Swap errors in spatial working memory are guesses

The neural basis of swap errors in working memory

When making decisions in a cluttered world, humans and other animals often have to hold multiple items in memory at once-such as the different items on a shopping list. Psychophysical experiments in humans and other animals have shown remembered stimuli can sometimes become confused, with participants reporting chimeric stimuli composed of features from different stimuli. In particular, subjects will often make "swap errors" where they misattribute a feature from one object as belonging to another object. While swap errors have been described behaviorally and theoretical explanations have been proposed, their neural mechanisms are unknown. Here, we elucidate these neural mechanisms by analyzing neural population recordings from monkeys performing two multistimulus working memory tasks. In these tasks, monkeys were cued to report the color of an item that either was previously shown at a corresponding location or will be shown at the corresponding location. Animals made swap errors in both tasks. In the neural data, we find evidence that the neural correlates of swap errors emerged when correctly remembered information is selected from working memory. This led to a representation of the distractor color as if it were the target color, underlying the eventual swap error. We did not find consistent evidence that swap errors arose from misinterpretation of the cue or errors during encoding or storage in working memory. These results provide evidence that swap errors emerge during selection of correctly remembered information from working memory, and highlight this selection as a crucial-yet surprisingly brittle-neural process.

Representation and computation in visual working memory

The ability to sustain internal representations of the sensory environment beyond immediate perception is a fundamental requirement of cognitive processing. In recent years, debates regarding the capacity and fidelity of the working memory (WM) system have advanced our understanding of the nature of these representations. In particular, there is growing recognition that WM representations are not merely imperfect copies of a perceived object or event. New experimental tools have revealed that observers possess richer information about the uncertainty in their memories and take advantage of environmental regularities to use limited memory resources optimally. Meanwhile, computational models of visuospatial WM formulated at different levels of implementation have converged on common principles relating capacity to variability and uncertainty. Here we review recent research on human WM from a computational perspective, including the neural mechanisms that support it.

Alpha phase-coding supports feature binding during working memory maintenance

The ability to successfully retain and manipulate information in working memory (WM) requires that objects' individual features are bound into cohesive representations; yet, the mechanisms supporting feature binding remain unclear. Binding (or swap) errors, where memorized features are erroneously associated with the wrong object, can provide a window into the intrinsic limits in capacity of WM that represent a key bottleneck in our cognitive ability. We tested the hypothesis that binding in WM is accomplished via neural phase synchrony and that swap errors result from perturbations in this synchrony. Using magnetoencephalography data collected from human subjects in a task designed to induce swap errors, we showed that swaps are characterized by reduced phase-locked oscillatory activity during memory retention, as predicted by an attractor model of spiking neural networks. Further, we found that this reduction arises from increased phase-coding variability in the alpha-band over a distributed network of sensorimotor areas. Our findings demonstrate that feature binding in WM is accomplished through phase-coding dynamics that emerge from the competition between different memories.

Effective connectivity of working memory performance: a DCM study of MEG data

Visual working memory (WM) engages several nodes of a large-scale network that includes frontal, parietal, and visual regions; however, little is understood about how these regions interact to support WM behavior. In particular, it is unclear whether network dynamics during WM maintenance primarily represent feedforward or feedback connections. This question has important implications for current debates about the relative roles of frontoparietal and visual regions in WM maintenance. In the current study, we investigated the network activity supporting WM using MEG data acquired while healthy subjects performed a multi-item delayed estimation WM task. We used computational modeling of behavior to discriminate correct responses (high accuracy trials) from two different types of incorrect responses (low accuracy and swap trials), and dynamic causal modeling of MEG data to measure effective connectivity. We observed behaviorally dependent changes in effective connectivity in a brain network comprising frontoparietal and early visual areas. In comparison with high accuracy trials, frontoparietal and frontooccipital networks showed disrupted signals depending on type of behavioral error. Low accuracy trials showed disrupted feedback signals during early portions of WM maintenance and disrupted feedforward signals during later portions of maintenance delay, while swap errors showed disrupted feedback signals during the whole delay period. These results support a distributed model of WM that emphasizes the role of visual regions in WM storage and where changes in large scale network configurations can have important consequences for memory-guided behavior.

