Reduced Recognition of Dynamic Facial Emotional Expressions in Children with ASD

Background: A vast number of researchers focused on the processing of facial emotional expressions in ASD, using a variety of paradigms and methodologies, and resulting in inconclusive and mixed findings. The Emotion Recognition Task is an emotion labelling task which already proved to be valuable in detecting subtle emotion recognition peculiarities in several clinical and non-clinical samples. Furthermore, this paradigm demonstrated its usefulness in identifying facial emotion recognition impairments in adolescents with a clinical diagnosis of ASD and in non-clinical adults with ASD traits. Given that neurocognitive studies often examine school-aged children without an intellectual disability, it seemed important to extend the findings with the Emotion Recognition Task to that population. In line with the recent interest in a dimensional perspective on ASD and the search for endophenotypes, the relationship between demographic factors and emotion recognition was also evaluated.

Objectives: The aim of this study was to examine emotion recognition abilities in a large sample of children with and without ASD. Extending previous studies, we also wanted to investigate the relationship between participant characteristics – apart from diagnostic classification – and emotion recognition performance.

Methods: Emotion recognition abilities were measured with the Emotion Recognition Task, which is an often-used emotion labelling task with dynamic facial expressions of six basic emotions, namely anger, disgust, fear, happiness, sadness, and surprise. Two levels of emotional intensity were used: an intermediate and a high emotional intensity level. Performance was compared between 45 children with ASD and 50 typically developing children, group-wise matched on intelligence and age. Participants were aged 6 to 14 years and none of them had an intellectual disability. ASD traits were measured with the Social Responsiveness Scale (SRS) and empathy was measured with the Empathy Quotient (EQ).

Results: Results showed that the ASD group generally performed worse than the comparison group, and this impairment appeared emotion-specific: the ASD group was especially outperformed by the typically developing group when labelling sadness, disgust and surprise. However, the typically developing group performed better when identifying fear. Both groups were better at recognising high intensity expressions than intermediate ones. After correcting for response biases, the differential effect of emotions disappeared, but the overall poorer emotion recognition performance in the ASD group remained significant. Age nor intelligence were associated with emotion recognition. However, emotion recognition was correlated with SRS scores and EQ scores.

Conclusions: Extending previous work using the Emotion Recognition Task, we showed a poorer emotion recognition performance in 6-to-14-year old children with ASD. Moreover, we revealed that response biases could (partially) attribute to the emotion-specificity of their impairment. In addition, emotion recognition abilities were associated with ASD traits and empathy, such that children with higher ASD traits and lower empathy, had a poorer performance on the emotion recognition task.