Shanok, Jones, Lucas, 2019

You might also like

Download as pdf or txt
Download as pdf or txt
You are on page 1of 7

Child Psychiatry & Human Development

https://doi.org/10.1007/s10578-019-00870-z

ORIGINAL ARTICLE

The Nature of Facial Emotion Recognition Impairments in Children


on the Autism Spectrum
Nathaniel A. Shanok1   · Nancy Aaron Jones1 · Nikola N. Lucas1

© Springer Science+Business Media, LLC, part of Springer Nature 2019

Abstract
This study examined socio-emotional skills, utilizing a facial emotion recognition (FER) task featuring unfamiliar and famil-
iar faces, in children with autism spectrum disorders (ASD) compared to typically developing (TD) children. Results showed
that the TD children were more proficient on the FER overall whereas ASD children recognized familiar expressions more
precisely than unfamiliar ones. Further, ASD children did not differ from TD children in recognizing happy expressions but
ASD children were less skilled with recognizing negative expressions. Findings suggest that ASD children possess more
adept FER abilities than previously thought especially for important social others. Ultimately, a task featuring an array of
positive and negative familiar and unfamiliar expressions may provide a more comprehensive assessment of socio-emotional
abilities in ASD children.

Keywords  Autism spectrum disorder · Facial emotion recognition · Developmental disorders · Socio-emotional
development · Emotional familiarity

Autism spectrum disorder (ASD) and autism are both gen- other studies have reported either marginal differences or
eral terms for a group of complex disorders characterized by no differences in performance between ASD and (typically
impaired social interactions, in which children are hindered developing) TD groups of similar ages [4, 11, 12]. Vari-
by a lack of interest in initiating conversation, atypical eye- ation in experimental methodologies, samples, cognitive
contact, and especially struggles in recognizing the emotions variants, and neurological factors could be responsible for
of others [1, 2]. The ability to infer emotions from facial the divergent findings obtained in prior studies. Many of the
expressions is critical for interpreting important social cues null findings have occurred in studies investigating high-
during conversation [3]. Facial emotion recognition (FER) functioning ASD (HF-ASD) samples, therefore this study
tasks have been devised to assess recognition ability and aimed to further evaluate HF-ASD emotion recognition
overall emotional intelligence, with the general aim being to ability during a key time-period of social development, i.e.,
match emotional labels to their corresponding facial expres- the period from preschool to elementary in which the child
sions [4, 5]. Performances on FER tasks are predictive of moves from predominantly family to school social experi-
individual differences in quality of social interactions [6]. ences (4–8-years-old).
While social and emotional impairments have been fre- Previous FER studies have explored the influence of vari-
quently observed in case studies on individuals with ASD ables like gender, type of emotional expression, and amount
[7], the literature regarding FER ability in both children of allotted time on emotional processing [13] however, only
and adults with ASD has yielded surprisingly inconsistent a few studies have looked at the factor of familiarity and its
results. Some studies have highlighted significant impair- effects on recognition ability in ASD [5, 14–16]. This gap in
ments in emotion recognition accuracy among clinically- the literature is surprising as familiar faces are representative
diagnosed children [8], adolescents [9], and adults [10], yet of a larger portion of an individual’s daily social interac-
tions, in that most interactions occur with family, friends,
and classmates [17]. It seems logical that an FER task featur-
* Nathaniel A. Shanok ing familiar and unfamiliar stimuli may serve as a better tool
nshanok@fau.edu for assessing emotion recognition than conventional FER
1 tasks which solely focus on the expressions of strangers.
Florida Atlantic University, Boca Raton, FL, USA

