Visual memory of central and marginal items in natural and complex scenes: influence of subjects’ expertise in road traffic.

Our purpose is to study the visual memory (VM) of natural complex scenes in function of subjects’ expertise. 15 subjects were divided into 3 groups according to their level of driving expertise (novices, <5 years and >5 years of experience) and were asked to memorize road images (divided into 3 levels of complexity). After each image presentation (5 second presentation), subjects were asked to answer 6 questions, 3 about central information for driving and 3 about marginal items (not relevant for driving) and to estimate their self-confidence. Our results showed an effect of detail type, subjects’ expertise and stimulus complexity: performance was significantly better for central items than for marginal information, for less complex images than for more complex images and with experts (level 2 and 3) than novices (level 1). We observed no interaction between these 3 variables (detail type, image complexity and subjects’ expertise). This finding suggests difference between central and marginal information in VM performance is stable independently of the image complexity and subjects’ expertise. It generalizes results from our previous study with experts in art history and from Melcher’s study (2006) that showed difference between central and marginal information was stable independently of stimulus presentation duration.