Abstract
Research on visual search shows that presenting a characteristic sound (e.g. barking) can reduce search time for a target object (e.g. a dog) that is among other isolated objects, even when the sound provides no spatial information about the location of the search target (Iordanescu, Guzman-Martinez, Grabowecky & Suzuki, 2008). Here we investigated whether characteristic sounds can facilitate visual search in the context of realistic scenes. Across trials, we manipulated sound congruency: congruent (e.g. hearing barking while searching for a dog in a living room), incongruent (e.g. hearing meowing while searching for a dog), white noise, or no sound. We also manipulated stimulus onset asynchrony (SOA) between the sound and the scene (-1000, -500, 0, +300 ms) and target presence (present or absent). On each trial, participants (N = 55) were presented with a word cue indicating the target object, and then either (a) heard a sound followed by a scene (negative SOA), (b) heard a sound simultaneous with the onset of the scene (SOA 0 ms) or (c) viewed a scene followed by a sound (positive SOA). Results indicated that reaction times (RTs) on target-present trials depended on both sound congruency and SOA. At negative SOAs, RTs were significantly shorter in congruent trials compared to incongruent or white noise trials. At SOA of 0 ms, RTs were significantly shorter in congruent compared to incongruent trials, but not compared to white noise trials. At SOA of +300 ms, visual search performance was impaired, with RTs higher across the three sound conditions compared to all other conditions in the experiment, but there was no congruency effect. Overall, our results extend previous findings and suggest that congruent characteristic sounds can facilitate finding target objects in realistic scenes, provided that the sounds are presented before the scene.