BlueSky has been trying to enforce some awkward age verification on Texas residents, so I couldn't log in until I was in Chicago for ICONS this week. Lo and behold, there are many neuroscience posts about new findings... of things that people have known forever. We really are the Sisyphus field🧠🧪🤖

AnD tHe BRaiN is JuST liKe aN AnN... JusT... goTTa ... FINd ... tHe ... bACkpRop... Seriously though... Neuromodulators are cooler than anything anyone in NeuroAI gives them credit for. Maybe norepinephrine and the LC's few hundred neurons is actually all you need.

Jeremiah Cohen@jeremiahycohen.bsky.social · 4mo ago

Delighted to share our discoveries about one of the brain's neurotransmitter systems: www.biorxiv.org/content/10.6... Together with colleagues at the @alleninstitute.org, we have learned a lot about a tiny cluster of neurons in the brainstem locus coeruleus (LC) that releases norepinephrine (NE). 1

We should fund science more not less. But the cynical part of me thinks this is an opportunity to maybe stop funding the same neuroscience questions with just a fancier microscope or just one region over time and time again. We need to cure diseases and fix AI. Not keep doing the same thing.

Excited to be in San Antonio for the UTSA AI Matrix THOR Neuromorphic Commons kickoff! THOR will be one of the first community resources fully dedicated for accessing scalable neuromorphic hardware! Check it out! #NeuroAI #Neuromorphic 🧪🧠🤖 www.neuromorphiccommons.com/events/thor_...

Am I alone in being skeptical when seeing "We need WORLD MODELS because that's what the BRAIN does!"? Won't this just be another 'AI tech bros use the brain to get attention and $$$ but ignore it at the first opportunity'? Why trust any of the AI crowd to talk about the brain? We need real #NeuroAI

Maybe cynical, but any time a scientist says "We should leave AI to industry" it is because they are scared about $$ going to something they don't work on I've heard for >10 years "let industry lead" and we have LLMs, power plants & data centers #NeuroAI needs research, not just venture capital

Brad Aimone@jbimaknee.bsky.social · 7mo ago

There is plenty of space to innovate in #NeuroAI. The issue has been that neuroscientists don't even try as they assume industry will do it. Deferring AI and neural computing to an industry that only cares about selling ads is not a way to help further our understanding of the brain.

My prediction / hope for 2026: this will be the year we start seeing the theoretical neuroscience and #NeuroAI fields start embracing spiking as something beyond a poor man's ReLU. Spikes aren't just the brain's activation function but they are fundamentally different. 1/ 🧠🧪🤖

I find it annoying that people who for many years have been at the forefront of "Scale, baby, scale!" are now lecturing on the importance of actually thinking. There is a community of people who have been thinking all along. Without expending terawatt-hours of electricity for no good reason

This gets to the heart of a big problem in neuroscience. Since we don't have a formal model of neural computation and since everything is important to people who study it, we are trapped in this "but my thing matters too!" mode Prioritization of what is important is necessary, even if harsh

Markus Meister@mameister4.bsky.social · 9mo ago

3. Compared to the other interactions (chemical synapses, electrical synapses, neuromodulation) the LFP is exceedingly weak and low-dimensional. 4. So, if you model the CNS by successive approximation, you can safely ignore extracellular fields in your first few attempts. 2/2

I can't stop reading Konrad's thread here, and it just drives something home that I'm increasingly appreciating: Neuroscientists really really really REALLY want the brain to be low dimensional. So much so that the low dimensionality is often assumed, and as Konrad gets to here, that causes problems

Konrad Kording@kordinglab.bsky.social · 9mo ago

I have so many issues with this podcast with @earlkmiller.bsky.social . I think that this podcast nicely shows why I have trouble with such approaches. Lets go through some of the claims.

Konrad is doing an amazing take-down here, and at least on this point I fully agree. The "brain is analog" argument is a terrible strawman that holds back neuro. Analog dynamics may help, but the theoretical scaling of neurons mostly come from event-driven parallelism. arxiv.org/abs/2507.17886

Neuromorphic Computing: A Theoretical Framework for Time, Space, and Energy Scaling

Neuromorphic computing (NMC) is increasingly viewed as a low-power alternative to conventional von Neumann architectures such as central processing units (CPUs) and graphics processing units (GPUs), h...

arxiv.org

Konrad Kording@kordinglab.bsky.social · 9mo ago

The whole “analog = parallel, digital = serial” contrast is a straw man built to make the brain look uniquely magical.

Our paper on a neuromorphic algorithm for solving finite element models is out, which is exciting for a number of reasons. But really I want to talk a bit about how this relates to understanding the brain and #NeuroAI. 🤖🧪🧠🧵 1/ www.nature.com/articles/s42...

Solving sparse finite element problems on neuromorphic hardware - Nature Machine Intelligence

Theilman and Aimone introduce a natively spiking algorithm for solving partial differential equations on large-scale neuromorphic computers and demonstrate the algorithm on Intel’s Loihi 2 neuromorphi...

nature.com

Nothing like flying out of a city when all the SfN folks are flying in. Satellite meetings are where it's at. 3 over 3 days was a little intense. But smaller meetings are more fun. Enjoy SfN everyone!

I was amused to find myself quoted, kind of, in the last paragraph of this; don't know if I've ever been the punchline quote before!

