Guillaume Bellec
@bellecguill.bsky.social
AI, Neuroscience and Music
Pinned
Pre-print 🧠🧪
Is mechanism modeling dead in the AI era?
ML models trained to predict neural activity fail to generalize to unseen opto perturbations. But mechanism modeling can solve that.
We say "perturbation testing" is the right way to evaluate mechanisms in data-constrained models
1/8
Is mechanism modeling dead in the AI era?
ML models trained to predict neural activity fail to generalize to unseen opto perturbations. But mechanism modeling can solve that.
We say "perturbation testing" is the right way to evaluate mechanisms in data-constrained models
1/8
Reposted by Guillaume Bellec
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.
academicpositions.harvard.edu/postings/15868
An opportunity to work with a great group of people across Harvard, MIT, and UC Berkeley.
February 10, 2026 at 7:36 PM
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.
academicpositions.harvard.edu/postings/15868
An opportunity to work with a great group of people across Harvard, MIT, and UC Berkeley.
Reposted by Guillaume Bellec
PhD Alert! We’re recruiting for a new, fully-funded #Neuroskyence PhD studentship to model multiple sclerosis in vitro, combining immunology with ephys. Led by the awesome @juliae.bsky.social
Please repost.
www.findaphd.com/phds/project...
Please repost.
www.findaphd.com/phds/project...
Modelling multiple sclerosis in a dish to test therapeutics at University of Glasgow on FindAPhD.com
PhD Project - Modelling multiple sclerosis in a dish to test therapeutics at University of Glasgow, listed on FindAPhD.com
www.findaphd.com
February 3, 2026 at 7:23 PM
PhD Alert! We’re recruiting for a new, fully-funded #Neuroskyence PhD studentship to model multiple sclerosis in vitro, combining immunology with ephys. Led by the awesome @juliae.bsky.social
Please repost.
www.findaphd.com/phds/project...
Please repost.
www.findaphd.com/phds/project...
Summer school on Neuro AI in Cambridge.
Registration Deadline: 16 Feb 2026
Speaker line up:
G. Bellec
A. Billard
R. Bogacz
R. Ponte Costa
W. Gerstner
M. Giugliano
L. Hunt
M. Sahani
P. Series
P. Tino
www.fens.org/news-activit...
Registration Deadline: 16 Feb 2026
Speaker line up:
G. Bellec
A. Billard
R. Bogacz
R. Ponte Costa
W. Gerstner
M. Giugliano
L. Hunt
M. Sahani
P. Series
P. Tino
www.fens.org/news-activit...
FENS-Chen Institute Summer Program on AI-Accelerated Neuroscience Discovery and Translation hosted by the University of Cambridge - Federation of European Neuroscience Societies
Application Submission Deadline: Monday 16th February 2026, midnight Applications will be open from Thursday 8th January to midnight on Monday 16th February. The application form and further details…
www.fens.org
January 30, 2026 at 4:47 PM
Summer school on Neuro AI in Cambridge.
Registration Deadline: 16 Feb 2026
Speaker line up:
G. Bellec
A. Billard
R. Bogacz
R. Ponte Costa
W. Gerstner
M. Giugliano
L. Hunt
M. Sahani
P. Series
P. Tino
www.fens.org/news-activit...
Registration Deadline: 16 Feb 2026
Speaker line up:
G. Bellec
A. Billard
R. Bogacz
R. Ponte Costa
W. Gerstner
M. Giugliano
L. Hunt
M. Sahani
P. Series
P. Tino
www.fens.org/news-activit...
I highly recommend this postdoc position in Paris.
Joao is amazing and it's a unique opportunity to do serious computational/ML analysis with human brain recordings in vivo. 🧪 🧠
Joao is amazing and it's a unique opportunity to do serious computational/ML analysis with human brain recordings in vivo. 🧪 🧠
Postdoc position in Paris: come help develop new generation human brain computer interfaces ⚡🧠💻
Interested? Contact me if you have experience with machine learning (e.g. simulation-based inference, RL, generative/diffusion models) or dynamical systems.
See below for + details and retweet 🙏
Interested? Contact me if you have experience with machine learning (e.g. simulation-based inference, RL, generative/diffusion models) or dynamical systems.
See below for + details and retweet 🙏
January 27, 2026 at 10:48 PM
I highly recommend this postdoc position in Paris.
