The binding pool: A model of shared neural resources for distinct items in visual working memory

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.

Applying Generative Artificial Intelligence to cognitive models of decision making

Introduction

Generative Artificial Intelligence has made significant impacts in many fields, including computational cognitive modeling of decision making, although these applications have not yet been theoretically related to each other. This work introduces a categorization of applications of Generative Artificial Intelligence to cognitive models of decision making.

Methods

This categorization is used to compare the existing literature and to provide insight into the design of an ablation study to evaluate our proposed model in three experimental paradigms. These experiments used for model comparison involve modeling human learning and decision making based on both visual information and natural language, in tasks that vary in realism and complexity. This comparison of applications takes as its basis Instance-Based Learning Theory, a theory of experiential decision making from which many models have emerged and been applied to a variety of domains and applications.

Results

The best performing model from the ablation we performed used a generative model to both create memory representations as well as predict participant actions. The results of this comparison demonstrates the importance of generative models in both forming memories and predicting actions in decision-modeling research.

Discussion

In this work, we present a model that integrates generative and cognitive models, using a variety of stimuli, applications, and training methods. These results can provide guidelines for cognitive modelers and decision making researchers interested in integrating Generative AI into their methods.

EZ-CDM: Fast, simple, robust, and accurate estimation of circular diffusion model parameters

The investigation of cognitive processes that form the basis of decision-making in paradigms involving continuous outcomes has gained the interest of modeling researchers who aim to develop a dynamic decision theory that accounts for both speed and accuracy. One of the most important of these continuous models is the circular diffusion model (CDM, Smith. Psychological Review, 123(4), 425. 2016), which posits a noisy accumulation process mathematically described as a stochastic two-dimensional Wiener process inside a disk. Despite the considerable benefits of this model, its mathematical intricacy has limited its utilization among scholars. Here, we propose a straightforward and user-friendly method for estimating the CDM parameters and fitting the model to continuous-scale data using simple formulas that can be readily computed and do not require theoretical knowledge of model fitting or extensive programming. Notwithstanding its simplicity, we demonstrate that the aforementioned method performs with a level of accuracy that is comparable to that of the maximum likelihood estimation method. Furthermore, a robust version of the method is presented, which maintains its simplicity while exhibiting a high degree of resistance to contaminant responses. Additionally, we show that the approach is capable of reliably measuring the key parameters of the CDM, even when these values are subject to across-trial variability. Finally, we demonstrate the practical application of the method on experimental data. Specifically, an illustrative example is presented wherein the method is employed along with estimating the probability of guessing. It is hoped that the straightforward methodology presented here will, on the one hand, help narrow the divide between theoretical constructs and empirical observations on continuous response tasks and, on the other hand, inspire cognitive psychology researchers to shift their laboratory investigations towards continuous response paradigms.

Mapping visual working memory models to a theoretical framework

The body of research on visual working memory (VWM)-the system often described as a limited memory store of visual information in service of ongoing tasks-is growing rapidly. The discovery of numerous related phenomena, and the many subtly different definitions of working memory, signify a challenge to maintain a coherent theoretical framework to discuss concepts, compare models and design studies. A lack of robust theory development has been a noteworthy concern in the psychological sciences, thought to be a precursor to the reproducibility crisis (Oberauer & Lewandowsky, Psychonomic Bulletin & Review, 26, 1596-1618, 2019). I review the theoretical landscape of the VWM field by examining two prominent debates-whether VWM is object-based or feature-based, and whether discrete-slots or variable-precision best describe VWM limits. I share my concerns about the dualistic nature of these debates and the lack of clear model specification that prevents fully determined empirical tests. In hopes of promoting theory development, I provide a working theory map by using the broadly encompassing memory for latent representations model (Hedayati et al., Nature Human Behaviour, 6, 5, 2022) as a scaffold for relevant phenomena and current theories. I illustrate how opposing viewpoints can be brought into accordance, situating leading models of VWM to better identify their differences and improve their comparison. The hope is that the theory map will help VWM researchers get on the same page-clarifying hidden intuitions and aligning varying definitions-and become a useful device for meaningful discussions, development of models, and definitive empirical tests of theories.

Electrophysiological hallmarks for event relations and event roles in working memory

The ability to maintain events (i.e., interactions between/among objects) in working memory is crucial for our everyday cognition, yet the format of this representation is poorly understood. The current ERP study was designed to answer two questions: How is maintaining events (e.g., the tiger hit the lion) neurally different from maintaining item coordinations (e.g., the tiger and the lion)? That is, how is the event relation (present in events but not coordinations) represented? And how is the agent, or initiator of the event encoded differently from the patient, or receiver of the event during maintenance? We used a novel picture-sentence match-across-delay approach in which the working memory representation was "pinged" during the delay, replicated across two ERP experiments with Chinese and English materials. We found that maintenance of events elicited a long-lasting late sustained difference in posterior-occipital electrodes relative to non-events. This effect resembled the negative slow wave reported in previous studies of working memory, suggesting that the maintenance of events in working memory may impose a higher cost compared to coordinations. Although we did not observe significant ERP differences associated with pinging the agent vs. the patient during the delay, we did find that the ping appeared to dampen the ongoing sustained difference, suggesting a shift from sustained activity to activity silent mechanisms. These results suggest a new method by which ERPs can be used to elucidate the format of neural representation for events in working memory.

