Abstract

It has been demonstrated that observers are able to extract the statistical properties of a set of objects, such as the mean size of a set of circles, with surprising efficiency. However, debate surrounds the role of attention in this process, with some theorists claiming that it is ‘automatic’ and done ‘in parallel’, arising from a ‘preattentive’ processing stage. Conversely, other theorists claim that there is a limited attentional capacity for the perception of summary statistics. Further, there is more specific debate surrounding the processes by which summary statistics are computed across different dimensions of visual features simultaneously, such as mean orientation and size. Some findings suggest a cost of encoding two dimensions simultaneously (Emmanouil & Treisman, 2008), whereas others have found evidence of unlimited capacity encoding (Attarha & Moore, 2015b). The Boolean map theory of visual attention would claim that the limiting attentional factor in perceiving multiple visual dimensions simultaneously is stimulus exposure duration. In the current study, participants were shown a set of stimuli which varied in colour, orientation, and size for a fixed exposure duration of 200ms. They were tasked with attending to either one, two, or all three dimensions present in the stimuli. Results indicated that psychophysical performance decreased in terms of both decreased accuracy and increased response times as the number of dimensions attended to increased. The results suggest a bottleneck of selective attention caused by a fixed and difficult exposure duration, consistent with the predictions of Boolean map theory.