Andre Marquand
@amarquand.bsky.social
Interested in data analysis, neuroscience, machine learning
A new study from Anthropic finds that gains in coding efficiency when relying on AI assistance did did not meet statistical significance; AI use noticeably degraded programmers’ understanding of what they were doing. Incredible.
Pavel@spavel.bsky.social · 6mo agoThe latest from Anthropic: using Anthropic's products makes you worse at your job
I am very excited about this new preprint! @likeajumprope.bsky.social presents an approach for charting growth velocity using normative models. It provides a fundamental framework for detecting individualised deviations from an assumed developmental trajectory. Enjoy! arxiv.org/abs/2601.07591
Charting the velocity of brain growth and development
Brain charts have emerged as a highly useful approach for understanding brain development and aging on the basis of brain imaging and have shown substantial utility in describing typical and atypical ...
arxiv.org
Also out today in PNAS: Anwesha Das shows in this thorough paper that "micro-offline gains" in sequence learning do not reflect consolidation of a motor memory - they also happen in random sequences. With @eazanon.bsky.social and Max-Philip Stenner (Magdeburg). www.pnas.org/doi/10.1073/...
... and a lifespan normative model of fractional anisotropy from Ramona Cirstian. An important contribution! linkinghub.elsevier.com/retrieve/pii...
Redirecting
linkinghub.elsevier.com
A bit belated, but here are some new papers! first, a conceptual framework for modelling extremes in neuroimaging data onlinelibrary.wiley.com/doi/full/10.... Great work from the inimitable Charlotte Fraza!
Using Extreme Value Statistics to Reconceptualize Psychopathology as Extreme Deviations From a Normative Reference Model
Overview of our approach for extreme value statistics. First, we fit a normative model to imaging phenotypes, before employing a ‘peaks-over-threshold’ approach widely used in meteorology and finance...
onlinelibrary.wiley.com
Come work with us! we are hiring a scientific software developer! www.radboudumc.nl/en/vacancies...
Vacancy - Research Software Engineer for Real-World Data
Join the forefront of healthcare innovation at Radboudumc as a research software engineer! We are seeking a passionate and driven individual to ...
eur01.safelinks.protection.outlook.com
Come work with us! we are recruiting for a talented PhD candidate for a new project www.radboudumc.nl/en/vacancies...
Vacancy - PhD candidate 'Nightly dance – Dynamics of mother, father, and baby night interactions and sleep'
We are looking for a highly motivated PhD candidate within the project Nightly dance – Dynamics of mother, father, and baby night interactions and ...
radboudumc.nl
[🆕RESEARCH] Increasing vegetation cover by 10%, 20%, and 30% could lower the global population-weighted mean temperature during the #warm season. These reductions could prevent approximately 0.86, 1.02, and 1.16 million deaths—corresponding to 27.16%, 32.22%, and 36.66% of all heat-related deaths.
... and another cool framework from Loran Knol! www.medrxiv.org/content/10.1...
Unobtrusive inference of circadian rhythms from smartphone data
Circadian rhythms are an integral feature of human cognition and psychopathology but are difficult to measure outside laboratory conditions and over long timescales. Smartphones and other wearables ar...
medrxiv.org
Very pleased to see this paper out! An analytical frameowrk for digital phenotyping data based on hidden Markov Models. Great stuff from Imogen Leaning! jmir.org/2025/1/e64007
Uncovering Social States in Healthy and Clinical Populations Using Digital Phenotyping and Hidden Markov Models: Observational Study
Background: Brain-related disorders are characterized by observable behavioral symptoms, for example, social withdrawal. Smartphones can passively collect behavioral data reflecting digital activities...
jmir.org
2/ Our first speaker was Prof. André Marquand (@amarquand.bsky.social), who studies how environmental challenges affect mental health. His talk reinforced that neuroscientists can also play a key role in researching impacts and effective ways to promote climate action engagement and policy-making! 🧠
In irony too funny for words, Anthropic asks that you please don’t use AI in your job application. No, really! open.substack.com/pub/garymarc...
Irony too funny for words
Oops
open.substack.com
A colleague today asked about Gaussian Processes on a sphere. Like spatial autocorrelation on the globe. I asked @avehtari.bsky.social. There are papers! This problem common enough to be worth having a case study I think? (1) Theory doi.org/10.1007/s112... (2) Application doi.org/10.48550/arX...
