Abstract
Previous research in reading has shown that information about a word is obtained from the parafovea before it is directly fixated, which speeds the processing of that word once it is subsequently fixated (Rayner, 1975). Further research has shown that contextual constraints (e.g. the predictability of a word) lead to an increase in information acquired from the parafovea (Balota, Rayner & Pollatsek, 1985; McClelland & O’Regan, 1981). Studies have also shown similar effects of contextual constraints on the processing of objects (Biederman, Mezzanotte & Rabinowitz, 1982; Henderson & Hollingworth, 1998). In the present study, we examine whether scene context constrains parafoveal processing of objects prior to fixation. Participants were shown a preview of the target object at either 4° (parafovea) or 10° (periphery) from the point of fixation in either a consistent or inconsistent scene context. The preview could be: (i) identical to the target; (ii) a different category with the same shape; (iii) a different category and different shape; or (iv) a black rectangle control. During the saccade towards the preview object, it changed to the target. In Experiment 1, participants had better performance for identical and similarly-shaped previews, with a stronger benefit for consistent contexts. In Experiment 2, participants had to verify whether the target matched a given object name seen before the trial commenced. Results showed a higher sensitivity (A-prime) for consistent vs. inconsistent contexts. In Experiment 3, participants verified the object name at the end of the trial. The added uncertainty of the target identity led to no difference in sensitivity between context conditions, but did demonstrate a difference in bias (B''D) toward ‘yes’ responses between consistent and inconsistent scene contexts. Across the three experiments, we show that scene context does have a constraining effect on preview processing of objects prior to fixation.
Meeting abstract presented at VSS 2013