Vocal emotion recognition (VER) indexed by behavior and neural response in adolescents with and without ASD

• In Event: Poster Session 02
  In Poster Session: PS 02 Topic - Developmental Disabilities

Abstract

Background: Emotion recognition is an essential component of navigating social relationships. Autism spectrum disorder (ASD) is characterized by impairments in social communication (APA, 2013) with some individuals demonstrating difficulty with emotion recognition (Uljarevic, 2012). While considerable research has examined facial emotion recognition abilities in children with ASD (Harms, 2010), there has been less focus on vocal emotion recognition (VER). Previous behavioral VER findings have been mixed, with results indicating impairment (Globerson, 2015), no impairment (Grossman, 2010), or only impairment in recognizing low intensity emotions (Golan, 2006). The few studies examining the time course of VER, utilizing event-related potentials (ERPs; N100, P200), suggest the N100 (indexing orientation of auditory attention) differs in ASD (Korpilathi, 2007) and is related to decreased behavioral performance (Lerner, 2013). The P200 amplitude indexes emotional, compared to neutral, voices (Paulmann, 2008; Wang, 2015), and some evidence suggests the P200 is modulated by specific emotions (Iredale, 2013). However, it is not clear whether VER ERPs differ by emotional intensity in individuals with ASD. This study examined differences in errors and reaction time, as well as N100 and P200 ERPs, during VER, and whether differences may be driven by low versus high intensity emotional prosody in adolescents with or without ASD.

Methods: A standardized VER task (Table 1) was completed by adolescents (N=161) with and without ASD (as determined by the Autism Diagnostic Observation Schedule-2 ADOS-2; Lord, 2012) while EEG was recorded. Group differences on categorical and continuous data were compared (Table 1).

Results: When identifying vocal emotions, adolescents with ASD made more errors (p=.002, d=.44) and reacted more slowly (p=.04, d=.31) compared to adolescents without ASD. Additionally, adolescents with ASD had a significantly slower N100 latency (Figure 1; p=.04, d=.32). No differences between N100 amplitude, P200 latency, or P200 amplitude emerged (ps>.15).

Regarding the intensity of VER, adolescents with ASD made more VER errors in both low (p=.002, d=.50) and high (p=.03, d=.27) intensity conditions. Adolescents with ASD demonstrated at least a marginally slower reaction time during low (p=.02, d=.37) and high (p=.07, d=.22) intensity VER. Compared to adolescents without ASD, adolescents with ASD demonstrated a marginally slower N100 latency to high (p=.08, d=.31), but not low (p=.58) intensity emotion.

Conclusions: Adolescents with ASD made more errors in identifying emotional prosodic cues, had a slower reaction time in their identification, and a slower N100 latency. While behavioral findings cut across emotional intensities, the N100 differences was specific to high intensity emotion. Unexpectedly, the N100, not the P200, differed by intensity. The N100 is conceptualized as a pre-emotion, sensory ERP (Schirmer, 2006) that is modulated by attention (Ho, 2015), so individuals with ASD may allocate reduced attentional resources to high intensity social stimuli. This may be due to reduced attention to high intensity voices throughout development shaping N100 latency (Social Motivation Hypothesis; Chevallier, 2012; Dawson, 2007).

Authors

• Talena Day, Stony Brook Unviersity
  Presenting Author

• Sydney Boateng, Stony Brook University
  Non-Presenting Author

• Matthew D Lerner, Stony Brook Unviersity
  Non-Presenting Author