13
Vol.:(0123456789)
Child Psychiatry & Human Development

The existing research on familiarity in FER studies is disgust) are more challenging for children to recognize
informative but exhibits important limitations that warrant in faces than positive emotions (happy, excitement, etc.)
further investigation of the topic. Together, these studies [19]. In addition, some studies have noted that children
suggest that children and adults with ASD show reduced with ASD are more likely to display impairments in rec-
eye-tracking behavior and fusiform gyrus/amygdala activa- ognizing negative emotions compared to positive ones [20,
tion to unfamiliar expressions compared to familiar ones 21]. Therefore, it was expected that ASD children would
[14–16]. However, none of the aforementioned studies pre- show lower proficiency on the FER task and would show
sented participants with a diverse range of emotions; the reduced skills in recognizing negative expressions (sad,
tasks consisted of only two types of expressions, “neutral” angry, fear) in comparison to the TD group. No hypothesis
and “emotional”. While these studies drew conclusions was formulated for the familiarity/unfamiliarity factor due
regarding the neurological and motor responses to emotional to the findings of past studies being less conclusive and
expressions in ASD, there was no assessment of recognition less theoretical than what is desirable [5, 8].
ability at the “conscious level” (i.e. self-report).
Thus far, one study has assessed group differences in rec-
ognition accuracy for a diverse range of emotions (happy,
sad, fear, and anger) [18] and surprisingly found that the Methods
ASD group had more proficient recognition for unfamiliar
facial expression stimuli than familiar ones. This finding, Participants
which is the opposite of what we would expect may be in
part attributable to two characteristics of the study. Primar- Twenty-eight, 4–8-years-old children (M age = 5.61 years,
ily, the familiar stimuli consisted of expressions generated SD = 1.23) were participants, with one group of children
by either the participant’s mother or teacher, which likely diagnosed with HF-ASD (N = 12) and the other a TD
represent varying degrees of familiarity to the participant. group (N = 16). See Table 1 for participant and family
Second, the participants consisted of a wide age-range of demographic characteristics. All of the HF-ASD children
children (4–15-years-old). had been previously diagnosed by a clinician at a nearby
Therefore, the aim of this paper was to further examine private practice. The participants in the ASD group were
the influence of familiarity on group performance (ASD, sampled from several autism schools on a university cam-
TD) in emotion recognition with a constant familiar pus and in a rural community while TD children were
stimulus (i.e. participant’s mother) as well as with a more recruited from university-based schools. This research
targeted age range (4–8-years-old). A second aim was to was reviewed and approved by the University’s IRB and
assess the variation in recognition ability for positive and participants were treated in accordance with the “Ethical
negative expression types between groups. Notably, there Principles of Psychologists and Code of Conduct” (Ameri-
is evidence that negative emotions (fear, anger, sadness, can Psychological Association, 1992).

Table 1  Comparing groups on Measurement ASD (N) ASD (%) TD (N) TD (%)


demographic variables
Age M = 5.75 SD = 0.97 M = 5.50 SD = 1.41
Gender
 Male 9 75.00 10 62.5
 Female 3 25.00 6 37.5
Race/ethnicity
 White/Caucasian 9 75.00 13 81.25
 Black/AA 1 8.33 0 0.00
 Hispanic/Latino 2 16.66 0 0.00
 Asian/Caucasian 0 0.00 3 18.75
GARS-2 score M = 55.50 SD = 6.09 M = 8.50 SD = 9.99
Family member with ASD
 Yes 3 25.00 0 0.00
 No 9 75.00 16 100.00
Family income (dollars) M = 96,000 SD = 33.38 M = 105,333 SD = 35.03
Mother’s education (years) M = 16.00 SD = 1.91 M = 16.38 SD = 1.67

13
Child Psychiatry & Human Development

Materials number of trials that a participant answered correctly by the


total number of trials. Percent scores were also calculated
To independently verify the reliability of the group status for separately for familiar stimuli, unfamiliar stimuli, and each
all children, the Gilliam autism rating scale: second edition of the four tested emotions (happy, sad, angry, fear). A criti-
(GARS-2) was administered to the mothers of participants cal goal was to include an adequate number of stimuli for
[22]. For the current study, the average composite GARS-2 experimental validity without overloading participants or
score in the ASD group was 55.50 (SD = 26.09), and 8.50 inducing a fatigue effect. The 24 trials were randomized so
(SD = 9.99) in the TD group. Notably, the GARS-2 is a that each participant completed them in a different order and
norm-referenced, 42-item instrument based on the diagnos- reliability was assessed across two experimenters (100%)
tic guidelines of the DSM-V (three subscales are: stereo- who recorded the responses for all 28 participants across
typed behaviors, communication, and social interaction) and the 24 stimuli.
the Autism Society of America. It is used to gauge the sever-
ity of autism symptoms in individuals from 3- to 22-years. Emotion Tutorial (Happy, Sad, Angry, Fear)

