And here’s ours -- fully reproducible code and materials:

🔗 github.com/alexisbreen/...

9/9

More great EWA modelling repos worth a shout-out:

🐒 @bjjbarrett.bsky.social et al. (capuchins):
🔗 github.com/bjbarrett/pa...

🐍 @mchimento.bsky.social et al. (Python-based theory):
🔗 github.com/michaelchime...

🧵8/n

Not a co-author, but a major influence -- big thanks to @dominikdeffner.bsky.social et al., whose prior modelling work and shared GitHub repo were foundational to this study:

🔗 github.com/DominikDeffn...

🧵7/n

Huge thanks to all co-authors (T. Eckersley, A. Camacho-Alpízar, C. Lambert, G. Balasubramanian, S. Healy, @rmcelreath.bsky.social, @laurenguillette.bsky.social), with extra thanks to @laurenguillette.bsky.social and @rmcelreath.bsky.social sky.social for encouraging this passion project 🙏

6/n

Overall, our study shows how computational modelling can link observed behaviour to hidden cognitive mechanisms, and probe their generalisability -- we hope this encourages wider use of these approaches in animal cognition research! 🌐📝⚙️🧠

🧵5/n

Using cognitive model estimates, we ran simulations where material rewards varied. Payoff structure -- not (dis)satisfaction -- drove social material choice. But bigger rewards didn’t mean more copying. So, preliminary insights into how material-use strategies may generalise e.g., in tool use.

🧵4/n

We link males' material choices to two asocial and two social latent learning parameters -- their decisions are dynamic! -- to provide the first formal mechanistic evidence for the cognitive basis of birds’ nest building ⚙️🧠

🧵3/n

Zebbie nest-builders seem to *copy-if-dissatisfied* with prior material-experience, particularly at 1st choice -- this suggests they recognise the utility of social information.

🧵2/n