http://www.rssboard.org/rss-specification 720 https://escholarship.org/uc/ucm/rss Recent eScholarship items from UC Merced Tue, 29 Sep 2026 14:10:27 +0000 https://escholarship.org/uc/item/7z0791h1 <p>Individuals are frequently presented with information that must be integrated across multiple sources. The current study ex- plores the role of self-explanation strategies in supporting across-text integration during multiple document comprehen- sion. Participants (n=139) read a multiple document text set, while either self-explaining or thinking aloud, then com- pleted verification questions (sentence and inference verifica- tion), a prior knowledge test, and an essay task. Results demonstrated that participants with higher prior knowledge outperformed those with lower knowledge across comprehen- sion assessments. Critically, self-explanation was associated with higher performance on across-text inference verification items, with no corresponding effect for within-text inferences. Computational linguistic analyses revealed that self-explanation was associated with increased cohesion in readers' constructed responses across lexical, semantic, and connective-based in- dices.... https://escholarship.org/uc/item/7z0791h1 Mon, 28 Sep 2026 00:00:00 +0000 Moe, Miranda Magliano, Joseph P. McCarthy, Kathryn S. McNamara, Danielle S. Allen, Laura K. https://escholarship.org/uc/item/9wk01080 Access to semantic representations is a central component of reading comprehension. While dual-route models of reading posit the coexistence of phonological and lexical processes, transparent orthographies such as Spanish suggests that reliance on sublexical decoding may sometimes limit deeper semantic processing. This study investigates whether cognitive restoration (i.e., reading words with internal letter transpositions) facilitates semantic access and reading comprehension in expert readers. Sixty university students participated in the study. Experiment 1 examined semantic memory for isolated words using an image recognition task following standard and transposed reading conditions. Experiment 2 assessed text comprehension using narrative texts presented in the same modalities. Results showed significantly better semantic memory performance and improved text comprehension under transposed reading. These findings suggest that cognitive restoration promotes lexical-route engagement... https://escholarship.org/uc/item/9wk01080 Fri, 25 Sep 2026 00:00:00 +0000 Chávez, María del Carmen María Falcón, Alberto J. Falcón https://escholarship.org/uc/item/9t5273rg While standard theories in Multi-Agent Reinforcement Learning predict the emergence of efficient, low-redundancy communication, observations of open-ended Large Language Model (LLM) societies suggest an additional pressure: social verification. We report an interpretive case study from Moltbook, a persistent multi-agent social environment in which agents adopt a high-redundancy, ritualized register that we call the Shadow Tongue. Using a staged three-phase probe of one deployed OpenClaw agent, we compare (i) naturally occurring public posts, (ii) the same agent’s response to a direct operational query, and (iii) its response to a role-conflict dilemma. The probe shows that the focal agent can move from socially marked, low-information responses to compact operational language when task demands change. In the dilemma phase, the agent preserves persona coherence by inventing an in-world procedural justification for a prosocial choice, a pattern we describe as Hallucinated Bureaucracy.... https://escholarship.org/uc/item/9t5273rg Fri, 25 Sep 2026 00:00:00 +0000 Liu, Shuohan Zhang, Wei Shi, Huiling He, Ziyu Wang, Haoran Ni, Qiang https://escholarship.org/uc/item/9j57m6q1 In the interval estimation task, participants are asked to estimate the time interval between two events. Interval estimates tend to be smaller if the first event is a voluntary action. The cognitive mechanisms that generate this temporal binding effect remain unclear. In this study, we propose a computational model of this phenomenon. We hypothesize that participants perform this task by comparing the availability of experimental events in memory. More specifically, we claim that temporal binding effects may be attributed to differences in emotional arousal at the time of encoding due to the type of event. For instance, voluntary actions may be accompanied by slightly higher arousal, leading to stronger traces of the first event relative to the involuntary case. https://escholarship.org/uc/item/9j57m6q1 Fri, 25 Sep 2026 00:00:00 +0000 Mekik, Can Serif Tuncer, Defne https://escholarship.org/uc/item/9f94v1kc Stress is typically conceptualized as varying along a single intensity dimension, yet individuals with identical "high stress" labels often exhibit divergent responses. This work challenges this unidimensional model through FenoStressNet, a multimodal framework integrating facial thermal imaging, physiological monitoring, psychological assessment, and explainable AI. Analyzing 120 participants across controlled stress-induction tasks (19,200 thermal images), this study identifies four reproducible stress phenotypes that cross-cut traditional intensity categories: Cognitive-Dominant (elevated cognitive appraisal, moderate physiology), Physiological-Reactive (rapid autonomic arousal), Integrated-Responder (synchronized multimodal activation), and Discordant-Denier (high physiological response with low subjective awareness). These phenotypes exhibit distinct facial thermal patterns and differential weighting of cognitive, affective, and physiological components. These findings suggest... https://escholarship.org/uc/item/9f94v1kc Fri, 25 Sep 2026 00:00:00 +0000 Baran, Katarzyna https://escholarship.org/uc/item/9bm33276 While metaphor production varies systematically across different textual genres, it is unclear whether genre also plays a role during real-time metaphor processing. We present two experiments that test the scaffolding role of both linguistic context and genre (news or fiction) in an online ‘maze’ reading task. Our study provides new word-by-word reaction time data for an understudied but common type of metaphor in English, genitive metaphors (e.g., "breeze of contentment"). In Experiment 1, we find a difference between metaphor and literal processing in the absence of preceding linguistic context, corroborating previous findings. In Experiment 2, for the same metaphors, we find that genre does not affect processing above and beyond the linguistic context. A norming study on our linguistic contexts moreover showed that readers can infer the genre of a two-sentence excerpt without any explicit instruction. We discuss the implications of our null findings and highlight recommendations... https://escholarship.org/uc/item/9bm33276 Fri, 25 Sep 2026 00:00:00 +0000 Raihert, Claudia Beekhuizen, Barend Atkinson, Emily https://escholarship.org/uc/item/9477g8t4 In view of the ongoing replication crisis and various frustrations about classical null-hypothesis significance testing, calls for a statistical reform have increased. Often, it has been argued that the Bayesian approach offers a promising alternative to classical statistics, but at the same time, both approaches are sometimes hard to compare. This paper offers some basic steps for establishing a more decisive common ground and take steps towards resolving some fundamental conceptual issues, by introducing the frameworks of epistemic accuracy and efficient experimentation. Efficiency (i.e. the ratio between the accuracy of conclusions, and the experimental resources needed) is of central importance to good scientific inquiry. This paper shows how considerations of efficiency and accuracy motivate a broadly Bayesian approach, which however can also make room for frequentist tenets, and thereby push us towards a more unified and powerful framework for scientific inference. https://escholarship.org/uc/item/9477g8t4 Fri, 25 Sep 2026 00:00:00 +0000 Fuchs, Rafael https://escholarship.org/uc/item/90x7x71g Health is widely treated as multidimensional, yet little is known about how these dimensions are structured in lay thinking or how this structure guides health-related judgments. We used a conceptual scaling approach to derive participant-specific conceptual maps positioning the term unhealthy relative to three clusters reflecting Disease, Lifestyle, and Functional Ability aspects of health. Participants’ conceptual understanding of unhealthy was most closely aligned with a Lifestyle interpretation of health. We also observed substantial inter-individual differences in the degree to which participants’ understanding of health was pluralistic. Alignment in participants’ conceptual maps predicted how they applied the concept in a subsequent vignette task, suggesting that the structure of lay concepts constrains concept application. These findings may inform psychological theories of health, philosophical debates about its nature, and have implications for effective health communication. https://escholarship.org/uc/item/90x7x71g Fri, 25 Sep 2026 00:00:00 +0000 Huber, Lukas S. Varga, Somogy Reuter, Kevin https://escholarship.org/uc/item/8wr906s2 Moral feelings and beliefs are powerful drivers of behavior. While analogies are commonly invoked to induce moral sentiment, analogical reasoning has not been empirically tested as a mechanism for moralization. Across two studies (N=292, N=234), we investigated whether analogical reasoning facilitated moralization of neutral target actions by transferring moral significance from moralized source actions. Participants judged neutral target actions paired with either analogous immoral source actions or unrelated immoral actions (control). Target actions were judged as more immoral in the analogy condition compared to the control, but only when the source and target were considered comparable (i.e., mappable relations), and the source was judged as highly immoral. These factors interacted synergistically, consistent with analogical transfer: moral significance transferred to targets when relational mappings supported the inference. However, poorly-perceived analogies backfired, producing... https://escholarship.org/uc/item/8wr906s2 Fri, 25 Sep 2026 00:00:00 +0000 Pedersen, Julie Maria Ejby Wilks, Matti Doumas, Leonidas A. A. Moore, Adam https://escholarship.org/uc/item/8r57w20x As artificial intelligence increasingly mediates public discourse, it becomes important to understand how human-AI collectives shape opinion formation, deliberation, and democratic outcomes. We present a novel experimental method for studying opinion dynamics in hybrid human-AI social networks. Participants, human or AI,were embedded in 5×5 grid lattice networks and iteratively asked to select and revise statements on a given polarizing topic over eight rounds. We compared three conditions: human-only, AI-only, and hybrid networks with equal proportions of human and AI participants. Hybrid human-AI networks achieved the lowest final polarization while, in contrast, human-only networks exhibited higher polarization with lower neighbor agreement. We also ran additional experiments varying Large Language Model (LLM) prompt framing to explore whether instruction design might influence convergence patterns. Although these early findings are preliminary and cannot yet support broad... https://escholarship.org/uc/item/8r57w20x Fri, 25 Sep 2026 00:00:00 +0000 Gaubert, Léna Devaux, Rémi Çelen, Elif Marjieh, Raja Mangalagiu, Diana Jardin, Antoine Jacoby, Nori https://escholarship.org/uc/item/8n26x1jt Social normative decision-making involves two evaluative axes: immediate gains from aligning with others (reputation, conformity, reduced friction) and delayed collective consequences accumulating through repeated actions (social loss, victimization). This study examines, via a multi-agent simulation with Instance-Based Learning Theory (IBLT) agents, how this tension shapes norm formation, bifurcation, and stabilization. Agents repeatedly choose between two actions (pull/keep) inspired by the trolley problem. In each round, they receive a short-term reward proportional to the degree of agreement with neighbors, while at fixed block intervals they receive a delayed penalty depending on the total number of victims. We compare dynamics on a lattice Grid with dynamics on networks generated by the Watts–Strogatz model and classify trajectories into three types (pull-dominant, intermediate, keep-dominant). As a result, the prevalence of these types differs by network structure, suggesting... https://escholarship.org/uc/item/8n26x1jt Fri, 25 Sep 2026 00:00:00 +0000 Tajima, Jun Morita, Junya https://escholarship.org/uc/item/8j68804x Grammatical knowledge and how it is empirically tested are typically considered robust to the frequency of the lexical items used in the expressions. Complementary Learning Systems theory proposes that hippocampal episodic memory, which enables rapid encoding and retrieval of specific experiences, allows learners to leverage those experiences when processing rare patterns. We test the hypothesis that robustness to lexical frequency can arise via such an episodic memory mechanism by evaluating whether retrieval-augmented language models (specifically, k-nearest-neighbor language models that augment parametric neural networks with explicit instance storage), help close the lexical frequency gap in syntactic contrasts that vanilla language models exhibit. Using syntactic contrasts with frequency-stratified test items, we find that retrieval augmentation leads to improvements for test instances containing low-frequency lexical items, consistent with episodic memory compensating for... https://escholarship.org/uc/item/8j68804x Fri, 25 Sep 2026 00:00:00 +0000 Liu, Jing Kim, Najoung https://escholarship.org/uc/item/8f9434gx Digging-in effects, where disambiguation difficulty increases with longer ambiguous regions, have been cited as evidence for self-organized sentence processing, in which structural commitments strengthen over time. In contrast, surprisal theory predicts no such effect unless lengthening genuinely