Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2024 Apr 22.
Published in final edited form as: J Autism Dev Disord. 2020 Jan;50(1):30–41. doi: 10.1007/s10803-019-04223-6

Attention Modification to Attenuate Facial Emotion Recognition Deficits in Children with Autism: A Pilot Study

Andrea Trubanova Wieckowski 1, Susan W White 2
PMCID: PMC11034769  NIHMSID: NIHMS1539897  PMID: 31520245

Abstract

Diminished attending to faces may contribute to the impairments in emotion recognition and expression in autism spectrum disorder (ASD). The current study evaluated the acceptability, feasibility, and preliminary efficacy of an attention modification intervention designed to attenuate deficits in facial emotion recognition (FER). During the 10-session experimental treatment, children (n=8) with ASD watched dynamic videos of people expressing different emotions with the facial features highlighted to guide children’s attention. Children and their parents generally rated the treatment as acceptable and helpful. Although FER improvement was not apparent on task-based measures, parents reported slight improvements and decreased socioemotional problems following treatment. Results suggest that further research on visual attention retraining for ASD, within an experimental therapeutic program, may be promising.

Keywords: Autism Spectrum Disorder, Facial Emotion Recognition, Eye-Tracking, Attention Training

Recognition of nonverbally expressed emotion is fundamental for successful social communication and interaction (Ekman, 1992; Nuske, Vivanti, & Dissanayake, 2013). Facial emotion recognition (FER) emerges early in life, with newborns being able to discriminate happy, sad, and surprised facial expressions (Field, Woodson, Greenberg, & Cohen, 1982). In addition, by the preschool years, children can accurately label most basic emotions (Widen & Russell, 2003) and it is during this time that FER deficits in autism become apparent, specifically with difficulty recognizing certain expressions (Rump, Giovannelli, Minshew, & Strauss, 2009). Scambler et al. (2007) found that children with ASD between 17 and 34 months already show less change in affect in response to facial expression than their typically developing peers. By 10 years of age, children with ASD lag behind typically developing peers at labeling basic expressions (e.g., Lindner & Rosén, 2006). Most studies show that adolescents and adults with ASD are not impaired in recognizing basic, prototypical emotions (Capps, Yirmiya, & Sigman, 1992; Grossman, Klin, Carter, & Volkmar, 2000), but impairment is apparent when stimuli are more subtle or complex, and when presented briefly (Humphreys, Minshew, Leonard, & Behrmann, 2007), such as the natural emotions that are encountered in everyday social interactions.

Deficits in facial emotion recognition and expression greatly affect quality of social interactions. These difficulties affect the ability of individuals to understand emotions of others as well as express emotions to others, both of which are essential for successful social interactions (Nuske, Vivanti, & Dissanayake, 2013). In a meta-analysis on the relationship between emotion recognition abilities and everyday social functioning in ASD, Trevisan and Birmingham (2016) found moderate but significant association between FER and social behavior in several studies. For example, Williams and Gray (2013) found that accuracy in recognition of sadness was associated with better socialization, above and beyond the influence of age, cognitive ability, or autism symptom severity. The reported associations between deficits in emotion recognition and social competence in children with ASD suggests the importance of addressing FER deficits in this population. Given advancements in technology and prior research suggesting that technology-based intervention might be more reinforcing to individuals with ASD (e.g., Goldsmith & LeBlanc, 2004), several neurotechnologies (cf. White et al., 2015), including computer-based programs, have been developed to address FER deficits in ASD. Studies have shown that such interventions are promising, though further research is needed in evaluation of their effectiveness (e.g., Kauo & Egel, 2016; Lee, Lam, Tsang, Yuen, & Ng, 2018).

While many studies have explored FER impairment in individuals with ASD, more recently research has focused on exploring mechanisms behind these impairments, utilizing eye-tracking and electroencephalography in order to gain insights into the attentional and neurological processes involved in FER. Results from these studies are inconsistent but, collectively, suggest impairment in some aspects of attention. In a review of eye tracking studies in ASD, Guillon and colleagues (2014) found that while the majority of the extant research indicates decreased visual attention to social stimuli in individuals with ASD, the degree of attention varies across contexts. Several studies show that the eye gaze patterns of people with ASD are characterized by more attending to inanimate objects and non-social features of the environment relative to socially relevant stimuli (e.g., faces and, in particular, the eye region) (Chevallier et al., 2015; Hanley et al., 2015; Klin, Jones, Schultz, Volkmar, & Cohen, 2002; Fletcher-Watson, Leekam, Benson, Frank, & Findlay, 2009). Notably, the mixed findings reported (e.g., Guillon et al., 2014), might be due to differences in the stimuli that are utilized. Hanley and colleagues (2013), for example, found that while individuals with ASD did not show diminished attention when facial stimuli were viewed in isolation, they showed diminished attention to the eyes, in comparison to typical peers, when the faces were part of social scenes.

Greater attending to non-social environmental stimuli and therefore lower fixation to the social stimuli, and in particular the eye region, is associated with lower social competence (e.g., Jones, Carr, & Klin, 2008), and avoidance of, and reduced orientation to, social stimuli have been shown to predict poorer emotional recognition (Kliemann, Dziobek, Hatri, Steimke, & Heekeren, 2010). This body of research suggests that individuals with ASD may have difficulties understanding others’ emotions if they do not attend to the important facial features. Two hypotheses have been presented to explain the reduced attention to other’s facial features, specifically eyes, in individuals with ASD. One hypothesis is that individuals with ASD purposefully look away because they find the eyes to be aversive (i.e., gaze aversion; Hutt & Ounsted, 1966; Kliemann et al., 2010), while the other hypothesis is gaze indifference, or looking less due to viewing the stimuli as unengaging or uninformative (Davies, Dapretto, Sigman, Sepeta, & Bookheimer, 2011; Senju & Johnson, 2009). Interventions that direct attention to the facial features may be successful in improving FER abilities in individuals with ASD, regardless of whether the root of the impairment is poor motivation (i.e., gaze indifference) or social aversion/arousal, as long as it is successful in increasing attention to the facial region.

Forms of attention modification have been utilized with populations outside of ASD, mainly with individuals with anxiety disorders (e.g., Hakamata et al., 2010; Linetzky, Pergamin-Hight, Pine, & Bar-Haim, 2015). Attention Bias Modification Treatment (ABMT), similar to the approach utilized in this pilot trial, attempts to guide the user’s visual attention. In ABMT, attention is focused toward very briefly presented benign or non-threatening stimuli, whereas in this intervention, the stimulus is presented for a longer period in an attempt to draw the youth’s overt visual attention. Results from the social anxiety literature suggest that gaze patterns can be altered (e.g., Linetzky et al., 2015). However, alteration of eye gaze patterns has not been directly explored in the service of improving FER in children with ASD. Therefore, the goal of the study was to utilize attention modification to increase focus on social features (i.e., face area), and eye gaze analyses to target the potential mechanism behind the FER impairment. We sought to develop and evaluate an attention modification program to address FER deficits in children with ASD.

