In this essay I try to articulate some thoughts about how meaning can arise in symbolic systems without external referents. The core idea: "A mathematical object is what it does inside a system of relations and operations"

venpopov.com/posts/2026/s...

Synthesizing scientific literature with retrieval-augmented language models

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

I met @russpoldrack.org after this effort and before I knew this particular story. I was blown away when I first read about it. Some scientific efforts “expand the domain of the understandable”. This was one of them. So great to see it profiled by @thetransmitter.bsky.social!

A unifying account of replay as context-driven memory reactivation A context-driven memory model simulates a wide range of characteristics of waking and sleeping hippocampal replay, providing a new account of how and why replay occurs.

Really thrilled that this paper led by @neurozz.bsky.social is now published in its final version in @elife.bsky.social!!

This is a memory-focused (as opposed to RL-focused) account of the detailed characteristics of forward and backward awake and sleep replay!

elifesciences.org/articles/99931

Be wary of assumptions in impossibility arguments A proof is only as good as its assumptions

When are impossibility proofs misleading? In infinitefaculty.substack.com/p/be-wary-of..., I discuss a common issue I see: proofs that are logically valid, but where the underlying assumptions are unjustified. I discuss ‘proofs’ that cognition cannot be tractably learned, and that LMs are 1/

I am happy to share that our preprint “𝗪𝗼𝗿𝗸𝗶𝗻𝗴 𝘄𝗶𝘁𝗵 𝗖𝗶𝗿𝗰𝘂𝗹𝗮𝗿 𝗗𝗮𝘁𝗮: 𝗔 𝗧𝘂𝘁𝗼𝗿𝗶𝗮𝗹 𝗳𝗼𝗿 𝗖𝗼𝗴𝗻𝗶𝘁𝗶𝘃𝗲 𝗮𝗻𝗱 𝗕𝗲𝗵𝗮𝘃𝗶𝗼𝗿𝗮𝗹 𝗥𝗲𝘀𝗲𝗮𝗿𝗰𝗵” is now out.

Huge thanks to @bayslab.org, Julie de Falco, Zahara, @cjungerius.bsky.social, @ivntmc.bsky.social, Adam, and Xiaolu for the lovely collaboration.

doi.org/10.31234/osf...

Infinite hidden Markov models can dissect the complexities of learning - Nature Neuroscience Bruijns et al. present a modeling tool that enables the tracking of learning dynamics across subjects to reveal how behaviors emerge and adapt. Applying the tool to a decision-making task in mice unco...

On a more positive note, this NN is worth a read. It takes a similar approach to Ashwood, Calhoun etc to explore diff behavioral states using HMM, but here using a hierarchical Dirichlet process to infer number of states www.nature.com/articles/s41...

New paper for #neurips2025!

AI models adjust millions of internal settings to get better at a task. But how are these adjustments determined? For decades, we've mostly figured this out through trial & error.

We took a different approach...🧵 (1/6)

🔗 openreview.net/forum?id=oMi...

New preprint 🎉 Psych constructs are complex. Symptoms overlap, people rarely fit neat categories, and patterns are non-linear. Most methods compromise this richness. Self-Organising Maps don't. We provide a step-by-step tutorial with annotated R code to make them accessible.

doi.org/10.31234/osf...

This was a great talk and project.

Short thread on today's HotFresh SJDM paper: bsky.app/profile/maxm...

Transitions in dynamical regime and neural mode during perceptual decisions - Nature Simultaneous recordings were made of hundreds of neurons in the rat frontal cortex and striatum, showing that decision commitment involves a rapid, coordinated transition in dynamical regime and neura...

How does the brain decide? 🧠

Our new @nature.com paper shows that neural activity switches from an 'evidence gathering' to a 'commitment' state at a precise moment we call nTc.

After nTc, new evidence is ignored, revealing a neural marker for the instant when the mind is made up.

rdcu.be/eGUrv

Inference in (neuro)cognitive systems | 9 | Neurocognitive Foundations Cognitive scientists ascribe inferential processes to (neuro)cognitive systems to explain many of their capacities. Since these ascriptions have different

Inferential theories are on the rise in cognitive science. But what does it mean to infer? Check out our take on inference across a variety of (neuro)cognitive systems.

Distorted learning from local metacognition supports transdiagnostic underconfidence - Nature Communications Individuals with symptoms of anxiety and depression exhibit persistent underconfidence. Here, the authors show that distortions in learning from local metacognition can explain how underconfidence is ...

Excited to see this now out in the world!

