Decision-making In Dynamic Environments Research Articles

Assessment of Airborne Transmitted Infection Risk in Classrooms Using Computational Fluid Dynamics and Machine Learning-Based Surrogate Modelling

In recent years, the assessment of airborne transmitted infection risk has been extensively performed using Computational Fluid Dynamics (CFD) simulations, especially in response to the COVID-19 pandemic. Nevertheless, the high computational demands and time-intensive nature of CFD simulations highlight the need for fast or real-time infection risk predictions. This capability is crucial for swift decision-making in dynamic environments where timely health interventions are critical. This paper presents a thorough analysis of airborne infection risks in classroom environments based on CFD simulations to understand key factors such as ventilation strategies, air change rates, occupant arrangements, source locations, and particle sizes. This study also employs data-driven supervised learning methods—specifically, Long Short-Term Memory (LSTM) and Artificial Neural Networks (ANN)—to generate surrogate models for predicting airborne infection risk. Key findings reveal that different ventilation strategies significantly affect airborne infection risk, reducing it by 49% to 77%. Moreover, the conventional Wells-Riley model was identified as lacking in its ability to accurately predict local infection risks. The study further challenges the assumption that higher air change rates are universally beneficial, considering that occupants seated in the back rows of a classroom experienced up to a 166% increased risk, despite elevating air change rates from 1.1 h-1 to 11 h-1. These results suggest that physical distancing alone may be insufficient and underscore the importance of considering other factors such as occupant arrangements. Regarding the model performance, the ANN-based surrogate model demonstrated varying prediction accuracy. For inhalable particle concentration predictions for susceptible occupants, R2 values ranged from 0.31 to 0.65 with CVRMSE values between 100% and 180%. In contrast, the model achieved an R2 of 0.79 and a CVRMSE of 34% for infectors. The insights and methodologies from this study can inform HVAC system design and operation strategies to better mitigate infectious disease transmission in densely occupied indoor environments.

Investigating agricultural water sustainability in arid regions with Bayesian network and water footprint theories

Water scarcity is a significant constraint in agricultural ecosystems of arid regions, necessitating sustainable development of agricultural water resources. This study innovatively combines Bayesian theory and Water Footprint (WF) to construct a Bayesian Network (BN). Water quantity and quality data were evaluated comprehensively by WF in agricultural production. This evaluation integrates WF and local water resources to establish a sustainability assessment framework. Selected nodes are incorporated into a BN and continuously updated through structural and parameter learning, resulting in a robust model. Results reveal a nearly threefold increase of WF in the arid regions of Northwest China from 1989 to 2019, averaging 189.95 × 108 m3 annually. The region's agricultural scale is expanding, and economic development is rapid, but the unsustainability of agricultural water use is increasing. Blue WF predominates in this region, with cotton having the highest WF among crops. The BN indicates a 70.1 % probability of unsustainable water use. Sensitivity analysis identifies anthropogenic factors as primary drivers influencing water resource sustainability. Scenario analysis underscores the need to reduce WF production and increase agricultural water supply for sustainable development in arid regions. Proposed strategies include improving irrigation methods, implementing integrated water-fertilizer management, and selecting drought-resistant, economically viable crops to optimize crop planting structures and enhance water use efficiency in current agricultural practices in arid regions. This study not only offers insights into water management in arid regions but also provides practical guidance for similar agricultural contexts. The BN model serves as a flexible tool for informed decision-making in dynamic environments.

Dynamic decisions of the manufacturer-led closed-loop supply chain considering altruistic behavior in EV battery

Behavioral studies have found that altruism can affect the decision of Closed-Loop Supply Chain (CLSC) significantly. Altruistic behaviors on optimal decision-making in dynamic environments are challenging problems. To solve these difficulties, we individually propose three types of cases and using differential game theory and Bellman theory: (a) a manufacturer behaving altruistically to third-party recyclers; (b) one manufacturer cooperating with other participants to have altruism; (c) the manufacturer collaborating with other participants to behave non-altruistically. We conduct optimal dynamic decisions for participants associated with quality of EV batteries, further explore the rule of time evolution considering altruistic behavior. We find three interesting results as follows: altruistic behavior of the manufacturer is advantageous to profitability of CLSC especially when it cooperates with a retailer and one of third-party recyclers. Altruistic behavior of the manufacturer is a behavior of “profit transferring” and moderate altruistic behavior is conducive to the profitability of CLSC and cooperation motivates activity of recycling. In dynamic situations, quality level of EV batteries is polarized. When the manufacturer has altruistic behavior to one or two third-party recyclers, long-term development of CLSC is affected by the deterioration of quality level of EV batteries.

