Learned face-voice pairings facilitate visual search.

Voices provide a rich source of information that is important for identifying individuals and for social interaction. During search for a face in a crowd, voices often accompany visual information, and they facilitate localization of the sought-after individual. However, it is unclear whether this facilitation occurs primarily because the voice cues the location of the face or because it also increases the salience of the associated face. Here we demonstrate that a voice that provides no location information nonetheless facilitates visual search for an associated face. We trained novel face-voice associations and verified learning using a two-alternative forced choice task in which participants had to correctly match a presented voice to the associated face. Following training, participants searched for a previously learned target face among other faces while hearing one of the following sounds (localized at the center of the display): a congruent learned voice, an incongruent but familiar voice, an unlearned and unfamiliar voice, or a time-reversed voice. Only the congruent learned voice speeded visual search for the associated face. This result suggests that voices facilitate the visual detection of associated faces, potentially by increasing their visual salience, and that the underlying crossmodal associations can be established through brief training.