A role for long term memory in search for spatial relations

vision, spatial relations, long term memory, working memory, visual search

Clevenger, Pamela E.

关键词: vision, spatial relations, long term memory, working memory, visual search;   

Others  :  https://www.ideals.illinois.edu/bitstream/handle/2142/99392/CLEVENGER-DISSERTATION-2017.pdf?sequence=1&isAllowed=y

美国|英语

来源: The Illinois Digital Environment for Access to Learning and Scholarship

【 摘 要 】

In the field of visual cognition, most studies focus on how we process features of objects, individual objects, or whole scenes, but relatively little research attention is paid to how we process relationships between objects and features. Our ability to appreciate relations between objects allows us to comprehend graphs, maps, and diagrams. Some argue that our ability to comprehend relations (visual and otherwise) is what makes us human (e.g. Penn, Holyoak, and Povinelli, 2008). Our capacity to maintain and manipulate relations is limited, though (Halford, Wilson, and Phillips, 1998; Halford, Baker, McCredden, and Bain, 2005) and experiments by Gordon Logan (1994) suggest that searching for a specified visual relation is attentionally- demanding. Dynamic binding, the mechanism that affords relational representations their power and flexibility, is subject to the limits of working memory and attention (e.g., Hummel and Biederman, 1992; Hummel, 2001; Hummel, 2003; Knowlton, Morisson, Hummel, and Holyoak, 2012). Logan (1994) found that searching for a target relation (i.e., a plus sign above a minus sign) among distractors did not show search slope improvements over thousands of trials even though searching for a single object among distractors did. His findings seem to suggest that there is no role for long term memory in the processing of relations. However, Woodman et al. (2013) showed that working memory demands for single objects held in working memory could be offloaded to long term memory as long as the target object was repeated for multiple trials. Could long term memory have an opportunity to offload working memory demands for relations if a target relation were repeated? Experiments 1, 2, and 3 presented here replicate and extend Logan’s finding that search slopes do not decrease when targets are defined by a spatial relation, but that they do decrease when the target relation is repeated on every single trial. Experiments 4, 5, and 6 show that participants’ search slopes decrease for relationally-defined targets even when visual features like shape, size, and location vary. Together, these experiments show that long term memory does have a role in processing visual relations that are repeated, and because it persists even when feature dimensions vary, the effect cannot be attributed to template matching.