When large language model meets optimization

Optimization algorithms and large language models (LLMs) enhance decision-making in dynamic environments by integrating artificial intelligence with traditional techniques. LLMs, with extensive domain knowledge, facilitate intelligent modeling and strategic decision-making in optimization, while optimization algorithms refine LLM architectures and output quality. This synergy offers novel approaches for advancing general AI, addressing both the computational challenges of complex problems and the application of LLMs in practical scenarios. This review outlines the progress and potential of combining LLMs with optimization algorithms, providing insights for future research directions.

Time and Distance Perception in Highly Dynamic Environments

Abstract Space- and time-based tasks, like flying airplanes, driving cars, or the manual controlled rendezvous and docking of spacecrafts can be very demanding for pilots, drivers, or astronauts, depending on selected trajectories, approaching speeds, or distracting information. One characteristic of highly dynamic environments is that persons have to consider their spatial position in relation to surrounding elements based on passing time, detected speed, and trajectories to avoid collisions with other objects. An inadequate perception of the own position in relation to other objects and passing time can lead to impaired decision-making processes, resulting in errors or collisions. A framework for the explanation of cognitive processes involved in perception, anticipation, and decision-making in dynamic environments is provided by the concept of situation awareness. Based on this framework persons must perceive relevant elements, integrate them in a coherent mental representation and anticipate the development of the current situation in the near future, to plan safe upcoming maneuvers. Thus, for safe maneuvers anticipation processes are playing a major role in decision-making and action selection. In this paper it is proposed that time and distance perceptions are central features of anticipation processes in dynamic situations. It is suggested to combine theories like the attentional-gate model and time-to-contact with the framework of situation awareness to explain anticipation processes in dynamic environments. The major aim is to develop precise cognitive models of persons acting in dynamic situations. Such cognitive models can be the basis for the development and parameterization of assistance systems, supporting anticipation in dynamic environments.

Automatic speech recognition using advanced deep learning approaches: A survey

Recent advancements in deep learning (DL) have posed a significant challenge for automatic speech recognition (ASR). ASR relies on extensive training datasets, including confidential ones, and demands substantial computational and storage resources. Enabling adaptive systems improves ASR performance in dynamic environments. DL techniques assume training and testing data originate from the same domain, which is not always true. Advanced DL techniques like deep transfer learning (DTL), federated learning (FL), and deep reinforcement learning (DRL) address these issues. DTL allows high-performance models using small yet related datasets, FL enables training on confidential data without dataset possession, and DRL optimizes decision-making in dynamic environments, reducing computation costs.This survey offers a comprehensive review of DTL, FL, and DRL-based ASR frameworks, aiming to provide insights into the latest developments and aid researchers and professionals in understanding the current challenges. Additionally, Transformers, which are advanced DL techniques heavily used in proposed ASR frameworks, are considered in this survey for their ability to capture extensive dependencies in the input ASR sequence. The paper starts by presenting the background of DTL, FL, DRL, and Transformers and then adopts a well-designed taxonomy to outline the state-of-the-art (SOTA) approaches. Subsequently, a critical analysis is conducted to identify the strengths and weaknesses of each framework. Additionally, a comparative study is presented to highlight the existing challenges, paving the way for future research opportunities.

Building Human-Like Artificial Agents: A General Cognitive Algorithm for Emulating Human Decision-Making in Dynamic Environments.

One of the early goals of artificial intelligence (AI) was to create algorithms that exhibited behavior indistinguishable from human behavior (i.e., human-like behavior). Today, AI has diverged, often aiming to excel in tasks inspired by human capabilities and outperform humans, rather than replicating human cogntion and action. In this paper, I explore the overarching question of whether computational algorithms have achieved this initial goal of AI. I focus on dynamic decision-making, approaching the question from the perspective of computational cognitive science. I present a general cognitive algorithm that intends to emulate human decision-making in dynamic environments, as defined in instance-based learning theory (IBLT). I use the cognitive steps proposed in IBLT to organize and discuss current evidence that supports some of the human-likeness of the decision-making mechanisms. I also highlight the significant gaps in research that are required to improve current models and to create higher fidelity in computational algorithms to represent human decision processes. I conclude with concrete steps toward advancing the construction of algorithms that exhibit human-like behavior with the ultimate goal of supporting human dynamic decision-making.

