Top down predictions moderate visual appearance of attended objects

Abstract

Early experiments on iconic memory suggested that people initially experience a detailed visual image but can report only a few cued items. We hypothesized that strong top down predictions would allow for visual experience of only task relevant information. In a change detection task we manipulated the appearance of visual stimuli in the critical part of the experiment by removing up to 70% of several items. Our results demonstrate that people are completely unaware of visual detail, even when their attention is focused directly on test items. In a control experiment we showed that the inability to discriminate was not the cause of the unawareness. While numerous studies on inattentional blindness have reported that people will miss objects in the visual field if their attention is directed elsewhere, our study demonstrates that in brief visual presentations top down predictions moderate visual appearance of attended, task relevant items.


Back to Table of Contents