Learning to trust a face: The time course of brain activation during a money game.

Evidence shows that human faces can rapidly produce impressions of trust or distrust on the basis of their facial features. However, trust is also built through repeated interactions in which an opposite party acts positively towards the subject in a consistent way. The dynamics of cortical activation of this form of interactively-experienced trust is unclear. The current study therefore investigated the electrophysiological response to trust/distrust, arising through interactions in an investment game. Using an ERP paradigm, participants took part in a money game in which they chose to entrust different amounts to fictitious players. Some of these players were associated with the higher probability of a positive outcome (trustworthy behaviour), others were associated with a higher negative outcome (untrustworthy behaviour), and yet others were neutral. Over the course of the game, a strong central positivity emerged between 450 and 650 ms for trustworthy faces, compared to both neutral and untrustworthy players. This time period thus reflects the window during which the trustworthiness of a face is processed, when based on prior interaction. In addition, by evidencing ERP modifications for trustworthy faces alone, these findings suggest that the "default mode" of processing is initially biased towards distrust.