, situations. The findings from this study also
Moore, & Wishart, 2010; Moore & George, 2014). suggest that the use of emotion labels might be
As important as establishing the cause of fear similarly helpful in supporting emotion recogni-
recognition difficulties identified in laboratory- tion. Research with typically developing children,
based studies is the need to establish whether this however, suggests that exploring the embedding of
difficulty is observed in the everyday lives of these labels in causal conversational contexts may
children with Down syndrome and, if so, whether be particularly productive in supporting the
this influences their behavior or social interac- development of emotional understanding (Brown
tions. Even if fear is difficult for some children & Dunn, 1996; see Salmon et al., 2013).
with Down syndrome to recognize, it may be that In terms of facilitating emotion recognition
this does not affect day-to-day interactions in any development, the use of exaggerated emotions and
substantive way. This is not to suggest that emotion labels may not be equally facilitative for
recognition of fearful expressions is unimportant all emotions, however. Results from the present
(see e.g., Knieps, Walden, & Baxter, 1994; Thur- study indicated that the fear recognition ability of
man & Mervis, 2013), but rather that if interven- the children with Down syndrome—which was at
tions to better support adaptive social functioning chance level—did not improve under either of
are to be identified, there is a need to understand these conditions. This might suggest that ap-
more fully the impact of emotion recognition proaches need to be individually tailored more
ability on the social experiences of individuals closely to specific emotion and developmental
with intellectual disability (Zaja & Rojahn, 2008). stage in order to be effective.

K. R. Cebula et al. 149


AMERICAN JOURNAL ON INTELLECTUAL AND DEVELOPMENTAL DISABILITIES ÓAAIDD
2017, Vol. 122, No. 2, 138–155 DOI: 10.1352/1944-7558-122.2.138

In contrast to our earlier work (e.g., Williams language measures, supports the view that, lan-
et al., 2005; Wishart et al., 2007), the present study guage in general and emotion words more
found that emotion recognition ability among specifically, play a key role in emotion recognition
children with Down syndrome was significantly (Barrett et al. 2007). Yet the lack of correlation
correlated with broader aspects of cognitive and between children’s ability to recognize emotions
language development, such as performance and to label them emphasizes again that emotion-
mental age, verbal mental age, and vocabulary al understanding is not a unidimensional aspect of
comprehension ability. The present finding sug- development (Widen, Pochedly, & Russell, 2015).
gests that emotion recognition ability does not When considering the findings from this
necessarily unfold in isolation from other devel- study, some inherent limitations need to be taken
opment domains. It is therefore important to into account. As with so many studies in this field,
consider whether strengthening domain general sample size was relatively small and the age range
aspects of development, such as language and of the Down syndrome and nonspecific intellec-
cognition, might also enhance emotion recogni- tual disability groups was quite broad. However,
tion skills (Rosenqvist, Lahti-Nuuttila, Laasonen, all three participant groups were closely matched,
& Korkman, 2014, see also Happe & Frith, 2014). and relatively homogeneous, in terms of develop-