Probabilistic and rich individual working memories revealed by a betting game

When asked to remember a color, do people remember a point estimate (e.g., a particular shade of red), a point estimate plus an uncertainty estimate, or are memory representations rich probabilistic distributions over feature space? We asked participants to report the color of a circle held in working memory. Rather than collecting a single report per trial, we had participants place multiple bets to create trialwise uncertainty distributions. Bet dispersion correlated with performance, indicating that internal uncertainty guided bet placement. While the first bet was on average the most precisely placed, the later bets systematically shifted the distribution closer to the target, resulting in asymmetrical distributions about the first bet. This resulted in memory performance improvements when averaging across bets, and overall suggests that memory representations contain more information than can be conveyed by a single response. The later bets contained target information even when the first response would generally be classified as a guess or report of an incorrect item, suggesting that such failures are not all-or-none. This paradigm provides multiple pieces of evidence that memory representations are rich and probabilistic. Crucially, standard discrete response paradigms underestimate the amount of information in memory representations.

The neural basis of swap errors in working memory

When making decisions in a cluttered world, humans and other animals often have to hold multiple items in memory at once - such as the different items on a shopping list. Psychophysical experiments in humans and other animals have shown remembered stimuli can sometimes become confused, with participants reporting chimeric stimuli composed of features from different stimuli. In particular, subjects will often make "swap errors" where they misattribute a feature from one object as belonging to another object. While swap errors have been described behaviorally, their neural mechanisms are unknown. Here, we elucidate these neural mechanisms through trial-by-trial analysis of neural population recordings from posterior and frontal brain regions while monkeys perform two multi-stimulus working memory tasks. In these tasks, monkeys were cued to report the color of an item that either was previously shown at a corresponding location (requiring selection from working memory) or will be shown at the corresponding location (requiring attention to a position). Animals made swap errors in both tasks. In the neural data, we find evidence that the neural correlates of swap errors emerged when correctly remembered information is selected incorrectly from working memory. This led to a representation of the distractor color as if it were the target color, underlying the eventual swap error. We did not find consistent evidence that swap errors arose from misinterpretation of the cue or errors during encoding or storage in working memory. These results suggest an alternative to established views on the neural origins of swap errors, and highlight selection from and manipulation in working memory as crucial - yet surprisingly brittle - neural processes.

Similarity-based clustering of multifeature objects in visual working memory

This study investigated the similarity-based clustering mechanism of multifeature stimuli, wherein items are separated or grouped based on their similarity in visual working memory (VWM). In particular, we investigated whether clustering occurred at an individual feature level or at an integrated object level when participants encoded objects with multiple features for VWM. To test this, we conducted two experiments in which participants remembered and reconstructed a randomly chosen feature (either color or orientation) from one of five presented stimuli. As a key manipulation, we kept the distributions of the two feature dimensions constant while controlling the conjunction between the two dimensions in two different conditions: congruent conjunction (CC) and incongruent conjunction (IC). With this manipulation, we expected to observe the same number of clusters regardless of the conjunction condition when clustering occurred at the feature level. However, we expected a different number of clusters for CC and IC conditions when clustering occurred at the object level. Across two experiments, we consistently observed evidence that favored feature-level clustering. Nevertheless, we found that the swap error rates increased in the IC condition only when two features had to be encoded in VWM. These results suggest that clustering occurs at the feature level in VWM and that feature-level clustering influences item-level feature binding. Therefore, our study demonstrated the flexibility of representational units in VWM.

