The time course of configural change detection for novel 3-D objects

The present study investigated the time course of visual information processing that is responsible for successful

object change detection involving the configuration and shape of 3-D novel object parts. Using a one-shot

change detection task, we manipulated stimulus and interstimulus mask durations (40–500 msec). Experiments

1A and 1B showed no change detection advantage for configuration at very short (40-msec) stimulus

durations, but the configural advantage did emerge with durations between 80 and 160 msec. In Experiment 2,

we showed that, at shorter stimulus durations, the number of parts changing was the best predictor of change

detection performance. Finally, in Experiment 3, with a stimulus duration of 160 msec, configuration change

detection was found to be highly accurate for each of the mask durations tested, suggesting a fast processing

speed for this kind of change information. However, switch and shape change detection reached peak levels of

accuracy only when mask durations were increased to 160 and 320 msec, respectively. We conclude that, with

very short stimulus exposures, successful object change detection depends primarily on quantitative measures

of change. However, with longer stimulus exposures, the qualitative nature of the change becomes progressively

more important, resulting in the well-known configural advantage for change detection.