Abstract
It takes less than 150 ms for the primate visual system to categorize a complex scene (Thorpe et al., 1996). Although such ultra-rapid categorization (URC) is fast and efficient, there is considerable debate regarding the mechanism that accounts for the process. The URC mechanism has been argued to operate rapidly outside of the focus of attention; yet, and in contrast, it has been demonstrated that scene classification is impaired when attention is allocated to a concurrent task (Walker et. al, 2008). Here, we address this apparent contradiction by examining which mode of attention (focused or global) is most compatible with the URC mechanism. Observers performed a Go/No-Go task in which they were asked to indicate whether or not a briefly presented natural image (32 ms) corresponded to a cued category. On some trials, observers completed a concurrent attentionally demanding task that required either global attention (indicate the orientation of a large box bordering the image) or focused attention (indicate the orientation of a small box located in the center of the image; cf. Chong & Treisman, 2005); in a control condition there was no attention task. Blocks were separated by unique cue / distractor pairings, allowing for the use of signal detection analysis to determine sensitivity and bias parameters between specific categories. Results indicated that categorization performance was worse when observers were asked to complete a concurrent attentionally demanding task than when they were not, corroborating previous data suggesting that attention affects the URC mechanism. Further elucidating these effects, results also indicated that categorization performance was higher when observers completed a concurrent focused attention task than a concurrent global attention task, suggesting that the URC mechanism is most affected by concurrent tasks that require global attention, and that this impairment is reduced when the categorization task requires focused attention.
Meeting abstract presented at VSS 2012