IRIS: Integrated Retinal Functionality in Image Sensors.

Neuromorphic image sensors draw inspiration from the biological retina to implement visual computations in electronic hardware. Gain control in phototransduction and temporal differentiation at the first retinal synapse inspired the first generation of neuromorphic sensors, but processing in downstream retinal circuits, much of which has been discovered in the past decade, has not been implemented in image sensor technology. We present a technology-circuit co-design solution that implements two motion computations-object motion sensitivity and looming detection-at the retina's output that could have wide applications for vision-based decision-making in dynamic environments. Our simulations on Globalfoundries 22 nm technology node show that the proposed retina-inspired circuits can be fabricated on image sensing platforms in existing semiconductor foundries by taking advantage of the recent advances in semiconductor chip stacking technology. Integrated Retinal Functionality in Image Sensors (IRIS) technology could drive advances in machine vision applications that demand energy-efficient and low-bandwidth real-time decision-making.

Real-time Data Processing with Machine Learning Algorithms

In the era of information abundance, organizations are faced with the challenge of harnessing real-time data streams to extract valuable insights swiftly. This research paper explores the intersection of real-time data processing and machine learning algorithms, aiming to develop a comprehensive understanding of their integration for efficient decision-making in dynamic environments. The paper begins by delineating the landscape of real-time data processing, emphasizing the significance of timely and accurate information in contemporary business scenarios. It delves into the challenges posed by the velocity and volume of data generated continuously, necessitating advanced processing mechanisms capable of handling data streams in real-time. As the focus shifts to machine learning algorithms, the research outlines the diverse range of algorithms suitable for real-time applications. From online learning methods to streaming algorithms, the exploration encompasses techniques tailored to adapt and evolve with incoming data. This section also addresses the trade-offs between accuracy and computational efficiency, crucial considerations in real-time processing environments. The core of the paper lies in the synthesis of real-time data processing and machine learning algorithms. It investigates how machine learning models can be seamlessly integrated into data processing pipelines to analyze and respond to streaming data instantaneously. Case studies and practical implementations exemplify instances where predictive analytics and anomaly detection algorithms contribute to real-time decision support. Ethical considerations and challenges related to the deployment of machine learning in real-time settings are also examined. The paper advocates for responsible and transparent use of algorithms, emphasizing the importance of mitigating biases and ensuring accountability in decision-making processes driven by machine learning insights. this research paper provides a roadmap for organizations seeking to harness the synergy between real-time data processing and machine learning. The insights gained from this exploration pave the way for advancements in adaptive decision-making systems, offering a competitive edge in industries where rapid response to evolving data is paramount.

New challenges in reinforcement learning: a survey of security and privacy

Reinforcement learning is one of the most important branches of AI. Due to its capacity for self-adaption and decision-making in dynamic environments, reinforcement learning has been widely applied in multiple areas, such as healthcare, data markets, autonomous driving, and robotics. However, some of these applications and systems have been shown to be vulnerable to security or privacy attacks, resulting in unreliable or unstable services. A large number of studies have focused on these security and privacy problems in reinforcement learning. However, few surveys have provided a systematic review and comparison of existing problems and state-of-the-art solutions to keep up with the pace of emerging threats. Accordingly, we herein present such a comprehensive review to explain and summarize the challenges associated with security and privacy in reinforcement learning from a new perspective, namely that of the Markov Decision Process (MDP). In this survey, we first introduce the key concepts related to this area. Next, we cover the security and privacy issues linked to the state, action, environment, and reward function of the MDP process, respectively. We further highlight the special characteristics of security and privacy methodologies related to reinforcement learning. Finally, we discuss the possible future research directions within this area.

Automatic Rule Generation for Decision-Making in Context-Aware Systems Using Machine Learning.

