Schematic representation of procedures used in the study.

<p>Participants had to memorize three groups of faces. A screen with a behavioral descriptor appeared before the presentation of each group of faces (prosocial, antisocial and neutral). The order of descriptors and correspondent group of faces was randomized and counterbalanced. Memorization was followed by a distracter task (consolidation) consisting of a series of multiplications. In the “yes/no” task each group had 4 faces and each face appeared in frontal and profile view (8 images/group; frontal view always to the left of its profile view). Participants had to answer whether they had seen the displayed face before (50% tagged faces and 50% new faces). In the 4-alternative forced-choice task each group also had 4 faces, but only in frontal view. In this task participants had to choose which of the faces they had memorized (25% tagged faces, 75% new). The display for the recall task was similar to the yes/no task, but participants had to answer whether the face they saw belonged to “cheaters”, “neutrals”, or “cooperators” instead (only tagged faces were presented). Tagged and new faces were interleaved and presented randomly.