We're now a team of 12 people in 8 countries, and preparing papers on future thinking (a bit of anticipatory clickbait: the results are genuinely stunning), Flashbulb Memories, and traumatic memories. More data collections are to come. Watch this space - and help Ukrainian defenders.

In a seminar, I had the chance to report how I started the collaboration with Ukrainian colleagues and how we collected data there. Weirdly, that story started in the russian city of Perm in 2013 - and the project somehow feels like the culmination of much of my adult life.

Once again, my admiration for Ukraine grew even more. It is inspiring to see how our colleagues have adapted to the situation and simply keep on doing their work - and how motivated they are to even make use of the situation to change the country and achieve stronger alignment with Europe.

The less shouty version can be accessed here: www.sciencedirect.com/science/arti...

(It's also the perfect summer read, as it is long enough to keep you busy even on more prolonged train rides)

A huge thanks to my co-authors (including @nliedtke.bsky.social, @ssiestrup.bsky.social, @moritzwurm.bsky.social, @schubotzlab.bsky.social) for their substantial contributions to this work.

You can read it in NeuroImage:
doi.org/10.1016/j.ne...

4) Both prior precision (how strong the prediction is) and prior accuracy (how good the prediction is) uniquely contribute to these processes with different time profiles.

3) These effects are implemented by reinstatement of the original episode (used for prediction) during the new and mismatching input. The more reinstatement, the more distinct encoding.

1) Larger PEs promote detailed memories that are more neurally distinct from the memories that were used for the prediction.
2) Medium PEs lead to less distinct memories, negatively impacting recognition performance.

Looks really nice and seems very useful especially now that quite a few students think it's not worth bothering learning a programming language if ChatGPT can do it (and then end up not being able to trouble-shoot anything). Thanks for putting this together!

Congrats, looking forward to reading!

Marius @mariusboeltzig.bsky.social presented his poster on the influence of prediction error size on memory - showing that large prediction errors lead to distinct encoding of new episodes, while keeping old ones intact.

Das heißt sofort Sondervermögen mit dem alten Bundestag, schnelle Koalitionsverhandlungen und maximale Unterstützung der Ukraine?

Ah, we'll dig trenches around Berlin in two years because a guy who suggested drinking disinfectant thinks we're not thankful? Cool cool, let me just sharpen the discussion of whether my results are in congruence of the Latent Cause Theory.