With the increasing interest devoted to dynamic environments, a crucial aspect is revealed in context-aware systems to deal with the rapid changes occurring in users' surrounding environments at runtime. However, most context-aware systems with predefined context-aware rules may not support effective decision-making in dynamic environments. These context-aware rules, which take into account different context information to reach an appropriate decision, could lose their efficiency at runtime. Therefore, a growing need is emerging to address the decision-making issue leveraged by dynamic environments. To tackle this issue, we present an approach that relies on improving decision-making in the wake of dynamic environments through automatically enriching a rule knowledge base with new context-aware rules discovered at runtime. The major features of the presented approach are as follows: (i) a hybridization of two machine learning algorithms for rule generation, (ii) an extended genetic algorithm (GA) for rule optimization, and (iii) a rule transformation for the knowledge base enrichment in an automated manner. Furthermore, extensive experiments on different datasets are performed to assess the effectiveness of the presented approach. The obtained experimental results depict that this approach exhibits better effectiveness compared to some algorithms and state-of-the-art works.

Decision-making Improvement in Dynamic Environments Using Machine Learning

The proliferation of ubiquitous computing with smartphones makes context models and information extremely rich and dynamic on account of highly dynamic environments. However, defining rules at design time may impair their efficiency and the decision-making process at runtime. Therefore, it is important to address decision-making problems leveraged by dynamic environments and context-model evolution. In this sense, a solution that could emerge is the continuous rule knowledge base evolution at runtime. In this paper, we propose a decision adaptation component that relies on generating new rules due to changes occurring around users at runtime. This component aims to support decision-making in dynamic environments and to alleviate the human effort needed to infer new rules. A case study was conducted to illustrate the implementation of the proposed component for the rule knowledge database enrichment and the decision-making improvement. Moreover, an experimental evaluation is provided to assess the effectiveness of the proposed component. The results show that this component exhibits better effectiveness than other well-known algorithms and state-of-the-art approaches. Doi: 10.28991/HEF-2022-03-01-04 Full Text: PDF

Hypersensitivity to negative feedback during dynamic risky-decision making in major depressive disorder: An event-related potential study

BackgroundPatients with major depressive disorder (MDD) exhibit a diminished ability to think or concentrate, indecisiveness, and altered sensitivity to reward and punishment. These impairments can influence complex risk-related decision-making in dynamic environments. The neurophysiological mechanisms mediating MDD effects on decision-making behavior are not well understood. MethodsPatients with MDD (N=50) and healthy controls (HC, N=40) were enrolled. They completed a series of psychometric tests. Event-related potentials (ERPs) were recorded during the performance of a well-validated modified version of Balloon Analogue Risk Task (BART). ResultsBART behavior data were similar across the two groups except the MDD patients showed more stability of risk aversion. Neurophysiologically, BART losses generated larger P3 amplitudes than wins, and MDD patients had larger feedback-related negativity (FRN) components than HCs in response to negative feedback (losses). Greater FRN amplitudes in response to losses correlated with higher levels of depressiveness, psychological pain, and anhedonia. A longer FRN latency in MDD patients was associated with more severe suicidal ideation. LimitationsThe findings are based on cross-sectional data, which are not powerful enough to make causal inferences. ConclusionMDD patients exhibit enhanced FRNs in the frontocentral region after receiving negative feedback in a risky decision-making task. FRN magnitude is associated with depressive symptom severity. Punishment hypersensitivity may contribute to the maintenance of depressive symptoms in MDD patients, and FRN may be a useful index of such hypersensitivity.

Neurogenesis-dependent transformation of hippocampal engrams

In humans and other mammals, memories of events are encoded by neuronal ensembles (or engrams) in the hippocampus. The mnemonic information stored in these engrams can then be used to guide future behavior, including prediction- and decision-making in dynamic environments. While some hippocampal engrams may be persistently stored, others are modified over time, suggesting that the represented memories may also be transformed. How might hippocampal engrams be modified through time? Adult hippocampal neurogenesis represents one process that continuously rewires hippocampal circuitry, presumably including stored hippocampal engrams. At intermediate stages, we propose that neurogenesis-mediated rewiring of hippocampal engram circuitry induces forgetting of specific stimulus attributes, and this less precise engram allows for generalization. At more advanced stages, we propose that neurogenesis-mediated rewiring of hippocampal engram circuitry leads to silencing of hippocampal engrams, rendering them no longer accessible by natural retrieval cues.

