Visual working memory items drift apart due to active, not passive, maintenance.

Abstract

How are humans capable of maintaining detailed representations of visual items in memory? When required to make fine discriminations, we sometimes implicitly differentiate memory representations away from each other to reduce inter-item confusion. However, this separation of representations can inadvertently lead memories to be recalled as biased away from other memory items, a phenomenon termed repulsion bias. Using a non-retinotopically specific working memory paradigm, we found stronger repulsion bias with longer working memory delays, but only when items were actively maintained. These results suggest that (1) repulsion bias can reflect a mnemonic phenomenon, distinct from perceptually driven observations of repulsion bias, and (2) mnemonic repulsion bias is ongoing during maintenance and dependent on attention to internally maintained memory items. These results support theories of working memory where items are represented interdependently and further reveals contexts where stronger attention to working memory items during maintenance increases repulsion bias between them.

An adaptive perspective on visual working memory distortions.

“…Importantly, these biases are not simply inherited from perceptual processing: as noise accumulates in memory over time (reducing the signal-tonoise if memory items), and the need to keep memoranda distinct grows, a corresponding increase in the repulsion bias is observed. Importantly, very recent work (performed since the first presentation of the experiments in the current article) has confirmed various key aspects of our framework: As memories get weaker, biases switch from repulsion to attraction (Lively et al, 2021), and repulsion biases increase with longer memory delays (Scotti et al, 2021).…”

“…Past work has found evidence for attraction biases (Brady et al, 2011; Brady & Alvarez, 2011; 2015; Dubé et al, 2014; Dubé & Sekuler, 2015; Huang & Sekuler, 2010; Lew & Vul, 2015; Lorenc et al, 2018; Utochkin & Brady, 2020), repulsion biases (O’Toole & Wenderoth, 1977; Rauber & Treue, 1998; Scotti et al, 2021; Suzuki & Cavanagh, 1997), or both (Bae & Luck, 2017; Golomb, 2015; Rademaker et al, 2015). Our model and empirical work identifies several key factors that drive these effects and provides evidence that both can arise even in similar paradigms.…”

“…While previous work in this domain has focused on selective attention to continuously present stimuli, a similar type of modulation in the domain of working memory might give rise to repulsive biases in mnemonic representations. Indeed, repulsion biases grow with delay only when a memory is actively held in mind (but disappears when an attention-demanding task is performed during the delay), suggesting that the repulsion bias is not a product of some passive process, but instead requires active maintenance (Scotti et al, 2021). While speculative, this type of adaptive neural modulation may map onto the psychological space framework, such that changes in the discriminability of stimuli in early visual cortex—either due to a lifetime of experience or to dynamic changes in the focus of attention—lead to a warping of perception and memory.…”

Errors in visuospatial working memory across space and time

“…Systematic errors may substantially increase after 5 s, the longest delay interval we used. Moreover, it has been shown that the mean recall errors for orientation 48 , 49 or color 50 increased over memory delays when multiple items were presented and when the items were actively maintained in memory 50 . More studies are needed to examine where, when, and how systematic and unsystematic errors occur.…”

“…In contrast, others showed that unsystematic error accumulated within the first 800 ms of the delay and stayed stable 6 . Similarly, some suggested that most systematic errors accumulated within the first second of the delay 6 , 8 , whereas others showed an increase in the systematic bias with delays up to 3 s 9 , 48 – 50 . One of the primary sources of VSWM errors across time may be noise accumulation over memory maintenance.…”