Fanny Poncet¹ (), Marie Smith², Roberto Caldara¹, Anne-Raphaëlle Richoz¹; ¹Eye and Brain Mapping Laboratory (iBMLab), Department of Psychology, University of Fribourg, Fribourg, Switzerland, ²School of Psychological Sciences, Birkbeck, University of London, London, United Kingdom

Distinguishing between fear and surprise expressions is difficult, as these emotions share several overlapping facial features. Prior research shows a dynamic advantage for recognizing surprise, but not fear, and age-related differences in eye-movement strategies. However, it remains unclear how the diagnostic information used to recognize these expressions varies across ages and stimulus modalities. Here, we used the “Bubbles” reverse-correlation technique to isolate the diagnostic facial information underlying static and dynamic facial expression recognition (FER) across ages. Participants were 120 young children (6–9-years), 110 older children (10–13-years) tested in either static or dynamic condition, and 45 young adults (18–30-years) tested in both modalities. The number of bubbles was set a priori and varied by age group to balance task difficulty. All groups performed better with dynamic presentations overall. At the expression level, recognition of both fear and surprise improved with dynamic cues in older groups, and confusion between the two expressions was significantly reduced in the dynamic condition, most strongly for surprise. Diagnostic information also varied as a function of age and modality. Six- to nine-year-olds primarily relied on the mouth but started to include the eyebrow information in the dynamic modality. Ten-to thirteen-year-olds and adults showed similar patterns: static recognition involved mouth, eyes and nose (children only), while dynamic recognition incorporated the eyebrows (with specific eyebrow regions depending on the expression) alongside the mouth. Interestingly, specific diagnostic patterns were observed when examining each expression separately, interacting with both modality and age. In conclusion, dynamic facial expressions modulate facial information use and improve FER performance. Our data reveal the facial features involved in recognizing fear and surprise, highlighting the strong influence of age and presentation modality. These findings provide novel insights into the developmental trajectory underlying the well-established confusion between these expressions.

Acknowledgements: This study, was supported by the grant no 10001C_201145 from the Swiss National Science Foundation awarded to RC and ARR.