Elisabeth Parés Pujolràs

Elisabeth Parés Pujolràs

@epares.bsky.social

Cognitive Neuroscience postdoc @ucddublin, previously @UCL_ICN. Interested in the science & philosophy of decision making, voluntary action and consciousness.

215 followers406 following25 posts

... and this is for people who *did* have advanced skills to begin with. It is important for students to realise you should demand an education that teaches *you* how to learn, understand and do things, rather than cheating you out of this opportunity www.nature.com/articles/d41...

Iris van Rooij 💭@irisvanrooij.bsky.social

What a surprise ….. “Is AI ruining our skills? Early results are in — and they’re not good Reliance on artificial-intelligence tools degrades the abilities of physicians and software engineers, studies show.” www.nature.com/articles/d41...

Elisabeth Parés Pujolràs reposted

New research assistant position available in our human intracranial EEG lab in NY (hbmlab-nyc.com). DM or email me if you're interested to learn more!

Elisabeth Parés Pujolràs reposted

Come join the lab @maynoothuniversity.ie and work with me and stellar collaborators @seanfw.bsky.social and @redmondoconnell.bsky.social! Exciting project at the intersection of multi-task RNNs, compositional cognition and human decision neuroscience. Details and how to apply at shorturl.at/FsTQ0

Neuroscience Ireland@neuroscienceirl.bsky.social

✅Funded PhD studentship in computational modelling of supra-modal evidence accumulation/decision-making! ⏰ Closing date: 1st July 2026! ➡️ More information: shorturl.at/FsTQ0 @redmondoconnell.bsky.social @seanfw.bsky.social @maynoothuniversity.ie @tcdpsychology.bsky.social

No reviewer should have to waste any time trying to make sense of something that was not actually *thought* by someone else. LLMs produce stuff that *looks* sensible, but is an empty shell.

Iris van Rooij 💭@irisvanrooij.bsky.social

“Feel free to adapt and reuse. No attribution needed (…) If you make an adaption that you also want to share, I’d love to know about it, too.” irisvanrooijcogsci.com/2026/04/08/r...

Poorly-phrased thoughts and also mistaken ones can drive science forward. When reading those, a reviewer can ask: "what do you *mean*?" and a discussion can ensue. But a reviewer cannot ask for the meaning or assess the value of an LLM-generated shell of a sentence, because it contains none.

A reviewer's job is to engage with all the complex nuances of actual thoughts, which convey *someone else's* interpretation of the world. Human effort should only be spent understanding human work. Else we will soon be asked to accept that AI-based tools "evaluate" our work - as if they could.

Strong agree - doing things faster is not doing them better. Especially if you're not doing them at all. I too refuse to join that ship and agree we should fight to change flawed incentive systems rather than cave to them by slowly surrendering our ability to think for ourselves.

Gabriel Weindel@gweindel.bsky.social

Pressure to use genAI for #ECR is quite high, I feel it weekly... ethical, ecological, economical, privacy- and cognitive-related worries keeps me from joining that ship. I need more time to research not something that creates the illusion of reasoning, if that costs grants or positions, let it be.

"If everyone thinks technological trends are inescapable, they probably will become so. That is why those serious about the human prospect should reject the rhetoric of fatalism and demand something more substantial."

Elisabeth Parés Pujolràs reposted

Pressure to use genAI for #ECR is quite high, I feel it weekly... ethical, ecological, economical, privacy- and cognitive-related worries keeps me from joining that ship. I need more time to research not something that creates the illusion of reasoning, if that costs grants or positions, let it be.

Elisabeth Parés Pujolràs reposted

Elisabeth Parés Pujolràs reposted

Elisabeth Parés Pujolràs reposted

Tired of AI hype posts? You might like my sober assessment of whether AI-generated summaries are suitable for studying and research. Spoiler alert, they are not. The text is primarily aimed at students and researchers, but has much broader relevance. So share freely!: www.tue.nl/en/our-unive...

tue.nlAre AI-generated summaries suitable for studying and research?Despite didactic, ethical, and environmental concerns, the use of GenAI is on the rise in academia. For most applications, the jury is still out on whether and how they will benefit education and rese...

Elisabeth Parés Pujolràs reposted

Elisabeth Parés Pujolràs reposted

Postdoc position in Paris: come help develop new generation human brain computer interfaces ⚡🧠💻 Interested? Contact me if you have experience with machine learning (e.g. simulation-based inference, RL, generative/diffusion models) or dynamical systems. See below for + details and retweet 🙏

Elisabeth Parés Pujolràs reposted

Elisabeth Parés Pujolràs reposted

Our #sEEG study is now published in Nature Communications: rdcu.be/eIkoG! 🧠 Key finding: We discovered neural evidence accumulation for visual perception that's independent of report preparation—recorded from >3000 channels across 3 experiments! #Neuroscience #Consciousness #OpenAccess

rdcu.beCortical evidence accumulation for visual perception occurs irrespective of reportsNature Communications - Perception can occur in the absence of report. Here, the authors use intracranial recordings to show that sequential sampling of sensory evidence in the ventral visual...
François Stockart@francoisstock.bsky.social

Our new sEEG study is out on BioRxiv: tinyurl.com/4afh3b9n! Using 3 face detection experiments and recording from >3000 channels, we found a neural code that reflected evidence accumulation for conscious perception and perceptual confidence! 👇

View full thread

Overall, our results highlight how prominent neural decision signals may take different shapes in different contexts, and how probing them using tasks that depart from conventional ones can provide important insights into their functional roles. Thanks for reading this far!