NimStim and FER Stimuli Prior to the presentation of the experimental task, a brief
tutorial was given on each of the four emotion labels tested
The unfamiliar FER stimuli used in this study were sampled in this study (happy, sad, angry, fear) to ensure that partici-
from the NimStim set [23]. The entire NimStim set consists pants had a conceptual understanding of these terms. The
of 646 experimentally-tested facial emotion images, includ- tutorial consisted of descriptions/examples of these emo-
ing images of happiness, sadness, anger, fear, disgust, sur- tions, and included a task requiring participants to match
prise, neutral, and calm. For the current study, the number of sample phrases to the target words. No images were used in
expressions was condensed down to five: happy, sad, anger, the training, just sentences and phrases. Overall, this tuto-
fear, and neutral, with neutral serving as a control. Both rial took 10 min to complete and all 28 participants showed
positive (happy) and negative emotions (anger, fear, sad) a proficient understanding of the emotion labels (90% or
were included because negative emotions are thought to be better).
more difficult to master in developing children [24]. From
the NimStim set (unfamiliar), three images were sampled Procedures
for each of the five emotional expressions used in this study,
resulting in 15 images. To create the familiar expressions, Testing took place in a research lab and all procedures
the 15 unfamiliar expressions were replicated by the par- were completed in one sitting. After parental consent was
ticipant’s mother. All images presented were of females to obtained, the mother began the process of creating the
control for any gender-induced effects. familiar facial emotion expressions for their child. While
The adjusted FER task was presented to participants the pictures were being taken, child participants were play-
via the PowerPoint application on a MacBook Pro 13-inch ing with a sensory toy or building blocks under the supervi-
computer screen. The trials included a target emotion word sion of a research assistant. Mothers were prompted to rep-
and only three images that each participant was directed to licate three of each expression, including happy, sad, angry,
choose among. The reason for this setup was to avoid over- fear, and neutral, to the best of their ability. An image was
loading participants with processing demands [5]. The three only deemed usable when two experimenters agreed that it
images included the correct matching image, an incorrect resembled the desired image. All images were captured, then
one, and a neutral facial expression to serve as a control. All cropped and finally fit into the FER PowerPoint presentation,
trials included three images from the same female whether utilizing the Photo Booth application on a MacBook Pro
it was the mother (familiar) or one of the actresses (unfa- laptop computer. After, mothers completed the demographic
miliar). No images of the same actress were used for more and the GARS-2 questionnaires while the child executed the
than one unfamiliar trial. Each set of emotional images were FER task.
paired with the other three emotion types an equal number During the FER task, children were instructed to touch
of times, and each possible order was presented to control the facial emotion expression on the computer screen that
for any priming or ordering effects. The total FER task they felt best matched the corresponding emotion word.
consisted of 24 trials with each emotion type (happy, sad, Each of the 24 trials were presented one at a time. The tar-
angry, fear) having the same number of correct trials (6). In get emotion label (happy, sad, angry, fear) appeared at the
addition, 12 of the trials were familiar (3 for each emotion) bottom of the screen and was read aloud to the participant.
and 12 were unfamiliar (3 for each emotion). The primary If participants changed their answer, then the last answer
dependent variable in this study was emotion recognition was recorded and used for analysis. Completion time was
percent score across all trials, calculated by dividing the approximately 1.5 h for each family.