Limitations on flexible allocation of visual short-term memory resources with multiple levels of goal-directed attentional prioritization

Studies suggest that visual short-term memory (VSTM) is a continuous resource that can be flexibly allocated using probabilistic cues that indicate test likelihood (i.e., goal-directed attentional priority to those items). Previous studies using simultaneous cues have not examined this flexible allocation beyond two distinct levels of priority. Moreover, previous studies have not examined whether there are individual differences in the ability to flexibly allocate VSTM resources, as well as whether this ability benefits from practice. The current study used a continuous report procedure to examine whether participants can use up to three levels of attentional priority to allocate VSTM resources via simultaneous probabilistic spatial cues. Three experiments were performed with differing priority levels, cues, and cue presentation times. Group level analysis demonstrated flexible allocation of VSTM resources; however, there was limited evidence that participants could use three goal-directed priority levels. A temporal analysis suggested that task fatigue, rather than practice effects, may interact with item priority. A Bayesian individual-differences analysis revealed that a minority of participants were using three levels of attentional priority, demonstrating that, while possible, it is not the predominant pattern of behaviour. Thus, we provided evidence that flexible allocation to three attention levels is possible under simultaneous cuing conditions for a minority of participants. Flexible allocation to three categories may be interpreted as a skill of high-performing participants akin to high memory capacity.

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 effect of memory load on object reconstruction: Insights from an online mouse-tracking task

Why can't we remember everything that we experience? Previous work in the domain of object memory has suggested that our ability to resolve interference between relevant and irrelevant object features may limit how much we can remember at any given moment. Here, we developed an online mouse-tracking task to study how memory load influences object reconstruction, testing participants synchronously over virtual conference calls. We first tested up to 18 participants concurrently, replicating memory findings from a condition where participants were tested individually. Next, we examined how memory load influenced mouse trajectories as participants reconstructed target objects. We found interference between the contents of working memory and what was perceived during object reconstruction, an effect that interacted with visual similarity and memory load. Furthermore, we found interference from previously studied but currently irrelevant objects, providing evidence of object-to-location binding errors. At the greatest memory load, participants were nearly three times more likely to move their mouse cursor over previously studied nontarget objects, an effect observed primarily during object reconstruction rather than in the period before the final response. As evidence of the dynamic interplay between working memory and perception, these results show that object reconstruction behavior may be altered by (i) interference between what is represented in mind and what is currently being viewed, and (ii) interference from previously studied but currently irrelevant information. Finally, we discuss how mouse tracking can provide a rich characterization of participant behavior at millisecond temporal resolution, enormously increasing power in cognitive psychology experiments.

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.

Interference between items stored for distinct tasks in visual working memory

The action perspective on working memory suggests that memory representations are coded according to their specific temporal and behavioral task demands. This stands in contrast to theories that assume representations are stored in a task-agnostic format within a "common workspace". Here, we tested whether visual items that are memorized for different tasks are stored separately from one another or show evidence of inter-item interference during concurrent maintenance, indicating a common storage. In two experiments, we combined a framing memory task (memorize a motion direction for continuous direction report) with an embedded memory task (memorize a motion direction for a binary direction discrimination) that was placed within the retention period of the framing task. Even though the temporal and action demands were item specific, we observed two types of interference effects between the items: The embedded motion direction was (1) repulsed away and (2) degraded in precision by the motion direction of the item in the framing task. Repulsion and precision degradation increased with item similarity when both items were concurrently held in working memory. In contrast, perceptual and iconic memory control conditions revealed weaker repulsion overall and no interference effect on precision during the stimulus processing stages prior to working memory consolidation. Thus, additional inter-item interference arose uniquely within working memory. Together, our results present evidence that items that are stored for distinct tasks to be performed at distinct points in time, reside in a common workspace in working memory.

Neural signature of flexible coding in prefrontal cortex

The ability of prefrontal cortex to quickly encode novel associations is crucial for adaptive behavior and central to working memory. Fast Hebbian changes in synaptic strength permit forming new associations, but neuronal signatures of this have been elusive. We devised a trialwise index of pattern similarity to look for rapid changes in population codes. Based on a computational model of working memory, we hypothesized that synaptic strength-and consequently, the tuning of neurons-could change if features of a subsequent stimulus need to be "reassociated," i.e., if bindings between features need to be broken to encode the new item. As a result, identical stimuli might elicit different neural responses. As predicted, neural response similarity dropped following rebinding, but only in prefrontal cortex. The history-dependent changes were expressed on top of traditional, fixed selectivity and were not explainable by carryover of previous firing into the current trial or by neural adaptation.

