A fast spiking interneuron sub-type in medial and lateral prefrontal cortex fires and gamma-synchronizes prominently during adaptive learning of reward values when outcomes are uncertain and choice options have similar values.
Pronounced cerebellar activation during unexpected omission of a potentially harmful event suggests that the cerebellum has to be added to the neural network processing prediction errors underlying emotional associative learning.
Dopamine neurons signal value prediction errors (VPEs) integrating information about both reward and aversion, in low reward contexts, whereas VPEs in some dopamine neurons are distorted in high reward contexts.
Veterans with PTSD show increased attention to a history of unexpected outcomes during loss learning, both as measured by computational model-derived behavioral parameters and in increased neural signaling in amygdala and insula.
Changes to sensory predictions are encoded by beta oscillations, surprise due to prediction violations by gamma oscillations, and alpha oscillations may have a role in controlling the precision of predictions.
fMRI evidence for off-task replay predicts subsequent replanning behavior in humans, suggesting that learning from simulated experience during replay helps update past policies in reinforcement learning.