Reward
Neuroanatomy of the reward system
Rewards and phasic dopamine
Reward prediction error (RPE) and uncertainty of outcome
Phasic dopamine and addiction
Phasic dopamine in learning
Addiction changes how the world looks like
rostral stratal region as a hub for vmPFC, OFC, dACC. to connect with dorsal. and lateral PFC. regions integrating motivational, reward. and cognitive control information
mesolimbic pathway & nigrostriatal pathway
OFC, mPFC
dorsal tier
coordinated activation of terminals in striatal areas would produce a unique combinatorial activation
value computations
encoding of reward / saliency prediction errors
group of cells in the midbrain that extend dorsolaterally and merge with the ventral segmental area (VTA), dendrites stretch to dorsal SNc (but not ventral parts)
ventral tier
dendrites extending into SNr
DA synapses on the dendrites of medium spiny neurons (MSN) to modulate incoming cortical information or exerting generalised effect in the striatum
midbrain - striatum projection
differential ratio of DA projections
inverse dorsal-ventral topographic organisation
dorsal DA cells project to ventral parts of striatum and vice versa
dorsal tier projects to ventral striatum (VS)
ventral tier projects centrally with most ventral part projecting primarily to dorsolateral (sensorimotor) striatum
input to ventral striatum is derived from most limited midbrain area
input to dorsolateral striatum derived from largest midbrain area
striatum-midbrain projection
inverse dorsal-ventral topography
but: ventral striatum projects to a large area in the midbrain
but dorsolateral striatum projects to a limited area in the midbrain
organisation of afferent and efferent projections (striate-nigro-striatal connections) in an ascending spiral interconnecting different functional regions of the striatum
click to edit
information from the limbic system reaching motor system through a series of connections - fig 6 (!)
mechanism by which motivation and cognition can influence motor decision-making and responses to environmental cues
the striato-nigro-striatal network
reciprocal connections linking up regions with similar functions
non-reciprocal connections linking up regions associated with different cortical basal ganglia circuits
feedback + feedforward flow of information
parallel processing? rather combination of motivational, cognitive, motor areas to produce goal-directed behaviour
reward - function is defined by action / influence on behaviour
search for reward neurons - midbrain-dopamine mediated signals that signify true reward value of objects
reward-prediction error
difference between obtained and predicted reward
Rescorla-Wagner learning rule
phasic activity in midbrain dopaminergic neurons
coding prediction error
animal learning theory
burst activity
fully predicted reward - no response
negative prediction error - omission of a predicted reward - depression
positive-prediction error - unpredicted reward - activation
tonic dopamine
sustained extracellular dopamine concentration
tonically stimulating D2 receptors
further regulated by processes such as dopamine reuptake, end-product control of synthesis and release and presynaptic influences from other neurotransmitters
uncertainty
reward prediction error hypothesis
uncertainty influences reward value
evokes low dopamine concentrations appropriate for stimulating high-affinity D2 receptors
conditioned-inhibition paradigm
blocking paradigm
following conditioned reward-predicting stimuli
responses vary with expected values of. the probability distributions of reward values
activation of dopaminergic neurons might reflect sensitivity to rewarding and physically salient stimuli but not punishers, negative surprises, conditioned inhibitors
mesolimbic pathway
ventral tegmental area
nucleus accumbens
phasic dopamine bursts dopamine release influences long-term potentiation in the striatum
three factor learning rule
medium spiny cells receive input from cortical neurons (glutaminergic) and from midbrain neurons (dopaminergic)
simultaneous input strengthens synaptic transmission - cortical neurons coding information about the environment, midbrain (dopaminergic) input singling reward prediction error - adjustments are necessary
strong pre- and postsynaptic activation and dopamine release induce LTP
dopaminergic midbrain neurons and targets responds at two timepoints
in response to the reward predicting stimulus
in response to the receipt or omission of the reward
proportional to the subjective value of the predicted reward
difference between expected and received reward
reinforcers
secondary
primary
partially conditioned
generalised conditioned
dopaminergic activation across a range of stimuli that act as different types of reinforcers
click to edit
VS
money
highest VS + NA