13
Child Psychiatry & Human Development

Results Two‑Way Interaction: Group X Familiarity and Group


X Emotion Type
Overall Group Differences on FER
The primary hypotheses were related to the two-way inter-
A 2 × 2 × 4 mixed model analysis of variance was performed, actions; such that it was expected that groups would show
with group (ASD, TD) as the between-subjects variable different results based on stimuli familiarity and emotion
and both familiarity (familiar, unfamiliar) and emotion type expression type. To understand the significant interaction
(happy, sad, angry, fear) as within-subjects variables. The effect of group (ASD, TD) and familiarity (familiar, unfa-
dependent variable was emotion recognition percent score, miliar) on recognition proficiency, two independent samples
computed as the percentage of correct responses out of 24 t tests were performed and revealed that the TD group scored
trials. significantly better on unfamiliar expressions compared to
Overall, there was a significant three-way interaction the ASD group, t(26) = 4.82, p < .001 yet there were no dif-
effect of group by familiarity by emotion type on emo- ferences between groups for the familiar stimuli, despite the
tion recognition percent scores, F(3,78) = 2.98, p < .05, TD group having a higher mean percent score, t(26) = 1.45,
η 2 = 0.10. The two-way interaction effect of group by p = .16. To control for the familywise error-rate in the two t
familiarity was also significant, F(1,26) = 10.11, p < .01, test comparisons, a Bonferonni correction was used here and
η2 = 0.28, however the two-way interaction effect of group in further analyses with the resulting alpha values p < .025
by emotion type was not significant, F(3,78) = 2.47, p = .07, (0.05/2). Significant group differences remained for unfa-
η 2 = 0.09. Finally, there were significant main effects miliar expressions. A paired samples t test revealed that the
for group, F(1,26) = 15.04, p < .001, η2 = 0.36, familiar- ASD group scored significantly better on familiar than on
ity, F(1,26) = 6.58, p < .05, η2 = 0.20, and emotion type, unfamiliar facial expressions, t(11) = − 2.53, p < .05. Sig-
F(3,78) = 7.13, p < .001, η2 = 0.22. To clarify, the specific nificant differences on these two measures did not exist for
ASD and TD group means for overall emotion recognition the TD group (Table 2).
scores are presented in Table 2. Follow-up analyses were The two-way interaction of group and emotion type
conducted for the significant three-way interaction effect as approximated significance, F(3,78) = 2.47, p = .07, η2 = 0.09.
well as the two-way interaction effects. For exploratory purposes, follow-up independent samples
t tests were run to determine if recognition of any specific
emotions differed between the two groups, using a Bonfer-
onni correction, p < .0125 (0.05/4). It was found that sig-
nificant group differences in emotion recognition percent
scores existed for sad expressions, t(26) = 4.521, p < .001
only, with no differences for happy, t(26) = 1.72, p = .10,

Table 2  Comparing groups on Measurement: percent score ASD mean % (SD) TD mean % (SD) Significant group
emotion recognition percent differences (T-tests)
scores for all stimuli types
Overall score 85.39 (7.44) 95.31 (4.78) *p < .001
 Unfamiliar score 79.86 (10.33) 94.79 (5.99) *p < .001
 Familiar score 90.97 (10.92) 95.83 (6.80) p > .05
Total happy score 97.22 (6.49) 100.00 (0) p > .05
 Unfamiliar happy 94.44 (12.98) 100.00 (0) p > .05
 Familiar happy 100.00 (0) 100.00 (0) p > .05
Total fear score 83.28 (15.96) 93.75 (10.32) *p = .04
 Unfamiliar fear 86.11 (17.17) 93.75 (13.44) p > .05
 Familiar fear 83.33 (22.48) 93.75 (13.44) p > .05
Total sad score 72.22 (16.41) 94.79 (10.04) *p < .001
 Unfamiliar sad 69.44 (22.49) 95.83 (11.39) *p < .001
 Familiar sad 86.11 (17.17) 93.75 (13.44) p > .05
Total anger score 87.47 (7.56) 92.71 (10.49) p = .16
 Unfamiliar anger 77.77 (16.42) 93.75 (11.39) *p = .009
 Familiar anger 97.22 (14.91) 91.67 (14.99) p > .05

