How do we discover and evaluate
what is worth doing?
Learning, Decisions, and Latent States
Goal-dependent behavior requires not only efficient execution of thoughts and actions, but also the ability to discover, organize, and evaluate what is worth doing.
We study how the brain navigates open-ended decision spaces, from constructing latent representations of value and assigning credit to outcomes, to generating and evaluating options worth pursuing.
Related publications
Matthews, J., Kikumoto, A., Miyamoto, K., & Shibata, K. (2026). Metacognition is mentally demanding. PsyArXiv. Manuscript submitted for publication. [LINK]
Lloyd, B., Wurm, F., Kikumoto, A., & Lluis-Vives, M. (2025). Nailed-it! The unique value of perfect predictions in reinforcement learning. RLDM, Dublin, Ireland.
Kikumoto, A., & Mayr, U. (2019). Balancing model-based and memory-free action selection under competitive pressure. eLife, 8:e48810.
Kikumoto, A., Hubbard, J., & Mayr, U. (2015). Dynamics of task-set carry-over: Evidence from eye-movement analyses. Psychonomic Bulletin & Review.