Abstract
Ensemble perception refers to the visual system’s ability to rapidly evaluate “gist” information from groups of stimuli. For example, participants can easily determine the average emotional tenor of a crowd (e.g., an angry mob of protestors, or mourners at funeral), even when they fail to recall individual emotions presented in the crowd (Haberman & Whitney, 2007, 2009). To extend prior findings, we tested ensemble perception during a change localization task. We used Mooney faces as our experimental stimuli. Mooney faces are two-tone, low-feature images which require observers to perceive holistically. First, participants viewed a crowd of Mooney faces. Next, participants viewed the same crowd again, with some of the faces either replaced by new faces or simply shuffled in locations. In the response phase, participants were asked to localize any one of the replacement/shuffled faces by clicking on the corresponding screen location. Participants were also asked to report whether the average emotions of the two displays were the same or different. Although the task was challenging, participants performed slightly, but significantly, above chance in both the ensemble and change localization tasks. Most surprisingly, participants were sensitive to the emotional tenor of the crowd even when their change localization failed. To ensure that participants were not just relying on information from a single display to probabilistically guess the answer, we conducted a control experiment where one of the two displays was occluded. As expected, performance in this control condition was at chance, demonstrating that participants relied on ensemble information from both displays to successfully discriminate the average expression in the crowd. These results reinforce and extend prior findings (Haberman & Whitney, 2011) by showing that holistic ensemble perception can remain intact even when individual objects that drive the ensemble are outside the focus of attention.