Multiple studies have reported greater visual latency to faces in individuals with ASD (e.g., Freeth, Foulsham, & Chapman, 2011; Wilson, Brock, & Palermo, 2010), indicating difficulty with orienting toward the social stimuli (i.e., faces). Therefore, this study aims to increase orienting toward the facial region in order to explore the effect of greater visual attending to stimuli on FER ability. We assert that greater attention to others’ faces should translate to improved FER, which may ultimately translate to reduction of core symptoms of ASD (i.e., greater social competence) in children with ASD.

The primary aim of this pilot study was to assess feasibility, operationalized as low attrition (i.e., no more than 1 person dropping out) and at least 80% of the child participants and their parents reporting the intervention to be acceptable. The primary therapeutic aim of this attention modification intervention was to alter the gaze patterns. To assess preliminary efficacy, we examined change in both the targeted mechanism (attention to social stimuli), operationalized as total fixation duration to facial features of the stimuli from pre to post intervention, and change in the distal outcome of interest – FER, as well as secondary clinical outcomes (i.e., socio-emotional problems, ASD severity).

Method

Participants

Participants included children (total n = 8) between the ages of 9 and 12 years, inclusive (see Table 1). In order to be eligible, all participants had to evidence difficulties with FER (based on parent report) and have at least low average intellectual abilities (IQ score greater than 75 on the 2-subtest Weschler Abbreviated Scale of Intelligence, 2nd Edition; Wechsler, 2011). In addition, participants were required to have received a clinical diagnosis of ASD, which was confirmed by a research reliable administration of the Autism Diagnostic Observation Schedule, Second Edition (ADOS-2; Lord et al., 2012). No participants presented with severe co-occurring psychopathology, such as psychotic symptoms or severe aggression. In addition, children who were receiving psychosocial therapy aimed at addressing FER at the time of study were excluded.

Table 1.

Demographic Characteristics (n = 8)

Mean (SD)
Age (in years) 10.93 (1.29)
FSIQ-2 110.89 (13.60)
SRS-2 78.13 (6.75)
ADOS Total Score 9.11 (2.20)
N (%)
Gender
 Male 5 (62.5)
 Female 3 (37.5)
Race
 Caucasian 8 (100)
 Hispanic/Latino 1 (12.5)
Diagnoses1
 ASD 8 (100)
 Anxiety 1 (12.5)
 OCD 0
 ADHD 3 (37.5)
 Depression 0
 ID/LD 1 (12.5)
Receiving Services2 5 (62.5)
Medications
 None 6 (75.0)
 ADHD 1 (12.5)
 Asthma/Allergy 1 (12.5)
FER/FEE Difficulties
 Difficulty Recognizing Emotions 8 (100)
 Limited Range of Facial Expressions 8 (100)
 Expressions not appropriate to situation 4 (50.0)
Open in a new tab
1

Diagnoses of non-ASD disorders were inferred based on parent report on the demographics questionnaire.

2

Services included peer support, Occupational Therapy, and counseling which did not address FER.

Measures

Affect Recognition subtest of the Developmental Neuropsychological Assessment (NEPSY-II: Korkman, Kirk, & Kemp, 2007).

The Affect Recognition (AR) subtest of the NEPSY-II assesses a child’s ability to recognize affect from colored photographs of children’s faces. AR has been used in prior studies to measure deficits in FER in children with ASD (e.g., Williams, Gray, & Tonge, 2012). It provides age-based standard scores, with lower scores indicating poorer FER ability. AR subtest was administered to participants during the pre-treatment and post-treatment sessions.

Emotion Regulation and Social Skills Questionnaire, Parent Version (ERSSQ; Beaumont & Sofronoff, 2008).

ERSSQ is a parent measure designed to evaluate emotion regulation and social skills in children and adolescents with ASD. It consists of 27 items evaluated in terms of frequency of skill or behavior of the child. Parents completed the ERSSQ during the weekly treatment sessions, in order to evaluate potential changes in emotion regulation and social skills throughout the treatment.

EU-Emotion Stimulus Set (O’Reilly et al., 2016).

The EU-Emotion Stimulus Set contains video clips of children and adults expressing different emotions and mental states. The set contains stimuli of facial expressions, body gesture scenes, and contextual social scenes. Only facial expressions were utilized for the current study, and only videos with at least 60% accuracy rating based on O’Reilly and colleagues (2016) were included. Twelve videos, two of each of the six basic emotions (i.e., fear, anger, disgust, happiness, sadness, and neutral) were shown to the participant during each session: one video of high intensity, and one of low intensity, and videos were randomized across and within the sessions. The order for the EU videos was the same, however, for each participant for a specific session.

Social Responsiveness Scale, Second Edition, Parent Version (SRS-2; Constantino & Gruber, 2012).

The SRS-2 is comprised of 65 items measuring parent report of ASD-related social impairments. SRS-2 assesses social awareness, social cognition, social communication, social motivation, and restricted interests and repetitive behavior. SRS-2 was completed by the parents during the intake and endpoint sessions. While SRS-2 asks the parents to focus on the behavior during the past 6 months, at the endpoint session, parents were asked to rate their child’s behavior since the start of the intervention.

Treatment Satisfaction Form.

Participants and their parents completed a treatment satisfaction form, on which they were asked to rate the statements regarding whether they found the intervention to be helpful, acceptable in terms of time and commitment, whether they would recommend the program to a friend, and whether overall, they liked the program on a 5-point scale, ranging from “strongly disagree” to “strongly agree.” In addition, parent and child were asked to rate whether the child’s ability to recognize emotions has “greatly decreased”, “slightly decreased”, “stayed the same”, “slightly increased” or “greatly increased.” Lastly, the parents were asked what changes they have noticed regarding their child’s FER skills.

Video Emotion Recognition Task.

In order to evaluate change in FER abilities, participants completed a video emotion recognition task, in which they viewed a video of an adult expressing one of six basic expressions (happiness, sadness, fear, anger, surprise, and disgust). After the presentation of each video, participants were asked to tell the examiner which emotion they saw portrayed in the video. Each expression was presented 6 times, for a total of 36 videos. Each video lasted 2.73 seconds (Wieckowski & White, 2017), provide further information regarding the task). The Video Emotion Recognition Task was administered to the participants during the pre-treatment and post-treatment sessions.

