Real-world objects are not stored in holistic representations in visual working memory

Abstract

When storing multiple objects in visual working memory, observers sometimes misattribute perceived features to incorrect locations or objects. These misattributions are called binding errors (or swaps) and have been previously demonstrated mostly in simple objects whose features are easy to encode independently and arbitrarily chosen, like colors and orientations. Here, we tested whether similar swaps can occur with real-world objects, where the connection between features is meaningful rather than arbitrary. In Experiments 1 and 2, observers were simultaneously shown four items from two object categories. Within a category, the two exemplars could be presented in either the same or different states (e.g., open/closed; full/empty). After a delay, both exemplars from one of the categories were probed, and participants had to recognize which exemplar went with which state. We found good memory for state information and exemplar information on their own, but a significant memory decrement for exemplar-state combinations, suggesting that binding was difficult for observers and “swap” errors occurred even for meaningful real-world objects. In Experiment 3, we used the same tasks, but on half of the trials, the locations of the exemplars were swapped at test. We found that participants ascribed incorrect states to exemplars more frequently when the locations of exemplars were swapped. We concluded that the internal features of real-world objects are not perfectly bound in working memory, and location updates impair the object representation. Overall, we provide evidence that even real-world objects are not stored in an entirely unitized format in working memory.

Demands on perceptual and mnemonic fidelity are a key determinant of age-related cognitive decline throughout the lifespan

“…It is therefore unclear whether the trial-by-trial association between fidelity of object and spatial memory follows the same pattens as that observed for course-grained mnemonic representations. Given prior work suggesting that stimulus features in working memory may be forgotten independently from one another (Bays et al, 2011;Fougnie & Alvarez, 2011;Markov et al, 2021), whereas pattern completion and reinstatement tend to result in holistic retrieval from long-term memory (Grande et al, 2019;Horner et al, 2015), we predicted a stronger association between fidelity of object and spatial information in LTM as compared to short-term memory (STM). Moreover, based on well-documented age-related declines in binding together elements in memory (Chen & Naveh-Benjamin, 2012;Henkel et al, 1998;Hou et al, 2019;James et al, 2019;Lyle et al, 2006;Naveh-Benjamin & Mayr, 2018;Ngo & Newcombe, 2021), we also hypothesised a weaker trial-by-trial association between object and spatial memory fidelity with age.…”

The best game in town: The reemergence of the language-of-thought hypothesis across the cognitive sciences

“…However, more striking evidence comes from the way in which featural information is "swapped" between objects. Participants often misremember a feature of one object as bound to another object (Bays, Catalao, & Husain 2009), including for real-world stimuli (Utochkin & Brady 2020;Markov et al 2021). Even during multiple-object tracking, a stored feature of one object (e.g., a previewed numeral) may be swapped with another object if they come too close to each other during tracking (Pylyshyn 2004).…”

Representation and computation in working memory

“…It has been proposed that binding in WM, as in visual perception, is achieved through feature maps over visual space, with different non-spatial features of an object bound to each other only indirectly via their shared location (Schneegans and Bays, 2017a). This account allows for independent resource pools for different non-spatial features while still employing inherently conjunctive memory representations, and it explains patterns of error correlations in dual-report paradigms (Bays et al, 2011;Fougnie and Alvarez, 2011;Kovacs and Harris, 2019;Markov et al, 2021; but see Sone et al, 2021 for an alternative account). More recent work further indicates that for sequentially presented stimuli, presentation time may take a similar role as location in binding visual features (Schneegans et al, 2021(Schneegans et al, , 2022Heuer and Rolfs, 2021).…”