Temporal difference learning
Change in environment over time causes change in expectations, which causes change in behaviour.
TDL model vs. R-W model
Dopamine codes prediction errors
Neurons that release dopamine appear to mimic the error function from temporal difference learning.
Reward coding
Dopamine neural response is proportional to R(pS)
Delay coding
The greater the delay from cue onset, the lower the intensity of dopamine neurons. When reward is provided, dopamine neurons fire.