Capacity limits in the detection and identification of change have implications for models of visual short term memory

Abstract

Observers typically have trouble reporting salient changes between two visual displays if they are presented in alternation, separated by a short blank interval. This phenomenon, commonly known as 'change blindness', suggests that the visual system maintains a relatively sparse representation of the world. Work in a number of research areas - transsaccadic memory, visual-tracking, and visual short-term memory - suggests that visual processing capacity is limited in many tasks to approximately 4-6 items. Here we report findings from an experiment in which observers were asked to detect and identify change in an array of coloured forms. In separate blocks of trials subjects were asked to detect and identify either a colour change (e.g., 'red' to 'blue') or a form change (e.g., 'L' to 'T') to one of the items in the array. As expected, the probability of detecting or identifying change in the form condition was much worse than in the colour condition. However, a lawful relationship was found between detection and identification rates in both conditions. These findings are inconsistent with a model in which limitations in the detection and identification of change are the result of a single underlying process, operating on a limited number of coherent objects held in a high-level working memory store. Instead, we suggest that detection and identification of change are separate processes that share a common low-level informational bottleneck.