Emotional Response Coherence in Toddlers with ASD

Background: The theory of emotions as tripartite complexes postulates that emotions are conveyed by facial expressions, vocalizations, and body postures (Izard, 1977). Expressive incoherence is the mismatch in the expression of emotions across channels. There is very limited data regarding coherence across expressive channels in ASD. A recent study suggests that in response to in-vivo probes eliciting frustration, school-age children with ASD show a mismatch between their facial and bodily expressions, especially during negatively-valenced events (Costa et al., 2017). This limited coherence across emotion expression channels could lead to problems in the parent-child dyad, as caretakers may have difficulty accurately reading the emotional state of children with ASD. Little is known about the existence of emotional expressive incoherence at the earliest stages when ASD can be reliably diagnosed.

Objectives: To examine if toddlers with ASD exhibit expressive incoherence compared with age-matched peers with developmental delays (DD) and typical development (TD) in response to in-vivo probes aimed at eliciting anger, fear, and joy.

Methods: 83 toddlers (ASD, n=43; TD n=40) ages 13-30 months (M=21) completed nine episodes of the Laboratory-Based Temperament Assessment Battery (Lab-TAB; Goldsmith & Rothbart, 1999). Episodes were designed to elicit Anger, Fear, and Joy. The assessments were videotaped and coded offline by blinded coders for intensity of emotional responses during Fear, Joy, and Anger probes separately in the facial, vocal, and musculoskeletal channels. Within each emotion, Pearson’s r correlations were used to compare the strength of associations across pairs of channels (facial-bodily, vocal-bodily, and facial-vocal) and Fisher r-to-z transformations were used to test the significance of the difference between correlations.

Results: In general, the correlations between the channels within ASD and TD groups were statistically significant with the magnitude of the effect ranging from r=.313 to r=.863 in the ASD group and from r=.339 to r=.692 in the TD group (Figure 1). Comparison of the magnitude to coherence (Pearson’s r) between the two groups identified only two differences, both in the Joy condition. Specifically, there were significant differences between the ASD and TD groups in two pairs: facial-vocal expression coherence (p=.02), and bodily-vocal expression coherence (p=.01). In both cases, the ASD toddlers evidenced higher emotional coherence than the TD toddlers.

Conclusions: Our results indicate that TD toddlers and toddlers with ASD largely evidence similar emotional coherence across the three channels of emotional expression during tasks designed to elicit anger, fear, and joy. Contrary to our hypotheses, the bodily-vocal channels and the facial-vocal channels appear more coherent among children with ASD than TD children in the Joy condition. Our data do not provide support for emotional response incoherence in ASD.