The Impact of Associated Anxiety Symptoms on Emotion Recognition in Individuals with Autism Spectrum Disorder

Background: Some autistic individuals experience emotion recognition difficulties, as compared to neurotypical populations. Furthermore, anxiety disorders have also been related to emotion recognition difficulties, potentially due to heightened state arousal in response to some expressions, such as high intensity emotions/threat cues. Despite the high prevalence of anxiety symptoms in autism – over 40-50% - few studies have examined the impact of co-occurring anxiety on emotion recognition.

Objectives: To investigate the relationship between anxiety symptoms and emotion recognition in autism.

Methods: 612 6-30-year-olds with (N=363) and without (N=249) Autism Spectrum Disorder (ASD) from the EU-AIMS Longitudinal European Autism Project were included. At Time 1, anxiety symptoms were measured using the Development and Wellbeing Assessment (DAWBA), with a ‘high anxiety’ cut-off implemented (see Goodman et al., 2010). 105 participants (29%) in the ASD group had high anxiety. Emotion recognition difficulties were assessed using an emotion-matching paradigm (Palermo et al., 2013), whereby participants selected the non-matching expression from an array of three faces, oriented left/right, or front-facing. At Time 2 (~18 months later) anxiety symptoms were re-assessed using the ‘emotional symptoms’ subscale from the Strengths and Difficulties Questionnaire (SDQ). Concordance between anxiety measures was good (r>0.63, p<2.2e-16). Emotion recognition difficulties were assessed in a basic emotion recognition paradigm using the Karolinska Directed Emotional Faces stimuli, including neutral, positive (e.g. happy) and negative (e.g. angry) expressions. We report effect sizes for Mann-Whitney U tests (r) and Spearman’s correlations (r_s).

Results: At Time 1, individuals with ASD and low, compared to high, DAWBA-rated anxiety did not differ on age (effect size r=0.01, p=0.92), nor IQ (effect size r=0.05, p=0.38). A higher proportion of females than males had high anxiety (X²=5.65, p=0.02, φ=0.12). Individuals with ASD and high anxiety were significantly less accurate at emotion-matching than those with lower anxiety on front-facing trials (effect size r=-0.30, p=0.01; Figure 1). In other words, the relationship between anxiety symptoms and emotion recognition – confirmed by simple correlations (r_s=-0.27, p=0.02) - was only present when stimuli were oriented to the front, with direct gaze. At Time 2, higher self-reported anxiety symptoms, according to the SDQ, were also marginally related to basic emotion recognition difficulties, specifically for trials depicting anger (r_s=-0.23, p=0.05), a socially and evolutionarily relevant threat cue. This effect was significant at the p<0.05 threshold but did not survive Bonferroni correction.

Conclusions: We found that co-occurring anxiety symptoms in ASD are related to emotion recognition accuracy for stimuli with direct gaze and displaying threat cues (anger). The ability to recognise and respond to the emotional expressions of others is fundamental for building successful interpersonal relationships. Therefore, our results emphasise that the management of co-occurring anxiety symptoms in ASD is important, not only for enhancing the quality of life of autistic individuals, but also facilitating social functioning and development. Future research should incorporate state arousal measures (e.g. heart rate) to further elucidate how anxiety symptoms modulate reactivity to social stimuli in ASD. This is particularly relevant given that stimuli with direct gaze/threat cues increase physiological responsiveness in neurotypical populations.