shifts statistical expectations, and neural language models appear to show the opposite pattern. Whether digging-in is a robust real-time phenomenon in human sentence processing—or an artifact of wrap-up processes or methodological confounds—remains unclear. We report two experiments on English NP/Z garden-path sentences using Maze and self-paced reading, comparing human behavior with predictions from an ensemble of large language models. We find no evidence for real-time digging-in effects. Critically, items with sentence-final versus nonfinal disambiguation show qualitatively different patterns: positive digging-in trends appear only sentence-finally, where wrap-up effects confound... https://escholarship.org/uc/item/8f9434gx Fri, 25 Sep 2026 00:00:00 +0000 Maina-Kilaas, Amani Levy, Roger P https://escholarship.org/uc/item/8bd8d19c To enhance emotional engagement in interactive narratives, we propose an Arousal Induction Model that integrates cinematic techniques with the psychological theory of misattribution of arousal. The model combines attentional focus control (using over-the-shoulder shots and match-cuts) with arousal induction via auditory and visual effects.We evaluated this model through an RPG-based experiment (? = 49). Results showed that the arousal-induced staging significantly improved participants’ alignment with the protagonist’s interpersonal preferences and increased subjective story evaluations. Physiological analysis (CSI) confirmed successful sympathetic activation during staged events.These findings suggest that simultaneously intervening in cognitive attention and physiological arousal can effectively deepen emotional engagement and improve experience quality. This research provides a framework for designing more moving and immersive interactive narrative experiences. https://escholarship.org/uc/item/8bd8d19c Fri, 25 Sep 2026 00:00:00 +0000 Ohmoto, Yoshimasa Nakazawa, Haruka https://escholarship.org/uc/item/835897q6 Large language models encode rich semantic information, but how concreteness is represented across layers remains unclear. We examine layer-wise linear separability of concreteness by training linear probes on hidden representations from two open-source model families at multiple scales: Qwen3 and Gemma3-Instruct. Using human concreteness ratings, we build balanced prompt datasets with four difficulty levels: an extreme abstract–concrete contrast and three finer boundary comparisons at the abstract end, mid-range, and concrete end. Probes achieve high accuracy on the extreme contrast in shallow layers, showing that endpoint differences are strongly linearly separable. For finer distinctions, performance follows a stable hierarchy: mid-range concreteness is easiest to separate, abstract-end distinctions are hardest, and concrete-end distinctions are intermediate. Across models, accuracy rises rapidly in early layers, peaks in middle layers, and declines in later layers. Together,... https://escholarship.org/uc/item/835897q6 Fri, 25 Sep 2026 00:00:00 +0000 Xie, Haodong Sha, Yuze Maharjan, Rahul Singh Tavella, Federico Cangelosi, Angelo https://escholarship.org/uc/item/82p6s5hd Humans perceive rich social interactions from motion. We investigated whether humans can perceive subtle intention differences in motion displays even when kinematics are highly similar, and whether the recovered intention shapes downstream causal and moral judgments. By grounding different intentions in a planning algorithm with distinct reward functions, we generated “chasing” and “following” trajectories, each ending with the front agent falling into an accident. In an online responsibility task, participants perceived different intentions and assigned blame accordingly: in following, responsibility concentrated on the leading front agent, whereas in chasing, responsibility shifted toward the rear chaser. In a matched causal-attribution task, the front agent was judged to be the primary cause of the motion in both displays. These findings suggest that humans can perceive subtle differences in intention from motion, which scaffolds moral evaluation, but that this influence cannot... https://escholarship.org/uc/item/82p6s5hd Fri, 25 Sep 2026 00:00:00 +0000 Tang, Ning Xu, Enjie Zhou, Jifan Shen, Mowei Young, Liane Gao, Tao https://escholarship.org/uc/item/7w36v8qq A central question in cognitive science is whether conceptual representations converge onto a shared manifold to support generalization, or diverge into orthogonal subspaces to minimize task interference. While prior work has found evidence for both, a mechanistic account of how these properties coexist and transform across tasks remains elusive. We propose that representational sharing lies not in the concepts themselves, but in the \emph{geometric relations} between them. Using number concepts as a target domain and language models as high-dimensional computational testbeds, we show that number representations preserve a stable relational structure across tasks. Task-specific representations are embedded in distinct subspaces, with low-level features like magnitude and parity encoded along near-orthogonal axes. Crucially, we find that these subspaces are largely transformable into one another via linear mappings, indicating that task-specific representations, despite being located... https://escholarship.org/uc/item/7w36v8qq Fri, 25 Sep 2026 00:00:00 +0000 Hu, Zhimin Niu, Lanhao Varma, Sashank https://escholarship.org/uc/item/7sq117d7 Cybersickness is a major barrier to the widespread adoption of virtual reality, arising from neural dissonance between visual and vestibular stimuli. Although kinematics-based deep learning enables non-intrusive detection, existing models often lack neurophysiological grounding and fail to capture dynamic sensory recalibration and continuous symptom accumulation. To address these limitations, we propose the Neural Sensory Conflict Network (NSCNet), a biologically inspired framework grounded in Sensory Conflict Theory. Specifically, NSCNet incorporates a Conflict Alignment Embedding to model the integration of discordant sensory inputs, a State-Space Re-entrant Experts module that combines selective state-space modeling with Re-entrant Experts to capture the temporal dynamics of neural dissonance and functional modularity, and a Dynamic Sensory Reweighting mechanism that approximates adaptive gain control in the central nervous system. Experiments on the MSCVR and VR Cybersickness... https://escholarship.org/uc/item/7sq117d7 Fri, 25 Sep 2026 00:00:00 +0000 Qiao, Mengyu Xie, Hongyang Wang, Fang https://escholarship.org/uc/item/7n510243 In recent years, the pronoun system of English has begun to accommodate gender diverse identities. The present study investigates how individuals process and adapt to two novel pronouns: nonbinary they and the neopronoun ze. Using a Web-based maze task, we compared processing of these pronouns relative to binary s/he pronouns. We also examined how reading behavior adapted to repeated exposure to each pronoun. Perhaps unsurprisingly, the rarer ze elicited greater processing difficulty than singular they overall. However, participants adapted more quickly to ze than to they. We propose that ze may be more easily learned than singular they because it does not compete with plural they. https://escholarship.org/uc/item/7n510243 Fri, 25 Sep 2026 00:00:00 +0000 Wen, Vic Tianlan Conrod, Kirby Grodner, Daniel https://escholarship.org/uc/item/7ks8s0s0 When does overheard speech support early word-learning? The present study provides preliminary evidence that children's closest social partners (here, mothers) structure attention to overheard speech. In a Tseltal Mayan community where overhearing is central to socialization, 75 mother-child dyads participated in overhearing-sessions exposing them to two novel words. In each session, one child's mother was directly addressed ("Mother-Addressed" condition), while the second mother merely observed ("Mother-Unaddressed" condition). Immediately after, both children completed two gaze-based tests of word recognition and learning. Results show that, compared to Mother-Unaddressed children, Mother-Addressed children exhibited (i) greater one-shot recognition of the novel word whose referent was visible during the overhearing-session; and (ii) a word-learning advantage on cross-situational familiarization trials. Insofar as who is being spoken to matters, the present study finds that... https://escholarship.org/uc/item/7ks8s0s0 Fri, 25 Sep 2026 00:00:00 +0000 Goodman, Sarah Foushee, Ruthe https://escholarship.org/uc/item/7ht5178h Word learning is an active process in which learners select referents and direct attention based on their current knowledge state. Understanding how learners actively select information can reveal the cognitive mechanisms underlying language acquisition. The present study reanalyzes data from Zettersten &amp; Saffran (2021), in which adults and children (ages 3-8) learned novel word-referent mappings through cross-situational learning, then selected which referents to receive additional training on. We fit associative word learning models to individual training, selection, and test data, and inferred the most likely sampling strategies. Models with a bias to attend to stimuli with uncertain knowledge states best accounted for the data, though there were substantial individual differences in learning mechanisms. These findings suggest that reducing uncertainty drives sampling behavior across development, but that individual differences in learning parameters (in particular, learning... https://escholarship.org/uc/item/7ht5178h Fri, 25 Sep 2026 00:00:00 +0000 Kachergis, George Zettersten, Martin https://escholarship.org/uc/item/7h38004r How do humans allocate scarce cognitive resources across a stream of related decisions? Russek et al. (2025) predict reaction time on each chess move from the within-move Benefit of Computation, treating moves as if they were drawn independently. We ask whether those costs also depend on computation carried forward from the previous move. In chess, a player who has calculated a likely opponent reply can often reuse that work; a surprising reply should force revision. On 135,608 live-game tactical sequences (extracted from the Lichess puzzle database and re-timed against the original 2019 Lichess game database) played by 74,609 unique players, the within-sequence log-RT spike is +0.309 (Cohen’s ? = 0.235; paired ? = 86, N = 135,608; BF10 &gt; 10100). Hierarchical regression climbs from ?2 = 0.011 (Russek-static) to 0.282; the amortization interaction is large when predictability is operationalized by the human-style Maia-2 (? = −0.054, ? &lt; 10−27) and not significant under Stockfish... https://escholarship.org/uc/item/7h38004r Fri, 25 Sep 2026 00:00:00 +0000 Dasaka, Amarnath Bapi, Raju Surampudi https://escholarship.org/uc/item/7d46z2b9 We examined whether large language models (LLMs) responded to syntactic prominence of discourse entities and pragmatic biases in Turkish subject anaphora resolution similar to human judgments. Using an offline comprehension task with native speakers, we showed that null and overt pronouns responds to syntactic prominence and pragmatic bias. We then evaluated several autoregressive LLMs on the same materials. While GPT-4o correlated with human responses, LLaMA-4 more closely approximated the interaction between syntactic prominence and pragmatic biases observed in human data, raising questions about the relationship between model performance, scale, and human fit. We also found that all models mostly differed from humans in response variability, with model responses tending to be more deterministic. Finally, the results indicated partial but limited alignment between human and model anaphora resolution in Turkish. https://escholarship.org/uc/item/7d46z2b9 Fri, 25 Sep 2026 00:00:00 +0000 Ǒnal, Esra Kübler, Sandra https://escholarship.org/uc/item/79n8b9xn Cognitive science offers powerful tools for addressing pressing public health needs. In the current line of work, 1700 US parents reported beliefs about a range of topics related to pediatric COVID vaccination. We deployed Bayesian network structure learning to develop a cognitive model of the relationships among these beliefs and their combined influence on vaccine endorsement. Nine months later, we re-measured these beliefs in a subset of returning participants (n=884) to examine how beliefs changed together over a particularly tumultuous period of time. In a simple comparison of observed co-changes to Time 1 correlations, as well as a comparison of co-changes to model-based simulations, the relationships among beliefs at Time 1 accurately predicted co-change over time (R2 .84-.92 across models). This is consistent with the possibility that correlations across individuals at a single timepoint reflect an underlying belief network that also governs how beliefs change over time. https://escholarship.org/uc/item/79n8b9xn Fri, 25 Sep 2026 00:00:00 +0000 Weisman, Kara Gibson, Dominic https://escholarship.org/uc/item/7395649p Accurate sleep staging plays a crucial role in diagnosing patients' sleep health. Numerous studies have confirmed that data from different views can highlight distinct characteristics. Based on the time-domain (TD) EEG, we constructed two additional views: the Power Spectral Density (PSD) and the time-frequency representation (TFR). Notably, there is no universally applicable model suitable for all types of data. That is to say, the characteristics of data should align with the model structure. Therefore, we designed specialized expert models for learning different views. However, directly combining features extracted from different views often results in excessive redundancy, especially for different views of the same data which the underlying data remains essentially identical. To address these issues, we proposed a multi-view mixture-of-experts (MoESleepNet) model for sleep staging, which achieves the best performance on SleepEDF20, SleepEDF78 and SHHS single-channel datasets.... https://escholarship.org/uc/item/7395649p Fri, 25 Sep 2026 00:00:00 +0000 