Sebastian Speer

We're a month further into the job market - how are things looking? The good news is that the market does seem to have been delayed: ~150 new listings have appeared since my last post. The bad news is that the total is still substantially lower than what it was during the covid dip.

Paper and R package for (more flexible) power analysis in multilevel: psycnet.apa.org/record/2024-...

As our lab started to build encoding 🧠 models, we were trying to figure out best practices in the field. So @neurotaha.bsky.social
built a library to easily compare design choices & model features across datasets!

We hope it will be useful to the community & plan to keep expanding it!
1/

Good Conversations Don’t Require Everybody to Agree, Neuroscience Shows Brain imaging is illuminating the patterns linked to productive, positive dialogue, and those insights could help people connect with others

@falklab.bsky.social wrote a nice summary of our hyperscanning work on how people form social connection and find agreement through conversation: www.scientificamerican.com/article/good...

Toward a unified taxonomy of information dynamics via Integrated Information Decomposition | PNAS Our ability to understand and control complex systems of many interacting parts remains limited. A key challenge is that we still do not know how b...

Finally published:
“Toward a unified taxonomy of information dynamics via Integrated Information Decomposition”
www.pnas.org/doi/10.1073/...

An information-theoretic framework to analyse complex multivariate dynamical processes ✨

“As AI tools become more capable, funding agencies and institutions may question why labs need dedicated computational staff. But these examples suggest the roles will become more important, not less”

Increasing alignment of large language models with language processing in the human brain - Nature Computational Science Larger LLMs’ self-attention more accurately predicts readers’ regressive saccades and fMRI responses in language regions, whereas instruction tuning adds no benefit.

📢Out now! @jixingli.bsky.social, @lamb-cityuhk.bsky.social and colleagues assess whether instruction tuning can enhance LLM's ability to capture linguistic information in the human brain. www.nature.com/articles/s43... #compneuro #ArtificialIntelligence

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

How to compute surprisal (Shannon Information) for every word in your language sample. Then --- correlate surprisal with other major lexical-semantic variables (concreteness, semantic neighborhoods, freq). reilly-lab.github.io/Surprisal.html

These are the most challenging times for early-career scientists and engineers.

My team is doing its bit to help by maintaining and even expanding our database of funding opportunities for early-career researchers.

We found 437 of them.

Download it freely here: research.jhu.edu/rdt/funding-...

CICADA: An automated and flexible tool for comprehensive fMRI noise reduction Abstract. Independent component analysis (ICA) denoising methods can be highly effective for reducing functional magnetic resonance imaging (fMRI) noise. ICA denoising method success heavily depends, ...

My brilliant labmate, Keith Dodd, created this non-aggressive denoising pipeline for fMRI data for his dissertation! It has been super useful on some of our high-motion datasets, and I highly suggest folks check it out for their own studies 🧠📈🪲:
direct.mit.edu/imag/article...

🔖 NEW PAPER ALERT 🔖 Out in PNAS – we used equation discovery algorithms to improve models of human reinforcement learning.

Led by the wonderful Kyle Lafollette with an amazing team Yanni Yuval @roeyschurr.bsky.social and @d-melnikoff.bsky.social

www.pnas.org/doi/epdf/10....

Feature-based reward learning shapes human social learning strategies - Nature Human Behaviour This research advances a mechanistic reward learning account of social learning strategies. Through experiments and simulations, it shows how individuals learn to learn from others, dynamically shapin...

If you want to read more about the nuts and bolts of our reward learning account of social learning—the Social Feature Learning (SFL) model—check out our recent @nathumbehav.nature.com paper: www.nature.com/articles/s41...

Our conversation analysis software is now on CRAN! Reach out with any questions. We hope it’s useful for all you language and social interaction folks out there.

"We propose a unified quantitative model that integrates decision-making and communication, revealing how verbal framing of probabilistic expression both reflects and shapes beliefs in speakers and listeners."

Read more here:
www.sciencedirect.com/science/arti...

Evolving general cooperation with a Bayesian theory of mind | PNAS Theories of the evolution of cooperation through reciprocity explain how unrelated self-interested individuals can accomplish more together than th...

Our new paper is out in PNAS: "Evolving general cooperation with a Bayesian theory of mind"!

Humans are the ultimate cooperators. We coordinate on a scale and scope no other species (nor AI) can match. What makes this possible? 🧵

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

Optimistic people have functional MRI commonality of neural processing for imagining the future (whereas pessimists have a diverse pattern)
www.pnas.org/doi/epdf/10....

Collaborative Rational Speech Act: Pragmatic Reasoning for Multi-Turn Dialog As AI systems take on collaborative roles, they must reason about shared goals and beliefs-not just generate fluent language. The Rational Speech Act (RSA) framework offers a principled approach to pr...

