Investigating the effect of real-world background images on spatial working memory representations

Recent work finds that the location of remembered stimuli can be reconstructed from patterns of EEG alpha-band activity throughout a 1 to 2 second memory delay (Foster et al., 2016; Sutterer et al., 2021). This foundational work involved objects being presented on a blank background. Thus, an open question is whether this is the way we use spatial working memory in the real world, where many background cues are present in addition to the target object. To answer this question, we used a delayed spatial estimation task in which participants remembered the location of a target dot presented around fixation while the background alternated block-by-block between a conventional blank background and an image of a real-world setting. We hypothesized that background cues could affect spatial memory representations in opposing ways. Background cues could enhance spatial memory representations by providing a physical cue that can be attended during the memory delay. Alternatively, background cues could serve as a placeholder that can be used in lieu of maintaining the target location throughout the delay. To test between these alternatives, we applied an Inverted Encoding Model (IEM) to the topography of EEG alpha-band power to measure the spatial selectivity for the target location for each background condition. We observed higher spatial selectivity throughout the trial for blank background compared to scene background trials. This observation is consistent with the idea that the presence of background information reduces demands on spatial working memory.