Samuel Recht

Pleased to say "Space, Time, and Memory", an academic book by Oxford University Press edited by the inimitable Lynn Nadel & Sara Aronowitz is now out.
I contributed a chapter, "Memory and Planning in Brains and Machines".
You can download the entire book for free:
library.oapen.org/bitstream/ha...

Metacognition and Uncertainty Communication in Humans and Large Language Models Metacognition, the capacity to monitor and evaluate one's own knowledge and performance, is foundational to human decision-making, learning, and communication. As large language models (LLMs) become i...

like a metacognitive evaluation? in my understanding metacog is at best weird, but more likely just not even present in most if not all LLMs. they don't do anything like "decide when more context is needed" anything like we do.

not an LLM expert but i do know abt metacog!

arxiv.org/abs/2504.14045

Human-like object concept representations emerge naturally in multimodal large language models - Nature Machine Intelligence Multimodal large language models are shown to develop object concept representations similar to those of humans. These representations closely align with neural activity in brain regions involved in o...

Great work by Changde Du from Huiguang He's lab at the Chinese Academy of Sciences. How similar are visual and conceptual representations in (multimodal) large language models to those found in humans? It turns out quite similar!
www.nature.com/articles/s42...

We have a PhD position for an upcoming project in metacognition research in Potsdam / Berlin. A great opportunity for those interested in cognitive modeling of confidence and EEG.

More information at coconeuro.github.io/phd2025

Kindly share this opportunity with potential candidates - Thanks!

Dissociable habits of response preparation versus response initiation - Nature Human Behaviour Du and Haith show that behaviour can become habitual in two different ways, involving response initiation and response preparation.

www.nature.com/articles/s41...

Harvard today, your institution tomorrow.

It's all part of the Project 2025 plan to destroy high education in America.

All leading US universities depend on federal funding and tuition dollars from international students. They severely curtailed the former. Now they're eliminating the latter.

Very happy this is finally out 🎉. In this paper, we show that conf. reports are biased by a visual illusion, an asymmetrical base rate and a payoff scheme. Crucially, we show that only the visual illusion affects subj. experience, suggesting decision bias leaks onto confidence reports.

Communication of perceptual predictions from the hippocampus to the deep layers of the parahippocampal cortex High-resolution neuroimaging reveals stimulus-specific predictions sent from hippocampus to the neocortex during perception.

Our study using layer fMRI to study the direction of communication between the hippocampus and cortex during perceptual predictions is finally out in Science Advances! Predicted-but-omitted shapes are represented in CA2/3 and correlate specifically with deep layers of PHC, suggesting feedback. 🧠🟦

a screenshot of the preprint first page, with the text: Between repulsion and attraction in serial biases: Replication of Chen & Bae, 2024 Juni B. Akselberg, Sara B. Cardona, Mikkel Dybvad, Lise Martine Karlstad, Malin Langemyr, Ingrid A. Mellingsæter-Jokic, Mats K. K. Moe, Amalie C. Solvang, and Andrey Chetverikov 1 Faculty of Psychology, University of Bergen 2 Department of Psychosocial Science, Faculty of Psychology, University of Bergen Abstract What you see depends on what you have seen before, and commonly your perception is drawn toward the past. Such attractive biases, known as serial dependence, are well established for many visual features. Interestingly, Chen and Bae (2024, Cognition) recently reported a repulsive serial bias in a pointing direction estimation task that switched to an attractive one in the presence of a distracting task. At the same time, an analysis of response trajectories revealed a repulsive bias during response execution, irrespective of the condition. These surprising findings prompted us to attempt a replication. We confirmed the main findings of Chen and Bae. However, we also demonstrated that the overall direction and magnitude of the bias are relatively stable for a given observer, regardless of the condition. Furthermore, we found that already the very first moment in the response trajectory differed between conditions, showing a predominantly attractive bias for trials that ended with attraction. The results confirm the robustness of the original findings and pose a challenge for a simple Bayesian model of serial dependence, highlighting the need for computational models that can explain both attractive and repulsive biases. Keywords: serial dependence, bias, direction estimation, orientation, visual working memory, Stroop, attraction, repulsion, mouse tracking, stimulus history effect, perceptual decision-making, replication

New preprint! Serial dependence is assumed to be attractive, but some studies consistently show repulsion. We tried to replicate a surprising repulsive serial bias that switches to an attractive one when people get distracted during the memory maintenance. It worked! osf.io/preprints/ps... 🧵

Alpha power indexes working memory load for durations Timing, that is estimating, comparing, or remembering how long events last, requires the temporary storage of durations. How durations are stored in working memory is unknown, despite the widely held ...

Episode II of how are durations stored in working memory:
Besides replicating our previous findings, we find that
alpha power reflects a universal signature of WM load and mediates recall precision, even for abstract information like duration
www.biorxiv.org/content/10.1...
🔽 co-authors below

Great to see our research, led by @johannakuci.bsky.social, covered in The Transmitter. One of the more surprising findings to have come out of the lab.