There Is no Theory-Free Measure of "Swaps" in Visual Working Memory Experiments

Visual working memory is highly limited, and its capacity is tied to many indices of cognitive function. For this reason, there is much interest in understanding its architecture and the sources of its limited capacity. As part of this research effort, researchers often attempt to decompose visual working memory errors into different kinds of errors, with different origins. One of the most common kinds of memory error is referred to as a "swap," where people report a value that closely resembles an item that was not probed (e.g., an incorrect, non-target item). This is typically assumed to reflect confusions, like location binding errors, which result in the wrong item being reported. Capturing swap rates reliably and validly is of great importance because it permits researchers to accurately decompose different sources of memory errors and elucidate the processes that give rise to them. Here, we ask whether different visual working memory models yield robust and consistent estimates of swap rates. This is a major gap in the literature because in both empirical and modeling work, researchers measure swaps without motivating their choice of swap model. Therefore, we use extensive parameter recovery simulations with three mainstream swap models to demonstrate how the choice of measurement model can result in very large differences in estimated swap rates. We find that these choices can have major implications for how swap rates are estimated to change across conditions. In particular, each of the three models we consider can lead to differential quantitative and qualitative interpretations of the data. Our work serves as a cautionary note to researchers as well as a guide for model-based measurement of visual working memory processes.

The spatiotemporal gradient of intrusion errors in continuous outcome source memory: Source retrieval is affected by both guessing and intrusions

Previous research has characterized source retrieval as a thresholded process, which fails on a proportion of trials and leads to guessing, as opposed to a continuous process, in which response precision varies across trials but is never zero. The thresholded view of source retrieval is largely based on the observation of heavy tailed distributions of response errors, thought to reflect a large proportion of "memoryless" trials. In this study, we investigate whether these errors might instead reflect systematic intrusions from other list items which can mimic source guessing. Using the circular diffusion model of decision making, which accounts for both response errors and RTs we found that intrusions account for some, but not all, errors in a continuous-report source memory task. We found that intrusion errors were more likely to come from items studied in nearby locations and times, and were well-described by a spatiotemporal gradient model, but not from semantically or perceptually similar cues. Our findings support a thresholded view of source retrieval but suggest that previous work has overestimated the proportion of guesses which have been conflated with intrusions.

Swap errors in visual working memory are fully explained by cue-feature variability

In cue-based recall from working memory, incorrectly reporting features of an uncued item may be referred to as a "swap" error. One account of these errors ascribes them to variability in memory for the cue features leading to erroneous selection of a non-target item, especially if it is similar to the target in the cue-feature dimension. However, alternative accounts of swap errors include cue-independent misbinding, and strategic guessing when the cued item is not in memory. Here we investigated the cause of swap errors by manipulating the variability with which either cue or report features (orientations in Exp 1; motion directions in Exp 2) were encoded. We found that swap errors increased with increasing variability in memory for the cue features, and their changing frequency could be quantitatively predicted based on recall variability when the same feature was used for report. These results are inconsistent with the hypothesis that swaps are a strategic response to forgotten items, and suggest that swap errors could be wholly accounted for by confusions due to cue-dimension variability. In a third experiment we examined whether spatial configuration of memory arrays in tasks with spatial cueing has an influence on swap error frequency. We observed a specific tendency to make swap errors to non-targets located precisely opposite to the cued location, suggesting that stimulus positions are partially encoded in a non-metric format.

Perception is rich and probabilistic

When we see a stimulus, e.g. a star-shaped object, our intuition is that we should perceive a single, coherent percept (even if it is inaccurate). But the neural processes that support perception are complex and probabilistic. Simple lines cause orientation-selective neurons across a population to fire in a probabilistic-like manner. Does probabilistic neural firing lead to non-probabilistic perception, or are the representations behind perception richer and more complex than intuition would suggest? To test this, we briefly presented a complex shape and had participants report the correct shape from a set of options. Rather than reporting a single value, we used a paradigm designed to encourage to directly report a representation over shape space—participants placed a series of Gaussian bets. We found that participants could report more than point-estimates of shape. The spread of responses was correlated with accuracy, suggesting that participants can convey a notion of relative imprecision. Critically, as participants placed more bets, the mean of responses show increased precision. The later bets were systematically biased towards the target rather than haphazardly placed around bet 1. These findings strongly indicate that participants were aware of more than just a point-estimate; Perceptual representations are rich and likely probabilistic.

