Studies of Novel Popout
Annual rept. 15 Aug 1992-14 Aug 1994
UTAH UNIV SALT LAKE CITY DEPT OF PSYCHOLOGY
Pagination or Media Count:
Familiar arrays of objects are perceived better than novel arrays, indicating a perceptual bias toward expected inputs. Yet a novel object in an otherwise familiar array attracts attention, indicating a perceptual bias toward unexpected inputs. These phenomena describe a highly adaptive system but pose a paradox How can the mind be biased simultaneously toward both what it most expects and what it least expects Our research on novel popout illuminates the empirical boundaries of this stabilityplasticity dilemma, and our computational model, called mismatch theory, provides a resolution. In this report we summarize the last two years of research on novel popout and the evolution of mismatch theory. Among other findings, we cite evidence that novel popout represents an automatic and conceptually-driven form of attention capture and that it is not attributable exclusively to simple feature analysis. We argue that the data undermine certain widespread concepts of attention but are in accord with mismatch theory. The general idea behind mismatch theory is that because the processing of expected inputs can be knowledge-based or conceptually-driven data-driven processing can be inhibited for expected inputs and thereby dedicated to any unexpected inputs. Mismatch theory accommodates our findings and resolves the stabilityplasticity dilemma without appealing to the concept of attention as a special gate-keeping device between preattentive and post-attentive processing. Instead, no distinction is drawn between pre-attention and post-attention, and attention is viewed as an emergent phenomenon of ordinary perceptual processes.