Positive, Negative, and Other Emotions in Young Autistic Children: The Importance of Context

Background:  Young autistic children are claimed to have excessively negative and dysregulated emotions, with pervasively reduced positive affect (Zwaigenbaum et al., 2013; Hirschler-Guttenberg et al., 2015). However, in preliminary data from the Montreal Stimulating Play Situation (MSPS; Jacques et al., 2015), young autistic and age-matched typical children displayed similar positive and negative emotions. These pilot findings raise questions about the importance of context and about autistic emotions perceived as “unknown” by typical observers (Yirmiya et al., 1989).

Objectives:  To assess larger groups of age-matched young autistic and typical children with MSPS, and to further document the context of their emotions.

Methods:  37 autistic (mean age=45.8 months, SD=10.5; MSEL=74.0, SD=27.7) and 39 typical (mean age=41.1 months, SD=14.1, p=0.124; MSEL=103.0, SD=24.8, p<0.001) children were assessed with MSPS. Four play periods (free-play 1, semi-free play, semi-structured play, free-play 2) with 40 objects of potential interest to autistic children were filmed by a trained cameraman. Two naïve typical raters coded positive, negative, neutral, and unknown emotions on Observer XT 11©. Duration, frequency, and proportion of children displaying each emotion were analysed for all play periods, and associated with object explorations and repetitive behaviors.

Results: For the full MSPS, there were no significant group differences in frequency, duration, and proportion of children displaying positive, negative, or neutral emotions (p’s>.09). Positive emotions were pervasive and negative emotions rare in both groups. However, emotions coded as unknown were observed in 43.2% of autistic vs 0% of typical children (p<.001). Autistic children thus displayed significantly greater frequency (mean=3.5, SD=9.0, p=.018) and duration (mean=14s, SD=37s, p=.019) of emotions coded as unknown, compared to their complete absence in typical children.

For individual play periods, there were no significant group differences for positive, negative, or neutral emotions (p’s>.12) except in semi-structured play, with greater duration of positive emotions in typical (mean=104s, SD=109s) vs autistic (mean=59s, SD=58s, p<.05) children. Unknown emotions were significantly more frequent in autistic children in semi-structured and free-play 2 periods (p’s<.05).

For emotions associated with objects explored by >75% of children, across both groups positive emotions were expressed 18.9-76.9% and negative emotions 0-5.1% of the time, with no significant group differences for any object (p’s>.06). Emotions coded as unknown, unique to autistics, were observed 2.7-13.5% of the time (p’s=.018-.301).

Finally, for emotions in autistic children associated with their most-observed repetitive behaviors, positive emotions were observed during arm movements, hand flapping, and close gaze at objects (16.2%, 24.3%, 16.2% of the time, respectively), as were unknown emotions (8.1%, 5.4%, 5.4%). Negative emotions were not observed during any of these behaviors.

Conclusions:  In a novel context of potential interest to them, autistic children expressed many positive and few negative emotions, particularly when exploring objects freely. Characteristically autistic repetitive behaviors co-occurred with positive, not negative, emotions. We did not find evidence of emotional dysregulation in young autistic children, compared to age-matched typical children with significantly higher MSEL scores. Autistic children uniquely expressed a range of emotions coded as unknown, suggesting there is room to improve our understanding of their full emotional repertoire.