Abstract
Various forces influence what we ultimately remember – such as how attention-grabbing an image is, how much effort we put into remembering that image, or how primed we are for that image. Recent work has identified an intrinsic, high-level perceptual stimulus property – memorability – that is highly consistent across viewers and predictive of whether an image is likely to be remembered or forgotten. Can memorability effects be explained by these other phenomena, such as attention, known to influence memory? We explore this question through a series of psychophysical experiments using classical paradigms well-tested in the literature (i.e., visual search, directed forgetting, perceptual priming) but using face and scene images pre-determined to have high and low memorability. These stimuli are controlled for any possible confounds, including low-level attributes (e.g., color, spatial frequency) and high-level attributes (e.g., gender or emotion for faces; indoor/outdoor or number of objects for scenes). We find that memorability remains resilient despite manipulations of attention and priming: memorability does not cause automatic bottom-up attention capture; one cannot intentionally forget a memorable face or remember a forgettable face; and memorability and perceptual priming are two separate phenomena. In sum, these results provide evidence for memorability as an independent, intrinsic attribute of an image that works in conjunction with other phenomena such as attention and priming to determine if an event will be ultimately remembered.
Meeting abstract presented at VSS 2017