Last week, Areti Majumdar (2nd year PhD student) presented @jerrytang.bsky.social and colleague’s seminal paper on reconstructing continuous language from non-invasive brain recordings. This 🧵 explores our thoughts and questions.

Proud mentor moment! My student @mujianing.bsky.social (co-supervised w/ @alexanderhuth.bsky.social) presents at #CEMS2026 today. Her work challenges the idea that event boundaries are remembered because they're surprising. What predicts memory is how much a moment shares with the rest of the story.

A recent paper by Epp et al. in NN claimed that ~40% of reported BOLD findings could be misinterpreted. In our reanalysis, we identified several statistical issues that, in our view, undermine these conclusions. www.biorxiv.org/content/10.6... 👇 1/11

Really excited about our new work on aphasia! Even in fairly profound aphasia, we can recover semantic maps through visual stimuli and use them to decode language. This is a big step! Language BCIs in aphasia might be possible!

@jerrytang.bsky.social · 4mo ago

We're excited to share our new study on decoding brain activity in participants with post-stroke aphasia! We think this is an important step towards cognitive brain-computer interfaces for patients with language disorders www.biorxiv.org/content/10.6... 1/8

Pycortex, our awesome brain viewer software, has a major new feature! Thanks to the hard work of Dr. Matteo Visconti and Aditya Vaidya, Pycortex now supports headless webgl viewers! You can create 3D inflated plots without needing a display, and in Jupyter notebooks! gallantlab.org/pycortex/aut...

Thank you for your responses, Valentin! On this: I tried to match the contrast values to your results — the "true" ΔCBF is based on your reported 7.7% from Table S1; others are from Figure 2b,c. The noise-free simulation shows ΔCBF ranging from -7..7%, well within the -15..30% range in your Fig. 3b.

Valentin Riedl@vavatin.bsky.social · 7mo ago

MRI-data are noisy, but your simulation uses error-terms and SNRs beyond real data quality (i’d guess your CBF signal is around 5x weaker than imaging data, the real T2* changes are around 5x higher), so sure, you’ll easily (intentionally?) get more noise propagation.

This paper had a pretty shocking headline result (40% of voxels!), so I dug into it, and I think it is wrong. Essentially: they compare two noisy measures and find that about 40% of voxels have different sign between the two. I think this is just noise!

Eiko Fried@eikofried.bsky.social · 7mo ago

Would love to hear expert views on this paper. It appears to show that the operationalization of brain activity the field has relied on for 3 decades—the BOLD response—is not actually a sensible measure of brain activity. www.nature.com/articles/s41...

📣 New preprint from the Braga Lab! 📣 The ventral visual stream for reading converges on the transmodal language network Congrats to Dr. Joe Salvo for this epic set of results Big Q: What brain systems support the translation of writing to concepts and meaning? Thread 🧵 ⬇️

Happy and proud to see @rjantonello.bsky.social’s work awarded by SNL!

MIT Press@mitpress.bsky.social · 11mo ago

🏆 Best Paper — Predictive Coding or Just Feature Discovery? An Alternative Account of Why Language Models Fit Brain Data: doi.org/10.1162/nol_... Richard Antonello and @alexanderhuth.bsky.social

Our latest paper outlining our ecosystem of tools for mining the neuroimaging literature, is finally officially published in eLife! doi.org/10.7554/eLif...

Mining the neuroimaging literature

New tools for literature mining, such as automated analysis of the research literature, are accessible, scalable, and reliable.

doi.org

New Open dataset alert: 🧠 Introducing "Spacetop" – a massive multimodal fMRI dataset that bridges naturalistic and experimental neuroscience! N = 101 x 6 hours each = 606 functional iso-hours combining movies, pain, faces, theory-of-mind and other cognitive tasks! 🧵below

New paper with @rjantonello.bsky.social @csinva.bsky.social, Suna Guo, Gavin Mischler, Jianfeng Gao, & Nima Mesgarani: We use LLMs to generate VERY interpretable embeddings where each dimension corresponds to a scientific theory, & then use these embeddings to predict fMRI and ECoG. It WORKS!

bioRxiv Neuroscience@biorxiv-neursci.bsky.social · last yr.

Evaluating scientific theories as predictive models in language neuroscience https://www.biorxiv.org/content/10.1101/2025.08.12.669958v1

In our new paper, we explore how we can build encoding models that are both powerful and understandable. Our model uses an LLM to answer 35 questions about a sentence's content. The answers linearly contribute to our prediction of how the brain will respond to that sentence. 1/6

The preprint of my 1st project in grad school is up 🙌 We propose a simple, information-theoretic model of how humans remember narratives. We tested it with the help of open-source LLMs. Plz check out this thread for details ➡️ Many thanks to my wonderful advisors! It's been a fun adventure!!

Alexander Huth@alexanderhuth.bsky.social · last yr.

New paper with @mujianing.bsky.social & @prestonlab.bsky.social! We propose a simple model for human memory of narratives: we uniformly sample incoming information at a constant rate. This explains behavioral data much better than variable-rate sampling triggered by event segmentation or surprisal.

