"Learning to be confident: How agents learn confidence based on prediction errors"! Now out in @cognitionjournal.bsky.social led by @pierreledenmat.bsky.social

Paper: desenderlab.com/wp-content/u... Thread ↓↓↓

#AcademicSky #PsychSciSky #Neuroscience #Neuroskyence

Introducing hMFC: A Bayesian hierarchical model of trial-to-trial fluctuations in decision criterion! Now out in @plos.org Comp Bio.
led by Robin Vloeberghs with @anne-urai.bsky.social Scott Linderman

Paper: desenderlab.com/wp-content/u... Thread ↓↓↓

#PsychSciSky #Neuroscience #Neuroskyence

Common neural choice signals reflect accumulated evidence, not confidence! Now out in @cerebralcortex.bsky.social w @helenevanmarcke.bsky.social @pierreledenmat.bsky.social @yfvisser.bsky.social @denizerdil.bsky.social a.o.

Paper: desenderlab.com/wp-content/u... Thread ↓↓↓

So, you know these "Read full article before sharing"-prompts? Yeah, not as effective as you would think! 🤔
But to hear our two cents on why this is the case and what you should do instead, you will have to (how ironic!) read the full article 🤓 (6/6)

‼️Issue 2: info seeking for fake news **did not** increase accurate fake news detection: participants remain no better than chance at detecting fake news! (5/6)

We find that confidence indeed drives info seeking (1️⃣, plot B) and confirmation biases (3️⃣), yet we identify two crucial issues.
❗Issue 1: confidence for fake news items does not reliably reflect accuracy (plot C), thus failing to drive information seeking towards the least accurate items. (4/6)

Here, we tested whether these 3 principles hold in the context of fake news: 314 American adults rated news headlines' veracity, gave their confidence and could decide to seek additional info (A). When seeking info, they were presented the full news article before finalising their decision (B).(3/6)

Fake news is on the rise, so how can we make people better at detecting it? Well, perceptual decision making research would tell us to simply encourage information seeking; which is
1️⃣driven by confidence,
2️⃣improves decision accuracy,
3️⃣cooccurs with confidence-driven confirmation biases. (2/6)

Warning: this is *not* fake news! New preprint out w/ S. Kunkle & @kobedesender.bsky.social on how confidence-driven information seeking is suboptimal in the context of fake news 📰🔍🧵 (1/6):
osf.io/preprints/ps...

Full-Force at #ccn2025 in Amsterdam. Come along for a chat if you're interested in metacognition, confidence, computational modelling, reasoning, etc. @yfvisser.bsky.social @jeremiebeucler.bsky.social @helenevanmarcke.bsky.social @alexandre-lietard.bsky.social @zoepurcell.bsky.social

Our findings suggest a less simple link between confidence and information seeking as previously thought, suggesting that confidence and information seeking are separately driven by beliefs about past performance versus perceived difficulty. 💭

Read all about it in our paper! (7/7)

In Exp.2, participants were only trained on one difficulty level (easy/hard) and experienced a sudden in-/decrease in trial difficulty in the testing phase, directly leading to an in-/decrease in overall information seeking to maintain performance. (6/7)

To explain this unexpected dissociation, we point towards the differences in training difficulty between experiments. In Exp.1, participants were trained on all difficulty levels, with no unexpected change in difficulty in the testing phase. Thus, info seeking was not directly affected. (5/7)

However, in Exp.2, an easy/hard training also induced over/underconfidence but respectively INcreased/DEcreased overall info seeking, an effect that was unmediated by confidence. At the trial level, info seeking was still driven by trial-level confidence. (4/7)

Strikingly, we found a dissociation between the type of confidence manipulation and the effect on information seeking. In Exp.1, positive/negative comparative feedback induced over/underconfidence and resp. decreased/increased info seeking, through a full mediation by trial-level confidence. (3/7)

Decision confidence is known to drive information seeking, but what happens when confidence is causally manipulated? In two experiments, we induced under- and overconfidence through a comparative feedback (Exp.1) or a training difficulty manipulation (Exp.2) and looked at information-seeking. (2/7)

Happy to share my latest work with @kobedesender.bsky.social on the context-dependent role of confidence in information-seeking 🔍now out in @cognitionjournal.bsky.social --
authors.elsevier.com/a/1lGLb2Hx2-... 🧵⬇️ (1/7)

PhD position in cognitive computational neuroscience! Join us, & investigate how we can endow domain-specific models of vision (eg DNNs) with domain-general processes such as metacognition or working memory.
All details => www.kuleuven.be/personeel/jo...
#PsychSciSky #Neuroscience #Neuroskyence

Why does it feel good to be sure? In new work @voalan.bsky.social @andero.bsky.social and I bridge confidence and affect. In perceptual DM, we found a remarkably consistent association between both measures! ↓↓↓

link: www.tandfonline.com/doi/full/10....
#PsychSciSky #Neuroscience #Neuroskyence