Hilbert space methods for reduced-rank Gaussian process regression - Statistics and Computing
This paper proposes a novel scheme for reduced-rank Gaussian process regression. The method is based on an approximate series expansion of the covariance function in terms of an eigenfunction expansio...
doi.org
My impression of this action is that it is quite deliberately intended to sow chaos, doubt, and destruction. wapo.st/4gfgcHQ
White House pauses all federal grants, sparking confusion
The Trump administration has put a hold on all federal financial grants and loans, affecting tens of billions of dollars in payments.
wapo.st
Week 3 homework for my stats course. "Causal salad": adding measurements to a regression until something is significant and interpreting every coefficient as a total causal effect. This is rarely sensible. (All course materials here github.com/rmcelreath/s...)
New paper in Imaging Neuroscience by Hossein Rafipoor, Saad Jbabdi, et al: Hierarchical modelling of crossing fibres in the white matter direct.mit.edu/imag/article...
“LLMs amplify existing security risks and introduce new ones” Even Microsoft sees it. New paper at arXiv, discussed briefly below. open.substack.com/pub/garymarc...
“LLMs amplify existing security risks and introduce new ones”
also: “the work of securing AI systems will never be complete”
open.substack.com
🎉 eLife is pleased to announce Timothy Behrens (@behrenstimb.bsky.social) as our new Editor-in-Chief! A distinguished neuroscientist and long-time supporter, Tim will lead our efforts in transforming research communication for all.
eLife appoints Timothy Behrens as Editor-in-Chief
Behrens will lead eLife’s efforts to improve research communication for all.
buff.ly
lol, a clip of a Tesla's vision system trying to understand a train labyrinth.zone/notice/Aq5V7...
“𝗧𝗵𝗿𝗲𝗲 𝗺𝗮𝘁𝗵𝗲𝗺𝗮𝘁𝗶𝗰𝗶𝗮𝗻𝘀 𝗮𝘁 𝘁𝗵𝗲 𝗯𝗲𝗮𝗰𝗵, 𝘄𝗶𝘁𝗵 𝘁𝗵𝗲𝗶𝗿 𝗵𝗲𝗮𝗱𝘀 𝗶𝗻 𝘁𝗵𝗲𝗶𝗿 𝘀𝗮𝗻𝗱”
Should also mention: associated OSF resource osf.io/d9r3x/ with links to code & demo videos. Again, huge tks to contributors incl. @oesteban.bsky.social @amarquand.bsky.social @engra.me @clarekelly.bsky.social @jivr.bsky.social @christiangaser.bsky.social
Methods for analyzing large neuroimaging datasets
A fully open-access book with a truly practical guide to analyzing large neuroimaging datasets. Get the book: https://link.springer.com/book/10.1007/978-1-0716-4260-3 Hosted on the Open Science...
osf.io
The man in charge of social media platforms used by billions should show some moral courage, writes @jemima.bsky.social www.ft.com/content/682c...
New paper! This is a very elegant way to assess longitudinal change against normative models estimated cross-sectionally! Great stuff from Barbora Rehák Bučková! elifesciences.org/reviewed-pre...
Using normative models pre-trained on cross-sectional data to evaluate intra-individual longitudinal changes in neuroimaging data
elifesciences.org
Calibration “resolves” epistemic uncertainty by giving predictions that are indistinguishable from the true probabilities. Why is this still unsatisfying? statmodeling.stat.columbia.edu/2024/12/31/c...
Calibration “resolves” epistemic uncertainty by giving predictions that are indistinguishable from the true probabilities. Why is this still unsatisfying? | Statistical Modeling, Causal Inference, an...
statmodeling.stat.columbia.edu
o3, #AGI, the art of the demo, and what you can expect in 2025 "I saw zero evidence that o3 could work reliably in open-ended domains... The most important question wasn’t really addressed." @garymarcus.bsky.social garymarcus.substack.... #generativeAI
🛠️ November's most-read Tools and Resources paper presents a new method to identify dynamic causal interactions in complex systems: https://buff.ly/3Zldxpt Learn more about publishing your tools or resources with us: https://buff.ly/4f2O0Ht