The role of ensemble average differs in working memory for depth and planar information

The representation of individual planar locations and features stored in working memory can be affected by the average representation. However, less is known about how the average representation affects the short-term storage of depth information. To evaluate the possible different roles of the ensemble average in working memory for planar and depth information, we used mathematical models to fit the data collected from one study on working memory for depth and 12 studies on working memory for planar information. The pattern of recalled depth was well captured by models assuming that there was a probability of reporting the average depth instead of the individual depth, compressing the recalled front-back distance of the stimulus ensemble compared to the perceived distance. However, when modeling the recalled planar information, we found that participants tended to report individual nontarget features when the target was not memorized, and the assumption of reporting average information improves the data fitting only in very few studies. These results provide evidence for our hypothesis that average depth information can be used as a substitution for individual depth information stored in working memory, but for planar visual features, the substitution of target with the average works under a constraint that the average of to-be-remembered features is readily accessible.

Working Memory Swap Errors Have Identifiable Neural Representations

Working memory is an essential component of cognition that facilitates goal-directed behavior. Famously, it is severely limited and performance suffers when memory load exceeds an individual's capacity. Modeling of visual working memory responses has identified two likely types of errors: guesses and swaps. Swap errors may arise from a misbinding between the features of different items. Alternatively, these errors could arise from memory noise in the feature dimension used for cueing a to-be-tested memory item, resulting in the wrong item being selected. Finally, it is possible that so-called swap errors actually reflect informed guessing, which could occur at the time of a cue, or alternatively, at the time of the response. Here, we combined behavioral response modeling and fMRI pattern analysis to test the hypothesis that swap errors involve the active maintenance of an incorrect memory item. After the encoding of six spatial locations, a retro-cue indicated which location would be tested after memory retention. On accurate trials, we could reconstruct a memory representation of the cued location in both early visual cortex and intraparietal sulcus. On swap error trials identified with mixture modeling, we were able to reconstruct a representation of the swapped location, but not of the cued location, suggesting the maintenance of the incorrect memory item before response. Moreover, participants subjectively responded with some level of confidence, rather than complete guessing, on a majority of swap error trials. Together, these results suggest that swap errors are not mere response-phase guesses, but instead result from failures of selection in working memory, contextual binding errors, or informed guesses, which produce active maintenance of incorrect memory representations.

Effects of item distinctiveness on the retrieval of objects and object-location bindings from visual working memory

Visual working memory (VWM) is prone to interference from stored items competing for its limited capacity. Distinctiveness or similarity of the items is acknowledged to affect this competition, such that poor item distinctiveness causes a failure to discriminate between items sharing common features. In three experiments, we studied how the distinctiveness of studied real-world objects (i.e., whether the objects belong to the same or different basic categories) affects the retrieval of objects themselves (simple recognition) and object-location conjunctions (information about which object was where in a display, cued recall). In Experiments 1 and 2, we found that distinctiveness did not affect memories for objects or for locations, but low-distinctive objects were more frequently reported at "swapped" locations that originally contained other objects, showing object-location memory swaps. In Experiments 3 we found that observers swapped the location of a tested object with another object from the same category more frequently than with any of the objects from another category. This suggests that more similar studied objects cause more retrieval competition in object-location judgments than in simple recognition. Additionally, we discuss a possible role of categorical labeling of locations that can support object-location retrieval when the studied objects are highly distinct.