Joao is amazing and it's a unique opportunity to do serious computational/ML analysis with human brain recordings in vivo. 🧪 🧠
Joao is amazing and it's a unique opportunity to do serious computational/ML analysis with human brain recordings in vivo. 🧪 🧠
These days I drink Iranian tea everyday so I don't forget.
Media coverage of these events have been very poor in my internet bubble so far.
Media coverage of these events have been very poor in my internet bubble so far.
January 26, 2026 at 1:29 PM
These days I drink Iranian tea everyday so I don't forget.
Media coverage of these events have been very poor in my internet bubble so far.
Media coverage of these events have been very poor in my internet bubble so far.
Reposted by Guillaume Bellec
Come join us for this first worshop of a three part series on the computational ingredients of reasoning in minds and AI. Reasoning is a complex term, especially in light of an exploding category of methods in LLMs. These workshops will explore reasoning’s multiple facets.
IVADO unveils the schedule of the first workshop "Cognitive Basis of #Reasoning (in Minds and #AI)", Jan 27-29, 2026, spearheaded by @taylorwwebb.bsky.social and Dhanya Sridhar.
🗓️ Schedule and speakers: ivado.ca/en/events/co...
📥 Registration: event.fourwaves.com/thematicseme...
🗓️ Schedule and speakers: ivado.ca/en/events/co...
📥 Registration: event.fourwaves.com/thematicseme...
Cognitive Basis of Reasoning (in Minds and AI) | IVADO
ivado.ca
January 20, 2026 at 10:34 PM
Come join us for this first worshop of a three part series on the computational ingredients of reasoning in minds and AI. Reasoning is a complex term, especially in light of an exploding category of methods in LLMs. These workshops will explore reasoning’s multiple facets.
I always have an issue with self proclaimed "foundation models in X".
A key defining feature of a foundation model is adoption by a community of users. If nobody has built on top of it. It's no foundation
A key defining feature of a foundation model is adoption by a community of users. If nobody has built on top of it. It's no foundation
January 13, 2026 at 11:30 PM
I always have an issue with self proclaimed "foundation models in X".
A key defining feature of a foundation model is adoption by a community of users. If nobody has built on top of it. It's no foundation
A key defining feature of a foundation model is adoption by a community of users. If nobody has built on top of it. It's no foundation
Neuromorphic camera for neural recordings 🤩🤩
So spiking cameras enabling sub millisecond optical imaging. Potentially enabling to record... Spikes? 🐸
Science history is just so ironic sometimes (50 years to close the loop)
So spiking cameras enabling sub millisecond optical imaging. Potentially enabling to record... Spikes? 🐸
Science history is just so ironic sometimes (50 years to close the loop)
For decades, imaging has forced the same trade-off:
Set a frame rate → sacrifice resolution or field of view to go fast.
We break that rule
Using an EVENT-BASED CAMERA, we record neural activity without frames. Signals are captured only where and when they occur:
www.biorxiv.org/content/10.6...
Set a frame rate → sacrifice resolution or field of view to go fast.
We break that rule
Using an EVENT-BASED CAMERA, we record neural activity without frames. Signals are captured only where and when they occur:
www.biorxiv.org/content/10.6...
Ultrafast Frame-Free Imaging of Neural Activity with Event Cameras
Frame-based fluorescence imaging has long defined how neural activity is optically measured. This approach requires acquiring all pixels within an image, regardless of whether they carry meaningful ne...
www.biorxiv.org
January 10, 2026 at 6:14 PM
Neuromorphic camera for neural recordings 🤩🤩
So spiking cameras enabling sub millisecond optical imaging. Potentially enabling to record... Spikes? 🐸
Science history is just so ironic sometimes (50 years to close the loop)
So spiking cameras enabling sub millisecond optical imaging. Potentially enabling to record... Spikes? 🐸
Science history is just so ironic sometimes (50 years to close the loop)
The kind of concept you hear once and you never forget.
If you know what the electrode records. You know what it stimulates.
If you know what the electrode records. You know what it stimulates.
It is not often I get an epiphany from our own research.
But this year Torbjørn Ness, Christof Koch, and I realized that when we know how to compute electric brain signals generated by a neuron, we also know how to electrically stimulate the same neuron.
journals.plos.org/ploscompbiol...
But this year Torbjørn Ness, Christof Koch, and I realized that when we know how to compute electric brain signals generated by a neuron, we also know how to electrically stimulate the same neuron.
journals.plos.org/ploscompbiol...