We identify a computational basis for how persistent underconfidence is maintained in the face of intact performance, finding that it is grounded in impaired updating of global self-beliefs from local metacognition

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

1/N

Beyond divisive normalization: Scalable feed-forward networks for multisensory integration across reference frames The integration of multiple sensory inputs is essential for human perception and action in uncertain environments. This process includes reference frame transformations as different sensory signals ar...

Our multi-sensory integration Neuro-AI paper is now published in J Neurosci

www.jneurosci.org/content/earl...

Super excited to share that Ven Popov & me published a new release of the #bmm R package: venpopov.github.io/bmm/

We have added the Memory Measurement Model for categorical #workingmemory tasks to the package!

Apart from that there are some minor fixes to already implemented models and functions.

Multiple brain activation patterns for the same perceptual decision-making task - Nature Communications Here, the authors show the brain uses multiple activation patterns to perform the same task. Even the default mode network, which is often inactive during focus, plays a role.

𝗖𝗮𝗻 𝘁𝗵𝗲 𝗯𝗿𝗮𝗶𝗻 𝗿𝗲𝘀𝗽𝗼𝗻𝗱 𝗶𝗻 𝗱𝗶𝗳𝗳𝗲𝗿𝗲𝗻𝘁 𝘄𝗮𝘆𝘀 𝘁𝗼 𝘁𝗵𝗲 𝘀𝗮𝗺𝗲 𝘁𝗮𝘀𝗸/𝗰𝗼𝗻𝗱𝗶𝘁𝗶𝗼𝗻𝘀?
Some papers suggesting that it can.
#neuroskyence
www.nature.com/articles/s41...

Are you a graduate student in #Ukraine interested in machine learning and neuroscience? My research lab at #UofT is now accepting applications for remote thesis supervision.

(1/3) #neuroAI #compneuro @vectorinstitute.ai @uoft.bsky.social @uoftcompsci.bsky.social @uhn.ca

Math and biology meet in the cerebellum There can be surprising differences between what neurons do and what neurons cause

"It is thus as if nature smiled and gave the cerebellum-studying neuroscientist the ideal tool for testing null space hypotheses..."

www.science.org/doi/10.1126/...

New work by Sudeep Bhatia @sdpbht.bsky.social, Simon van Baal @svanbaal.bsky.social, Feiyi Wang, and @lukaszwalasek.bsky.social Lukasz Walasek -- now out in Proceedings of the National Academy of Sciences!

www.pnas.org/doi/10.1073/...

Hey #AcademicSky #PsySky. All is ready for the start of the PG Research Day @warwickpsych.bsky.social @uni-of-warwick.bsky.social we’ll be covering the event live from 10am UK time

🚀🚀 Very excited about this new preprint with @yunxiao-li.bsky.social and @asanborn.bsky.social!

Months ago we released the samplr package on CRAN (helps you use sampling algorithms + cogn. models for human data). Here we explain the theoretical background and show how to use the pkg

osf.io/ax8hm

And a poster with collaborators from South University of Science and Technology and University of Birmingham Chen Wei, Jiachen Zou and Quanying Liu where we apply their new AI model (CoCoG arxiv.org/pdf/2404.16482) to generate novel real-world stimuli for visual memory reproduction tasks.

A poster with @timbrady.bsky.social on a model that provides a unifying framework for generalizing from temporal to strength (repetition) manipulation effects on visual memory.

Excited to share some work presented at @vssmtg.bsky.social this year #VSS2025. A great talk by Anxin Miao on a model codeveloped with @timbrady.bsky.social that bridges Bayesian and representational theories of memory, and makes parameter free predictions of visual memory biases.

The Brady Lab @timbrady.bsky.social will be at #VSS2025 @vssmtg.bsky.social this year! Here's a thread with some of the cool work we're coming to share:

On Sat., 05/17 at 2:30PM in Visual Memory: General, former undergrad Anxin Miao will tell us how to predict gist biases in memory APRIORI through independent measurements of memory biases at the level of items and ensembles. CC: @mmrobinson93.bsky.social

Extreme-Value Signal Detection Theory for RecognitionMemory: The Parametric Road Not Taken Signal Detection Theory has long served as a cornerstone of psychological research, particularly in recognition memory. Yet its conventional application hinges almost exclusively on the Gaussian…

Honey, we fixed Signal Detection Theory (SDT)! In this preprint, Constantin Meyer-Grant, David Kellen, Sam Harding, and I critically evaluate the (unequal-variance) Gaussian SDT model in recognition memory and pursue the Gumbel-min model as a principled alternative: doi.org/10.31234/osf...
🧵

If you’re into absurd movies check out my brothers deeply uncomfortable but awesome (I think) short