Do Toddlers with ASD Express Emotions That Are Incongruent with Valence of Real-World Events?

Background: Affective expressions are critical for communication and social interaction (Ekman, 1982). Although early conceptualizations suggested diminished affective expressions in children with ASD (DSM-III-R, 1987), subsequent research revealed that children with autism are not less expressive (Capps et al., 1993); rather, they may display emotions that are incongruent with the environmental context (McGee, Feldman, & Chernin, 1991; Reddy, Williams, & Vaughn, 2002). Still, little is known regarding whether congruency of emotional expression depends on the valence of environmental events and whether incongruent responses are present regardless of the emotion expression channel (e.g., facial or vocal).

Objectives: 1) To examine intensity of incongruent facial and vocal expressions during positively- (Joy) and negatively- (Frustration and Fear) valenced probes of the Laboratory Temperament Assessment Battery (Lab-TAB; Goldsmith & Rothbart, 1996) among toddlers with ASD, developmental delays (DD), and typical development (TD); 2) To examine the relationship between the intensity of incongruent emotional expressions and severity of ASD symptoms.

Methods: Participants included 99 toddlers (ASD, n=43; DD, n=16; TD, n=40) (M_age=21 months). Nine in-vivo probes were presented to elicit three emotions: Frustration, Joy, and Fear. Sessions were coded offline by blinded coders for peak intensity of emotional expression from two channels: vocal and facial.

Results: Intensity of facial expressions and vocalizations of joy during fear- and anger-eliciting probes were contrasted between the three groups of children via linear mixed effects models. Vocal distress and facial expressions of fear and anger during joy eliciting probes were analyzed between the diagnostic groups using univariate ANOVAs. No group differences in facial expression of joy during negatively-valenced probes (F(2, 143)=2.17, p=.14) or group by task interaction were found. Similarly, no group differences in facial expression of anger (F(1,98)=.93, p=.40) or fear (F(1,98)=.99, p=.38) during Joy probes were found. However, when examining vocalizations, differences between groups emerged. A main effect of diagnosis (F(1, 130)=6.57, p=.002) showed that toddlers with ASD expressed more intense positive vocalizations than TD (p=.001, d=.48) and DD (p=.01, d=.49) toddlers during negatively-valenced probes and produced more intense distress vocalizations than TD (p=.03, d=.49) and DD (p=.05, d=.67) toddlers during Joy probes. In the ASD group, intensity of incongruent distress vocalizations during Joy probes was associated with ADOS Social Affect scores (r=.37, p=.02).

Conclusions: While the intensity of incongruous facial expressions of toddlers with ASD during positively and negatively-valenced probes did not differ from those observed in age-matched DD and TD peers, their vocalizations appeared more atypical. Specifically, their probe-incongruous distressed or positive vocalizations were more intense than those observed in controls. The results suggest ambiguity in vocal expressions in toddlers with ASD, which may result in affective miscommunication between the children and their caretakers and peers and hamper their efforts to convey their needs and interact socially. Examining the mechanisms underlying incongruous emotional expressivity in toddlers with ASD constitutes a novel and highly clinically and theoretically fruitful area of inquiry in autism research.