Predicting neural responses to intra- and extra-cranial electric brain stimulation by means of the reciprocity theorem
Author summary Electric brain stimulation is widely used in neuroscience and medicine, from mapping brain function during surgery to treating disorders such as Parkinson’s disease and depression. Yet ...
journals.plos.org
December 27, 2025 at 9:17 PM
The kind of concept you hear once and you never forget.
If you know what the electrode records. You know what it stimulates.
If you know what the electrode records. You know what it stimulates.
Reposted by Guillaume Bellec
Applications are open for the 2026 Paris Spring School of Imaging and Electrophysiology! It’s a great way to discover Paris (❤️) while learning a pretty wide range of techniques for neuroscience, in a fantastic environment. Share widely: parisneuro.ovh
Paris NeuroCourse | Optical Imaging and Electrophysiological Recording in Neuroscience
parisneuro.ovh
December 24, 2025 at 9:20 AM
Applications are open for the 2026 Paris Spring School of Imaging and Electrophysiology! It’s a great way to discover Paris (❤️) while learning a pretty wide range of techniques for neuroscience, in a fantastic environment. Share widely: parisneuro.ovh
Reposted by Guillaume Bellec
Sequences are everywhere! In every brain region. And are written in stone.
Invariant Activity Sequences Across the Mouse Brain.
Out today, by Célian Bimbard, with @kenneth-harris.bsky.social.
Based on data by Célian and by @intlbrainlab.bsky.social.
www.biorxiv.org/content/10.6...
Invariant Activity Sequences Across the Mouse Brain.
Out today, by Célian Bimbard, with @kenneth-harris.bsky.social.
Based on data by Célian and by @intlbrainlab.bsky.social.
www.biorxiv.org/content/10.6...
December 22, 2025 at 11:59 AM
Sequences are everywhere! In every brain region. And are written in stone.
Invariant Activity Sequences Across the Mouse Brain.
Out today, by Célian Bimbard, with @kenneth-harris.bsky.social.
Based on data by Célian and by @intlbrainlab.bsky.social.
www.biorxiv.org/content/10.6...
Invariant Activity Sequences Across the Mouse Brain.
Out today, by Célian Bimbard, with @kenneth-harris.bsky.social.
Based on data by Célian and by @intlbrainlab.bsky.social.
www.biorxiv.org/content/10.6...
Very important
This report in Nature on the costs of competing for & administering scientific grants is shocking: "In other words, European taxpayers will have spent more on the funding process than on the funding itself, and the scientific ecosystem has been drained." www.nature.com/articles/d41... 🧪
Point of no returns: researchers are crossing a threshold in the fight for funding
With so little money to go round, the costs of competing for grants can exceed what the grants are worth. When that happens, nobody wins.
www.nature.com
December 20, 2025 at 9:47 AM
Very important
Reposted by Guillaume Bellec
This report in Nature on the costs of competing for & administering scientific grants is shocking: "In other words, European taxpayers will have spent more on the funding process than on the funding itself, and the scientific ecosystem has been drained." www.nature.com/articles/d41... 🧪
Point of no returns: researchers are crossing a threshold in the fight for funding
With so little money to go round, the costs of competing for grants can exceed what the grants are worth. When that happens, nobody wins.
www.nature.com
December 19, 2025 at 6:46 PM
This report in Nature on the costs of competing for & administering scientific grants is shocking: "In other words, European taxpayers will have spent more on the funding process than on the funding itself, and the scientific ecosystem has been drained." www.nature.com/articles/d41... 🧪
Reposted by Guillaume Bellec
🔬 We're hiring a Junior Group Leader for Data-Driven Digital Twins/System Models in medicine/life sciences at @uni_goettingen! Perfect opportunity for early-career researchers who want to:
- Lead their own research group
- Work at the intersection of AI, simulation, and biomedicine
- Lead their own research group
- Work at the intersection of AI, simulation, and biomedicine
December 19, 2025 at 8:02 AM
🔬 We're hiring a Junior Group Leader for Data-Driven Digital Twins/System Models in medicine/life sciences at @uni_goettingen! Perfect opportunity for early-career researchers who want to:
- Lead their own research group
- Work at the intersection of AI, simulation, and biomedicine
- Lead their own research group
- Work at the intersection of AI, simulation, and biomedicine
Our paper on data constrained RNN that generalize to optogenetic perturbations now citable on eLife:
doi.org/10.7554/eLif...
doi.org/10.7554/eLif...
