Judgments of trustworthiness from faces determine basic approach/avoidance responses and approximate the valence evaluation of faces that runs across multiple person judgments. Here, based on trustworthiness judgments and using a computer model for face representation, we built a model for representing face trustworthiness (study 1). Using this model, we generated novel faces with an increased range of trustworthiness and used these faces as stimuli in a functional Magnetic Resonance Imaging study (study 2). Although participants did not engage in explicit evaluation of the faces, the amygdala response changed as a function of face trustworthiness. An area in the right amygdala showed a negative linear response-as the untrustworthiness of faces increased so did the amygdala response. Areas in the left and right putamen, the latter area extended into the anterior insula, showed a similar negative linear response. The response in the left amygdala was quadratic--strongest for faces on both extremes of the trustworthiness dimension. The medial prefrontal cortex and precuneus also showed a quadratic response, but their response was strongest to faces in the middle range of the trustworthiness dimension.

Previous neuroimaging research has shown amygdala sensitivity to the perceived trustworthiness of neutral faces, with greater responses to untrustworthy compared with trustworthy faces. This observation is consistent with the common view that the amygdala encodes fear and is preferentially responsive to negative stimuli. However, some studies have shown greater amygdala activation to positive compared with neutral stimuli. The first goal of this study was to more fully characterize the amygdala response to face trustworthiness by modeling its activation with both linear and nonlinear predictors. Using fMRI, we report a nonmonotonic response profile, such that the amygdala responds strongest to highly trustworthy and highly untrustworthy faces. This finding complicates future attempts to make inferences about mental states based on activation in the amygdala. The second goal of the study was to test for modulatory effects of image spatial frequency filtering on the amygdala response. We predicted greater amygdala sensitivity to face trustworthiness for low spatial frequency images compared with high spatial frequency images. Instead, we found that both frequency ranges provided sufficient information for the amygdala to differentiate faces on trustworthiness. This finding is consistent with behavioral results and suggests that trustworthiness information may reach the amygdala through pathways carrying both coarse and fine resolution visual signals.

While positive behavioral information is diagnostic when evaluating a person's abilities, negative information is diagnostic when evaluating morality. Although social psychology has considered these two domains as orthogonal and distinct from one another, we demonstrate that this asymmetry in diagnosticity can be explained by a single parsimonious principle-the perceived frequency of behaviors in these domains. Less frequent behaviors (e.g., high ability and low morality) are weighed more heavily in evaluations. We show that this statistical principle of frequency-derived diagnosticity is evident in human participants at both behavioral and neural levels of analysis. Specifically, activity in right ventrolateral prefrontal cortex increased preferentially when participants updated impressions based on diagnostic behaviors, and further, activity in this region covaried parametrically with the perceived frequency of behaviors. Activity in left ventrolateral PFC, left inferior frontal gyrus, and left superior temporal sulcus showed similar patterns of diagnosticity and sensitivity, though additional analyses confirmed that these regions responded primarily to updates based on immoral behaviors.

No abstract

We explored the neural correlates of learning about people when the affective value of both facial appearance and behavioral information is manipulated. Participants were presented with faces that were either rated as high or low on trustworthiness. Subsequently, we paired these faces with positive, negative, or no behavioral information. Prior to forming face-behavior associations, a cluster in the right amygdala responded more strongly to untrustworthy than to trustworthy faces. During learning, a cluster in the dorsomedial prefrontal cortex (dmPFC) responded more strongly to faces paired with behaviors than faces not paired with behaviors. We also observed that the activity in the dmPFC was correlated with behavioral learning performance assessed after scanning. Interestingly, individual differences in the initial amygdala response to face trustworthiness prior to learning modulated the relationship between dmPFC activity and learning. This finding suggests that the activity of the amygdala can affect the interaction between dmPFC activity and learning.