Cognition

Duration reproduction under memory pressure: Modeling the roles of visual memory set size in duration encoding and reproduction Duration estimates are systematically biased toward the mean of recently sampled intervals, a central-tendency effect typically attributed to the inte…

Ever feel like time flies when you're juggling too much? It's not just a feeling. Our new study shows that holding more items in memory literally warps how your brain encodes and reproduces short durations. ⌛️🧠

www.sciencedirect.com/science/arti...

A dynamic affective surprise signal influences episodic memory Enhanced memory for salient events can be driven by affective responses and by surprising outcomes. Recent work has suggested that deviations in expec…

"A dynamic affective surprise signal influences episodic memory"

www.sciencedirect.com/science/arti...

This work integrates affective and predictive accounts of memory enhancements and suggests that affective surprise may capture dynamic affective prediction errors that shape memory.

@rohini-kumar.bsky.social
@tejassavalia.bsky.social
davidclewett.bsky.social
@aliocohen.bsky.social

Changes in our feelings, or "affective surprise," may act as a learning signal that influences what we remember. Large magnitude deviations in experienced valence during encoding relate to better long-term associative memory.

New paper exploring the flexibility of control mechanisms in bilinguals is out in @cognitionjournal.bsky.social l! w/ @kalinkatimmer.bsky.social, Jakub Szewczyk, and Zofia Wodniecka.

We ask if language context can affect language control in bilinguals.

authors.elsevier.com/a/1meRF2Hx32...
👇🧵

These findings show that statistical learning is sometimes grounded in the statistics of agent-environment interactions, not in the statistics of the world per se. They align with theories linking attention to action. Our study also highlights the importance of dynamic experimental setups.

Across experiments, we found that implicit distractor-location learning is viewer dependent when embedded in active behavior. That is, spatial inhibition cannot be abstracted from the agent moving through a 3D world and how they can suppress sampling the world from their perspective.

We showed this in 3 experiments with a novel setup: participants performed an additional singleton task projected on a table top, while switching standing position. This way, we could manipulate whether a high probably distractor location was fixed in the world or with respect to their viewpoint.

It is typically thought that distractor inhibition entails inhibiting the distractor location. Yet, here we show that distractor inhibition takes into account how one can suppress attentional sampling in space from their viewpoint to prevent distraction. It operates egocentrically.

"Grounding distractor inhibition in action control: Implicit distractor-location learning is viewer dependent"

📢New paper from: Litian Chen, Freek van Ede, Chris Jungerius, Heleen A. Slagter

www.sciencedirect.com/science/arti...

Context-dependent effects of branches in decisions under risk This paper investigates how the number of branches in a prospect influences decision makers' preferences. I propose that individuals may use differenc…

"Context-dependent effects of branches in decisions under risk"

📢New paper from: Ioannis Evangelidis

www.sciencedirect.com/science/arti...

Design/ethics takeaway: branching can nudge choices without improving expected outcomes. If you want better decisions, highlight expected value (or expected lives saved), not just the number of pathways.

Practical implication: people may overweight “number of chances” vs value. Financial products or policies can be made more attractive by splitting outcomes—especially when compared against another risky option.

Theory: branch-splitting isn’t a fixed property of a prospect. Models that treat branching as context-invariant (e.g., classic configural-weight approaches) miss that branch effects depend on what else is on the menu.

Generalizes beyond money: similar branching effects appear in policy-style choices about lives saved. Branch structure can shift preference even when expected lives saved is lower.

Losses mirror: with probabilistic losses, more “pathways to losing” reduces choice share. More branches helps for gains (in joint risky choice) but hurts for losses.

Context test 2: evaluate the target alone and branching reverses—people become LESS willing to play the more-branched gamble. In isolation, absolute value/EV-type thinking seems to dominate.

Context test 1: pair the risky target with a sure gain and the branching effect vanishes. With no competing risky option, “more pathways” no longer works as a justification.

Moderator: similarity matters. The branching boost is stronger when split outcomes are close in magnitude; large dispersion weakens the effect (likely because perceived risk dominates “pathways” reasoning).

Boundary condition: the effect requires explicit branches. If you only state a payoff range (without listing distinct outcomes + probabilities), the “more pathways” advantage disappears—and can even backfire.

Diminishing sensitivity: going from 1→2 gain branches increases attractiveness, but adding more branches yields little/no additional lift. Extra branching doesn’t keep paying off.

In joint choice between risky gains, the option with more “pathways” to win attracts higher preference—even when expected value (and CPT value) is lower. Branch count becomes a reason or justification for choice.

New paper: the effect of adding branches in risky prospects is context-dependent. More branches boost choice when comparing two risky gains—but not when a sure option is present, and the effect reverses in separate evaluation.

Is a linguistic model needed to build abstract event representations? A central question in cognitive development is whether language simply expresses pre-existing event concepts or plays a critical role in their constru…

"Is a linguistic model needed to build abstract event representations?"

📢New paper from: Irene Canudas-Grabolosa, Madeline Quam, Marie Coppola, Jesse Snedeker, & Annemarie Kocab

www.sciencedirect.com/science/arti...

These findings challenge the claim that learning a language is necessary to represent relational concepts, such as those involving asymmetrical roles between participants.

Crucially, there were no differences between groups or between one- and two-participant events, showing that role binding is not uniquely supported by external linguistic input.

Homesigners accurately generalized two-participant events, consistently mapping the correct kinds to agent and patient roles—performing just as well as English-speaking 5-year-olds.

We tested this claim by studying adult homesigners—deaf individuals with no exposure to a conventional language—using a nonverbal imitation task involving one- and two-participant events.

Does language merely express pre-existing event concepts, or does it provide the tools needed to build them?

Some accounts argue that abstract two-participant events (who does what to whom) can only be learned through acquiring a language, by mastering transitive syntax.

Subject islands do not reduce to construction-specific discourse function A central question about our shared capacity for language is how it is integrated with other cognitive systems. One important debate focuses on the ex…

"Subject islands do not reduce to construction-specific discourse function"

📢New paper from: Mandy Cartner, Matthew Kogan, Nikolas Webster, Matt Wagers & Ivy Sichel

www.sciencedirect.com/science/arti...