Naturalistic Assessment of Empathy and Social Cognition in Adolescent ASD – Eye-Tracking As Predictor of Performance and Behavioral Phenotypes in Clinical and General Populations

Background: Empathy and social cognition are the basis of our social functioning and they are deviating in autism spectrum disorders (ASD; Harms et al., 2010; Bons et al., 2013; Senju, 2013). Questionnaire measures or abstract tests often fail to capture deviating empathy and social cognition in adolescent ASD, which can be attributed to compensating strategies in lab experiments that are revealed by eye-tracking methodology (Senju et al., 2009; Chevallier et al., 2015). Aberrant empathy in ASD may also reflect the endpoint of a trait continuum in the normal population that is different to aberrant empathy in conduct disorder

Objectives: The Multifaceted Empathy Test – Junior Revised (MET-JR) and the Movie for the Assessment of Social Cognition (MASC) were adapted for adolescent age ranges and investigated in three independent samples (MET-JR clinical sample: n = 57, MASC clinical sample: n = 56, MET-JR community sample: n = 215).

Methods: The original MET is a picture-based measure of cognitive and emotional empathy (Dziobek et al., 2007), while the MASC is a video-based behavioral test of inferring others' mental and emotional states (Dziobek et al., 2006). The MASC was assessed with concurrent eye-tracking. We conducted factor and psychometric analysis of our naturalistic measures in comparison to common questionnaires and tests (MASC vs. SRS, RMET, and EQ; MET-JR vs. SRS, GEM, and ICU).

Results: The MET-JR and the MASC were validated as ecologically-valid estimate of empathy (α = .71-.96) or social cognition (α = .74-.84) as they correlated significantly with related measures (r = .39 - .53) and associated constructs (r = -.37 - .59), but not significantly with cognitive ability or age. Both measures were able to differentiate ASD from clinical control groups, with AUC = .87 and AUC = .75, and delivered large effects by comparing ASD and non-ASD groups, with d = 1.1 – 2.4 and d = 0.9. Concerning the MET-JR, distinct clinical empathy profiles found for ASD and conduct disorder were replicated as disjoint empathy continua in the community sample. Concerning the MASC, smaller pupil sizes were observed for ASD (d = 0.6) and fixation on eyes positively predicted performance (f² = 0.15). However, exploratory factor analysis retrieved a single-factor solution for the MASC with R² = .36, while confirmatory factor analysis of the MET-JR partially delivered insufficient fit indices (RMSEA = .066).

Conclusions: We delivered naturalistic assessment tools of the interrelated constructs of empathy (MET-JR) and social cognition (MASC). We showed aberrant performance in both measures for adolescent ASD that can be related to empathy trait continua in the normal population and be predicted by gaze behavior. Nonetheless, factor analyses revealed that we have not sufficiently understood the latent construct structure of empathy and social cognition that also show conceptual overlap (Schaafsma et al., 2015). This could be overcome by a replication with a larger sample assessed with both naturalistic measures, the MET-JR and the MASC.