Department of Cognitive Sciences

We build upon Haueis and Colaço's account in two ways. First, we argue that the backgrounding of metabolism was a pragmatic response to a data granularity problem faced by traditional cognitive modelers. Then, we suggest that equivalence demonstrations offer a promising pathway for incrementally introducing metabolic considerations into extant modeling practice.

Open science advocates aim to address epistemic injustice and global power asymmetries in research infrastructures yet underestimate structural vulnerabilities and risks embedded in data sharing practices. Drawing on an ethnographic study of the GCRF South–South Migration Inequality and Development Hub’s (MIDEQ), we argue that even well-intentioned systems, such as the MIDEQ hub, do not reach these vulnerabilities. To capture these overlooked harms, we employ the concept of epistemic disadvantage, a category coextensive with, but separate from epistemic injustice. Epistemic disadvantage includes ethically justified data withholding and misappropriation within consensual systems, which is not captured by existing notions of extractivism as those focus on the monetisation of epistemic resources without regard for the loss incurred by the originating community. To better account for these harms, we draw on two recent epistemic attitudes: the Book of Truths (BoT), which sees scientific truths as emerging through peer review, and the Book of Conversations (BoC), which values data for the interactions and reflections it provokes. Even when epistemic values recognise knowledge as relational and situated, extractivist harms persist. We offer an ameliorated concept of extractivism to include subtler, systemic harms rooted in epistemic disadvantage.

Shiffrin et al. argue that scientific practice often produces illusions of understanding; situations in which familiar inferential tools generate misplaced confidence. We describe a related illusion that arises not from statistical misapplications, but from ordinary theory testing itself. Here we highlight a different source of epistemic uncertainty or overconfidence; one that emerges even when statistical tools are used correctly, effect sizes are precisely estimated, findings replicate without controversy, and heterogeneity is explicitly acknowledged within hierarchical models.

Most widely studied option evaluation strategies rely on knowledge accumulated across repeated experiences. But how should options be evaluated in unfamiliar environments, in which knowledge is sparse? In these situations, how do decision makers make efficient use of limited past experience to guide their choices? One possible strategy is episodic sampling, in which a decision maker retrieves a small number of past decisions from memory to estimate the value of present options. By virtue of their age, children and adolescents have less experience than adults, making episodic sampling a particularly useful strategy for them. At the same time, the effectiveness of episodic sampling derives from memory's precision and context sensitivity-properties that continue to develop into adolescence and young adulthood. This tension raises a key question: are developmental differences in episodic memory associated with differences in episodic sampling? To address this question, 106 participants, ages 8 to 25, completed a 2-d choice task that dissociated the influence of a single episodic memory from the influence of multiple episodes sharing a common context. At all ages, single episodes biased choices. But, only adults were sensitive to the broader evoked context. Further clarifying the relationship between episodic memory and decision making, differences in memory precision predicted differences in episodic sampling, even after taking into account age, while episodic sampling, in turn, accounted for individual differences in forward planning. Together, these findings suggest that episodic memory guides decision making throughout development, but the character of its influence evolves as memory becomes more precise and richly structured.

Purpose

Method

Findings

Conclusions

Shared memories of event details are crucial to eyewitness testimony. When different people encode or recall the same event, similar scene-specific neural activity patterns emerge across individual brains. However, it remains unclear whether these patterns are specific to event details and how test expectancy (i.e., expecting free recall or general memory tests) and misinformation affect them. In this study, 100 participants were randomly assigned to view one of two versions of each event. Both versions featured identical scenarios, but with different details. About half of the participants were informed about the upcoming free recall before viewing events, while the others were told to expect a general memory test. Functional magnetic resonance imaging was used to record their brain activity during four stages: viewing original events, initial free recall, reading misinformation, and final free recall of original events. The neuroimaging data were analyzed based on the similarity of neural patterns across participants. Test expectancy increased the similarity of detail-specific neural activity patterns between individuals when they viewed original events in brain regions relevant for visual attention. Misinformation increased the likelihood of people forming shared false memories of event details. People who formed shared false memories exhibited similar detail-specific patterns of activity in the dorsomedial prefrontal cortex when reading misinformation. People who formed shared true memories exhibited similar detail-specific patterns of activity in the inferior parietal lobe when viewing original events, as well as in the ventrolateral prefrontal cortex and middle temporal gyrus when recalling them after exposure to misinformation. Our findings revealed that different brain regions of the default mode network play distinct roles in the encoding and recall of event details shared by individuals.

