Carolina Feher da Silva

A Computational Toolkit for Designing and Analysing Repeated Binary Choice Experiments

This paper introduces a computational framework for analyzing repeated binary choice experiments, revealing hidden decision patterns through advanced time series analysis methods.

A new take on model-based and model-free influences on mental effort and striatal prediction errors

This study challenges the assumption that model-free learning is automatic by showing that model-based strategies can reduce mental effort without increasing it.

Humans are primarily model-based learners in the two-stage task

This study shows that detailed task instructions lead humans to primarily use model-based learning in two-stage tasks, with minimal model-free influence, and highlights misconceptions that can create false appearances of mixed learning.

Can model-free reinforcement learning operate over information stored in working-memory?

This study investigates whether model-free reinforcement learning can operate on information stored in working memory using human participants in a reward learning task with sequential and simultaneous visual stimuli.

A note on the analysis of two-stage task results: how changes in task structure affect what model-free and model-based strategies predict about the effects of reward and transition on the stay probability

The paper demonstrates how small changes in two-stage task design can alter model-free and model-based agent behavior, emphasizing the need for simulations in experimental design.

Exploration and recency as the main proximate causes of probability matching: a reinforcement learning analysis

The paper uses reinforcement learning to analyze why humans probability match instead of maximize in a learning task, finding that pattern search, exploration, and recency significantly influence behavior.