Abstract
Theories relating change blindness and visual memory have relied mainly on findings from intentional change detection and memory tasks. In the current experiment, incidental change detection and long-term recognition of object tokens were examined by using two cover tasks. In the “Object” task, subjects were told to make go/no-go responses based on whether a cue captured an object (go) or not (no-go) within an array. In the more attentionally demanding “Inversion” task, participants responded based on objects' orientation (upright: go, inverted: no-go). In each trial, the 8 locations of an array were cued once before a screen flicker, and once after. As the trials progressed, subjects experienced occasional unexpected change-trials in which a flicker and change occurred in the middle of a series of cues. Even though arrays in the Inversion condition contained more objects, change detection and localization for the first unexpected-change trial were more accurate than in the Object condition, in which localization accuracy was at chance. But, in later trials change detection improved only in the Object condition. This result is consistent with theories proposing a primary role for focused attention in on-line change detection, as the attentionally demanding task helped incidental detection, but may have tied-up resources preventing improvement. Long-term recognition for no-change arrays was equivalent and above chance in both conditions, even for arrays prior to change detection trials, and was not correlated with change detection. These results have implications for current theories of change blindness and visual memory. Theories proposing that visual details are not retained in memory can not easily account for these results, nor can theories positing similar retrieval mechanisms for on-line change detection and long-term recognition