Storage in Visual Working Memory Recruits a Content-Independent Pointer System

Past work has shown that storage in working memory elicits stimulus-specific neural activity that tracks the stored content. Here, we present evidence for a distinct class of load-sensitive neural activity that indexes items without representing their contents per se. We recorded electroencephalogram (EEG) activity while adult human subjects stored varying numbers of items in visual working memory. Multivariate analysis of the scalp topography of EEG voltage enabled precise tracking of the number of individuated items stored and robustly predicted individual differences in working memory capacity. Critically, this signature of working memory load generalized across variations in both the type and number of visual features stored about each item, suggesting that it tracked the number of individuated memory representations and not the content of those memories. We hypothesize that these findings reflect the operation of a capacity-limited pointer system that supports on-line storage and attentive tracking.

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.

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.

Modelling the simultaneous encoding/serial experience theory of the perceptual moment: a blink of meta-experience

One way to understand a system is to explore how its behaviour degrades when it is overloaded. This approach can be applied to understanding conscious perception by presenting stimuli in rapid succession in the 'same' perceptual event/moment. In previous work, we have identified a striking dissociation during the perceptual moment, between what is encoded into working memory [Lag-1 sparing in the attentional blink (AB)] and what is consciously perceived (Lag-1 impairing in the experiential blink). This paper links this dissociation to what, taking inspiration from the metacognition literature, could be called meta-experience; i.e. how the ability to track and comment on one's visual experience with subjectivity ratings reflects objective performance. Specifically, we provide evidence that the information (in bits) associated with an encoding into working memory decouples from the experiential reflection upon that perceptual/encoding event and that this decoupling is largest when there is the greatest perceptual overload. This is the meta-experiential blink. Meta-experiential self-observation is common to many computational models, including connectionist interpretations of consciousness, Bayesian observers and the readout-enhanced simultaneous type/serial token (reSTST) model. We assess how our meta-experiential blink data could be modelled using the concept of self-observation, providing model fits to behavioural and electroencephalogram responses in the reSTST model. We discuss the implications of our computational modelling of parallel encoding but serial experience for theories of conscious perception. Specifically, we (i) inform theories of Lag-1 sparing during the AB and (ii) consider the implications for the global workspace theory of conscious perception and higher-order theories of consciousness.

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.

A model of working memory for latent representations

We propose a mechanistic explanation of how working memories are built and reconstructed from the latent representations of visual knowledge. The proposed model features a variational autoencoder with an architecture that corresponds broadly to the human visual system and an activation-based binding pool of neurons that links latent space activities to tokenized representations. The simulation results revealed that new pictures of familiar types of items can be encoded and retrieved efficiently from higher levels of the visual hierarchy, whereas truly novel patterns are better stored using only early layers. Moreover, a given stimulus in working memory can have multiple codes, which allows representation of visual detail in addition to categorical information. Finally, we validated our model's assumptions by testing a series of predictions against behavioural results obtained from working memory tasks. The model provides a demonstration of how visual knowledge yields compact visual representation for efficient memory encoding.

How Working Memory and Reinforcement Learning Are Intertwined: A Cognitive, Neural, and Computational Perspective

Reinforcement learning and working memory are two core processes of human cognition and are often considered cognitively, neuroscientifically, and algorithmically distinct. Here, we show that the brain networks that support them actually overlap significantly and that they are less distinct cognitive processes than often assumed. We review literature demonstrating the benefits of considering each process to explain properties of the other and highlight recent work investigating their more complex interactions. We discuss how future research in both computational and cognitive sciences can benefit from one another, suggesting that a key missing piece for artificial agents to learn to behave with more human-like efficiency is taking working memory's role in learning seriously. This review highlights the risks of neglecting the interplay between different processes when studying human behavior (in particular when considering individual differences). We emphasize the importance of investigating these dynamics to build a comprehensive understanding of human cognition.

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.

Capturing the objects of vision with neural networks

Human visual perception carves a scene at its physical joints, decomposing the world into objects, which are selectively attended, tracked and predicted as we engage our surroundings. Object representations emancipate perception from the sensory input, enabling us to keep in mind that which is out of sight and to use perceptual content as a basis for action and symbolic cognition. Human behavioural studies have documented how object representations emerge through grouping, amodal completion, proto-objects and object files. By contrast, deep neural network models of visual object recognition remain largely tethered to sensory input, despite achieving human-level performance at labelling objects. Here, we review related work in both fields and examine how these fields can help each other. The cognitive literature provides a starting point for the development of new experimental tasks that reveal mechanisms of human object perception and serve as benchmarks driving the development of deep neural network models that will put the object into object recognition.

Understanding occipital and parietal contributions to visual working memory: Commentary on Xu (2020)

In her commentary, Xu (2020) admonishes the reader that "To have a full understanding of the cognitive mechanisms underlying VWM [visual working memory], both behavioral and neural evidence needs to be taken into account. This is a must, and not a choice, for any study that attempts to capture the nature of VWM" (p. 11). Although we don't disagree with this statement, our overall assessment of this commentary is that it, itself, fails to satisfy several "musts" and, consequently, does not pose a serious challenge for the sensory recruitment framework for understanding visual working memory. These "musts" include accurately characterizing the framework being critiqued, not favoring verbal models and intuition at the expense of formal quantitative models, and providing even-handed interpretation of the work of others. We'll conclude with a summary of how the sensory recruitment framework can be incorporated into a broader working model of visual working memory.