Supporting more general language development is mental ability. It also has to be noted that the use
a common intervention goal for children with of video analysis to measure response speeds may
Down syndrome (Fidler, Philofsky, & Hepburn, have been somewhat less accurate than if the trials
2007), but might also provide some concomitant had been presented via touchscreen computer.
support for emotion recognition. Support of other However, our previous work with these participant
core cognitive skills, such as verbal short-term groups has found that children at this develop-
memory, working memory, and visual selective mental level sometimes touch a number of
attention, all known to be areas of difficulty for
photographs before settling on their final answer,
those with Down syndrome (e.g., Breckenridge,
making the use of touchscreens potentially
Braddick, Anker, Woodhouse, & Atkinson, 2013;
problematic. In addition, our study employed
Brock & Jarrold, 2005; Cornish, Scerif, & Karmil-
archetypal black and white static images of facial
off-Smith, 2007; Costanzo et al., 2013), and also
expressions as stimuli. Clearly, in everyday life, as
linked to emotion recognition (Buitelaaer, van der
well as being dynamic and in context, emotions
Wees, Swaab-Barnveld, & van der Gaag, 1999;
Matthersul et al., 2009; Rosenqvist et al., 2014), can often be more subtle and ambiguous (Barrett,
might also be beneficial. Given the role of one’s Lindquist, & Gendron, 2007). It is acknowledged
own emotional experiences, and that of close that the use of static images in the present study
family members, on emotion recognition abilities limits its ecological validity, but it did have the
(e.g., Kujawa et al., 2014; Loi, Vaidya, & Paradiso, advantage of reducing the memory load for the
2013; Whittington & Holland, 2011), taking into participating children. The stimuli also had high
account the broader environmental context of the proven inter-rater reliability and included precisely
child may also be informative. measured exaggerated emotions. Further work to
In all participant groups, general language explore the ability of children with Down
measures (VMA and vocabulary comprehension) syndrome to recognize fearful expressions using
were correlated with emotion recognition ability, more naturalistic stimuli is nevertheless clearly
but the production of emotion labels was not. In desirable. The recent findings of Channell et al.
contrast, Herba et al. (2008) found that typically (2014) are informative here in suggesting that
developing children’s labelling ability was signif- emotion recognition may present as less of a
icantly correlated with emotion recognition accu- difficulty in tasks using color and dynamic stimuli,
racy. However, their task involved defining and also in highlighting the need for research into
emotion labels and label-to-emotion matching, the use of emotion knowledge in real time social
rather than label production, as in our study. In interactions. Use of more recent threshold ap-
the present study there is a complex pattern of proaches to measuring emotion discrimination
relationships between emotion recognition and (e.g., Rodger, Vizioli, Ouyang, & Caldara, 2015)
language: our finding that the use of emotion would also be helpful in providing a finer grained
labels improved recognition accuracy, along with understanding of emotion recognition in Down
the correlations between the tasks and the syndrome, though the suitability of such lengthy