An interference model for visual working memory: Applications to the change detection task

Most studies of visual-working memory employ one of two experimental paradigms: change-detection or continuous-stimulus reproduction. In this study, we extended the Interference Model (IM; Oberauer & Lin, 2017), which was designed for continuous reproduction, to the single-probe change-detection task. In continuous reproduction, participants occasionally report the non-target items instead of the target. The presence of non-target response is predicted by the Interference Model, which relies in part on the interference of non-target items to explain the set-size effect. By presenting a probe matching a non-target item, we can investigate the amount of interference from non-target items in change detection. As predicted by the Interference Model, we observed poorer performance in rejecting a probe matching a non-target item compared to a new probe (i.e., a cost due to intrusions from non-targets). We fitted the IM along with the Variable Precision, the Slot-Averaging, and the Neural-Population model to the data from two change-detection experiments. The models were equipped with a Bayesian decision rule based on the one used in Keshvari, van den Berg, and Ma (2013). The Interference Model and the Neural-Population model successfully predicted the set-size effect and the non-target intrusion cost, whereas the Variable Precision (VP) and Slot-Averaging (SA) models failed to predict the intrusion cost at all. Even with additional assumptions enabling VP and SA to produce intrusion costs, the IM still performed better than the competing models quantitatively.

The influence of attentional biases on multiple working memory precision parameters for children and adults

Working memory (WM) improves dramatically during childhood but what drives this improvement is not well understood. One influential account thus far has proposed a simple increase in storage capacity. However, recent findings have shown that multiple factors, such as differences in the ability to use attention to enhance the maintenance of internal representations, as well as changes in WM precision, also interact in influencing age-related differences in WM capacity. We aimed to examine whether and how the developing ability to orient attention retrospectively to internal representations influences WM precision. To do so, we employed a paradigm that combined the continuous-recall WM task with the partial-cueing report task. Specifically, 7-year-olds and young adults were asked to reproduce the colour of a probe item in a colour wheel. The initial memory array, which included the probe item, could be followed by a spatial cue that directed participants' attention to a location in the memory array (a 'retro-cue'). Results showed that attentional biases engendered by retro-cues facilitated overall precision compared to uncued baseline performance, for both age groups, although to a smaller degree in 7-year-olds compared to adults. Importantly, investigation of modelling parameters suggested that children demonstrate lower representational quality of items in WM but that spatial attentional cues improve overall precision by increasing the probability of target storage, maintenance and recall, and by reducing misbinding errors as well as random guessing, not by changing representational quality. These results add significantly to our knowledge on the relation between retrospective attention and WM development.

Across-Area Synchronization Supports Feature Integration in a Biophysical Network Model of Working Memory

Working memory function is severely limited. One key limitation that constrains the ability to maintain multiple items in working memory simultaneously is so-called swap errors. These errors occur when an inaccurate response is in fact accurate relative to a non-target stimulus, reflecting the failure to maintain the appropriate association or "binding" between the features that define one object (e.g., color and location). The mechanisms underlying feature binding in working memory remain unknown. Here, we tested the hypothesis that features are bound in memory through synchrony across feature-specific neural assemblies. We built a biophysical neural network model composed of two one-dimensional attractor networks - one for color and one for location - simulating feature storage in different cortical areas. Within each area, gamma oscillations were induced during bump attractor activity through the interplay of fast recurrent excitation and slower feedback inhibition. As a result, different memorized items were held at different phases of the network's oscillation. These two areas were then reciprocally connected via weak cortico-cortical excitation, accomplishing binding between color and location through the synchronization of pairs of bumps across the two areas. Encoding and decoding of color-location associations was accomplished through rate coding, overcoming a long-standing limitation of binding through synchrony. In some simulations, swap errors arose: "color bumps" abruptly changed their phase relationship with "location bumps." This model, which leverages the explanatory power of similar attractor models, specifies a plausible mechanism for feature binding and makes specific predictions about swap errors that are testable at behavioral and neurophysiological levels.

Statistical learning as a reference point for memory distortions: Swap and shift errors

Humans use regularities in the environment to facilitate learning, often without awareness or intent. How might such regularities distort long-term memory? Here, participants studied and reported the colors of objects in a long-term memory paradigm, uninformed that certain colors were sampled more frequently overall. When participants misreported an object's color, these errors were often centered around the average studied color (i.e., "Rich" color), demonstrating swap errors in long-term memory due to imposed statistical regularities. We observed such swap errors regardless of memory load, explicit knowledge, or the distance in color space between the correct color of the tested object and the Rich color. An explicit guessing strategy where participants intentionally made swap errors when uncertain could not fully account for our results. We discuss other potential sources of observed swap errors such as false memory and implicit biased guessing. Although less robust than swap errors, evidence was also observed for subtle shift errors towards or away from the Rich color dependent on the color distance between the correct color and the Rich color. Together, these findings of swap and shift errors provide converging evidence for memory distortion mechanisms induced by a reference point, bridging a gap in the literature between how attention to regularities similarly influences visual working memory and visual long-term memory.

