Thomas E. Akam

This paper proposes a mechanistic theory of planning in prefrontal cortex using spacetime attractors, showing that recurrent neural networks trained on planning tasks recapitulate this structure.

This study shows mice shift from model-free to structure-informed reinforcement learning strategies during early training in a dynamic decision-making task.

This paper investigates temporal predictive coding (tPC) as a framework for learning temporal dependencies in recurrent neural networks, showing its equivalence to BPTT variants and its ability to handle complex tasks with local plasticity.

This study identifies two distinct neural representations in mouse medial frontal cortex during flexible navigation, showing how they encode position and goal distance to guide behavior.

The paper develops a behavioural assay to study flexible navigation in mice, showing they learn maze structures quickly and navigate efficiently.

This paper proposes that dopamine ramps result from a dual-process learning system and demonstrates that the model reproduces key dopamine ramp phenomena.

The paper identifies a neuronal mechanism in mice that maps abstract behavioural structure, enabling zero-shot learning of new tasks through goal-progress cells in the medial frontal cortex.

This study shows that ACC neurons encode chosen actions and their outcomes rather than cognitive states like attention or impulsivity in mice performing a task.

This paper introduces a theory of hippocampal replay that explains how replay builds a cognitive map for navigation, reconciling conflicting hypotheses about its function and predicting new experimental outcomes.

This study identifies claustrum neurons and their connectivity patterns, showing they integrate inputs from multiple cortical regions and contribute to multimodal sensory processing.

pyControl provides open-source hardware and software for designing and running complex behavioural experiments with improved reproducibility and scalability.

The paper investigates how neural variability in the somatosensory cortex affects the perception and propagation of activity between brain regions in mice.

This study shows that the retrosplenial cortex is necessary for both spatial and non-spatial latent learning in mice, specifically during the preconditioning phase.

This study investigates how dopamine clearance mechanisms, DAT and COMT, affect behavioral flexibility in mice using pharmacological modulation and behavioral tasks.

The paper shows that dopamine-independent state inference and dopamine-mediated reinforcement learning interact to guide reward-based decisions in mice.

The paper decodes working memory-related neural connectivity in mice and humans across tasks, finding task-specific and distributed patterns with high prediction accuracy.

The paper shows that prefrontal cortex abstracts problem structure while hippocampus maps it to specific situations in mice during reversal learning.

This study shows that the anterior cingulate cortex (ACC) represents action-state transitions and mediates model-based reinforcement learning in mice by influencing subsequent choices based on these transitions.

pyPhotometry provides open-source hardware and software for fiber photometry data acquisition, offering better signal-to-noise ratio and lower cost than commercial alternatives.

This study uses closed-loop optogenetic modulation to show that gamma frequency and amplitude in hippocampal pyramidal cells can be independently altered by adjusting feedback timing.