150 Emotion Recognition and Down Syndrome


AMERICAN JOURNAL ON INTELLECTUAL AND DEVELOPMENTAL DISABILITIES ÓAAIDD
2017, Vol. 122, No. 2, 138–155 DOI: 10.1352/1944-7558-122.2.138

tasks for participants with intellectual disability Benton, A. L., Sivan, A. B., Hamsher, K. de S.,
has yet to be established. Varney, N. R., & Spreen, O. (1994). Contribu-
In future studies it might also be helpful to tions to neuropsychological assessment (2nd ed.).
explore fear recognition in Down syndrome across New York, NY: Oxford University Press.
a wider age range. In typically developing adults Blair, R. J. R., Colledge, E., Murray, L., &
there is a slight, but significant, decrease in the Mitchell, D. G. V. (2001). A selective
ability to recognize fearful expressions with age impairment in the processing of sad and
(Calder et al., 2003). The neurological changes fearful expressions in children with psycho-
associated with dementia have been found early in pathic tendencies. Journal of Abnormal Child
life in Down syndrome (for overviews see Karmil- Psychology, 29, 491–498.
off-Smith et al., 2016; Zigman & Lott, 2007), but
Breckenridge, K., Braddick, O., Anker, S., Wood-
it would be premature to associate such changes at
house, M., & Atkinson, J. (2013). Attention in
a neurological level with a difficulty in fear
recognition at a behavioral level in children and Williams syndrome and Down’s syndrome:
young adults: searching for developmental expla- Performance on the new early childhood
nations of recognition difficulties drawing on attention battery. British Journal of Develop-
environmental and/or neurological correlates of mental Psychology, 31, 257–269.
Down syndrome itself might be more appropriate. Brock J., & Jarrold, C. (2005). Serial order
Few studies to date encompass both adults and reconstruction in Down syndrome: Evidence
children and with only a few notable exceptions, for a selective deficit in verbal short-term
such as Kasari et al. (2001) and Pochon and memory. Journal of Child Psychology & Psychi-
Declercq (2013, 2014), little work in this field has atry, 46, 304–316.
been longitudinal. It seems likely therefore that Brown, J. R., & Dunn, J. (1996). Continuities in
research across a wider age-span, with accompany- emotion understanding from 3-6 years. Child
ing profiles on more global aspects of develop- Development, 67, 789–802.
mental and neurological functioning, could be Buitelaar, J. K., van der Wees, M., Swaab-
helpful in furthering our understanding of how Barneveld, H., & van der Gaag, R. J. (1999).
face processing and emotion recognition develop Verbal memory and performance IQ predict
with age in Down syndrome, pinpointing theory of mind and emotion recognition
strengths and weaknesses and better informing ability in children with autistic spectrum
intervention strategies where appropriate. disorders and in psychiatric control children.
Journal of Child Psychology and Psychiatry, 40,
References 869–881.
Aylward, E., Li, Q., Honeycutt, N., Warren, A., Calder, A. J., Keane, J., Manly, T., Sprengelmeyer,
Pulsifer, M., Barta, P., & Pearlson, G. (1999). R., Scott, S., Nimmo-Smith, I., & Young, A.
MRI volumes of the hippocampus and W. (2003) Facial expression recognition across
amygdala in adults with Down’s syndrome the adult life span. Neuropsychologia, 41, 195–
with and without dementia. American Journal 202.
of Psychiatry, 156(4), 564–568. Calder, A. J., Rowland, D., Young, A. W.,
Baron-Cohen, S., Golan, O., Wheelwright, S., & Nimmo-Smith, I., Keane, J., & Perrett, D. I.
Hill, J. J. (2004). Mindreading: An interactive (2000). Caricaturing facial expressions. Cogni-
guide to emotions. London, UK: Jessica Kings- tion, 76, 105–146.
ley. Calder, A. J., Young, A. W., Rowland, D., &
Barrett, L.F., Lindquist, K.A., & Gendrom, M. Perrett, D. I. (1997). Computer-enhanced
(2007). Language as context for the perception emotion in facial expressions. Proceedings of
of emotion. Trends in Cognitive Science, 11, the Royal Society of London Series B: Biological
327–332. Sciences, 264, 919–925.
Benton, A. L., Sivan, A. B., Hamsher, K. de S., Carvajal, F., Ferández-Alcaraz, C., Rueda, M., &
Varney, N. R., & Spreen, O. (1983). Benton Sarrión, L. (2012). Processing of facial expres-
facial recognition test. New York, NY: Oxford sions of emotion by adults with Down
University Press. syndrome and moderate intellectual disability.

K. R. Cebula et al. 151


AMERICAN JOURNAL ON INTELLECTUAL AND DEVELOPMENTAL DISABILITIES ÓAAIDD
2017, Vol. 122, No. 2, 138–155 DOI: 10.1352/1944-7558-122.2.138

