Re-examining emotion recognition in children with autism spectrum disorder using multimodal and dimensional strategies

• In Event: Social and Emotional Wellbeing for Differently-abled Individuals and their Families

Abstract

The unique social and communication profile central to autism spectrum disorder (ASD) is often ostensibly associated with a different style of emotion recognition (ER) compared to the general population (Uljarevic & Hamilton, 2012). The evidence for this suggested area of difference has typically been based on results of studies where emotions are presented in faces and voices and in which the participants respond using forced-choice verbal label responses. The objectives of this study were to examine whether these differences would be found when the stimuli were presented 1) in the context of music, an area of interest and potential strength among persons with ASD, and 2) with a paradigm that involved dimensional responses to emotional arousal and valence in addition to the traditional forced-choice responses.

Twenty-five children with ASD and 23 typically-developing (TD) children between the ages of 6-13 years (M=10.39, SD=1.52) with a wide range of verbal IQ scores (51-128) completed tasks associated with identifying emotions in faces, vocalizations, and music. Each modality included 8 happy, 8 sad, and 8 fearful emotional expressions, presented for 1.5-2 seconds, from validated stimuli sets for a total of 72 stimuli (Montreal Affective Voices: Belin et al., 2008; Musical Emotional Bursts: Paquette et al., 2013; NimStim Set of Facial Expressions: Tottenham et al., 2009). In the forced-choice response condition, the participants selected the verbal label that best described the stimulus from among these three possible expressions. In the dimensional response condition, the participants provided arousal and valence ratings on emoticon-based Likert scales.

Overall median accuracy ratings in the forced-choice response condition did not differ significantly between the groups (77-92% accuracy), suggesting that within this paradigm, ER was not an area of difficulty for children with ASD. Additional analyses revealed that overall, the participants with ASD and the TD participants more accurately identified emotions from voices than faces or music (ps < .01), and that happy and sad emotions were more accurately identified than fearful emotions (ps < .01). Similarly, in the dimensional response condition, both groups generally showed similar patterns of rating the overall emotional arousal and valence of the stimuli, across emotions and modalities. Both groups rated happy stimuli as more energetic (arousal) and more positive (valence) than fear and sad (ps < .0001). Exceptionally, the children with ASD rated happy stimuli as more positively valenced than did the TD children. Finally, accounting for verbal mental age, which differed at the group level, did not significantly influence the results. Overall, these findings are contrary to the generally accepted literature on reduced ER capabilities of children with ASD, and necessitate a re-examination of ER capabilities consistent with a shifting paradigm in which the diversity of styles of processing and ways of being of persons with ASD are emphasized.

Authors

• Shalini Sivathasan, University of Pittsburgh School of Medicine
  Presenting Author

• Hadas Dahary, McGill University
  Non-Presenting Author

• Jacob Ari (Jake) Burack, McGill University
  Non-Presenting Author

• Eve-Marie Quintin, McGill University
  Non-Presenting Author