Sign in

Steve Fleming

@smfleming.bsky.social
6.6K followers 1.1K following 351 posts

Psychologist and neuroscientist at UCL metacoglab.org Author, Know Thyself (2021) metacoglab.org/book Dad and assistant to the Diplomat

PostsRepliesMedia
Reposted by Steve Fleming
Kerri Smith @kerri-smith.bsky.social · 17/09/2026
How do minds think about themselves, and how do we study it? New review in @natrevneuro.nature.com www.nature.com/articles/s41...
nature.com
Towards an integrative neuroscience of metacognition - Nature Reviews Neuroscience
The capacity to self-evaluate and control one’s own cognitive performance, known as metacognition, underpins adaptive learning, social behaviours and self-belief. Stephen Fleming outlines our current…
192
Steve Fleming @smfleming.bsky.social · 5h
thanks for sharing Kerri!
000
Steve Fleming @smfleming.bsky.social · 10h
Useful thread about how to navigate the (rapidly changing) threat of bots in online behavioural studies. A move to unpaid citizen science platforms might end up being the best long-term strategy here
0112
Reposted by Steve Fleming
Christoph Strauch @cstrauch.bsky.social · 02/10/2026
How much of a risk are LLM-generated bots for online behavioral experiments? We present results of an entirely natural language prompt-generated agent, capable of absolving any of the 26 experiments we tested it on. Most tasks are passed with data that can pass as human osf.io/preprints/ps... (1/)
59143
Reposted by Steve Fleming
Mel Goodale @action-brain.bsky.social · 22h
Good to see that our opinion piece on grand theories of consciousness is out. #consciousness #neuroskyence #philsci #philsky🧠 #neuroscience #brain 🧪
12813
Reposted by Steve Fleming
Tim Behrens @behrenstimb.bsky.social · 22h
Best place to do neuroscience in the world.
0161
Reposted by Steve Fleming
megan peters 🧠 @meganakpeters.bsky.social · 06/10/2026
I am able to take on new doctoral students through this amazing initiative! If you are interested please reach out to me by email to discuss.
14923
Reposted by Steve Fleming
Heleen Slagter @haslagter.bsky.social · 06/10/2026
Reposting, in case you missed it! 8-) Reskeet appreciated!
11919
Reposted by Steve Fleming
Department of Imaging Neuroscience @imagingneuroucl.bsky.social · 06/10/2026
#BrainMeeting 🧠 Alert! 🎺 This Friday, October 9th, the Brain Meeting speaker will be Valentin Wyart giving a talk entitled "Zeroing in on human information seeking". Join us in person or online. Read more: www.in.fil.ion.ucl.ac.uk/events/ygazm...
0134
Steve Fleming @smfleming.bsky.social · 06/10/2026
Yes! We have capacity to cover international fees
060
Reposted by Steve Fleming
Nora Bradford @norabradford.bsky.social · 05/10/2026
love to see metacognition getting public attention:) www.thetransmitter.org/decision-mak...
thetransmitter.org
Is confidence a second thought—or part of the first one?
Neural circuits may sometimes calculate confidence in a decision simultaneously with the choice itself.
0125
Steve Fleming @smfleming.bsky.social · 05/10/2026
Nobel Prize for optogenetics! www.theguardian.com/science/2026...
theguardian.com
Nobel prize in medicine 2026 awarded for research into mysteries of brain
Three scientists to share 12m Swedish kronor prize for their work on ‘light-gated ion channels and optogenetics’
0425
Steve Fleming @smfleming.bsky.social · 05/10/2026
Excited to be launching our new PhD program in Minds and Machines at UCL with co-Directors Tali Sharot and @nadinedijkstra.bsky.social To apply, find out more here: www.ucl.ac.uk/brain-scienc... Our launch event is on Nov 10th, with a cracking lineup of speakers: luma.com/cikzcth0
lnkd.in
LinkedIn
This link will take you to a page that’s not on LinkedIn
29751
Reposted by Steve Fleming
Stefano Palminteri @stepalminteri.bsky.social · 05/10/2026
New preprint! Can LLMs really know when they know? We trained 10 open models to predict their own accuracy and found a striking limit: out of distribution, confidence tracks consistency more than correctness. Led by @nicolasyax.bsky.social with @pyoudeyer.bsky.social arxiv.org/pdf/2609.33886
0298
Reposted by Steve Fleming
Earl K. Miller @earlkmiller.bsky.social · 03/10/2026
This is a two-fer: 1. IT cortex encodes not just what objects are, but also where they are. 2. Current AI does not work like IT cortex. www.cell.com/current-biol... #neuroscience
cell.com
The macaque IT cortex but not current artificial vision networks encode object position in perceptually aligned coordinates
Yakubovskaya et al. show that motion adaptation shifts macaque IT population codes for object position in a direction that mirrors human perceptual mislocalization. Current feedforward, recurrent, and...