Research in Developmental Disabilities, 33, 783– Clinical and Experimental Neuropsychology, 24,
790. 200–213.
Carvajal, F., & Iglesias, J. (1997). Mother and Dunn, L. M., Dunn, L. M., Whetton, C., &
infant smiling exchanges during face-to-face Burley, J. (1997). British picture vocabulary scale
interaction in infants with and without Down (2nd ed.). Windsor, UK: NFER-Nelson.
syndrome. Developmental Psychobiology, 31, Ekman, P., & Friesen, W. V. (1976). Pictures of
277–286. facial affect. Palo Alto, CA: Consulting Psy-
Castelli, F. (2005). Understanding emotion from chologists Press.
standardized facial expressions in autism and Fidler, D., & Nadel, L. (2007). Education and
normal development. Autism, 9, 428–449. children with Down syndrome: Neuroscience,
Cebula, K. R., Moore, D. G., & Wishart, J. G. development and intervention. Mental Retar-
(2010). Social cognition in children with dation and Developmental Disabilities Research
Down’s syndrome: Challenges to research Reviews, 13, 262–271.
and theory building. Journal of Intellectual Fidler, D. J., Philofsky, A., & Hepburn, S. L.
Disability Research, 54, 113–134. (2007). Language phenotypes and interven-
Cebula, K. R., & Wishart, J. G. (2008). Social tion planning: Bridging research and practice.
cognition in children with Down syndrome in Mental Retardation and Developmental Disabili-
L. Glidden (Ed.), International Review of ties Research Reviews, 13, 47–57.
Research in Mental Retardation (Vol. 35, pp. Gao, X., & Maurer, D. (2009). Influence of
43–86). New York, NY: Academic Press. intensity on children’s sensitivity to happy,
Celani, G., Battacchi, M. W., & Arcidiacono, L. sad, and fearful facial expressions. Journal of
(1999). The understanding of the emotional
Experimental Child Psychology, 102, 503–521.
meaning of facial expressions in people with
Golan, O., Ashwin, E., Granader, Y., McClintock,
autism. Journal of Autism and Developmental
S., Day, K., Leggett, V., & Baron-Cohen, S.
Disabilities, 29, 57–66.
(2010). Enhancing emotion recognition in
Channell, M. M., Conners, F. A., & Barth, J. M.
children with autism spectrum conditions: An
(2014). Emotion knowledge in children and
intervention using animated vehicles with real
adolescents with Down syndrome: A new
emotional faces. Journal of Autism and Devel-
methodological approach. American Journal of
opmental Disorders, 40, 269–279.
Intellectual and Developmental Disabilities, 119,
405–421. Guyer, A. E., Monk C. S., McClure-Tone, E. B.,
Constanzo, F., Varussa, C., Menghini, D., Addo- Nelson, E. E., Roberson-Nay, R., Adler, A. D.,
na, F., Gianesini, T., & Vicari, S. (2013). . . . Ernst, M. (2008). A developmental
Executive function in intellectual disabilities: examination of amygdala response to facial
A comparison between Williams syndrome expressions. Journal of Cognitive Neuroscience,
and Down syndrome. Research in Developmen- 20, 1565–1582.
tal Disabilities, 34, 1770–1780. Happé, F., & Frith, U. (2014). Annual research
Contestabile, A., Benfenati, F., & Gasparini, L. review: Towards a developmental neurosci-
(2010). Communication breaks-Down: From ence of atypical social cognition. Journal of
neurodevelopment defects to cognitive dis- Child Psychology and Psychiatry, 55, 553–577.
abilities in Down syndrome. Progress in Herba, C., & Phillips, M. (2004). Annotation:
Neurobiology, 91(1), 1–22. Development of facial expression recognition
Cornish, K., Scerif, G., & Karmiloff-Smith, A. from childhood to adolescence: Behavioural
(2007). Tracing syndrome-specific trajectories and neurological perspectives. Journal of Child
of attention across the lifespan. Cortex, 43, Psychology and Psychiatry, 45, 1185–1198.
672–685. Herba, C. M., Benson, P., Landau, S., Russell, T.,
DeSonneville, L. M. J., Verschoor, C. A., Njio- Goodwin, C., Lemche, E., . . . Phillips, M.
kiktjien, C., Op het Veld, V., Toorenaar, N., (2008). Impact of familiarity upon children’s
& Vranken, M. (2002). Facial identity and developing facial expression recognition. Jour-
facial emotions: Speed, accuracy, and process- nal of Child Psychology and Psychiatry, 49, 201–
ing strategies in children and adults. Journal of 210.

152 Emotion Recognition and Down Syndrome


AMERICAN JOURNAL ON INTELLECTUAL AND DEVELOPMENTAL DISABILITIES ÓAAIDD
2017, Vol. 122, No. 2, 138–155 DOI: 10.1352/1944-7558-122.2.138

