You can find the original article here (open access).
In this study, we made participants like and dislike other people by using a social interactive game, and then we checked how their brains responded to pictures of the co-players. In fact, there were no other players, just algorithms designed to play “fair” or not. However, we made sure the subjects were convinced otherwise: we took pics of them, “called” other labs, etc.
All participants believed they played with other people. One even tried to a co-player’s number to ask her out! Of course, after the experiment they were all informed about the deception. Then, in a Posner task, the gaze of co-players gave cues for the location of future targets. Turns out, already after 90 ms the negative player showed enhanced perceptual processing!
Why is this study interesting?
(1) Because of increased ecological validity: we used a naturalistic social learning in an interactive exchange game to induce affective knowledge about others
(2) Because we showed *very early* (from 90 ms) modulation of brain responses by affective knowledge: emotionally charged interactions increase the saliency of the interaction partners’ faces
So, simplifying, when watching someone’s face, our brains know we dislike them before we even know what we’re seeing! It’s fascinating how higher processing (semantic knowledge) penetrates perception.
MagdaLena Matyjek, Ph.D. 
One thought on “Socially induced negative affective knowledge modulates early face perception but not gaze cueing of attention”