Dynamic modulation of confidence based on the metacognitive skills of collaborators! Now out in @cognitionjournal.bsky.social New work from
@felixhermans.bsky.social Simon Gaia @majafr.bsky.social
and me! Paper here: pubmed.ncbi.nlm.nih.gov/40262423/ Details ↓↓↓

I'm super excited that our new preprint titled "Intrinsic rewards guide visual resource allocation via reinforcement learning" with Rodrigo Raimundo and Paul Bays (@bayslab.org) is finally out! www.biorxiv.org/content/10.1...
1/n
#visionscience #compneurosky #neuroskyence

relationship between sensory evidence and confidence evidence

If you work with perceptual confidence judgments, you may be interested in our CNCB model of confidence ratings. Joint work with Vincent de Gardelle.

Uncorrected proofs here:
dx.plos.org/10.1371/jour...

(🧵 1/4)

The tilt illusion arises from an efficient reallocation of neural coding resources at the contextual boundary | PNAS The tilt illusion—a bias in the perceived orientation of a center stimulus induced by an oriented surround—illustrates how context shapes visual pe...

Very happy to see our work finally in print!
www.pnas.org/doi/10.1073/...

TLDR: Tilt illusion is not a bug, but a feature of a well-designed visual system that maximizes information capacity adaptively based on spatial context. (1/6)

Attention induced perceptual traveling waves in binocular rivalry Cortical traveling waves -smooth changes of phase over time across the cortical surface- have been proposed to modulate perception periodically as they travel through retinotopic cortex. Yet, little i...

Very proud of João VX Cardoso from the #DugueLab for the first preprint of his PhD on a computational model of traveling waves. Work in collaboration with David J. Heeger and Hsin-Hung Li.
@erc.europa.eu

www.biorxiv.org/content/10.1...

a close up of a man wearing glasses and a hat with abc family on the bottom ALT: a close up of a man wearing glasses and a hat with abc family on the bottom

Hey Bluesky! I’m excited to share my new preprint with @matanmazor.bsky.social and @ruvidar.bsky.social, where we show that surface-level questionnaire filling behaviours drive correlations between mental health and metacognition (confidence ratings). osf.io/preprints/ps... [1/14]

The title of the paper, "Beliefs about Perception Shape Perceptual Inference: An Ideal Observer Model of Detection", with the author names, and Fig. 1: a schematic of an inverse optics account of vision.

Our "I would have seen it if it were there" paper — a collaboration with @ranimo.bsky.social and @clarepress.bsky.social — is now out in Psych. Review.

There’s a lot in this paper, but here are what I see as the 3 main takeaways:

psycnet.apa.org/fulltext/202...

Going ‘meta’: a systematic review of metacognition and functional neurological disorder Sadnicka et al., overview experimental data that has examined metacognition in patients with functional neurological disorders (FND). Interestingly, most s

Lovely/comprehensive review of metacognition and functional neurological disorder; a condition that demands a metacognitive interpretation: doi.org/10.1093/brai...

The emerging picture aligns with findings in psychopathology more broadly—metacognitive performance is often equivalent to controls.

Metacognitive confidence and affect – two sides of the same coin? Decision confidence is a prototypical metacognitive representation that is thought to approximate the probability that a decision is correct. The perception of being correct has also been associate...

Why does it feel good to be sure? In new work @voalan.bsky.social @andero.bsky.social and I bridge confidence and affect. In perceptual DM, we found a remarkably consistent association between both measures! ↓↓↓

link: www.tandfonline.com/doi/full/10....
#PsychSciSky #Neuroscience #Neuroskyence

A comprehensive assessment of current methods for measuring metacognition - Nature Communications Measuring metacognitive ability is one of the enduring challenges in cognitive science. The current paper develops formal tests of the quality of different measures and assesses how current metrics pe...

My article "A comprehensive assessment of current methods for measuring metacognition" is finally out in Nature Communications 🎉 If you work on metacognition and think you know the psychometric properties of your favorite measure, you may be surprised.

www.nature.com/articles/s41...

Thanks a lot, Marlene! 🙂

How aggregated opinions shape beliefs

Review by Kerem Oktar & Tania Lombrozo

Web: go.nature.com/40lJX56
PDF: rdcu.be/d5tor

Please, spread!

🧠 Competitive PhD funding is available for neuroscience projects in the Parisian region. Consider your options 👇

dim-cbrains.fr/en/phd-progr...

@c-brains.bsky.social

How can sharing our uncertainty with others alter our confidence when we're alone?

Delighted to share the lab's new paper, with Einar Andreassen and @cdfrith.bsky.social. Particularly pleased as it's Einar's first!

Link: osf.io/preprints/ps...

🧵👇

#PsychSciSky
#neuroskyence

My first paper out 🎉 Huge thanks to project supervisor & lead author @sam.re and my collaborators!

Humans are curious not just about the world, but also about their own minds. In our new paper, we describe a specific form of curiosity in which people strategically seek information not only about their decisions but also about the accuracy of their self-evaluations. psycnet.apa.org/fulltext/202...

Meta-analysis (n=52 studies, 195 effect sizes) confirms the incubation effect: taking breaks away from a problem leads to more creative solutions (d=0.39)

"The Illusion-Illusion: Vision Language Models See
Illusions Where There are None"

osf.io/preprints/ps...

(this is a more systematic examination of the thing I was looking at a few days ago)

Check out @bkhmsi.bsky.social's summary of our new paper:

We identify "language network" units in LLMs using neuroscience approaches and show that ablating these units (but not random ones) drastically impair LLM language performance--moreover, these units better align with human brain data.