Does empathy predict autistic traits? A multi-modal investigation

The empathy-autistic trait relationship is complex and controversial. Empathy research is often undermined by unidimensional approaches to defining and measuring a multidimensional construct. This study regresses autistic trait level variation onto cognitive and emotional empathy variables across modalities (self-report, behavioural, and physiological) using a non-clinical adult sample (N = 123). We also account for several important covariates (age, sex, IQ, years of education and visual attentional behaviour). Self-report (trait-based) cognitive and emotional subscales predicted AQ more strongly and frequently than lab-based measures (behavioural/physiological; state-based), in directions consistent with prior findings. Emotional empathy questionnaire subscales differed in predictive directionality; a finding discussed with reference to state-trait measurement divergences and problematic subscales such as the personal distress scale of the Interpersonal Reactivity Index, which may contribute inconsistency in emotional empathy findings. Visual (in)attention was a consistent covariate predictor of AQ, highlighting the importance of accounting for gaze in autism trait studies to avoid erroneous conclusions. Findings are discussed with reference to the challenge of bridging the gaps between lab, theory, and application. Results further support the broader autism phenotype notion and contribute to debate concerning the relationship between empathy and autistic traits.