Idiosyncratic face exploration strategy is influenced by face’s emotion type and observer’s empathic profile

Previous studies have shown that people have different face exploration strategies, and that different facial emotions elicit different gaze patterns. However, little is known about the factors driving these differences. Here, we hypothesized that the type of emotional facial expression has a stronger effect on the gaze patterns of more empathic individuals. We recorded the eye movements of 165 observers looking at neutral, sad, disgusted, and surprised faces and asked them to complete empathy questionnaires. First, we predefined regions-of-interest (ROI) and counted their fixation rates. We showed that the type of emotion had an effect on ROIs fixation rate and that more empathic participants looked more at the eye region. Then, we modeled participants’ scanpaths with Hidden Markov Models (HMMs), which encapsulate the dynamic and individualistic dimensions of gaze behavior, and do not rely on predefined ROIs. We found that the Kullback–Leibler divergence (KLD) between HMMs representing scanpaths elicited by disgust (disgust-HMM) and neutral-HMM was bigger than for any other emotion, and that this KLD was linked to the empathy score. This suggest that the effect of empathy on gaze patterns depends on the emotion type.