Set size effects on working memory precision are not due to an averaging of slots

Visual working memory is often characterized as a discrete system, where an item is either stored in memory or it is lost completely. As this theory predicts, increasing memory load primarily affects the probability that an item is in memory. However, the precision of items successfully stored in memory also decreases with memory load. The prominent explanation for this effect is the "slots-plus-averaging" model, which proposes that an item can be stored in replicate across multiple memory slots. Here, however, precision declined with set size even in iconic memory tasks that did not require working memory storage, ruling out such storage accounts. Moreover, whereas the slots-plus-averaging model predicts that precision effects should plateau at working memory capacity limits, precision continued to decline well beyond these limits in an iconic memory task, where the number of items available at test was far greater than working memory capacity. Precision also declined in tasks that did not require study items to be encoded simultaneously, ruling out perceptual limitations as the cause of set size effects on memory precision. Taken together, these results imply that set size effects on working memory precision do not stem from working memory storage processes, such as an averaging of slots, and are not due to perceptual limitations. This rejection of the prominent slots-plus-averaging model has implications for how contemporary models of discrete capacities theories can be improved, and how they might be rejected.

Distinguishing target biases and strategic guesses in visual working memory

Previous studies have shown that when there are statistical regularities in the items stored in visual working memory, the responses are biased toward the ensemble average. This statistical-regularity-induced bias could happen in two ways: (1) a target bias, where the individual memory representations are pulled toward the ensemble average; or (2) a strategic guess, for items that are not memorized, other information in the ensemble (e.g., another item) is reported as a substitute. Here, these two mechanisms are distinguished on the basis of a three-part model (target responses + swap responses + random guesses; e.g., Bays, Catalao, & Husain, 2009, Journal of Vision, 9, 7). The strategic guess is operationalized as swap responses, whereas the target bias is reflected by a bias parameter in the target responses. This model was applied on 8 data sets (22 observers each). In this model, contributions of target biases and strategic guesses can be clearly distinguished from each other because they lead to distinctive patterns in the distribution of responses. In the present results, strategic guesses always contributed substantially to the statistical-regularity-induced biases, whereas target biases were limited to specific conditions. All in all, the Bayesian inference in visual working memory is much more limited than what is previously advocated.

The Graded Fate of Unattended Stimulus Representations in Visuospatial Working Memory

As in visual perception, information can be selected for prioritized processing at the expense of unattended representations in visual working memory (VWM). However, what is not clear is whether and how this prioritization degrades the unattended representations. We addressed two hypotheses. First, the representational quality of unattended items could be degraded as a function of the spatial distance to attended information in VWM. Second, the strength with which an item is bound to its location is degraded as a function of the spatial distance to attended information in VWM. To disentangle these possibilities, we designed an experiment in which participants performed a continuous production task in which they memorized a visual array with colored discs, one of which was spatially retro-cued, informing the target location of an impending probe that was to be recalled (Experiment 1). We systematically varied the spatial distance between the cued and probed locations and obtained model-based estimates of the representational quality and binding strengths at varying cue-probe distances. Although the representational quality of the unattended representations remained unaffected by the cue-probe distance, spatially graded binding strengths were observed, as reflected in more spatial confusions at smaller cue-probe distances. These graded binding strengths were further replicated with a model-free approach in a categorical version of the production task in which stimuli and responses consisted of easily discriminable colors (Experiment 2). These results demonstrate that unattended representations are prone to spatial confusions due to spatial degradation of binding strengths in WM, even though they are stored with the same representational quality.