*p < .05

13
Child Psychiatry & Human Development

fear, t(26) = 2.11, p = .04, nor angry, t(26) = 1.47, p = .16 Herba and colleagues [18] who found that surprisingly chil-
(Table 2). Still these results must be taken with caution given dren with ASD were more adept in recognizing unfamiliar
the non-significant two-way interaction effect. expressions than familiar ones. To extend this line of work, it
may be prudent to determine if children with ASD also have
Three‑Way Interaction: Group X Emotion Type X enhanced emotion recognition when processing expressions
Familiarity generated by other individuals in their lives, such as fathers,
siblings, peers, and teachers. Taken together, these findings
To understand the significant three-way interaction effect, indicate that children with ASD are perceptive to their moth-
eight independent samples t tests were run (unfamiliar er’s emotions, which may indicate greater potential to learn
happy, familiar happy, unfamiliar sad, familiar sad, unfa- and socialize with close others rather than unfamiliar people.
miliar fear, familiar fear, unfamiliar angry, familiar angry), An additional goal of this study was to understand rec-
utilizing the Bonferonni correction, p < .00625 (0.05/8). ognition abilities to specific emotions (happy, sad, angry,
Group differences remained significant only for unfamiliar fear). Interestingly, TD children performed significantly bet-
sad expressions, t(26) = 4.10, p < .001 (Table 2). ter (than the ASD children) on two negative emotions (sad,
fear), but group differences were not uncovered for happy
and angry expressions. In children with or without develop-
Discussion mental complications, it is often more challenging to master
negative than positive emotion recognition [29]. For chil-
To expand the literature regarding the facial emotion rec- dren with ASD, further difficulties in recognizing negative
ognition abilities of individuals with HF-ASD, the current or complex facial emotional expressions may be due in part
study compared the FER abilities of 4–8-year-old children to variations in eye-contact and eye tracking behavior, with
on the ASD spectrum with TD children. In essence, the goal studies demonstrating that ASD children focus less on the
was to assess whether children with ASD differed in their eyes and more on other facial areas that are less informative
recognition of emotions using an FER task that included of an individual’s emotional state [4, 30]. Schurgin et al. [31]
unfamiliar expressions as well as familiar ones (i.e., those found evidence that eye fixations in normative individuals
expressions exhibited by the child’s mother). The familiar vary depending on which emotional expression is displayed
and unfamiliar images portrayed three negative emotions by a face. For example, when examining sad, angry, shame,
(sad, angry, fear) and one positive emotion (happy). Nearly and disgust expressions, fixation of eyes occurs most in eye
all past investigations on this topic have presented images regions whereas for happy expressions, individuals tend to
of only unfamiliar adults, failing to appreciate and assess fixate more on the mouth region. The finding that children
a major portion of a child’s social interactions, ones that with ASD tend to make less eye-contact during socio-emo-
commonly occur with family, friends, and teachers [5, 14, tional processing may explain their further struggles with
18]. The results demonstrated that both groups have similar recognizing sad and other negative expressions [32], and
ability to recognize familiar expressions and that the TD also why they rarely struggle with recognizing happy expres-
children show additional proficiencies in recognizing unfa- sions, as was the case for this study too.
miliar expressions [13, 17, 25]. This study was not without limitations, as there was (1)
The finding that ASD children performed similarly to a restricted range of outcomes, (2) expected confines in the
TD children on familiar expressions is an interesting aspect duration of attention within the child-clinical sample and
of the study as this result suggests children with ASD are (3) reduced sample size. Regarding the latter, a recent meta-
as in-tune with their mother’s feelings and emotions as TD analysis on the topic [33] analyzed 51 studies on the topic
children [16, 26, 27]. Moreover, these findings show that of emotion recognition and autism and found that only 20%
additional research is needed to understand how children of the studies (10 studies) had group sample sizes over 20
with ASD form emotional bonds or attachments with their participants. The sample size in the current investigation
family members. The mother–child relationships may appear was admittedly small but comparable to previous investiga-
to be slightly unconventional, but it is likely that quality tions. In addition, we examined a more targeted age-range
interactions in these relationships is vital for the well-being (4–8-years-old). Thus, replication of findings is still nec-
of children with ASD. The results obtained here are in-line essary to enhance validity. Moreover, it is also important
with previous neurological and eye-tracking investigations to acknowledge that ceiling effects were present in the TD
on the topic [14, 15, 28]. For example, research has shown group, especially with the happy faces. Ceiling effects are a
that children with ASD showed normative pupillary reac- concern in that there is an increased risk in making a type 1
tions to familiar people showing fear, yet both attenuated error or finding false significance [34]. To control for ceiling
eye-tracking and EEG N400 neural responses to unfamiliar effects, it may be sensible to present the task in a manner that
people [28]. However, the results obtained here contrast with is more challenging, either by: (1) presenting participants

