Poster #117 - Recognition of Emotion from Lexical Meaning and Tone of Voice

• In Event: 2-110 - Poster Session 07
  In Poster Session: PS 07 Section - Language, Communication

Integrative Statement

Young children are known to have so-called “lexical bias”: the bias towards lexical meaning of emotion words rather than affective tone of voice to infer the emotion of the speaker (Morton & Trehub, 2001). Existing studies suggest that unlike adults, when the prosodic and lexical information are incongruent, children between 4 and 10 years of age have difficulty in inferring the emotional state of the speaker from tone of voice (Friend, 2000).

Autism spectrum disorder (ASD) is characterized by difficulties in everyday social interaction including recognition of emotion. Investigation on emotion recognition in the auditory domain among individuals with ASD, however, is still scarce compared to the visual domain, and their findings are mixed. In some studies, adults with ASD were found to have difficulty in understanding emotions by tone of voice when the lexical content was emotion-neutral (e.g. Philip et al. 2010), but not in other studies (e.g. Jones et al. 2011). In addition, adolescents with ASD were found to have strong lexical bias: their recognition of the speaker’s emotion was based on the lexical meaning than tone of voice (Stewart et al. 2013).

Currently, little is known about the recognition of vocal emotion in children with ASD. Our aim in the present study was to find out how their vocal emotion recognition differs from that of typically developing children. Twenty-five children with ASD (mean age=9.2 years) and language-matched 25 typically developing children (mean age=8.7 years) participated in the study. The stimuli consisted of 6 positive emotion words (PEW) (e.g. happy), 6 negative emotion words (NEW) (e.g. sad) and 6 emotion-neutral words (ENW) (e.g. carrot). These words were accompanied by either positive tone of voice (PTV) or negative tone of voice (NTV). Thus, the stimuli was made up of the following 6 conditions, each of which had 6 trials: PEW-PTV, PEW-NTV, NEW-PTV, NEW-NTV, ENW-PTV, ENW-NTV. Each stimulus word was presented aurally in random order while children were shown a pair of faces of a woman on the computer screen, one with positive emotion and the other with negative emotion. Children were asked to indicate which one of the faces was the speaker’s face when she uttered the word they just heard.

Our results suggest that the recognition of emotion from tone of voice by children with ASD was on a par with that of typically developing children (a) when the tone of voice and the meaning of emotion words was congruent and (b) when the word meaning was emotionally neutral. Both groups of children chose the facial expressions that matched tone of voice significantly more frequently than the alternative. By contrast, when the meaning of emotion words and the tone of voice were incongruent, the performance of the two groups was significantly different (p <.005). Typically developing children chose the face that matched the tone of voice whereas children with ASD chose the face that matched the meaning of the emotion words. The finding indicates that children with ASD have stronger lexical bias than typically developing children.

Authors

• Tomoko Matsui, Tokyo Gakugei University
  Presenting Author

• Naotake Tsukidate, Yamanashi-Eiwa University
  Non-Presenting Author

• Hiroshi Fujno, Tokyo Gakugei University
  Non-Presenting Author