Herba, C. M., Landau, S., Russell, T., Ecker, C., & Lindquist, K. A., & Gendron, M. (2013). What’s
Phillips, M. L. (2006). The development of in a word? Language constructs emotion
emotion-processing in children: Effects of age, perception. Emotion Review, 5, 66–71.
emotion, and intensity. Journal of Child Lindquist, K., Wager, T., Kober, H., Bliss-Moreau,
Psychology and Psychiatry, 47, 1098–1106. E., & Barrett, L. (2012). The brain basis of
Hippolyte, L., Barisnikov, K., Van der Linden, M., emotion: A meta-analytic review. Behavioral
& Detraux, J. J. (2009). From facial emotional and Brain Sciences, 35(3), 121–143.
recognition abilities to emotional attribution: Loi, F., Vaidya, J. G., & Paradiso, S. (2013).
A study in Down syndrome. Research in Recognition of emotion from body language
Developmental Disabilities, 30, 1007–1022. among patients with unipolar depression.
Karmiloff-Smith, A., Al-Janabi, T., D’Souza, H. Psychiatry Research, 209, 40–49.
Groet, J., Esha Massand, E., Mok, K., . . . Mathersul., D., Palmer, D. M., Gur, R. C., Gur, R.
Strydom, A. (2016). The importance of E., Cooper, N., Gordon, E., & Williams, L.
understanding individual differences in Down M. (2009). Explicit identification and implicit
syndrome. F1000Research, 5. emotion recognition of facial emotions: II.
Kasari, C., Freeman, S. F. N., & Hughes, M. A. Core domains and relationships with general
(2001). Emotion recognition by children with cognition. Journal of Clinical and Experimental
Down syndrome. American Journal on Mental Neuropsychology, 31, 278–291.
Retardation, 106, 59–72. McKone, E., Crookes, K., Jeffery, L., & Dilks, D.
Knieps, L. J., Walden, T., A. & Baxter, A. (1994). (2012). A critical review of the development of
Affective expressions of toddlers with and face recognition: Experience is less important
without Down syndrome in a social referenc- than previously believed. Cognitive Neuropsy-
ing context. American Journal on Mental chology, 29(12), 174–212.
Retardation, 99, 301–312. Moore, D. G., & George, R. (2014). ACORNS: A
Krasuski, J. S., Alexander, G. E., Horwitz, B., tool for visual modelling of causes and
Rapoport, S. I., & Schapiro, M. B. (2002). outcomes in neurodevelopmental disorders.
Relation of medial temporal lobe volumes to In J. Van Herwegen & D. Riby (Eds.),
age and memory function in non-demented Neurodevelopmental disorders: research challenges
adults with Down’s syndrome: Implications and solutions. London, UK: Psychology Press.
for the prodromal phase of Alzheimer’s Morris, J., Friston, K., Buchel, C., Frith, C.,
disease. American Journal of Psychiatry, 159(1), Young, A., Calder, A., & Dolan, R. (1998). A
74–81. neuromodulatory role for the human amyg-
Kuchuk, A., Vibbert, M., & Bornstein, M. (1986). dala in processing emotional facial expres-
The perception of smiling and its experiential sions. Brain, 121(1), 47–57.
correlates in three-month-old infants. Child Morris, J. S., Frith, C. D., Perrett, D. I., Rowland,
Development, 57(4), 1054–1061. D., Young, A. W., Calder, A. J., & Dolan, R. J.
Kujawa, A., Dougherty, L., Durbin, E., Laptook, (1996). A differential neural response in the
R., Torpey, D., & Klein, D. N. (2014). human amygdala to fearful and happy facial
Emotion recognition in preschool children: expressions. Nature, 383(6603), 812–815.
Associations with maternal depression and Noldus, L. P. J. J., Trienes, R. J. H., Hendriksen, A.
early parenting. Development and Psychopathol- H. M., Jansen, H., & Jansen, R. G. (2000). The
ogy, 26, 159–170. Observer Video-Pro: New software for the
Kumfor, F., Miller, L., Lah, S., Hsieh, S., Savage, collection, management and presentation of
S., Hodges, J., & Piguet, O. (2011). Are you time-structured data from videotapes and
really angry? The effect of intensity on facial digital media files. Behavior Research Methods,
emotion recognition in frontotemporal de- Instruments and Computers, 32, 197–206.
mentia. Social Neuroscience, 6(56), 502–514. Pochon, R., & Declercq, C. (2013). Emotion
Lindquist, K., Barrett, L., Bliss-Moreau, E., recognition by children with Down syn-
Russell, J., Davidson, R. J., & Scherer, K. R. drome: A longitudinal study. Journal of
(2006). Language and the perception of Intellectual and Developmental Disability, 38,
emotion. Emotion, 6(1), 125–138. 332–343.

