News

New frameworks for biologically constrained learning: from similarities in humans and transformers and sparse, contrained RNNs to a unified theory of hippocampal replay

Unifying computational principles underly flexible cognition—from the orbitofrontal cortex to compositional task representations, working-memory–habit mixtures, and recurrent network subspaces.

From connectome-constrained dynamics and projection-specific population codes to human–ANN continual learning, and advances in hippocampal replay theory.

Neural studies reveal dynamic neural manifolds in arousal, decision-making, spatial coding, cortical development, and object-centered navigation.

A tour through the brain’s code: interneuron teamwork, speed-boosted navigation, dopamine-shaped learning, compositional maps, and value-sensitive choices

Neural Flexibility in Action: How the Brain Adapts Memory, Navigation, and Decisions to Changing Contexts

How Tiny RNNs, Inter-Brain Synchrony, and a Foundation Model Are Transforming Our Understanding of Cognition