A random intuitionistic fuzzy factor analysis model for complex multi-attribute large group decision-making in dynamic environments

The challenge of complex multi-attribute large group decision-making (CMALGDM) is reflected from three perspectives: interrelated attributes, large group decision makers (DMs) and dynamic decision environments. However, there are few decision techniques that can address the three perspectives simultaneously. This paper proposes a random intuitionistic fuzzy factor analysis model, aiming to address the challenge of CMALGDM from the three perspectives. The proposed method effectively reduces the dimensionality of the original data and takes into account the underlying random environmental factors which may affect the performances of alternatives. The development of this method follows three steps. First, the random intuitionistic fuzzy variables are developed to deal with a hybrid uncertain situation where fuzziness and randomness co-exist. Second, a novel factor analysis model for random intuitionistic fuzzy variables is proposed. This model uses specific mappings or functions to define the way in which evaluations are affected by the dynamic environment vector through data learning or prior distributions. Third, multiple correlated attribute variables and DM variables are transformed into fewer independent factors by a two-step procedure using the proposed model. In addition, the objective classifications and weights for attributes and DMs are obtained from the results of orthogonal rotated factor loading. An illustrative case and detailed comparisons of decision results in different environmental conditions are demonstrated to test the feasibility and validity of the proposed method.

Analyzing dynamic decision-making models using Chapman-Kolmogorov equations.

Decision-making in dynamic environments typically requires adaptive evidence accumulation that weights new evidence more heavily than old observations. Recent experimental studies of dynamic decision tasks require subjects to make decisions for which the correct choice switches stochastically throughout a single trial. In such cases, an ideal observer's belief is described by an evolution equation that is doubly stochastic, reflecting stochasticity in the both observations and environmental changes. In these contexts, we show that the probability density of the belief can be represented using differential Chapman-Kolmogorov equations, allowing efficient computation of ensemble statistics. This allows us to reliably compare normative models to near-normative approximations using, as model performance metrics, decision response accuracy and Kullback-Leibler divergence of the belief distributions. Such belief distributions could be obtained empirically from subjects by asking them to report their decision confidence. We also study how response accuracy is affected by additional internal noise, showing optimality requires longer integration timescales as more noise is added. Lastly, we demonstrate that our method can be applied to tasks in which evidence arrives in a discrete, pulsatile fashion, rather than continuously.

Individual Neurons in the Cingulate Cortex Encode Action Monitoring, Not Selection, during Adaptive Decision-Making.

The cingulate cortex contributes to complex, adaptive behaviors, but the exact nature of its contributions remains unresolved. Proposals from previous studies, including evaluating past actions or selecting future ones, have been difficult to distinguish in part because of an incomplete understanding of the task-relevant variables that are encoded by individual cingulate neurons. In this study, we recorded from individual neurons in parts of both the anterior cingulate cortex (ACC) and posterior cingulate cortex (PCC) in 2 male rhesus monkeys performing a saccadic reward task. The task required them to use adaptive, feedback-driven strategies to infer the spatial location of a rewarded saccade target in the presence of different forms of uncertainty. We found that task-relevant, spatially selective feedback signals were encoded by the activity of individual neurons in both brain regions, with stronger selectivity for spatial choice and reward-target signals in PCC and stronger selectivity for feedback in ACC. Moreover, neurons in both regions were sensitive to sequential effects of feedback that partly reflected sequential behavioral patterns. However, neither brain region exhibited systematic modulations by the blockwise conditions that governed the reliability of the trial-by-trial feedback and drove adaptive behavioral patterns. There was also little evidence that single-neuron responses in either brain region directly predicted the extent to which feedback and contextual information were used to inform choices on the subsequent trial. Thus, certain cingulate neurons encode diverse, evaluative signals needed for adaptive, feedback-driven decision-making, but those signals may be integrated elsewhere in the brain to guide actions.SIGNIFICANCE STATEMENT Effective decision-making in dynamic environments requires adapting to changes in feedback and context. The anterior and posterior cingulate cortex have been implicated in adaptive decision-making, but the exact nature of their respective roles remains unresolved. Here we compare patterns of task-driven activity of subsets of individual neurons from parts of the two brain regions in monkeys performing a saccadic task with dynamically changing reward locations. We find evidence for regional specializations in neural representations of choice and feedback, including task-relevant modulations of activity that could be used for performance monitoring. However, we find little evidence that these neural representations are used directly to adjust choice behavior, which thus likely requires integration of these signals elsewhere in the brain.