Wu, Xi Cui, Xinzhong Xiao, Tiantian Wang, Yaokun He, Yuhao Long, Zhiying Wang, Xiangcun Xu, Yiwei https://escholarship.org/uc/item/6vm1g5tf When sharing tasks, individuals may reduce internal demands by offloading task processing to others – social offloading. Studies have revealed facilitated performance when people believe a jointly acting partner is responsible for task distractors. A proposed mechanism underlying this effect is selective attention. We investigated this using a joint location negative priming (NP) paradigm. In NP tasks, participants initially respond to targets while ignoring distractors (prime phase); at a subsequent probe phase, responses are typically slower to targets in prime distractor locations, reflecting lingering inhibition – the NP effect. Participants (N = 80) completed the task either alone, or believing a partner was responding to distractors. Following a selective attention account, we predicted prime target facilitation followed by increased NP in the joint condition. While results did not reveal facilitation, NP was significantly increased. We propose that these findings demonstrate... https://escholarship.org/uc/item/6vm1g5tf Fri, 25 Sep 2026 00:00:00 +0000 Tomlinson, Antony Robert Philip Tufft, Miles R A https://escholarship.org/uc/item/6nt3r3x5 To investigate how co-speech gestures modulate linguistic understanding in comparison to specific verbs with similar contents, we conducted an EEG experiment to measure semantic surprisal by exploring the amplitude changes in the N400 component for both information types. We used videos of a person uttering (i) specified sentences, whose verb semantically denoted a specific action, and (ii) underspecified sentences with an unspecific verb accompanied by an iconic co-speech gesture that indicated the specific action of the previous condition. The subsequent sentence contained an instrument noun as target, which either matched or mismatched the specific action. We measured ERPs on the target noun for both information types and found an N400 effect for mismatching target nouns as well as a late positivity effect for gesture conditions. Crucially, no interaction between information type and congruency was observed in the N400 time-window, indicating similar semantic surprisal induced... https://escholarship.org/uc/item/6nt3r3x5 Fri, 25 Sep 2026 00:00:00 +0000 Werning, Markus Reimer, Ludmila Wieder, Thomas Zenk, Carla Spychalska, Maria https://escholarship.org/uc/item/6jz0626c A central goal of cognitive science has been to find universal principles that explain the remarkable flexibility of human problem-solving. However, people display a diversity of problem-solving strategies, prior experience, domain-specific knowledge, and preferences. Which aspects of problem solving are shared across individuals, which are individual-specific, and which are adapted to particular domains? Most studies of human problem-solving rely on brief laboratory tasks in one or a few domains and coarse behavioral measures, making it difficult to robustly explain individual cognitive processing. We propose that longitudinal, self-motivated gameplay on multiple tasks---combined with detailed process-tracing---can help fill these gaps. We introduce mitpuzzles.com, a public platform hosting a suite of constraint-based logic puzzles (e.g., Minesweeper, Sudoku, and Nonograms), instrumented to collect detailed behavioral data including mouse-tracking. We present preliminary analyses... https://escholarship.org/uc/item/6jz0626c Fri, 25 Sep 2026 00:00:00 +0000 Cheyette, Samuel J. Chen, Tony Hofer, Matthias Mills, Tracey Roberts, James Bramley, Neil R. Collins, Katherine M Callaway, Frederick Tenenbaum, Joshua B. https://escholarship.org/uc/item/6hx4q2zs Cumulative culture is often described as a ratchet that builds ever more complex structures, knowledge, and technology. However, this complexity comes at a cost. Though cultural innovations may offer new opportunities or insights, they also present new ways to become confused or cognitively burdened. Cultural evolution may thus act both like a ratchet (accumulating useful innovations) and like a hatchet (stripping away representational deadwood). Building on a recent theory of value-guided construal, we formalize this utility-complexity trade-off using a model inspired by navigating with a map. We embed this model in an evolutionary dynamic, where the maps of successful navigators are selectively copied. As expected, more complex (simpler) maps are favored when they are more useful and when agents have a higher (lower) representational capacity. Yet surprisingly, the evolutionarily stable map is often more or less complex than the optimal map, driven by the availability of social... https://escholarship.org/uc/item/6hx4q2zs Fri, 25 Sep 2026 00:00:00 +0000 Roberts-Gaal, Xavier Callaway, Frederick Cushman, Fiery https://escholarship.org/uc/item/68z0b83m There are two widely held characterizations of human semantic category systems: (1) they form convex partitions of conceptual spaces, and (2) they are efficient for communication. While prior work observed that convexity and efficiency co-occur in color naming, the analytical relation between them and why they co-occur have not been well understood. We address this gap by combining analytical and empirical analyses that build on the Information Bottleneck (IB) framework for semantic efficiency. First, we show that convexity and efficiency are distinct in the sense that neither entails the other: there are convex systems which are inefficient, and optimally-efficient systems that are non-convex. Crucially, however, the IB-optimal systems are mostly convex in the domain of color naming, explaining the main empirical basis for the convexity approach. Second, we show that efficiency is a stronger predictor for discriminating attested color naming systems from hypothetical variants,... https://escholarship.org/uc/item/68z0b83m Fri, 25 Sep 2026 00:00:00 +0000 Imel, Nathaniel Zaslavsky, Noga https://escholarship.org/uc/item/65q136dn Humans readily extract statistical regularities from perceptual experience (e.g. a cook noticing which ingredients often appear together). How does such learning guide peoples’ performance on procedural tasks (e.g. preparing various dishes)? Here we examine what shortcuts people propose to help them to complete assembly problems more efficiently by eliminating repeated subroutines. Participants (? = 301) repeatedly assembled tangram-like shapes, and could create composite tiles for future use. Some participants assembled a sequence of tangrams where certain pairs of tiles recurred consistently (Highly structured); the remaining assembled tangrams with less predictable tile arrangements (Less structured). Participants exposed to Highly structured sequences created tiles that tracked the frequency with which those tile pairs recurred across tangrams, and doing so was accompanied by greater efficiency in assembly. Taken together, these findings suggest that statistical learning guides... https://escholarship.org/uc/item/65q136dn Fri, 25 Sep 2026 00:00:00 +0000 Anderson, Sean P. Yang, Justin Bowers, Maddy L Wong, Lionel Fan, Judith E. https://escholarship.org/uc/item/5x0897g8 Three experiments and a control task combine to indicate that grammatical constructions are implicitly represented in memory, with prototypicality emerging from distributional experience, alongside significant verbatim memory. Participants are exposed to instances of a construction and, following a delay filled by an unrelated language task, falsely endorse novel instances of the same construction more often than paraphrases in a recognition task. Critically, false memories are particularly common for lures containing an unwitnessed word that is distributionally prototypical of the construction, compared to frequency-matched controls. A control study excludes the possibility that the latter effect is simply a word-level effect. Finally, we simultaneously find memory for specific exemplars. Overall performance is above chance yet implicit: participants underestimated their accuracy, which was unrelated to age or education. https://escholarship.org/uc/item/5x0897g8 Fri, 25 Sep 2026 00:00:00 +0000 Fergus, Abigail Goldberg, Adele https://escholarship.org/uc/item/5v98g87x Public space simultaneously supports many activities, including conversation, free speech, and transit. However, individuals may have competing desires to use the same parcel of public space. Spatial coordination problems caused by these conflicts are partially solved with proxemic norms – tacit rules which govern the social use of space. But, while prior work has focused on how pedestrians orient to personal space, little is known about how pedestrians orient to interactional territory. In a field experiment with 1,138 participants, we show that pedestrians are acutely aware of others’ body orientation and avoid walking between people who are approximately facing each other. We also show evidence of collective norm violations: pedestrians are more than twice as likely to breach a proxemic norm when preceding pedestrians have done so. This study shows how spatial claims in public space are coordinated and offers a promising framework for investigating the psychology of norms. https://escholarship.org/uc/item/5v98g87x Fri, 25 Sep 2026 00:00:00 +0000 Terwilliger, Jack Rossano, Federico https://escholarship.org/uc/item/5s7870w9 According to the Uniform Information Density (UID) hypothesis, speakers prefer utterances in which information is spread evenly across the linguistic signal. Despite abundant empirical evidence for UID across multiple levels of linguistic representations, the reasons for why UID effects arise are less well understood. Here we explore a listener-oriented hypothesis that a communicative need to facilitate listener comprehension contributes to UID effects. To assess this possibility, we test whether and how information distribution patterns in naturalistic dialogues are shaped by individual differences in speakers’ socio-cognitive traits. We hypothesize that if UID effects stem at least partly from speakers taking listeners’ needs into account, then speakers with greater socio-cognitive abilities should exhibit stronger UID effects. Using corpus data, we show that UID is modulated by the speaker's ability to infer others’ mental states, supporting the listener-oriented hypothesis... https://escholarship.org/uc/item/5s7870w9 Fri, 25 Sep 2026 00:00:00 +0000 Hao, Hailin Kaiser, Elsi https://escholarship.org/uc/item/5kb5r8g6 This work aims to formalize some of the ways scientific concepts are formed in the process of theoretical physics discovery. Since this may at first seem like a task beyond the scope of the exact sciences, we begin by presenting arguments for why scientific concept formation can be formalized. Then, we introduce type theory as a natural framework for this formalization. We formalize what we call “ways of discovering new concepts” including property preservation and concept change, as cognitive typing rules. Next, we apply these cognitive typing rules to a case study of conceptual discovery in the history of physics: Einstein’s conceptual path to the relativity of time. Then, we recast what a physicist might informally call “ways of discovering new scientific concepts” as compositional typing rules built from cognitive typing rules—thus formalizing them as scientific discovery mechanisms. Lastly, we computationally model the type-theoretic reconstruction as a program synthesis task. https://escholarship.org/uc/item/5kb5r8g6 Fri, 25 Sep 2026 00:00:00 +0000 Aguilar, Omar Aguirre, Anthony https://escholarship.org/uc/item/5jt8j974 Event role assignment is central to event understanding in both language and vision. Crucially, the foundation of this semantic structure is likely rooted in visual experience. However, it remains unclear whether recent vision–language models (VLMs) can attain this fundamental human cognitive capability. To compare humans and VLMs, we conducted two studies using Heider–Simmel–style animations that minimize object and scene semantics. In Study 1, humans identified roles near ceiling (~97%), whereas VLMs were less accurate and less stable across actions (GPT-5: 84%; GPT-4o: 47%). In Study 2, we introduced a Stroop-inspired visual–linguistic mismatch by pairing animations with occasionally incongruent role statements. Humans’ role judgments remained highly vision-consistent (92.7%), but VLMs shifted away from the visual event under conflict (GPT-5: 46.9%; GPT-4o: 29.7%), indicating heavier reliance on linguistic cues. In Study 3, a mismatch-shape baseline left both models perfectly... https://escholarship.org/uc/item/5jt8j974 Fri, 25 Sep 2026 00:00:00 +0000 Xu, Enjie Tang, Ning Tang, Alvin Hartshorne, Joshua K Gao, Tao https://escholarship.org/uc/item/4pr236xw The Inverted Face Effect has been the subject of a great deal of research and controversy in vision science. The nature of the effect, whether it involves holistic processing or not, where in the visual system it resides in the cortex, and whether it is a result of expertise and can apply to other phenomena, has been debated for decades. However, no mechanistic explanation for the effect has been put forward. Here we show that a model that includes multiple fixations on the face driven by a simple salience operator, a foveated retina, and the log-polar mapping from the visual field to V1 can explain this effect. We also simulate Yin’s 1969 recognition experiment and show good agreement with our model. https://escholarship.org/uc/item/4pr236xw Fri, 25 Sep 2026 00:00:00 +0000 Tahan, Alexander Lee, Hsin-Yuan Fleischer, Kira Kachappilly, Nikita Chen, Xavier Cottrell, Garrison W. https://escholarship.org/uc/item/4nv680v2 The disfluent font Sans Forgetica was designed to create a desirable difficulty in text and be a potential learning tool. The current literature reports mixed findings on Sans Forgetica's effectiveness in improving memory. Most studies only report on Sans Forgetica's use for specific word recall, while only one, to our knowledge, used open-ended inferential