Étienne Koechlin

The paper investigates how task-set learning relies on unsupervised temporal chunking of stimulus-response associations using behavioral and neural data.

The paper shows that imprecise neural computations following the Weber Law enable near-optimal adaptive behavior in volatile environments, outperforming higher-order models and reinforcement learning.

The paper introduces a model of causal inference that explains how people perceive their causal power through counterfactual reasoning, showing it better explains behavior than alternative models.

The paper argues that model simulation is essential to support claims about behavioral function in computational cognitive neuroscience and proposes guidelines combining model comparison and simulation.