Brokoslaw Laschowski

Compact deep neural network models of the visual cortex - Nature Parsimonious deep neural network models can be used for prediction of visual neuron responses.

Nature research paper: Compact deep neural network models of the visual cortex

go.nature.com/3OKRXZU

I am totally pumped about this new work . "Task-trained RNNs" are a powerful and influential framework in neuroscience, but have lacked a firm theoretical footing. This work provides one, and makes direct contact with the classical theory of random RNNs:
www.biorxiv.org/content/10.6...

Great work @cpehlevan.bsky.social and team

Building neuroscience-inspired AI. Follow @compneurolab.bsky.social for updates.

#neuroAI #compneuro @utoronto.ca

Vectorized instructive signals in cortical dendrites - Nature Mice learning a neurofeedback brain–computer interface task show neuron-specific teaching signals in cortical dendrites, consistent with a vectorized solution for credit assignment in the brain.

This paper on how the brain may do gradient descent is very cool: www.nature.com/articles/s41...

Can we predict a thought before it happens?

To know what one neuron will do next, you have to know what the entire brain is doing right now.

In our latest @kempnerinstitute.bsky.social Deeper Learning blog, @duranrin.bsky.social introduces POCO, a tool paving the way for adaptive neurotechnology.

Interesting

Congrats @dlevenstein.bsky.social

U of T Engineering News Brains, minds & machines: A new algorithm for decoding intelligence [photo of Laschowski in a lab with a machine and a whiteboard with formulas]

🧠 Imagine being able to control machines by thinking.

@drlaschowski.bsky.social (MIE) and his Computational Neuroscience Lab are working to make it possible. They've developed a new algorithm that could make brain decoding more accurate and efficient.

Read the story: uofteng.ca/25h7c8

Our paper is out in @natneuro.nature.com!

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

We develop a geometric theory of how neural populations support generalization across many tasks.

@zuckermanbrain.bsky.social
@flatironinstitute.org
@kempnerinstitute.bsky.social

1/14

If you work at the intersection of computational neuroscience and machine learning, consider applying for this postdoc position (January 2027 start date):
academicpositions.harvard.edu/postings/15868
An opportunity to work with a great group of people across Harvard, MIT, and UC Berkeley.

Nice work!

Hybrid neural–cognitive models reveal how memory shapes human reward learning Nature Human Behaviour, Published online: 05 February 2026; doi:10.1038/s41562-025-02324-0Using artificial neural networks applied to human data, Eckstein et al. show that good models of reinforcement learning require memory components that track representations of the past.

Hybrid neural–cognitive models reveal how memory shapes human reward learning

Forecasting the Brain: Scalable Neural Prediction with POCO - Kempner Institute Predicting future neural activity is a critical step toward achieving real-time, closed-loop neurotechnologies. To this end, we introduce POCO, a unified forecasting model trained on diverse calcium i...

🧠📈A new foundation model for #neuroscience! In the latest Deeper Learning blog, @duranrin.bsky.social and @kanakarajanphd.bsky.social describe their state-of-the-art POCO model, which accurately forecasts neural dynamics across individuals and species.

bit.ly/4afqx5A #NeuroAI #neuroscience

Congrats @kanakarajanphd.bsky.social

Our mission: build a computational brain. Follow @compneurolab.bsky.social for updates.

#neuroAI #compneuro @utoronto.ca @uoftcompsci.bsky.social @kbi-uhn.bsky.social @vectorinstitute.ai

Brown University professor John Donoghue wins Queen Elizabeth Prize for Engineering Donoghue was awarded the prize, considered among the most prestigious honors in engineering, for pioneering work in developing brain-computer interfaces, which enable the restoration of voluntary comm...

Professor John Donoghue has been awarded a Queen Elizabeth Prize for Engineering. Read the story here: www.brown.edu/news/2026-02-03/donoghue-qeprize

Learning Abstractions for Hierarchical Planning in Program-Synthesis Agents Humans learn abstractions and use them to plan efficiently to quickly generalize across tasks -- an ability that remains challenging for state-of-the-art large language model (LLM) agents and deep rei...

A new and improved version of TheoryCoder, which learns to play video games in a human-like way by synthesizing both high-level abstractions and a low-level model of game mechanics:
arxiv.org/abs/2602.00929

OpenAI’s brain implant would use ultrasound to read minds. Does the science stand up? Spin-out Merge Labs aims to rival Elon Musk’s firm Neuralink. But researchers say the technology is still at an early stage.

Brain implants are beginning to help people with severe disabilities to speak and even sing in near-real time

go.nature.com/49RH292

BI 219 Xaq Pitkow: Principles and Constraints of Cognition YouTube video by Brain Inspired

On the latest Brain Inspired podcast episode, Xaq Pitkow shares his principles to study cognition in our imperfect brains and bodies, and how AI and machine learning are contributing to our efforts to understand brains and minds.

youtu.be/ugorctvkCa0?...

New paper from our lab on the behavioral significance of high-dimensional neural representations!

Biological Brains Inspire a New Building Block for Artificial Neural Networks Biological Brains Inspire a New Building Block for Artificial Neural Networks on Simons Foundation

While AI systems have advanced tremendously, they still lag behind real brains in reliability & efficiency. A new computational unit developed at #FlatironCCN could help close that gap: https://www.simonsfoundation.org/biological-brains-inspire-a-new-building-block-for-artificial-neural-networks

(PDF) Complete functional characterization of sensory neurons by system identification PDF | System identification is a growing approach to sensory neurophysiology that facilitates the development of quantitative functional models of... | Find, read and cite all the research you need on...

Voxelwise Encoding Models do some things differently from other GLM-related methods because the roots of VEM are in neurophysiology, not psychology. This 2006 paper describes system identification for neurophysiology. Change "neurons" to "voxels" and it all still applies.
tinyurl.com/wu-etal-2006

🚨 #CCN2026 Proceedings submissions are open!
CCN 2026 again features an 8-page Proceedings track (alongside extended abstracts). Accepted papers will appear in CCN-Proceedings (CCN‑P) with DOIs on OpenReview.

New paper with @deanpospisil.bsky.social , in which we introduce a new estimator for the "signal eigenspectrum" (i.e., the eigenvalues of the noiseless population responses). We re-analyze data from Stringer et al 2019 and show eigenvalues of mouse V1 are well explained by a broken power.

now accepted at ICLR! 🐺🥳🐺

arxiv.org/abs/2506.20666

Sadly, we can relate. 🇺🇦

Looking forward to presenting at the #AAAI #NeuroAI workshop; including 3 projects that were just accepted to ICLR! arxiv.org/abs/2509.24597, arxiv.org/abs/2510.03684, arxiv.org/abs/2506.13331 🧪🧠🤖

Summer Workshop on the Dynamic Brain 2026 The Summer Workshop on the Dynamic Brain is an intensive, project-based residential course with a focus on the neurobiology of sensory processing, coding, and neural population dynamics.

You have less than a week to apply for the Summer Workshop on the Dynamic Brain in the gorgeous San Juan Islands!

Apply by February 1 to be considered for this 2-week residential course and deepen your data analysis and computational skills.

🔗 https://bit.ly/4samKOZ

Using artificial neural networks to reveal the human confidence computation Author summary Human decisions are accompanied by a sense of confidence which reflects the decision accuracy. Conventionally, human confidence has been studied using two-choice tasks with simple stimu...

How do people compute a sense of confidence? This question is usually addressed using very simple images because we don't know how complex stimuli are represented internally. In a new paper, we addressed this question using artificial neural networks (ANNs).

journals.plos.org/ploscompbiol...