Proceedings of the National Academy of Sciences@pnas.org · 10mo ago

Can #NeuromorphicComputing help reduce AI’s high #energy cost? Researchers see big potential in #EnergyEfficient systems inspired by the #HumanBrain. A PNAS Core Concept explainer: https://ow.ly/45rk50XkYt5 #AI #ArtificialIntelligence #LLMs #ChatGPT #NeuralNetwork #DataCenter

This is a great thread, and I think it hits on one of the biggest challenges in neuroscience that I hope NeuroAI can impact. Until we can rigorously define what computations are occurring in the brain, we can't make any real progress in our functional understanding. We need constraints. 🧠🤖🧪

Sam Gershman@gershbrain.bsky.social · 11mo ago

I once saw a (very interesting) talk about sleep in which the speaker started by saying that we don't really know how to define sleep, and then proceeded to operationalize sleep in flies as basically periods when they are still for a long time. This got me thinking...

We put the FlyWire connectome on the Loihi 2 neuromorphic platform 🤖🧠🧪🪰 arxiv.org/abs/2508.16792

Neuromorphic Simulation of Drosophila Melanogaster Brain Connectome on Loihi 2

We demonstrate the first-ever nontrivial, biologically realistic connectome simulated on neuromorphic computing hardware. Specifically, we implement the whole-brain connectome of the adult Drosophila ...

arxiv.org

An interesting comment and discussion that gets to the deeper question of whether NeuroAI is a move towards something new or repackaging of old tired approaches with a shiny AI paint job.

Doug Crawford@jdcrawford.bsky.social · last yr.

This statement is frustrating because systems neuroscientists (e.g. @gunnarblohm.bsky.social) have spent many years trying to build biologically realistic, mechanistic network models that are largely ignored outside a small community. And now NeuroAI is going to discover this is important?

Traffic in Seattle is really bad...

I'm a big fan of computing with neurons, stem cells, organoids, and energy efficient AI. But we really have to stop coming up with systems (bio or inorganic) that are a few thousand neurons and claiming wins. AI models have BILLIONS of neurons. GPUs are quite efficient for the models they run 🤖🧠🧪

IEEE Spectrum@spectrum.ieee.org · last yr.

As AI's energy demands soar, biochips might hold the answer. How do lab-grown neurons redefine computing efficiency?

Spiking neural networks people, this message is for you! The annual SNUFA workshop is now open for abstract submission (deadline Sept 26) and (free) registration. This year's speakers include Elisabetta Chicca, Jason Eshraghian, Tomoki Fukai, Chengcheng Huang, and... you? snufa.net/2025/ 🤖🧠🧪

For a few years I have said that neuromorphic is specialized general purpose, like GPUs, but with different advantages. In this preprint I try to put some substance to that claim. There are real theoretical advantages, but they aren't obvious. 🧪🧠🤖 www.arxiv.org/abs/2507.17886

Neuromorphic Computing: A Theoretical Framework for Time, Space, and Energy Scaling

Neuromorphic computing (NMC) is increasingly viewed as a low-power alternative to conventional von Neumann architectures such as central processing units (CPUs) and graphics processing units (GPUs), h...

arxiv.org

This is an interesting thread getting to th heart of the neuroscience / AI disconnect. I think we need better "comp neuro for dummies" options, but there is a widely held view among engineers and computing people that neuroscientists obsess over details for no reason.

Shahab Bakhtiari@shahabbakht.bsky.social · last yr.

For trainees entering computational neuroscience or NeuroAI from an engineering background, where do you direct them to learn some neuroscience these days? Books, courses, ...? And no... I'm not interested in scaring them off with Kandel! 🧠🤖, 🧠📈

Dear editors: If you want me to review a paper, give me a form with two entries: comments for authors, optional comments for editor. Maybe (maybe) ask for accept/revise/reject recommendation, but that is really your job. Please don't ask me to answer 12!!! separate questions... 🧪

Lately my feed has been full of stories like "The brain's learning is more complex than Hebb thought!" and "Does the adult brain have new neurons?". Hebb was almost 100 years ago and the neurogenesis "debate" is 30 years old. We are worse than Hollywood in terms of rehashing the same old stories 🧠🧪

Bill Dally presenting the rise in necessary compute for AI I see why this is good for Nvidia. But why would this be considered good for anyone else?

At the @cra-hq.bsky.social CCC Computing Futures Symposium this week - exciting (and interesting) times to be talking about the future of computing research! As part of this event, last night we demo'd neuromorphic and its potential impact on ModSim and #NeuroAI at the US Senate!

With the demise of Twitter, I greatly miss the often critical but honest takes on new results. LinkedIn's AI community is all hype and often wrong. I would like to see Bsky prioritize honesty and truth over becoming an echo chamber, particularly with #NeuroAI new results that risk being misused 🧠🧪🤖

Konrad Kording@kordinglab.bsky.social · last yr.

There is a strong anti-negative-comment bias on here. And how can you discuss if the negative part of the spectrum is unacceptable?

This RealID chaos is so strange. I am quite certain I have had a compliant ID for about 10, maybe even 15 years. What is going on that some states are so backwards that they can't figure this out? And I used to live in some pretty dysfunctional states that could even mange this.

I'm unsure whether I agree with this. On the one hand, questions matter the most. On the other hand, I'm not convinced we (as neuroscientists) know the right questions to ask.

Grace Lindsay@neurograce.bsky.social · last yr.

While I would say my research is part of #NeuroAI, I don't actually thing it's very useful to define a research field as a constellation of methodologies. We should organize primarily around questions, and then use whatever methods best answer them.

I'm not a fan of either, but at least with "ANNs to describe the brain" the idea is new. For 50+ years physicists have been trying to shoehorn the brain into mean field approaches with little justification beyond "it would be so convenient" If you don't want to think about the brain, don't study it