0308
Reposted by Steve Fleming
Eric Schwitzgebel @eschwitz.bsky.social · 02/10/2026
The Appeal of (Semi-?) Biological Naturalism [link in thread] Biological naturalism about consciousness will, I predict, soon be hot. As AI systems start to look more and more like conscious or sentient entities, at least superficially, doubters will be increasingly drawn to 1/3
A Mark Tobey painting
4103
Steve Fleming @smfleming.bsky.social · 29/09/2026
Amazing news @felipedebrigard.bsky.social, many congratulations!!
110
Reposted by Steve Fleming
Jakob Hohwy @hohwy.bsky.social · 26/09/2026
Wanja is an important voice in the science of consciousness. Read his illuminating commentary on our target paper - placing it in the wider context of theories of consciousness, and pushing toward new insights.
1175
Steve Fleming @smfleming.bsky.social · 23/09/2026
Wow that’s awesome Blake @tyrellturing.bsky.social, congratulations!!
120
Reposted by Steve Fleming
Jeni Kubota @jenikubota.bsky.social · 22/09/2026
📣 Please submit to this year's conference in Montréal!🧠 We would love to include all kinds of measures that can inform our understanding of the social brain, including physiological and neuro measures, computational models, theory pieces, and behavioral data!
1126
Steve Fleming @smfleming.bsky.social · 22/09/2026
I'm increasingly leaning towards in-person testing for this reason. For the studies that can't be done this way, e.g. big N individual differences, for now we appear reliant on the repetitional cost of subjects being called out and banned from Prolific for AI use. Unclear how sustainable this is!
120
Reposted by Steve Fleming
Richard O'Farrell @richieof.bsky.social · 22/09/2026
Recently ran a prolific study asking ppts to describe images in <20 words. With 50 captions per person, we were able to pass each participant’s full text into Pangram. It detected 2 out of 200 with high confidence as fully AI, and about 6/7 others as borderline AI/human. FWIW :)
253
Reposted by Steve Fleming
European Society for Cognitive Psychology @escop.bsky.social · 21/09/2026
Two Research Assistant Positions in the Consciousness Group, BCBL, Basque Center on Cognition, Brain & Language, Donostia-San Sebastián, Spain Application deadline: 8 October 2026
bcbl.eu
TWO RESEARCH ASSISTANT POSITIONS IN THE CONSCIOUSNESS Group
at the BCBL- Basque Center on Cognition Brain and Language (San Sebastián, Basque Country, Spain) www.bcbl.eu       The Basque Center on Cognition Brain and Language (BCBL) in San Sebastian (Basque Co...
062
Reposted by Steve Fleming
Wiktoria Kozyra @wiktoriakozyra.bsky.social · 19/09/2026
I’m very grateful to @smfleming.bsky.social for his supervision and @kevingoneill.github.io for a great collaboration. We show that the PE bias in confidence judgments can be a Bayes-optimal consequence of inference in high-dimensional spaces, and in CNNs, the bias also increases with dimensionality
1157
Reposted by Steve Fleming
Jorge Morales @jorgemorales.me · 18/09/2026
The Psychology & Neuroscience Department at Duke is looking to hire a Full Professor in clinical psychology (Distinguished Chair, endowed). Specific research area is open, but preference will be given to applicants employing human neuroscience to advance and improve mental health. Deadline Nov. 15.
academicjobsonline.org
Duke University, Psychology and Neuroscience
Job #AJO32753, Professor (tenured), Psychology and Neuroscience, Duke University, Durham, North Carolina, US
096
Reposted by Steve Fleming
Stanislas Dehaene @standehaene.bsky.social · 18/09/2026
An important new paper from the lab : in a few years, vision models have progressed to such an extent that they almost match human perception in the domain of geometry (but may still fall short of having a genuine language of shapes).
0176
Steve Fleming @smfleming.bsky.social · 18/09/2026
What if consciousness isn't real? By @timbayne.bsky.social in @sciam.bsky.social "we should ask whether bees, babies or bots are conscious. But in asking that question, we should be alive to the possibility that the question itself might need replacing." www.scientificamerican.com/article/what...
scientificamerican.com
What if ‘consciousness’ isn’t real?
Our understanding of consciousness could be fundamentally wrong
2158
Steve Fleming @smfleming.bsky.social · 18/09/2026
Thank you Alex!
010
Steve Fleming @smfleming.bsky.social · 18/09/2026
Awesome!
000
Steve Fleming @smfleming.bsky.social · 18/09/2026
Thanks Maxime! Yes- some ongoing work on deploying common tasks/models across species between labs here at UCL and potentially further afield
110
Reposted by Steve Fleming
Anil Seth @anilseth.bsky.social · 17/09/2026
5/ My response "The stuff matters: consciousness, computation, and biology" was published just now, and is here: url.uk.m.mimecastprotect.com/s/xZqFCN05NT...
1195
Steve Fleming @smfleming.bsky.social · 17/09/2026
thank you Micah!
100
Reposted by Steve Fleming
Anil Seth @anilseth.bsky.social · 16/09/2026
1/6 An extraordinary @nature.com paper has just landed from the lab of @sergiuppasca.bsky.social. They found a way to create "xenocortical mice" which have ~92% of their cortex occupied by a cortical organoid derived from human stem cells.