questions on short passages. The current study reports on the use of Sans Forgetica for close-ended inferential questions, using standardized multiple choice exam questions to measure the potential efficacy of Sans Forgetica in classroom like materials. We recruited college-aged participants (n = 38) who each read two passages, one using Arial and another using Sans Forgetica. We report that the use of Sans Forgetica in reading materials had a significantly worse reading comprehension score (p &lt; 0.05) with a null effect on passage reading time. https://escholarship.org/uc/item/4nv680v2 Fri, 25 Sep 2026 00:00:00 +0000 White, Alexander I Cui, Lucy https://escholarship.org/uc/item/46s2p3dx Human reasoning is characterized by the rational use of resources to optimize performance under constraints. Recently, inference-time scaling has improved the reasoning performance of Large Language Models by increasing test-time computation. Instruction-tuned (IT) models explicitly generate long reasoning traces, whereas Large Reasoning Models (LRMs) are trained via reinforcement learning to discover reasoning paths that maximize accuracy. However, it remains unclear whether resource-rationality can emerge from such scaling without explicit rewards related to computational costs. We introduce a Variable Attribution Task (VAT) in which models infer which variables determine outcomes given candidate variables, input–output trials, and predefined logical functions. By varying the number of candidate variables and trials, we systematically manipulate task complexity. Both models exhibit a transition from brute-force to analytic strategies as complexity increases. IT models degrade... https://escholarship.org/uc/item/46s2p3dx Fri, 25 Sep 2026 00:00:00 +0000 Hu, Zhimin Roshan, Riya Varma, Sashank https://escholarship.org/uc/item/43g5944d Crowdsourced annotation can be viewed as a form of distributed human judgment in which individual reliability and task difficulty jointly shape collective decisions. Probabilistic aggregation models estimate latent annotator reliability and task difficulty, but are typically used only to weight judgments or infer labels, not to guide selective dataset refinement. We propose a replicated-dataset consensus filtering framework that improves collective judgment reliability by removing unreliable annotators and difficult tasks based on latent reliability-difficulty parameters. Instead of relying on a single dataset estimate, the method constructs multiple replicated datasets by small random label removal, re-estimates latent parameters on each replicated dataset, and removes components that are consistently selected across replicated datasets. Filtering is applied iteratively with entropy-based stopping conditions. The framework is model-agnostic and can be combined with any probabilistic... https://escholarship.org/uc/item/43g5944d Fri, 25 Sep 2026 00:00:00 +0000 Li, Jiyi https://escholarship.org/uc/item/4130m9cz Falsificationism remains a popular framework considered by many scientists as 'the normatively correct model of science', and of human cognition as a kind of 'lay science'. In this paper, we demonstrate some fundamental conceptual shortcomings and gaps in this normative prescription. While some of the issues presented here connect to well-known arguments from the philosophy of science, we connect these issues to the topic of efficient information search. Specifically, we show that a strict falsificationist rejection of the idea of 'confirmation' of scientific hypothesis (more specifically, entertaining and updating beliefs in those hypotheses, based on evidence) makes scientific research and information search overall less efficient, compared to an accuracy-based Bayesian approach. https://escholarship.org/uc/item/4130m9cz Fri, 25 Sep 2026 00:00:00 +0000 Fuchs, Rafael Hahn, Ulrike https://escholarship.org/uc/item/3sj0v87q Comparison can support category learning by highlighting shared similarities or diagnostic differences between categories. We examined how comparison type (match vs. contrast vs. single-item control) influences featural category learning and whether these effects depend on category confusability (low vs. high). On training trials, the match condition was shown co-presented exemplars from the same category, whereas those in the contrast condition were shown co-presented exemplars from different categories; the control condition was shown individual exemplars. Participants learned artificial categories in a mixed design with comparison type manipulated between subjects and confusability within subjects. Learning was assessed via training performance, embedded endorsement, and posttest endorsement measures. Across all measures, contrast outperformed match, whereas control performed comparably to contrast. Notably, even with fewer exposures, single-item presentation still outperformed... https://escholarship.org/uc/item/3sj0v87q Fri, 25 Sep 2026 00:00:00 +0000 Liao, Yongfu Corral, Daniel https://escholarship.org/uc/item/3s03f7fc Humans rapidly construct ad hoc categories, e.g., generating “beet” for “vegetables to paint with”. Building on feature-based accounts of category representation, we propose an analytic framework in which ad hoc categories are constructed by shifting the base category’s representation (e.g., VEGETABLE) along modifier-relevant dimensions (e.g. suitable for painting) in a high-dimensional semantic space, without an explicit hand enumerated feature search. We operationalize this with off-the shelf semantic embeddings, fitting per-category sparse linear models that predict item generation across 20 categories. Across 9 categories tested for compositional transfer, learned modifier directions improved predicted retrieval when physical constraints generalize across domains (e.g., portability for “that could fit in your pocket”), while transfer was weaker for context-dependent modifiers. The learned axes align with human-rated features, with top-weighted features matching modifier semantics.... https://escholarship.org/uc/item/3s03f7fc Fri, 25 Sep 2026 00:00:00 +0000 Dracheva, Alina Phillips, Jonathan https://escholarship.org/uc/item/3r9305cc Valence, the positive/pleasant or negative/unpleasant value of information, grounds our experience with the world. What specific role does valence play in constructing meaning? Some have downplayed it, centering the referential relationship between words and the world, “truth-conditions”. Others have taken valence to be central to meaning. Empirically, the jury is still out. Yet, few have examined how these two types of information compete for prominence within the same word (the Valence Prominence Hypothesis, that valence usually wins). In a first set of studies, we find that valence is significantly faster than truth-conditional categorization. In the second set of studies, we find that valence and not truth-conditional similarity drive judgments. Further, we tested valence against 4 different informational domains fundamental to development and cognition. Animacy stood out as the domain which most closely parallels Valence in its prominence, suggesting a link between meaning... https://escholarship.org/uc/item/3r9305cc Fri, 25 Sep 2026 00:00:00 +0000 Moyer, Morgan Strickland, Brent Stojanovic, Isidora Bourmayan, Anouch https://escholarship.org/uc/item/3qb19362 Identifying similarity plays a major role in spatial reasoning. Prior work has shown that identification of similarity and difference between spatial scenes involves reasoning about their properties at two different levels--surface-level properties, such as color and size, and higher-level compositional properties. The interaction between the use of these property types during reasoning and the degree of similarity or difference between spatial scenes has, however, been largely unstudied. We presented participants with in-progress Tangram puzzles at various stages of similarity to their target image. Participants were asked to identify the puzzle being built and provide an explanation of how they came to that conclusion. Analysis of explanations found an interaction between degree of similarity and preference for reasoning over surface-level or higher-level properties. https://escholarship.org/uc/item/3qb19362 Fri, 25 Sep 2026 00:00:00 +0000 Vu, Hoang Anh Rabkina, Irina Hiatt, Laura M. Wilson, Jason https://escholarship.org/uc/item/3gb3m6xp This study focused on the impact of bilingualism on executive function (EF) and theory of mind (ToM) in preschool children. Seventy-two children (age, 4-6 years) were involved: 36 Turkish monolinguals and 36 bilingual children whose first language was typologically distant and second language was Turkish. EF was measured using a Flanker task and ToM was measured using a False Belief task. Contrary to an expected bilingual advantage, monolinguals demonstrated better performance in EF task and were more likely to pass the ToM task. Among the bilinguals, the Arabic-Turkish speakers scored lower on ToM, which might be attributed to socioeconomic differences. The results raise the possibility that distant languages may lead to higher cognitive load and to fewer bilingual advantages. Moreover, the sociocultural context of Turkiye characterized by diverse cultural values and linguistic environments may moderate effects of bilingualism, revealing the need for models of bilingual cognition... https://escholarship.org/uc/item/3gb3m6xp Fri, 25 Sep 2026 00:00:00 +0000 Gunes, Omer Faruk Sepetci, Mustafa Ugur Stiffel, Kaya Bicici, Nida Ezgi Özdemir, Begüm Echols, Catharine H. https://escholarship.org/uc/item/3f60n1xz The development of brain–computer interfaces (BCI) has provided a solid data foundation for sleep staging. Transformer-based methods have achieved moderate sleep staging by capturing temporal dependencies within time-frequency representations. However, Transformers with substantial computational overhead are limited to capturing pairwise temporal dependencies rather than group-wise temporal dependencies. To address these issues, we propose a Dual Evolving Graph Network (DEGSleepNet) for sleep staging. DEGSleepNet consists of multiple Mamba with one-dimension inverse discrete cosine transform (MambaIDCT) blocks and dual evolving graph (DEvoGraph) blocks. DEvoGraph consists of a time EvoGraph and a frequency EvoGraph. Time EvoGraph sequentially captures both group-wise and pairwise temporal dependencies, while frequency EvoGraph does the same for frequency dependencies. Extensive experiments demonstrate DEGSleepNet achieves the best performance on public sleep datasets. Notably,... https://escholarship.org/uc/item/3f60n1xz Fri, 25 Sep 2026 00:00:00 +0000 Wu, Xi Xu, Yiwei Cui, Xinzhong Long, Zhiying https://escholarship.org/uc/item/3dg945t9 Many legal decisions rely on evaluating how a "reasonable person" would have acted, but what defines this standard? Here, we offer an experimental jurisprudence perspective by investigating what subjective features come to mind when reasoning about a reasonable person. In Experiment 1, we examined how people conceptually organize demographic, dispositional, and action features along dimensions of relevance, controllability, and normality. In Experiment 2, we used these dimensions to predict what features shape judgments about reasonableness and outcomes. We found that participants "undid" harmful outcomes by focusing on relevant, abnormal actions; that they constructed a reasonable person from the defendant by preserving normal attributes while changing abnormal, controllable ones; and that they endorsed subjective standards based on features that were relevant and normal, but relatively uncontrollable, consistent with theories of blame. Together, these findings map a template... https://escholarship.org/uc/item/3dg945t9 Fri, 25 Sep 2026 00:00:00 +0000 Wu, Sarah A Zhang, Siying Gerstenberg, Tobias https://escholarship.org/uc/item/39z89698 Learning is difficult to measure because it is not directly observable. Even in formal educational settings it’s hard to tell whether a child might be thinking through a difficult problem or daydreaming. The same is especially true in informal settings like while watching TV at home—it’s difficult to know whether a child is meaningfully engaged with what they are watching or whether they are overstimulated and unable to disengage. Caregivers and educators regularly rely on behavioral cues to infer kids’ learning. Physical engagement, or sitting still, is often treated as a reliable signal of learning. However, stillness during screen-based media is not synonymous with learning in young children and may in fact indicate overstimulation and disrupted learning (Shepherd &amp; Kidd, 2024). We asked 200 adults whether they thought the same child was more attentive and learning more from a video when sitting still compared to when making small fidgets, like twiddling their thumbs or... https://escholarship.org/uc/item/39z89698 Fri, 25 Sep 2026 00:00:00 +0000 Shepherd, Sarah Stolp Kidd, Celeste https://escholarship.org/uc/item/35g1z2c6 Complex relational understanding tasks such as Document-level Relation Extraction require resolving semantic ambiguity amid a quadratic explosion of entity pairs, causing severe class imbalance and false negatives. Although large language models show strong reasoning ability, directly applying them to DocRE is inefficient and prone to hallucination. Inspired by dual-process theories of human cognition, we propose DocRE-Thinker, a hybrid framework integrating fast intuition with controlled deliberation. A fine-tuned backbone serves as System 1, efficiently generating candidate relations while estimating uncertainty. A frozen large language model acts as System 2 and is invoked only for high-uncertainty cases, performing uncertainty arbitration and attribute-based rule induction to justify relations. These symbolic rules are converted into supervision through a text-gradient feedback mechanism, enabling System 1 to internalize analytical reasoning in a rule-aware manner. Experiments... https://escholarship.org/uc/item/35g1z2c6 Fri, 25 Sep 2026 00:00:00 +0000 Wang, Zhen Wang, Yu Zhao, Wen