Zeb L. Kurth-Nelson

Functional reorganization of motor cortex connectivity during learning

This study shows how rapid learning in mice involves structured changes in motor cortex connectivity, rerouting preparatory activity to control a conditioned neuron.

Reasoning with programs in replay

The paper investigates how the brain executes programs by analyzing neural activity during mental program execution, finding that fast neural sequences represent candidate solutions.

Neural mechanisms of credit assignment for delayed outcomes during contingent learning

The paper investigates neural mechanisms of credit assignment for delayed outcomes using fMRI and an adaptive learning task, identifying roles for lOFC, HC, and lFPC in tracking causal choices over time.

Neural representations supporting reuse of abstract structural knowledge in humans

The paper investigates how neural representations support the reuse of abstract structural knowledge in humans during sequence learning tasks.

Distributional reinforcement learning in prefrontal cortex

This study shows distributional reinforcement learning better explains prefrontal cortex responses compared to classic reinforcement learning.

Cognitive Model Discovery via Disentangled RNNs

This paper learns cognitive models from behavioral data using disentangled RNNs, recovering interpretable models from synthetic and rat behavioral data.

Generative replay for compositional visual understanding in the prefrontal-hippocampal circuit

The paper shows that visual understanding involves relational and compositional representations in the prefrontal-hippocampal circuit, with replay sequences forming hypotheses about scene configurations.

Replay bursts coincide with activation of the default mode and parietal alpha network

The paper shows replay bursts occur during activation of the default mode and parietal alpha networks, coordinating information flow between cortical areas.

Measuring Sequences of Representations with Temporally Delayed Linear Modelling

This paper introduces TDLM, a toolkit for analyzing neural activity sequences, showing its ability to detect human replay during rest using MEG data and parallels to rodent hippocampal replay.

The value of what’s to come: neural mechanisms coupling prediction error and reward anticipation

The paper investigates how neural systems compute anticipatory value from future rewards, showing three brain regions coordinate this process.

What is a cognitive map? Organising knowledge for flexible behaviour

The paper reviews evidence and theories suggesting cognitive maps organize knowledge across domains, supporting flexible behavior and abstract learning.

Episodic Control as Meta-Reinforcement Learning

The paper extends a unified account of reinforcement learning to include episodic learning by introducing a meta-learner that coordinates different learning algorithms.

Prefrontal cortex as a meta-reinforcement learning system

The paper proposes a new theory of reward-based learning where the prefrontal cortex functions as a meta-reinforcement learning system trained by the dopamine system.