Youth Top Problems (YTP).

The YTP measure was administered to the parents during the intake session and during weekly treatment sessions, in order to evaluate changes in identified social difficulties during treatment. During the intake session, parents were asked to identify three behaviors related to nonverbal socioemotional interaction (e.g., FER) that are causing the child most difficulty. During the treatment sessions, the parents were asked to provide ratings from 0 to 10, to indicate ‘how big of a problem’ the identified difficulties were for the child since the prior visit (i.e., within the last 72 hours).

Apparatus and Stimuli

Eye gaze.

Eye-tracking was completed using a Tobii T60 XL eye tracker. Participants sat approximately 60 cm from the eye tracker and they were instructed to look at the video stimuli on the screen. Prior to displaying each stimulus, a centered “X” (1.5 cm × 1.5 cm wide) appeared on screen and stimuli were only presented after the participant successfully attended to the centered “X”. The eye-tracking system was calibrated to each participant’s eyes prior to data collection using a five-point calibration procedure (i.e., a red circle moving to five predefined locations across the screen, including the four corners and the center of the screen). The experimenter visually inspected the tracking to all five points and corrected any missing or excessive error tracking before advancing the participant to the task. Eye-tracking data were collected during the treatment session for the EU-Emotion Stimulus Set and the Intervention Task (described below). Participant’s eye gaze patterns and fixations were collected through Tobii studio and analyzed using a Matlab code (MatlabR2014b, Mathworks Inc., MA).

Attention Modification Intervention.

Video clips for the intervention task were selected if they met the following criteria: 1) publicly available; 2) at least one actor expressed one of the following emotions: happiness, sadness, anger, disgust, fear, and neutral expression (surprise was not shown as outcome FER measures did not assess surprise); 3) taken from television shows and movies not currently in mainstream child entertainment, in order to decrease chances of familiarity with the clip. A total of 290 clips were identified. Prior to use in this study, videos were validated, using an undergraduate sample, to ensure they adequately captured the target emotion. Video stimuli with 60% or better agreement on the emotion shown were included as the treatment videos, except for three disgust videos. For disgust, because of lower overall mean agreement (M = 72.87%), a slightly lower threshold of 55% was used.

The final stimuli set for the intervention task consisted of 170 clips of social interactions where at least one character was expressing one of the six emotions. Seventeen clips (three per emotion except for disgust, which had 2 clips per session) were presented per session and each clip was only presented once during treatment; the order of the stimuli remained constant within session across participants, so that all participants viewed the same order of the videos for each session. The videos were exactly five seconds in length followed by a black screen with options for the emotions displayed in order to allow participants enough time to tell the examiner which emotion was displayed in the video. The order was semi-random in terms of each session having high/medium/low rated videos (based on agreement from a stimuli validation study). However, the videos were randomized within and across sessions.

The faces of the actors in the video were surrounded by a square box (white dotted line) in order to draw attention to the socially relevant information (i.e., the social stimulus that conveys the target emotion), consistent with prior research using this approach (e.g., Richler, Floyd, & Gauthier, 2014) as well technological approaches being developed (e.g., Kinect system highlighting the face in a square) to “read” the facial expression. As the child proceeded through the program, the number of videos with a square box decreased so that for each subsequent session, the percentage of videos with the cue was less than in the prior session. Therefore, while in the first session all videos (100%) had the facial area highlighted, the last session included only two videos (12%) with the visual cue. This fading procedure is adapted from behavioral intervention techniques for individuals with development delays, to eliminate need for continued prompting and therefore reduce prompt dependency (Lovaas, 2003). At the start of each session, participants were instructed to watch each video and were informed that they were going to be asked questions regarding the video afterward. At the end of each video they watched, the participants were asked which emotion was displayed in the clip, with options presented in print on the screen. The participants verbally labelled the emotion which the experimenter recorded for later analyses. While viewing the stimuli, participants’ eye gaze was recorded using a Tobii T60 XL eye-tracker, which tracked their eye-gaze in real time.

Procedure

Following a brief phone screen, participants were scheduled for an in-person appointment to assess preliminary eligibility. All eligible participants attended 10 twice-weekly individual sessions, each one lasting approximately 20 minutes (including welcoming, treatment session, parent measures). This timing has been shown to be tolerable to participants for the ABMT protocols (e.g., Ollendick et al., 2019). In addition to the intake session, data were collected at the start of treatment (i.e., pre-treatment assessment) and at the end of treatment (i.e., post-treatment assessment). Participants were compensated $20 for completion of the intake, pre-treatment and post-treatment assessment sessions, in addition to $20 for completion of all treatment sessions, for a maximum total of $80.

Data Analyses

For the eye tracking data collected during the intervention sessions, Tobii Studio was used to analyze the data. The Areas of Interest (AOIs) included the face region, predefined using the oval-shaped AOI tool, and background region, predefined using a rectangle-shaped AOI tool, available in the Tobii T60 (Studio Professional) platform. In addition, the box highlighting the facial region was predefined, to allow for analyses regarding whether participants looked within the box, as instructed. Total duration of fixations made to these regions was calculated. Any trial showing a major loss of tracking was excluded from data analyses. On-task percentage score above 50% is a common benchmark for inclusion of participants within ASD eye-tracking studies (e.g., Fischer et al., 2014; Swanson, Serlin, & Siller, 2013). Therefore, on-task percentage scores above 25% for the entire task and above 50% per stimulus were used. Eye gaze was analyzed within session and within person, in addition to across sessions, in order to explore change in eye gaze to face area. Duration of fixations made to these regions was calculated using in-house Matlab code.

Descriptive statistics were computed using means and standard deviations for continuous variables, and frequencies and proportions for categorical variables. Descriptive statistics were utilized to assess feasibility and acceptability of the program. Reliable Change Indices (RCI; Jacobson & Truax, 1991) were calculated in order to determine whether the amount of change in FER (i.e., AR task from NEPSY-II) and other clinically relevant outcome measures (i.e., SRS-2, ERSSQ) was statistically significant at the individual level. RCI assesses for the amount of change and whether the change reflects more than just the fluctuation due to an imprecise measurement. The recommended cutoff of the RCI is 1.96, with scores above 1.96 indicating statistically significant change. The scores of the outcome measures were averaged for each time point separately, and then these scores were used to calculate the RCI for each distinct outcome measure. A Wilcoxon Signed Rank Test, a non-parametric statistical test, was used to compare initial and endpoint scores for all participants. The Wilcoxon Signed Rank Test is an alternative to a paired t-test, which assumes normal distribution. Given the small sample size, normal distribution cannot be assumed. An effect size was calculated using Excel for all Wilcoxon Signed Rank Test analyses with the following formula: r = Z/sqrt(N) where N is the total number of samples (Rosenthal, 1994). Given the small sample size and preliminary nature of the study, in addition to significance testing, we also focus on effect size interpretation. Per Rosenthal’s guidelines (1994), a small effect size [r] is 0.1, medium effect size is 0.3, and large effect size is 0.5. Data were analyzed with IBM SPSS Statistics (SPSS 24.0).

