Nicolas Leenaerts

When you submit a bunch of jobs to the compute cluster...

The rise of reinforcement learning algorithms in psychology and psychiatry must integrate open science practices!

I am therefore proud to present the preregistration of our study using off-policy reinforcement learning to improve JITAIs for weight loss: osf.io/6wpu4/overview

Picture from the MAPCI training institute I was lucky to attend. A great initiative from the @d3center.bsky.social. If you're interested in learning about adaptive interventions for cancer control, be sure to attend next year!

Spirituality and Harmful or Hazardous Alcohol and Other Drug Use This meta-analysis synthesizes findings about spirituality and alcohol and other drug use to estimate the overall effect size of the associated risk reduction.

Meta-analysis finds a relation between spirituality and a reduced risk of harmful alcohol and drug use.

However, is there a difference between religions banning alcohol and drugs and those that do not? Also, this does not prove that promoting spirituality reduces the risk of substance use.

Saw another paper reporting out-of-sample R² as the squared correlation between predictions and outcomes. That’s wrong!

Correlation measures association, not accuracy. It doesn’t tell you how close predictions are.

#machinelearning #ML

Yes indeed! The variables you want to match up need to be binary. Categorical variables need to be transformed into binary variables first (e.g., through dummy coding). We are currently working on papers on this methodology, but you can check an old preprint for more info: osf.io/preprints/ps...

Have you been interested in any of the set theory stuff I've been doing recently? Wondered at all how you can get in on the fun? Here's an R package @nicolasleenaerts.bsky.social and I built called 'setweaver.' This vignette should get you up and running. Yay :)
cran.r-project.org/web/packages...

Found where he lives!

Telemann: Complete Tafelmusik YouTube video by Brilliant Classics

If you want some music playing in the background while you work (something stimulating but not too much), I can’t recommend Telemann’s Tafelmusik highly enough. It was literally written to be background music!

www.youtube.com/watch?v=xNFk...

As somebody with a mole near their left tear duct, let me tell you this hit a little too close to home 🫠

And to hear this for the first time at 32 😵

Really interesting paper for everybody dealing with missing data.

Watch our symposium at #TIPS2025 where I,
@emmadeschuyteneer.bsky.social, and Chunzhuo Wang presented our work on using #AI to treat #eatingdisorders and #obesity.

www.youtube.com/watch?v=TAVS...

I'm having some disappointing initial results using LLM's/foundation models for intensive longitudinal data in patients with a psychiatric disorder. Hit me up if you are having more luck or want to commiserate!

Harvard in the snow...

Oh and I'm also in the picture.

Take it as a compliment! It means the algorithm thinks you've got 10K laying around to pay for the APC.

Analysis pipeline using off-policy offline RL to improve JITAIs for weight loss is done! Will release the preregistration by the end of the month! Be on the lookout!

WARN-D machine learning competition is live » Eiko Fried If you share one single thing of our team in 2026—on social media or per email with your colleagues—please let it be this machine learning competition. It was half a decade of work to get here, especi...

After 5 years of data collection, our WARN-D machine learning competition to forecast depression onset is now LIVE! We hope many of you will participate—we have incredibly rich data.

If you share a single thing of my lab this year, please make it this competition.

eiko-fried.com/warn-d-machi...

Snapshots from our symposium at #TIPS2025 where I, @emmadeschuyteneer.bsky.social, and Chunzhuo Wang presented our work on using #AI to treat #eatingdisorders and #obesity.

Want life tips? Well, you won’t get them from me.
But if you want to hear me talk about using AI to address eating behaviors, join me at #TIPS2025.

Here @mcleanhospital.bsky.social with Tori Votaw to give a presentation on using reinforcement learning to improve JITAIs for cannabis use!

We used LLM's to {insert topic here}.

Join us in Vienna next year for #SAA2026. Abstract submissions are now open: univie.eventsair.com/saa2026/subm...

‘It being Monday’ is showing up as the strongest predictor for a negative outcome in my dataset.

Who am I modeling? Garfield? 🐈

Got selected as a MAPCI 2026 fellow! Really excited to learn how to develop JITAIs for cancer survivors struggling with substance use!

Only 3 days left to apply! MAPCI is a free, comprehensive training program (Feb 24–27, 2026 • Ann Arbor) for cancer control researchers developing adaptive interventions (including JITAIs in mobile health). Significant travel & lodging support. Apply by Nov 14, 2025 → mapci.space

Our experimental study on the effects of negative mood and alcohol intoxication on reinforcement learning in the Iowa Gambling Task is now published OA in Judgment and Decision Making: doi.org/10.1017/jdm.... as always, anonymized data, materials, and code can be found on the OSF: osf.io/ky3aj/

Surely we can all agree that having Black Friday deals for GLP-1 agonists is just not OK.

I love that LLMs can help the new generation with coding, but it's making me feel like an old guy every time I say: 'Back in my day, we had to go to Stack Overflow and hope somebody already posted about this problem, and that someone else actually had a helpful response.'

Are you interested in how different AI methods can be used in the treatment of eating disorders and obesity? Come to the symposium of me, @emmadeschuyteneer.bsky.social, and Chunzhuo Wang at #TIPS2025!

Must We Always Go Idiographic? - Cognitive Therapy and Research Purpose While psychological change processes are increasingly assumed to be “non-ergodic”, prompting a shift toward idiographic approaches, the assumption of ergodicity is often accepted a priori rath...

🚀 New paper (open access) out in Cognitive Therapy & Research: we (@herzog.bsky.social, @evalottabrakemeier.bsky.social, @hudsongolino.bsky.social and me) ask whether group‑level symptom‑change networks in CBT actually capture what happens inside each patient. #Psychology #CBT