Social hierarchy is an essential and pervasive feature of group living in many species, including humans, where hierarchical differentiation has a significant influence on behavior, motivation, and health 1,2. High-ranking animals tend to have more access to food resources and territory, and also have a higher chance of reproductive success 3. In humans, assessing the relative rank of conspecifics is crucial to successfully navigating our complex social environments. One fundamental distinction concerning social hierarchy representations is that they can be assessed according to dominance cues e.g., facial features, physical attributes such as body size, posture and aggressive expressions) to rapidly evaluate the strength of potential competitors, and to avoid costly physical conflict 4. Dominance hierarchies can also be learned by observation or through direct competitive dyadic interactions against rivals 5. Each of these processes has been shown to engage specific brain networks 6,7.
Here, we focus on the ability to assess socially dominant individuals based on facial features, which is an essential skill to avoid costly competitions leading to social defeats 8. Dominance is one of two trait dimensions underpinning social judgments of human faces 9–12. From an evolutionary perspective, identifying the dominance status from features of faces is important for reproductive success. For example, a number of studies have found associations between adult men’s facial width-to-height ratio (fWHR) and perceived likelihood of dominance and aggression 13,14. For this reason, increasing efforts have been devoted to understanding how individuals recognize or infer dominance hierarchies based on facial dominance evaluation in humans 6. Recent progress in social neuroscience research advances our understanding of this matter by exploring the temporal dynamics of brain activities associated with such processing, with the aim of delineating the neurocognitive subprocesses of facial dominance evaluation 15.
Several electrophysiological studies have consistently found modulation of the Late Positive Potentials (LPP) during facial dominance evaluation 16–21. The LPP has been interpreted to reflect evaluations of faces on social dimensions during face processing, suggesting that the inference of dominance hierarchies occurs during the late stage of face processing. However, a discrepancy remains in the literature concerning whether facial dominance judgment also occurs during the early stage of face processing, as indexed by the face-sensitive N170 component. Specifically, a modulation of the N170 by facial dominance has consistently been reported in EEG studies where facial dominance was conveyed through facial expression 22 or simultaneous presentation of dominance-conveying symbols (different number of stars) with emotionally neutral faces 18,19. However, this was not the case in studies where facial dominance was learned either via the direct experience of competitive interactions, 16 or when occupational labels were presented immediately before the faces 17. This discrepancy has been argued to result from methodological differences in manipulating facial dominance 17. Specifically, in the second line of research, facial dominance was manipulated through either learning from the outcome of direct interactions or associating professional dominance with the faces. Thus, the inference of facial dominance would not rely on perceptual identification of explicit dominance-conveying cues but rather the memory of dominance hierarchies established during the learning phase, or the knowledge of professional dominance. Consequently, participants’ facial dominance evaluation would occur during the late phase of face processing, usually reflected by an increased amplitude of the LPP. Previous EEG findings identifying the dynamic processing of neurocognitive subprocesses of facial dominance evaluation have been limited by the fact that they did not use faces with intrinsically different dominance levels as developed by Todorov 10. The use of extrinsic symbols as dominance cues introduces a confounding effect by allowing participants to determine hierarchical dominance independently of face processing. For this reason, the nature of neurocognitive subprocesses associated with facial dominance evaluation still needs to be elucidated.
The present ERPs study was designed to improve our understanding of this issue. Since facial dominance judgments have been revealed to be sensitive to facial features signaling physical strength/weakness 10, we manipulated facial dominance by varying only facial features to control for the confounding effect described above. Faces that varied on the dominance dimension in terms of physical strength/weakness were taken from a validated, computer-generated face database 23. This database contains 25 different faces that can each be morphed to different dominance levels along a dominance scale ranging from − 3 (most submissive) to 0 (neutral) to + 3 (most dominant). We developed a dominance perception task where participants passively viewed these faces. To confirm that participants were sensitive to different levels of facial dominance, passive trials were randomly interleaved with active trials where participants had to identify which one of two faces, of a single identity, was more dominant. Drawing on recent ERP findings 16–19, 21,22, we focused on two ERP components related to facial dominance processing, the N170 and the LPP. These two ERP components can help to explore how facial dominance related to physical strength/weakness exerts a temporally dynamic influence on face processing.