For the multiple time point data collected during the intervention phase (i.e., ERSSQ, Intervention Task Accuracy), linear mixed effect model was fit. R package (Package, http://www.R-project.org/) was used to fit a linear mixed model, which can accommodate both the variation among measurements within individuals and the individual-to-individual variation. The model was used to compare treatment effects across the intervention time points. In addition, for the YTP, as each parent reported different problems and therefore the scores are not equivalent across participants, a linear model was fit to understand the relationship between time point and problem rating for each of the three problems that parents provided for each participant individually. An estimate and variance were calculated for each participant, for each problem rating, in addition to the t-test statistic and p value to detect change in rating over time.

Results

Feasibility

In terms of viability of the intervention, results support feasibility and acceptability of the program. Nine eligible participants were enrolled, eight of whom completed the study. One participant dropped after the first treatment session due to difficulty getting to the center (~1 hour each way) for the treatment twice a week. In addition, one of the eight participants who completed the intervention sessions (ID 8) completed the study in a non-protocol timeframe with 3 weeks of treatment instead of 5, due to the child travelling out of the country for an extended period of time. Sessions were still kept on non-consecutive days, but this participant averaged three sessions, instead of two sessions, per week. Analyses were completed with and without this participant to see whether the abbreviated intervention timeframe effects the results. Since all of the results remained unchanged with the exclusion of this participant, the reported results include this participant. The other seven participants completed the program as designed, with two sessions a week, for five weeks, with no consecutive session days. Table 1 displays descriptive statistics for the 8 participants that completed the intervention. In terms of acceptability, all children and parents provided generally positive ratings on the Treatment Satisfaction form regarding acceptability of the program (Child: M = 3.88, SD = 0.35; Parent: M = 4.38, SD = 0.52) as well as other questions related to acceptability. See Table 2 for a summary.

Table 2.

Number of Responses and Average Ratings on the Treatment Satisfaction Form

Question Responder Response Selected Average Rating
Neutral Agree Strongly Agree
“This program is acceptable” Child 1 7 M = 3.88
SD = 0.35
Parent 5 3 M = 4.38
SD = 0.52
“Overall, this program seems helpful” Child 5 3 M = 3.38
SD = 0.52
Parent 2 5 1 M = 3.88
SD = 0.64
“I would recommend this program to a friend” Child 3 4 1 M = 3.75
SD = 0.71
Parent 1 6 1 M = 4.00
SD = 0.53
“Overall, I like the program” Child 4 3 1 M = 3.63
SD = 0.74
Parent 7 1 M = 4.13
SD = 0.35
Stayed the Same Slightly Increased Greatly Increased
“At this point, my ability to recognize emotions has:” Child 2 5 1 M = 3.88
SD = 0.64
“At this point, my child’s ability to recognize emotions has:” Parent 3 5 M = 3.63
SD = 0.52
Open in a new tab

Note. Number of children or parents that selected each response is reported for each question. No children or parents selected “strongly disagree” or “disagree” option for the first four questions or “greatly decreased” or “decreased” options for the last question, and therefore those options are not portrayed. Responses were averaged by assigning a point value for each question on a 5-point scale, ranging from “1 = strongly disagree” to “5 = strongly agree” for first four questions and “1 = greatly decreased” to “5 = greatly increased” for the last question.

Preliminary Efficacy – Change in Amount of Fixation Duration to Facial Features

Although all participants attained calibration prior to eye-tracking, not all of the participants showed acceptable on-task percentage scores. Across the six participants for whom tracking was available during the attention modification intervention, on average, while looking at the screen when the box was present, participants looked within the box 84.93% of the time, suggesting that participants followed the instruction.

During the EU task, eye tracking was collected for all participants and every participant showed at least a 50% on-task percentage for at least one of the treatment sessions. Only one participant (ID 2) was excluded as his last treatment session eye tracking fell below threshold. Wilcoxon signed-rank test indicated that fixation durations to the face area for the last treatment session were not statistically higher than those for the first treatment session (Z = −1.69, p = .09). The medium to large effect size (r = −.45), however, indicates that a change in eye gaze to face occurred. This change was in the unanticipated direction, with lower fixation duration to the face as the sessions progressed.

Similarly, looking at analyses across all sessions, the linear mixed effect model also did not show a significant linear relationship for fixation duration to the face region (β1 = −43.75, SE = 40.65, t = −1.08, p =.29). In order to evaluate the relationship between fixation duration to the face region and time for each participant separately, linear models were fit. The relationship between fixation duration to the face region and time was not significant for any of the participants (all t < 1.83, all p > .13).

Change in FER

A Wilcoxon signed-rank test indicated that scores on the NEPSY II AR task at the post-intervention assessment were not statistically higher than those in the pre-intervention assessment (Z = −.17, p = .86, r = .04). However, RCI analyses showed that one participant (ID 6) showed reliable increase on the AR task (RCI = 2.57). None of the other seven participants showed a reliable change (Table 3). A Wilcoxon signed-rank test indicated no significant difference between pre-treatment and post-treatment scores on the Video Emotion Recognition task (Z = −.70, p = .48, r = .18). In addition, no significant difference was observed for any of the individual emotions (all p > .06). No individual subject analyses were completed for the Video Emotion Recognition task as no standardized norms exist for this measure.

Table 3.