December 18, 2025 at 11:07 PM
Our paper on data constrained RNN that generalize to optogenetic perturbations now citable on eLife:
doi.org/10.7554/eLif...
doi.org/10.7554/eLif...
Reposted by Guillaume Bellec
We should just use standardized test scores + grades to admit students into college.
Everything else is way too easy to game, allowing the extremely rich to dominate access to prestigious colleges. www.nytimes.com/interactive/...
I've been saying this for a long time, but the data makes it clear.
Everything else is way too easy to game, allowing the extremely rich to dominate access to prestigious colleges. www.nytimes.com/interactive/...
I've been saying this for a long time, but the data makes it clear.
December 6, 2025 at 6:07 PM
We should just use standardized test scores + grades to admit students into college.
Everything else is way too easy to game, allowing the extremely rich to dominate access to prestigious colleges. www.nytimes.com/interactive/...
I've been saying this for a long time, but the data makes it clear.
Everything else is way too easy to game, allowing the extremely rich to dominate access to prestigious colleges. www.nytimes.com/interactive/...
I've been saying this for a long time, but the data makes it clear.
Reposted by Guillaume Bellec
1/ Why does RL struggle with social dilemmas? How can we ensure that AI learns to cooperate rather than compete?
Introducing our new framework: MUPI (Embedded Universal Predictive Intelligence) which provides a theoretical basis for new cooperative solutions in RL.
Preprint🧵👇
(Paper link below.)
Introducing our new framework: MUPI (Embedded Universal Predictive Intelligence) which provides a theoretical basis for new cooperative solutions in RL.
Preprint🧵👇
(Paper link below.)
December 3, 2025 at 7:19 PM
1/ Why does RL struggle with social dilemmas? How can we ensure that AI learns to cooperate rather than compete?
Introducing our new framework: MUPI (Embedded Universal Predictive Intelligence) which provides a theoretical basis for new cooperative solutions in RL.
Preprint🧵👇
(Paper link below.)
Introducing our new framework: MUPI (Embedded Universal Predictive Intelligence) which provides a theoretical basis for new cooperative solutions in RL.
Preprint🧵👇
(Paper link below.)
Reposted by Guillaume Bellec
How I contributed to rejecting one of my favorite papers of all times, Yes, I teach it to students daily, and refer to it in lots of papers. Sorry. open.substack.com/pub/kording/...
How I contributed to rejecting one of my favorite papers of all time
I believe we should talk about the mistakes we make.
open.substack.com
December 2, 2025 at 1:27 AM
How I contributed to rejecting one of my favorite papers of all times, Yes, I teach it to students daily, and refer to it in lots of papers. Sorry. open.substack.com/pub/kording/...
Reposted by Guillaume Bellec
The big AI labs continue to be infuriatingly opaque about the actual figures for their total electricity and water consumption
The best report I have seen from them so far is this one from Mistral in July, but it still left unanswered questions for me simonwillison.net/2025/Jul/22/...
The best report I have seen from them so far is this one from Mistral in July, but it still left unanswered questions for me simonwillison.net/2025/Jul/22/...
Our contribution to a global environmental standard for AI
Mistral have released environmental impact numbers for their largest model, Mistral Large 2, in more detail than I have seen from any of the other large AI labs. The methodology …
simonwillison.net
November 29, 2025 at 3:56 PM
The big AI labs continue to be infuriatingly opaque about the actual figures for their total electricity and water consumption
The best report I have seen from them so far is this one from Mistral in July, but it still left unanswered questions for me simonwillison.net/2025/Jul/22/...
The best report I have seen from them so far is this one from Mistral in July, but it still left unanswered questions for me simonwillison.net/2025/Jul/22/...
I am impressed by the improvement with Gemini 3 and nano 🍌.
I used nano banana to make scientific figures. With 2.5 it was repetitively putting a human brain inside the 🐭 head.
Now it draws an accurate mouse brain anatomy even seems to locate correctly the cortical areas. Big jump imo
I used nano banana to make scientific figures. With 2.5 it was repetitively putting a human brain inside the 🐭 head.
Now it draws an accurate mouse brain anatomy even seems to locate correctly the cortical areas. Big jump imo
November 21, 2025 at 11:32 AM
I am impressed by the improvement with Gemini 3 and nano 🍌.
I used nano banana to make scientific figures. With 2.5 it was repetitively putting a human brain inside the 🐭 head.