Decoding Concurrent Representations of Pitch and Location in Auditory Working Memory

Multivariate analyses of hemodynamic signals serve to identify the storage of specific stimulus contents in working memory (WM). Representations of visual stimuli have been demonstrated both in sensory regions and in higher cortical areas. While previous research has typically focused on the WM maintenance of a single content feature, it remains unclear whether two separate features of a single object can be decoded concurrently. Also, much less evidence exists for representations of auditory compared with visual stimulus features. To address these issues, human participants had to memorize both pitch and perceived location of one of two sample sounds. After a delay phase, they were asked to reproduce either pitch or location. At recall, both features showed comparable levels of discriminability. Region of interest (ROI)-based decoding of functional magnetic resonance imaging (fMRI) data during the delay phase revealed feature-selective activity for both pitch and location of a memorized sound in auditory cortex and superior parietal lobule. The latter region showed higher decoding accuracy for location than pitch. In addition, location could be decoded from angular and supramarginal gyrus and both superior and inferior frontal gyrus. The latter region also showed a trend for decoding of pitch. We found no region exclusively coding pitch memory information. In summary, the present study yielded evidence for concurrent representations of pitch and location of a single object both in sensory cortex and in hierarchically higher regions, pointing toward representation formats that enable feature integration within the same anatomic brain regions.SIGNIFICANCE STATEMENT Decoding of hemodynamic signals serves to identify brain regions involved in the storage of stimulus-specific information in working memory (WM). While to-be-remembered information typically consists of several features, most previous investigations have focused on the maintenance of one memorized feature belonging to one visual object. The present study assessed the concurrent storage of two features of the same object in auditory WM. We found that both pitch and location of memorized sounds were decodable both in early sensory areas, in higher-level superior parietal cortex and, to a lesser extent, in inferior frontal cortex. While auditory cortex is known to process different features in parallel, their concurrent representation in parietal regions may support the integration of object features in WM.

The Sync-Fire/deSync model: Modelling the reactivation of dynamic memories from cortical alpha oscillations

We propose a neural network model to explore how humans can learn and accurately retrieve temporal sequences, such as melodies, movies, or other dynamic content. We identify target memories by their neural oscillatory signatures, as shown in recent human episodic memory paradigms. Our model comprises three plausible components for the binding of temporal content, where each component imposes unique limitations on the encoding and representation of that content. A cortical component actively represents sequences through the disruption of an intrinsically generated alpha rhythm, where a desynchronisation marks information-rich operations as the literature predicts. A binding component converts each event into a discrete index, enabling repetitions through a sparse encoding of events. A timing component - consisting of an oscillatory "ticking clock" made up of hierarchical synfire chains - discretely indexes a moment in time. By encoding the absolute timing between discretised events, we show how one can use cortical desynchronisations to dynamically detect unique temporal signatures as they are reactivated in the brain. We validate this model by simulating a series of events where sequences are uniquely identifiable by analysing phasic information, as several recent EEG/MEG studies have shown. As such, we show how one can encode and retrieve complete episodic memories where the quality of such memories is modulated by the following: alpha gate keepers to content representation; binding limitations that induce a blink in temporal perception; and nested oscillations that provide preferential learning phases in order to temporally sequence events.

Location-independent feature binding in visual working memory for sequentially presented objects

Spatial location is believed to have a privileged role in binding features held in visual working memory. Supporting this view, Pertzov and Husain (Attention, Perception, & Psychophysics, 76(7), 1914–1924, 2014) reported that recall of bindings between visual features was selectively impaired when items were presented sequentially at the same location compared to sequentially at different locations. We replicated their experiment, but additionally tested whether the observed impairment could be explained by perceptual interference during encoding. Participants viewed four oriented bars in highly discriminable colors presented sequentially either at the same or different locations, and after a brief delay were cued with one color to reproduce the associated orientation. When we used the same timing as the original study, we reproduced its key finding of impaired binding memory in the same-location condition. Critically, however, this effect was significantly modulated by the duration of the inter-stimulus interval, and disappeared if memoranda were presented with longer delays between them. In a second experiment, we tested whether the effect generalized to other visual features, namely reporting of colors cued by stimulus shape. While we found performance deficits in the same-location condition, these did not selectively affect binding memory. We argue that the observed effects are best explained by encoding interference, and that memory for feature binding is not necessarily impaired when memoranda share the same location.