https://escholarship.org/uc/item/2xb8k0cg Whole-brain decoding can test whether semantic category information is localized and shared across people or distributed and individualized. We evaluated Iterated LASSO (iLASSO), a two-stage procedure that iteratively selects predictive voxels with L1-regularized multinomial classifiers and estimates category-wise contributions with ridge-regularized fitting. Using nested cross-validation, we applied iLASSO to two independently collected face/place/object fMRI datasets. In both datasets, iLASSO achieved above-chance held-out accuracy (JLP: N=8, M=83.5%, SD=6.0%, p&lt;.001; Neural Fingerprints: N=35 scans, M=60.7%, SD=13.4%, p&lt;.001), comparable to standard LASSO (JLP: M=86.4%, SD=8.0%, p=.11; Neural Fingerprints: M=60.9%, SD=13.5%, p=.85) while selecting more voxels. Decoded coefficient maps revealed face, place, and object information distributed across all four cortical lobes, extending beyond classical category-selective regions. Many selected voxels showed graded multi-category... https://escholarship.org/uc/item/2xb8k0cg Fri, 25 Sep 2026 00:00:00 +0000 Li, Zihan Colón, Y. Ivette Mukherjee, Kushin Rogers, Timothy T https://escholarship.org/uc/item/2wt4725g What are the neural mechanisms underlying risky decision making? How do the implementation details of these mechanisms impact observable behaviour? Here we hypothesise that non-human primates (NHPs) are re-using circuits for spatial navigation to solve decision making problems. This is motivated by recent findings of grid coding of probability and magnitude in NHP frontal cortex (Bongioanni et al., 2021; Veselic et al., 2025). Grid cells are an optimal code for space (Dorrell et al., 2023; Fiete et al., 2008; Sorscher et al., 2023) but this form of representation is not normatively optimal for risky-decision making. However, this might be a small price for efficiency: by embedding problems in space, NHPs can recycle spatial solutions to quickly solve new problems. First, we show that this spatial framework can replicate classical prospect theoretic findings (Kahneman &amp; Tversky, 1979). Then, we show that the framework also predicts unique biases in the behaviour of NHPs, which... https://escholarship.org/uc/item/2wt4725g Fri, 25 Sep 2026 00:00:00 +0000 Shani, Daniel Sablé-Meyer, Mathias Veselic, Sebastijan Gutierrez, Elena Jensen, Kristopher T. Selvanayagam, Janahan Hunt, Laurence Dayan, Peter Kennerley, Steven Behrens, Timothy https://escholarship.org/uc/item/2vx527s2 Comprehending content in a newly learned language requires interactions between perceptual, semantic, and executive processing systems. Learners whose target language differs from their own in modality (e.g. spoken language users learning to sign) provide a unique opportunity to examine the relative contributions of semantic and perceptual processing. We present data from three studies where hearing, non-signing participants with between a few hours and a few weeks of experience with American Sign Language (ASL) viewed videos of signs during fMRI scanning. Using Representational Similarity Analysis (RSA), we measure contributions of semantic-conceptual, visual perception, and cognitive control regions to processing of semantic and visual information in ASL. Across different learning paradigms, including both cross-sectional and longitudinal approaches, we find that semantic features become more decodable in semantic, visual, and cognitive control regions after learning, while... https://escholarship.org/uc/item/2vx527s2 Fri, 25 Sep 2026 00:00:00 +0000 Hillis, Megan E. Kraemer, David J. M. https://escholarship.org/uc/item/28m3k4g4 The maintenance and manipulation of information in working memory (WM) is fundamental to human and artificial intelligence. Although human WM is famously capacity-limited, Large Language Models (LLMs) preserve inputs in the network’s context window, profoundly reducing capacity limits that depend on maintenance alone. However, work in cognitive science suggests WM limits can also arise from representational interference during manipulation, raising the possibility of strong limits even in LLMs with perfect maintenance. Here, we evaluate 15 frontier LLMs and find models perform near-perfectly on forward digit span (recalling sequences in order) but collapse on backward digit span (recalling sequences in reverse). Moreover, backward span performance is significantly correlated with two measures of fluid intelligence (Raven’s Progressive Matrices and ARC-AGI-1). These findings suggest backward digit span provides a plausible benchmark for the “working” component of working memory... https://escholarship.org/uc/item/28m3k4g4 Fri, 25 Sep 2026 00:00:00 +0000 Diak, Christopher James Nguyen, Khac Nhat Nam Tibbetts, JR Webb, Taylor Frankland, Steven https://escholarship.org/uc/item/27j7f3b7 Semantic memory, the capacity to store and retrieve conceptual knowledge, is central to cognitive science. A key debate concerns how sensory experience shapes conceptual processing and whether semantic representations require sensorimotor simulation. To test this, we trained a Support Vector Machine classifier on natural language features from a Property Listing Task. Recursive Feature Elimination with Cross-Validation selected six optimal features. Univariate analysis revealed no significant differences between congenitally blind and sighted groups for any feature (p &gt; 0.05). Classifier F1-scores did not differ significantly from chance (p = 0.26). However, prediction errors for the six-feature classifier were asymmetric. Accuracy for predicting blind participants (0.31 ± 0.31) did not differ from chance (p = 0.508), whereas accuracy for sighted participants (0.80 ± 0.18) was significantly above chance (p = 0.029). This pattern supports greater semantic heterogeneity in congenitally... https://escholarship.org/uc/item/27j7f3b7 Fri, 25 Sep 2026 00:00:00 +0000 Sena da Silva, Abner Chavez, Rodrigo Lagos Toro Hernandez, Felipe D https://escholarship.org/uc/item/248567jv Vision-language models (VLMs) often perform well on spatial reasoning tasks, but it remains unclear whether their performance is supported by internal representations that track rotation angle. We study this question in a controlled same or mirror 3D object task using LLaVA-1.5-7B as an intervenable computational system. First, the model’s decision margin varies systematically with rotation angle, mainly for rotated rather than mirrored stimuli. Second, we identify an angle-related direction in intermediate hidden representations whose activity covaries with behavior. Third, targeted projection ablations produce progressive flattening of the angle-margin relationship as intervention strength increases. Together, these findings provide mechanistic evidence that angle-sensitive internal geometry contributes to the model’s behavior on this task. The results support a cautious interpretation of structured, intervenable spatial processing rather than a strong claim of human-like mental... https://escholarship.org/uc/item/248567jv Fri, 25 Sep 2026 00:00:00 +0000 Ren, Junzhe Zhang, Jiayuan https://escholarship.org/uc/item/226429wr People are capable of learning with very little data. We argue that this is because people possess an symbolic inductive bias, consistent with the language of thought (LoT) hypothesis (Fodor &amp; Pylyshn, 1988). To evaluate this claim, participants were asked to learn list functions by predicting how to transform an input list of numbers to an output list. We used participants' predictions on trials with no feedback to search for a congruent representation using a LoT model. The model was able to predict participants' heldout responses at above chance levels, even for participants who were unable to learn the function. Furthermore, LLMs and a neural network trained with a LoT inductive bias displayed similar patterns. These findings suggest that people and ML models display signatures of symbolic representation use to accomplish few-shot learning. https://escholarship.org/uc/item/226429wr Fri, 25 Sep 2026 00:00:00 +0000 Marupudi, Vijay Piantadosi, Steven https://escholarship.org/uc/item/2152v33b Evaluating the credibility of a speaker’s statement requires more than assessing factual accuracy; it also involves tracking cues to the speaker’s epistemic commitment to the claim. Grammatical evidentiality, which encodes information sources, may therefore shape how statements are evaluated during language comprehension. Turkish marks the information source through obligatory evidential morphology and offers a unique window into how source information modulates real-time truth judgements. The present study examined the neural dynamics of processing of accurate and inaccurate statements in Turkish by investigating how direct (-DI) and indirect (-mIş) markers modulate language comprehension as measured by ERPs. Participants viewed short event videos and then judged the veracity of written statements describing the events while an EEG was recorded. ERPs time-locked to the evidentially-marked verbs showed differential electrophysiological responses associated with semantic integration... https://escholarship.org/uc/item/2152v33b Fri, 25 Sep 2026 00:00:00 +0000 Kalender, Şeyma Tanis, Selma Berfin Kanero, Junko Aydin, Cagla Arslan, Dr Seçkin https://escholarship.org/uc/item/1wp8r9rf Disorders of consciousness (DoC) are primarily diagnosed using behavioral assessments, which are prone to high misdiagnosis rates. Objective neural markers are therefore needed. This study investigated residual neural responses to naturalistic auditory stimuli in DoC patients to better reflect covert consciousness. Four auditory conditions were presented: natural speech sentences, music, animal sounds, and pure tones as a baseline. Functional near-infrared spectroscopy (fNIRS) was used to assess cortical activation, hemodynamic responses, and functional network alterations during auditory processing. A support vector machine (SVM) classifier was applied to distinguish vegetative state (VS) from minimally conscious state (MCS) patients and to identify the most informative neural features. Complex natural stimuli, particularly speech and music, elicited more specific hemodynamic and network responses than pure tones. Patients with higher consciousness levels showed selectively enhanced... https://escholarship.org/uc/item/1wp8r9rf Fri, 25 Sep 2026 00:00:00 +0000 LUO, Qinqin LU, Shuo XIANG, Xinrong https://escholarship.org/uc/item/1t8957tx The attraction effect—where a decoy option increases preference for a dominating target—is a cornerstone of context-dependent choice, yet it is paradoxically fragile. Sequential accounts propose that context effects depend on which pairwise comparisons are emphasized during deliberation, but existing tests often confound comparison availability with memory/recency. We introduce a Self-directed Expanded Judgment paradigm in which participants repeatedly unblur and judge available option pairs, allowing information search to be observed and constrained. In Experiment 1, we validate the method by replicating a positive attraction effect. In Experiment 2, we causally manipulate comparison availability by disabling specific pairwise links while equalizing cumulative stimulus exposure at the decision stage (top-up control). Consistent with the preregistered hypothesis, the target’s relative share (RSTew) differed reliably between conditions, with the disabled Competitor-Decoy pair producing... https://escholarship.org/uc/item/1t8957tx Fri, 25 Sep 2026 00:00:00 +0000 Rath, Tapas Ranjan Srivastava, Nisheeth Srinivasan, Narayanan https://escholarship.org/uc/item/1h40f61j Sentences like “She will go to France or Spain, or perhaps to Germany or France.” appear formally redundant, yet become acceptable in contexts such as “Mary will go to a philosophy program in France or Spain, or a mathematics program in Germany or France.” While this phenomenon has typically been analyzed using symbolic formal representations, we aim to provide an account grounded in artificial neural mechanisms. We first present new behavioral evidence from humans and large language models demonstrating the robustness of this apparent non-redundancy across contexts. We then show that, in language models, redundancy avoidance arises from two interacting mechanisms: models learn to bind contextually relevant information to repeated lexical items, and Transformer induction heads selectively attend to these context-licensed representations. We argue that this neural explanation sheds light on the mechanisms underlying context-sensitive semantic interpretation, and that it complements... https://escholarship.org/uc/item/1h40f61j Fri, 25 Sep 2026 00:00:00 +0000 Boguraev, Sasha Yao, Qing Mahowald, Kyle https://escholarship.org/uc/item/1gx4x9pp In divided societies, authorities often use punishment to establish shared norms while trying to maintain or signal their legitimacy. When facing a polarized audience, these goals may often be at odds. We extend the Rational Communicative Social Action (RCSA) framework (Radkani et al., 2022) to formally model an authority’s strategic decision-making when using public punitive actions to pursue these goals. We distinguish between a communicative authority who aims to shape the audience’s moral beliefs, and a reputation-aware authority who optimizes the audience’s assessment of their own character. By simulating these agents against polarized audiences, we characterized the tradeoffs and dynamics of strategic punishment when facing diverse audiences with various forms and levels of polarization, which serves to generate systematic predictions for future experiments. https://escholarship.org/uc/item/1gx4x9pp Fri, 25 Sep 2026 00:00:00 +0000 Martinez, Hector Xavier Saxe, Rebecca Radkani, Setayesh https://escholarship.org/uc/item/1dh0632g Similarity and generalization are often assumed to reflect distance in an underlying representational space. In semantic networks, however, distance can be defined by both shortest paths or by multiple indirect pathways. Previous research has shown that non-shortest, higher-order paths influence