Abstract
Purpose. To assess the influence of target context on performance in a change detection task. Method. On each trial, the observer viewed two briefly flashed, 6-by-6, random checkerboards separated by a 1 sec. ISI and judged (with feedback) whether they were different. The observer was aware that the difference (if there was one) was always in the contrast polarity of a single square. We fit a model in which probability of detection depended on several features of the local context: Feature1 was present if the squares abutting the target on the right and left were equal in contrast; Feature2, if the squares abutting the target above and below were equal in contrast; Feature3, if the squares diagonally adjacent to the target on the upper left and lower right were equal in contrast, and Feature4, if the squares diagonally adjacent to the target on the lower left and upper right were equal in contrast. Results. Features 1 and 2 strongly increased the probability of detection. Features 3 and 4 moderately facilitated detection provided Features 1 and 2 were absent; however, they suppressed detection if either of Features 1 or 2 were present. Conclusion: Detection of a change in a random checkerboard is strongly influenced by the local context in which the change occurs. In particular, changes that alter the topology of the local pattern are better detected than those that do not.