Incidental encoding of visual information in temporal reference frames in working memory

Visual events are structured in space and time, yet models of visual working memory (VWM) have largely relied on tasks emphasizing spatial aspects. Here, we show that temporal properties of visual events are incidentally encoded along with spatial properties. In five experiments, participants performed change-detection tasks, in which items had unique spatial and temporal coordinates at encoding. Crucially, neither space nor time was task-relevant. The key manipulation concerned the retrieval context: The test array was identical to the memory array either in its entire spatiotemporal structure, or only its spatial or temporal structure. Removing spatial or temporal information at retrieval resulted in costs, indicating that memory relied on both spatial and temporal context in which items were initially perceived. Encoding of spatiotemporal structure occurred incidentally, not strategically, as it was robust even when the retrieval context was perfectly predictable. However, spatial and temporal inter-item spacings influenced the weighting of spatial and temporal information: It favoured the domain in which items were more widely spaced, facilitating their individuation and, likely, access to representations. Across individuals, the weighting of spatial and temporal information varied substantially, but it remained consistent across sessions, suggesting stable preferences for coding in the spatial or temporal domain. No comparable incidental encoding occurred for other task-irrelevant feature dimensions (size or colour). We propose that temporal structure serves as fundamental a function in VWM as spatial structure, scaffolding events that unfold over time.

The Role of Location-Context Binding in Nonspatial Visual Working Memory

Successful retrieval of an item from visual working memory (VWM) often requires an associated representation of the trial-unique context in which that item was presented. In experiment 1, fMRI of 16 male and female humans replicated a previous dissociation of the effects of manipulating memory load in comparison to the effects of manipulating context binding, by comparing VWM for one oriented line versus for three lines individuated by their location versus for three “heterogeneous” items drawn from different categories (orientation, color, and luminance): delay-period fMRI signal in frontal cortex and intraparietal sulcus (IPS) was sensitive to stimulus homogeneity rather than to memory load per se. Additionally, inspection of behavioral performance revealed a broad range of individual differences in the probability of responses to nontargets (also known as “swap errors”), and a post hoc comparison of high swap-error versus low swap-error groups generated several intriguing results: at recall, high swap-error subjects were seen to represent both the orientation and the location of the probed item less strongly, and with less differentiation from nonprobed items, and delay-period signal in IPS predicted behavioral and neural correlates of context binding at recall. In experiment 2, which was a preregistered replication, the 27 male and female humans were grouped into low and high swap-error groups by median split, and the results were broadly consistent with experiment 1. These results present a neural correlate of swap errors, and suggest that delay-period activity of the IPS may be more important for the operation of context binding than for representation per se of stimulus identity.

A delay in sampling information from temporally autocorrelated visual stimuli

Much of our world changes smoothly in time, yet the allocation of attention is typically studied with sudden changes - transients. A sizeable lag in selecting feature information is seen when stimuli change smoothly. Yet this lag is not seen with temporally uncorrelated rapid serial visual presentation (RSVP) stimuli. This suggests that temporal autocorrelation of a feature paradoxically increases the latency at which information is sampled. To test this, participants are asked to report the color of a disk when a cue was presented. There is an increase in selection latency when the disk's color changed smoothly compared to randomly. This increase is due to the smooth color change presented after the cue rather than extrapolated predictions based on the color changes presented before. These results support an attentional drag theory, whereby attentional engagement is prolonged when features change smoothly. A computational model provides insights into the potential underlying neural mechanisms.

Cognitive architecture and capacity of the cognitive system responsible for Same - Different judgments

Participants tend to match identical pairs of stimuli faster than different pairs. Despite many endeavours to explain this fast-same effect, there is still no theoretical consensus. A potential reason for the lack of consensus is that the cognitive architecture and capacity underlying such phenomenon is assumed and not formally tested. For example, the dual-process approach suggests that Same responses arise from a parallel treatment, whereas Different responses arise from a serial treatment. It also suggests that in both conditions, the capacity of the process is unaffected by workload (unlimited capacity). Alternative approaches argue that the fast-same effect can be explained by parallel or coactive architectures with channels working in either limited or super capacity. In this study, we formally assess the architecture (three possibilities: serial, parallel and coactive) and the capacity (three possibilities: unlimited, limited and super-capacity) of the cognitive system in a Same-Different task using Systems Factorial Technology (SFT). We recruited twenty participants to perform a double-factorial task lasting four sessions. Because of the lack of effectiveness of the blurring manipulation, we cannot draw a strong conclusion about the cognitive architecture. As for the capacity, the results show that it is mostly limited for the majority of participants. However, between 300 and 500 ms, participants tend to have a much stronger processing capacity in the Same condition compared to the Different condition. This short but strong burst of activity for identical stimuli might explain the fast-same effect.

Training neural networks to encode symbols enables combinatorial generalization

Combinatorial generalization-the ability to understand and produce novel combinations of already familiar elements-is considered to be a core capacity of the human mind and a major challenge to neural network models. A significant body of research suggests that conventional neural networks cannot solve this problem unless they are endowed with mechanisms specifically engineered for the purpose of representing symbols. In this paper, we introduce a novel way of representing symbolic structures in connectionist terms-the vectors approach to representing symbols (VARS), which allows training standard neural architectures to encode symbolic knowledge explicitly at their output layers. In two simulations, we show that neural networks not only can learn to produce VARS representations, but in doing so they achieve combinatorial generalization in their symbolic and non-symbolic output. This adds to other recent work that has shown improved combinatorial generalization under some training conditions, and raises the question of whether specific mechanisms or training routines are needed to support symbolic processing. This article is part of the theme issue 'Towards mechanistic models of meaning composition'.