Reliable Change Indices for Outcome Variables [Baseline to Endpoint]

ID AR SRS-2 ERSSQ
1 BL 11 112 44
EP 10 108 47
RCI −.37 −.30 .63
2 BL 12 114 64
EP 13 107 73
RCI .45 −.53 1.89
3 BL 9 108 49
EP 9 102 51
RCI 0 −.45 0.42
4 BL 10 74 52
EP 12 73 48
RCI .90 −.08 −.83
5 BL 13 99 69
EP 12 73 83
RCI −.45 −1.97* 2.93*
6 BL 3 99 56
EP 10 77 56
RCI 2.57* −1.67 0
7 BL 10 92 42
EP 9 52 78
RCI −.37 −3.03* 7.55*
8 BL 12 108 53
EP 10 97 71
RCI −.90 −.83 3.77*
Open in a new tab

Note. BL = baseline, EP = endpoint, RCI = Reliable Change Index Norms for RCI for AR obtained from Korkman et al., 2007, SRS-2 (Norms obtained from Constantino & Gruber, 2012), ERSSQ (Norms from Beaumont & Sofronoff, 2008; alpha was used instead of test-retest),

*

Reliable improvement (RCI ≤ or ≥ 1.96)

For the EU-Emotion Stimulus Set Task, across all individuals, a Wilcoxon signed-rank test indicated no significant difference between the last treatment session and the first treatment session scores (Z = −1.05, p = .29, r = .26). Across all ten sessions, a linear mixed effect model did not show a significant linear relationship between task accuracy and time points. While the intercept was significant, the slope fixed effects were not (β0 = 9.58, SE = 0.80, t = 11.91, p <.001; β1 = 0.08, SE = 0.04, t = 1.78, p = 0.08), indicating that accuracy on the EU-Emotion Stimulus Set Task and time point does not have a significant positive linear relationship with slope across participants. In order to evaluate the relationship between accuracy on the EU-Emotion Stimulus Set Task and time point for each participant separately, a linear model was fit. None of the individual participant relationships were significant (all t < 1.73, p > .12).

With respect to FER accuracy on the Intervention Video Task, a Wilcoxon signed-rank test indicated that accuracy scores at the last treatment session were statistically higher than those in the first treatment session (Z = −2.23, p = .03, with large effect r = .56). Looking across all ten sessions, a linear mixed effect model showed a significant linear relationship between the accuracy on the task and time points across participants. Both intercept and slope fixed effects were significant (β0 = 10.87, SE = 0.81, t = 13.39, p <.001; β1 = 0.23, SE = 0.07, t = 3.43, p <.01). These results suggest that accuracy on the Intervention Video Task and time point have a significant positive linear relationship with slope across participants. A liner model showed that none of the individual participant relationships between accuracy on the Intervention Task and time point were significant (all t < 1.73; p > .12), even though combining across all participants, the model was significant, accounting for the variability across subjects. The individual variability likely resulted in a larger error term, which deemed the individual patterns to not be significant.

On the Treatment Satisfaction Form, participants and their parents indicated generally positive ratings regarding perceived changes in the child’s FER (Child: M = 3.88, SD = 0.64; Parent: M = 3.63, SD = 0.52; see Table 2 for details). For the question of “What changes (if any) have you noticed regarding your child’s emotion recognition skills?” two parents did not provide an answer, one parent noted they have not noticed any differences, while the other five saw changes from “more facial expression done now”, “seems better at noticing non-verbal cues”, “is better able to recognize his own emotions and those of others”, “begun to ask family members about their feelings more and react to other’s reactions more appropriately”, and “seems to be more thoughtful – seems to have a better recognition that (child) has separate, distinct emotions.”

Change in Secondary Clinical Outcomes

A Wilcoxon signed-rank test indicated that the total scores on the SRS-2 at the post-intervention assessment were statistically lower than those in the pre-intervention assessment (Z = −2.37, p = .02, r = .59). RCI analyses show that only two participants showed reliable decrease (corresponding to lesser social impairment). Although all participants had lower endpoint SRS-2 scores than at baseline, none of the other participants showed a reliable change (Table 3).

For the ERSSQ, a Wilcoxon signed-rank test indicated that accuracy scores for the intervention videos at the last treatment session were not statistically higher than those in the first treatment session (Z = −1.86, p = .06, r = −.46). The effect sizes, however, suggest moderate effect of the intervention on the ERSSQ scores. In addition, RCI analyses revealed that three of the participants showed significantly improved accuracy from the first session to the last treatment session (Table 3). In addition, looking at analyses across all ten sessions, linear mixed effect model showed a significant linear relationship between the scores on the ERSSQ and time point. Both intercept and slope fixed effects were significant (β0 = 52.46, SE = 3.77, t = 13.90, p <.001; β1 = 1.18, SE = 0.42, t = 2.78, p <.01, respectively). These results suggest that ERSSQ score and time point have a significant positive linear relationship across participants. The linear model, evaluating the relationship between ERSSQ and time for each participant separately, revelated that the relationship between ERSSQ and time was significant for four participants (ID 2: t = 3.10, p = .01; ID 5: t = 4.84; p <.01; ID 7: t = 4.19, p < .01; ID 8: t = 4.29, p < .01).

Each parent was able to identify three difficulties on the YTP. The identified problems fell into four categories: 1) Recognition of emotions and social cues (e.g., recognition of peer reaction, difficulty recognizing nonverbal cues, not recognizing emotions), 2) Emotion expression (e.g., flat affect, overexpression of emotion), 3) Emotion regulation (e.g., managing frustration, overreaction to other’s words and emotions), and 4) Social skills (e.g., difficulty with back and forth conversation, unable to talk to strangers). Linear model results for each problem identified, for each parent, are displayed in Table 4. As can be seen from the table, all but two of the participants show a significant relationship between problem rating and time point for at least one problem behavior, with three participants showing a significant relationship for all three identified problems. These results suggest that parent-reported problems with social-emotional behavior improved as treatment progressed.

Table 4.

Relationship Between Problem Behaviors Identified on the YTP and Time for Each Participant

ID Problem Estimate SE t-value p-value
1 1. EE 0.00 0.00 1.73 .12
2. Rec −0.02 0.08 −0.24 .81
3. SS 0.01 0.03 0.30 .77
2 1. ER 0.15 0.13 1.18 .27
2. EE −0.10 0.05 −1.98 .08
3. Rec −0.17 0.05 −3.30 .01*
3 1. Rec −0.31 0.05 −6.53 <.001*
2. Rec −0.25 0.05 −5.11 <.001*
3. EE −0.18 0.05 −4.05 <.01*
4 1. ER −0.16 0.04 −4.63 <.01*
2. Rec −0.15 0.06 −2.46 .04*
3. Rec −0.29 0.08 −3.49 <.01*
5 1. Rec 0.02 0.04 0.41 .69
2. ER −0.04 0.05 −0.71 .49
3. Rec 0.03 0.05 0.55 .60
6 1. Rec −0.18 0.06 −3.16 .01*
2. EE −0.08 0.09 −0.91 .39
3. SS −0.23 .08 −2.80 .02*
7 1. Rec −0.33 0.04 −7.56 <.001*
2. EE −0.37 0.07 −5.40 <.001*
3. Rec −0.31 0.07 −4.15 <.01*
8 1. SS −0.22 0.08 −2.60 .03*
2. ER −0.41 0.06 −6.33 <001*
3. EE −0.11 0.08 −1.36 .21
Open in a new tab

Note. EE = emotion expression; Rec = recognition of emotions and social cues; ER = emotion regulation; SS = social skills.

