Berna Devezer

devezer.bsky.social

09 July 2026 12:54

Keep hoping that metascience will pivot toward properly contextualizing studies within their design, assumptions, domain, literature... instead of treating every standardized, decontextualized data point as meaningful evidence wrt the state of science but we're getting farther away from that w AI.

Chris Baldassano

chrisbaldassano.bsky.social

17 June 2026 19:47

This work is now published in Nature Communications! We were able to train participants to have near-perfect memory for lists of 40 words - they added creative details to their memory representations that we could measure in their behavior and brains 🧠 doi.org/10.1038/s414...

Sam Verschooren

samversc.bsky.social

05 June 2026 16:27

🎉New paper out today in Nature Reviews Psychology🎉 with @mjdahl.bsky.social, @mariamaly.bsky.social, and @thiasmittner.bsky.social. We've been working on a unified framework for attentional states and the dynamics of transitions between them. 🧵

Berna Devezer

devezer.bsky.social

07 May 2026 20:48

There are a few more metascientific implications I'd like to highlight. 1. When reporting replicability rates, there are multiple levels of uncertainty that needs to be modeled in a hierarchical way, and metascientific practice flattens all to the lowest level, masking much of it in the process.

Chong Zhao

zhaochong.bsky.social

05 May 2026 20:20

In collaboration with @monicarosenb.bsky.social , we showed that individual diffs in LTM encoding is uniquely predicted by inter-electrode correlations even controlling for working memory abilities. This suggests that WM & LTM encoding are separate abilities coded by different neural signatures! 1/n