Learning how to exploit sources of information

How is our strategy for forming memories shaped by experience with a task? Previous work using surprise questions (i.e., unexpected by the participant) has shown a remarkable inability to report attributes of an attended target in a search display. This representational poverty presumably reflects a form of information exploitation, in which control processes specialize the conversion of available information into memory representations. We hypothesize that such control is refined by repeated experience with a task, and as a result, memory representations will specialize as task experience accrues, such that report accuracy for an unexpected question will progressively worsen as the number of preceding trials increases. To test this, subjects were asked to report the location of a letter among three digits. The ability to respond correctly to a surprise question about the identity of that letter became worse as the experiment progressed. A follow-up study evaluated whether this incremental worsening of report accuracy could be explained as a buildup of proactive interference by varying the set of letters for the surprise test. The results were unchanged relative to the original experiment, which argues against a primary contribution of proactive interference in the worsening performance. The effect was replicated in a similar paradigm using color disks. These findings illustrate that repeated performance of a prescriptive task engages an adaptive modification of control processes that focus information processing on specific attributes of a stimulus that are expected to be necessary in the future, regardless of their immediate task relevance.

Visual Working Memory for Faces and Facial Expressions as a Useful "Tool" for Understanding Social and Affective Cognition

Visual working memory (VWM) is one of the most investigated cognitive systems functioning as a hub between low- and high-level processes. Remarkably, its role in human cognitive architecture makes it a stage of crucial importance for the study of socio-affective cognition, also in relation with psychopathology such as anxiety. Among socio-affective stimuli, faces occupy a place of first importance. How faces and facial expressions are encoded and maintained in VWM is the focus of this review. Within the main theoretical VWM models, we will review research comparing VWM representations of faces and of other classes of stimuli. We will further present previous work investigating if and how both static (i.e., ethnicity, trustworthiness and identity) and changeable (i.e., facial expressions) facial features are represented in VWM. Finally, we will examine research showing qualitative differences in VWM for face representations as a function of psychopathology and personality traits. The findings that we will review are not always coherent with each other, and for this reason we will highlight the main methodological differences as the main source of inconsistency. Finally, we will provide some suggestions for future research in this field in order to foster our understanding of representation of faces in VWM and its potential role in supporting socio-affective cognition.

Hierarchical recruitment of competition alleviates working memory overload in a frontoparietal model

The storage limitations of visual working memory have been the subject of intense research interest for several decades, but few studies have systematically investigated the dependence of these limitations on memory load that exceeds our retention abilities. Under this real-world scenario, performance typically declines beyond a critical load among low-performing subjects, a phenomenon known as working memory overload. We used a frontoparietal cortical model to test the hypothesis that high-performing subjects select a manageable number of items for storage, thereby avoiding overload. The model accounts for behavioral and electrophysiological data from high-performing subjects in a parameter regime where competitive encoding in its prefrontal network selects items for storage, interareal projections sustain their representations after stimulus offset, and weak dynamics in its parietal network limit their mutual interference. Violation of these principles accounts for these data among low-performing subjects, implying that poor visual working memory performance reflects poor control over frontoparietal circuitry, making testable predictions for experiments.

Perceptual consciousness and cognitive access from the perspective of capacity-unlimited working memory

Theories of consciousness divide over whether perceptual consciousness is rich or sparse in specific representational content and whether it requires cognitive access. These two issues are often treated in tandem because of a shared assumption that the representational capacity of cognitive access is fairly limited. Recent research on working memory challenges this shared assumption. This paper argues that abandoning the assumption undermines post-cue-based 'overflow' arguments, according to which perceptual consciousness is rich and does not require cognitive access. Abandoning it also dissociates the rich/sparse debate from the access question. The paper then explores attempts to reformulate overflow theses in ways that do not require the assumption of limited capacity. Finally, it discusses the problem of relating seemingly non-probabilistic perceptual consciousness to the probabilistic representations posited by the models that challenge conceptions of cognitive access as capacity-limited.This article is part of the theme issue 'Perceptual consciousness and cognitive access'.

Neural mechanisms of attending to items in working memory

Working memory, the ability to keep recently accessed information available for immediate manipulation, has been proposed to rely on two mechanisms that appear difficult to reconcile: self-sustained neural firing, or the opposite-activity-silent synaptic traces. Here we review and contrast models of these two mechanisms, and then show that both phenomena can co-exist within a unified system in which neurons hold information in both activity and synapses. Rapid plasticity in flexibly-coding neurons allows features to be bound together into objects, with an important emergent property being the focus of attention. One memory item is held by persistent activity in an attended or "focused" state, and is thus remembered better than other items. Other, previously attended items can remain in memory but in the background, encoded in activity-silent synaptic traces. This dual functional architecture provides a unified common mechanism accounting for a diversity of perplexing attention and memory effects that have been hitherto difficult to explain in a single theoretical framework.