*

designates significant relationship

Discussion

This study sought to examine the feasibility and preliminary impact of a novel attention retraining program to alter gaze toward social stimuli (faces) in order to attenuate FER deficits in children with ASD. Results suggest that the new attention modification program is feasible and acceptable to children as well as their parents. In terms of feasibility, 89% of children successfully completed the program and all but one child rated the program to be acceptable (one child indicated ‘neutral’) and all parents noted that the program is acceptable.

In terms of preliminary efficacy, or impact on gaze to socially relevant cues (target engagement), we did not find statistically significant change in gaze over treatment. While the paradigm manipulation, which highlighted the facial area in the videos, was successful in terms of children looking within the area when the box was presented, the manipulation did not successfully alter the gaze during other viewing tasks. This may be due to a host of factors including the stimuli used for the study, or the chosen prompt withdrawal timeframe. The stimuli were dynamic, naturalistic, and of varying intensity levels, so that the intervention stimuli approximated emotions encountered in natural interactions. However, while the preliminary validation of the stimuli ensured that the stimuli depicted a specific emotion, it is unclear whether the stimuli were sufficiently interesting, as no pilot data was collected from the participants to determine salience or level of interest. In addition, it is possible that the withdrawal timeframe, or fading procedure was too brief, as only the first session included a visual cue for all of the videos, with the tenth session including only two videos with the cue. It is also possible that there is perhaps other mechanisms behind the change in FER. Of note, participants were not provided feedback on their responses during the intervention in order to determine whether gaze training alone may be sufficient to exert an effect on FER. If the participants were given feedback, it would not be possible to disentangle if FER impairment is rooted in attentional processes or deficiency in reward processes. In this study, we wanted to test the hypothesis that there is deficiency in attention to social stimuli leading to FER impairment (e.g., Dadds, El Masry, Wimalaweera, & Guastella, 2008), and therefore participants were not provided feedback.

While we hypothesized that there would be a significant improvement in participants’ FER ability from pre to post intervention, the results are not consistent across the different measures. On the only standardized measure of FER utilized in this study (AR subtest of NEPSY-II), one participant showed a reliable increase in FER score. Importantly however, this participant was the only child in the study to score below average on this measure prior to the intervention. All other participants scored in the average or above average range prior to the intervention; therefore, ceiling effects may have impacted our ability to see movement on this measure. At the group level, performance on the Video Emotion Recognition task did not change. Although participants significantly increased their performance on the Intervention Video Task, the individual participant trajectories showed that no participants showed a consistent increase in accuracy over time. This could be due to variable difficulty of emotion identification throughout the sessions or due to actual fluctuation of the scores (i.e., instability of the measure). While the results suggest general improvement on the group level for some of the FER measures, the subject-level data suggest the intervention was not successful at increasing FER on these measures. Overall, given these results, it is possible that this type of an intervention, focusing on increasing gaze to faces, may only be useful for children who exhibit low FER ability across measures, as ceiling effects may impact how much change in gaze may lead to an increase in FER.

Although the majority of the participants did not show impairment in FER on the AR subtest, all parents indicated impairment in FER at the start of treatment and 5 of the 8 reported improvement in FER at endpoint. While parent report is not always accurate, prior studies show that parents of children with ASD are variable in their reports of emotional difficulties (e.g., (White et al., 2018), suggesting usefulness of parent report in assessing for the deficits. FER as a construct may vary and present differently both across and within individuals, suggesting that discrete behaviors related to FER (e.g., response time) are needed. Results from this study as well as variable findings in prior studies suggest that current FER measures may lack ecological validity, as all parents in the study reported their children showing FER impairment.

Although change in FER was variable across measures, all eight subjects showed a decrease in the SRS-2 total score, with a large effect (r = −.59), indicating decreased social impairment after intervention; however, the decrease was reliable, per RCI, for only two of the participants. On the ERSSQ, which detects emotion recognition and social skills, four parents reported improvement over the 10 sessions. In addition, on the YTP, which assesses parent identified behavioral change, six parents rated at least two of the identified problems to have decreased over the course of the study. As with the FER measures, this result suggests the importance of incorporating parent ratings of problems they identify is most relevant to their children in addition to the normed measures, as the most consistent change occurred for the questions that were tailored to each child (i.e., identified by the parent). While most of these outcome measures showed a significant improvement for at least one child, it is important to point out that improvements on different measures were found for different children.

These findings should be considered in light of the limitations of this study, of which the primary one is lack of any objective measure of FER impairment for inclusion. We relied solely on parent report of difficulty, and found average to above average FER abilities for all but one participant on the standardized measure of FER (i.e., AR task on the NEPSY-II). Even though all parents reported FER difficulties, these difficulties were not observed on the standardized measure, limiting the ability to observe any significant change. This is not a design failure per se, as there are no established measures that are sensitive to FER change in this population (Wieckowski, Flynn, Richey, Gracanin, & White, 2019). In addition, it is possible that the observed changes in FER, mainly on the parent report, occurred due to potential increased awareness of the FER difficulties given enrollment in the intervention. Since no changes in gaze occurred, it is possible that the change in FER occurred due to other mechanism. Further exploration with larger samples and a control condition is needed to more fully understand the factors leading to such changes. Additionally, within an experimental therapeutic approach, it is important to show that it is possible to change the most proximal mechanism, which theoretically leads to secondary changes. We made a preliminary step toward this by showing that participants looked within the attention getting area (i.e., box) the majority of the time, but lack of control prohibits firmly concluding movement in the targeted mechanism. Finally, feasibility in this study was assessed only with consumer acceptability and attrition. While these are the primary and most common indicators of feasibility, there are other indicators of feasibility that need to be explored in future research (e.g., therapist training, transportability, cost).

While the findings of this pilot study indicate feasibility and acceptability of a novel attention retraining intervention to improve FER in children with ASD, the impact on gaze patterns, FER, and clinical outcomes is less clear. In addition to addressing the limitations highlighted above, given the variability in FER deficiency among children with ASD, it is important to ensure stability of FER impairment which may require creation of new measures to assess FER impairment in individuals with ASD that are able to detect actual change in the process following an intervention.

