Ensemble Perception Under Rapid Serial Visual Presentation

A number of recent experiments demonstrate the role of ensemble summary statistics in perception (see Ariely 2001, Alvarez 2011 for review). It was found that summary statistics information does not require attentional resources, and can be processed rapidly and in parallel (Chong, Treisman, 2003; 2005b). Many other studies found that summary statistic information does require attentional resources (Whiting, Oriet, 2011; de Fockert, Marchant, 2008). In the present study, we investigate resources needed to process summary statistic information. We used a Rapid Serial Visual Presentation (RSVP) paradigm with secondary memory task for a target white letter. 29 participants viewed a rapid presentation of letters in the center of the screen. One letter in a stream was (A, T or H) followed by a letter surrounded by white circles. Circle sizes were normally distributed in each trial. There were thee size means: 16, 26 and 36 pixels with standard deviation 6 pixels. After the stream, participants adjusted one central circle to the mean size of circles in the stream, then responded to white letter. We fix the circle time presentation in RSVP and vary T1 time occurrence (100, 200, 300ms, 400ms, 500ms, 700 ms). As a control in the second stage of the experiment participants were instructed to ignore the T1 task; also the response to this task was excluded. In another experiment, we replicate previous results. And in the last experiment replicate classical Raymond (1992) results. We found a significant lag difference at 100-ms and 200-ms, with 700-ms lag. Similar results were found in the control stage. Two possible explanations are discussed. First, summary statistic perception requires attentional resources. Second, a perception of summary statistic processing is influenced by other information, even if participants are told to ignore everything except the ensemble task. This may be attentional capture effect.