"We introduce Collaborative Rational Speech Act (CRSA), an information-theoretic (IT) extension of RSA that models multi-turn dialog by optimizing a gain function adapted from rate-distortion theory."

arxiv.org/abs/2507.14063

Been really enjoying the Complexity Science Discussion Group hosted by Paul Middlebrooks, where we're working our way through the Santa Fe Institute's "Classic Papers in Complexity Science". The group is open for anyone interested:

braininspired.co/complexity-g...

Major updates to our ConversationAlign R package - submitted to CRAN and JOSS today. If you're interested in computing alignment and main effects for >40 psycholinguistic dimensions in naturalistic conversation, this is the package for you reilly-conceptscognitionlab.github.io/Conversation...

Seamless Interaction: Dyadic Audiovisual Motion Modeling and Large-Scale Dataset Human communication involves a complex interplay of verbal and nonverbal signals, essential for conveying meaning and achieving interpersonal goals. To develop socially intelligent AI technologies, it...

Seamless Interaction dataset with 4000+ hrs of face-to-face dyadic interactions! 🤯

Seems like it could be a game changer for studying embodied turn-taking & mutual adaptation complementary with CANDOR, etc...

arxiv.org/abs/2506.22554

Discovering cognitive strategies with tiny recurrent neural networks - Nature Modelling biological decision-making with tiny recurrent neural networks enables more accurate predictions of animal choices than classical cognitive models and offers insights into the underlying cog...

Thrilled to see our TinyRNN paper in @nature! We show how tiny RNNs predict choices of individual subjects accurately while staying fully interpretable. This approach can transform how we model cognitive processes in both healthy and disordered decisions. doi.org/10.1038/s415...

Disagreement drives metacognitive development Metacognition improves significantly over childhood, but the mechanisms underlying this development are poorly understood. We first review recent rese…

Check out our new TICS paper on disagreement and metacognition! We argue that disagreement drives metacognitive development by expanding children's consideration sets.
With Antonia Langenhoff, Bill Thompson @wdt.bsky.social and Mahesh Srinivasan
www.sciencedirect.com/science/arti...

What a fantastic resource! Super valuable for those of us teaching experimental methods for undergrads (not to mention staying crystal clear on our own procedures) 👇

Movie-watching evokes ripple-like activity within events and at event boundaries Nature Communications - The neural processes involved in memory formation for realistic experiences remain poorly understood. Here, the authors found that ripple-like activity in the human...

🧠 Paper out!

We investigated how hippocampal and cortical ripples support memory during movie watching. We found that:

🎬 Hippocampal ripples mark event boundaries
🧩 Cortical ripples predict later recall

Ripples may help transform real-life experiences into lasting memories!

rdcu.be/eui9l

Does linguistic diversity when talking about emotions track wellbeing? In a new pre-print with @eriknook.bsky.social, we link emotion vocabularies to mental health in a large real-world psychotherapy dataset. Highlight: therapist emo vocab may help clients get better over time! shorturl.at/TKPN1

A screenshot of the first page of a newly published forum article entitled "Advancing GWAS of human communication" authored by Rosa S. Gisladottir in the journal Trends in Genetics. The abstract reads as follows: The last decade has seen an explosion in genome-wide association studies (GWAS) on almost any imaginable phenotype. Unfortunately, humanity’s most distinctive trait – communication, broadly construed – has been underserved. In this forum article I review recent advances and promising avenues that may help us understand the genetics and evolution of human communication.

Highly recommend this superb forum article on genomics of human communication traits, free to read in @cp-trendsgenetics.bsky.social. Beautifully succinct & accessible overview by @rosagisladottir.bsky.social on where the field is now & where it's heading. 🗣️💬🧠🧬🧪
www.sciencedirect.com/science/arti...

Friends writing a paper you're going to cite for years is a cause for celebration!

@elisabaek.bsky.social @rzhprfshr.bsky.social & @joebayer.bsky.social map how social connection is defined & measured, point to their false equivalencies + consider how media/constant connectivity changes things ♥️

If you fancy the idea that emotions are inferences we make together, have a read and fire back with critiques, questions, or priors of your own.

arxiv.org/abs/2506.08599

#neuroskyence #hyperscanning

Conserved brain-wide emergence of emotional response from sensory experience in humans and mice Emotional responses to sensory experience are central to the human condition in health and disease. We hypothesized that principles governing the emergence of emotion from sensation might be discovera...

This! This the type of work that makes me so optimistic that the next 3 decades of brain research will be more impactful that the last for understanding/treating mental conditions. We finally have the right measurement tools & analyses (eg for emotion).

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