13
Child Psychiatry & Human Development

with more choices per trial; (2) evaluating additional expres- from additional familiar individuals (i.e. fathers, siblings,
sions such as surprise, disgust, or embarrassment; or (3) classmates).
examining dynamic changes in facial expression recogni-
tion in stimuli that gradually morph into specific expressions
thereby assessing the threshold for recognition. It is also Compliance with Ethical Standards 
recommended for future studies to include a formal measure
for assessing the quality of created stimuli (familiar), how- Conflict of interest  All author declares that they have no conflict of
interest.
ever this was not likely a factor in this study as both groups
showed more proficient recognition of their mother’s images Ethical Approval  All procedures performed in studies involving human
compared to the actress’s images. participants were in accordance with the ethical standards of the insti-
To conclude, the current study highlights the impor- tutional and/or national research committee and with the 1964 Helsinki
declaration and its later amendments or comparable ethical standards.
tance of understanding FER skills in children on the autism
spectrum and potentially enhancing FER tasks to include Informed Consent  Informed consent was obtained from all individual
expressions from both familiar and unfamiliar individuals. participants included in the study.
Administering this type of task in children that vary in socio-
emotional processing skills may help developmental scien-
tists and clinicians to understand the types of social situa-
tions that are more challenging for this group. The results of References
this study illustrate that children with ASD potentially have
1. Aldred C, Green J, Adams C (2004) A new social communication
adept and competent emotional processing when viewing intervention for children with autism: pilot randomized controlled
their mother’s faces, but are less skilled at processing emo- treatment study suggesting effectiveness. J Child Psychol Psychia-
tional information, especially sad faces, when viewing the try 45:1420–1430
2. Dawson G, Meltzoff AN, Osterling J, Rinaldi J, Brown E (1998)
faces of strangers. It is unclear whether the latter is due to
Children with autism fail to orient to naturally occurring social
genuine deficits in processing abilities of ASD children, the stimuli. J Autism Dev Disord 28:479–485
aversive nature of negative emotions, or due to a general pat- 3. Baron-Cohen S, Ring HA, Wheelwright S, Bullmore ET, Bram-
tern of disinterest in negative emotion displays or unfamiliar mer MJ, Simmons A, Williams SC (1999) Social intelligence
in the normal and autistic brain: an fMRI study. Eur J Neurosci
people. Future research should continue to scrutinize the
11:1891–1898
socio-cognitive and neurodevelopmental processes involved 4. Leppänen JM, Nelson CA (2006) The development and neu-
in emotion recognition, especially in children on the autism ral bases of facial emotion recognition. Adv Child Dev Behav
spectrum. This study is further evidence that an FER task 34:207–246
5. Harms MB, Martin A, Wallace GL (2010) Facial emotion recog-
can be an extremely useful resource in not only assessing
nition in autism spectrum disorders: a review of behavioral and
emotional intelligence in developing children, but also track- neuroimaging studies. Neuropsychol Rev 20:290–322
ing the range of individual differences in the development of 6. Balconi M, Amenta S, Ferrari C (2012) Emotional decoding in
socio-emotional abilities over time. facial expression, scripts and videos: a comparison between nor-
mal, autistic and Asperger children. Res Autism Spectr Disord
6:193–203
7. Lombardo MV, Chakrabarti B, Bullmore ET, Baron-Cohen S
Summary (2011) Specialization of right temporo-parietal junction for men-
talizing and its relation to social impairments in autism. Neuroim-
age 56:1832–1838
ASD and autism are common developmental disorders that
8. Dawson G, Carver L, Meltzoff AN, Panagiotides H, McPartland J,
typically effect social and behavioral functioning, including Webb SJ (2002) Neural correlates of face and object recognition
the ability to comprehend other’s emotions and intentions in young children with autism spectrum disorder, developmental
though FER. This study provides evidence that TD children delay, and typical development. Child Dev 73:700–717
9. Lahaie A, Mottron L, Arguin M, Berthiaume C, Jemel B, Saum-
are more proficient in recognizing unfamiliar facial emotion
ier D (2006) Face perception in high-functioning autistic adults:
expressions than children with ASD, particularly for nega- evidence for superior processing of face parts, not for a configural
tive emotions (sad, fear). Preliminary evidence also demon- face-processing deficit. Neuropsychology 20:30–41
strated that children with ASD have adept and competent 10. Celani G, Battacchi MW, Arcidiacono L (1999) The understand-
ing of the emotional meaning of facial expressions in people with
emotional processing when viewing their mother’s faces, as
Autism. J Autism Dev Disord 29:57–66
the two groups did not differ in recognizing familiar expres- 11. Capps L, Yirmiya N, Sigman M (1992) Understanding of simple
sions. This result is insightful as previous neurological and and complex emotions in non-retarded children with autism. J
eye-tracking investigations have hinted at this possibility, Child Psychol Psychiatry 33:1169–1182
12. Robel L, Ennouri K, Piana H, Vaivre-Douret L, Perier A, Flament
but it has rarely been demonstrated at the conscious level
MF et al (2004) Discrimination of face identities and expressions
(i.e. self-report). Future large-scale studies should seek in children with autism: same or different? Eur Child Adolesc
to determine if this effect is also pertinent to expressions Psychiatry 13:227–233