K. R. Cebula et al. 153


AMERICAN JOURNAL ON INTELLECTUAL AND DEVELOPMENTAL DISABILITIES ÓAAIDD
2017, Vol. 122, No. 2, 138–155 DOI: 10.1352/1944-7558-122.2.138

Pochon, R., & Declercq, C. (2014). Emotional Székely, E., Tiemeier, H., Arends, L. R., Jaddoe, V.
lexicon understanding and emotion recogni- W. V., Hofman, A., Verhulst, F. C., & Herba,
tion: A longitudinal study in children with C. M. (2011). Recognition of facial expres-
Down syndrome. Journal of Developmental and sions of emotions by 3-year-olds. Emotion, 11,
Physical Disabilities, 26, 549–563. 425–435.
Porter, M. A., Coltheart, M., & Langdon, R. Thurman, A. J., & Mervis, C. B. (2013). The
(2007). The neuropsychological basis of hy- regulatory function of social referencing in
persociability in Williams and Down syn- preschoolers with Down syndrome or Wil-
drome. Neuropsychologia, 45, 2839–2849. liams syndrome. Journal of Neurodevelopmental
Reed, P., & Steed, I. (2015). Interference with Disorders, 5, 2.
facial emotion recognition by verbal but not Tingley, E. C., Gleason, J. B., & Hooshyar, N.
visual loads. Research in Developmental Disabil- (1994). Mothers’ lexicon of internal words in
ities, 47, 441–50. speech to children with Down syndrome and to
Ridgeway, D., Waters, E., & Kuczaj, S. A. (1985). nonhandicapped children at mealtime. Journal
Acquisition of emotion-descriptive language: of Communication Disorders, 27, 135–155.
Receptive and productive vocabulary norms Turk, J., & Cornish, K. (1998). Face recognition
for ages 18 months to 6 years. Developmental and emotion perception in boys with fragile-X
Psychology, 21, 901–908. syndrome. Journal of Intellectual Disability
Rodger, H., Vizioli, L., Ouyang, X., & Caldara, R. Research, 42, 490–499.
(2015). Mapping the development of facial Vicari, S., Reilly, J. S., Pasqualetti, P., Vizzotto, A.,
expression recognition. Developmental Science, & Caltagirone, C. (2000). Recognition of
18(6), 926–939. facial expressions of emotion in school-age
Rogoff, B. (1990). Apprenticeship in thinking: children: the intersection of perceptual and
Cognitive development in social context. Oxford, semantic categories. Acta Paediatr, 89, 836–45.
UK: Oxford University Press. Virji-Babul, N., Watt, K., Nathoo, F., & Johnson,
Rosenqvist, J., Lahti-Nuuttila, P., Laasonen, M., & P. (2012). Recognition of facial expressions of
Korkman, M. (2014). Preschoolers’ recogni- emotion in adults with Down syndrome.
tion of emotional expressions: Relationships Physical and Occupational Therapy in Pediatrics,
with other neurocognitive capacities. Child 32, 333–343.
Neuropsychology: A Journal on Normal and Way, E., & Rojahn, L. (2012). Psycho-social charac-
Abnormal Development in Childhood and Ado- teristics of children with prenatal alcohol expo-
lescence, 20, 281–302. sure, compared to children with Down
Russell, J. A. & Widen, S. C. (2002). A label syndrome and typical children. Journal of Devel-
superiority effect in children’s categorization of opmental and Physical Disabilities, 24(3), 247–268
facial expressions. Social Development, 11, 30–52. Wechsler, D. (1990). Wechsler pre-school and primary
Salmon, K., Evans, I. M., Moskowitz, S., Grou- scales of intelligence (revised UK ed.). London,
den, M., Parkes, F., & Miller, E. (2013). The UK: The Psychological Corporation.
components of young children’s emotion White, N. S, Alkire, M. T., & Haier, R. J. (2003). A
knowledge: Which are enhanced by adult voxel-based morphometric study of nonde-
emotion talk? Social Development, 22, 94–110. mented adults with Down Syndrome. Neuro-
Sattler, J. M. (1992). Assessment of children: WISC image, 20(1), 393–403.
and WPPSI-R supplement. San Diego, CA: Whittington, J., & Holland, T. (2011). Recogni-
Jerome M. Sattler. tion of emotion in facial expression by people
Silver, M., & Oakes, P. (2001). Evaluation of a new with Prader-Willi syndrome. Journal of Intellec-
computer intervention to teach people with tual Disability Research, 55, 75–84.
autism or Asperger syndrome to recognize and Widen, S. C., Pochedly, J. T., & Russell, J. A.
predict emotions in others. Autism, 5, 299–316. (2015). The development of emotion con-
Stern, D. (1974). The goal and structure of cepts: A story superiority effect in older
mother-infant play. Journal of the American children and adolescents. Journal of Experimen-
Academy of Child Psychiatry, 13(3), 402–421. tal Child Psychology, 131, 186–192.