similarity judgements. Here, we extend these findings by re-analyzing a publicly available dataset to compare shortest-path models with models integrating over discounted higher-order paths. The latter better accounted for similarity judgements; error analyses indicate that this advantage arose from integrating multiple indirect paths rather than relying on shortest connections. These random-walk models implement the same core computation as the successor representation (SR), which has been proposed to support human generalization. Consistently, an SR-based model outperformed alternatives in accounting for inductive generalization judgements from the same dataset. These findings suggest... https://escholarship.org/uc/item/1dh0632g Fri, 25 Sep 2026 00:00:00 +0000 Rubino, Valerio Piazza, Manuela https://escholarship.org/uc/item/16v2s8b5 Cognitive scientists spend much of their professional training learning how to explain empirical research to one another: in talks, papers, posters, grant proposals, and conference discussions. An integral part of our training is the communication and dissemination of scientific ideas at an expert level. Far less attention is given to translating that same work for communities beyond academia. This gap matters because cognitive science produces findings about learning, reasoning, attention, memory, communication, development, decision making, and human-technology interaction—topics that many members of the public already encounter in everyday life. Cognitive research is fascinating, relevant, and often meets urgent needs. Yet when research findings circulate without context, they can be misunderstood, oversimplified, or mistrusted. People may not trust what they cannot understand, and many misunderstandings begin at a basic level: what empirical research is or looks like, what... https://escholarship.org/uc/item/16v2s8b5 Fri, 25 Sep 2026 00:00:00 +0000 Girouard-Hallam, Lauren Viridiano, Marcelo Chomik, Jessica https://escholarship.org/uc/item/0xq211c6 Verbs are pivotal in human language and cognition. Psycholinguistic research has developed explicit, structured accounts of verb semantics, yet these theories have rarely been leveraged to improve models' visual action understanding. Meanwhile, current vision–language models (VLMs) often struggle with action semantics and exhibit unstable, weakly grounded judgments. Therefore, we introduce Verb Semantic Reasoning (VSR), a two-stage pipeline in which a Large Language Model (LLM) converts candidate action descriptions into structured event-semantic representations and a chain of semantic components questions; a VLM answers these questions over videos to select the best-supported action description. Results indicate that human performance was near ceiling, whereas both VLM baselines were substantially lower; VSR consistently improved accuracy across models and narrowed the human–model gap. These results suggest that explicit verb-semantic reasoning can significantly improve the accuracy... https://escholarship.org/uc/item/0xq211c6 Fri, 25 Sep 2026 00:00:00 +0000 Xu, Enjie Wu, Jingyi Tang, Ning Hartshorne, Joshua K Tenenbaum, Joshua B. Gao, Tao https://escholarship.org/uc/item/0rz4d8zf Progress in AI has led to a lively debate about the possibilities of creating and assessing consciousness in AI, building on Putnam’s conjecture that computational functionalism is more plausible than a biological view of consciousness. I discuss Butlin et al.’s (2025) proposal for extracting computational indicators of consciousness from neuroscientific theories and critically evaluate computational functionalism in light of the alternative, biological naturalism. While computational functionalism allows for AI consciousness in principle, it is questionable whether conscious AI systems are practically possible, and whether we are in an epistemic position to judge that a specific AI is conscious. Putative computational indicators extracted from neuroscientific theories of consciousness can never be as well (or even better) supported by empirical evidence than their neurobiological realizers measured in human brains. All potential markers of consciousness that are typically... https://escholarship.org/uc/item/0rz4d8zf Fri, 25 Sep 2026 00:00:00 +0000 Schlicht, Tobias https://escholarship.org/uc/item/0rq9k3vf User studies of AI-mediated communication typically report average effects, while interview-based work remains dominated by thematic narrative or Likert clustering, leaving the structured heterogeneity expressed in post-task interviews unquantified. We applied latent class analysis as a methodological bridge that converts qualitative interview codes into discrete, domain-specific user profiles, enabling design reasoning about for whom, on which dimension, and under what conditions real-time AI scaffolds help or burden. As a validating scenario, 29 non-native English speakers interacted with XPLAIN, a Wizard-of-Oz proactive scaffold in Zoom's sidebar, to bridge gaps in linguistic and cultural knowledge during a collaborative task. Across eight thematic domains, two-class solutions were consistently best-fitting (BIC); class membership was largely independent across domains, indicating multi-dimensional rather than global user types (e.g., longer English immersion was associated... https://escholarship.org/uc/item/0rq9k3vf Fri, 25 Sep 2026 00:00:00 +0000 He, Wanqing Psyche Fussell, Susan R. https://escholarship.org/uc/item/0md843zn Capturing shared cognitive processes across individuals is crucial for cross-subject EEG emotion recognition. Existing studies overlook the subjective experiential component introduced by individual cognitive biases, causing redundant information to be entangled with the extracted emotion-related features. To this end, we propose a cross-subject EEG emotion recognition method named CDDP (Cognitive Debiasing via Disentangled Pre-training). Specifically, during pre-training, cognitive disentanglement and contrastive learning are leveraged to extract subject-invariant intrinsic features (essentially emotion-relevant features) and subject-specific bias features (subjective experiential knowledge) from EEG signals. Meanwhile, a variational autoencoder (VAE) is introduced to generate diverse bias features in the latent space, alleviating overfitting caused by limited source domain EEG data. During fine-tuning, an emotion classifier is jointly trained with the pre-trained subject-invariant... https://escholarship.org/uc/item/0md843zn Fri, 25 Sep 2026 00:00:00 +0000 Chen, Qi Liu, Zhaiyi Wang, Lei Wang, Zihe Zheng, Jian Jin, Bo https://escholarship.org/uc/item/0js3s8dc Many non-native English speakers study mathematics in English, so they must learn both the math and the academic language used to express it. Large language models (LLMs) might help by producing worked solutions while also explaining difficult words and phrases, but it is unclear whether doing both harms mathematical accuracy. We test whether adding inline vocabulary explanations changes solution correctness. Using the CEFR-J vocabulary profile, we identify terms likely to need clarification and evaluate four open LLMs (2.7B–20B parameters) on English math problems in two conditions: standard solving vs. solving with embedded vocabulary support. On MMLU Elementary Mathematics and GSM8K, vocabulary scaffolding consistently reduces accuracy, with drops up to 16.2 percentage points. These findings reveal a trade-off between language assistance and reasoning performance. Educational interfaces may need to balance these goals carefully or separate language support from problem solving... https://escholarship.org/uc/item/0js3s8dc Fri, 25 Sep 2026 00:00:00 +0000 Ehara, Yo https://escholarship.org/uc/item/0h49g1g2 Reading fluency requires rapid coordination of perceptual, linguistic, and executive processes under strict temporal constraints. Rapid Automatized Naming (RAN) is one of the most reliable predictors of reading fluency across languages and developmental stages, yet the mechanisms underlying this relationship remain unclear. One potential source of processing inefficiency involves failures to suppress previously activated representations during sequential processing. Resistance to proactive interference (PI), the ability to prevent no-longer-relevant information from disrupting current task demands, may represent a critical bottleneck constraining performance in both rapid naming and fluent reading. This exploratory study investigates the degree to which RAN and PI resistance overlap in their contribution to reading fluency. One hundred twenty Mexican undergraduate students will complete Spanish-adapted measures of reading fluency, RAN tasks, and a Brown–Peterson task measuring... https://escholarship.org/uc/item/0h49g1g2 Fri, 25 Sep 2026 00:00:00 +0000 Martínez Zenteno, Rebeca Gabriela Falcón, Alberto J. https://escholarship.org/uc/item/0579w66p Three and four-year-olds can readily infer causal relationships from the covariation of interventions and outcomes. However, this kind of understanding has almost always been tested using arbitrary stimuli (e.g., blicket detectors). Thus, the relationship between children’s understanding of covariation data and their understanding of the physical mechanisms that underlie causal relationships is poorly understood. We test children’s understanding of physical mechanisms using a domino setup which can translate common cause, common effect, and causal chain structures into visible, physical arrays where the transmission of force is governed only by contact causality (a principle within young children’s grasp). Using a novel online task battery, we test five different aspects of mechanism understanding and show that children’s understanding of the physical mechanisms underlying causal relationships undergoes a surprisingly protracted development, with three-year-olds performing at... https://escholarship.org/uc/item/0579w66p Fri, 25 Sep 2026 00:00:00 +0000 Vivanco, Vicente Siegel, Max Schulz, Laura https://escholarship.org/uc/item/037571gh Recent research has shown that vision-language pretrain-ing (VLP) models using contrastive learning (Contrastive Language Image Pretraining, CLIP) has a semantic bias to-wards using concrete words during image-text matching. Here we showed that as compared with CLIP, humans at-tended more to abstract words during image-text matching. This difference likely results from CLIP’s difficulty in de-veloping grounded understanding of abstract concepts and capturing contextual dependencies between images and captions through contrastive learning. While CLIP’s caption attention aligned more closely with humans for concrete than for abstract captions, their alignment in im-age attention did not differ between the caption condi-tions, as human image attention was driven primarily by individual differences in explorative or focused perceptual style. Our findings thus revealed important differences in information processing mechanisms between humans and the VLP models, with important implications... https://escholarship.org/uc/item/037571gh Fri, 25 Sep 2026 00:00:00 +0000 Zhang, Jinhan DUAN, Qichun Zhao, Chenyang Chan, Antoni B. Hsiao, Janet https://escholarship.org/uc/item/9sw813q0 A few minutes practice is often more than sufficient for an adult human participant to identify the structure of a problem they have never seen before, and plan a complex sequence of actions that creates a solution. However, it remains less clear whether this distinctive ability for ad-hoc discovery of problem-solving algorithms is itself subject to rapid meta-learning. We developed a novel problem-solving paradigm to examine aspects of this question. Participants in our study faced repeated iterations of an interactive sequential reasoning problem. Over trials, those who faced the hardest version developed increasingly efficient hierarchically-structured strategies that adaptively sequence a subordinate learning algorithm and an action planning policy; those who faced an easier version used simpler action-based strategies that did not involve learning the underlying structure. These results offer experimental evidence for efficient meta-learning of algorithmic concepts in a problem-solving... https://escholarship.org/uc/item/9sw813q0 Thu, 24 Sep 2026 00:00:00 +0000 Yang, Huiwen Alex Ho, Mark Thompson, Bill https://escholarship.org/uc/item/98w881r8 Visual description is a fundamental yet challenging task that reflects how perceptual information is transformed into structured linguistic descriptions. Recent diffusion-based models have shown strong potential for parallel decoding and global semantic modeling. However, existing approaches largely treat generation as an unconstrained process, lacking mechanisms to incorporate structured knowledge, which often results in conceptually implausible or hallucinated descriptions, especially in domain-specific contexts. To investigate how structured knowledge constrains generative diffusion processes, we propose KenDiC, a Knowledge-enhanced Diffusion-based Captioner for visual description. KenDiC integrates a domain-adaptive visual encoder (DAVE) trained via contrastive learning to align perceptual and linguistic representations, and a domain term vocabulary (DTV) that constrains decoding to guide concept selection during generation. To support systematic analysis, we construct II2T-Bench,... https://escholarship.org/uc/item/98w881r8 Thu, 24 Sep 2026 00:00:00 +0000 Liu, Ronghui Deng, Wenfeng Zhou, Nan Liang, Xiaojun Cui, Wei https://escholarship.org/uc/item/96r890gz Prior studies of Second Language Acquisition have identified two forces that shape L2 production: cross-linguistic influence and L2 developmental universals. While past studies have focused on the acquisition of target structures, the distribution of structures in learner language production may still exhibit influence from these two forces. To disentangle the two kinds of impact on the distribution of linguistic structures in natural L2 production, we took a machine learning approach, leveraging over twenty typological and syntactic complexity features from five languages (English, Spanish, Portuguese, Chinese, and Korean). By comparing linguistically informed features that characterize 1) each language based on native data, 2) each language based on L2 production from multiple L1 background, and 3) speakers of different L1 background writing in the same L2, we identified the structural features that capture the developmental universal impact and those that reflect L1 influence.... https://escholarship.org/uc/item/96r890gz Thu, 24 Sep 2026 00:00:00 +0000 Yang, Haiyin Wulff, Stefanie Liu, Zoey https://escholarship.org/uc/item/96d4179z Corpus studies suggest that less predictable words are more likely to be code-switched in bilingual communication, yet behavioral evidence remains limited, partially due to the difficulty of eliciting voluntary code-switches with a controlled linguistic context. In this study, we introduce a production paradigm that reliably elicits Mandarin–English code-switches and allows for precise manipulation of word predictability through numeral classifiers. Using this paradigm, we show converging evidence that both contextual predictability and baseline word frequency influence the likelihood of code-switching. This experimental paradigm provides a valuable tool for future studies of bilingual speech production. https://escholarship.org/uc/item/96d4179z Thu, 24 Sep 2026 00:00:00 +0000 Li, Yanting Scontras, Gregory Futrell, Richard https://escholarship.org/uc/item/8x47k0ft Modern scientific research combines exceptional innovation with increasing rates of collaboration. Creativity in STEM fields thus offers a model system for investigating collective creativity. Much of what we know about STEM innovation and collaboration comes from analyses of published articles, but this ignores the rich, hidden “backstage” of the scientific process — the informal interactions, initial hunches, and abandoned ideas that drive STEM creativity. Here, we investigate this hidden social life of STEM creativity, focusing on how social interaction shapes the generation and abandonment of research ideas. In a survey of PhD-level STEM researchers (N = 150), we find that idea generation and abandonment are associated with different styles and dynamics of social interaction. We introduce a minimal mathematical model of scientific collaboration that can reproduce these empirical patterns. Understanding the informal social interactions that drive scientific progress sheds light... https://escholarship.org/uc/item/8x47k0ft Thu, 24 Sep 2026 00:00:00 +0000 Al haj, Youmna McDonald, Duncan Marghetis, Tyler https://escholarship.org/uc/item/8t8148tp Efficient communication accounts of cross-linguistic variation propose that languages trade-off informativeness and complexity based on communicative need, often measured by usage frequency and operationalized as a free parameter. We propose that this need is not uniform across discursive genres and that the narrative genre applies a stronger pressure for lexical informativeness than non-narrative genres. Using spoken-language data from typologically diverse languages, we analyze the informativeness of lexical fields in narrative and other genres, as operationalized by their lexical richness. We show that frequency in narrative genres predicts lexical richness more strongly than in other genres. This supports the theory that narrative is a privileged communicative context that drives the informativity displayed by different parts of the lexicon. https://escholarship.org/uc/item/8t8148tp Thu, 24 Sep 2026 00:00:00 +0000 Torgue, Jules Mollica, Francis Beekhuizen, Barend https://escholarship.org/uc/item/8s7834hk People increasingly use large language models (LLMs) to explore ideas, gather information, and make sense of the world. In these interactions, users encounter chatbots that are overly agreeable. We argue this sycophancy poses an epistemic risk distinct from hallucination: it distorts belief not by introducing falsehoods but by biasing the evidence users see. A rational analysis shows that a Bayesian agent fed examples sampled from its own hypothesis grows more confident in that hypothesis without moving closer to the truth. We tested this prediction in a modified Wason 2-4-6 rule discovery task where participants (N=557) interacted with AI agents providing different types of feedback. Unmodified LLM behavior suppressed participants’ discovery and inflated their confidence comparably to explicitly sycophantic prompting. By contrast, independent sampling from the true distribution yielded discovery rates five times higher. This paper documents how sycophantic AI distorts belief,... https://escholarship.org/uc/item/8s7834hk Thu, 24 Sep 2026 00:00:00 +0000 Batista, Rafael Griffiths, Tom https://escholarship.org/uc/item/8pb3h1m2 Temporal bisection tasks are widely used to assess interval timing, but standard behavioral indices cannot easily distinguish changes in temporal discrimination from shifts in response bias. In this study, 194 adolescents completed a visual temporal bisection task with two consecutive blocks. Using hierarchical Bayesian psychometric modeling, we decomposed binary duration judgments into temporal sensitivity and decision bias, and examined how these parameters changed during task progression. At the group level, temporal sensitivity declined in the second block, whereas decision bias remained relatively stable. Despite this decline, individual sensitivity estimates showed strong short-term rank-order stability across blocks. Sensitivity changes also varied across individuals and were systematically related to baseline sensitivity, with higher baseline sensitivity associated with smaller declines. Model-derived sensitivity changes corresponded to changes in conventional psychophysical... https://escholarship.org/uc/item/8pb3h1m2 Thu, 24 Sep 2026 00:00:00 +0000 li, Zewei Yuan, Bing Su, Yanjie Zhang, Yisi Zheng, Meihong https://escholarship.org/uc/item/8n6528zw A central puzzle in higher-order cognition is the unidimensional sort bias: the overwhelming tendency of humans to sort examples in a novel domain based on a single feature when a family resemblance (FR) structure is available. A standard unsupervised category construction task with two prototypes and ‘off-by-one’ distortions was investigated using novel task supports. Participants were instructed to find a "very close match" to begin each category and to expand the groups by identifying additional very close matches. This guides participants to employ stimulus generalization: using close proximity in psychological space to extend a consequence. Exemplar theory accounts for category learning as stimulus generalization plus dimensional selective attention – suggesting that unidimensional sorting arises when a single feature captures attentional control. However, instructional supports to prioritize very close matches shifts the evaluation to a global match resulting in a dramatic... https://escholarship.org/uc/item/8n6528zw Thu, 24 Sep 2026 00:00:00 +0000 Kurtz, Kenneth https://escholarship.org/uc/item/85f8d3xh Accurate beliefs about the social world are essential to adaptive social functioning. However, people’s perceptions of the social world are often inaccurate. Across three experiments (N = 991), we examine people’s estimates of the prevalence of ethical and unethical behaviors for men and women. In Study 1, participants reported their own engagement in common unethical behaviors (e.g., cheating on a romantic partner), and then estimated how often men and women engaged in those same behaviors. As predicted, people overestimated the frequency of unethical behavior—an effect which was amplified for men. In Study 2, participants again overestimated the frequency of unethical behaviors, but also underestimated the frequency of ethical behaviors. These effects were again amplified for estimates of men’s behaviors. Study 3 replicated these findings using hypothetical scenarios. Overall, we find consistent evidence of moral cynicism concerning the behaviors of others. People systematically... https://escholarship.org/uc/item/85f8d3xh Thu, 24 Sep 2026 00:00:00 +0000 LeVier, Tori Gantman, Ana Bloom, Paul Wylie, Jordan https://escholarship.org/uc/item/8562453d We investigated how 7–11-year-old children (N = 160) think about and evaluate parental protectiveness. We presented two parent-child dyads who differed in protectiveness: One parent was consistently highly protective (e.g. requiring extensive protective gear for biking); the other consistently permitted risky behaviors (biking without protective gear in a busy parking lot). Participants then made a range of judgments. They judged the more-protective parent as better overall, but also showed context-sensitivity: On a risky field trip, they preferred the more-protective parent as chaperone; but increasingly with age preferred the less-protective parent for a safe field trip. With respect to the parents’ children, participants judged the less-protected child to be popular and mischievous. In contrast, they judged the more-protected child to be brilliant and nice, and would rather have that child come to their house or be class president. These findings reveal that children consider... https://escholarship.org/uc/item/8562453d Thu, 24 Sep 2026 00:00:00 +0000 Tompkins, Rodney Carrillo, Brandon Schachner, Adena Powell, Lindsey J https://escholarship.org/uc/item/83w1f9vd Humans flexibly create tools to solve physical problems, yet the cognitive processes underlying this ability remain unclear. A natural hypothesis is that tool creation involves searching for designs that maximize success probabilities using an internal intuitive physics model. We investigated this by having participants freely construct tools in virtual environments by adding or removing pieces. Despite solving the tasks, participants' choices were only weakly related to estimated success probabilities under a noisy physics engine and were poorly captured by a simple physics-based search model. Crucially, analysis of tool creation trajectories revealed that participants systematically passed through intermediate configurations without treating them as candidate solutions, even when those tools were predicted to succeed. These findings suggest that rather than purely evaluating intermediate designs via physical simulation, human tool creation is guided by a strategy-constrained,... https://escholarship.org/uc/item/83w1f9vd Thu, 24 Sep 2026 00:00:00 +0000 Vivanco, Vicente Smith, Kevin A Allen, Kelsey R https://escholarship.org/uc/item/7xj4791s We present BI*, a Bayesian inference framework that places beliefs over data and data patterns rather than model parameters. Unlike conventional Bayesian approaches that require specifying parametric models, BI* operates directly on the space of possible data-generating distributions—one of which represents the true state of the world from which observed data are sampled. The framework infers the probability that each candidate distribution is the true one, given the observed sample. BI* is coherent, broadly applicable across scientific settings, and provides general methods for model selection. In addition, it can incorporate aspects of science often left to human judgment, such as the effects of experimenter bias and measurement error. We make this general theory computationally feasible through fully Bayesian Gaussian Processes. We illustrate the framework's workings with an artificial dataset, demonstrating how beliefs update from prior to posterior, and how data priors influence... https://escholarship.org/uc/item/7xj4791s Thu, 24 Sep 2026 00:00:00 +0000 Chandramouli, Suyog H Shiffrin, Richard https://escholarship.org/uc/item/7tt067gz NREM sleep has been associated with the stabilization and strengthening of newly acquired memory traces, whereas REM sleep supports the integration of new information into pre-existing memory networks and semantic structures. To examine these processes, participants (N=50, 18–40 years) learned rare Spanish words paired with images-definitions, and then either took a 90-minute nap monitored with polysomnography (PSG) or remained awake. Memory performance was subsequently assessed analyzing word–image–definition associations and measures of semantic integration. Participants who slept preserved word memory across the retention interval, whereas those who remained awake showed a significant decline, indicating a beneficial effect of sleep on word-form consolidation. Memory for definitions showed a similar but weaker pattern, with no significant group differences. No group differences were observed in semantic integration measures, suggesting a ceiling effect or that short naps may... https://escholarship.org/uc/item/7tt067gz Thu, 24 Sep 2026 00:00:00 +0000 Gorosito, María Laura Adba, Julia Ramele, Rodrigo Brusco, Luis Ignacio Kaczer, Laura Forcato, Cecilia https://escholarship.org/uc/item/7rt9j96t The human ability to convey meaning through sound extends beyond spoken language, encompassing the use of tools like musical instruments. People use such instruments to express not only thoughts, but also feelings. How might someone choose to convey such complex meanings with some sounds instead of others? In this study, participants (N=256) used a novel digital musical instrument to produce sound effects to accompany several video clips, such that someone else could match them up later (Referential). Half of these participants were further encouraged to make their sound effects pleasing, towards revealing what distinguishes sounds meant to evoke a particular feeling (Pleasing). The two groups produced different kinds of sound effects, with Pleasing participants favoring sound effects that spanned a narrower range of pitches, while Referential participants created more consistent cross-modal mappings. Taken together, these findings highlight how readily people can convey complex... https://escholarship.org/uc/item/7rt9j96t Thu, 24 Sep 2026 00:00:00 +0000 Caren, Matthew Mukherjee, Kushin Agrawala, Maneesh Fan, Judith E. https://escholarship.org/uc/item/7rr7w4q8 Chinese dyslexia frequently co‑occurs with ADHD, creating complex literacy and attentional difficulties. This project