Neurochemical and Behavioral Dissections of Decision-Making in a Rodent Multistage Task

Flexible decision-making in dynamic environments requires both retrospective appraisal of reinforced actions and prospective reasoning about the consequences of actions. These complementary reinforcement-learning systems can be characterized computationally with model-free and model-based algorithms, but how these processes interact at a neurobehavioral level in normal and pathological states is unknown. Here, we developed a translationally analogous multistage decision-making (MSDM) task to independently quantify model-free and model-based behavioral mechanisms in rats. We provide the first direct evidence that male rats, similar to humans, use both model-free and model-based learning when making value-based choices in the MSDM task and provide novel analytic approaches for independently quantifying these reinforcement-learning strategies. Furthermore, we report that ex vivo dopamine tone in the ventral striatum and orbitofrontal cortex correlate with model-based, but not model-free, strategies, indicating that the biological mechanisms mediating decision-making in the multistage task are conserved in rats and humans. This new multistage task provides a unique behavioral platform for conducting systems-level analyses of decision-making in normal and pathological states.SIGNIFICANCE STATEMENT Decision-making is influenced by both a retrospective "model-free" system and a prospective "model-based" system in humans, but the biobehavioral mechanisms mediating these learning systems in normal and disease states are unknown. Here, we describe a translationally analogous multistage decision-making task to provide a behavioral platform for conducting neuroscience studies of decision-making in rats. We provide the first evidence that choice behavior in rats is influenced by model-free and model-based systems and demonstrate that model-based, but not model-free, learning is associated with corticostriatal dopamine tone. This novel behavioral paradigm has the potential to yield critical insights into the mechanisms mediating decision-making alterations in mental disorders.

Modeling Individual Strategies in Dynamic Decision-making with ACT-R: A Task Toward Decision-making Assistance in HCI

New findings from cognitive science, computer science, and psychology should be used to develop better artificial intelligence (AI). One of the important goals in AI development is the accurate understanding and prediction of the behaviors and decision-making processes of humans. It is especially demanding to achieve this for real dynamic settings, characterized by constant changes. Individual differences in decision-making and behavior make this even more challenging. The area of human-computer interaction looks at a series of decisions and multifactor situations which are influenced by corresponding feedback. Cognitive modeling provides us with a method to understand and explain how such dynamic decisions are made. This work is a demonstration of how cognitive modeling allows to flexible simulate decision-making in dynamic environments for different individual strategies. In this work an empirical study of an improved complex category learning task is presented, the study is based on previous work [8]. The task requires participants to categorize tones (consisting of different features) by applying acoustic strategies to define a target category and adapt to a reversal of feedback. Thus, a model based on the cognitive architecture ACT-R is developed. This model firstly tries out one-feature strategies (e.g. frequency) and then switches to two-feature strategies (e.g. frequency + volume) as a result of negative feedback. However, after comparing the model fit data and analyzing each individual’s data and answers, there is a great variance among some individuals and the first model which only considers acoustic feature strategies and cannot predict individuals who consider the uncertain correct button representing target tones. The upgraded second model contains two independent threshold count mechanisms for these two factors’ learning process. The result of the second model provides a better approximation of the values with the empirical data of those subjects who prefer to consider multi-factors in the tasks. It proves the extensibility of this ACT-R cognitive modeling approach for the different individual cases. A great potential of our approach is, that it can be applied to other HCI tasks and thus it can contribute to related AI approaches and help build AIs with a better understanding of human decision-making.

Fronto-parietal Cortical Circuits Encode Accumulated Evidence with a Diversity of Timescales

Decision-making in dynamic environments often involves accumulation of evidence, in which new information is used to update beliefs and select future actions. Using invivo cellular resolution imaging in voluntarily head-restrained rats, we examined the responses of neurons in frontal and parietal cortices during a pulse-based accumulation of evidence task. Neurons exhibited activity that predicted the animal's upcoming choice, previous choice, and graded responses that reflected the strength of the accumulated evidence. The pulsatile nature of the stimuli enabled characterization of the responses of neurons to a single quantum (pulse) of evidence. Across the population, individual neurons displayed extensive heterogeneity in the dynamics of responses to pulses. The diversity of responses was sufficiently rich to form a temporal basis for accumulated evidence estimated from a latent variable model. These results suggest that heterogeneous, often transient sensory responses distributed across the fronto-parietal cortex may support working memory on behavioral timescales. VIDEO ABSTRACT.