154 Emotion Recognition and Down Syndrome


AMERICAN JOURNAL ON INTELLECTUAL AND DEVELOPMENTAL DISABILITIES ÓAAIDD
2017, Vol. 122, No. 2, 138–155 DOI: 10.1352/1944-7558-122.2.138

Widen, S. C., & Russell, J. A. (2003). A closer look Zigman, W. B., & Lott, I. T. (2007). Alzheimer’s
at preschoolers’ freely produced labels for disease in Down syndrome: Neurobiology
facial expressions. Developmental Psychology, 39, and risk. Mental Retardation and Developmental
114–128. Disabilities Research Reviews, 13(3), 237–246.
Widen, S. C., & Russell, J. A. (2004). The relative
power of an emotion’s facial expression, label
and behavioral consequence to evoke pre- Received 3/30/2015, accepted 7/13/2016.
schoolers’ knowledge of its cause. Cognitive
Development, 19, 111–125.
Widen, S. C., & Russell, J. A. (2008). Children This research was funded by the Medical Research
acquire emotion categories gradually. Cogni- Council of Great Britain (Project Grant No.
tive Development, 23, 291–312. G0000325). Many thanks are due to the schools and
Williams, K. R., Wishart, J. G., Pitcairn, T. K., & children who took part in this study, and to Down’s
Willis, D. (2005). Emotion recognition by Syndrome Scotland for their continuing support of our
children with Down syndrome: Investigation of program of research. Thanks are also extended to Debra
specific impairments and error patterns. Ameri- Bowyer and Judith Scott (clinical coordinators), as well
can Journal on Mental Retardation, 110, 378–392. as to Pat Jackson and Lothian Primary Care Trust
Wishart, J. G., Cebula, K. R., Willis, D. S., & Pitcairn, (confirmation of participant diagnoses).
T. K., (2007). Understanding of facial expres-
sions of emotion by children with intellectual
disabilities of differing aetiology. Journal of Authors:
Intellectual Disability Research, 51, 551–563. Katie R. Cebula, University of Edinburgh, School
Wishart, J. G., & Pitcairn, T. K. (2000). Recogni- of Education; Jennifer G. Wishart, University of
tion of identity and expression in faces by Edinburgh; Diane S. Willis, Edinburgh Napier
children with Down syndrome. American University; and Tom K. Pitcairn, formerly of
Journal on Mental Retardation, 105, 466–479. University of Edinburgh.
Young, A., Perrett, D., Calder, A., Sprengelmeyer,
R., & Ekman, P. (2002). Facial expressions of
emotion – stimuli and tests (FEEST). Bury St. Correspondence concerning this article should be
Edmunds, UK: Thames Valley Test Company. addressed to Katie R. Cebula, University of
Zaja, R. H., & Rojahn, J. (2008). Facial emotion Edinburgh, School of Education, Holyrood
recognition in intellectual disabilities. Current Road, Edinburgh, EH8 8AQ, UK (e-mail: Katie.
Opinion in Psychiatry, 21, 441–444. Cebula@ed.ac.uk).

K. R. Cebula et al. 155

Вам также может понравиться