Rik Henson

Excited to share a new paper spearheaded by the wonderful @baror-shira.bsky.social:
tinyurl.com/bd8xdcum
@erc.europa.eu @nathumbehav.nature.com

We test the link between serial dependence (as an index of continuity) and event boundaries (indexing segmentation). A few key findings in the thread:

Gradually, then suddenly: the precarious position of UK preclinical neuroscience Mark Walton argues that while the UK has a long and distinguished record of supporting brain research, this now risks being eroded by funding insufficienci

Gradually, then suddenly: the precarious position of UK preclinical #neuroscience

Professor Mark Walton’s new opinion piece in @brain1878.bsky.social highlights the funding and workforce pressures facing UK preclinical neuroscience.

academic.oup.com/brain/advanc...

#PreclinicalResearch

Over half of Britons say it would be a bad outcome if the assisted dying Bill runs out of time in the Lords.

Over a quarter of opponents agree.

The public don’t want the Bill blocked by delay.
Add your name to the petition urging the Government to intervene:
www.dignityindying.org.uk/petition

A bit old now, but great Registered Report by Maya Raza (not on here)

In a vague attempt to lighten mood (and possibly even with educational benefit), I have added a couple more cartoons about the tools of cognitive neuroscience (with brief explanations): www.mrc-cbu.cam.ac.uk/people/rik.h.... Thanks to @iancooketapia.bsky.social for prettifying.

Really important points…

Open Science, Psychology, and the Art of Not Quite Claiming Causality with Julia Rohrer - Decoding the Gurus In a rare departure from our usual diet of online weirdos, this episode features an academic who is very much not a guru. We’re joined by Julia Rohrer...

Do you sometimes think "oh boy, I would really like to hear Julia ramble some more about the topics about which she doesn't stop talking to begin with?"

The wait is finally over! @guruspod.bsky.social and had a chat about open science, causal inference, and apparently birth order effects.

Important reminder that the relationship between functional connectivity and a trait is likely to be state-dependent, and hence it may be unwise to rely on resting state only

An image reading the statement "The Future is Diamond" alongside the logo of the Open Journals Collective.

We've officially launched! With our new #DiamondOpenAccess investment campaign to help libraries build a sustainable, community‑led future for scholarly publishing & support journals flipping away from costly subscription models. Read the Press release here drive.google.com/file/d/10Hgf...

Great work by Roni Tibon (not on BlueSky) - surprising that negligible difference in fMRI correlates of semantic vs episodic retrieval?

Want to help make replications easier to find, share, and credit? Join us in Germany for a hackathon to do just that. Applications now open ⬇️

Trump's jibes are wearing thin for many of Europe's leaders The ball is in the court of European leaders ahead of Thursday's emergency EU meeting in Brussels, writes Nick Beake.

www.bbc.com/news/article...

Brilliant!

4.1 Topics covered: Experimental vs Field Designs, Multiple Regression, Correlated Predictors, Range effects, Measurement noise, Sequential Orthogonalisation (Type II vs Type III sum of squares)

4. (don't worry; stopping soon...) Statistical Power for Interactions: github.com/RikHenson/Po.... Is it true that interactions are always harder to detect than main effects? Not always... available as Rmd and HTML (no Python or Matlab yet)

3.1 Topics: Effect of SOA, Power Spectra, Filtering, HRF Models, Trial Order, Null Events, Correlated Predictors, Working Memory trials, State/Item effects, Single-trial estimation (LSA), Regularised LSA (Beta-Series Regression), LSS, Trial:Scan Variance, Trial:Scan Spatial Covariance (MVPA).

3. Efficient fMRI experiments: github.com/RikHenson/fM.... How to design an fMRI experiment, as a function of trial/HRF modelling, trial interval and trial order, that is efficient (sensitive) to a specific contrast (hypothesis), available as Matlab Livescript (or plain HTML; no Python or Rmd yet)

2.1 Topics: Multiple Regression versus Path Models, Mediation, Moderation, Structural Equation Models, Measurement Invariance, Longitudinal designs, Mixed Effects models, Practice effects, Attrition, Cross-Lagged Panel Models, Latent Curve Models, Parallel Process Models.

2. Relating Age, Brain and Cognition: github.com/RikHenson/Ag.... How to model cross-sectional and longitudinal relationships between 3 or more variables (ABC), available as Rmd (or plain HTML; no Python or Matlab yet).

Trains are good for reminiscing. I forgot to congratulate @adamjcurtis.bsky.social for passing his PhD viva with no corrections. Only downside is that as his external, I felt both old and odd, examining an “academic grandchild” (via @aidanhorner.bsky.social)…

Oops - correction - just realised that speed varies with location (and # users, obvs!) - sorry Eurostar!

WiFi speed on Eurostar: 30 Mbps download and 5 upload…
WiFi speed on DB ICE train: 100 Mbps download and upload…

I like Stewart Lee’s story about being unable to buy a ticket for an near-empty bus he had boarded because you could only buy tickets online, but the bus was late, so according to the company’s “AI” system, the bus was not in fact there, but several miles ahead, so buying a ticket was not possible…

I agree with this sentiment for empirical papers, which generally benefit from stronger evidence, eg more replications across multiple experiments within paper. But I’m less sure new theoretical ideas need to be “slowed down”? (unless half-baked just to get a publication)

"Artificial-intelligence systems are feeding on Wikipedia without giving back, and academic indifference is threatening the survival of what is arguably the most widely used reference work on the planet." www.nature.com/articles/d41... Check out the WikiJournal of Science, created to address this.

1.1 Topics: Central Limit Theorem, Randomisation, NHST (error rate/power), Bayes Factors, Smoothing, Height/Cluster/Mass Family-wise Error, False Discovery Rate, General Linear Model, T/F-tests, ANOVAs, Nonsphericity, Regression, fMRI timeseries (HRF, filtering, autocorrelation), Mixed Effects

1. Statistics for Neuroimaging: github.com/RikHenson/St.... A simulation-based introduction to statistical concepts used in univariate fMRI and M/EEG analyses, available as Jupyter or Matlab Livescript (or plain HTML; no Rmd yet).

Feel free to play with them, and/or re-post their existence. Comments/corrections also welcomed, as would any conversions from one type of notebook (eg Matlab) to another (eg Rmd)!

Over the years, I have written a few Jupyter/Rmd/Matlab notebooks that attempt to teach some statistical concepts, particularly in neuroimaging. You can find them here: www.mrc-cbu.cam.ac.uk/people/rik.h..., though I will say a bit more about each one in a number of posts over next few days.

A journal that supports a society, and run by brilliant colleagues like Kate!