Psychological Review

Dynamic retrieval of events and associations from memory: An integrated account of item and associative recognition.
Memory theories distinguish between item and associative information, which are engaged by different tasks: item recognition uses item information to decide whether an event occurred in a particular context; associative recognition uses associative information to decide whether two events occurred together. Associative recognition is slower and less accurate than item recognition, suggesting that item and associative information may be represented in different forms and retrieved using different processes. Instead, I show how a dynamic model (Cox & Criss, 2020; Cox & Shiffrin, 2017) accounts for accuracy and response time distributions in both item and associative recognition with the same set of representations and processes. Item and associative information are both represented as vectors of features. Item and associative recognition both depend on comparing traces in memory with probes of memory in which item and associative features gradually accumulate. Associative features are slower to accumulate, but largely because they emerge from conjunctions of already-accumulated item features. I apply the model to data from 453 participants, each of whom performed an item and performed associative recognition following identical study conditions (Cox et al., 2018). Comparisons among restricted versions of the model show that its account of associative feature formation, coupled with limits on the rate at which features accumulate from multiple items, explains how and why the dynamics of associative recognition differ from those of item recognition even while both tasks rely on the same underlying representations. (PsycInfo Database Record (c) 2024 APA, all rights reserved)
Memory out of context: Spacing effects and decontextualization in a computational model of the medial temporal lobe.
Some neural representations gradually change across multiple timescales. Here we argue that modeling this “drift” could help explain the spacing effect (the long-term benefit of distributed learning), whereby differences between stored and current temporal context activity patterns produce greater error-driven learning. We trained a neurobiologically realistic model of the entorhinal cortex and hippocampus to learn paired associates alongside temporal context vectors that drifted between learning episodes and/or before final retention intervals. In line with spacing effects, greater drift led to better model recall after longer retention intervals. Dissecting model mechanisms revealed that greater drift increased error-driven learning, strengthened weights in slower drifting temporal context neurons (temporal abstraction), and improved direct cue–target associations (decontextualization). Intriguingly, these results suggest that decontextualization—generally ascribed only to the neocortex—can occur within the hippocampus itself. Altogether, our findings provide a mechanistic formalization for established learning concepts such as spacing effects and errors during learning. (PsycInfo Database Record (c) 2024 APA, all rights reserved)
Inductive reasoning in minds and machines.
Induction—the ability to generalize from existing knowledge—is the cornerstone of intelligence. Cognitive models of human induction are largely limited to toy problems and cannot make quantitative predictions for the thousands of different induction arguments that have been studied by researchers, or to the countless induction arguments that could be encountered in everyday life. Leading large language models (LLMs) go beyond toy problems but fail to mimic observed patterns of human induction. In this article, we combine rich knowledge representations obtained from LLMs with theories of human inductive reasoning developed by cognitive psychologists. We show that this integrative approach can capture several benchmark empirical findings on human induction and generate human-like responses to natural language arguments with thousands of common categories and properties. These findings shed light on the cognitive mechanisms at play in human induction and show how existing theories in psychology and cognitive science can be integrated with new methods in artificial intelligence, to successfully model high-level human cognition. (PsycInfo Database Record (c) 2024 APA, all rights reserved)
Violations of transitive preference: A comparison of compensatory and noncompensatory accounts.
Violations of transitive preference can be accounted for by both the noncompensatory lexicographic semiorder heuristic and the compensatory additive difference model. However, the two have not been directly compared. Here, we fully develop a simplified additive difference (SAD) model, which includes a graphical analysis of precisely which parameter values are consistent with adherence to, or violation of, transitive preference, as specified by weak stochastic transitivity (WST) and triangle inequalities (TI). The model is compatible with compensatory, within-dimension evaluation. We also develop a stochastic difference threshold model that also predicts intransitive preferences and encompasses a stochastic lexicographic semiorder model. We apply frequentist methods to compare the goodness of fit of both of these models to Tversky’s (1969) data and four replications and Bayes factor methods to determine the strength of evidence for each model. We find that the two methods of analysis converge and that, for two thirds of the participants for whom predictions can be made, one of these models predicting violations of WST has a good and the best fit and has strong Bayesian support relative to an encompassing model. Furthermore, for about 20% of all participants, the SAD model (consistent with violations of WST or TI) is significantly better-fitting and has stronger Bayesian support than the stochastic difference threshold model. Finally, Bayes factor analysis finds strong evidence against transitive models for most participants for whom the SAD model consistent with violation of WST or TI is strongly supported. (PsycInfo Database Record (c) 2024 APA, all rights reserved)
The dual role of culture for reconstructing early sapiens cognition.