A Flexible Model of Working Memory

Working memory is fundamental to cognition, allowing one to hold information "in mind." A defining characteristic of working memory is its flexibility: we can hold anything in mind. However, typical models of working memory rely on finely tuned, content-specific attractors to persistently maintain neural activity and therefore do not allow for the flexibility observed in behavior. Here, we present a flexible model of working memory that maintains representations through random recurrent connections between two layers of neurons: a structured "sensory" layer and a randomly connected, unstructured layer. As the interactions are untuned with respect to the content being stored, the network maintains any arbitrary input. However, in our model, this flexibility comes at a cost: the random connections overlap, leading to interference between representations and limiting the memory capacity of the network. Additionally, our model captures several other key behavioral and neurophysiological characteristics of working memory.

Chunking as a rational strategy for lossy data compression in visual working memory

The nature of capacity limits for visual working memory has been the subject of an intense debate that has relied on models that assume items are encoded independently. Here we propose that instead, similar features are jointly encoded through a "chunking" process to optimize performance on visual working memory tasks. We show that such chunking can: (a) facilitate performance improvements for abstract capacity-limited systems, (b) be optimized through reinforcement, (c) be implemented by center-surround dynamics, and (d) increase effective storage capacity at the expense of recall precision. Human performance on a variant of a canonical working memory task demonstrated performance advantages, precision detriments, interitem dependencies, and trial-to-trial behavioral adjustments diagnostic of performance optimization through center-surround chunking. Models incorporating center-surround chunking provided a better quantitative description of human performance in our study as well as in a meta-analytic dataset, and apparent differences in working memory capacity across individuals were attributable to individual differences in the implementation of chunking. Our results reveal a normative rationale for center-surround connectivity in working memory circuitry, call for reevaluation of memory performance differences that have previously been attributed to differences in capacity, and support a more nuanced view of visual working memory capacity limitations: strategic tradeoff between storage capacity and memory precision through chunking contribute to flexible capacity limitations that include both discrete and continuous aspects. (PsycINFO Database Record

Current directions in visual working memory research: An introduction and emerging insights

Visual working memory (VWM) is a core construct in the cognitive (neuro-)sciences, assumed to serve as a hub for information exchange and thus supporting a multitude of cognitive functions related to processing visual information. Here, we give an introduction into key terms and paradigms and an overview of ongoing debates in the field, to which the articles collected in this Special Issue on 'Current Directions in Visual Working Memory Research' contribute. Our aim is to extract, from this overview, some 'emerging' theoretical insights concerning questions such as the optimal way to examine VWM, which types of mental representations contribute to performance on VWM tasks, and how VWM keeps features from the same object together and apart from features of concurrently maintained objects (the binding problem).

Fast and flexible sequence induction in spiking neural networks via rapid excitability changes

Cognitive flexibility likely depends on modulation of the dynamics underlying how biological neural networks process information. While dynamics can be reshaped by gradually modifying connectivity, less is known about mechanisms operating on faster timescales. A compelling entrypoint to this problem is the observation that exploratory behaviors can rapidly cause selective hippocampal sequences to 'replay' during rest. Using a spiking network model, we asked whether simplified replay could arise from three biological components: fixed recurrent connectivity; stochastic 'gating' inputs; and rapid gating input scaling via long-term potentiation of intrinsic excitability (LTP-IE). Indeed, these enabled both forward and reverse replay of recent sensorimotor-evoked sequences, despite unchanged recurrent weights. LTP-IE 'tags' specific neurons with increased spiking probability under gating input, and ordering is reconstructed from recurrent connectivity. We further show how LTP-IE can implement temporary stimulus-response mappings. This elucidates a novel combination of mechanisms that might play a role in rapid cognitive flexibility.

A two-level hierarchical framework of visual short-term memory

Over the last couple of decades, a vast amount of research has been dedicated to understanding the nature and the architecture of visual short-term memory (VSTM), the mechanism by which currently relevant visual information is maintained. According to discrete-capacity models, VSTM is constrained by a limited number of discrete representations held simultaneously. In contrast, shared-resource models regard VSTM as limited in resources, which can be distributed flexibly between varying numbers of representations; and a new interference model posits that capacity is limited by interference among items. In this article, we begin by reviewing benchmark findings regarding the debate over VSTM limitations, focusing on whether VSTM storage is all-or-none and on whether object complexity affects capacity. After that, we put forward a hybrid framework of VSTM architecture, arguing that this system is composed of a two-level hierarchy of memory stores, each containing a different set of representations: (1) perceptual memory, a resourcelike level containing analog automatically formed representations of visual stimuli in varying degrees of activation, and (2) visual working memory, in which a subset of three to four items from perceptual memory are bound to conceptual representations and to their locations, thus conveying discrete (digital/symbolic) information which appears quantized. While perceptual memory has a large capacity and is relatively nonselective, visual working memory is restricted in the number of items that can be maintained simultaneously, and its content is regulated by a gating mechanism.