69140
Reposted by Steve Fleming
Nature Reviews Neuroscience @natrevneuro.nature.com · 16/09/2026
Towards an integrative neuroscience of metacognition — a Review by Stephen M. Fleming #neuroscience #neuroskyence www.nature.com/articles/s41...
nature.com
Towards an integrative neuroscience of metacognition - Nature Reviews Neuroscience
The capacity to self-evaluate and control one’s own cognitive performance, known as metacognition, underpins adaptive learning, social behaviours and self-belief. Stephen Fleming outlines our current ...
04511
Steve Fleming @smfleming.bsky.social · 16/09/2026
Full text link is here: rdcu.be/9YK3mF7dYhFa
rdcu.be
051
Steve Fleming @smfleming.bsky.social · 16/09/2026
Thanks to Kaiden Stewart, @tianqizhan.bsky.social, @kevingoneill.github.io, @nadinedijkstra.bsky.social and @seowxft.bsky.social for comments during the writing, Chris Fetsch, Robbe Goris, Jongmin Moon, @jonasobleser.bsky.social and an anonymous reviewer, and Katherine Whalley for editorial guidance
160
Steve Fleming @smfleming.bsky.social · 16/09/2026
It also includes some exploration of how metacognitive mechanisms such as reality monitoring might contribute to functional explanations of conscious experience
140
Steve Fleming @smfleming.bsky.social · 16/09/2026
A key aim was to provide a synthesis of the current experimental work that can illuminate metacognitive function and dysfunction in relation to psychiatric and neurological disorders, and contextualise the very fast-moving world of exposing metacognitive mechanisms in AI systems (not just LLMs!)
261
Steve Fleming @smfleming.bsky.social · 16/09/2026
We now have an increasingly detailed understanding (although with much more still to discover) of how metacognition is supported by a collection of neural and computational building blocks in human and animal brains
192
Steve Fleming @smfleming.bsky.social · 16/09/2026
This is unusually fertile ground for connecting systems, computational and cognitive neuroscience in an understanding of how minds of all shapes and sizes think about themselves
140
Steve Fleming @smfleming.bsky.social · 16/09/2026
I try to articulate the potential of this field to integrate multiple perspectives on how brains and minds work. Measures of metacognition connect to subjective experience via links to introspection, while being grounded in tasks and models that can be applied in humans, animals and AI systems
150
Steve Fleming @smfleming.bsky.social · 16/09/2026
📖 New review out today in @natrevneuro.nature.com: Towards an integrative neuroscience of metacognition 🧠🧪 Full-text link to the paper is here: rdcu.be/9YK3mF7dYhFa
414058
Reposted by Steve Fleming
Kevin O’Neill @kevingoneill.github.io · 15/09/2026
delighted to have been able to chip in on this newly published paper by @wiktoriakozyra.bsky.social & @smfleming.bsky.social on how confidence “biases” arise from normative Bayesian models!
1157
Reposted by Steve Fleming
Claire Gillan @clairegillan.bsky.social · 14/09/2026
Hey our psychology department is recruiting an assistant professor in developmental psychology - please check it out universityvacancies.com/trinity-coll...
universityvacancies.com
University & College Vacancies
Connecting talent with opportunity in higher education
01016
Reposted by Steve Fleming
Renzo Lanfranco @renzolanfranco.bsky.social · 14/09/2026
How much information about a stimulus is available to a perceptual system, and how much of that information becomes consciously accessible? In this new Perspective, “Perceptual awareness as an information-distribution problem,” I propose an information-theoretic framework. 🧵Preprint: osf.io/zmf5h
osf.io
OSF
1206
Steve Fleming @smfleming.bsky.social · 14/09/2026
Link again to paper here direct.mit.edu/opmi/article...
direct.mit.edu
Confidence Is Detection-Like in High-Dimensional Spaces
Abstract. Confidence estimates are often “detection-like”—driven by positive evidence in favour of a decision. This empirical observation has been interpreted as showing that human metacognition is li...
010
Steve Fleming @smfleming.bsky.social · 14/09/2026
A broader implication is that, in higher dimensional spaces, confidence for detection and identification become more similar, and strict dissociations (e.g. in blindsight and related conditions) may be a consequence of the adoption of narrower stimulus spaces
110
Steve Fleming @smfleming.bsky.social · 14/09/2026
This is still circumstantial evidence, and the proposal needs a more direct test – but it’s at least consistent with the model
110
Steve Fleming @smfleming.bsky.social · 14/09/2026
One consequence is that it should be easy to “flip” the effect of the observer’s goal on the PEB. This is what we found recently in empirical work with Prady Sepulveda and Benedetto De Martino www.biorxiv.org/content/10.6...
biorxiv.org
Goals shape human confidence
Confidence guides our decisions: it determines when we commit to a choice, gather more information, or change our minds. Prominent models treat confidence as a readout of the evidence in support of a ...
120
Steve Fleming @smfleming.bsky.social · 14/09/2026
The intuition is that as dimensions are added to the stimulus space, they compress the effect of distractors on confidence due to normalisation. This means that confidence is more sensitive to the chosen evidence relative to unchosen evidence, which is the hallmark of the positive evidence bias
110