While AI coding tools have demonstrated potential to accelerate software development, their use in scientific computing raises critical questions about code quality and scientific validity. In this paper, we provide twelve practical tips for AI-assisted coding that balance the capabilities of AI with the demands of scientific and methodological rigor. We address how AI can be leveraged strategically throughout the development cycle with four key themes: problem preparation and understanding, managing context and interaction, testing and validation, and code quality assurance and iterative improvement. These principles serve to emphasize maintaining human agency in coding decisions, establishing robust validation procedures, and preserving the domain expertise essential for methodologically sound research. These tips are intended to help researchers harness AI's transformative potential for faster software development while ensuring that their code meets the standards of reliability, reproducibility, and scientific validity that research integrity demands.

Background

Objective

Methods

Results

Conclusions

Early in development, children can infer latent structure in the world from sparse and ambiguous evidence. Through a process known as structure learning, they extract statistical regularities, construct causal models from those regularities, and use those models to arbitrate between exploiting known options and exploring novel alternatives. In turn, each decision and its outcomes refine the model that produced them. Despite the clear reciprocal relationship between structure learning and decision-making in the real world, developmental research has largely examined these processes separately. To address this gap, we compared how children, adolescents, and adults behaved in a patch-foraging task designed to reveal how structure learning shapes exploratory decisions in a richly structured, dynamic environment. We found that younger participants left patches sooner than adults, enabling them to explore the environment more broadly within the fixed time window of the study. Computational modeling demonstrated that this difference in exploration arose from differences in participants' causal models of the environments. Younger participants grouped all patches into a single category despite large differences in richness, whereas older participants separated them into distinct categories. Despite differences in representation, participants of all ages used their uncertainty about the environment to guide their decisions. Together, our findings suggest that structure learning undergoes protracted development, but uncertainty-sensitive decision-making emerges earlier and can support adaptive behavior even when representations remain imprecise. SUMMARY: Children and adolescents formed less granular representations of environmental structure than adults. Despite these imprecise representations, younger participants showed adult-like sensitivity to uncertainty, planning further ahead when more confident in their internal models of the environment.

Cognitive maps support inference and planning by representing associations between experiences encoded in memory. These map-like representations are thought to carry information not only about directly observed links but also about longer paths. The ability to make judgments based on multi-step associations varies with one's experience in an environment and with changes in memory abilities across the lifespan. However, it remains unclear exactly how representations of associative structure are influenced by learning curricula and memory constraints. Prior studies have suggested a tradeoff: memory representations can either be more integrated to improve inference, or more separated to recall distinct direct associations. Whether overlapping associations are experienced nearby in time (interleaved) or spaced apart (blocked) can bias memory representations toward integration or separation. However, key recent findings about how blocked versus interleaved experience bias integration or separation have been inconsistent. Here, we introduce a computational framework that helps reconcile these apparent discrepancies. Using neural network simulations of three separate memory-guided inference tasks, we show that variations in memory capacity and the sparsity of neural codes interact with learning sequence to shape network representations. Specifically, blocked learning promotes integration when memory capacity is low, while interleaved learning promotes integration when memory capacity is high. Integration is especially likely to result from representations formed when neural codes are both sparse and distributed. These results offer a principled computational account of how flexible, map-like representations can arise from experience and suggest avenues for individualized memory interventions to improve inference, generalization, and planning.