Abstract
We have measured observers' eye movements in a directed visual search task, with a varied complexity of the search matrix. The stimuli were complex 20×20 × matrices in which distraction items were T-characters rotated by steps of 45 , and target items were arrows pointing to four directions (up, down, left, right). Both items were randomly placed in straight lines and rows. The task was to find a path consisting of 15 arrows. Each path started from the middle of the left side of the matrix and ended in the middle of right side of the matrix. The observers had to find the paths by moving their gaze to the directions pointed by the arrows. The matrix complexity was varied in four conditions. Either the background of the matrix or the target-distractor confusability was varied. The background was either grey Gaussian noise or contained randomly rotated small strokes added to the background. In target-distractor confusability conditions, the T-characters were either rotated by random angles or the sizes of the T-characters varied. It was found that reaction times, fixation durations and the number of fixations were significantly higher in background conditions than in distraction item conditions. The results indicate that the background difficulty has stronger effects on processing efficiency than target-distractor confusability. Noisy background apparently distracts saccadic planning more than the confusability of targets and distractors.