Conclusion

This study was the first attempt to develop and implement a brief attention retraining program to attenuate FER deficits in children with ASD. Results indicate the program is feasible to implement as planned and acceptable to families. Moreover, we see improved parent-reported FER and diminished parent-identified socioemotional problem behaviors. Change in FER as measured via behavioral tasks is not apparent, except when FER was considerably below average prior to the intervention. In addition, there was no change in gaze to socially-relevant cues following the intervention, which is inconsistent with the hypothesized mechanism behind the intervention in increasing gaze to socially relevant stimuli. Further exploration into FER, and externally valid measurement approaches, is necessary. Prior to further assessment of feasibility of attention retraining in individuals with ASD, it may also be beneficial to assess the degree to which FER is modifiable in a tightly controlled experimental design to assure that the construct is able to be altered.

Acknowledgments

This work was funded by Organization for Autism Research Graduate Research Grant and the Routh Research and Dissertation Grant through Society of Clinical Child and Adolescent Psychology, APA Division 53. We are grateful to the children and families who participated in this study. We also greatly appreciate the help from Stephanie Roldan, who has aided in the development of the Matlab code to analyze the eye-tracking data.

Footnotes

Publisher's Disclaimer: This Author Accepted Manuscript is a PDF file of an unedited peer-reviewed manuscript that has been accepted for publication but has not been copyedited or corrected. The official version of record that is published in the journal is kept up to date and so may therefore differ from this version.

The authors report no potential conflicts of interest. The research involved human participants. Informed consent was obtained prior to data collection.

