Jr. Hollerman et W. Schultz, DOPAMINE NEURONS REPORT AN ERROR IN THE TEMPORAL PREDICTION OF REWARDDURING LEARNING, NATURE NEUROSCIENCE, 1(4), 1998, pp. 304-309
Many behaviors are affected by rewards, undergoing long-term changes w
hen rewards are different than predicted but remaining unchanged when
rewards occur exactly as predicted. The discrepancy between reward occ
urrence and reward prediction is termed an 'error in reward prediction
'. Dopamine neurons in the substantia nigra and the ventral tegmental
area are believed to be involved in reward-dependent behaviors. Consis
tent with this role, they are activated by rewards, and because they a
re activated more strongly by unpredicted than by predicted rewards th
ey may play a role in learning. The present study investigated whether
monkey dopamine neurons code an error in reward prediction during the
course of learning. Dopamine neuron responses reflected the changes i
n reward prediction during individual learning episodes; dopamine neur
ons were activated by rewards during early trials, when errors were fr
equent and rewards unpredictable, but activation was progressively red
uced as performance was consolidated and rewards became more predictab
le. These neurons were also activated when rewards occurred at unpredi
cted times and were depressed when rewards were omitted at the predict
ed times. Thus, dopamine neurons code errors in the prediction of both
the occurrence and the time of rewards. In this respect, their respon
ses resemble the teaching signals that have been employed in particula
rly efficient computational learning models.