Dopamine signals in the ventral, dorsomedial, and dorsolateral striatum are modulated by various variables, such as stimulus-associated value, choice, confidence, but these modulations can be inclusively explained by TD errors.
Midbrain dopamine neurons in rats signal discrepancies between predicted and actual rewards, regardless of whether the rewards are predicted on the basis of experience or inference.
Human learning relies on short-term memories (eligibility traces) which provide a mechanism to reinforce sequences of actions from a single reward (one-shot).
Perception in autism is sensitive to absolute rather than to relative metrics of the environment, encoding changes in the environment without calibrating the changes relative to reference stimulation.
The brain has a central cortico-striatal learning circuit that suppresses ongoing pain after injury when actively learning about things that could remove the cause of the pain.
Self-reactive B cells downregulate the IgM but not the IgD B cell receptor, and this serves as a critical tolerance mechanism because IgD is less sensitive to bona fide endogenous antigens than IgM.
Electrophysiological recordings in monkeys reveal that cerebellar complex spikes encode future reward size when reward information is first made available, but not during reward delivery or smooth pursuit eye movement.
Confidence-dependent reinforcement learning is active and produces trial-to-trial choice updating even in well-learned perceptual decisions without explicit reward biases, across species and sensory modalities.
Functional hypoconnectivity between ‘social brain’ default mode circuitry and visual association cortex underpins a subtype of autistic toddlers with a strong preference to attend to the non-social visual world.
Children with autism often 'tune out' the voices in their environment and new results show that impaired processing of voices in the brain's reward system may underlie this social behavior.