Now it draws an accurate mouse brain anatomy even seems to locate correctly the cortical areas. Big jump imo
I used nano banana to make scientific figures. With 2.5 it was repetitively putting a human brain inside the 🐭 head.
Now it draws an accurate mouse brain anatomy even seems to locate correctly the cortical areas. Big jump imo
Reposted by Guillaume Bellec
Our work on training biophysical models with Jaxley is now out in @natmethods.nature.com. Led by @deismic.bsky.social, with @philipp.hertie.ai, @ppjgoncalves.bsky.social & @jakhmack.bsky.social et al.
Paper: www.nature.com/articles/s41...
Paper: www.nature.com/articles/s41...
Jaxley: differentiable simulation enables large-scale training of detailed biophysical models of neural dynamics - Nature Methods
Jaxley is a versatile platform for biophysical modeling in neuroscience. It allows efficiently simulating large-scale biophysical models on CPUs, GPUs and TPUs. Model parameters can be optimized with ...
www.nature.com
November 13, 2025 at 1:08 PM
Our work on training biophysical models with Jaxley is now out in @natmethods.nature.com. Led by @deismic.bsky.social, with @philipp.hertie.ai, @ppjgoncalves.bsky.social & @jakhmack.bsky.social et al.
Paper: www.nature.com/articles/s41...
Paper: www.nature.com/articles/s41...
Whoopsy
We used the great Infinigram from Jiacheng Liu and found numerous hints of test set leakage in DCLM, which is used in OLMo-2
For instance, the fraction of MMLU questions that are leaked in pretraining had gone from ~1% to 24% between OLMo-1 and 2 😬
For instance, the fraction of MMLU questions that are leaked in pretraining had gone from ~1% to 24% between OLMo-1 and 2 😬
November 9, 2025 at 1:09 PM
Whoopsy
Reposted by Guillaume Bellec
I’m looking for interns to join our lab for a project on foundation models in neuroscience.
Funded by @ivado.bsky.social and in collaboration with the IVADO regroupement 1 (AI and Neuroscience: ivado.ca/en/regroupem...).
Interested? See the details in the comments. (1/3)
🧠🤖
Funded by @ivado.bsky.social and in collaboration with the IVADO regroupement 1 (AI and Neuroscience: ivado.ca/en/regroupem...).
Interested? See the details in the comments. (1/3)
🧠🤖
AI and Neuroscience | IVADO
ivado.ca
November 7, 2025 at 1:52 PM
I’m looking for interns to join our lab for a project on foundation models in neuroscience.
Funded by @ivado.bsky.social and in collaboration with the IVADO regroupement 1 (AI and Neuroscience: ivado.ca/en/regroupem...).
Interested? See the details in the comments. (1/3)
🧠🤖
Funded by @ivado.bsky.social and in collaboration with the IVADO regroupement 1 (AI and Neuroscience: ivado.ca/en/regroupem...).
Interested? See the details in the comments. (1/3)
🧠🤖
Reposted by Guillaume Bellec
Introducing CorText: a framework that fuses brain data directly into a large language model, allowing for interactive neural readout using natural language.
tl;dr: you can now chat with a brain scan 🧠💬
1/n
tl;dr: you can now chat with a brain scan 🧠💬
1/n
November 3, 2025 at 3:17 PM
Introducing CorText: a framework that fuses brain data directly into a large language model, allowing for interactive neural readout using natural language.
tl;dr: you can now chat with a brain scan 🧠💬
1/n
tl;dr: you can now chat with a brain scan 🧠💬
1/n
Reposted by Guillaume Bellec
Reminder this is happening this Wed/Thu. Free spiking neural network conference - registration required (see below).
Spiking NN fans - the #SNUFA workshop (Nov 5-6) agenda is finalised and online now. Make sure to register (free) soon. (Note you can register for either day and come to both.)
Agenda: snufa.net/2025/
Registration: www.eventbrite.co.uk/e/snufa-2025...
Thanks to all who voted on abstracts!
🤖🧠🧪
Agenda: snufa.net/2025/
Registration: www.eventbrite.co.uk/e/snufa-2025...
Thanks to all who voted on abstracts!
🤖🧠🧪
SNUFA 2025
Spiking Neural networks as Universal Function Approximators
snufa.net
November 3, 2025 at 3:28 PM
Reminder this is happening this Wed/Thu. Free spiking neural network conference - registration required (see below).