Questions on early sapiens cognition, the cognitive abilities of our ancestors, are intriguing but notoriously hard to tackle. Leaving no hard traces in the archeological record, these abilities need to be inferred from indirect evidence, informed by our understanding of present-day cognition. Most of such attempts acknowledge the role that culture, as a faculty, has played for human evolution, but they underrate or even disregard the role of distinct cultural traditions and the ensuing diversity, both in present-day humans and as a dimension of past cognition. We argue that culture has exerted a profound impact on human cognition from the start in a dual manner: It scaffolds cognition through both development and evolution, and it thereby continually diversifies the form and content of human thinking. To unveil early sapiens cognition and retrace its evolutionary trajectories, this cognitive diversity must be considered. We present two strategies to achieve this: large-scale extrapolation and phylogenetic comparison. The former aims at filtering out diversity to determine what is basic and universal versus culturally shaped (illustrated for theory of mind abilities). The latter capitalizes on the diversity to reconstruct evolutionary trajectories (illustrated for religious beliefs). The two methods, in combination, advance our understanding of the cognitive abilities of our early sapiens ancestors and of how these abilities emerged and evolved. To conclude, we discuss the implications of this approach for our insights into early cognition itself and its scientific investigation. (PsycInfo Database Record (c) 2024 APA, all rights reserved)
A flexible threshold theory of change perception in self, others, and the world.
I propose a flexible threshold theory of change perception in self and social judgment. Traditionally, change perception is viewed as a basic cognitive process entailing the act of discriminating informational differences. This article takes a more dynamic view of change perception, highlighting people’s motivations in interpreting those differences. Specifically, I propose people’s change perceptions depend not only on the salience and quality of the evidence for change but they also depend on the adaptation implications of the change, as people are sensitive to whether their prompted response would be worth it. Variables that exacerbate perceived adaptation implications should thus lead people to contract their change perception thresholds (people should become less open to concluding things have changed and so less likely to act), while variables that alleviate perceived adaptation implications should thus lead people to expand their change perception thresholds (people should become more open to concluding things have changed and so more likely to act), all else equal in the evidence. Moreover, these effects should emerge for perceiving declines and improvements alike so long as change bears on adaptation implications. I review support for these proposals and use the theory to generate novel predictions, contributions, and applications. The theory can explain anew why people respond (or fail to respond) to changing climates and economies, worsening personal health, growing social progress, and many other self and social phenomena. Change perception is more than an act of discriminating differences—it also entails people’s threshold judgments of whether and how these differences matter. (PsycInfo Database Record (c) 2024 APA, all rights reserved)
Longtime nemeses or cordial allies? How individuals mentally relate science and religion.
Science and religion are influential social forces, and their interplay has been subject to many public and scholarly debates. The present article addresses how people mentally conceptualize the relationship between science and religion and how these conceptualizations can be systematized. To that end, we provide a comprehensive, integrative review of the pertinent literature. Moreover, we discuss how cognitive (in particular, epistemic beliefs) and motivational factors (in particular, epistemic needs, identity, and moral beliefs), as well as personality and contextual factors (e.g., rearing practices and cross-cultural exposure), are related to these mental conceptualizations. And finally, we provide a flowchart detailing the psychological processes leading to these mental conceptualizations. A comprehensive understanding of how individuals perceive the science–religion relationship is interesting in and of itself and practically relevant for managing societal challenges, such as science denial. (PsycInfo Database Record (c) 2024 APA, all rights reserved)
Advancing the network theory of mental disorders: A computational model of panic disorder.
The network theory of psychopathology posits that mental disorders are systems of mutually reinforcing symptoms. This framework has proven highly generative but does not specify precisely how any specific mental disorder operates as such a system. Cognitive behavioral theories of mental disorders provide considerable insight into how these systems may operate. However, the development of cognitive behavioral theories has itself been stagnant in recent years. In this article, we advance both theoretical frameworks by developing a network theory of panic disorder rooted in cognitive behavioral theory and formalized as a computational model. We use this computational model to evaluate the theory’s ability to explain five fundamental panic disorder-related phenomena. Our results demonstrate that the network theory of panic disorder can explain core panic disorder phenomena. In addition, by formalizing this theory as a computational model and using the model to evaluate the theory’s implications, we reveal gaps in the empirical literature and shortcomings in theories of panic disorder. We use these limitations to develop a novel, theory-driven agenda for panic disorder research. This agenda departs from current research practices and places its focus on (a) addressing areas in need of more rigorous descriptive research, (b) investigating novel phenomena predicted by the computational model, and (c) ongoing collaborative development of formal theories of panic disorder, with explanation as a central criterion for theory evaluation. We conclude with a discussion of the implications of this work for research investigating mental disorders as complex systems. (PsycInfo Database Record (c) 2024 APA, all rights reserved)
The meaning of attention control.
Attention control has been proposed as an ability construct that explains individual differences in fluid intelligence. Evaluating this hypothesis is complicated by a lack of clarity in the definition of attention control. Here, I propose a definition of attention control, based on experimental research and computational models of what guides attention, and how cognitive processes are controlled. Attention is the selection of mental representations for prioritized processing, and the ability to control attention is the ability to prioritize those representations that are relevant for the person’s current goal, thereby enabling them to think and act in accordance with their intentions. This definition can be used to identify appropriate and less appropriate ways to measure individual differences in attention control. An analysis of various approaches to measurement reveals that the current practice of measuring attention control leaves room for improvement. Aligning our psychometric measurements with a clear, theoretically grounded concept of attention control can lead to more valid measures of that construct. (PsycInfo Database Record (c) 2024 APA, all rights reserved)

Psychological Review - Vol 131, Iss 6