References

  1. Beaumont R, & Sofronoff K (2008). A multi-component social skills intervention for children with Asperger syndrome: The junior detective training program. Journal of Child Psychology and Psychiatry, 49(7), 743–753. doi: 10.1111/j.1469-7610.2008.01920 [DOI] [PubMed] [Google Scholar]
  2. Capps L, Yirmiya N, & Sigman M (1992). Understanding of simple and complex emotions in non-retarded children with autism. Journal of Child Psychology and Psychiatry, 33, 1169–1182. [DOI] [PubMed] [Google Scholar]
  3. Chevallier C, Parish-Morris J, McVey A, Rump KM, Sasson NJ, Herrington JD, & Schultz RT (2015). Measuring social attention and motivation in autism spectrum disorder using eye-tracking: Stimulus type matters. Autism Research, 8(5), 620–628. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Constantino JN, & Gruber C (2012). The Social Responsiveness Scale – Second edition (SRS-2). Torrance, CA: Western Psychological Services. [Google Scholar]
  5. Dadds MR, El Masry Y, Wimalaweera S, & Guastella AJ (2008). Reduced eye gaze explains “fear blindness” in childhood psychopathic traits. Journal of the American Academy of Child & Adolescent Psychiatry, 47(4), 455–463. [DOI] [PubMed] [Google Scholar]
  6. Davies MS, Dapretto M, Sigman M, Sepeta L, & Bookheimer SY (2011). Neural bases of gaze and emotion processing in children with autism spectrum disorders. Brain and Behavior, 1(1), 1–11. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Ekman P (1992). An argument for basic emotions. Cognition & Emotion, 6(3–4), 169–200. [Google Scholar]
  8. Field TM, Woodson RW, Greenberg R, & Cohen C (1982). Discrimination and imitation of facial expressions by neonates. Science, 218, 179–181. [DOI] [PubMed] [Google Scholar]
  9. Fischer J, Koldewyn K, Jiang YV, & Kanwisher N (2014). Unimpaired attentional disengagement and social orienting in children with autism. Clinical Psychological Science, 2, 214–223. doi: 10.1177/2167702613496242 [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Fletcher-Watson S, Leekam SR, Benson V, Frank MC, & Findlay JM (2009). Eye-movements reveal attention to social information in autism spectrum disorder. Neuropsychologia, 47, 248–257. [DOI] [PubMed] [Google Scholar]
  11. Freeth M, Foulsham T, & Chapman P (2011). The influence of visual saliency on fixation patterns in individuals with Autism Spectrum Disorders. Neuropsychologia, 49(1), 156–160. [DOI] [PubMed] [Google Scholar]
  12. Goldsmith TR, & LeBlanc LA (2004). Use of technology in interventions for children with autism. Journal of Early and Intensive Behavior Intervention, 1(2), 166–177. [Google Scholar]
  13. Grossman JB, Klin A, Carter AS, & Volkmar FR (2000). Verbal bias in recognition of facial emotions in children with Asperger syndrome. Journal of Child Psychology and Psychiatry, 41, 369–379. [PubMed] [Google Scholar]
  14. Guillon Q, Hadjikhani N, Baduel S, & Rogé B (2014). Visual social attention in autism spectrum disorder: Insights from eye tracking studies. Neuroscience and Biobehavioral Reviews, 42, 279–297. [DOI] [PubMed] [Google Scholar]
  15. Hakamata Y, Lissek S, Bar-Haim Y, Britton JC, Fox NA, Leibenluft E, … & Pine DS (2010). Attention bias modification treatment: A meta-analysis toward the establishment of novel treatment for anxiety. Biological Psychiatry, 68(11), 982–990. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Hanley M, McPhillips M, Mulhern G, & Riby DM (2013). Spontaneous attention to faces in Asperger syndrome using ecologically valid static stimuli. Autism, 17(6), 754–761. [DOI] [PubMed] [Google Scholar]
  17. Hanley M, Riby DM, Carty C, McAteer AM, Kennedy A, McPhillips M (2015). The use of eye-tracking to explore social difficulties in cognitively able students with autism spectrum disorder: A pilot investigation. Autism, 1–6. doi: 10.1177/1362361315580767 [DOI] [PubMed] [Google Scholar]
  18. Humphreys K, Minshew N, Leonard GL, & Behrmann M (2007). A fine-grained analysis of facial expression processing in high-functioning adults with autism. Neuropsychologia, 45, 685–695. [DOI] [PubMed] [Google Scholar]
  19. Hutt C, & Ounsted C (1966). The biological significance of gaze aversion with particular reference to the syndrome of infantile autism. Behavioral Science, 11(5), 346–356. [DOI] [PubMed] [Google Scholar]
  20. Jacobson NS, & Truax P (1991). Clinical significance: A statistical approach to defining meaningful change in psychotherapy research. Journal of Consulting and Clinical Psychology, 59, 12–19. [DOI] [PubMed] [Google Scholar]
  21. Jones W, Carr K, & Klin A (2008). Absence of preferential looking to the eyes of approaching audits predicts level of social disability in 2-year-old toddlers with autism spectrum disorder. Archives of General Psychiatry, 65(8), 946–954. [DOI] [PubMed] [Google Scholar]
  22. Kauo JL, & Egel AL (2016). The effectiveness of interventions in teaching emotion recognition to children with autism spectrum disorders. Review Journal of Autism and Developmental Disorders, 3, 254:265. [Google Scholar]
  23. Kliemann D, Dziobek I, Hatri A, Steimke R, & Heekeren HR (2010). Atypical reflexive gaze patterns on emotional faces in autism spectrum disorders. The Journal of Neuroscience, 30(37), 12281–12287. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Klin A, Jones W, Schultz R, Volkmar F, & Cohen D (2002). Visual fixation patterns during viewing of naturalistic social situations as predictors of social competence in individuals with autism. Archives of General Psychiatry, 59, 809–816. doi: 10.1001/archpsyc.59.9.809 [DOI] [PubMed] [Google Scholar]
  25. Korkman M, Kirk U, & Kemp S (2007). NEPSY-II: A developmental neuropsychological assessment. San Antonio, TX: The Psychological Corporation. [Google Scholar]
  26. Lee CS, Lam SH, Tsang ST, Yuen CM, & Ng CK (2018). The effectiveness of technology-based intervention in improving emotion recognition through facial expression in people with autism spectrum disorder: A systematic review. Review Journal of Autism and Developmental Disorders, 5(2), 91–104. [Google Scholar]
  27. Lindner JL, & Rosen LA (2006). Decoding of emotion through facial expression, prosody, and verbal content in children and adolescents with Asperger’s syndrome. Journal of Autism and Developmental Disorders, 36, 769–777. [DOI] [PubMed] [Google Scholar]
  28. Linetzky M, Pergamin-Hight L, Pine DS, & Bar-Haim Y (2015). Quantitative evaluation of the clinical efficacy of attention bias modification treatment for anxiety disorders. Depression and Anxiety, 32(6), 383–391. [DOI] [PubMed] [Google Scholar]
  29. Lord C, Rutter M, DiLavore PC, Risi S, Gotham K, & Bishop SL (2012). Autism diagnostic observation schedule (ADOS-2) (2nd ed.). Torrence, CA: Western Psychological Services. [Google Scholar]
  30. Lovaas OI (2003). Teaching individuals with developmental delays: Basic intervention techniques. Austin, TX: Pro-Ed. [Google Scholar]
  31. Nuske HJ, Vivanti G, Dissanayake C (2013). Are emotion impairments unique to, universal, or specific in autism spectrum disorder? A comprehensive review. Cognition & Emotion, 27(6). [DOI] [PubMed] [Google Scholar]
  32. Ollendick TH, White SW, Richey J, Kim-Spoon J, Ryan SM, Wieckowski AT, … & Smith M (2019). Attention bias modification treatment for adolescents with social anxiety disorder. Behavior Therapy, 50(1), 126–139. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. O’Reilly H, Pigat D, Fridenson S, Berggren S, Tal S, Golan O, … & Lundqvist D (2016). The EU-emotion stimulus set: A validation study. Behavior Research Methods, 48(2), 567–576. [DOI] [PubMed] [Google Scholar]
  34. Richler JJ, Floyd RJ, & Gauthier I (2014). The Vanderbilt Holistic Face Processing Test: A short and reliable measure of holistic face processing. Journal of Vision, 14(11), 10–10. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Rosenthal R (1994). Parametric measures of effect size. In Cooper H & Hedges LV (Eds.), The Handbook of Research Synthesis (231–244). New York, NY: Sage. [Google Scholar]
  36. Rump KM, Giovannelli JL, Minshew NJ, & Strauss MS (2009). The development of emotion recognition in individuals with autism. Child Development, 80(5), 1434–1447. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Scambler DJ, Hepburn S, Rutherford MD, Wehner EA, & Rogers SJ (2007). Emotional responsivity in children with autism, children with other developmental disabilities, and children with typical development. Journal of Autism and Developmental Disorders, 37(3), 553–563. [DOI] [PubMed] [Google Scholar]
  38. Senju A, & Johnson MH (2009). Atypical eye contact in autism: Models, mechanisms and development. Neuroscience & Biobehavioral Reviews, 33(8), 1204–1214. [DOI] [PubMed] [Google Scholar]
  39. Swanson MR, Serlin GC, & Siller M (2013). Broad autism phenotype in typically developing children predicts performance on an eye-tracking measure of joint attention. Journal of Autism and Developmental Disorders, 34, 707–718. doi: 10.1007/s10803-012-1616-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Trevisan DA, & Birmingham E (2016). Are emotion recognition abilities related to everyday social functioning in ASD? A meta-analysis. Research in Autism Spectrum Disorders, 32, 24–42. [Google Scholar]
  41. Wechsler D (2011). Wechsler Abbreviated Scale of Intelligence–Second Edition (WASI-II). San Antonio, TX: NCS Pearson. [Google Scholar]
  42. White SW, Abbott L, Wieckowski AT, Capriola-Hall NN, Aly S, & Youssef A (2018). Feasibility of automated training for facial emotion expression and recognition in autism. Behavior Therapy. [DOI] [PubMed] [Google Scholar]
  43. White SW, Richey JA, Gracanin D, Bell MA, LaConte S, Coffman M, … & Kim I (2015). The promise of neurotechnology in clinical translational science. Clinical Psychological Science, 3(5), 797–815. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Widen SC, & Russell JA (2003). A closer look at preschoolers’ freely produced labels for facial expressions. Developmental Psychology, 39(1), 114–128. [DOI] [PubMed] [Google Scholar]
  45. Wieckowski AT, Flynn LT, Richey JA, Gracanin D, & White SW (2019). Measuring change in facial emotion recognition in individuals with Autism Spectrum Disorder: A systematic review. Manuscript submitted for publication. [DOI] [PMC free article] [PubMed]
  46. Wieckowski AT, & White SW (2017). Eye-gaze analysis of facial emotion recognition and expression in adolescents with ASD. Journal of Clinical Child & Adolescent Psychology, 46(1), 110–124. [DOI] [PubMed] [Google Scholar]
  47. Williams BT, & Gray KM (2013). The relationship between emotion recognition ability and social skills in young children with autism. Autism, 17(6), 762–768. [DOI] [PubMed] [Google Scholar]
  48. Williams BT, Gray KM, & Tonge BJ (2012). Teaching emotion recognition skills to young children with autism: A randomised controlled trial of an emotion training programme. Journal of Child Psychology and Psychiatry, 53(12), 1268–1276. [DOI] [PubMed] [Google Scholar]
  49. Wilson CE, Brock J, & Palermo R (2010). Attention to social stimuli and facial identity recognition skills in autism spectrum disorder. Journal of Intellectual Disability Research, 54(12), 1104–1115. [DOI] [PubMed] [Google Scholar]

RESOURCES