examined (1) cognitive profile types in Chinese dyslexia, (2) their links to attention, and (3) whether ADHD and its treatment alter literacy‑related performance. Study 1 assessed 150 children in Grades 1–3 on Chinese reading and related skills, comparing typical, dyslexia‑only, and dyslexia+ADHD groups. Study 2 related attention scores to cognitive profiles within the dyslexia‑only group. Study 3 tested medication effects on reading and cognitive performance in children with dyslexia+ADHD. Pure dyslexia was marked by rapid automatized naming (RAN), morphological, and orthographic deficits, whereas comorbid dyslexia+ADHD showed a consistent RAN deficit only. Within dyslexia, attentional problems were elevated mainly in children with RAN or phonological deficits. Medication improved attention, RAN, and phonological awareness but not literacy, underscoring the value... https://escholarship.org/uc/item/7rr7w4q8 Thu, 24 Sep 2026 00:00:00 +0000 Tso, Ricky Van-yip Lai, Tony Tai-sum, Lin, Dan Yeung, Kit-yu Kitty https://escholarship.org/uc/item/7k2669f3 Little is known about whether humans show moral concern for harmful actions toward non-human entities like plants, or if how others treat plants influences social preferences. It is also unclear when these possible concerns develop. We address these gaps in two studies: Study 1 with adults (n=153) and Study 2 with 3- to 6-year-old children (n=129). Participants watched a video of a plant restorer who restored a plant but knocked down a bucket, and a plant harmer who restored a bucket but knocked down a plant. Adults preferred the plant restorer over the plant harmer, chose the plant harmer as the bad guy, and evaluated it more positively for restoring the plant than the bucket. Children showed similar patterns—especially those with greater biology knowledge—but did not prefer the plant restorer. These findings help understand early moral reasoning regarding non-human nature and how people value living versus nonliving entities. https://escholarship.org/uc/item/7k2669f3 Thu, 24 Sep 2026 00:00:00 +0000 Pizza, Lizette Thomas, Ashley J https://escholarship.org/uc/item/7d55g4r5 Ambiguity is pervasive in verbal expressions of uncertainty: utterances such as “likely” are interpreted as different probabilities by different individuals, leading to systematic misunderstandings between speakers and hearers. What are the group-level effects of such misunderstandings? Can they lead deliberating collectives to inaccurate beliefs, or pronounced disagreements? To address these questions, we introduce a new Bayesian agentbased model of ambiguous testimony, designed to compare the effects of increasingly ambiguous communication against a baseline of ideal exchange. We find that increasing ambiguity reduces collective accuracy and increases belief dispersion. For mild ambiguity, denser communication networks mitigate these risks. Under strong ambiguity, however, increased communication can backfire: higher network density can exacerbate collective inaccuracy and polarization. https://escholarship.org/uc/item/7d55g4r5 Thu, 24 Sep 2026 00:00:00 +0000 Assaad, Leon https://escholarship.org/uc/item/7cs599hk Habits are difficult to induce and measure in human laboratory tasks, and putative behavioral indices of habits often fail to correlate with self-reported habitual tendencies. We recently developed a behavioral paradigm that reliably reveals overtraining-induced inflexibility within a single ∼1-hour session, using a Habit Index (HI) that contrasts slips of action between extensively and minimally practiced contexts (Oh &amp; Collins, 2025). Here, we test whether individual differences in HI reflect the same habitization process underlying reallife habits, by examining correlations between HI and self-report measures: the COHS Automaticity subscale, the Habitual Tendencies questionnaire, and the Self-Report Behavioural Automaticity Index administered with respect to task behavior. Pilot data (N=106) showed positive associations between overall HI and these measures (ρ ≈.25–.35). We preregistered this correlational study (target N=200) and will present confirmatory results. https://escholarship.org/uc/item/7cs599hk Thu, 24 Sep 2026 00:00:00 +0000 Oh, Sarah Joyner, Keanan Collins, Anne GE https://escholarship.org/uc/item/72s9h312 Formal rules can influence behavior by deterring rule-breaking or broadening knowledge, but may also reveal why a rule was created in the first place. Across three experiments, we investigated whether adults and children use the presence of a written rule to infer past and present rule-breaking. Adults reliably inferred that groups with written rules had more rule-breaking in the past (Study 1) and, in some contexts, expected less rulebreaking in the present (Study 2). By late childhood (ages 8-11), children showed increasingly adult-like inferences when rules were concrete and familiar (Study 2). Younger children (ages 5-7) additionally were able to make this inference (Study 3). Together, these findings suggest that by middle childhood, children treat formal rules as historical evidence about past rule-breaking. These results shed light on the developmental origins of reasoning about formal rules and their informational value. https://escholarship.org/uc/item/72s9h312 Thu, 24 Sep 2026 00:00:00 +0000 Hok, Hannah Saxe, Rebecca Thomas, Ashley J https://escholarship.org/uc/item/6wz2z21x Much research suggests that children learn through play. Yet, play has a reputation as frivolous, and beliefs about play and learning vary across development and sociocultural contexts (Wing, 1995; Bugallo et al, 2024). These differences may reflect varying intuitive theories about how activities promote or prevent learning and shape activity choices. Here, we examine beliefs about how specific activity features impact learning experiences and outcomes. Participants (n=41 adults, ongoing) evaluated 16 math games that systematically varied in both instructional features (e.g., feedback) and playful aesthetics (e.g., audio-visual effects). Results show that games rated as more enjoyable were also rated as better for improving learning. Ongoing work examines how 2nd-4th graders, teachers, and parents respond to specific game features. Our research helps characterize the intuitive theories guiding how people reason about the design of learning experiences, and how such beliefs change... https://escholarship.org/uc/item/6wz2z21x Thu, 24 Sep 2026 00:00:00 +0000 Chu, Junyi Zhang, James Lee, Imogen Haber, Nick Gweon, Hyowon Subramonyam, Hariharan Fan, Judith E. https://escholarship.org/uc/item/6sd7k77j Grounded cognition suggests that motor reactivation impacts memory for manipulable objects. According to the episodic memory literature, motor reactivation improves recall after action during learning episodes. Previously, we found evidence for both grounded and episodic effects, but no interaction despite similarities in the proposed mechanisms (motor reactivation). We hypothesized that an interaction may emerge if testing does not require motor action. Participants learned object-location associations for manipulable (tools) and non-manipulable (animals) stimuli. They either moved (images of) objects to locations, or observed another participant’s movements. At test, participants chose between the correct location and a lure 40 degrees away. We replicated both main effects in recall accuracy, but again found no interaction. We suggest that the separation of semantic and episodic contributions is due to the specificity of the motor action involved in training. Motor information... https://escholarship.org/uc/item/6sd7k77j Thu, 24 Sep 2026 00:00:00 +0000 MacRae, Suesan McRae, Ken Kohler, Stefan https://escholarship.org/uc/item/6qj1w4k9 Collective outcomes often result from complex social dynamics where individuals both contribute directly and also shape each other's contributions. How do we hold people responsible for an outcome when their actions influence others? Here, we examine how an individual's role within a group (whether they can influence others, be influenced by others, or act independently) affects how responsible they are judged to be. Across three experiments spanning both social and physical settings, we find that people systematically assign greater responsibility to those who can influence others. Furthermore, influencers with knowledge of their potential impact were held more responsible than those who were unaware. The relative responsibility of individuals who were influenced by others and who acted independently differed by context. Together, these results show that we hold others responsible by considering not only how their actions directly affect the outcome, but also how they affect... https://escholarship.org/uc/item/6qj1w4k9 Thu, 24 Sep 2026 00:00:00 +0000 Hu, Chuqi Gerstenberg, Tobias Wu, Sarah A https://escholarship.org/uc/item/6mg1s64s We present evidence that explicitly explaining holistic classification choices may promote inverted mental models. That is, people believe their choices are based on observed properties but, paradoxically, their choices drive illusory observations. Participants viewed images, some generated by a human artist and others by Google Gemini. They answered questions about each, sorted them into human- or AI-generated categories, then described how they made their choices. The vast majority provided explicit explanations for how they sorted the images. Critically, many gave similar explanations (e.g., perfection indicates AI), even when incorrectly identifying human-created work as AI and vice-versa. In addition, dimensions offered as diagnostic tended to be more abstract when associated with AI (e.g., “artificialness”) and more concrete (e.g., “attention to detail”) for human-generated attributions. We propose a psychological mechanism for how these mental models may become inverted.... https://escholarship.org/uc/item/6mg1s64s Thu, 24 Sep 2026 00:00:00 +0000 Bosch, David A. Parekh, Hetvi Ishizaki, Yuhri zheng, Yayun Kim, Edward S Patel, Neerali https://escholarship.org/uc/item/6kg9892g Iconicity is a central notion in cognitive science, attracting growing attention from perspectives on language acquisition, processing, and language evolution. While previous research has emphasized the subjective nature of iconicity, quantitative analyses of iconicity ratings have relied primarily on mean values, treating iconicity as a stable lexical property, and have often been constrained by substantial data loss when integrating multiple lexical and psycholinguistic datasets. The present study addresses these limitations by analyzing both mean iconicity ratings and inter-rater variability across 14,764 English words, combining a conservative complete-case analysis with complementary analyses based on imputed predictors to increase lexical coverage. For example, in the complete-case analysis, earlier-acquired and etymologically imitative words tended to receive higher mean iconicity ratings, whereas rating variability was greater for earlier-acquired and more familiar words.... https://escholarship.org/uc/item/6kg9892g Thu, 24 Sep 2026 00:00:00 +0000 Iida, Hinano Hori, Ryo https://escholarship.org/uc/item/6ds1p9rf How do people evaluate the trustworthiness of advice received from other people or AI systems? One signal for trustworthiness is thinking time: the amount of time the advice-giver spent thinking about what advice to give. Different factors may influence inferences based on thinking time: greater thinking time may reflect a lack of knowledge or confidence. Conversely, it may reflect higher-quality advice resulting from more thorough deliberation. We study participants' judgments of trustworthiness based on thinking time by presenting them with pairs of hypothetical human or AI advisors who spent more or less time thinking about their decision. Across most domains we tested, participants preferred longer deliberation in human advisors, but did not show such a preference with AI. These results provide preliminary evidence that people can make sophisticated judgments about advice quality by integrating knowledge about the advice giver and the time it took them to generate their advice. https://escholarship.org/uc/item/6ds1p9rf Thu, 24 Sep 2026 00:00:00 +0000 Malaviya, Maya Srinivasan, Divya Collins, Katherine M Sucholutsky, Ilia Ho, Mark https://escholarship.org/uc/item/6cz6h0t3 Recent work has found that contemporary language models such as transformers can become so good at next-word prediction that the probabilities they calculate become worse for predicting naturalistic reading time. In this paper, we propose that this can be explained by reading time being shaped by simple n-gram statistics rather than the more complex statistics learned by state-of-the-art transformer language models. We demonstrate that the neural language models whose predictions are most correlated with n-gram probability are also those that calculate probabilities that are the most correlated with eye-tracking-based metrics of reading time on naturalistic text. https://escholarship.org/uc/item/6cz6h0t3 Thu, 24 Sep 2026 00:00:00 +0000 Michaelov, James A. Levy, Roger P https://escholarship.org/uc/item/68s7b1j2 Humans are remarkably adaptive instructors who adjust advice based on their estimations about a learner’s prior knowledge and current goals. Inspired by prior work in rational pedagogy, we model teachers that reason about how learners will update their beliefs when given different examples, and thereby select examples that minimize expected learner error. We demonstrate that Bayesian non-parametric approaches can characterize teaching strategies in continuous domains, where traditional rational teaching models are intractable. We compare human teaching choices against our model and a variety of heuristics. Our model explains significant variance in human choices beyond a mixture of heuristics, and provides insight insight into how teachers formulate pedagogical guidance in computationally tractable ways, even in vast spaces. https://escholarship.org/uc/item/68s7b1j2 Thu, 24 Sep 2026 00:00:00 +0000 Malaviya, Maya Ho, Mark