Emotion Perception during An Audio-Visual Emotion Perception Task: Differences Between Forced-Choice and Free Response Formats in Individuals with Autism Spectrum Disorders and Typically-Developing Individuals

Background: The accurate perception of facial expression and prosody are critical for successful social interactions. When information is perceived from multiple modalities, typically-developing individuals (TD) consistently show a bias toward facial emotion during forced-choice tasks (Massaro et al., 1996; Santorelli & Robins, 2006). Studies have shown that individuals with autism spectrum disorders (ASD) have difficulties understanding emotions expressed by others, even when emotion labels are provided for them (Lindner & Rosen, 2006). Individuals with ASD have also been shown to be less efficient than TD in facial perception tasks with visual-only and combined auditory-visual stimuli (Gepner et al., 1996). In another study, children with ASD are able to infer emotional states when provided with specific prompts, but they do not spontaneously use emotion terms to describe others (Serra et al., 1999). Taken together, these findings suggest that individuals with ASD may possess the ability to accurately interpret and identify emotions portrayed by others, but do not spontaneously look for or evaluate nonverbal cues.

Objectives: The current study investigates whether individuals with ASD show the same pattern of responses as TD individuals on free-response and forced-choice versions of a dynamic audio-visual emotion perception task. In addition, this study examines whether individuals with ASD, like TD individuals, are biased toward information from the visual modality when presented with conflicting emotional information from the auditory modality on a forced-choice format.

Methods: Participants were recruited as part of a larger study. The ASD group included 10 individuals (M = 19.8, range 12-49). The TD group included 8 individuals (M = 15.2, range 9-30). On the forced choice task, participants were presented with an emotion word cue (e.g. happy, angry) and asked to determine whether or not it correctly described the affect portrayed in the subsequent movie clip (i.e., yes/no). During the free response task, participants were presented with stimuli and subsequently asked to label the affect portrayed by the actor using a single emotion word. Only incongruent (i.e. happy face/angry voice) movie stimuli were utilized for this study.

Results: Paired-samples t-tests were conducted. When presented with incongruent movies in the forced-choice format, TD participants showed a strong bias towards the facial expression (t=3.000, p=.020, ŋ²=.56). The ASD group did not demonstrate such bias, instead equally relying on facial expression and prosody (t=1.432, p=.186, ŋ²=.19). When asked to provide one emotion word to describe the actor’s affect during a free response task, neither the TD group (t=.640, p=.543, ŋ²=.06) nor the ASD group (t= -.469, p=.651, ŋ²=-.03) showed a bias toward one modality.

Conclusions: Findings suggest that complex emotion perception impairments in individuals with ASD may result from an under-reliance on interpreting facial expressions when asked about a specific emotion. However, this may only be the case in forced-choice emotion perception tasks, which is a format frequently used in emotion research. Neither group showed a modality bias during the open-response format. Future research should consider potential differences in response formats in order to influence intervention strategies addressing the social difficulties in ASD.