View full thread

Our participants sometimes terminated decisions early, tending to use the second pulse less than the first one, and CPP amplitudes following each pulse reflected this differential weight on choice (i.e. small CPP, small weight on choice), also capturing condition-specific differences in the effect.

We show that when noisy evidence for a single decision comes in temporally separate pulses, the centroparietal positivity (CPP) tracks multiple rounds of evidence accumulation, falling back down to baseline during the intervals in between pulses when no relevant info is presented.

Elisabeth Parés Pujolràs reposted

📢Fully Funded PhD position in Barcelona! I'm excited to announce that I’m opening a PhD position at @idibaps.bsky.social, Barcelona! We'll investigate the role of bodily signals in autobiographical memory, using virtual reality, EEG and TMS, in healthy volunteers and patient populations.

Excellent piece. We know how to improve writing ability: it's by doing more, not less of it. "LLMs do not improve one’s writing ability much like taking a taxi does not improve one’s driving ability. Students should hone their writing, thinking, and other academic skills at every opportunity."

Finally! 🤩 Our position piece: Against the Uncritical Adoption of 'AI' Technologies in Academia: doi.org/10.5281/zeno... We unpick the tech industry’s marketing, hype, & harm; and we argue for safeguarding higher education, critical thinking, expertise, academic freedom, & scientific integrity. 1/n

Elisabeth Parés Pujolràs reposted

New paper from the lab indicates some surprising differences in visual cortical contributions to two popular VEP types. In it, Kieran uses some of his brilliant tricks from last year's paper:

direct.mit.eduVisual cortical area contributions to the transient, multifocal and steady-state VEP: A forward model-informed analysisAbstract. Central to our understanding of how visual-evoked potentials (VEPs) contribute to visual processing is the question of where their anatomical sources are. Three well-established measures of ...
kiemohr.bsky.social@kiemohr.bsky.social

Delighted to share this paper, now published in @arvoinfo.bsky.social. With @spk3lly.bsky.social‬ and Anna Geuzebroek we explored differences in visual cortical responses for pattern pulses and pattern reversals. Here's the link: jov.arvojournals.org/article.aspx...

Elisabeth Parés Pujolràs reposted

Elisabeth Parés Pujolràs reposted

CONVERSATIONS ON CONSCIOUSNESS How the CCN Community Can Contribute #CCN2025 on Wednesday 13th (10am 🇪🇺) Livestream link below ➡️ Exploring the possibility of Computational Consciousness Science ➡️ Discussing three Templeton World Charity Foundation Adversarial Collaborations

Elisabeth Parés Pujolràs reposted

Elisabeth Parés Pujolràs reposted

The way that the CPP behaves in this task seems very different from its well-established characteristics in static contexts with continuous evidence (e.g. classic random dot motion tasks), where it reflects the absolute cumulative sum of evidence.

We believe our results hint at a flexible neural architecture that can adapt to different decision contexts, and highlight how the CPP may play a key intermediary role in adaptive evidence accumulation. Stay tuned for upcoming related work, and thanks for reading!

Further, we found that the centroparietal positivity (CPP) tracks effective evidence (i.e. it reflects how much a DV changes following a token, rather than just stimulus information), but it does not keep a sustained representation of the DV.

We used an expanded judgement task where participants watched a slow sequence of discrete tokens (400 ms per token, up to 10 tokens) which favoured one of two possible choices. The task was volatile, which means that the choice supported by the tokens may change within a trial.

Because the normative solution to this task does not prescribe perfect accumulation, we could dissociate objective evidence presented on the screen from effective evidence (i.e. the belief updates that each token triggered), and map these quantities onto neural signals.

Elisabeth Parés Pujolràs reposted

Elisabeth Parés Pujolràs reposted

1/7 Our paper on individual variability in decision-making is finally out in @nature.com! Inspired by the classic work by Mante and Sussillo, we trained many rats to solve context-dependent decision-making, and we found that different brains use different neural mechanisms to solve the same task!

nature.comIndividual variability of neural computations underlying flexible decisions - NatureBehavioural experiments to study decision-making in response to context-dependent accumulation of evidence provide testable models that are consistent with the heterogeneity in neural signatures among...