Variables entered into the respective models.

Empathy is a core feature of the human social experience, underpinned by the sharing of and understanding of others’ states. However, we know relatively little about its early development. Here, we used the emerging technique of infrared thermography, combined with behavioural observations, to investigate how young children (1–3 years old) respond to others’ distress. We measured the emotional reactions – including changes in facial skin temperature and behavioural responses – of N = 30 typically developing children from British nurseries while they watched short video-vignettes of familiar and unfamiliar caregivers displaying emotional distress or neutral expressions. We hypothesised that children’s facial temperature change and behavioural responses (including facial expressions and self-directed behaviours) would be stronger to distress than to neutral stimuli. Based on the social bias of empathy, we hypothesised that responses would be stronger towards familiar individuals as compared to unfamiliar. Results confirmed that children’s thermal response at the nose bridge was stronger when observing someone in distress compared to the neutral control. Although model familiarity had only a weak effect, there was an effect of age, with younger infants showing stronger thermal responses to distress stimuli compared to older children. We found no effect on behavioural responses. We discuss implications of these results, including whether thermal-imagery may be a more sensitive measure of affective responding than behavioural measures, and the importance of using multiple methods. By combining physiological and behavioural measures, this study contributes to theoretical and methodological advances into understanding how internal affective processes map onto external measures in early childhood, shedding light on the development of empathy.