New paper with @mujianing.bsky.social & @prestonlab.bsky.social! We propose a simple model for human memory of narratives: we uniformly sample incoming information at a constant rate. This explains behavioral data much better than variable-rate sampling triggered by event segmentation or surprisal.

bioRxiv Neuroscience@biorxiv-neursci.bsky.social · last yr.

Efficient uniform sampling explains non-uniform memory of narrative stories https://www.biorxiv.org/content/10.1101/2025.07.31.667952v1

Hey, if you need a breather from the news, we have pretty brain pics to distract you! We made an interactive viewer showing images optimized to elicit responses from different places in the brain. Link to viewer in 🧵

Mark Lescroart@neuromdl.bsky.social · last yr.

In a new paper led by @matthewshinkle.bsky.social, we use a pre-trained deep neural network (DNN) to model brain activity and to make detailed visualizations of feature selectivity across many brain areas. In plain English: we make pretty pictures of what different parts of the brain like to see.

Our latest brain-to-image decoding model is now available on HuggingFace: "Dynadiff: Single-stage Decoding of Images from Continuously Evolving fMRI", led by Marlène Careil and Yohann Benchetrit: - Paper: arxiv.org/pdf/2505.14556 - Github: github.com/facebookrese... - Thread: 👇

people need to internalize, very quickly, that federal research grants are a hypercompetitive contracting process not charity, and that what Uncle Sam gets in return for that money is American dominance in the future

Tell Congress to protect crucial medical research funding #saveNIHresearch ucal.us/saveNIHresearch @uofcalifornia.bsky.social @ucsantabarbara.bsky.social

Proud to be a signatory of this statement from 1900 members of NASEM: We call on the administration to cease its wholesale assault on U.S. science, and we urge the public to join this call. docs.google.com/document/d/1...

Public Statement on Supporting Science for the Benefit of All Citizens

TO THE AMERICAN PEOPLE We all rely on science. Science gave us the smartphones in our pockets, the navigation systems in our cars, and life-saving medical care. We count on engineers when we drive acr...

docs.google.com

Excited to share on my first post on Bluesky our new paper in NHB examine the innate and developing aspects of the wiring of the visual system. Congratulations to @emilykubota.bsky.social and the baby MRI team on this important work

Emily Kubota@emilykubota.bsky.social · last yr.

The latest paper from my PhD is now out in Nature Human Behavior! rdcu.be/edRwQ

This was an interesting commentary to write on work by Binder et al. regarding impaired acoustic phonetic perception after unilateral left hemisphere stroke. academic.oup.com/brain/advanc...

Converging and conflicting evidence for left temporal lobe regions in acoustic-phonetic perception

This scientific commentary refers to ‘Lesion correlates of impaired acoustic-phonetic perception after unilateral left hemisphere stroke’ by Binder et al.

academic.oup.com

I am incredibly proud to share my first, first-author paper as a postdoc with @benhayden.bsky.social . How does the human hippocampus, known for encoding concepts, represent the meanings of words while listening to narrative speech? www.biorxiv.org/content/10.1...

Trends in Cognitive Sciences Ignoring the cerebellum is hindering progress in neuroscience www.cell.com/trends/cogni...

Ignoring the cerebellum is hindering progress in neuroscience

Traditionally considered a motor structure, the cerebellum has been shown to play a key role in several cognitive functions. However, for decades, the cerebellum has been largely overlooked and even d...

cell.com

Our paper on how neural codes track prior events in a narrative and predict subsequent memory for details, led by @collinsilvy.bsky.social, is now out in Communications Psychology! rdcu.be/d93Vc #neuroskyence #psychscisky

Neural codes track prior events in a narrative and predict subsequent memory for details

Communications Psychology - Context-dependent temporal structures are represented in multiple ways in parallel in the human brain during processing of temporally extended events. The neural...

rdcu.be

Many cognitive neuroscientists ignore the cerebellum but it's long past time to pay more attention to it (and fund more cerebellum grants). Not just for basic science: the cerebellum is commonly affected by pediatric brain tumors, leading to lifelong motor AND cognitive deficits. Must-read paper 👇🏼

Ignoring the cerebellum is hindering progress in neuroscience

Traditionally considered a motor structure, the cerebellum has been shown to play a key role in several cognitive functions. However, for decades, the…

sciencedirect.com

Very pleased to share our recent work, in which we use LLMs for automated discovery of interpretable models of animal behavior 🪰🐀🕵️‍♀️ that take the form of Python programs 🐍 See below for a summary of key results by @pcastr.bsky.social!

Pablo Samuel Castro@pcastr.bsky.social · 2y ago

Can LLMs be used to discover interpretable models of human and animal behavior?🤔 Turns out: yes! Thrilled to share our latest preprint where we used FunSearch to automatically discover symbolic cognitive models of behavior. 1/12

Two new studies from our team we're particularly happy about: Study 1: ai.meta.com/research/pub... Study 2: ai.meta.com/research/pub... Blog 3: ai.meta.com/blog/brain-a...

I'm excited to share our new paper (with @alexanderhuth.bsky.social) on transferring language decoders across participants and modalities! authors.elsevier.com/a/1kZRD3QW8S... 1/5