🎉 Bonus thread addition: Lots of open questions. How do children learn to predict temporally distal states? To what extent can children leverage these predictions in more complex environments? How do early experiences in predictable environments modulate children's reliance on the SR? Stay tuned!

What about when mental shortcuts, like more temporally abstracted, predictive representations can be used, circumventing the need for costly simulation? Here we show that, in these cases, children use knowledge of their environments to guide their decisions to a comparable degree as adults! 🧒🧑 (4/4)

We suggest that this dissociation may emerge because leveraging structured knowledge often requires mentally simulating sequences of potential actions and their potential outcomes, which requires both time and computational resources that may be particularly costly for children. 🧠⌛(3/4)

❓ There's a puzzling dissociation in the developmental literature: In many contexts, children seem to effectively learn the structure of their environments, but fail to use this structured knowledge to guide their choices. ❓ (2/4)

Our findings suggest that differences in the dynamics of learning can account for age-related changes in structured knowledge acquisition.

Lots more details (and exciting open questions) in the paper!