Recurrent neural network and its readout (cortex–striatum) can learn state representation and value using online random-weight feedback of temporal-difference reward-prediction-error (dopamine) through feedback alignment or biological non-negative-weight constraint-induced loose alignment.
Rab10 and Rab4 bidirectionally regulate AMPA receptor trafficking and spine structural plasticity during long-term potentiation by exhibiting opposing and temporally distinct activity patterns in stimulated dendritic spines.
Christopher M Kim, Carson C Chow, Bruno B Averbeck
Neural activity during reversal learning encodes decision-related evidence integrated across trials and shows substantial dynamics during each trial, suggesting an extension of the line attractor model that incorporates non-stationary dynamics.
To parse foreground and background sounds, the auditory cortex of humans and ferrets share a similar hierarchical organization, but the underlying computational mechanisms are fundamentally different.