13
Child Psychiatry & Human Development

13. Begeer S, de Rosnay M, Fink E, Wierda M, Koot HM (2014) 25. Kaufmann JM, Schweinberger SR (2004) Expression influences
Accuracy and response time for the recognition of facial emotions the recognition of familiar faces. Perception 33:399–408
in a large sample of children with autism spectrum disorders. J 26. Yeung MK, Sze SL, Chan AS, Han YMY (2014) Altered right
Autism Dev Disord 44:2363 frontal cortical connectivity during facial emotion recognition in
14. Pierce K, Redcay E (2008) Fusiform function in children with an children with autism spectrum disorders. Res Autism Spectr Dis-
autism spectrum disorder is a matter of “who”. Biol Psychiatry ord 8:1567–1577
64:552–560 27. Tantam D, Monaghan L, Nicholson H, Stirling J (1989) Autistic
15. Dalton KM, Nacewicz BM, Johnstone T, Schaefer HS, Gerns- children’s ability to interpret faces: a research note. J Child Psy-
bacher MA, Goldsmith HH, Davidson RJ (2005) Gaze fixation chol Psychiatry 30:623–630
and the neural circuitry of face processing in autism. Nat Neurosci 28. Nuske HJ, Vivanti G, Dissanayake C (2014) Reactivity to fearful
8:519 expressions of familiar and unfamiliar people in children with
16. Pierce K, Haist F, Sedaghat F, Courchesne E (2004) The brain autism. J Neurodev Disord 6:14
response to personally familiar faces in autism: findings of fusi- 29. Widen SC, Russell JA (2010) Differentiation in preschooler’s cat-
form activity and beyond. Brain 127:2703–2716 egories of emotion. Emotion 10:651–661
17. Enticott PG, Kennedy HA, Johnston PJ, Rinehart NJ, Tonge BJ, 30. Kliemann D, Dziobek I, Hatri A, Steimke R, Heekeren HR (2010)
Taffe JR et al (2014) Emotion recognition of static and dynamic Atypical reflexive gaze patterns on emotional faces in autism
faces in autism spectrum disorder. Cognit Emot 28:1110–1118 spectrum disorders. J Neurosci 30:12281–12287
18. Herba CM, Benson P, Landau S, Russell T, Goodwin C, Lemche E 31. Schurgin MW, Nelson J, Iida S, Ohira H, Chiao JY, Franconeri
et al (2008) Impact of familiarity upon children’s developing facial SL (2014) Eye movements during emotion recognition in faces. J
expression recognition. Clin Child Psychol Psychiatry 49:201–210 Vision 14:14–15
19. Gagnon M, Gosselin P, Maassarani R (2014) Children’s ability to 32. Rydell AM, Berlin L, Bohlin G (2003) Emotionality, emotion
recognize emotions from partial and complete facial expressions. regulation, and adaptation among 5-to 8-year-old children. Emo-
J Genet Psychol 175:416–430 tion 3:30
20. Dawson G, Webb SJ, Carver L, Panagiotides H, McPartland J 33. Uljarevic M, Hamilton A (2013) Recognition of emotions
(2004) Young children with autism show atypical brain responses in autism: a formal meta-analysis. J Autism Dev Disord
to fearful versus neutral facial expressions of emotion. Dev Sci 43:1517–1526
7:340–359 34. Austin PC, Brunner LJ (2003) Type I error inflation in the pres-
21. Wong TK, Fung PC, Chua SE, McAlonan GM (2008) Abnor- ence of a ceiling effect. Am Stat 57:97–104
mal spatiotemporal processing of emotional facial expressions in
childhood autism: dipole source analysis of event-related poten- Publisher’s Note Springer Nature remains neutral with regard to
tials. Eur J Neurosci 28:407–416 jurisdictional claims in published maps and institutional affiliations.
22. Gilliam JE (2006) Gilliam autism rating scale: GARS 2. Pro-ed,
Austin, TX
23. Tottenham N, Tanaka JW, Leon AC, McCarry T, Nurse M, Hare
TA et al (2009) The NimStim set of facial expressions: judgments
from untrained research participants. Psychiatry Res 168:242–249
24. Mancini G, Agnoli S, Baldaro B, Ricci Bitti PE, Surcinelli P
(2013) Facial expressions of emotions: recognition accuracy and
affective reactions during late childhood. J Psychol 147:599–617

13

You might also like