Chunking in working memory via content-free labels

A recent study found that visual working memory performance was enhanced when pairs of colors were predictably paired, and it was interpreted as a form of “memory compression” which implies that more colors could be stored online in a more efficient format. Here we propose an alternative hypothesis that does not entail any increase in the number of individuated representations stored online. Instead, familiar ensembles of items may be attached to a content-free label (e.g., remembering red-white-blue as “American flag”) that can be used to retrieve the constituents of a chunk when they are needed to guide a response. If accessing “compressed” memories requires an additional retrieval process, then access to compressed items should be slower than for uncompressed items. Indeed, Experiments 1 (visual) and 2 (verbal) showed that response times were substantially longer in patterned (i.e., compressed) than in control conditions. In Experiments 3 and 4, regularity-based advantages were eliminated with brief (1000 or 875 ms) response deadlines, in line with our hypothesis that accessing compressed memories requires a slow retrieval process. In sum, while statistical regularities can enable access to larger amounts of information, this information may not be available “online” in the same way as singleton items.

Interactions between visual working memory representations

We investigated whether the representations of different objects are maintained independently in working memory or interact with each other. Observers were shown two sequentially presented orientations and required to reproduce each orientation after a delay. The sequential presentation minimized perceptual interactions so that we could isolate interactions between memory representations per se. We found that similar orientations were repelled from each other whereas dissimilar orientations were attracted to each other. In addition, when one of the items was given greater attentional priority by means of a cue, the representation of the high-priority item was not influenced very much by the orientation of the low-priority item, but the representation of the low-priority item was strongly influenced by the orientation of the high-priority item. This indicates that attention modulates the interactions between working memory representations. In addition, errors in the reported orientations of the two objects were positively correlated under some conditions, suggesting that representations of distinct objects may become grouped together in memory. Together, these results demonstrate that working-memory representations are not independent but instead interact with each other in a manner that depends on attentional priority.

In search of the focus of attention in working memory: 13 years of the retro-cue effect

The concept of attention has a prominent place in cognitive psychology. Attention can be directed not only to perceptual information, but also to information in working memory (WM). Evidence for an internal focus of attention has come from the retro-cue effect: Performance in tests of visual WM is improved when attention is guided to the test-relevant contents of WM ahead of testing them. The retro-cue paradigm has served as a test bed to empirically investigate the functions and limits of the focus of attention in WM. In this article, we review the growing body of (behavioral) studies on the retro-cue effect. We evaluate the degrees of experimental support for six hypotheses about what causes the retro-cue effect: (1) Attention protects representations from decay, (2) attention prioritizes the selected WM contents for comparison with a probe display, (3) attended representations are strengthened in WM, (4) not-attended representations are removed from WM, (5) a retro-cue to the retrieval target provides a head start for its retrieval before decision making, and (6) attention protects the selected representation from perceptual interference. The extant evidence provides support for the last four of these hypotheses.

Competitive interactions affect working memory performance for both simultaneous and sequential stimulus presentation

Competition between simultaneously presented visual stimuli lengthens reaction time and reduces both the BOLD response and neural firing. In contrast, conditions of sequential presentation have been assumed to be free from competition. Here we manipulated the spatial proximity of stimuli (Near versus Far conditions) to examine the effects of simultaneous and sequential competition on different measures of working memory (WM) for colour. With simultaneous presentation, the measure of WM precision was significantly lower for Near items, and participants reported the colour of the wrong item more often. These effects were preserved when the second stimulus immediately followed the first, disappeared when they were separated by 500 ms, and were partly recovered (evident for our measure of mis-binding but not WM precision) when the task was altered to encourage participants to maintain the sequentially presented items together in WM. Our results show, for the first time, that competition affects the measure of WM precision, and challenge the assumption that sequential presentation removes competition.

Working memory representations persist in the face of unexpected task alterations

It is well known that information can be held in memory while performing other tasks concurrently, such as remembering a color or number during a separate visual search task. However, it is not clear what happens to stored information in the face of unexpected tasks, such as the surprise questions that are often used in experiments related to inattentional and change blindness. Does the unpredicted shift in task context cause memory representations to be cleared in anticipation of new information? To answer this question, we ran two experiments where the task unexpectedly switched partway through the experiment with a surprise question. Half of the participants were asked to report the same attribute (Exp. 1 = Identity, Exp. 2 = Color) of a target stimulus in both presurprise and postsurprise trials, while for the other half, the reported attribute switched from identity to color (Exp. 1) or vice versa (Exp. 2). Importantly, all participants had to read an unexpected set of instructions and respond differently on the surprise trial. Accuracy on the surprise trial was higher for the same-attribute groups than the different-attribute groups. Furthermore, there was no difference in reaction time on the surprise trial between the two groups. These results suggest that information participants expected to report can survive an encounter with an unexpected task. The implication is that failures to report information on a surprise trial in many experiments reflect genuine differences in memory encoding, rather than forgetting or overwriting induced by the surprise question.