Abstract
Visuospatial working memory enables us to maintain access to visual information for processing even when a stimulus is no longer present, due to occlusion, our own movements, or transience of the stimulus. In the real world, the visual stimuli we need to remember are rarely seen in isolation. Instead, at any given moment, there are usually innumerable other objects visible in a scene that can act as reference points ("visual landmarks") for improving spatial localization (Byrne and Crawford, 2010; Deubel, 2004; Obhi and Goodale, 2005). However, given the limited resources of visuospatial working memory (Bays & Husain, 2008; Zhang & Luck, 2008), what are the consequences of remembering additional object-relative information? To investigate this question, we had participants reproduce the spatial location of a stimulus from a memory array after a short delay. Persistent landmark stimuli were presented on some trials, potentially enabling participants to encode additional relative spatial information. We observed that, when landmarks were present, there was a substantial decrease in localisation variability for stimuli in their vicinity, with the effects decreasing as distance from the landmark increased. Critically, this improvement was not associated with any cost to the precision of egocentric estimates, suggesting that relative spatial information represents an additional, independent spatial cue for memory recall. An optimal integration model, in which relative (allocentric) spatial information was combined with absolute (egocentric) spatial information, accurately captured human performance. In the presence of two landmarks, enhanced localisation precision was evident near both, but the magnitude was reduced compared to the single landmark case. This suggests that spatial information encoded relative to each landmark competed for the same limited resource. In complex scenes, with many potential landmarks present, this highlights the importance of landmark selection for the dynamic allocation of memory resources.
Meeting abstract presented at VSS 2018