Abstract
Visual search performance can be facilitated by incidentally learned associations between contextual regularities and target locations within repeated displays, effect known as contextual cueing (CC). Robust CC effect has been found in numerous studies, but the exact mechanisms underlying CC are still not well understood. Here, we investigated the factors that best account for differential CC effects across repeated stimulus configurations. Twenty-three participants searched a T-shape target among L-shape distractors and reported the target’s orientation in a standard CC task. We performed a principal component analysis on 15 variables with a sample of N = 276 pseudo-random configurations to explore the factor structure of the repeated configurations. The 15 variables characterized information about inter-stimulus distance (Mean and SD target-distractor distance, Mean and SD distance between stimuli), target position (X and Y coordinates, N boarder tiles around the target), and target-distractor proximity and similarity (N neighbors, N neighbors differing from the target’s orientation, N neighbors differing from the target’s color, N different colors among neighbors, each variable considered either as a function of direct adjacency to the target or within the target’s quadrant). Based on parallel analysis, we extracted three factors, accounting for 61% of the total explained variance. A factor analysis based on Oblimin rotation to account for the intercorrelation between factors allowed us to retain 13 variables showing high loadings (> 0.40) for one of the three factors. The three factors in their order of importance could be expressed as adjacent neighborhood, inter-stimulus distance, and quadrant neighborhood. The regression model selection for the factors onto CC effects configuration-by-configuration indicated that the best-fitting model with the lowest AIC retained information about the target’s quadrant neighborhood. These results support the importance of local context in the differential effect of CC across spatial configurations.